CurrentAnthropology

T H E W E N N E R - G R E N S Y M P O S I U M S E R I E S

ALTERNATIVE PATHWAYS TO COMPLEXITY: EVOLUTIONARY TRAJECTORIES IN THE MIDDLE PALEOLITHIC AND MIDDLE STONE AGE

V O L U M E 5 4

Current Anthropology is sponsored by �eWenner-Gren Foundation for AnthropologicalResearch, a foundation endowed for scientific,educational, and charitable purposes. �eFoundation, however, is not to be understood asendorsing, by virtue of its financial support, any ofthe statements made, or views expressed, herein.

Forthcoming Current Anthropology Wenner-Gren SymposiumSupplementary Issues (in order of appearance)

Cu

rrent

An

throp

ologyD

ecemb

er 2013Volum

e54

Supp

lemen

t 8Pag

esS173–S406

T H E U N I V E R S I T Y O F C H I C A G O P R E S S

S p o n s o r e d b y t h e We n n e r - G r e n F o u n d a t i o n f o r A n t h r o p o l o g i c a l R e s e a r c h

S U P P L E M E N T 8 D E C E M B E R 2 0 1 3

Alternative Pathways to Complexity

Mediterranean and Red Sea Paleoclimate

Neanderthal Demographic Estimates

Agreements and Misunderstandings among Three Scientific Fields

Hominin Evolution in the Middle-Late Pleistocene

Variability in the Middle Stone Age of Eastern Africa

Roots of the Middle Paleolithic in Eurasia

Middle Stone Age Hunting Strategies and Diet Breadth

Trends versus Conservatism in the Predatory Niche

Technological Trends in the Middle Stone Age of South Africa

Change and Stasis in the Iberian Middle Paleolithic

Lessons from the Levantine Middle Paleolithic Record

Paleolithic Cultures in China

Mechanisms behind MP and MSA Cultural Trajectories

Population Size and the Paleolithic Archaeological Record

Measuring the Complexity of Lithic Technology

GUEST EDITORS: STEVEN L. KUHN AND ERELLA HOVERS

Crisis, Value, and Hope: Rethinking the Economy. Susana Narotzky and Niko Besnier, eds.

�e Anthropology of Christianity: Unity, Diversity, New Directions. Joel Robbins and Naomi Haynes, eds.

Politics of the Urban Poor. Veena Das and Shalini Randeria, eds.

Working Memory: Beyond Language and Symbolism. �omas Wynn and Frederick L. Coolidge, eds.

Engaged Anthropology: Diversity and Dilemmas. Setha M. Low and Sally Engle Merry, eds.

Corporate Lives: New Perspectives on the Social Life of the Corporate Form. Damani Partridge, Marina Welker, and Rebecca Hardin, eds.

�e Origins of Agriculture: New Data, New Ideas. T. Douglas Price and Ofer Bar-Yosef, eds.

�e Biological Anthropology of Living Human Populations: World Histories, National Styles, and International Networks. Susan Lindee and Ricardo Ventura Santos, eds.

Human Biology and the Origins of Homo. Susan Antón and Leslie C. Aiello, eds.

Potentiality and Humanness: Revisiting the Anthropological Object in Contemporary Biomedicine. Klaus Hoeyer and Karen-Sue Taussig, eds.

Previously Published Supplementary Issues

Wenner-Gren Symposium Series Editor: Leslie AielloWenner-Gren Symposium Series Managing Editors: Laurie Obbink and Daniel SalasCurrent Anthropology Editor: Mark AldenderferCurrent Anthropology Managing Editor: Lisa McKamyBook Reviews Editor: Holley MoyesCorresponding Editors: Claudia Briones (IIDyPCa-Universidad Nacional de Rıo Negro, Argentina; briones@gmail.com), MichalisKontopodis (Humboldt Universitat zu Berlin, Germany; michaliskonto@googlemail.com), Jose Luis Lanata (Universidad Nacionalde Rıo Negro San Carlos de Bariloche, Argentina; jllanta@gmail.com), David Palmer (Hong Kong University, China;palmer19@hku.hk), Anne de Sales (Centre National de la Recherche Scientifique, France; desales.anne@wanadoo.fr), Zhang Yinong(Shanghai University, China; yz36edu@gmail.com)

Please send all editorial correspondence toMark AldenderferSchool of Social Sciences, Humanities, and ArtsUniversity of California, Merced5200 North Lake RoadMerced, CA 95343, U.S.A.(fax: 209-228-4007; e-mail: maldenderfer@ucmerced.edu)

Individual subscription rates for 2014: $76 print � elec-tronic, $45 print-only, $44 e-only. Institutional print � elec-tronic and e-only subscriptions are available through JSTOR’sCurrent Scholarship Program and include unlimited onlineaccess; rates are tiered according to an institution’s type andresearch output: $314 to $628 (print � electronic), $273 to$546 (e-only). Institutional print-only is $332. For additionalrates, including single copy rates and print-only or electronic-only subscriptions, please visit www.journals.uchicago.edu/CA. Additional taxes and/or postage for non-US subscriptionsmay apply. Free or deeply discounted access is available toreaders in most developing nations through the ChicagoEmerging Nations Initiative (www.journals.uchicago.edu/ceni/).

Please direct subscription inquiries, back-issue requests,and address changes to the University of Chicago Press, Jour-nals Division, P.O. Box 37005, Chicago, IL 60637. Telephone:(773) 753-3347 or toll-free in the United States and Canada(877) 705-1878. Fax: (773) 753-0811 or toll-free (877) 705-1879. E-mail: subscriptions@press.uchicago.edu

Reasons of practicality or law make it necessary or desirableto circulate Current Anthropology without charge in certainportions of the world; it is hoped, however, that recipients ofthis journal without charge will individually or collectively invarious groups apply funds or time and energy to the worldgood of humankind through the human sciences. Informationconcerning applicable countries is available on request.

� 2013 by The Wenner-Gren Foundation for AnthropologicalResearch. All rights reserved. Current Anthropology (issn0011-3204) is published bimonthly in February, April, June,August, October, and December by The University of ChicagoPress, 1427 East 60th Street, Chicago, IL 60637-2954.Periodicals postage paid at Chicago, IL, and at additionalmailing offices. Postmaster: Send address changes toCurrent Anthropology, P.O. Box 37005, Chicago, IL 60637.

Current AnthropologyVolume 54 Supplement 8 December 2013

http://www.journals.uchicago.edu/CA

Alternative Pathways to Complexity:Evolutionary Trajectories in the MiddlePaleolithic and Middle Stone Age

Leslie C. AielloAlternative Pathways to Complexity: EvolutionaryTrajectories in the Middle Paleolithic and Middle StoneAge: Wenner-Gren Symposium Supplement 8 S173

Steven L. Kuhn and Erella HoversAlternative Pathways to Complexity: EvolutionaryTrajectories in the Middle Paleolithic and Middle StoneAge: An Introduction to Supplement 8 S176

Eelco J. Rohling, Katharine M. Grant, Andrew P. Roberts,and Juan-Cruz LarrasoanaPaleoclimate Variability in the Mediterranean and Red SeaRegions during the Last 500,000 Years: Implications forHominin Migrations S183

Jean-Pierre Bocquet-Appel and Anna DegioanniNeanderthal Demographic Estimates S202

Carles Lalueza-FoxAgreements and Misunderstandings among ThreeScientific Fields: Paleogenomics, Archaeology, and HumanPaleontology S214

Osbjorn M. PearsonHominin Evolution in the Middle-Late Pleistocene:Fossils, Adaptive Scenarios, and Alternatives S221

Christian A. Tryon and J. Tyler FaithVariability in the Middle Stone Age of Eastern Africa S234

Steven L. KuhnRoots of the Middle Paleolithic in Eurasia S255

Jamie L. Clark and Andrew W. KandelThe Evolutionary Implications of Variation in HumanHunting Strategies and Diet Breadth during the MiddleStone Age of Southern Africa S269

Mary C. StinerAn Unshakable Middle Paleolithic? Trends versusConservatism in the Predatory Niche and Their SocialRamifications S288

Sarah WurzTechnological Trends in the Middle Stone Age of SouthAfrica between MIS 7 and MIS 3 S305

Ignacio de la Torre, Jorge Martınez-Moreno, andRafael MoraChange and Stasis in the Iberian Middle Paleolithic:Considerations on the Significance of MousterianTechnological Variability S320

Erella Hovers and Anna Belfer-CohenOn Variability and Complexity: Lessons from theLevantine Middle Paleolithic Record S337

Xing GaoPaleolithic Cultures in China: Uniqueness and Divergence S358

Francesco d’Errico and William E. BanksIdentifying Mechanisms behind Middle Paleolithic andMiddle Stone Age Cultural Trajectories S371

Mark Collard, Briggs Buchanan, and Michael J. O’BrienPopulation Size as an Explanation for Patterns in thePaleolithic Archaeological Record: More Caution IsNeeded S388

Charles Perreault, P. Jeffrey Brantingham, Steven L. Kuhn,Sarah Wurz, and Xing GaoMeasuring the Complexity of Lithic Technology S397

Current Anthropology Volume 54, Supplement 8, December 2013 S173

� 2013 by The Wenner-Gren Foundation for Anthropological Research. All rights reserved. 0011-3204/2013/54S8-0001$10.00. DOI: 10.1086/673284

Alternative Pathways to Complexity:Evolutionary Trajectories in the

Middle Paleolithic andMiddle Stone Age

Wenner-Gren Symposium Supplement 8

by Leslie C. Aiello

Figure 1. Seated: Leslie Aiello, Xing Gao, Francesco d’Errico, Christian Tryon, Erella Hovers. Standing: Ignacio de la Torre, JamieClark, Sarah Wurz, Anna Degioanni, Eelco Rohling, Mary Stiner, Osbjorn Pearson, Mark Collard, Charles Perreault, Steve Kuhn,Ariel Malinsky-Buller. A color version of this photo appears in the online edition of Current Anthropology.

The Wenner-Gren Foundation for Anthropological Researchhas a long tradition of organizing symposia that deal with the“big” questions in contemporary anthropology. AlternativePathways to Complexity: Evolutionary Trajectories in the MiddlePaleolithic and Middle Stone Age is the 145th in the sympo-sium series and the eighth to be published as an open-access

Leslie C. Aiello is President of the Wenner-Gren Foundation forAnthropological Research (470 Park Avenue South, 8th Floor North,New York, New York 10016, U.S.A.).

supplement of Current Anthropology (see http://www.wennergren.org/history/conferences-seminars-symposia/wenner-gren-symposia for a complete list of symposia andthe history of the symposium program). The Alternative Path-ways symposium was organized by Steven L. Kuhn (Universityof Arizona, U.S.A.) and Erella Hovers (Hebrew University,Israel) and was held at Haringe Slott, Stockholm, Sweden,June 1–8, 2012 (fig. 1).

Haringe Slott was the country home of Axel Wenner-Grenfrom the late 1930s until his death in 1961, and it was par-

S174 Current Anthropology Volume 54, Supplement 8, December 2013

Figure 2. Haringe Slott, the summer home of Axel Wenner-Gren, located in Vasterhaninge, Sweden, approximately 25 km southof Stockholm. A color version of this photo appears in the online edition of Current Anthropology.

ticularly enjoyable to spend an academically stimulating weekin a beautiful environment relevant to the history of the Foun-dation (fig. 2). It is also noteworthy that this symposiumfollowed on from a workshop cosponsored by the Wenner-Gren Foundation for Anthropological Research and the Swed-ish Wenner-Gren Foundations (the Wenner-Gren CenterFoundation for Scientific Research, the Axel Wenner-GrenFoundation for International Exchange of Scientists, and theFoundation Wenner-Grenska Samfundet). The meeting wastitled Reality and Myth: A Symposium on Axel Wenner-Grenand was held at the Wenner-Gren Center in Stockholm, Swe-den (May 30–31, 2012). It provided the first opportunity inrecent years for the independent U.S. and Swedish founda-tions to meet and explore the fascinating history of theirmutual founder. Information on this workshop and the lifeof Axel Wenner-Gren can be found at http://wennergren.org/history/-story-and-people-wenner-gren/people-wenner-gren/axel-wenner-gren.

Alternative Pathways is the most recent symposium in theFoundation’s long history of interest in and support for re-search on human origins. One of the first large meetingsorganized by the Foundation (in 1956 in collaboration withthe city of Dusseldorf, Germany, and H. R. von Koenigswald,Wilhelm Gieseler, and Horst Sieloff), marked the centenaryof the discovery of Neanderthal Man (Neanderthal 1; fig. 3;and see http://www.wennergren.org/history/conferences-seminars-symposia/wenner-gren-symposia/wenner-gren-symposia1950-1970).

Other Wenner-Gren Symposia dealing with human origins

are summarized in the introduction to Human Biology andthe Origins of Homo (Aiello 2012), a recent Current Anthro-pology supplementary issue on human origins (Anton andAiello 2012). The majority of these deal with the earlier phasesof the human evolutionary record, reflecting the focus onaustralopith research in eastern and southern Africa in the1960s and 1970s. A notable exception is After the Australo-pithecines: Stratigraphy, Ecology and Culture Change in theMiddle Pleistocene (Butzer and Isaac 1975). In more recentyears the Foundation has supported meetings on broader hu-man origins topics, for example, Ancestors: The Hard Evidence(Delson 1985), and on broader topics, Tools, Language andCognition in Human Evolution (Gibson and Ingold 1993),Roots of Human Sociality: Culture, Cognition and Interaction(Enfield and Levinson 2006), and Working Memory: BeyondLanguage and Symbolism (Wynn and Coolidge 2010).

Alternative Pathways combines past interests in data-richdiscussions of human origins with more theoretical discus-sions of the significance of paleoanthropology for understand-ing the course of human evolution. As discussed in the in-troduction to this issue (Kuhn and Hovers 2013), MiddlePaleolithic Neanderthals and Middle Stone Age humans pro-vide a unique opportunity “to study the evolution of differentkinds of intelligence and different modes of behavioral com-plexity in sister taxa, using multiple lines of archeological andbiological evidence.” One purpose of the meeting was to in-form our understanding of why Neanderthals went to ex-tinction and modern humans were able to spread throughoutthe world. However, going beyond this, the intention was to

Aiello Alternative Pathways to Complexity: Evolutionary Trajectories in the Middle Paleolithic and Middle Stone Age S175

Figure 3. From the left: Gunter Behm-Blanke, G. H. R. vonKoenigswald, and Abbe H. Breuil examine the Ehringsdorf fossilremains.

explore the alternative elaborate and flexible cultural adaptivesystems associated with two forms of cognitively sophisticatedhominins.

As with most successful meetings, more questions wereraised than answers provided. However, collaborative projects

emerged from the symposium, and these will take the research

forward. Additionally, the papers provide an assessment of

state-of-the art knowledge in the Middle Paleolithic of Eurasia

as well as the Middle Stone Age in Africa, along with relevant

paleoclimatological, genetic, demographic, and biological per-

spectives. This collection is destined to be an important re-

source for the future.

The Wenner-Gren Foundation is always looking for big

questions and innovative new ideas in all areas of anthro-

pology for future Foundation-sponsored and Foundation-

organized symposia and eventual CA publication. We en-

courage anthropologists to contact us with their proposals for

future meetings. Information about the Wenner-Gren Foun-

dation, the symposium program, application procedures and

deadlines, and what constitutes a good symposium topic can

be found on the Foundation’s website (http://wennergren.org

/programs/international-symposia).

References Cited

Aiello, Leslie C. 2012. Human biology and the origins of Homo: Wenner-Grensymposium supplement 6. Current Anthropology 53(suppl. 6):S267–S268.

Anton, Susan C., and Leslie C. Aiello, eds. 2012. Human biology and the originsof “Homo.” Current Anthropology, vol. 53, suppl. 6.

Butzer, Karl W., and Glynn Isaac, eds. 1975. After the Australopithecines: stra-tigraphy, ecology and culture change in the Middle Pleistocene. World An-thropology Series. The Hague: Mouton.

Delson, Eric, ed. 1985. Ancestors: the hard evidence. New York: Liss.Enfield, Nicholas J., and Stephen C. Levinson, eds. 2006. Roots of human

sociality: culture, cognition and interaction. Oxford: Berg.Gibson, Kathleen, and Tim Ingold, eds. 1993. Tools, language and cognition

in human evolution. Cambridge: Cambridge University Press.Kuhn, Steven L., and Erella Hovers. 2013. Alternative pathways to complexity:

evolutionary trajectories in the Middle Paleolithic and Middle Stone Age:an introduction to supplement 8. Current Anthropology 54(suppl. 8):S176–S182.

Wynn, Thomas, and Frederick L. Coolidge, eds. 2010. Working memory: beyondlanguage and symbolism. Current Anthropology, vol. 51, suppl. 1.

S176 Current Anthropology Volume 54, Supplement 8, December 2013

� 2013 by The Wenner-Gren Foundation for Anthropological Research. All rights reserved. 0011-3204/2013/54S8-0002$10.00. DOI: 10.1086/673501

Alternative Pathways to Complexity:Evolutionary Trajectories in the Middle

Paleolithic and Middle Stone AgeAn Introduction to Supplement 8

by Steven L. Kuhn and Erella Hovers

The 145th symposium of the Wenner-Gren Foundation took place June 1–8, 2012, in Haringe Slott near Stockholm,Sweden. The primary goal of the symposium was to reframe discussions of behavioral evolution among Neanderthalsand early modern humans. We hoped to replace conventions of a single scale of evolutionary progress (in whichthe primary benchmark is “modern human behavior”) with a more Darwinian framework that could allow forindependent evolutionary trajectories in different areas. The 15 participants included archaeologists researchingmaterial culture and subsistence in Eurasia, Africa, and China; physical anthropologists; a demographer; a geneticist;modelers of cultural evolution; and a climatologist. Participants were asked to draw on evidence in their areas ofexpertise, focusing on evolutionary trends in both modal tendencies and levels of variation/diversity within variousregions during the interval in which the Middle Stone Age and Middle Paleolithic developed, spread, and eventuallydisappeared. It was agreed that there is compelling evidence for very different trajectories of cultural evolution indifferent parts of the world but that we are not yet in a position to fully evaluate and understand the outcomes ofthe parallel cultural evolutionary pathways among modern Homo sapiens in Africa and Neanderthals in Europe.Answering questions this large in scope requires synthesis on a large geographic scale comparable to studies byclimate scientists and biogeographers. Conventional approaches to collecting, reporting, and analyzing archaeologicaland skeletal data do not lend themselves to rigorous tests of alternative evolutionary models. At the same time, theintellectual tools needed to research these questions are well developed, and answers are within reach.

In his 1989 book Wonderful Life (Gould 1989), the biologistS. J. Gould proposed a singular thought experiment, the ideaof “replaying the tape of life.” Gould’s aim was to raise ques-tions about the role of contingency in the evolution of lifeon earth, to consider whether the same initial conditionscould have led to radically different forms of plants and an-imals because of slight differences in external conditions.Scholars disagree about the likely outcomes of the exercise.Gould (1989:51) believed that repeatedly playing the tape oflife forward would produce very different outcomes, whereasDennett (1995) argues the opposite. Regardless of the specificviews, such a line of inquiry perforce remains a thought ex-periment. And yet for more limited arenas there are naturalsituations that partially replicate Gould’s mental exercise. One

Steven L. Kuhn is Professor at the School of Anthropology of theUniversity of Arizona (Building 30, Tucson, Arizona 85721-0030,U.S.A. [skuhn@email.arizona.edu]). Erella Hovers is Professor at theInstitute of Archaeology of the Hebrew University of Jerusalem(Mount Scopus, Jerusalem 91905, Israel). This paper was submitted3 VII 13, accepted 14 VIII 13, and electronically published 20 XII13.

of these was the subject of the workshop from which thepapers of this volume derive.

Genetic evidence (Endicott, Ho, and Stringer 2010; Lalueza-Fox 2013) show that Neanderthals and anatomically modernHomo sapiens (AMHs) last shared a common ancestor roughly400–500 kya. Although they may have remained geneticallycompatible (Sankararaman et al. 2012), by the time of the mostrecent of the human dispersals out of Africa and into westernEurasia, the two metapopulations had been partially if notwholly isolated from one another for at least 300,000 yr. Be-tween the time the two lineages separated and the time theycame into contact again in marine isotope stage (MIS) 4, acertain amount of encephalization and a great deal of culturalevolution had occurred in both of them. The long period ofgeographic vicariance rendered Neanderthals and African H.sapiens both behaviorally and anatomically differently deriveddescendants of their last common ancestor. By the time theymet again in Eurasia, both Neanderthals and African H. sap-iens had developed culturally aided adaptations that weremore complex than those of any other organism in the historyof the planet, in essence creating new regimes of evolutionary

Kuhn and Hovers Alternative Pathways to Complexity S177

change. Thus, the two taxa can be seen as the outcomesof parallel evolutionary “experiments” resulting in large-brained, behaviorally complex, yet quite distinct hominins.The parallel biological and cultural evolution in Europeand Africa certainly differs from Gould’s original exampleof the Burgess Shale fauna in that it involved evolutionaryprocesses unfolding in different habitats and leading todiversification, whereas Gould was referring to loss of di-versity. Still, the underlying principle that one can get verydifferently evolved outcomes from the same starting pointremains the same.

For most of the history of archaeology and paleoan-thropology, studies of Neanderthals and early AMHs havetypically adopted a very different perspective on the re-lationship between the two taxa than the one implied byGould’s metaphor. Comparisons of material culture evi-dence from Eurasia and Africa have focused mainly on theevolutionary status of the Middle Paleolithic (MP) andMiddle Stone Age (MSA) in relation to the presumed com-plexity of later time periods. The behavioral evolution ofboth taxa has been evaluated using a single scale, definedin terms of a shifting set of traits referred to as “modernhuman behavior.” Studies tend to focus on the manifes-tations of particular behavioral traits (ornaments, bonetools, small-game hunting) among Neanderthals or AMHsas testament to what members of the two taxa could orcould not do. This evidence is not treated symmetrically.For the most part, similarities are equated with equalitybetween the two populations, but differences tend to beinferred as deficiencies on the part of Neanderthals. Be-cause the field remains focused on the question of whathappened to the Neanderthals and why H. sapiens cameto be the only hominin on the planet, differences or sim-ilarities in turn are used to justify or undermine hypothesesabout the behavioral, cultural, or cognitive superiority ofAMHs.

The differing behavioral tendencies of MP and MSA hom-inins may or may not help us explain the historical circum-stances of the Neanderthal’s disappearance and the rapid dis-persal and persistence of modern humans across the globe.However, we believe that attempts to establish the relativepositions of the MP and MSA in a single progressive, devel-opmental scheme culminating in “behavioral modernity” aremisguided (Hovers 2009a, 2009b; Hovers and Belfer-Cohen2006; Kuhn and Hovers 2006). They represent a holdoverfrom the anagenetic mind-sets of early twentieth-century cul-tural evolutionists, for whom all of human cultural variationcould be fit into a single set of developmental stages thatmarked the gradual elevation of the human race from savageryto civilization or from acephalous bands to complexly or-ganized states. Crucially, attempts to organize the MP andMSA into a linear framework of change toward “modernity”distract us from a unique and more valuable opportunitycreated by the long separation of two human lineages involvedin authoring these records. The differences between the be-

havioral records of the two taxa represent alternative elaborateand flexible cultural adaptive systems associated with twoforms of cognitively sophisticated hominins.

The notion of biological and cultural evolution in Africaand Eurasia as parallel experiments has significance beyondexplaining the record of hominin evolution in both areas.Comparison of the evolutionary trajectories taken by Nean-derthals and AMHs affords a rare prospect in the history oflife on earth to study the evolution of different kinds of in-telligence and different modes of behavioral complexity insister taxa, using multiple lines of archaeological and biolog-ical evidence. There is growing interest across the biologicaland social sciences in topics such as the evolution of cultureand niche construction. Humans are typically held up as nicheconstructors, and modern humans are regarded as the ulti-mate expression of a culturally adapted organism with noequal in the animal kingdom. Because they are so differentfrom other living organisms, they are regarded as an extremecase. Treating the records of Neanderthals and AMHs as twoindependent pathways of cognitive and behavioral evolutionopens the prospect of a much richer and more nuanced un-derstanding of the potential for different kinds of culturallyaided niche construction in cognitively sophisticated organ-isms. It could shift the discussion from the human culturalniche to diverse sorts of cultural niches created and occupiedby contemporaneous hominins.

In organizing the Wenner-Gren symposium, we were in-terested in creating a forum for empirically based, theoreticallyinformed discussions of different evolutionary trajectories ofNeanderthals, African H. sapiens, and other hominins inde-pendent of the eventual fate of the population. The goal wasnot to formulate a consensus about which population firstreached some threshold of “behavioral modernity” (a conceptthat is epistemologically and practically problematic; see, e.g.,Belfer-Cohen and Hovers 2010; Henshilwood and Marean2003; Hovers 2009a, 2009b; Hovers and Belfer-Cohen 2006;Kuhn and Hovers 2006; Shea 2011). From the perspectivethat we adopt here, that particular question is anecdotal. In-stead, the specific aims of the symposium were to discuss andevaluate the evidence for independent histories of behavioralevolution in western Eurasia and southern Africa. We werenot so ambitious or so naive as to expect a final, compre-hensive account of the outcome and implications of the Eur-asian and African evolutionary “experiments.” Our aim wasconsiderably more modest. A previous symposium and editedvolume (Hovers and Kuhn 2006) explored the evidence forbehavioral change within the MP and MSA. The more recentworkshop turned to the question of outcomes—what hap-pened in the two macroregions—and how variables such asenvironmental change and demographic dynamics shaped(pre)historical trajectories. Equally, we hoped to reframe thediscussion and to reach a clearer understanding of the kindsof information still needed to answer the big questions wewere posing.

With these questions in mind, the 145th symposium of

S178 Current Anthropology Volume 54, Supplement 8, December 2013

the Wenner-Gren Foundation was convened June 1–8,2012, in Haringe Slott near Stockholm, Sweden (befittingly,a former residence of the foundation’s financial benefactorand namesake, Axel Wenner-Gren). The location and tim-ing enabled participants to enjoy (or suffer) 20 hours ofsunlight per day (on the days when the sun made an ap-pearance, that is). Invitees were asked to draw on evidencein their areas of expertise, focusing on evolutionary trendsin both modal tendencies and levels of variation/diversitywithin various regions. As dictated by the nature of thequestions asked, the 15 participants brought to the dis-cussion a diversity of training and expertise. The groupincluded paleoanthropologists who study material cultureand subsistence in Eurasia, Africa, and China; physical an-thropologists; a demographer; a geneticist; modelers of cul-tural evolution; and a climatologist. The papers and dis-cussions focused on the time period between 400 and 40ka (MIS 9–3), the interval in which the MSA and MPdeveloped, spread, and eventually disappeared.1 That manyof the papers presented in the formal session were co-authored by two or more researchers is perhaps a reflectionof the scope and scale of the questions addressed.

The first part of the workshop was devoted to establishingthe empirical evidence for evolutionary trajectories in Eurasiaand southern Africa. Though differing in the details, the his-tories of the African MSA and European MP seem to besimilar, with a gradual coalescence of behavioral character-istics over the second half of the Middle Pleistocene (Kuhn2013; Tryon and Faith 2013). Common features such as Le-vallois technology are likely to represent convergent evolutionrather than diffusion: in fact, there may be multiple Levalloisorigins within both Africa and Eurasia (Villa 2001; White andAshton 2003:605). Contrary to some characterizations, boththe MSA and the MP show a notable level of technologicalvariability, particularly in methods for producing flake blanks.Interestingly, the sorts of regional variability long noted inMP sequences from Eurasia are not as strongly expressed inthe East African record studied by Tryon and Faith (2013).Although substantial differences in sample size potentiallytemper any conclusions drawn from these observations, suchgeographic variability as exists among East African MSA andEuropean assemblages may correspond with broadly definedhabitat types (e.g., d’Errico and Banks 2013). The grain ofvariation in behavioral evidence in both Africa and Eurasiamay reflect the structure of populations on the landscape,which in turn is strongly influenced by the scale of environ-mental variation and presence of ecological barriers (Hoversand Belfer-Cohen 2013; Tryon and Faith 2013). Cultural di-versity in western Eurasia may also have deeper chronologicalroots, as there are strong regional differences in technologyas early as 400 kya. In contrast, the African MSA seems to

1. The student moderator for the meeting, Mr. Ariel Malinsky-Buller,is currently writing his PhD thesis on the transition from the late Lowerto the early Middle Paleolithic in Eurasia.

have developed everywhere out of local variants of the Acheu-lean.

Stronger contrasts emerge when comparing the archae-ological records of the later MSA and MP. Sites from acrosssouthern Africa show similar patterns of cultural successionduring MIS 4, with MSA 1 and 2 followed by distinctivebut short-lived Still Bay, Howiesons Poort, and “post–How-iesons Poort” technocomplexes (but see Brown et al. 2012,who suggest that the punctuated nature of technologicalpractices associated with these cultural phenomena is anartifact of sampling). The Still Bay and Howiesons Poorthave received a great deal of attention in recent years becauseof evidence for “complex” behaviors such as bone-tool mak-ing, personal ornaments, heat treatment of raw materials,and abstract symbolic renderings (Wurz 2013). The orderlysequencing of assemblage types across southern Africa alsofits many researchers’ expectations for cumulative, progres-sive cultural evolution or “cultural ratcheting” (Tennie, Call,and Tomasello 2009; Tomasello 1999) similar to what isobserved in the Eurasian Upper Paleolithic (UP).

Well-studied areas of southwest Europe show rather di-vergent evolutionary patterns during the late MP. As de laTorre and colleagues (2013) demonstrate, there are no strongdirectional trends within the northern Spanish MP recordthrough the Upper Pleistocene, whereas the southwesternFrench Mousterian shows a distinct sequence of industriesover time (Delagnes, Jaubert, and Meignen 2007; Delagnesand Meignen 2006). The presence of very divergent regionaltrajectories in neighboring regions such as northern Iberiaand southern France exemplifies the geographically parti-tioned nature of the European MP record. These observationsalso highlight one apparent difference in the Eurasian andAfrican records, namely the apparently more progressive pat-tern of culture change in the MSA of southern Africa. Thisraises the question of whether the Still Bay and HowiesonsPoort represent the first evidence of cumulative, directionalhuman cultural evolution, as some have argued, or whetherthey simply reflect more rapid diffusion of novel ideas acrosssouthern African landscapes. Conversely, the more localizedand divergent patterns in southern Europe could be evidencefor different cultural capacities of the hominins for acquisitionand diffusion of cultural information, but they could alsosimply be a consequence of more fragmented habitats andless thoroughly interconnected populations.

The contribution by Gao (2013) brings into the discussiona very different record. The workshop focused on the MSAand western Eurasian MP simply because the preponderanceof the evidence comes from those macroregions. A substantialbody of archaeological and fossil evidence is now availablefrom China, even if it is not well known to Western scholars.In contrast to Africa and western Eurasia, where lithic tech-nology provides robust evidence of culture change and di-versification during the Paleolithic, the Chinese lithic recordis remarkably stable from the Early Pleistocene through thelate Pleistocene or even into the early Holocene. It makes little

Kuhn and Hovers Alternative Pathways to Complexity S179

sense to talk about a distinct “MP” phenomenon, and overmost of that vast country even early UP stone artifacts differremarkably little from much older Lower Paleolithic ones (seeBar-Yosef and Wang 2012). At the same time, elements ofmaterial culture such as ornaments and bone and antler ar-tifacts made their way into the behavioral repertoire of Chi-nese hominins during the Upper Pleistocene. The Middle andUpper Pleistocene archaeology of China strongly suggests thatthis region may represent yet another independent trajectoryin the evolution of behavioral complexity during the UpperPleistocene. Recent analyses of ancient and modern DNAindicate that northeast Asia was also home to a distinct pop-ulation of hominins, further demonstrating its independencefrom “the West.” We suspect that there are other parallelsituations in macroregions such as south Asia and possiblyNorth Africa not covered in this workshop.

It was clear from presentations and discussions during thesymposium that demography and its effects on processes ofcultural transmission have become important to many ac-counts of MP/MSA technological variation and foraging be-haviors. The majority of participants subscribed to the viewthat Neanderthals were as a rule spread very thin on thelandscape. Demography can influence the flow of culturalinformation just as it does the flow of genetic information,and many interesting hypotheses about variation and inno-vation in cultural practices as well as trajectories of biologicalchange can be derived from this single premise. Interestingly,the physical anthropologist who participated in the discussion(Pearson 2013) also concludes that demographic effects suchas multiple bottlenecks may explain many of the anatomicalcharacteristic of native Eurasian hominins (i.e., Neanderthals)better than models of adaptation.

As compelling as demographic explanations may be, theyare still in their infancy. To date, models of demographiceffects on Paleolithic cultural dynamics have been verified(checked for internal consistency) but have not been validated(i.e., checked against real-world evidence; Collard, Buchanan,and O’Brien 2013). The papers by Boucquet-Appel and De-gioanni (2013) and by Lalueza-Fox (2013) point out some ofthe difficulties in testing these models. Estimates of demo-graphic potentials based on life-history characteristics are no-toriously sensitive to component variables and often produceconflicting predictions, whereas estimates of effective popu-lation size from genetic data are not easily translated intocensus population size. Attempts to reconstruct census pop-ulation levels based on archaeological evidence (number ofdated sites, site area, etc.; see, e.g., Mellars and French 2011)are uncertain (Dogandzic and McPherron 2013). Thus, whilethere is both empirical and theoretical reason to believe thatNeanderthals lived at comparatively low population densities,we have a very coarse understanding of actual MP populationsizes or of the relative sizes of contemporaneous or later H.sapiens populations.

Other sources of information may provide proxy evidencefor changes in demography if not actual population sizes.

Evidence of subsistence intensification in the form either ofexpanding diet breadth or intensified resource processing hasbeen widely associated with changes in population-resourcebalances. When the effects of changing climates are filteredout, the MP archaeofaunal record of Eurasia (summarized byStiner 2013) provides very little evidence for sustained sub-sistence intensification. The hominins producing MP indus-tries seem to have focused their meat-oriented foraging effortson high-ranked, large-body-sized ungulate prey. There is reg-ular use of some specific kinds of small game and sporadicevidence for expansion of the meat diet to include more costlyand lower-ranked items such as birds and small mammalseven in the early MP. However, these cases seem isolated andso far do not resolve into evidence for sustained intensificationof subsistence. The more rarefied archaeofaunal record fromsouthern Africa (synthesized by Clark and Kandel 2013)shows a clear response to climate, especially evident in faunasdating to MIS 4. Still, while shellfish exploitation appears quiteearly in southern Africa (as it does in Iberia and other partsof the Mediterranean), there are no long-term trends indic-ative of continuous pressure on high-ranked resources unlessthe gradual increase in suid exploitation is considered as such.Thus, the South African faunal evidence at least provides littlesupport for the notion that demographic conditions were verydifferent there than in Eurasia.

Vegetable foods may have played an important, even dom-inant role in the diets of hominins in many times and places,but macroscopic plant remains seldom preserve over Paleo-lithic timescales. While occasional discoveries of preservedseeds from archaeological sites as well as recent findings ofphytoliths and starch grains adhering to teeth or evidence forplant residues on artifacts (e.g., Barton et al. 1999; Hardy andMoncel 2011; Henry, Brooks, and Piperno 2010; Lev, Kislev,and Bar-Yosef 2005) open a new window on the diets ofNeanderthals and other hominins, the unfortunate fact re-mains that various sources of data are incommensurate. An-imal bones are common and quantifiable but reveal only partof the diet. Stable isotope data provide a better proportionalestimate of the importance of plant and animal foods, butonly as regards their contributions to protein intake. Mean-while, emerging but still sparse data on plant foods in thePaleolithic tell us about diversity but not quantity; that is, weknow what people might have eaten but not how much. In-tegrating these disparate sources into a more coherent pictureof dietary variation is a challenge yet to be addressed by thefield. The work of nutritionists and ecologists, who have iden-tified a number of patterns of the use of fish, meat, and plantresources among recent foraging populations, can be helpfulin contextualizing some of the archaeological data into a morenuanced view of dietary variations in various regions (e.g.,Cordain et al. 2000; Hockett 2012; Hockett and Haws 2005;Kelly 1995; see review in Speth 2010). On the other hand,these observations are themselves contingent on the particularhistorical, climatic, and demographic circumstances of the lastfew hundred years. Another line of evidence for dietary in-

S180 Current Anthropology Volume 54, Supplement 8, December 2013

tensification—and indirectly, population-resource imbal-ances—could come from technology. Maximally efficient useof plant foods such as seeds and nuts often requires intensiveprocessing to achieve maximum nutrient yields. Moreover,the technological accoutrements of this sort of food process-ing—grinding or pounding stones—are durable and large,unlike the botanical remains themselves. Reports of grind-stones from MSA sites in southern Africa may hint at anunexplored dimension of subsistence intensification in theremote past, but until these artifacts are systematically ana-lyzed in Africa as well as in Eurasia, they remain little morethan tantalizing anecdotes.

Overall, the similarities in subsistence regimes between latePleistocene H. sapiens in southern Africa and late PleistoceneNeanderthals in western Eurasia far outweigh the differences.This observation in turn highlights another interesting dif-ference in the outcomes of independent trajectories of culturalevolution in the two regions. Recent research has producedindisputable evidence that both MSA and MP hominins reg-ularly produced composite, multipart tools, some of whichwere weapons (Rots 2013; Tryon and Faith 2013; Villa andLenoir 2008; Villa et al. 2009; Wilkins et al. 2012; Wurz 2013).However, the frequency, variety, and elaboration of projectileelements seems to be higher in the later MSA than the lateMP even though the Pleistocene environments of temperateEurasia should have favored greater dependence on large gamethan the subtropical and tropical habitats in which MSA pop-ulations operated. Among recent foragers, greater dependenceon large game has been shown to correlate with more in-vestment in hunting technology (see Collard, Buchanan, andO’Brien 2013). The divergence was not caused by the absenceof key innovations or some cognitive difference, as the samebasic procedures and techniques had long existed in bothareas. One potential explanation is that MP diets in fact in-volved less use of large game than is normally thought. Analternate explanation is that MP hominins used means suchas group-wide social cooperation to improve hunting effec-tiveness. Finally, unstable demographic conditions in Eurasiacould have prevented MP hominins from reliably passing oncomplex technological procedures. These kinds of contrastspoint to possible fundamental divergences in the culturalniches of African MSA and Eurasian MP hominins. At thesame time they reveal anomalies that must be resolved forthe field to move forward.

In the end, understanding how the sizes and structuresof Neanderthal and AMH populations may have influencedcultural evolutionary processes will require more than justmore precise estimates of population sizes. The resultingmodels must also be made empirically testable using ar-chaeological data. Accomplishing this will require moreprecise definitions and replicable measures of key param-eters such as complexity and diversity of cultural behaviorsanalogous but not identical to some measures in geneticsand ecology. During the last days of the meeting, partic-ipants worked toward using the empirical data and theo-

retical models that had been discussed to formulate testablehypotheses about rates of innovation, change, and cu-mulative cultural evolution. One working group developeda framework for evaluating complexity of technical pro-cesses as a means of assessing cumulative cultural evolu-tion. The paper by Perrault et al. (2013) contains an initialapplication of this framework to Paleolithic lithic tech-nological evidence, a “proof of concept” for this analyticaland methodological approach. A second group exploredstrategies for comparing assemblages of material cultureacross regions with very different histories as a means ofevaluating both cultural phylogenies and adaptive varia-tion.

In many respects, the rise of population-based argumentsin paleoanthropology today is similar to what occurred inpopulation genetics in the 1960s or in ecology during the late1990s, when neutral theory was first introduced into thosedisciplines by Kimura (1968, 1983) and Hubbell (2001), re-spectively. Recognition of the importance of replicator dy-namics led to an increasing emphasis on stochastic, neutralprocesses such as drift or dispersal at the expense of selectionor adaptation. In fact, both stochastic and directional/selectiveprocesses influence most genetic and cultural systems. Thebig challenge for archaeology and human paleontology alikeis to sort out what is neutral from what is selectively oradaptively salient in a particular case. Papers by Collard,Buchanan, and O’Brien (2013) and Pearson (2013) show howthis could be approached for both anatomy and material cul-ture.

Obviously, climate is a key element in any evolutionaryscenario. A third theme that figured prominently both in thisworkshop and in the current literature concerns the effectsof climatic events on the evolutionary histories of differenthominin species and different regions. Attempts to matchtrends in anatomical and archaeological data to global climatecurves are basically exercises in pattern recognition, underlainby the explicit assumption that association equals causation,rather than concerted attempts at explanation. As Rohling etal. (2013) emphasize, matching scales of observation and anal-ysis is a crucial prerequisite to integrating data from diversefields. When paleoanthropology turns to climate sciences todefine the environmental envelope of MSA/MP behaviors invarious regions, spatial and temporal scales are often incom-patible. Climate variability in the Upper Pleistocene is oftendocumented on a millennial, centennial, or (rarely) decadalscale of resolution, whereas dating of archaeological sites doesnot normally achieve such accuracy. At the same time, sitesare by nature of restricted spatial scales whereas major climaticshifts occur globally yet manifest themselves differently atsmaller geographic scales. Hominins in Eurasia and southernAfrica were exposed to the same global climatic events, butthey experienced them in very different ways. Features of thelandscape (as well as technological repertoires and anatomy)play a major role in determining how these global events aremanifest at the scale of sites or research areas. An important

Kuhn and Hovers Alternative Pathways to Complexity S181

focus of discussion during the latter part of the workshopwas developing strategies for resolving such geographic andtemporal incongruities, ranging from development of localrecords to focusing on periods of high and low variabilityrather than specific sets of conditions (Rohling et al. 2013).The notion of a “Cohesive Adaptive System,” or CAS, asoutlined by d’Errico and Banks here and in other papers(Banks et al. 2006), is one approach to analyzing culture andenvironmental variation on similar spatial and temporalscales.

Differences between southern Africa and western Eurasiawith respect to the magnitude of climate change over timeand spatial grain of environmental variation will make itdifficult though not impossible to fully exploit the naturalexperiments that inspired this workshop. However, otherareas may provide unique controls over these variables.The Levant offers a particularly interesting and useful case.Because of the strong habitat zonation and limited mag-nitude of temperature and rainfall variation when com-pared with more extreme latitudes, the region provided arelatively stable environmental setting over the course ofthe Pleistocene compared with Europe. It also has a“checkered” history in terms of hominin dispersals, withmultiple taxa, including both early AMHs and Neander-thals, coming and going. Thanks to many highly detailedstudies of lithic material culture and of nonlithic relatedbehaviors combined with high resolution terrestrial cli-matic records, research on the Levantine MP may serve asa first approximation for constructing scale-appropriateresearch programs that are suitable to address the specificquestions raised in the meeting and a basis on which toincorporate new methods of analysis. The paper by Hoversand Belfer-Cohen (2013) explores some of the apparentconsequences of this unique environment for the behaviorof successive hominin populations. South Africa andsouthwest Europe (northern Iberia, southern France) werealso nominated as areas of roughly comparable size anddata density, although they have very different environ-mental histories.

One conclusion from the meeting was that we are not yetin a position to fully evaluate and understand the outcomesof the parallel evolutionary trajectories represented by AMHsand the MSA in Africa and Neanderthals and the MP inEurope. The meeting has underlined the fact that whereas thebasic notion is sound and some of the available evidencecompelling, conventional approaches to collecting, reporting,and analyzing data do not lend themselves to rigorous testsof alternative evolutionary models. Answering questions thislarge in scope requires synthesis on a vast geographic scalesimilar to efforts undertaken by climate scientists and bio-geographers. At the same time, the participants agreed thatthe intellectual tools needed to research these questions arewell within reach. Various members of the group have beguncollaborative efforts in this direction, including the designand implementation of pilot studies of cumulative evolution

using material culture data and attempts to identify phylo-genetic relationships between MP and MSA cultural entities.This fulfills one final aim of the workshop, which was to openup new avenues for collaboration among participants andtheir colleagues in a range of disciplines.

References CitedBanks, William E., Francesco d’Errico, Harold L. Dibble, Leonard Krishtalka,

Dixie West, Deborah I. Olszewski, A. Townsend Peterson, et al. 2006. Eco-cultural niche modeling: a new tool for reconstructing the geography andecology of past human populations. Paleoanthropology 2006:68–83.

Barton, R. N. E., Andrew Currant, Yolanda Fernandez-Jalvo, Clive Finlayson,Paul Goldberg, Richard Macphail, Paul Pettit, and Chris Stringer. 1999.Gibraltar Neanderthals and results of recent excavations in Gorham’s, Van-guard, and Ibex caves. Antiquity 73:13–24.

Bar-Yosef, Ofer, and Youping Wang. 2012. Paleolithic archaeology in China.Annual Review of Anthropology 41:319–335.

Belfer-Cohen, Anna, and Erella Hovers. 2010. Modernity, enhanced workingmemory, and the Middle to Upper Paleolithic record in the Levant. CurrentAnthropology 51(suppl. 1):S167–S175.

Bocquet-Appel, Jean-Pierre, and Anna Degioanni. 2013. Neanderthal demo-graphic estimates. Current Anthropology 54(suppl. 8):S202–S213.

Brown, Kyle S., Curtis W. Marean, Zenobia Jacobs, Benjamin J. Schoville,Simen Oestmo, Erich C. Fisher, Jocelyn Bernatchez, Panageorgis Karkanas,and Thalassa Matthews. 2012. An early and enduring advanced technologyoriginating 71,000 years ago in South Africa. Nature 491:590–593.

Clark, Jamie L., and Andrew W. Kandel. 2013. The evolutionary implicationsof variation in human hunting strategies and diet breadth during the MiddleStone Age of southern Africa. Current Anthropology 54(suppl. 8):S269–S287.

Collard, Mark, Briggs Buchanan, and Michael J. O’Brien. 2013. Populationsize as an explanation for patterns in the Paleolithic archaeological record:more caution is needed. Current Anthropology 54(suppl. 8):S388–S396.

Cordain, Lauren, Janette B. Miller, S. Boyd Eaton, Neil Mann, Susanne H.Holt, and John D. Speth. 2000. Plant-animal subsistence ratios and mac-ronutrient energy estimations in worldwide hunter-gatherer diets. AmericanJournal of Clinical Nutrition 71:682–692.

Delagnes, Anne, Jacques Jaubert, and Liliane Meignen. 2007. Les technocom-plexes du Paleolithique moyen en Europe occidentale dans leur cadre dia-chronique et geographique. In Les neandertaliens: biologie et cultures. Ber-nard Vandermeersh and Bruno Maureille, eds. Pp. 213–229. DocumentsPrehistoriques 23. Paris: Editions du Comite des Travaux Historiques etScientifiques.

Delagnes, Anne, and Liliane Meignen. 2006. Diversity of lithic productionsystems during the Middle Palaeolithic in France: are there any chrono-logical trends? In Transitions before the transition: evolution and stability inthe Middle Paleolithic and Middle Stone Age. Erella Hovers and Steven L.Kuhn, eds. Pp. 85–107. New York: Springer.

de la Torre, Ignacio, Jorge Martınez-Moreno, and Rafael Mora. 2013. Changeand stasis in the Iberian Middle Paleolithic: considerations on the signifi-cance of Mousterian technological variability. Current Anthropology54(suppl. 8):S320–S336.

Dennett, Daniel C. 1995. Darwin’s dangerous idea: evolution and the meaningsof life. New York: Simon & Schuster.

d’Errico, Francesco, and William E. Banks. 2013. Identifying mechanismsbehind Middle Paleolithic and Middle Stone Age cultural trajectories. Cur-rent Anthropology 54(suppl. 8):S371–S387.

Dogandzic, T., and S. P. McPherron 2013. Demography and the demise ofNeandertals: a comment on “Tenfold population increase in Western Europeat the Neandertal-to-modern human transition.” Journal of Human Evo-lution 64:311–313.

Endicott, Phillip, Simon Ho, and Chris Stringer. 2010. Using genetic evidenceto evaluate four palaeoanthropological hypotheses for the timing of Ne-anderthal and modern human origins. Journal of Human Evolution 59:87–95.

Gao, Xing. 2013. Paleolithic cultures in China: uniqueness and divergence.Current Anthropology 54(suppl. 8):S358–S370.

Gould, Stephen J. 1989. Wonderful life: the Burgess Shale and the nature ofhistory. New York: Norton.

S182 Current Anthropology Volume 54, Supplement 8, December 2013

Hardy, Bruce L., and Marie-Helene Moncel. 2011. Neanderthal use of fish,mammals, birds, starchy plants and wood 125–250,000 years ago. PLoSONE 6:e23768.

Henry, Amanda G., Alison S. Brooks, and Dolores R. Piperno. 2010. Micro-fossils in calculus demonstrate consumption of plants and cooked foods inNeanderthal diets (Shanidar III, Iraq; Spy I and II, Belgium). Proceedingsof the National Academy of Sciences of the USA 108:486–491.

Henshilwood, Christopher S., and Curtis W. Marean. 2003. The origin ofmodern human behavior: critique of the models and their test implications.Current Anthropology 44:627–651.

Hockett, Bryan. 2012. The consequences of Middle Paleolithic diets on preg-nant Neanderthal women. Quaternary International 264:78–82.

Hockett, Bryan, and Jonathan A. Haws. 2005. Nutritional ecology and thehuman demography of Neandertal extinction. Quaternary International 137:21–34.

Hovers, Erella. 2009a. The lithic assemblages of Qafzeh Cave. New York: OxfordUniversity Press.

———. 2009b. The Middle-to-Upper Paleolithic transition: what news? InSourcebook of Paleolithic transitions: methods, theories, and interpretations.Marta Camps and Parth Chauhan, eds. Pp. 455–462. New York: Springer.

Hovers, Erella, and Anna Belfer-Cohen. 2006. “Now you see it, now youdon’t”: modern human behavior in the Middle Paleolithic. In Transitionsbefore the transition: evolution and stability in the Middle Paleolithic andMiddle Stone Age. Erella Hovers and Steven L. Kuhn, eds. Pp. 295–304.New York: Springer.

———. 2013. On variability and complexity: lessons from the LevantineMiddle Paleolithic record. Current Anthropology 54(suppl. 8):S337–S357.

Hovers, Erella, and Steven L. Kuhn, eds. 2006. Transitions before the transition:evolution and stability in the Middle Paleolithic and Middle Stone Age. NewYork: Springer.

Hubbell, Stephen P. 2001. The unified neutral theory of biodiversity and bio-geography. Princeton, NJ: Princeton University Press.

Kelly, Robert L. 1995. The foraging spectrum: diversity in hunter-gatherer life-ways. Washington, DC: Smithsonian Institution.

Kimura, Motoo. 1968. Evolutionary rate at the molecular level. Nature217(5129):624–626.

———. 1983. The neutral theory of molecular evolution. Cambridge: Cam-bridge University Press.

Kuhn, Steven L. 2013. Roots of the Middle Paleolithic in Eurasia. CurrentAnthropology 54(suppl. 8):S255–S268.

Kuhn, Steven L., and Erella Hovers. 2006. General introduction. In Transitionsbefore the transition: evolution and stability in the Middle Paleolithic andMiddle Stone Age. Erella Hovers and Steven L. Kuhn, eds. Pp. 1–11. NewYork: Springer.

Lalueza-Fox, Carles. 2013. Agreements and misunderstandings among threescientific fields: paleogenomics, archaeology, and human paleontology. Cur-rent Anthropology 54(suppl. 8):S214–S220.

Lev, Ephraim, Mordechai E. Kislev, and Ofer Bar-Yosef. 2005. Mousterianvegetal food in Kebara Cave, Mt. Carmel. Journal of Archaeological Science32:475–484.

Mellars, Paul, and Jennifer C. French. 2011. Tenfold population increase in

Western Europe at the Neandertal-modern human transition. Science 333:623–627.

Pearson, Osbjorn M. 2013. Hominin evolution in the Middle-Late Pleistocene:fossils, adaptive scenarios, and alternatives. Current Anthropology 54(suppl.8):S221–S233.

Perreault, Charles, P. Jeffrey Brantingham, Steven L. Kuhn, Sarah Wurz, andXing Gao. 2013. Measuring the complexity of lithic technology. CurrentAnthropology 54(suppl. 8):S397–S406.

Rohling, Eelco J., Katharine M. Grant, Andrew P. Roberts, and Juan-CruzLarrasoana. 2013. Paleoclimate variability in the Mediterranean and RedSea regions during the last 500,000 years: implications for hominin migra-tions. Current Anthropology 54(suppl. 8):S183–S201.

Rots, Veerle. 2013. Insights into early Middle Palaeolithic tool use and haftingin Western Europe: the functional analysis of level IIa of the early MiddlePalaeolithic site of Biache-Saint-Vaast (France). Journal of ArchaeologicalScience 40:497–506.

Sankararaman, Sriram, Nick Patterson, Heng Li, Svante Paabo, and DavidReich. 2012. The date of interbreeding between Neandertals and modernhumans. PLoS Genetics 8(10):e1002947, doi:10.1371/journal.pgen.1002947.

Shea, John J. 2011. Homo sapiens is as Homo sapiens was: behavioral variabilityversus “behavioral modernity” in Paleolithic archaeology. Current Anthro-pology 52:1–35.

Speth, John D. 2010. The paleoanthropology and archaeology of big-game hunt-ing: protein, fat, or politics? New York: Springer.

Stiner, Mary C. 2013. An unshakable Middle Paleolithic? trends versus con-servatism in the predatory niche and their social ramifications. CurrentAnthropology 54(suppl. 8):S288–S304.

Tennie, Claudio, Josep Call, and Michael Tomasello. 2009. Ratcheting up theratchet: on the evolution of cumulative culture. Philosophical Transactionsof the Royal Society B 364(1528):2405–2415.

Tomasello, Michael. 1999. The cultural origins of human cognition. Cambridge,MA: Harvard University Press.

Tryon, Christian A., and J. Tyler Faith. 2013. Variability in the Middle StoneAge of eastern Africa. Current Anthropology 54(suppl. 8):S234–S254.

Villa, Paola. 2001. Early Italy and the colonization of Western Europe. Qua-ternary International 75:113–130.

Villa, Paola, Paolo Boscato, Filomena Ranaldo, and Annamaria Ronchitelli.2009. Stone tools for the hunt: points with impact scars from a MiddlePaleolithic site in southern Italy. Journal of Archaeological Science 36(3):850–859.

Villa, Paola, and Michel Lenoir. 2008. Hunting and hunting weapons of theLower and Middle Paleolithic of Europe. In The evolution of hominin diets.Jean-Jacques Hublin and Michael Richards, eds. Pp. 59–87. Dordrecht:Springer.

White, Mark, and Nick Ashton. 2003. Lower Paleolithic core technology andthe origins of the Levallois method in north-western Europe. Current An-thropology 44:598–608.

Wilkins, Jayne, Benjamin J. Schoville, Kyle S. Brown, and Michael Chazan.2012. Evidence for early hafted hunting technology. Science 338:942–946.

Wurz, Sarah. 2013. Technological trends in the Middle Stone Age of SouthAfrica between MIS 7 and MIS 3. Current Anthropology 54(suppl. 8):S305–S319.

Current Anthropology Volume 54, Supplement 8, December 2013 S183

� 2013 by The Wenner-Gren Foundation for Anthropological Research. All rights reserved. 0011-3204/2013/54S8-0003$10.00. DOI: 10.1086/673882

Paleoclimate Variability in the Mediterraneanand Red Sea Regions during the

Last 500,000 YearsImplications for Hominin Migrations

by Eelco J. Rohling, Katharine M. Grant, Andrew P. Roberts,and Juan-Cruz Larrasoana

The Mediterranean–Red Sea region has been critical to dispersal of hominids and other species between Africa andthe rest of the world, and climate and sea level are thought to be key controls on migration pathways. Assessingclimate variations, we highlight increased millennial-scale variability at 480–460, 440–400, 380–360, 340–320, 260–220, 200–160, 140–120, and 80–40 thousand years ago (ka), which likely caused intermittent habitat fragmentation.We also find that passageways across the Sahara Desert and the northern out-of-Africa route (from Egypt into theLevant) were intermittently open during pluvials associated with orbital insolation maxima. No such relationshipis apparent for the southern out-of-Africa route (across the Red Sea). Instead, we present a novel interpretation ofcombined sea-level and regional climate control on potential migrations via the southern route, with “windows ofopportunity” at 458–448, 345–340, 272–265, 145–140, and 70–65 ka. The 145–140 ka window seems relevant forearly colonization of Arabia at 127 � 16 ka, and the 70–65 ka window agrees with estimates of 65 �5/�8 ka forthe final out-of-Africa migration by the anatomically modern human founder group of all non-Africans. Once theyreached Eurasian Mediterranean margins, populations benefited from a rich diversity of terrain and microclimates,with persistent favorable conditions in lowlands and potential to occupy higher elevations during milder periods.

Introduction

The Mediterranean–Red Sea region occupies a zone that isinfluenced by four major climate systems. From northwest tosoutheast, these are (1) the temperate westerlies that affectEurope, the western Mediterranean, and the northern sectorof the eastern Mediterranean; (2) the dry subtropical con-ditions that dominate the southern and eastern sectors of theMediterranean basin as well as the entire Red Sea region; (3)the African monsoon, which affects Mediterranean conditionsthrough inflow of major rivers (Nile, and in the past otherNorth African drainage systems) and that has caused past

Eelco J. Rohling is Professor, Katharine M. Grant is PostdoctoralResearcher, and Andrew P. Roberts is Professor at the ResearchSchool of Earth Sciences of the Australian National University(Canberra, Australian Capital Territory 0200, Australia[eelco.rohling@anu.edu.au; katharine.grant@anu.edu.au; andrew.roberts@anu.edu.au]). Juan-Cruz Larrasoana is Staff Scientist at theInstituto Geologico y Minero de Espana of Unidad de Zaragoza(Zaragoza 50006, Spain [jc.larra@igme.es]). This paper wassubmitted 3 VII 13, accepted 4 IX 13, and electronically published18 XII 13.

contractions and expansions of the Sahara desert; and (4) theIndian Ocean monsoon, which causes seasonal wind reversalsover the southern Red Sea region up to latitudes of 20�–25�N.No significant rainfall is associated with the Indian Oceanmonsoon over the Red Sea and Arabian Peninsula today, butthe monsoon’s summer rainfall domain may have shifted ontothe southeast margin of the Arabian Peninsula (i.e., Yemenand Oman) during insolation-driven monsoon maxima suchas those of the Early-Middle Holocene (Conroy and Overpeck2011).

Climatic gradients over the Mediterranean and Red Searegion are well illustrated by the updated Koppen-Geiger cli-mate classification (Kottek et al. 2006; fig. 1). Important mea-sures of climate variability through time are obtained from avariety of methods, such as pollen data, lake levels, stableisotopes from a variety of sedimentary archives, faunalchanges, and others. These data especially supply informationabout changes in regional temperature and precipitation re-gimes. In addition, the Red Sea has in the past decade becomea key region for reconstruction of continuous records of sea-level fluctuations (Grant et al. 2012; Rohling et al. 1998a,2008a, 2008b, 2009b, 2010; Siddall et al. 2003, 2004). Finally,the presence of vast deserts to the south of the Mediterranean

S184 Current Anthropology Volume 54, Supplement 8, December 2013

Figure 1. Summary map of the main climate zones in the study region using the Koppen-Geiger climate classification. A pequatorial; B p arid; C p warm temperate; D p snow; s p summer dry; f p fully humid; S p steppe; W p desert; w p winterdry; m p monsoonal; a p hot summer; b p warm summer; k p cold arid; h p hot arid. Modified after Kottek et al. (2006).A color version of this figure is available in the online edition of Current Anthropology.

and around the Red Sea gives rise to large windblown dustfluxes into these basins, and reconstructions of those fluxesthrough time—notably using marine sediment cores—alsoreveal important changes in regional climate conditions.

In this paper, we present an overview of the modern climateof the region and of the main changes that have been recon-structed from paleoclimate proxy records for the last 500,000yr. This overview provides a context for consideration of therole of climate variability in anthropological and archaeolog-ical developments at the interface of Africa and Europe.

Modern Climatic Setting

Mediterranean

We summarize modern Mediterranean climate conditions fol-lowing the recent review of Rohling et al. (2009a). The clas-sical Mediterranean climate is characterized by warm and drysummers and mild and wet winters. Mean annual precipi-tation along the Mediterranean ranges from less than 0.12 min North Africa, to over 2.00 m in portions of southwestTurkey and in the eastern Adriatic Sea along the slopes of theDinaric Alps (Naval Oceanography Command 1987). Total

evaporation in the entire Mediterranean increases toward theeast, with an average of 1.45 m y�1 (Malanotte-Rizzoli andBergamasco 1991) to 1.57 m y�1 (Bethoux and Gentili 1994).

The classical Mediterranean climate is a result of the re-gion’s location on the transition between temperate westerliesthat dominate over central and northern Europe and the sub-tropical high-pressure belt over North Africa (fig. 2; Boucher1975; Lolis, Bartzokas, and Katsoulis 2002). In summer, sub-tropical high-pressure conditions (and drought) extend fromthe southeast in a northwestward direction over most of theMediterranean. Polar-front depressions may still reach thewestern Mediterranean, but they only exceptionally penetratethe eastern Mediterranean (Rohling and Hilgen 1991). Duringwinter, the subtropical conditions shift southward, and thenorthern sector of the Mediterranean becomes influenced bythe temperate westerlies with associated Atlantic depressionsthat track eastward over Europe. These depression influencesextend from the Mediterranean southeastward across the Le-vant and into the northernmost sector of the Red Sea (e.g.,Arz et al. 2003a; Bar-Matthews et al. 2003; Goodfriend 1991;Matthews, Ayalon, and Bar-Matthews 2000; McGarry et al.2004; Trommer et al. 2010).

Rohling et al. Mediterranean and Red Sea Paleoclimate S185

Figure 2. Atmospheric circulation pattern during Northern Hemisphere summer. The main winds are indicated by arrows.ITCZ p intertropical convergence zone; H p areas of high sea-level pressure; L p areas of low sea-level pressure. After an adaptationin Rohling et al. (2009), which compiled information from Rossignol-Strick (1985) and Reichart (1997).

Polar and continental air masses over Europe are channeledinto the Mediterranean basin through gaps in the mountain-ous topography of the northern Mediterranean margin. Dur-ing winter and spring, intense cold and dry air flows throughthe lower Rhone Valley to reach the Gulf of Lions (the “mis-tral”), and similar flows extend over the Adriatic and AegeanSeas (the “Bora” and “Vardar”), where they cause strong evap-oration and sea surface cooling (e.g., Casford et al. 2003;Leaman and Schott 1991; Maheras et al. 1999; Poulos, Drak-opoulos, and Collins 1997; Saaroni et al. 1996; and referencestherein). The northerly air flows into the western and easternMediterranean are determined by interaction between an in-tense low over the central or eastern Mediterranean andnortheastward extension of the Azores High (over Iberia,France, and southern Britain) or westward ridging of theSiberian High toward northwestern Europe and southernScandinavia (Lolis, Bartzokas, and Katsoulis 2002; Maheraset al. 1999). Persistent winter low-pressure conditions overthe region result from high Mediterranean sea surface tem-peratures (Lolis, Bartzokas, and Katsoulis 2002).

The most pronounced basin-wide cold winter events com-plement cold conditions over Europe and develop in asso-ciation with positive sea-level pressure anomalies to the westor northwest of the British Isles and particularly low pressureover the Mediterranean. An important aspect of winter var-iability concerns cyclogenesis (formation of new depressions),which governs precipitation in the northeastern and south-central sectors of the Mediterranean. Some Atlantic depres-sions may enter the (western) basin, but most cyclones ob-served in the Mediterranean form over the basin itself(Rumney 1968; Trigo, Davies, and Bigg 1999), when cold andrelatively dry northerly air flows extend over warm sea sur-

faces in the northern sectors of the basin. Thus, winter cy-clones are linked to North Atlantic systems, given that theyrepresent either (occasional) direct entries of Atlantic synopticsystems into the Mediterranean basin or secondary lowsformed when Atlantic systems interact with the Alps and leadto cyclogenesis within the basin (Trigo, Davies, and Bigg2000).

Over the Mediterranean Sea, cyclogenesis is most frequentover the Gulf of Genoa and the Ligurian Sea, but the AegeanSea is also a major center for winter cyclogenesis (Boucher1975; Cantu 1977; Rumney 1968; Trewartha 1966; Trigo, Da-vies, and Bigg 1999). Most Genoan depressions track overItaly, thereby affecting the Adriatic region, and thence in agenerally eastward direction toward the Aegean Sea and/ornorthern Levantine seas (Lolis, Bartzokas, and Katsoulis 2002;Rumney 1968; Trewartha 1966; Trigo, Davies, and Bigg 1999).These depressions, along with those that develop over othercenters of cyclogenesis, cause the winter precipitation that ischaracteristic of modern Mediterranean climate. Geologicalarchives indicate that Mediterranean depressions have con-trolled Mediterranean climate in the Levant as an enduringfeature over glacial-interglacial timescales (Bar-Matthewset al. 2003; Goodfriend 1991; Matthews, Ayalon, and Bar-Matthews 2000; McGarry et al. 2004; Rohling 2013; and ref-erences therein).

Summer rainfall is low around the Mediterranean region,especially in eastern and southeastern sectors. Some cyclo-genesis occurs around Cyprus and the Middle East in summer,but adiabatic descent in the upper troposphere—related tothe Asian summer monsoon—precludes deep convection overthe region and so causes the prevalence of dry summer con-

S186 Current Anthropology Volume 54, Supplement 8, December 2013

ditions (Rodwell and Hoskins 1996; Trigo, Davies, and Bigg1999).

The African monsoon does not reach directly into the Med-iterranean basin, and there is no evidence that it ever didduring the Quaternary. It does, however, have (or more ap-propriately used to have) a “remote” influence on the basinthrough Nile River discharge. Before the anthropogenic con-trol of the Nile, its average discharge was 8.4 # 1010 m3 yr�1

(4.5 # 1010 m3 yr�1 in low-flood years to 15.0 # 1010 m3

yr�1 in high-flood years), which from the mid-1960s has dwin-dled to nearly nothing (Bethoux 1984; Nof 1979; Rohling andBryden 1992; Said 1981; Wahby and Bishara 1981). Duringthe predamming instrumental era, a strong (threefold) inter-annual variability has been noted between high and low dis-charge years, mainly due to variability in the monsoon-fedcontribution of the Blue Nile and Atbara rivers (see datasummary in Rohling et al. 2009a).

The Nile River comprises two different systems: the WhiteNile, which drains the equatorial uplands of Uganda in aregular, permanent manner; and the Blue Nile and Atbara,which drain highly seasonal (summer) African monsoon pre-cipitation from the Ethiopian highlands. Summarizing thepredamming Nile hydrology after Adamson et al. (1980) andWilliams et al. (2000), it appears that up to 30% of the annualdischarge of the Nile originated from the White Nile and aminimum of 70% of the annual discharge from the Blue Nile/Atbara. The winter flow was dominated (83%) by the steadyWhite Nile contribution, and the Blue Nile/Atbara compo-nent provided 90% of the summer flow (with a peak overAugust–October). The White Nile discharge has a muchsmaller ratio of change between its annual peak and lowestmonthly value, with a maximum between late September andJanuary.

It should be noted that the Nile has not always been theonly route for drainage of African monsoon precipitation intothe Mediterranean. During past monsoon maxima (related toorbitally induced insolation maxima), the extent of the Saharadesert was much reduced, and there was northward routingof drainage from the central Saharan watershed into the Med-iterranean along the wider North African margin (see belowfor details).

Red Sea

The Red Sea basin is entirely situated in an arid zone withvery low humidity. Coastal stations record annual rainfallfigures of less than 20 mm in the north and 50–100 mm inthe south (Pedgley 1974). Riverine flow into the basin is neg-ligible because of the basin’s small watershed (Maillard andSoliman 1986; Morcos 1970; Siddall et al. 2004; fig. 3). Oneof the larger systems that drains into the Red Sea is the Baraka(Tokar) wadi in Sudan, which today is active 40–70 d yr�1

(mainly during autumn). Wadi Baraka discharges 200–970#106 m3 water at 18.5�N into the Red Sea (Trommer etal. 2011; Whiteman 1971), which is equivalent to a maximum

of only 2 mm y�1 when distributed over the entire Red Seasurface area. Thus, river inflow and precipitation are negligiblethroughout the Red Sea region, especially when contrastedwith the high rates of evaporation, which seasonally reach 2m y�1 or more (Maillard and Soliman 1986; Morcos 1970;Pedgley 1974; Privett 1959; see also Fenton et al. 2000; Siddallet al. 2004). Evaporation over the Red Sea is seasonally af-fected by atmospheric circulation changes that are related tothe Indian Ocean monsoon. The most notable feature is aseasonal wind reversal over the southern Red Sea, the generalnature of which is discussed below. For more details andmodeling, see Jiang et al. (2009).

During winter (October–May), reduction of sensible heatradiation occurs over the cold landmass of central Asia thatis enhanced by snow-induced high albedo (reflection of in-solation) over the Tibetan Plateau and Himalayas. This resultsin a quasi-stable high-pressure system that extends from Mon-golia to central Europe, Turkey, and Arabia. The cold, de-scending air leads to a radial outflow of cold dry air towardlow-pressure areas over the relatively warm Indian Ocean.The intertropical convergence zone (ITCZ) is displaced south-ward in winter, to 20�S over East Africa, and the winter north-east monsoon blows across the Arabian Sea and the Gulf ofAden toward central Africa (Morcos 1970). A general south-easterly wind circulation results between a continental de-pression over central Africa and a continental anticyclone thatextends from Asia to Arabia. Channeling by the rift geometryof the Red Sea coast then causes strong (6.7–9.3 m s�1) windsto blow from the south or south-southeast over the southernhalf of the Red Sea up to about 20�N (e.g., Morcos 1970;Patzert 1974). Throughout October–December, there is a con-vergence between these south-southeasterlies and the north-northwesterly winds that prevail year-round over the northernRed Sea. This convergence occupies a north-south zone acrossthe Red Sea at around 20�N (Jiang et al. 2009; Pedgley 1974),which varies in size and is characterized by low-pressure calms(Morcos 1970). Jiang et al. (2009) demonstrated that the po-sition of the convergence zone may be determined by gapsin the topography, notably the Tokar Gap some 50 km inlandfrom the Tokar Delta on the Sudan coast, which acts as anoutlet of surface wind from the Red Sea basin in winter (andas an inlet in summer; see below).

During summer (June–September), the monsoon systemis reversed. The spring melt reduces albedo over central Asia,and insolation increases, thereby causing the landmass towarm. This warms the air above, causing it to rise, creatinga quasi-stable low-pressure system over northern India (!995mb; Morcos 1970), which extends over Pakistan to the PersianGulf. The updraft causes humid, relatively cool, maritime airto flow in from the area of higher pressure over the nowrelatively cool Indian Ocean. Latent heat release from con-densation/precipitation against the flanks of the Himalayasfuels further deepening of the continental low-pressure celland consequently more inflow of maritime air, which rein-forces the system. The ITCZ reaches its most northerly po-

Rohling et al. Mediterranean and Red Sea Paleoclimate S187

Figure 3. Topography of the Red Sea basin. The dashed line delineates the Red Sea watershed. Numbers refer to key sediment corelocations (not used here). After Siddall et al. (2004). A color version of this figure is available in the online edition of CurrentAnthropology.

sition (20�N) in July and becomes identified with the frontof the southwest monsoon. This passes north of Aden, andthe southwest monsoon flows in a clockwise direction overEast Africa, the Gulf of Aden, and the Arabian Sea towardthe main monsoon low of northern India (Morcos 1970). Atthis time, a general northwesterly circulation is set up on thewestern side of the summer Asiatic low-pressure cell. Thiscauses relatively weak (2.4–4.4 m s�1) north-northwesterly ornorthwesterly winds to dominate over the entire length of theRed Sea (e.g., Morcos 1970; Patzert 1974).

Jiang et al. (2009) analyzed regional patterns superimposedon the general along-axis wind systems, which are more zonal(west to east or east to west) in nature and that are relatedto gaps in the topography along the basin, such as the TokarGap. There are eastward-blowing wind jets in summer(mainly through the Tokar Gap) and westward-blowing windjets in winter from the Saudi Arabian margin (mainly overthe northern Red Sea). Wind speeds in these surface jets canreach 10–15 m s�1, and Jiang et al. (2009) provide clear evi-dence of atmospheric dust entrainment over the Red Sea.Jiang et al. (2009:5) further note that “other strong zonalwinds . . . blow from the Egyptian coast eastward across the

Red Sea longitudinal axis. . . . This . . . can also drive duststorms.”

In the northernmost Red Sea, records of past climatechange have detected an influence of southeast Mediterraneanclimate influences that extend across the Middle East (Arz etal. 2003a; Legge, Mutterlose, and Arz 2006; Trommer et al.2010). Contemporary climatology suggests that this link func-tions through winter frontal rainfall associated with Cypruslows (El-Fandy 1946; Morcos 1970).

Past Climate Variability

Glacial-Interglacial Changes

Throughout the last three million years, climate variabilityover the entire study region has been dominated by the effectsof global ice-age cycles, which were particularly prominentduring the last 500,000 yr (e.g., Lisiecki and Raymo 2005),with variability that is forced by cyclic changes in the earth-sun orbital configuration. The astronomical forcing of climatetakes place because of three main processes, namely: changesin the eccentricity of the earth’s orbit around the sun, with

S188 Current Anthropology Volume 54, Supplement 8, December 2013

Figure 4. Comparison between three approaches for reconstructing continuous records of sea-level variations. Black is Waelbroecket al. (2002), based on coral-calibrated deep-sea stable oxygen isotope data, with uncertainties in gray. The dashed line is thereconstruction of De Boer et al. (2010, 2011) from a model-based deconvolution of deep-sea stable oxygen isotope records into atemperature and a sea-level component following the method of Bintanja, van de Wal, and Oerlemans (2005). Dots with (2σ)uncertainty bars represent the Red Sea reconstruction on a U-Th-adjusted chronology (Rohling et al. 2009b, 2010). Modified afterRohling et al. (2012). A color version of this figure is available in the online edition of Current Anthropology.

climate effects in approximately 100- and 400-kyr cycles;changes in the tilt of the earth’s rotational axis, with effectson climate in 41-kyr cycles; and precession of the equinoxes,with climate effects in cycles of 19- and 23-kyr durations (e.g.,Berger 1977; Hays, Imbrie, and Shackleton 1976; Imbrie andImbrie 1980, 1986; Milankovitch 1941).

Glaciations were strongly focused on the North Americanand northern Eurasian locations of continental ice sheets, butthe effects on climate were global. This had global implicationsfor sea level, with fluctuations between interglacial highstandsup to �10 m (Kopp et al. 2009; Muhs et al. 2011; Rohlinget al. 2008b, 2009b) and glacial lowstands at �120 m or lowerrelative to present sea level (Rohling et al. 1998a; Waelbroecket al. 2002). Methods for constructing continuous records ofsea-level variability now exist for the last half-million years(Rohling et al. 2009b, 2010; Siddall et al. 2003; Waelbroecket al. 2002). Model-based deconvolutions of deep-sea stableoxygen isotope records have also been used to extend therecord over many millions of years (Bintanja, van de Wal,and Oerlemans 2005; De Boer et al. 2010, 2011), but thesestill require independent validation (fig. 4).

Along with enlarged land-ice volumes, another defin-ing characteristic of glacials was widespread cooling (e.g.,CLIMAP Project Members 1976; MARGO Project Members2009; Rohling et al. 2012; Schneider von Deimling et al. 2006).In the study region, cooling produced a strong gradient fromwest to east through the Mediterranean and then to the south-east along the axis of the Red Sea. Relative to the present,glacial sea surface temperatures in summer were reduced by8� � 2�C in the westernmost Mediterranean, 5� � 2�Caround Sicily, and 3� � 2�C in the easternmost Mediterra-nean; winter sea surface temperatures were reduced by 5� �

2�C, 3� � 1�C, and 0�–2.5�C, respectively (Hayes et al. 2005).

In the Red Sea, glacial-interglacial differences in sea surfacetemperature were of the order of 3�–4�C (Arz et al. 2003b;Trommer et al. 2011), with a bias toward summer. Theseestimates for the Red Sea are in agreement with estimates forthe western Arabian Sea (see compilation of Rohling et al.2012).

Terrestrial temperature contrasts between glacials and in-terglacials are typically around 1.3 or 1.5 times larger thansea surface temperature contrasts (Braconnot et al. 2007;Laıne et al. 2009). This agrees with estimates by Kuhlemannet al. (2008) of terrestrial temperature reductions of 12�C inthe northwest Mediterranean region, 7.5�C around Sicily, and6�C or less to the south and east of Crete. Thus, mean glacial-interglacial temperature contrasts are expected on landaround the Red Sea. The inferred summer bias in Arz et al.(2003b) and Trommer et al. (2011) suggests that the 4�–6�Ccontrast estimated here may also have some bias toward sum-mer (i.e., winter contrasts may have been smaller).

A notable southward displacement of climate zones alsooccurred in glacials on both orbital and suborbital timescales(e.g., Wang et al. 2004, 2006), as well as global lowering ofthe snow line and associated vertical compression of vege-tation zones (e.g., Barmawidjaja et al. 1993; Broecker andDenton 1989; Klein, Isacks, and Bloom 1995; Kuhlemann etal. 2008). Such snow line and vegetation displacements wereparticularly strongly expressed in the mountain ranges alongthe northern margin of the Mediterranean (Kuhlemann et al.2008).

In a general sense, summer circulation/monsoons wereweakened in the Northern Hemisphere during glacial periods,and winter circulation/monsoons were strengthened (e.g.,Rohling, Mayewski, and Challenor 2003; Rohling et al. 2009cand references therein).

Rohling et al. Mediterranean and Red Sea Paleoclimate S189

Figure 5. Major global dust sources and locations of dust records. Dust flux contours (mg m�2 yr�1) are shown in oceans surroundingdust sources (after Duce et al. 1991). Locations of dust records discussed are indicated (ocean drilling program sites 659: Tiedemann,Sarntheim, and Shackleton 1994; 721: deMenocal, Bloemendal, and King 1991; 967: Larrasoana et al. 2003; and the Zhaojiachuanand Lingtai loess sections: Sun et al. 2006). Site KL09 is the Red Sea record presented in Rohling et al. (2008a, 2009b) and Robertset al. (2011). Modified after Roberts et al. (2011). A color version of this figure is available in the online edition of CurrentAnthropology.

Overall, atmospheric dust transport was strongly intensifiedduring glacials (e.g., Lambert et al. 2008; Larrasoana et al.2003; Mayewski et al. 1997; Roberts et al. 2011; Rohling,Mayewski, and Challenor 2003; Ruth et al. 2007; Trauth, Lar-rasoana, and Mudelsee 2009; Winckler et al. 2008), whichattests to increased aridity, stronger winds, and reduced veg-etation cover (reduced soil cohesion). However, even withinthe Mediterranean and Red Sea region, there can be consid-erable spatial differences in dust flux histories because of spa-tially different conditions in the various dust source areas.The western Mediterranean receives dust from northwest Af-rica/western Sahara, while the eastern Mediterranean receivesdust from the eastern Sahara (Libya, Egypt). The Red Seareceives influxes of windblown dust from the easternmostSahara, Sudan, and Saudi Arabia (Hickey and Goudie 2007;Jiang et al. 2009; Middleton and Goudie 2001; fig. 5). Onmillennial timescales, dust variability from various source ar-eas varies considerably within and around the Mediterraneanand Red Seas; these variations do not seem to be systematicbetween the two basins (Roberts et al. 2011; fig. 6). Thissuggests considerable regional differences in the temporal var-iability of vegetation cover, soil cohesion, and wind patterns/intensities.

Ice-core records from Greenland and Antarctica reveal thatglacial periods were characterized by strong temperature fluc-tuations on millennial timescales (fig. 7). Antarctic ice-corerecords reveal climate variability that was less abrupt and of

smaller amplitude than that observed in Greenland (Blunieret al. 1998; EPICA Community Members 2006). Continuousrecords of sea-level variability, developed from Red Sea oxygenisotope records, indicate that within the last glacial cycle,global ice volume fluctuated on millennial timescales with arhythm close to that of variability observed in Antarctic cli-mate records (Grant et al. 2012; Rohling et al. 2004a, 2009b;Siddall et al. 2003, 2008).

Greenland ice-core records and North Atlantic marine sed-iment records provide evidence of particularly strong climatefluctuations that have become known as Dansgaard-Oeschgercycles, which include the particularly cold Heinrich events(e.g., Broecker 2000; Dansgaard et al. 1993; Grootes et al.1993; Hemming 2004). In many records, these millennial-scale fluctuations appear as an alternation between more in-tense and less intense glacial conditions, and the Heinrichevents are often particularly cold/intense. Western Mediter-ranean sea surface temperature strongly fluctuated in closeagreement with the Dansgaard-Oeschger cycles and Heinrichevents (Cacho et al. 1999, 2000, 2001; Frigola et al. 2007;Martrat et al. 2004; Rohling et al. 1998b).

Dansgaard-Oeschger oscillations in the Northern Hemi-sphere were related to Antarctic (Southern Hemisphere) tem-perature cycles through a systematic out-of-phase relationship(e.g., Blunier and Brook 2001; Blunier et al. 1998; EPICACommunity Members 2006; Stocker and Johnsen 2003). Inthis relationship, which has become known as the “bipolar

Figure 6. Comparison of sea-level and dust records for the Red Sea, circum-Sahara region, and Chinese loess plateau. a, CentralRed Sea sea-level reconstruction (Rohling et al. 2009b, 2010). b, Environmental magnetic IRM900 mT@AF120 mT proxy forwindblown hematite (Hem.) in Red Sea core KL09 (gray) compared with the stacked Chinese loess grain size record from Zhaojiachuan

Rohling et al. Mediterranean and Red Sea Paleoclimate S191

and Lingtai, Chinese loess plateau (from Sun et al. 2006, with minor age adjustments in Roberts et al. 2011). c, d, e, and f, respectively,compare the Red Sea dust record (gray) with dust records from EPICA Dome C, Antarctica (Lambert et al. 2008); ODP Site 659,off northwest Africa (Tiedemann, Sarntheim, and Shackleton 1994); ODP Site 967, eastern Mediterranean Sea (Larrasoana et al.2003); and ODP Site 721, Arabian Sea (deMenocal, Bloemendal, and King 1991). Vertical dashed lines coincide with Red Sea dustpeaks at the glacial terminations and are shown to assist comparisons between panels. The Red Sea records are shown for consistencyon the same chronology (Rohling et al. 2009b, 2010) as used in Roberts et al. (2011). This chronology is (subtly) updated withinthe last 150,000 yr in figures 7 and 8 based on the latest age controls developed by Grant et al. (2012). A color version of thisfigure is available in the online edition of Current Anthropology.

temperature seesaw,” the magnitude of warming in the South-ern Hemisphere is proportional to the duration of cold ep-isodes in the Northern Hemisphere (e.g., EPICA CommunityMembers 2006; Stocker and Johnsen 2003). Siddall et al.(2010) identified during which intervals of the last 500,000yr such millennial-scale climate variability has been particu-larly pronounced and found that this was the case at 480–460, 440–400, 380–360, 340–320, 260–220, 200–160, 140–120,and 80–40 ka (fig. 7). Because of the global nature of thisvariability, it may be expected that these intervals of generallyenhanced climate variability will be noticeable in records fromAfrica and Eurasia even if the exact nature of the variabilitymay differ between regions.

Northern high-latitude cooling events particularly affectedthe northern sectors of the Mediterranean region because ofcold air outbreaks that were channeled toward the basinthrough gaps in the mountain ranges along its northern limits,which also triggered atmospheric instability over the Medi-terranean and had implications for regional precipitation re-gimes (e.g., Casford et al. 2003; Frigola et al. 2007; Kuhlemannet al. 2008; Rohling et al. 1998b, 2002b). Thus, vegetation(pollen) records from the western Mediterranean and north-ern sector of the eastern Mediterranean also reflect the strongeffects of northerly (Greenland-style) climate influences (e.g.,Allen et al. 1999; Kotthoff et al. 2008; Moreno et al. 2002;Muller and Pross 2007; Sanchez-Goni et al. 2002; Tzedakis1999, 2009; Tzedakis et al. 2004).

From the Red Sea area, relatively little is known aboutclimate variability. It was always dry, with high windblowndust input, but this input nevertheless reveals intensity var-iations that are closely similar to those observed on the Chi-nese loess plateau (Roberts et al. 2011). This suggests that themain climate influences over the central Red Sea (with respectto windblown dust input) were dominated by atmosphericcirculation/wind changes that reflect the larger westerlies-dominated Northern Hemisphere climate variability andIndian-Asian monsoon variability (e.g., Porter and An 1995;Rohling, Mayewski, and Challenor 2003).

African Monsoon Changes

Superimposed on the glacial cycles, the entire Mediterraneanregion is strongly affected by monsoon intensity variations,which are dominated by Northern Hemisphere insolationchanges that mainly reflect the influences of precession and

eccentricity (e.g., COHMAP Members 1988; Kutzbach andGallimore 1988; Kutzbach and Guetter 1986; Kutzbach andStreet-Perrott 1985; Rossignol-Strick 1985). African monsoonmaxima are well known to have been associated with inso-lation maxima. Freshwater flooding from the African margininto the Mediterranean during these times caused collapse ofdeepwater formation, eventually resulting in deepwater an-oxia. These events are easily recognized in eastern Mediter-ranean sedimentary sequences because organic carbon pres-ervation led to the formation of characteristic black organicrich layers, called sapropels (e.g., Rohling 1994 and referencestherein).

The Mediterranean sapropel record reflects the more in-tense African monsoon maxima, and individual sapropelshave been dated (Emeis et al. 2000; Hilgen 1991; Hilgen etal. 1993; Kroon et al. 1998; Lourens, Wehausen, and Brumsack2001; Lourens et al. 1996) based on the ages of precession-driven insolation maxima that are known from astronomicalsolutions (Berger 1977; Laskar et al. 2004; Milankovitch 1941).A particularly straightforward summary table of ages is givenfor the main sapropels by Kroon et al. (1998), after Lourenset al. (1996). These datings may be important for archaeo-logical, anthropological, and biogeographical studies becausethey provide a chronological framework that may help us todate and understand, for example, past migration pulsesthrough the Sahara desert region (e.g., Drake et al. 2011;Osborne et al. 2008) or past humid phases associated withfossil hominid finds (e.g., Brown, McDougall, and Fleagle2012; McDougall, Brown, and Fleagle 2005, 2008). The sap-ropel record indicates that there have been more than a hun-dred of such strongly developed African monsoon maximaover the past couple of million years (Larrasoana, Roberts,and Rohling 2013).

During times with intensified African monsoon circulation,the spatial extent of the Sahara Desert was much reduced,which has become known as “greening of the Sahara,” whenthe African monsoon penetrated farther northward than to-day (Larrasoana, Roberts, and Rohling 2013). This penetra-tion was partly due to orbital forcing and partly resulted fromnorthward expansion of vegetation into the previously morereflective desert, which in turn triggered further northwardpenetration of the monsoon front in a “vegetation-albedofeedback process” (e.g., Brovkin et al. 1998; Claussen et al.1998; Foley et al. 2003). During intense monsoon maxima,

S192 Current Anthropology Volume 54, Supplement 8, December 2013

Figure 7. Comparison of different records of climate variability including polar ice-core records, sea level, and Mediterraneantemperature variability. a, EPICA Dome C ice-core stable hydrogen isotope proxy for temperature, Antarctica (Jouzel et al. 2007).b, The Red Sea sea-level record in the pre-150 ka interval on the chronology of Rohling et al. (2009b, 2010) and in the post-150ka interval on the latest chronology using the age constraints developed by Grant et al. (2012). c, Sea surface temperature in theAlboran Sea, westernmost Mediterranean (Martrat et al. 2004). d, Stacked North-GRIP ice-core oxygen isotope proxy for temperature,Greenland (gray; Wolff et al. 2010), and—for clarity—a 21-point moving average smoothing (black). Bars at the bottom indicateintervals of particularly pronounced millennial-scale climate variability after the analysis of Siddall et al. (2010). A color version ofthis figure is available in the online edition of Current Anthropology.

the monsoon front appears to have penetrated northward pastthe central Saharan watershed (at about 21�N), and seasonalrunoff occurred from the central Saharan mountains into theeastern Mediterranean along the wider North African margin(Drake et al. 2011; Osborne et al. 2008; Paillou et al. 2009;Rohling et al. 2002a, 2004a, 2004b). Such drainages will havepresented green corridors along which humans and animalsmay have migrated across the otherwise arid region (Drakeet al. 2011; Osborne et al. 2008). These ideas integrate a widevariety of evidence for past increases in precipitation andwater availability throughout the currently hyperarid Sahara,including the distributions of animal and human fossils, sed-imentological data, lake-level reconstructions, etc. (e.g., Ar-mitage et al. 2007; Cremaschi 2002; Drake et al. 2011; Gasse2000; Gaven et al. 1981; Kuper and Kropelin 2006; Larrasoana,Roberts, and Rohling 2013; Mandell and Simmons 2001; Pa-chur 2001; Pachur and Altmann 2006; Pachur and Braun1980; Szabo, Haynes, and Maxwell 1995). Records of Africanmonsoon variability reveal that centennial- to millennial-scale

decreases/collapses in monsoon intensity occurred during sev-eral monsoon maxima and that these roughly coincided withintrusions of northerly cooling events into the Mediterraneanbasin (e.g., Casford et al. 2003; Gasse 2000; Mercone et al.2001; Osborne et al. 2008; Rohling et al. 2002a, 2002b, 2004b;Scrivner, Vance, and Rohling 2004).

It is important to note that there is no evidence that theAfrican monsoon penetrated at any time directly into theMediterranean basin. Rainfall gradient and isotope recon-structions in the Levant indicate that rain in that region wasalways sourced from the north and west, from the Mediter-ranean, even during monsoon maxima (Bar-Matthews et al.2003; Goodfriend 1991; Matthews, Ayalon, and Bar-Matthews2000; McGarry et al. 2004; Vaks et al. 2007). Moreover, Tze-dakis (2009) presented a compelling case that times ofinsolation-driven African monsoon maxima were not char-acterized by enhanced summer precipitation around thenorthern and eastern Mediterranean, as had been often sug-gested before (e.g., Rohling and Hilgen 1991; Tzedakis 2009;

Rohling et al. Mediterranean and Red Sea Paleoclimate S193

and references therein). Pollen data indicate that summeraridity was enhanced at these times and that any enhancedrainfall likely took place in winter (with some possible regionalexceptions); that is, typical Mediterranean climate conditionswere intensified (Tzedakis 2009).

During opposite phases of the precession cycle, the mon-soon was weak, and the (hyper)arid Sahara desert was spatiallyextended, similar to the present, without watercoursescrossing it to the wider North African margin. Orbital cyclicityin the African monsoon is obvious in records of windblowndust fluxes from the Sahara (e.g., Larrasoana et al. 2003; Lour-ens, Wehausen, and Brumsack 2001; Trauth, Larrasoana,Mudelsee 2009; Wehausen and Brumsack 2000), which re-flects insolation forcing of the African monsoon.

Indian Monsoon Changes

Variations in the Indian Ocean monsoon circulation, whichaffect the wind field over the southern sector of the Red Searegion, have been documented especially by marine sedi-mentary records from the Arabian Sea (e.g., Almogi-Labin etal. 2000; Clemens and Prell 1990, 2003; Clemens et al. 1991;Ivanochko et al. 2005; Prell and Kutzbach 1987; Reichart,Lourens, Zachariasse 1998; Rostek et al. 1997; Schmiedl andLeuschner 2005; Schultz, von Rad, and Erlenkeuser 1998; Sir-ocko et al. 1993) and by speleothem records from Oman andSocotra (Burns et al. 2003, 2004; Fleitmann et al. 2003a,2003b, 2004, 2007). Long Arabian Sea records provide con-vincing evidence for a predominant control of orbital pre-cession and thus insolation on the Indian monsoon intensity,similar to the African monsoon. Speleothem data from Oman(Fleitmann et al. 2003a, 2003b, 2004, 2007) indicate that theprecipitation regime of the Indian Ocean monsoon expandedto affect the southeastern margin of the Arabian Peninsuladuring the Early to Middle Holocene summer southwestMonsoon maximum, whereas that region currently (duringthe opposite precession phase) is unaffected (Conroy andOverpeck 2011).

Trommer et al. (2011) described the timing of a somewhatmore humid interval in the central Red Sea Bakala (Tokar)wadi catchment and found that this started immediately fol-lowing the last interglacial sea-level highstand and lasted sev-eral thousand years. This relative timing agrees with the periodof deposition of the deep-sea anoxic event known as sapropelS5 in the eastern Mediterranean (Grant et al. 2012), which isalso recognized as a humid interval in the speleothem recordof Soreq cave, Israel, which reflects the last interglacial Africanmonsoon maximum (Bar-Matthews, Ayalon, and Kaufman1997, 2000; Bar-Matthews et al. 1999, 2003). U-series datingof the Soreq cave record demonstrates that the last interglacialAfrican monsoon maximum dates to 128–120 ka. This agreeswith datings from Dongge cave, China, for the last interglacialmaximum of the Asian monsoon, from 129.3 � 0.9 to 119.6� 0.6 ka (Yuan et al. 2004). On millennial scales, therefore,the timing of monsoon maxima seems to be roughly similar

for all three major monsoon systems (African, Indian, andSoutheast Asian).

Red Sea records contain no evidence for any major pre-cipitation/vegetation changes associated with monsoon max-ima; the area appears to have remained (hyper)arid. Monsoonvariability, however, affected Red Sea oceanography throughchanges in the wind field over the basin (Biton et al. 2010;Trommer et al. 2011), and strong windblown dust variationsover time support the notion of important fluctuations inwind forcing over the basin (Roberts et al. 2011; Rohling etal. 2008a).

Millennial-scale variability in windblown dust records ofthe central Red Sea is coherent with millennial-scale changesin Arabian Sea productivity (Rohling et al. 2008a; Schultz,von Rad, Erlenkeuser 1998) as well as with Chinese loessrecords (Porter and An 1995; Roberts et al. 2011; Sun et al.2006). This links Arabia with winter-dominated climate var-iability as recorded in Greenland ice-core records (Rohling,Mayewski, and Challenor 2003). In short, there is good evi-dence that colder conditions in Greenland coincided withintensified winter-type atmospheric circulation over Asia,which also affected Arabia, possibly through the winter(northeast) monsoon.

Implications and Conclusions

Conditions Relevant to Proposed Migration Routesacross the Southern Red Sea

Sea level. The intense ice-age cycles of the last 500,000 yrhave been associated with important variability in global sealevel over a range of 120 m or more below the present toperhaps 10 m above the present (fig. 4). The Strait of Bab-el-Mandab in the southern Red Sea, which connects the basinwith the open ocean, is highly sensitive to sea-level changebecause it is (today) only 137 m deep. This is of the sameorder as past sea-level drops during glacial maxima, and therehave consequently been many proposals of a potential mi-gration route across the southern Red Sea, assuming that apassage between Africa and Arabia may have emerged duringtimes of maximum glacial sea-level lowstands.

Fernandes, Rohling, and Siddall (2006) evaluated the con-cept of emergence of a southern land bridge between Africaand Arabia and concluded that there is no evidence for it atany time during the last 500,000 yr. Emergence of a landbridge in the strait would rapidly lead to desiccation withinthe highly evaporative Red Sea, with deposition of evaporiteswithin a matter of several centuries in the open basin andwithin a matter of years to decades in shallow coastal envi-ronments. Moreover, there is evidence of substantial persistentlocal water depths above the sill in the strait of at least 15 m(Fernandes, Rohling, and Siddall 2006) and up to 35 m (Biton,Gildor, and Peltier 2008); for the Last Glacial Maximum, thepassage depth has been estimated as up to 25 � 4 m (Lambecket al. 2011). For lower water depths, inflow from the open

S194 Current Anthropology Volume 54, Supplement 8, December 2013

ocean would have become sufficiently restricted for basin-wide development of extreme salinities in excess of about 75practical salinity units (psu). There is evidence that salinitiesat times rose above 49 psu, causing local extinction of plank-tonic foraminifera (Fenton et al. 2000 and references therein).However, salinities remained below 75 psu given that highervalues would have also caused the local extinction of all pter-opods and all benthic foraminifera, which did not happen(Fenton et al. 2000; Rohling et al. 1998a).

Regardless of the above, the strait passage was much shal-lower and narrower than today during glacial sea-level low-stands (Lambeck et al. 2011; Rohling et al. 1998a; Siddall etal. 2004). This is especially clearly illustrated by a recent ad-vanced reconstruction of strait morphology during the LastGlacial Maximum (Lambeck et al. 2011), which takes intoaccount detailed hydrographic data for the present-day strait,sea-level change during the Last Glacial Maximum, and amodel for isostatic change components. Lambeck et al. (2011)also present arguments about older glacial lowstands and sup-port the notion that an open passage remained in existence.In summary, it is evident from the combined studies that anymigration across the southern Red Sea during glacial sea-levellowstands would have (a) benefited from the fact that themarine passage was strongly reduced in width, and (b) def-initely included some element of swimming, rafting, or nav-igation.

Climate conditions. Another important control on potentialmigrations across the southern Red Sea concerns regionalclimatic conditions. For example, was enough water and foodavailable to sustain migrating animals/humans on either sideof the strait? This question has come to the fore because ofa recent proposal (Armitage et al. 2011) that an early waveof human migration out of Africa occurred across the south-ern Red Sea, with migration toward and across the straitregion during the penultimate glacial maximum (the “Saa-lian”) and subsequent spreading across/along the ArabianPeninsula during the monsoon maximum of the last inter-glacial (about 128–120 ka).

Red Sea sedimentary records provide important insightsinto regional climatic conditions. Windblown dust concen-trations have been measured in exactly the same sedimentarysequences that were used to reconstruct the Red Sea sea-levelrecord. This dust record clearly illustrates that the transitionbetween the Saalian glacial and the last interglacial was char-acterized by extreme fluxes of windblown dust, which reflecthigh winds and pronounced aridity in the region (Roberts etal. 2011). Organic geochemical data, also from the same sam-ple sequence, demonstrate that (lightly) enhanced humidityin the Red Sea region—the most likely local expression of thelast interglacial summer monsoon maximum—first developedonly after the sea-level highstand had peaked and sea levelhad started to drop again (Trommer et al. 2011). Data fromOman suggest that relatively more humid conditions mayhave started to develop earlier, from about 135 ka (Fleitmann

et al. 2003b; Vaks et al. 2007), a little later than in the easternSahara (from 140 ka; Osmond and Dabous 2004; Szabo,Haynes, and Maxwell 1995; Vaks et al. 2007).

It would be an oversimplification to assume a simple suc-cession from low sea level to favorable climate for migrationacross the Arabian Peninsula during the transition from theSaalian glacial to the last interglacial. The data instead revealconsiderable complexity across that transition, highlightingthat the sea-level lowstand and the climatically more favorableconditions were separated by at least 5,000 yr of arid regionalconditions and sharp sea-level rise (Roberts et al. 2011). Thisdoes not exclude the possibility that the southern route outof Africa was employed at this time. But if it was, then theanimals/humans involved must have been resilient to signif-icantly adverse environmental conditions. We note that a sim-ilar sequence of events is observed for almost all glacial ter-minations. In the next section, we elaborate a new view ofmore promising migration intervals.

A new concept: “windows of opportunity” for southern mi-gration out of Africa. Our sea-level and windblown dust rec-ords, sampled from the same central Red Sea sediment se-quence (Rohling et al. 2008a, 2009b, 2010; Roberts et al. 2011)are shown in figure 8. The last 150,000 yr are shown on anew chronology (Grant et al. 2012, which is tightly con-strained relative to the U-Th dated Soreq Cave record fromIsrael; Bar-Matthews, Ayalon, and Kaufman 1997; Bar-Matthews et al. 1999, 2000, 2003). Before 150 ka, the recordsare shown on a chronology developed (Rohling et al. 2009b)by correlation with the European Project for Ice Coring inAntarctica (EPICA) Dome C temperature proxy record onthe EDC3 chronology (Jouzel et al. 2007) with an adjustmentthat accounts for radiometric datings of past sea-level high-stands (Rohling et al. 2010).

The horizontal light gray bar in figure 8 highlights relativelylow dust fluxes similar to those of the Holocene. During timeswith such relatively low dust fluxes, we suggest that the re-gional climate may have been most favorable for habitation/migration in contrast to the intervening intervals with highdust fluxes, which attest to more intense winds and lower soilcohesion due to even lower soil moisture and even morescarce vegetation cover than today. The vertical light gray barsidentify periods when both (a) sea level stood 100 m or morebelow the present, so that the marine connection to be crossedin the Strait of Bab el Mandab would have been only some6 km wide (cf. Last Glacial Maximum reconstruction ofLambeck et al. 2011); and (b) dust fluxes were relatively low.These highlighted intervals therefore represent periods withrelatively favorable conditions (windows of opportunity) forpotential migrations between northeast Africa and southwestArabia via a southern route. There is increasing evidence thatthe southern route was used for the final migration out ofAfrica by AMHs (e.g., Derricourt 2005; Fernandes et al. 2012).

The highlighted periods of significant sea-level lowstandswere also characterized by emerged continental shelves

Rohling et al. Mediterranean and Red Sea Paleoclimate S195

Figure 8. Direct comparison between (a) the latest version of the Red Sea sea-level reconstruction, with the pre-150 ka interval onthe chronology of Rohling et al. (2009b, 2010) and the post-150 ka interval on the new chronology of Grant et al. (2012), and (b)central Red Sea dust proxy data, including the Ti/Ca ratio from core-scanning XRF analysis (gray) and hematite concentration datafrom environmental magnetic analyses (smooth line; see also Roberts et al. 2011; Rohling et al. 2008a). Data originate from a singlesampling of the same sedimentary sequence, which ensures unambiguous phase relationships between the various records. Lightgray bars are explained in the text. Horizontal bars at bottom indicate intervals of particularly pronounced millennial-scale climatevariability, after the analysis of Siddall et al. (2010; see also fig. 7). A color version of this figure is available in the online editionof Current Anthropology.

around the Arabian Peninsula. These provided excellent hab-itation and migration potential for animals and humans, es-pecially if the enhanced hydraulic head due to sea-level low-ering led to enhanced freshwater seepage on the emergedshelves from groundwater reservoirs (e.g., Parker and Rose2008). Such seepage exists even today on the shelves, in sub-marine form (Ghoneim 2008; Parker and Rose 2008). Theemerged shelves may therefore have presented an excellenthabitat with gentle topography, freshwater availability, and anabundance of coastal/marine resources. If emerged shelveswere key migration/habitation zones, then much of the an-thropological record may now be under water.

Several windows of opportunity for migration along thesouthern route out of Africa within the last half-million yearsare highlighted in figure 8. These windows date to about 458–448, 345–340, 272–265, 145–140, and 70–65 ka. The intervalof 145–140 ka may be relevant with respect to a tentative earlymigration of AMHs that led to early inhabitation at the Hor-muz region of the Arabian Gulf with an oldest dating of 127� 16 ka (Armitage et al. 2011). That population is thoughtto have been “unsuccessful” in that it did not leave any de-scendants (Fernandes et al. 2012). Instead, all non-Africanhumans originate from a more recent migration out of Africadated with various molecular clock approaches at around 57–65 ka, with an upper age bound of 70–65 ka based on EastAfrican data (Fernandes et al. 2012). This age range of 65

�5/�8 ka is in remarkable agreement with our window ofopportunity of 70–65 ka for a southern route migration outof Africa. The chronological uncertainties for our records inthat interval are less than 1,000 yr (Grant et al. 2012). Hap-lotype dispersal maps suggest that the migration at around65 ka followed the southern route out of Africa and thatArabia was the first staging post in the spread of AMHsaround the world (Fernandes et al. 2012).

Conditions Relevant to Habitation of and Migrationthrough the Sahara

Migration potential through the Sahara Desert was compre-hensively assessed by Drake et al. (2011). Their abstract saysit all:

Both animals and humans populated it [the Sahara] during

past humid phases. . . . More animals crossed via this route

than used the Nile corridor . . . [and] many of these species

are aquatic. This dispersal was possible because during the

Holocene humid period the region contained a series of

linked lakes, rivers, and inland deltas comprising a large

interlinked waterway, channeling water and animals into

and across the Sahara. . . . This system was last active in

the early Holocene when many species appear to have oc-

cupied the entire Sahara. Human dispersals were influenced

by this distribution. . . . Lacustrine sediments show that the

S196 Current Anthropology Volume 54, Supplement 8, December 2013

“green Sahara” also existed during the last interglacial (∼125

ka) and provided green corridors that could have formed

dispersal routes at a likely time for the migration of modern

humans out of Africa.

Both periods highlighted in that study are well-knowninsolation-driven African monsoon maxima. The Holoceneperiod is known to correspond to eastern Mediterranean sap-ropel S1 and the last interglacial period to sapropel S5 (e.g.,Larrasoana, Roberts, and Rohling 2013; Osborne et al. 2008;Rohling and Hilgen 1991; Rohling et al. 2002a, 2004b; Scriv-ner, Vance, and Rohling 2004).

Archaeological observations around exclusively rain-fed de-pressions on the Libyan Plateau suggest that monsoonal sum-mer rains from central Africa periodically penetrated at leastas far north as Kharga (roughly 25�N) during the Holocenemonsoon maximum despite the fact that conditions duringthat pluvial phase seem to have remained drier than duringearlier Quaternary pluvial phases (Mandel and Simmons2001). This suggestion that the Holocene monsoon maximumwas of a relatively low intensity compared with previous Qua-ternary monsoon maxima has been corroborated by quan-titative reconstructions of effects of the Holocene and lastinterglacial monsoon maxima in the eastern Mediterranean(Rohling 1999; Rohling et al. 2004b). Hence, past monsoonmaxima—datings for which can be obtained from the astro-nomical ages of insolation-driven monsoon maxima (Emeiset al. 2000; Hilgen 1991; Hilgen et al. 1993; Kroon et al. 1998;Lourens et al. 1996, 2001)—represent times of enhanced hu-midity that may have been crucial for migrations through theotherwise hyperarid Sahara “barrier” between sub-SaharanAfrica and the Mediterranean/Levantine regions (Larrasoana,Roberts, and Rohling 2013).

As mentioned above, several African monsoon maxima(e.g., the Holocene and last interglacial) have been interruptedby centennial- to millennial-scale periods of reduced mon-soon intensity. Hence, potential routes for migration throughthe Sahara region remained intermittent, subject to periodicreturns of harsh conditions.

Finally, hypotheses that invoke an importance of periodsof rapid, millennial-scale climate variability for developmentsand/or migrations of hominins and their food sources—forexample, through intermittent habitat fragmentation—maybenefit from the objectively identified episodes of generally(globally) enhanced millennial-scale climate variability (Sid-dall et al. 2010). We have indicated these intervals (480–460,440–400, 380–360, 340–320, 260–220, 200–160, 140–120, and80–40 ka) in figures 7 and 8.

Conditions Relevant to Migrations through the Levant

Conditions for migrations through the Levant are normallypoor because of hyperarid conditions in the Sinai-Negev re-gion. However, data from a north–south array of caves high-light a window of time between 140 and 110 ka when this

region was intermittently more humid (especially between 133and 122 ka) and thus more hospitable (Vaks et al. 2007). Thisclosely matches data that indicate more humid conditions inthe Egyptian Sahara (Osmond and Dabous 2004), which sug-gests a high probability that a pathway existed through theLevant for migrations out of Africa between about 140 and110 ka, with possible extension to 85 ka, after which hyperaridconditions were reestablished (Vaks et al. 2007).

Derricourt (2005) reviewed climatic and archaeologicalstudies to conclude that the Levantine route was the mostlikely route of early (pre-85 ka) migrations out of Africa incontrast to the youngest (post-85 ka) migration, which helinks more to a southern route across the Bab-el-MandabStrait (see above). Early AMH finds from Qafzeh and Skhulin the Near East date to between 119 � 18 and 81 � 13 ka(e.g., Armitage et al. 2011; Grun et al. 2005; Petraglia 2011;Shea 2008). Genetic data indicate that this migration did notleave any descendants in the modern human population out-side Africa, which instead can be traced back to a migrationout of Africa that took place around 65 ka, likely via a south-ern route (Fernandes et al. 2012; see above).

For Levantine migration routes, the key control may havebeen exerted by development of pluvial episodes as recordedand accurately dated in cave speleothem deposits (e.g., Der-ricourt 2005; Shea 2008; Vaks et al. 2007). These intermittentlymore humid conditions in the Levant are closely related tomonsoon maxima associated with Northern Hemisphere in-solation maxima. It should be emphasized that monsoonmaxima, with “greening of the Sahara,” were not limited toonly interglacial insolation maxima but occurred also duringinsolation maxima within glacial times (e.g., Larrasoana, Rob-erts, and Rohling 2013; Larrasoana et al. 2003; Liu et al. 2012).

The control by pluvials that applies to the Levantine routeis considerably more straightforward than the windows ofopportunity control we have proposed in “A new concept:‘windows of opportunity’ for southern migration out of Af-rica” for the southern route, which requires a combinationof low sea level and concomitant interludes of relatively fa-vorable climate conditions. These windows of opportunityare not simply aligned with monsoon maxima but insteadreflect millennial-scale episodes of relatively favorable climatewithin glacial maxima.

Conditions along the European Margin of the Mediterranean

As recorded in Greenland ice cores, the European margin ofthe Mediterranean has been strongly affected by intense cli-mate swings that originate in the North Atlantic region. Coldevents were transmitted to the Mediterranean by northerlyair outbreaks through gaps in the mountainous topographyaround the northern margin of the Mediterranean, especiallyover the northwest Mediterranean, the Adriatic Sea, and theAegean (e.g., Casford et al. 2003; Frigola et al. 2007; Kuhle-mann et al. 2008; Moreno et al. 2002; Rohling et al. 1998b,2002b). During the Holocene, a strong temporal coincidence

Rohling et al. Mediterranean and Red Sea Paleoclimate S197

has been found between the northerly cooling events andarchaeological transitions in the northeastern sector of theMediterranean region (Clare et al. 2008). Because of resultantatmospheric instabilities within the Mediterranean basin,northerly cooling events may have had an amplified effect onsnow line lowerings and vegetation-zone migrations in theMediterranean region (e.g., Kuhlemann et al. 2008 and ref-erences therein).

Despite this glacial variability, the Mediterranean regionduring glacial times was a refugium with relatively mild con-ditions for plants and animals, avoiding the much more severeclimatic stress over the main Eurasian landmass to the northof the Alps/Carpathians (e.g., Blondel 2009; Tzedakis 2009).The region’s great diversity of terrain and microclimates,drainage patterns, and accentuated relief with many rock shel-ters and caves would have been conducive to a relatively richvariety of exploitable plant and animal food sources and tocamp locations. From a climatic point of view, therefore, itis likely that habitation and migration were possible through-out this region even during severe glacials, especially in low-lands and coastal plains, and that there was the potential torange into higher elevations during periods with milder con-ditions.

Acknowledgments

Thanks to the Wenner-Gren Foundation for organizing theconference and arranging for publication of the papers pre-sented there and to Erella Hovers and Steve Kuhn and twoanonymous reviewers for their constructive assessments ofearlier versions of the manuscript. This paper contributes tothe United Kingdom Natural Environment Research Councilprojects NE/E01531X/1 (RESET), NE/I009906/1 (iGlass), andNE/H004424/1 as well as E. J. Rohling’s Royal Society–Wolfson Research Merit Award and Australian Laureate Fel-lowship FL120100050.

References CitedAdamson, D. A., F. Gasse, F. A. Street, and M. A. J. Williams. 1980. Late

Quaternary history of the Nile. Nature 288:50–55.Allen, J. R. M., U. Brandt, A. Brauer, H. W. Hubberten, B. Huntley, J. Keller,

M. Kraml, et al. 1999. Rapid environmental changes in southern Europeduring the last glacial period. Nature 400:740–743.

Almogi-Labin, A., G. Schmiedl, C. Hemleben, R. Siman-Tov, M. Segl, and D.Meischner. 2000. The influence of the NE winter monsoon on productivitychanges in the Gulf of Aden, NW Arabian Sea, during the last 530 ka asrecorded by benthic foraminifera. Marine Micropaleontology 40:295–319.

Armitage, S. J., N. A. Drake, S. Stokes, A. El-Hawat, M. Salem, K. White, P.Turner, and S. J. McLaren. 2007. Multiple phases of North African humidityrecorded in lacustrine sediments from the Fazzan Basin, Libyan Sahara.Quaternary Geochronology 2:181–186.

Armitage, S. J., S. A. Jasim, A. E. Marks, A. G. Parker, V. I. Usik, and H. P.Uerpmann. 2011. The southern route “Out of Africa”: evidence for an earlyexpansion of modern humans into Arabia. Science 331:453–456.

Arz, H. W., F. Lamy, J. Patzold, P. J. Muller, and M. Prins. 2003a. Mediter-ranean moisture source for an Early-Holocene humid period in theNorthern Red Sea. Science 300:118–121.

Arz, H. W., J. Patzold, P. J. Muller, and M. O. Moammar. 2003b. Influenceof Northern Hemisphere climate and global sea level rise on the restricted

Red Sea marine environment during termination I. Paleoceanography 18:1053, doi:10.1029/2002PA000864.

Bar-Matthews, M., A. Ayalon, M. Gilmour, A. Matthews, and C. J. Hawkes-worth. 2003. Sea-land oxygen isotopic relationships from planktonic fo-raminifera and speleothems in the Eastern Mediterranean region and theirimplication for paleorainfall during interglacial intervals. Geochimica et Cos-mochimica Acta 67:3181–3199.

Bar-Matthews, M., A. Ayalon, and A. Kaufman. 1997. Late Quaternary pa-leoclimate in the eastern Mediterranean region from stable isotope analysisof speleothems at Soreq Cave, Israel. Quaternary Research 47:155–168.

———. 2000. Timing and hydrological conditions of sapropel events in theEastern Mediterranean, as evident from speleothems, Soreq Cave, Israel.Chemical Geology 169:145–156.

Bar-Matthews, M., A. Ayalon, A. Kaufman, and G. J. Wasserburg. 1999. TheEastern Mediterranean paleoclimate as a reflection of regional events: SoreqCave, Israel. Earth and Planetary Science Letters 166:85–95.

Barmawidjaja, D. M., E. J. Rohling, W. A. Van der Kaars, C. Vergnaud-Graz-zini, and W. J. Zachariasse. 1993. Glacial conditions in the northern MoluccaSea region (Indonesia). Palaeogeography, Palaeoclimatology, Palaeoecology101:147–167.

Berger, A. 1977. Support for the astronomical theory of climatic change.Nature 268:44–45.

Bethoux, J. P. 1984. Paleo-hydrologie de la Mer Mediterranee au cours desderniers 20 000 ans. Oceanologica Acta 7:43–48.

Bethoux, J. P., and B. Gentili. 1994. The Mediterranean Sea: a test area formarine and climatic interactions. In Ocean processes in climate dynamics:global and Mediterranean examples. P. Malanotte-Rizzoli and A. R. Rob-inson, eds. Pp. 239–254. Dordrecht: Kluwer.

Bintanja, R., R. S. W. van de Wal, and J. Oerlemans. 2005. Modelled atmo-spheric temperatures and global sea levels over the past million years. Nature437:125–128.

Biton, E., H. Gildor, and W. R. Peltier. 2008. Relative sea level reduction atthe Red Sea during the Last Glacial Maximum. Paleoceanography 23:PA1214,doi:10.1029/2007PA001431.

Biton, E., H. Gildor, G. Trommer, M. Siccha, M. Kucera, M. T. J. van derMeer, and S. Schouten. 2010. Sensitivity of Red Sea circulation to monsoonalvariability during the Holocene: an integrated data and modeling study.Paleoceanography 25:PA4209, doi:10.1029/2009PA001876.

Blondel, J. 2009. The nature and origin of the vertebrate fauna. In The physicalgeography of the Mediterranean. J. C. Woodward, ed. Pp. 139–163. Oxford:Oxford University Press.

Blunier, T., and E. Brook. 2001. Timing of millennial-scale climate change inAntarctica and Greenland during the last glacial period. Science 291:109–112.

Blunier, T., J. Chapellaz, J. Schwander, A. Dallenbach, B. Stauffer, T. F. Stocker,D. Raynaud, et al. 1998. Asynchrony of antarctic and Greenland climatechange during the last glacial period. Nature 394:739–743.

Boucher, K. 1975. Global climate. London: English University Press.Braconnot, P., B. Otto-Bliesner, S. Harrison, S. Joussaume, J. Y. Peterchmitt,

A. Abe-Ouchi, M. Crucifix, et al. 2007. Results of PMIP2 coupled simu-lations of the mid-Holocene and Last Glacial Maximum. 1. Experimentsand large-scale features. Climate of the Past 3:261–277.

Broecker, W. S. 2000. Abrupt climate change: causal constraints provided bythe paleoclimate record. Earth-Science Reviews 51:137–154.

Broecker, W. S., and G. H. Denton. 1989. The role of ocean-atmospherereorganizations in glacial cycles. Geochimica et Cosmochimica Acta 53:2465–2501.

Brovkin, V., M. Claussen, V. Petoukhov, and A. Ganopolski. 1998. On thestability of the atmosphere-vegetation system in the Sahara/Sahel region.Journal of Geophysical Research 103:31613–31624.

Brown, F. H., I. McDougall, and J. G. Fleagle. 2012. Correlation of the KHSTuff of the Kibish Formation to volcanic ash layers at other sites, and theage of early Homo sapiens (Omo I and Omo II). Journal of Human Evolution63:577–585.

Burns, S. J., D. Fleitmann, A. Matter, J. Kramers, and A. A. Al-Subbary. 2003.Indian Ocean climate and an absolute chronology over Dansgaard/Oeschgerevents 9 to 13. Science 301:1365–1367.

———. 2004. Indian Ocean climate and an absolute chronology over Dans-gaard/Oeschger events 9 to 13: correction. Science 305:1567.

Cacho, I., J. O. Grimalt, M. Canals, L. Sbaffi, N. J. Shackleton, J. Schonfeld,and R. Zahn. 2001. Variability of the western Mediterranean Sea surfacetemperatures during the last 25,000 years and its connection with theNorthern Hemisphere climatic changes. Paleoceanography 16:40–52.

S198 Current Anthropology Volume 54, Supplement 8, December 2013

Cacho, I., J. O. Grimalt, C. Pelejero, M. Canals, F. J. Sierro, J. A. Flores, andN. J. Shackleton. 1999. Dansgaard-Oeschger and Heinrich event imprintsin Alboran Sea paleotemperatures. Paleoceanography 14:698–705.

Cacho, I., J. O. Grimalt, F. J. Sierro, N. J. Shackleton, and M. Canals. 2000.Evidence of enhanced Mediterranean thermohaline circulation during rapidclimate coolings. Earth and Planetary Science Letters 183:417–429.

Cantu, V. 1977. The climate of Italy. In Climates of central and southern Europe.C. C. Wallen, ed. Pp. 127–183. World Survey of Climatology, vol. 6. Am-sterdam: Elsevier.

Casford, J. S. L., E. J. Rohling, R. H. Abu-Zied, F. J. Jorissen, M. Leng, andJ. Thomson. 2003. A dynamic concept for eastern Mediterranean circulationand oxygenation during sapropel formation. Palaeogeography, Palaeocli-matology, Palaeoecology 190:103–119.

Clare, L., E. J. Rohling, B. Weninger, and J. Hilpert. 2008. Warfare in LateNeolithic/Early Chalcolithic Pisidia, southwestern Turkey: climate inducedsocial unrest in the late 7th millennium cal BC. Documenta Praehistorica35:65–92.

Claussen, M., V. Brovkin, A. Ganopolski, C. Kubatski, and V. Petoukhov.1998. Modelling global terrestrial vegetation-climate interaction. Philosoph-ical Transactions of the Royal Society B 353:53–63.

Clemens, S. C., and W. L. Prell. 1990. Late Pleistocene variability of ArabianSea summer monsoon winds and continental aridity: eolian records fromthe lithogenic component of deep-sea sediments. Paleoceanography 5:109–145.

———. 2003. A 350,000 year summer-monsoon multi-proxy stack from theOwen Ridge, northern Arabian Sea. Marine Geology 201:35–51.

Clemens, S. C., W. L. Prell, D. Murray, G. Shimmield, and G. Weedon. 1991.Forcing mechanisms of the Indian Ocean monsoon. Nature 353:720–725.

CLIMAP Project Members. 1976. The surface of the ice-age earth. Science191:1131–1137.

COHMAP Members. 1988. Climatic changes of the last 18,000 years: obser-vations and model simulations. Science 241:1043–1052.

Conroy, J. L., and J. T. Overpeck. 2011. Regionalization of present-day pre-cipitation in the greater monsoon region of Asia. Journal of Climate 24:4073–4095.

Cremaschi, M. 2002. Late Pleistocene and Holocene climatic changes in thecentral Sahara: the case study of the south-western Fezzan, Libya. InDroughts, food and culture: ecological change and food security in Africa’slater prehistory. F. A. Hassan, ed. Pp. 65–81. New York: Kluwer/Plenum.

Dansgaard, W., S. J. Johnsen, H. B. Clausen, D. Dahl-Jensen, N. S. Gundestrup,C. U. Hammer, C. S. Hvidberg, et al. 1993. Evidence for general instabilityof past climate from a 250 kyr ice core. Nature 364:218–219.

De Boer, B., R. S. W. van de Wal, R. Bintanja, L. J. Lourens, and E. Tuenter.2010. Cenozoic global ice-volume and temperature simulations with 1-Dice-sheet models forced by benthic δ18O records. Annals of Glaciology 51:23–33.

De Boer, B., R. S. W. van de Wal, L. J. Lourens, and R. Bintanja. 2011. Transientnature of the earth’s climate and the implications for the interpretation ofbenthic records. Palaeogeography, Palaeoclimatology, Palaeoecology 335/336:4–11.

deMenocal, P., J. Bloemendal, and J. King. 1991. A rock-magnetic record ofmonsoonal dust deposition to the Arabian Sea: evidence for a shift in themode of deposition at 2.4 Ma. In Proceedings of the Ocean Drilling Program,Scientific Results, vol. 117. Pp. 389–407. College Station, TX: Ocean DrillingProgram.

Derricourt, R. 2005. Getting “out of Africa”: sea crossings, land crossings andculture in the hominin migrations. Journal of World Prehistory 19:119–132.

Drake, N. A., R. M. Blench, S. J. Armitage, C. S. Bristow, and K. H. White.2011. Ancient watercourses and biogeography of the Sahara explain thepeopling of the desert. Proceedings of the National Academy of Sciences ofthe USA 108:458–462.

Duce, R. A., P. S. Liss, J. T. Merrill, E. L. Atlas, P. Buat-Menard, B. B. Hicks,J. M. Miller, et al. 1991. The atmospheric input of trace species to the worldocean. Global Biogeochemical Cycles 5:193–259.

El-Fandy, M. G. 1946. Barometric lows of Cyprus. Quarterly Journal of theRoyal Meteorological Society 72:291–306.

Emeis, K. C., T. Sakamoto, R. Wehausen, and H. J. Brumsack. 2000. Thesapropel record of the eastern Mediterranean Sea: results of Ocean DrillingProgram Leg 160. Palaeogeography, Palaeoclimatology, Palaeoecology 158:371–395.

EPICA (European Project for Ice Coring in Antarctica) Community Members.2006. One-to-one coupling of glacial climate variability in Greenland andAntarctica. Nature 444:195–197.

Fenton, M., S. Geiselhart, E. J. Rohling, and C. Hemleben. 2000. Aplanktoniczones in the Red Sea. Marine Micropaleontology 40:277–294.

Fernandes, C., E. J. Rohling, and M. Siddall. 2006. Absence of QuaternaryRed Sea land bridges: biogeographic implications. Journal of Biogeography33:961–966.

Fernandes, V., F. Alshamali, M. Alves, M. D. Costa, J. B. Pereira, N. M. Silva,L. Cherni, et al. 2012. The Arabian cradle: mitochondrial relicts of the firststeps along the southern route out of Africa. American Journal of HumanGenetics 90:1–9.

Fleitmann, D., S. J. Burns, A. Mangini, M. Mudelsee, J. Kramers, I. Villa, U.Neff, et al. 2007. Holocene ITCZ and Indian monsoon dynamics recordedin stalagmites from Oman and Yemen (Socotra). Quaternary Science Reviews26:170–188.

Fleitmann, D., S. J. Burns, M. Mudelsee, U. Neff, J. Kramers, A. Mangini,and A. Matter. 2003a. Holocene forcing of the Indian monsoon recordedin a stalagmite from Southern Oman. Science 300:1737–1739.

Fleitmann, D., S. J. Burns, U. Neff, A. Mangini, and A. Matter. 2003b. Chang-ing moisture sources over the last 330,000 years in Northern Oman fromfluid-inclusion evidence in speleothems. Quaternary Research 60:223–232.

Fleitmann, D., S. J. Burns, U. Neff, M. Mudelsee, A. Mangini, and A. Matter.2004. Palaeoclimatic interpretation of high-resolution oxygen isotope pro-files derived from annually laminated speleothems from Southern Oman.Quaternary Science Reviews 23:935–945.

Foley, J. A., M. T. Coe, M. Scheffer, and G. Wang. 2003. Regime shifts in theSahara and Sahel: interactions between ecological and climatic systems inNorthern Africa. Ecosystems 6:524–539.

Frigola, J., A. Moreno, I. Cacho, M. Canals, F. J. Sierro, J. A. Flores, J. O.Grimalt, D. A. Hodell, and J. H. Curtis. 2007. Holocene climate variabilityin the western Mediterranean region from a deepwater sediment record.Paleoceanography 22:PA2209, doi:10.1029/2006PA001307.

Gasse, F. 2000. Hydrological changes in the African tropics since the LastGlacial Maximum. Quaternary Science Reviews 19:189–211.

Gaven, C., C. Hillaire-Marcel, and N. Petit-Maire. 1981. A Pleistocene lacus-trine episode in southeastern Libya. Nature 290:131–133.

Ghoneim, E. 2008. Optimum groundwater locations in the northern UnitedArab Emirates. International Journal of Remote Sensing 29:5879–5906.

Goodfriend, G. A. 1991. Holocene trends in 18O in land snail shells from theNegev Desert and their implications for changes in rainfall source areas.Quaternary Research 35:417–426.

Grant, K. M., E. J. Rohling, M. Bar-Matthews, A. Ayalon, M. Medina-Elizalde,C. Bronk Ramsey, C. Satow, and A. P. Roberts. 2012. Rapid coupling be-tween ice volume and polar temperature over the past 150 kyr. Nature 491:744–747.

Grootes, P. M., M. Stuiver, J. W. C. White, S. Johnsen, and J. Jouzel. 1993.Comparison of oxygen isotope records from the GISP2 and GRIP Greenlandice cores. Nature 366:552–554.

Grun, R., C. Stringer, F. McDermott, R. Nathan, N. Porat, S. Robertson, L.Taylor, G. Mortimer, S. Eggins, and M. McCulloch. 2005. U-series and ESRanalyses of bones and teeth relating to the human burials from Skhul.Journal of Human Evolution 49:316–334.

Hayes, A., M. Kucera, N. Kallel, L. Sbaffi, and E. J. Rohling. 2005. GlacialMediterranean sea surface temperatures reconstructed from planktonic fo-raminiferal assemblages. Quaternary Science Reviews 24:999–1016.

Hays, J. D., J. Imbrie, and N. J. Shackleton. 1976. Variations in the earth’sorbit: pacemaker of the ice ages. Science 194:1121–1132.

Hemming, S. R. 2004. Heinrich events: massive Late Pleistocene detritus layersof the North Atlantic and their global imprint. Reviews of Geophysics 42:1–43.

Hickey, B., and A. S. Goudie. 2007. The use of TOMS and MODIS to identifydust storm source areas: the Tokar delta (Sudan) and the Seistan basin(south west Asia). In Geomorphological variations. A. S. Goudie, and J.Kalvoda, J., eds. Pp. 37–57. Prague: P3K.

Hilgen, F. J. 1991. Astronomical calibration of Gauss to Matuyama sapropelsin the Mediterranean and implication for the geomagnetic polarity timescale. Earth and Planetary Science Letters 104:226�244.

Hilgen, F. J., L. J. Lourens, A. Berger, and M. F. Loutre. 1993. Evaluation ofthe astronomically calibrated time-scale for the late Pliocene and earliestPleistocene. Paleoceanography 8:549�565.

Imbrie, J., and J. Z. Imbrie. 1980. Modeling the climatic response to orbitalvariations. Science 207:943–953.

Imbrie, J., and K. P Imbrie. 1986. Ice ages: solving the mystery. Cambridge,MA: Harvard University Press.

Ivanochko, T. S., R. S. Ganeshram, G. J. A. Brummer, G. Ganssen, S. J. A.

Rohling et al. Mediterranean and Red Sea Paleoclimate S199

Jung, S. G. Moreton, and D. Kroon. 2005. Variations in tropical convectionas an amplifier of global climate change at the millennial scale. Earth andPlanetary Science Letters 235:302–314.

Jiang, H., J. T. Farrar, R. C. Beardsley, R. Chen, and C. Chen. 2009. Zonalsurface wind jets across the Red Sea due to mountain gap forcing alongboth sides of the Red Sea. Geophysical Research Letters 36:L19605, doi:10.1029/2009GL040008.

Jouzel, J., V. Masson-Delmotte, O. Cattani, G. Dreyfus, S. Falourd, G. Hoff-mann, B. Minster, et al. 2007. Orbital and millennial Antarctic climatevariability over the past 800,000 years. Science 317:793–797.

Klein, A. G., B. Isacks, and A. L. Bloom. 1995. Modern and Last GlacialMaximum snowline in Peru and Bolivia: implications for regional climaticchange. Bulletin de l’Institut Francais d’Etudes Andines 24:607–617.

Kopp, R. E., F. J. Simons, J. X. Mitrovica, A. C. Maloof, and M. Oppenheimer.2009. Probabilistic assessment of sea level during the last interglacial stage.Nature 462:863–868.

Kottek, M., J. Grieser, C. Beck, B. Rudolf, and F. Rubel. 2006. World map ofthe Koppen-Geiger climate classification updated. Meteorologische Zeitschrift15:259–263.

Kotthoff, U., U. C. Muller, J. Pross, G. Schmiedl, I. T. Lawson, and H. Schulz.2008. Late glacial and Holocene vegetation dynamics in the Aegean region:an integrated view based on pollen data from marine and terrestrial archives.Holocene 18:1019–1032.

Kroon, D., I. Alexander, M. Little, L. J. Lourens, A. Matthewson, A. H. F.Robertson, and T. Sakamoto. 1998. Oxygen isotope and sapropel stratig-raphy in the Eastern Mediterranean during the last 3.2 million years. InProceedings of the Ocean Drilling Program, Scientific Results, vol. 160. Pp.181–190. College Station, TX: Ocean Drilling Program.

Kuhlemann, J., E. J. Rohling, I. Kumrei, P. Kubik, S. Ivy-Ochs, and M. Kucera.2008. Regional synthesis of Mediterranean atmospheric circulation duringthe Last Glacial Maximum. Science 321:1338–1340.

Kuper, R., and S. Kropelin. 2006. Climate-controlled Holocene occupation inthe Sahara: motor of Africa’s evolution. Science 313:803–807.

Kutzbach, J. E., and R. G. Gallimore. 1988. Sensitivity of a coupled atmo-sphere/mixed layer ocean model to changes in orbital forcing at 9000 years.Journal of Geophysical Research 93:803–821.

Kutzbach, J. E., and P. J. Guetter. 1986. The influence of changing orbitalparameters and surface boundary conditions on climate simulations for thepast 18,000 years. Journal of Atmospheric Science 43:1726–1759.

Kutzbach, J. E., and F. A. Street-Perrott. 1985. Milankovitch forcing in thelevel of tropical lakes from 18 to 0 kyr. Nature 317:130–134.

Laıne, A., M. Kageyama, P. Braconnot, and R. Alkama. 2009. Impact of green-house gas concentration changes on surface energetics in IPSL-CM4: re-gional warming patterns, land-sea warming ratios, and glacial-interglacialdifferences. Journal of Climate 22:4621–4635.

Lambeck, K., A. Purcell, N. C. Fleming, C. Vita-Finzi, A. M. Alsharekh, andG. M. Bailey. 2011. Sea level and shoreline reconstructions for the Red Sea:isostatic and tectonic considerations and implications for hominin migra-tion out of Africa. Quaternary Science Reviews 30:3542–3574.

Lambert, F., B. Delmonte, J. R. Petit, M. Bigler, P. R. Kaufmann, M. A. Hutterli,T. F. Stocker, U. Ruth, J. P. Steffensen, and V. Maggi. 2008. Dust-climatecouplings over the past 800,000 years from the EPICA Dome C ice core.Nature 452:616–619.

Larrasoana, J. C., A. P. Roberts, and E. J. Rohling. 2013. Dynamics of greenSahara periods and their role in hominin evolution. PLoS One 8:e76514,doi:10.1371/journal.pone.0076514.

Larrasoana, J. C., A. P. Roberts, E. J. Rohling, M. Winklhofer, and R. Wehausen.2003. Three million years of monsoon variability over the northern Sahara.Climate Dynamics 21:689–698.

Laskar, J., M. Gastineau, F. Joutel, P. Robutel, B. Levrard, et al. 2004. A long-term numerical solution for the insolation quantities of the Earth. Astron-omy and Astrophysics 428:261–285.

Leaman, K. D., and F. A. Schott. 1991. Hydrographic structure of the con-vection regime in the Gulf of Lions: winter 1987. Journal of Physical Ocean-ography 21:575–598.

Legge, H. L., J. Mutterlose, and H. W. Arz. 2006. Climatic changes in thenorthern Red Sea during the last 22,000 years as recorded by calcareousnannofossils. Paleoceanography 21:PA1003, doi:10.1029/2005PA001142.

Lisiecki, L. E., and M. E. Raymo. 2005. A Plio-Pleistocene stack of 57 globallydistributed benthic δ18O records. Paleoceanography 20:PA1003, doi:10.1029/2004PA001071.

Liu, Q., J. C. Larrasoana, J. Torrent, A. P. Roberts, E. J. Rohling, Z. Liu, andZ. Jiang. 2012. New constraints on climate forcing and variability in the

circum-Mediterranean region from magnetic and geochemical observationsof sapropels S1, S5 and S6. Palaeogeography, Palaeoclimatology, Palaeoecology333/334:1–12.

Lolis, C. J., A. Bartzokas, and B. D. Katsoulis. 2002. Spatial and temporal 850hPa air temperature and sea-surface temperature covariances in the Med-iterranean region and their connection to atmospheric circulation. Inter-national Journal of Climatology 22:663–676.

Lourens, L. J., A. Antonarakou, F. J. Hilgen, A. A. M. van Hoof, C. Vergnaud-Grazzini, and W. J. Zachariasse. 1996. Evaluation of the Plio-Pleistoceneastronomical timescale. Paleoceanography 11:391–413.

Lourens, L. J., R. Wehausen, and H. J. Brumsack. 2001. Geological constraintson tidal dissipation and dynamical ellipticity of the earth over the past threemillion years. Nature 409:1029–1033.

Maheras, P., E. Xoplaki, T. Davies, J. Martin-Vide, M. Bariendos, and M. J.Alcoforado. 1999. Warm and cold monthly anomalies across the Mediter-ranean basin and their relationship with circulation 1860–1990. Interna-tional Journal of Climatology 19:1697–1715.

Maillard, C., and G. Soliman. 1986. Hydrography of the Red Sea and exchangeswith the Indian Ocean in summer. Oceanologica Acta 9:249–269.

Malanotte-Rizzoli, P., and A. Bergamasco. 1991. The wind and thermallydriven circulation of the eastern Mediterranean Sea. 2. The baroclinic case.Dynamics of Atmospheres and Oceans 15:355–419.

Mandel, R. D., and A. H. Simmons. 2001. Prehistoric occupation of LateQuaternary landscapes near Kharga Oasis, western desert of Egypt. Geo-archaeology 16:95–117.

MARGO Project Members. 2009. Constraints on the magnitude and patternsof ocean cooling at the Last Glacial Maximum. Nature Geoscience 2:127–132.

Martrat, B., J. O. Grimalt, C. Lopez-Martinez, I. Cacho, F. J. Sierro, J. A.Flores, R. Zahn, M. Canals, J. H. Curtis, and D. A. Hodell. 2004. Abrupttemperature changes in the western Mediterranean over the past 250,000years. Science 306:1762–1765.

Matthews, A., A. Ayalon, and M. Bar-Matthews. 2000. D/H ratios of fluidinclusions of Soreq Cave (Israel) speleothems as a guide to the EasternMediterranean Meteoric Line relationships in the last 120 ky. ChemicalGeology 166:183–191.

Mayewski, P. A., L. D. Meeker, M. S. Twickler, S. Whitlow, Q. Yang, W. B.Lyons, and M. Prentice. 1997. Major features and forcing of high-latitudeNorthern Hemisphere atmospheric circulation using a 110,000-year-longglaciochemical series. Journal of Geophysical Research 102:26345–26366.

McDougall, I., F. H. Brown, and J. G. Fleagle. 2005. Stratigraphic placementand age of modern humans from Kibish, Ethiopia. Nature 433:733–736.

———. 2008. Sapropels and the age of hominins Omo I and II, Kibish,Ethiopia. Journal of Human Evolution 55:409–420.

McGarry, S., M. Bar-Matthews, A. Matthews, A. Vaks, B. Schilman, and A.Ayalon. 2004. Constraints on hydrological and paleotemperature variationsin the eastern Mediterranean region in the last 140 ka given by the δDvalues of speleothem fluid inclusions. Quaternary Science Reviews 23:919–934.

Mercone, D., J. Thomson, R. H. Abu-Zied, I. W. Croudace, and E. J. Rohling.2001. High-resolution geochemical and micropalaeontological profiling ofthe most recent eastern Mediterranean sapropel. Marine Geology 177:25–44.

Middleton, N. J., and A. S. Goudie. 2001. Saharan dust: sources and trajec-tories. Transactions of the Institute of British Geographers 26:165–181.

Milankovitch, M. 1941. Kanon der Erdbestrahlung und seine Anwendung aufdas Eiszeitenproblem. Royal Serbian Academy Special Publications, vol. 132;Section Mathematics and Natural Sciences, vol. 33. Belgrade: Royal SerbianAcademy.

Morcos, S. A. 1970. Physical and chemical oceanography of the Red Sea.Oceanography and Marine Biology 8:73–202.

Moreno, A., I. Cacho, M. Canals, M. A. Prins, M. F. Sanchez-Goni, J. O.Grimalt, and G. J. Weltje. 2002. Saharan dust transport and high-latitudeglacial climatic variability: the Alboran Sea record. Quaternary Research 58:318–328.

Muhs, D., K. R. Simmons, R. R. Schumann, and R. B. Halley. 2011. Sea-levelhistory of the past two interglacial periods: new evidence from U-seriesdating of reef corals from south Florida. Quaternary Science Reviews 30:570–590.

Muller, U., and J. Pross. 2007. Lesson from the past: present insolation min-imum holds potential for glacial inception. Quaternary Science Reviews 26:3025–3029.

S200 Current Anthropology Volume 54, Supplement 8, December 2013

Naval Oceanography Command. 1987. U.S. Navy climatic study of the Med-iterranean Sea. Ashville, NC: Naval Oceanography Command Detachment.

Nof, D. 1979. On man-induced variations in the circulation of the Mediter-ranean Sea. Tellus 31:558–564.

Osborne, A. H., D. Vance, E. J. Rohling, N. Barton, M. Rogerson, and N.Fello. 2008. A humid corridor across the Sahara for the migration “out ofAfrica” of early modern humans 120,000 years ago. Proceedings of the Na-tional Academy of Sciences of the USA 105:16444–16447.

Osmond, J. K., and A. A. Dabous. 2004. Timing and intensity of groundwatermovement during Egyptian Sahara pluvial periods by U-series analysis ofsecondary U in ores and carbonates. Quaternary Research 61:85–94.

Pachur, H. J. 2001. Holozane Klimawechsel in den nordlichen Subtropen.Nova Acta Leopoldina NF88 331:109–131.

Pachur, H. J., and N. Altmann. 2006. Die Ostsahara im Spatquartar. Berlin:Springer.

Pachur, H. J., and G. Braun. 1980. The paleoclimate of the central Sahara,Libya and the Libyan Desert. Palaeoecology of Africa 12:351–363.

Paillou, P., M. Schuster, S. Tooth, T. Farr, A. Rosenqvist, S. Lopez, and J. M.Malezieux. 2009. Mapping of a major paleodrainage system in eastern Libyausing orbital imaging radar: the Kufrah River. Earth and Planetary ScienceLetters 277:327–333.

Parker, A. G., and J. I. Rose. 2008. Climate change and human origins insoutheastern Arabia. Proceedings of the Seminar on Arabian Studies 38:25–42.

Patzert, W. C. 1974. Wind induced reversal in Red Sea circulation. Deep-SeaResearch 21:109–121.

Pedgley, D. E. 1974. An outline of the weather and climate of the Red Sea.Publications Centre National pour l’Exploitation des Oceans, Actes du Col-loques 2:9–21.

Petraglia, M. D. 2011. Trailblazers across Arabia. Nature 470:50–51.Porter, S. C., and Z. An. 1995. Correlation between climate events in the

North Atlantic and China during the last glaciation. Nature 375:305–308.Poulos, S. E., P. G. Drakopoulos, and M. B. Collins. 1997. Seasonal variability

in sea surface oceanographic conditions in the Aegean Sea (eastern Med-iterranean): an overview. Journal of Marine Systems 13:225–244.

Prell, W. L., and J. E. Kutzbach. 1987. Monsoon variability over the past150,000 years. Journal of Geophysical Research 92:8411–8425.

Privett, D. W. 1959. Monthly charts of evaporation from the North IndianOcean, including the Red Sea and Persian Gulf. Quarterly Journal of theMeteorological Society 85:424–428.

Reichart, G. J. 1997. Late Quaternary variability of the Arabian Sea monsoonand oxygen minimum zone. Geologica Ultraiectina, no. 154. PhD thesis,Universiteit Utrecht.

Reichart, G. J., L. J. Lourens, and W. J. Zachariasse. 1998. Temporal variabilityin the northern Arabian Sea oxygen minimum zone (OMZ) during the last225,000 years. Paleoceanography 13:607–621.

Roberts, A. P., E. J. Rohling, K. M. Grant, J. C. Larrasoana, and Q. Liu. 2011.Atmospheric dust variability from major global source regions over the last500,000 years. Quaternary Science Reviews 30:3537–3541.

Rodwell, M. J., and B. J. Hoskins. 1996. Monsoons and the dynamics of deserts.Quarterly Journal of the Royal Meteorological Society 122:1385–1404.

Rohling, E. J. 1994. Review and new aspects concerning the formation ofMediterranean sapropels Marine Geology 122:1–28.

———. 1999. Environmental controls on salinity and δ18O in the Mediter-ranean. Paleoceanography 14:706–715.

———. 2013. Quantitative assessment of glacial fluctuations in the level ofLake Lisan, Dead Sea rift. Quaternary Science Reviews 70:63–72.

Rohling, E. J., R. Abu-Zied, J. S. L. Casford, A. Hayes, and B. A. A. Hoogakker.2009a. The marine environment: present and past. In The physical geographyof the Mediterranean. J. C. Woodward, ed. Pp. 33–67. Oxford: Oxford Uni-versity Press.

Rohling, E. J., K. Braun, K. Grant, M. Kucera, A. P. Roberts, M. Siddall, andG. Trommer. 2010. Comparison between Holocene and marine isotopestage-11 sea-level histories. Earth and Planetary Science Letters 291:97–105.

Rohling, E. J., and H. L. Bryden. 1992. Man-induced salinity and temperatureincreases in Western Mediterranean Deep Water. Journal of GeophysicalResearch 97:11191–11198.

Rohling, E. J., T. R. Cane, S. Cooke, M. Sprovieri, I. Bouloubassi, K. C. Emeis,R. Schiebel, et al. 2002a. African monsoon variability during the previousinterglacial maximum. Earth and Planetary Science Letters 202:61–75.

Rohling, E. J., M. Fenton, F. J. Jorissen, P. Bertrand, G. Ganssen, and J. P.Caulet. 1998a. Magnitudes of sea-level lowstands of the past 500,000 years.Nature 394:162–165.

Rohling, E. J., K. Grant, M. Bolshaw, A. P. Roberts, M. Siddall, C. Hemleben,

and M. Kucera. 2009b. Antarctic temperature and global sea level closelycoupled over the past five glacial cycles. Nature Geoscience 2:500–504.

Rohling, E. J., K. Grant, C. Hemleben, M. Kucera, A. P. Roberts, I. Schmeltzer,H. Schulz, M. Siccha, M. Siddall, and G. Trommer. 2008a. New constraintson the timing and amplitude of sea level fluctuations during Marine IsotopeStage 3. Paleoceanography 23:PA3219, doi:10.1029/2008PA001617.

Rohling, E. J., K. Grant, C. Hemleben, M. Siddall, B. A. A. Hoogakker, M.Bolshaw, and M. Kucera. 2008b. High rates of sea-level rise during the lastinterglacial period. Nature Geoscience 1:38–42.

Rohling, E. J., A. Hayes, D. Kroon, S. De Rijk, and W. J. Zachariasse. 1998b.Abrupt cold spells in the NW Mediterranean. Paleoceanography 13:316–322.

Rohling, E. J., and F. J. Hilgen. 1991. The eastern Mediterranean climate attimes of sapropel formation: a review. Geologie en Mijnbouw 70:253–264.

Rohling, E. J., Q. Liu, A. P. Roberts, J. D. Stanford, S. O. Rasmussen, P. L.Langen, and M. Siddall. 2009c. Controls on the East Asian monsoon duringthe last glacial cycle, based on comparison between Hulu Cave and polarice-core records. Quaternary Science Reviews 28:3291–3302.

Rohling, E. J., R. Marsh, N. J. Wells, M. Siddall, and N. Edwards. 2004a.Similar melt-water contributions to glacial sea-level variability from ant-arctic and northern ice sheets. Nature 430:1016–1021.

Rohling, E. J., P. A. Mayewski, and P. Challenor. 2003. On the timing andmechanism of millennial-scale climate variability during the last glacialcycle. Climate Dynamics 20:257–267.

Rohling, E. J., P. A. Mayewski, A. Hayes, R. H. Abu-Zied, and J. S. L. Casford.2002b. Holocene atmosphere-ocean interactions: records from Greenlandand the Aegean Sea. Climate Dynamics 18:587–593.

Rohling, E. J., M. Medina-Elizalde, J. G. Shepherd, M. Siddall, and J. D.Stanford. 2012. Sea surface and high-latitude temperature sensitivity toradiative forcing of climate over several glacial cycles. Journal of Climate25:1635–1656.

Rohling, E. J., M. Sprovieri, T. R. Cane, J. S. L. Casford, S. Cooke, I. Bou-loubassi, and K. C. Emeis. 2004b. Reconstructing past planktic foraminiferalhabitats using stable isotope data: a case history for Mediterranean sapropelS5. Marine Micropaleontology 50:89–123.

Rossignol-Strick, M. 1985. Mediterranean Quaternary sapropels, an immediateresponse of the African monsoon to variation of insolation. Palaeogeography,Palaeoclimatology, Palaeoecology 49:237–263.

Rostek, F., E. Bard, L. Beaufort, C. Sonzogni, and G. Ganssen. 1997. Seasurface temperature and productivity records for the past 240 kyr in theArabian Sea. Deep Sea Research II 44:1461–1480.

Rumney, G. R. 1968. Climatology and the world’s climates. New York: Mac-millan.

Ruth, U., M. Bigler, R. Rothlisberger, M. L. Siggaard-Andersen, J. Kipfstuhl,K. Goto-Azuma, M. E. Hansson, S. J. Johnsen, H. Lu, and J. P. Steffensen.2007. Ice core evidence for a very tight link between North Atlantic andEast Asian glacial climate. Geophysical Research Letters 34:L03706, doi:10.1029/2006GL027876.

Saaroni, H., A. Bitan, P. Alpert, and B. Ziv. 1996. Continental polar outbreaksinto the Levant and Eastern Mediterranean. International Journal of Cli-matology 16:1175–1191.

Said, R. 1981. The geological evolution of the River Nile. New York: Springer.Sanchez-Goni, M. F., I. Cacho, J. L. Turon, J. Guiot, F. J. Sierro, J. P. Pey-

pouquet, J. O. Grimalt, and N. J. Shackleton. 2002. Synchroneity betweenmarine and terrestrial responses to millennial scale climatic variability dur-ing the last glacial period in the Mediterranean region. Climate Dynamics19:95–105.

Schmiedl, G., and D. C. Leuschner. 2005. Oxygenation changes in the deepwestern Arabian Sea during the last 190,000 years: productivity versus deep-water circulation. Paleoceanography 20:A2008, doi:10.1029/2004PA001044.

Schneider von Deimling, T., A. Ganopolski, H. Held and S. Rahmstorf. 2006.How cold was the Last Glacial Maximum? Geophysical Research Letters 33:L14709, doi:10.1029/2006GL026484.

Schulz, H., U. von Rad, and H. Erlenkeuser. 1998. Correlation between Ara-bian Sea and Greenland climate oscillations of the past 110,000 years. Nature393:54–57.

Scrivner, A. E., D. Vance, and E. J. Rohling. 2004. New neodymium isotopedata quantify Nile involvement in Mediterranean anoxic episodes. Geology32:565–568.

Shea, J. J. 2008. Transitions or turnovers? climatically-forced extinctions ofHomo sapiens and Neanderthals in the east Mediterranean Levant. Qua-ternary Science Reviews 27:2253–2270.

Siddall, M., E. J. Rohling, A. Almogi-Labin, C. Hemleben, D. Meischner, I.

Rohling et al. Mediterranean and Red Sea Paleoclimate S201

Schmeltzer, and D. A. Smeed. 2003. Sea-level fluctuations during the lastglacial cycle. Nature 423:853–858.

Siddall, M., E. J. Rohling, T. Blunier, and R. Spahni. 2010. Patterns of millennialvariability over the last 500 ka. Climate of the Past 6:295–303.

Siddall, M., E. J. Rohling, W. G. Thompson, and C. Waelbroeck. 2008. Marineisotope stage 3 sea level fluctuations: data synthesis and new outlook. Re-views of Geophysics 46:RG4003, doi:10.1029/2007RG000226.

Siddall, M., D. A. Smeed, C. Hemleben, E. J. Rohling, I. Schmeltzer, and W.R. Peltier. 2004. Understanding the Red Sea response to sea level. Earthand Planetary Science Letters 225:421–434.

Sirocko, F., M. Sarnthein, H. Erlenkeuser, H. Lange, M. Arnold, and J. C.Duplessy. 1993. Century-scale events in monsoonal climate over the past24,000 years. Nature 364:322–324.

Stocker, T. F., and S. J. Johnsen. 2003. A minimum thermodynamicmodel for the bipolar seesaw. Paleoceanography 18:1087, doi:10.1029/2003PA000920.

Sun, Y., S. C. Clemens, Z. An, and Z. Yu. 2006. Astronomic timescale andpalaeoclimatic implication of stacked 3.6-myr monsoon records from theChinese loess plateau. Quaternary Science Reviews 25:33–48.

Szabo, B. J., C. V. Haynes Jr., and T. A. Maxwell. 1995. Ages of Quaternarypluvial episodes determined by uranium-series and radiocarbon dating oflacustrine deposits of eastern Sahara. Palaeogeography, Palaeoclimatology,Palaeoecology 113:227–242.

Tiedemann, R., M. Sarnthein, and N. J. Shackleton. 1994. Astronomical time-scale for the Pliocene Atlantic δ18O and dust flux records of Ocean DrillingProgram site 659. Paleoceanography 9:619–638.

Trauth, M. H., J. C. Larrasoana, and M. Mudelsee. 2009. Trends, rhythms andevents in Plio-Pleistocene African climate. Quaternary Science Reviews 28:399–411.

Trewartha, G. T. 1966. The earth’s problem climates. London: Methuen.Trigo, I. F., T. D. Davies, and G. R. Bigg. 1999. Objective climatology in the

Mediterranean region. Journal of Climate 12:1685–1696.———. 2000. Decline in Mediterranean rainfall caused by weakening of Med-

iterranean cyclones. Geophysical Research Letters 27:2913–2916.Trommer, G., M. Siccha, E. J. Rohling, K. Grant, M. T. J. van der Meer, S.

Schouten, U. Baranowski, and M. Kucera. 2011. Sensitivity of Red Seacirculation to sea level and insolation forcing during the last interglacial.Climate of the Past 7:941–955.

Trommer, G., M. Siccha, E. J. Rohling, K. Grant, M. T. J. van der Meer, S.Schouten, C. Hemleben, and M. Kucera. 2010. Millennial scale variabilityin Red Sea circulation in response to Holocene insolation forcing. Paleo-ceanography 25:PA3203, doi:10.1029/2009PA001826.

Tzedakis, P. C. 1999. The last climatic cycle at Kopais, central Greece. Journalof the Geological Society, London 156:425–434.

———. 2009. Cenozoic climate and vegetation change. In The physical ge-ography of the Mediterranean. J. C. Woodward, ed. Pp. 89–137. Oxford:Oxford University Press.

Tzedakis, P. C., M. R. Frogley, I. T. Lawson, R. C. Preece, I. Cacho, and L.de Abreu. 2004. Ecological thresholds and patterns of millennial-scale cli-mate variability: the response of vegetation in Greece during the last glacialperiod. Geology 32:109–112.

Vaks, A., M. Bar-Matthews, A. Ayalon, A. Matthews, L. Halicz, and A. Frumkin.2007. Desert speleothems reveal climatic window for African exodus ofearly modern humans. Geology 35:831–834.

Waelbroeck, C., L. Labeyrie, E. Michel, J. C. Duplessy, J. F. McManus, K.Lambeck, E. Balbon, and M. Labracherie. 2002. Sea-level and deep watertemperature changes derived from benthic foraminifera isotopic records.Quaternary Science Reviews 21:295–305.

Wahby, S. D., and N. F. Bishara. 1981. The effect of the river Nile on Med-iterranean water, before and after the construction of the High Dam atAswan. In River inputs to ocean systems. J. M. Martin, J. D. Burton, and D.Eisma, eds. Pp. 311–318. Rome: UNEP-UNESCO.

Wang, X., A. S. Auler, R. L. Edwards, H. Cheng, E. Ito, and M. Solheid. 2006.Interhemispheric anti-phasing of rainfall during the last glacial period. Qua-ternary Science Reviews 25:3391–3403.

Wang, X. F., A. S. Auler, R. L. Edwards, H. Cheng, P. S. Cristalli, P. L. Smart,D. A. Richards, and C. C. Shen. 2004. Wet periods in northeastern Brazilover the past 210 ka linked to distant climate anomalies. Nature 432:740–743.

Wehausen, R., and H. J. Brumsack. 2000. Chemical cycles in Pliocene sapropel-bearing and sapropel-barren eastern Mediterranean sediments. Palaeogeog-raphy, Palaeoclimatology, Palaeoecology 158:325–352.

Whiteman, A. 1971. The geology of the Sudan Republic. Oxford: Clarendon.Williams, M. A. J., D. A. Adamson, B. Cock, and R. McEvedy. 2000. Late

Quaternary environments in the White Nile region, Sudan. Global andPlanetary Change 26:305–316.

Winckler, G., R. F. Anderson, M. Q. Fleisher, D. McGee, and N. Mahowald.2008. Covariant glacial-interglacial dust fluxes in the equatorial Pacific andAntarctica. Science 320:93–96.

Wolff, E. W., J. Chappellaz, T. Blunier, S. O. Rasmussen, and A. Svensson.2010. Millennial-scale variability during the last glacial: the ice core record.Quaternary Science Reviews 29:2828–2838.

Yuan, D., H. Cheng, R. L. Edwards, C. A. Dykoski, M. J. Kelly, M. Zhang, J.Qing, et al. 2004. Timing, duration, and transitions of the last interglacialAsian monsoon. Science 304:575–578.

S202 Current Anthropology Volume 54, Supplement 8, December 2013

� 2013 by The Wenner-Gren Foundation for Anthropological Research. All rights reserved. 0011-3204/2013/54S8-0004$10.00. DOI: 10.1086/673725

Neanderthal Demographic Estimates

by Jean-Pierre Bocquet-Appel and Anna Degioanni

CA� Online-Only Material: Supplement A

This article offers a critical review of population estimates for the Neanderthal metapopulation based on (paleo-)biological, archaeological, climatic, and genetic data. What do these data tell us about putative Neanderthal de-mography? Biological data suggest a similar demographic frame (life-history traits, such as potential maximumlongevity, age at menarche, and duration of gestation) between Neanderthals and modern humans. Archaeologicaldata have revealed a contradiction between the mortality pattern corresponding to 45� yr in Neanderthals and thelongevity displayed by the manifest continuum of extant mammals, including primates. Paleoclimatic data suggestthat the demography of Neanderthals, living as they did under highly fluctuating climatic conditions, was subjectto frequent bottlenecks. This demographic instability combined with the fragmentation of geographical areas andvariations in their distribution and extent could account for the fact that potential for technical creativity in theNeanderthal metapopulation would have been limited precisely because of its small numbers, leading it into whatis known as a “Boserupian trap” in macrodemographic theory. Finally, genetic literature reports different—butalways very low—estimations of the effective size (Ne) of the Neanderthal metapopulation. It is not easy to relateNe to the census size of a population, but by combining different demographic values, this study produced ninedifferent scenarios that were used to obtain an order of magnitude ranging from 5,000 to 70,000 individuals. Thecause of the cultural limitation of the Neanderthal metapopulation, compared with that of modern humans, maywell have resided in its small numbers alone.

In population biology, it is known that the number of indi-viduals of a species (its metapopulation)—that is, its demo-graphic size in terms of census data—is a measure of itsadaptive success. Thus, a population explosion is the resultof successful adaptation. We also know that this number ofindividuals determines the rate of evolutionary change. Therate is fast when the population size is small, initiating astochastic evolutionary direction; it is slow when a large pop-ulation pushes through the filter of purifying selection andagainst the resistance of niche construction (Kendal, Tehrani,and Odling-Smee 2011). This shows the value of linking de-mography to paleoanthropology. But there are many diffi-culties, mainly due to the nature of nondedicated information

Jean-Pierre Bocquet-Appel is Professor at the Ecole Pratique desHautes Etudes (Paris) and Research Director at the Centre Nationalde la Recherche Scientifique (UPR2147 44, rue de l’Amiral Mouchez,75014 Paris, France [jean-pierre.bocquet-appel@evolhum.cnrs.fr]).Anna Degioanni is Assistant Professor in the Department ofAnthropology at Aix-Marseille Universite, Centre National de laRecherche Scientifique (UMR 7269, Centre National de la RechercheScientifique, Maison Mediterraneenne des Sciences de l’Homme,Laboratoire Mediterraneen de Prehistoire [Europe-Afrique], BP 647,5 rue du Chateau de l’Horloge, 13094 Aix-en-Provence cedex 2,France). This paper was submitted 3 VII 13, accepted 13 VIII 13,and electronically published 19 XII 13.

of every kind that one must try to interpret demographicallyand to the amount of this information, which decreases dras-tically with temporal depth. In this paper, inferences aboutthe Neanderthal metapopulation are critically reviewed on thebasis of (paleo-) biological, climatic, archaeological, and pa-leogenetic data. What do these data tell us about the putativedemography of Neanderthals?

Demographic Inferences from (Paleo-)Biological Data: The Frame ofReference of Extant Mammals

Estimation of Potential Maximum Longevity in Mammals

For nearly 50 yr, thanks to the research initiated by Sacher(1959, 1975) and Sacher and Staffeldt (1974), a close rela-tionship has been known in extant mammals, including pri-mates, between maximum potential longevity (L) and thebiometric characters of brain weight (E) and body weight (P;Cutler 1975; Hofman 1993; for a review, see Hawkes 2006).This relationship is based on the biological quasi continuumbetween related species, which are connected by their phy-logenies, and the many similarities they generate. Used as anestimator of L for fossil hominins, this relationship has pro-duced values of 52, 78, 93, and 94 yr for Homo habilis, Homo

Bocquet-Appel and Degioanni Neanderthal Demographic Estimates S203

erectus, Neanderthals, and modern humans, respectively, aswell as age at sexual maturity taken as one-fifth of L, that is,12–13, 13–14, 18–19, and 18–19 yr, respectively (Cutler 1975;Sacher 1975). These estimates were incorporated into paleo-demography 35 yr ago as parameters influencing the shapeof the death distribution of fossil hominins (Bocquet andMasset 1977, 1982; Bocquet-Appel 1982). Updated estimatesof L, from biometric (estimated) data in more recent literatureusing the Hofman regression (1993),1 have produced 111.7and 111.2 yr for Neanderthals and modern humans, respec-tively, that is, identical figures between the two metapopu-lations and figures similar to those of Sacher (1975) and Cutler(1975). These estimates also suggest that other important de-terminants of life history, such as age at menarche or durationof gestation, were similar between Neanderthals and modernhumans, which allows them to be set within a common de-mographic frame.

Paleodemographic Estimates

Two death distributions by age, pre-Neanderthal (Homo hei-delbergensis; Bermudez de Castro and Nicolas 1997; Bocquet-Appel and Arsuaga 1999; minimum number of individuals p32, now 29; Bermudez de Castro et al. 2004) and Neanderthal(Trinkaus 1995; N p 206 registration numbers), were obtained,along with life span estimates of the australopithecines, un-til modern humans (Caspari and Lee 2004). The pre-Neanderthal distribution was obtained by a technique of ageestimation by self-reference2 based on dental attrition developedby Miles (1963, 2001; Bocquet-Appel and Arsuaga 1999). TheNeanderthal age distribution was obtained using the usual sta-tistical techniques exploiting morphological age indicators foradults (Trinkaus 1995), which are known to be biased towarda younger age (Bocquet and Masset 1982). These two sampledpaleontological distributions produce very high proportions ofyoung and mature adults (in the Neanderthal sample fromTrinkaus 1995, 80% of adults aged 20 yr and over die beforethe age of 40, not counting the question of children under 5yr of age, who are still significantly underrepresented in paleo-demographic data). These proportions have no equivalent inthe many controlled distributions of attritional death in otherprimates (Rawlins and Kessler 1987; Richard 1985), apes(Courtenay and Santow 1989; Hill et al. 2001; Teleki, Hunt,

1. Brain weight E p 1.036 # cranial capacity (Isler et al. (2008);average cranial capacity 1,519 cc, n p 9 (Trinkaus and Tompkins 1990);body weight P, 17 males, 77.6 � 4.5 kg; 9 females, 66.4 � 4.8 kg;averaged, 72 � 4.6 kg (Ruff, Trinkaus, and Holliday 1997, supplementaldata); Hofman regression no. 4, given with no standard deviation, R p0.896 (Hofman 1993:214).

2. This technique is described as “self-referencing” because the sampleanalyzed is its own anthropological age/indicator reference sample, unlikein other estimation techniques that require reference samples that areexternal to the sample analyzed. Moreover, the self-referencing approachis the technique of choice in paleoanthropology (Mann 1968), where itis impossible to make up anthropological age/indicator reference samplesof extinct hominins.

and Pfifferling 1976), preindustrial humans (Henry and Blayo1975; Jannetta and Preston 1991), or ethnographic humans(Hill and Hurtado 1996; Howell 1979).3

Caspari and Lee (2004, 2005a, 2005b, 2006) have detected“increased longevity, expressed as the number of individualssurviving to adulthood” (Caspari and Lee 2004:10895) fromthe ratios of old (30� yr) to young (15–29 yr) adults (notedOY, which is written in demographic notation as ωd30/15d15 andread as ωd30, the number of deaths at age 30 plus a number ofyears coinciding with the end of life, noted ω, over 15d15, thenumber of deaths at age 15, plus 15 yr (i.e., between 15 and29.9 yr). With samples of australopithecines, early Homo, Ne-anderthals, and Early Upper Paleolithic (modern human), Cas-pari and Lee obtained OY values of 0.12, 0.25, 0.39, and 2.08,respectively, with an abrupt change between Neanderthals andanatomically modern humans. If the information provided bythe data is true, then these ratios, with an increasing numberof older individuals relative to younger ones, express a trendtoward an increasing average life span in the metapopulationof extinct hominins (Caspari and Lee 2004, 2005a, 2005b,2006). The robustness of Caspari and Lee’s demonstration re-sides especially in the relatively large sizes of the paleontologicalsamples collected and in the use of the age estimation techniqueof Miles, which is homogeneous between groups and allowshuman groups to be set into a common comparative framewithout which this demographic signal, based on the OY in-dicator alone, would not have been detected.

Derived OY values express another interesting point of in-formation that combined with the estimates of potential lon-gevity L given above (111.7 and 111.2) produces the outlinesof living population pyramids. Assuming that populations arestable on average, the pyramids for these populations areregular and can be represented here by a triangular polygonwhere the height represents the ages and the width the agedistribution of the living population. The technical details forthe construction of the population pyramid from Caspari andLee’s data and an observed statistical relationship between OYvalues for dead and living in sample life tables of modernhuman populations (table 1; fig. 1) are given in CA� onlinesupplement A. What is obtained is a highly schematic rep-resentation of a population pyramid at 15� yr. The pyramidsfor the two groups are represented in figure 2. They show theflattening effect produced by the departure of the pyramid atthe point of the statistical longevity ω rather than at the

3. It is this observation of the systematic structural deviation betweencontrolled demographic patterns in primates, including humans, and thepatterns obtained most frequently from paleodemographic data thatbrought the results of the latter into question 40 yr ago (Bocquet andMasset 1982, 1985; Bocquet-Appel 1986; Masset 1973). The alternativeposition was to accept the paleodemographic distributions as true, toconsider them as ancestral patterns of the demographic distributions ofextant primate populations, and to consider the conflicting references ofcontrolled distributions of extant primates as expressing the tyranny ofactualist demography. This was the main thrust of the arguments givenby the American Journal of Physical Anthropology to one of us in 1977in rejecting a manuscript that discussed past orthodoxy.

S204 Current Anthropology Volume 54, Supplement 8, December 2013

Table 1. Paleodemographic statistical data

GroupDeath at 15 � 29 yr p

Y p 15d15

Death at 15� yr pO p ωd15

Y/(Y � O) dead p

15d15/ωd15

Estimated OYlivinga p 15L30/15L15

Estimated maximumlongevity L

Australopithecinesb 316 353 .895 .505 51.5c

Early Homo 166 208 .798 .539 73.5d

Neanderthal 96 133 .721 .569 111a,b

Early Upper Paleolithic 24 74 .324 .805 111a,b

a From , , , preindustrial life tables with low life expectancy2OY living p 0.47274 � 0.29547 # log [Y/(Y � O)] dead � 0.01904 R p 0.940 N p 44at birth, excluding Yanomamo, stationary population (see Bocquet and Masset 1977; Bocquet-Appel 2002), , , .F p 658.1 df p 1, 42 P ! .00001b Number of Australopithecines to Early Upper Paleolithic, aged by Miles’s technique (Caspari and Lee 2004).c Australopithecine and early Homo (Cutler 1975, table 2, excluding Homo habilis; see Caspari and Lee 2004).d Homo erectus, average.

Figure 1. Relationship between OY living (15L30/15L15, vertical axis)and the variable Y/(Y � O) dead (15d15/ωd15, horizontal axis) ina sample of 44 preindustrial life tables representing stationarypopulations (r p 0).

maximum longevity L, which gives the population pyramidsa more acute angle.

The first pyramid for the australopithecines, compared withthe others, resembles those known for the great apes. Partic-ularly notable is the similarity of the Neanderthal and EarlyUpper Paleolithic pyramids, which both have a base width at15 yr similar to that for early Homo, estimated from pale-ontological data, and a height identical to the Early UpperPaleolithic, estimated from zoological data. If L and ω arecorrectly estimated, then in the Neanderthal paleontologicaldata, the proportion of deaths corresponding to 45 � ω ofthe population pyramid is either missing or not recognizedamong the skeletal remains. The similarity in the represen-tation of population pyramids for Neanderthals and the EarlyUpper Paleolithic highlights the contradiction between thebiodemographic pattern of longevity of extant mammals, in-

cluding primates, and the paleodemographic signal detectedby Caspari and Lee that does not fit in the frame. We haveno solution to this contradiction, but it cannot be ignored(see also Trinkaus 2011). It raises two issues: one is technicaland concerns the recurrent problems of paleodemographicage estimation, while the other concerns population biologyand the assumption of faster maturation in Neanderthals thanmodern humans, which we will now discuss.

Miles’s technique is based on a self-reference in the ana-lyzed sample. It therefore seems to escape from the a prioriprobabilities that are encysted in the anthropological ages/indicators reference sample used in other techniques, whichpredetermine the age estimates and for which there is now arange of solutions currently under evaluation (Bocquet-Appeland Bacro 2008; Caussinus and Courgeau 2010; Hoppa andVaupel 2002; Lucy, Ackroyd, and Pollard 2002; Lucy et al.1996). Nevertheless, it may seem surprising that Miles’s tech-nique, which has been used many times in paleoanthropologysince its publication in 1963 (for an overview, see Miles 2001),has never, to the best of our knowledge, been tested with asample of skeletons of known ages, with its share of immatureindividuals to calibrate the standards and its share of adultsto apply them to. An old idea is that tooth wear is not linearwith chronological age, as implied in the technique, but de-creases asymptotically in older individuals. As long as thisvalidation test has not been done, in particular to estimatethe confidence intervals (CIs), this technique should be con-sidered conservatively, as indicated by Miles himself, as “anart, not a precise science” (Miles 2001:980).

The hypothesis of significantly faster biological maturationin Neanderthals than in modern humans has been put for-ward in recent years as estimated from the growth of toothenamel (Guatelli-Steinberg et al. 2005; Ramirez-Rozzi andBermudez de Castro 2004; Smith et al. 2007). As there is acorrelation between tooth development and other life-historytraits, then there are grounds for asking whether the life spanof Neanderthals, despite their large brains, could also havebeen shorter, as revealed by the growth of tooth enamel. Ifthis is indeed the case, it would suggest a death distributionat a younger age, on average, than in anatomically modernhumans. But there is no consensus over this hypothesis offaster maturation for reasons to do with the dental techniques

Bocquet-Appel and Degioanni Neanderthal Demographic Estimates S205

Figure 2. Schematic representation of the population pyramids for ages 15� yr of living populations of australopithecines, earlyHomo, Neanderthals, and Early Upper Paleolithic (modern humans) estimated from the 15L30/15L15 ratios, derived values of OY(Caspari and Lee 2004), and estimated maximum potential longevity (regression: Hofman 1993). Regarding the australopithecines,because the statistical demographic longevity ω is lower than the upper limit of the area representing the living at 30–45 yr of age,of unit density, this limit of 15L30 was set at ω.

used, the sizes of samples counted in units, and the failureto consider interindividual variations in Neanderthals andmodern human in comparisons (for a summary, see Guatelli-Steinberg 2009). Moreover, the hypothesis of rapid matura-tion does not imply, in paleodemographic distributions, anabsence of individuals with recognized aging markers, asseems to be the case in the data. These faster-maturing in-dividuals would also have these markers of aging, but theywould simply be chronologically younger than their modernhuman counterparts. Finally, ethnographic assumptions(Trinkaus 1995, 2011), or assumptions by analogy with eth-nohistorical environmental crises (Bocquet-Appel and Arsu-aga 1999), have been put forward to explain the apparentdeficit of adults over 40 yr old and even over 25 yr old inpre-Neanderthal and Neanderthal samples.

To summarize, the mortality pattern of Neanderthals, asrevealed by paleontological data, does not correspond to thepromise of longevity displayed by the manifest continuum ofextant mammals, including primates. Could one of the twobe wrong? This question, which is beyond the scope of thispaper, cannot be evaded.

Demographic Inferences from Paleoclimaticand Archaeological Data

We know there is a connection between climatic variationsin isotopic stages and the frequency of Mousterian archaeo-

logical sites in Europe (van Andel, Davies, and Weninger2003). This link is determined—via the amount of biomedistribution of primary biomass—by the biomass of ungulatepopulations available for hunter populations. From these rel-atively long-cycle climatic variations in the isotopic stages ofcelestial mechanics, it gradually becomes clear that medium-cycle cold fluctuations (Heinrich events 10–8 kyr [H]) aresuperimposed over relatively short-cycle and short-durationtemperate-cold fluctuations (Oeshger-Dansgaar events 1.5 kyr[DO]). These DO changes, which occur rapidly over time,significantly amplify the predominant climate in the isotopicstages. From oxygen isotope stage (OIS) 5 of the Eemianonward, 25 DO events occurred in Europe (Sanchez-Goni etal. 2008). Others would be expected to appear in earlier pe-riods when Neanderthals lived but are so far unknown forlack of data. A correlation between the frequency of UpperPaleolithic archaeological sites and the variation in DO eventshas been interpreted in terms of demographic and culturalchanges in Europe, which roughly correspond via changes inthe ecological zones of primary and secondary ungulate bio-mass (d’Errico, Sanchez-Goni, and Vanhaeren 2006). Howdid local populations react to these significant variationsacross their expansion area? Rather than a model of the ebband flow of small local populations on the geographical mar-gins of the metapopulation to and from attested refuge zonesover evolutionary time—which were, in fact, already occupied

S206 Current Anthropology Volume 54, Supplement 8, December 2013

by the cores of the metapopulation—a model of regionalextinction in situ is proposed (Hublin and Roebroeks 2009).

Other consequences of large-amplitude and long-durationclimatic fluctuations—in their low points of glacial temper-ature, moreover amplified by H and DO events—were likelyto create demographic bottlenecks for the Neanderthals, whoavoided permafrost zones (Aiello and Wheller 2003), thatdetermine stochastic genetic drifts (Bruner and Manzi 2006).But it is not certain, on the other hand, that the peak inter-glacial temperatures (isotopic stage 5e and 7e), which weresimilar to current temperatures and corresponded to a plantrecolonization on the latitudes, also corresponded to periodsof population growth. Recolonization can occur with a verylow population density. This is because for the interglacialperiods, account must be taken of the result, in terms ofungulate biomass, of the combined effects of (i) the expansionof habitable areas due to the withdrawal of ice masses anddespite the rise in sea level (3–7 m above the current level),and (ii) the decrease in the area of the steppe tundra as itshifted toward the high latitudes up to the pole. The con-ditions that brought this expansion of habitable areas werefavorable to forests (van Andel and Tedzakis 1996; Kukla etal. 2002), where ungulate biomass was very low. The north-ward shift of steppe tundra toward the high latitudes resultednot only in a smaller area, as mentioned above, but also inlower productivity as available sunlight diminished. Until allnet effects relative to the distribution of the major biomesduring the Eemian have been simulated, taking into accountthe conflicting effects described above, it will not be possibleto equate the interglacial with a demographic expansion ofthe Neanderthal metapopulation (see also Frenzel 1985; Gam-ble 1987).

Bottleneck situations and purifying selection are both chro-nologically identifiable. These correspond to the low (Brunerand Manzi 2006) and likely midpoints of climate series. Nev-ertheless, and underlying the anatomical traits that appearedin some glacial episodes because of drift (Hublin 1998), even-tually producing the classic Neanderthal morphotype de-scribed as “hyperarctic” (Aiello and Wheeler 2003; Holliday1997), the metapopulation had to reach a significant size inthese cold and harsh environments for the filter of purifyingselection to act on the phenotypes carrying selected genotypes,in particular via Bergman and Allen’s rule. These periods,when populations were relatively more numerous, may havematched phases of OIS 8 and 6, in which long, cold, but notextreme periods emerge (except H and DO events, which arenot yet listed).

Finally, from the differential distribution of well-docu-mented archaeological remains in the Perigordian region—between the Chatelperronian 45,000–40,000 yr ago and theAurignacian 40,000–35,000 yr ago, attributed respectively toNeanderthals and to anatomically modern humans—Mellarsand French (2011) have estimated vestige quantity as 10 timeslarger for the modern human population. The authors equatethis relative difference in the amount of remains in this area

to the demographic sizes of both populations. This area is arefuge zone on the evolutionary timescale. It should be re-membered that the size of the Aurignacian population in thissame area has been estimated at 795–12,980 individuals witha 95% CI (Bocquet-Appel et al. 2005:1665, fig. 5). By applyingthe magnitude of the relative difference estimated by Mellarsand French (2011) to this Aurignacian population size, weobtain a local Neanderthal population of 80–1,300 individualsin the Perigordian refuge area before the time of contact.

Hypothesis of a Boserupian NeanderthalPopulation Trap

The homogeneity of the lithic cultural remains of Neander-thals during their last 150 kyr is striking, except, apparently,at the point of their extinction (Bocquet-Appel and Tuffreau2009), suggesting very low technical elasticity despite the sig-nificant pressure of the environmental hazards summarizedabove, which should have favored innovations because of theoverall change in the ungulate biomass. Questions about thecognitive efficiency of Neanderthals are thereby raised (Neu-bauer and Hublin 2012). But the hypothesis of technical lim-itations in the production system of Neanderthal hunters dueto a demographic trap (Bocquet-Appel and Tuffreau 2009)must also be put forward again but more properly defined.

In any population, the production of innovations dependsnot only on its cognitive biological capacities but also on itsdemographic size. With the same cognitive capacity, underthe simple assumption that innovations are produced at a lowfrequency in any population, then the most demographicallynumerous population, in absolute terms, will produce thegreatest number of innovations (Kremer 1993; Kuznets 1973;Simon 1977; see also Powell, Shennan, and Thomas 2009;Shennan 2001). If the size of the Neanderthal metapopulationremained very low in terms of carrying capacity—that is tosay, the maximum number of mouths that it was possible tofeed per square kilometer of ungulate biomass given its pro-duction system (the technical and social relationships in anenvironment)—then the metapopulation could have main-tained itself in a state of demographic equilibrium at a “criticallevel of density” (Boserup 1965:33), but its potential technicalcreativity would have been strongly limited in what is knownas a “Boserupian trap” in macrodemography theory.

The technical and social characteristics of the Neanderthalproduction system might, in central and northern Europe,have operated in an open environment: high residential mo-bility but not to great distances (Conarda, Bolusc, and Mun-zeld 2012) and targeted to large gregarious herbivores andresources (horses, bison, reindeer, and ibex) with a smalleramount of many other larger game species typically domi-nating the assemblages (Conarda, Bolusc, and Munzeld 2012;Gamble 1999; Patou-Mathis 2000); consumption of shellfish,birds, and turtles in the peripheral southern latitudinal zonesof the expansion area (Finlayson et al. 2001); direct and dan-gerous contact with prey animals by killing with lances rather

Bocquet-Appel and Degioanni Neanderthal Demographic Estimates S207

than killing at a distance using projectiles (spears; Gamble1999) with the aid of beaters and with no division of laborby gender (or by age? Kuhn and Stiner 2006) between huntingand gathering, as observed ethnographically, that is, with bothmales and females working as hunters and beaters. The car-rying capacity of this hunter-gatherer production systemalong with high incidental mortality (Trinkaus 1995) shouldbe lower than in other systems in which clearly more efficienthunter-gatherer hunting techniques (spears, bows) were used.In addition, the energy balance of this putative productionsystem, combining high energy expenditure due to mobility(following herds) and energy gain from a low calorie diet(essentially game), determines long birth intervals in modernhuman females and, therefore, low fertility (see Bocquet-Appel 2008). By analogy, it may be thought that the energybalance/fertility reaction norm was similar in Neanderthalsand that their female fertility (total fertility rate) thereforetended, on average, toward low values (!Kung: 5 children andless) rather than high values. The relatively low fertility andhigh mortality of the Neanderthal hunter-gatherer productionsystem would have been a contributing factor in locking theminto the Boserupian trap.

Although ethnographic demographic control data are lack-ing on this point, there are two conceivable circumstances inwhich a population of hunters may have escaped stagnationin a Boserupian trap: (i) the outcome of the rate of innovation,which would sporadically increased the carrying capacity, and(ii) favorable environmental change (Wood 1998:112). Unlikethe realization of the rate of innovation, which is a simplestatistical function of the rate and the metapopulation sizeand cannot be located precisely in time, favorable environ-mental changes were a certainty in western Eurasia, especiallyover the relatively long durations of OIS 5 and 7. These fa-vorable environmental changes resulted in the extension ofsteppe savannah and grassland areas together with the cor-responding ungulate biomass and its carnivorous predators,which included the increasing Neanderthal metapopulation.Along with this demographic growth, the population expan-sion area was gradually homogenized, particularly through adefragmentation of the geographical structure inherited fromthe previous cold period, the increase in interpopulation mi-gration, and the redensification of local populations, all ofwhich favored the spread of innovations.

With no change in hunting technique (“in the herd”: Gam-ble 1999) but simply by virtue of the increase in the overallbiomass of ungulates during OIS 5 and 7, the logistic mobilityof the Neanderthal hunter system could have diminished. Theeffect of reduced logistic mobility (with the correspondingdecrease in energy expenditure) is to improve the energy bal-ance, which results in a rise in female fertility through areduction in the birth interval (Bocquet-Appel 2008). An in-crease in the metapopulation or local populations could thenoccur through an opportunistic adjustment of female fertilityto the new carrying capacity rather than a “technological”reduction in mortality due to hunting accidents (via killing

at distance). But 150 kyr of apparent Neanderthal techno-logical stability do not make the case for an escape from theBoserupian trap during the climate windows of OIS 5 and 7.

Demographic Inference from AncientNeanderthal DNA

While there has been general agreement since the 1970s overthe effective population size (Ne) of 10,000 individuals formodern humans (Hammer and Zegura 1996; Harding et al.1997; Nei and Graur 1984; Rogers and Jorde 1995; Takahata1993; Takahata, Satta, and Klein 1995; Wilson et al. 1985),there is no such consensus over the Neanderthal Ne. Thereis now enough data from Neanderthal DNA analysis to pro-vide an outline. Analysis of Neanderthal DNA sequences be-gan in 1997 (Krings et al. 1997) from the original specimendiscovered in 1856, but information has since become avail-able on mitochondrial sequences from 14 additional Nean-derthal specimens (see Degioanni, Fabre, and Condemi 2011for a review), including complete mitochondrial genome se-quences from nine Neanderthal individuals (Briggs et al. 2009;Green et al. 2008, 2010). Neanderthal genomic DNA has alsobeen sequenced since 2006 (Green et al. 2006; Noonan et al.2006: 65,000 and 1,000,000 base pairs, respectively). Finally,in 2010, the genes of three Neanderthal individuals to 1.3-fold genomic coverage were sequenced by the NeanderthalGenome Project (Green et al. 2010). This project mainly aimsto determine the complete sequence of Neanderthal DNA, toanswer the question of recent interbreeding between Nean-derthals and modern humans, and to provide a catalog ofdifferences between the human and Neanderthal genomes.The Neanderthal Genome Project has also revealed the se-quence of several Neanderthal protein-coding genes that haverecently evolved adaptively in modern humans (Green et al.2010). Several publications on Neanderthal DNA sequencesfocus on the time of the most recent common ancestor ofNeanderthals and modern humans, and several publicationsreport estimations for Neanderthal Ne.

The following is a brief digest of the demographic findingsof these studies. A first study comparing three short mtDNAsequences (Krings et al. 2000) suggested that Neanderthalshad expanded from a small population to explain the lowerdiversity of Neanderthal mtDNA than in the great apes andthe same order of magnitude in modern humans. This resultwas confirmed by analyzing five Neanderthal mtDNA ge-nomes (Briggs et al. 2009): the Ne was small and probablyincluded fewer than 3,500 females (mean Ne p 1,476; 268 to3,510, 95% HPD), and the authors proposed that the lowmtDNA diversity might reflect a low Neanderthal Ne over alarge part of their history. Estimating the θ p Neμ parameter(θ is the nucleotide diversity among sequences, and μ is thesubstitution rate per generation), assuming constant popu-lation size and 20 yr per generation over a short sequence of9 HVI mtDNA, Lalueza and colleagues (2005) propose an Ne

ranging from 5,000 to 9,000 individuals and confirm that this

S208 Current Anthropology Volume 54, Supplement 8, December 2013

population would have been constant over time. The com-plete Vindija 33.16 mitochondrial genome sequence (Greenet al. 2008) showed a significantly higher ratio of nonsynon-ymous to synonymous evolutionary (dN/dS) rates in Nean-derthals versus modern humans. This result could be ex-plained by a smaller Neanderthal Ne and contrasts with HVI1 sequence results from Teshik Tash and Okladnikov indi-viduals (Krause et al. 2007)—based on mean pairwise differ-ences, which suggested that Neanderthals had an Ne similarto that of modern Europeans or Asians but lower than thatof modern Africans—and also contrasts with the results ofOvchinnikov and Kholina (2010), where the dN/dS ratio in-dicates that the Ne for their common ancestor (the human-Neanderthal branch) tends to be larger than in either Ne-anderthals or modern humans.

On mtDNA data again, but using a modeling approach,Fabre, Condemi, and Degioanni (2009) found that the de-mographic scenario that best explains the variability of Ne-anderthal is a population with an Ne ranging from 3,000 to25,000 individuals that will grow in size up to 50,000 yr BPand then decline slowly until extinction. A larger distributionvalue for Neanderthal Ne, whose median corresponds to32,263 individuals, was proposed instead by the best scenarioanalyzing a possible admixture process between Neanderthalsand early European modern humans (Ghirotto et al. 2011).Ne values from nuclear DNA are very rare but confirm thatNeanderthals derived from a very small ancestral populationwith an Ne of about 3,000 ranging up to 12,000 (Green et al.2006). This estimated Ne turns out to be similar to that pro-posed for a population of modern humans: this “small pop-ulation” character seems to be a feature of both Neanderthaland modern human evolution.

As we have seen, the literature offers different Ne values,but all publications agree that the Neanderthal Ne would bevery low. But what is the Ne? Ne is defined as the size of anideal population, a Wright-Fisher population (Fisher 1930;Wright 1931) that has the same rate of change of allele fre-quencies or heterozygosity as the population under study. Itis not easy to relate the Ne to the census size of a population(Nc). The Ne is almost always smaller than the actual size Nc.Some authors (Belle et al. 2006; Ray 2003; Wood 1987; for areview, see Hawk 2008) propose that the Ne is approximatelyone-half of the census size corresponding to the reproductiveindividuals of the population. Moreover, many parameterscan affect the Ne value; this raises the question of whether Ne

is in fact a reliable predictor of Nc.In the field of animal species conservation, where it is es-

sential not only to maintain a sufficient population size forthe survival of the species but also some degree of geneticvariability, researchers were already focusing in the 1970s onthe relationship between Ne and Nc, for which Hill (1972)proposed an initial formula. This formula, modified by Nun-ney (1993) and widely used in animal studies, is written

N 4r(1 � r)Te p ,N [rA (1 � I ) � (1 � r)A (1 � I ) � (1 � r)I � rI ]f A m A bm bff m

where Ne is the effective size, N the population size, r the sexratio; A is the adult life span; T is the generation time; IA isthe standardized variance (variance/mean2) in the adult lifespan; Ib is the standardized variance (variance/mean2) in thereproductive success of sex; m indicates male and f female.N corresponds exactly to the size of the population (Nc) ofanimals with early sexual maturity and to the size of the adultpopulation only (Na) in late-maturing species (see Nunneyand Elan 1994). To obtain the N census value for Neander-thals, being presumably also late maturing like modern hu-mans, the fraction of the juvenile population (0–19.9 yr old,or 20L0 in demographic notation) must therefore be added toNa. While these parameters are readily available for ecologicalstudies on living animals (e.g., livestock), estimating Na forpast populations, for which none of the parameters of theformula are known, is much more complicated. It is never-theless possible to put forward known values for living hunter-gatherer populations and to use these values in the knowledgethat although they are not “real” values for Neanderthal pop-ulations, it is probable that the real values would be in thesuggested range. Our aim is to obtain an order of magnituderather than exact values. By cross-referencing the values, sev-eral scenarios are obtained. To reduce their number, and sincewe have no evidence to the contrary at present, we considerthat the population consisted of equal numbers of men andwomen (r p 0.5), that the men and women had the sameadult life span (Af p Am p Ai) and the same reproductivesuccess (Ibm p Ibf p Ibi). Regarding the adult life span, thereare only two detailed demographic studies of hunter-gatherers(Hill and Hurtado 1996; Howell 1979) giving this information(for individuals of 20� yr, both sexes). The adult life spansfor both sexes are virtually identical, at 36.5 for !Kungs and37 yr for Aches. These two poulations live in the tropics,unlike the Neanderthals, who lived in temperate and, mostfrequently, subarctic regions. These studies include inconsis-tencies, which are discussed by the authors (e.g., Hill andHurtado 1996:258; Howell 1979:116). However, in the ab-sence of any other information, these demographic data onliving hunter-gatherers are used in this article to estimate arange of adult life span (�3 yr: 33–40 yr), which are extendedto lower values for the lower class boundary to take the effectof a cool or cold climate into account as well as the demo-graphic reality possibly reflected by the very small proportionof adult Neanderthal skeletons found. The result is a relativelybroad range of 25–40 yr for the hypothetical life span ofNeanderthal adults, with an estimated SD (deviation/mean2)in the adult life span (both sexes) of 38.88888/(25 or 30 or

.240)Combining these values produced nine different scenarios.

For each scenario, we apply the formula to estimate the Na

value of the Neanderthal population given the extreme values

Bocquet-Appel and Degioanni Neanderthal Demographic Estimates S209

Table 2. Values of the parameters used to calculate theNe/Na ratio

Parameter Value

r p sex ratio .5T p generation time 17.5, 20, 25Ai p Af p Am p adult life span 25, 30, 40a

IAi p IAf p IAm p standardized variance(variance/mean2) in the adult lifespan of sex i

38.8888/Ai p .06222,.04320, .02430

Ibi p Ibf p Ibm p standardized variance(variance/mean2) in the reproductivesuccess of sex i

6/22 p 1.5

a Based on Hill and Hurtado (1996:196) and Howell (1979:88).

of Ne estimated from genomic DNA (Ne range from 3,000 to12,000).

We used Ne values estimated from the analysis of genomicDNA for two reasons. First, because the formulas availableto calculate Na are more suited to this type of data, and second,because we believe that the information derived from theentire genome is less biased than the information containedin mitochondrial DNA alone (females only, very wide chro-nological dispersion of sequences).

We also tested three different proportions (juvenileL 200

population): 40%, 50%, or 60% of the total population. Thedata used for the nine scenarios are shown in table 2, andthe results are shown in figure 3. The figure clearly showsthat the Ne/Na ratio reaches its maximum when the generationtime (T) is the longest and life expectancy (Ai) the shortest:in this case the N (Na and Nc) values are the lowest, evenwith 60% of young individuals in the population, rangingfrom 5,000 to 50,000. On the other hand, when the ratio Ne/Na is the lowest, that is to say, when T is short and Ai is long,Na and Nc are the highest, with a maximum of 70,000 indi-viduals.

An estimated population size of 5,000 to 70,000 individualsshould not be considered as an exact value but rather as anorder of magnitude. It should be remembered, furthermore,that the formula proposed by Nunney applies to a populationwith no subdivisions and no generation overlap. If this is notthe case, then generation overlap can reduce Ne to 25%–75%of Na (Felsenstein 1971). Reproductive variance (variation inthe contribution to the next generation) between males andfemales (Ai) implies that the Ne of portions of the genomewith different inheritance patterns can be different (the higherthe reproductive variance, the lower the Ne). This means thatNe estimated for the same population but with different mark-ers (Y chromosome, mtDNA, and autosomal markers) canbe very different and therefore difficult to compare. Nonran-dom mating, in particular assortative mating (mate chosenon the basis of phenotypic similarities) decreases the Ne andincreases the genetic drift. In a subdivided population, non-random mating can have a greater effect on members of thesame subpopulation: the Ne of a subdivided population canbe different (lower) compared with a randomly mating pop-ulation of the same size.

Finally, but crucially, it is important to keep in mind thatthe average Ne over the long term is not the “classical” arith-metic mean but rather the harmonic mean over several gen-erations (Crow and Kimura 1970; Wright 1938). This meansthat the Ne is strongly affected by the smaller Ne values andwill be close to the smallest Ne over several generations; thatis, bottlenecks can mask previously high Ne values.

The question we must ask is to what period the value ofNe calculated from Ne corresponds: beginning, average value,or end? The current proliferation of studies geared to theconservation and management of endangered or exploitedspecies (Gomez-Uchida et al. 2013; Serbezov et al. 2012;Whiteley et al. 2012) has produced new methods for esti-

mating long-term and short-term Ne. This research field willtherefore provide answers to our question within a short time.

Concluding Remarks

Paleodemographic data are eclectic by nature, but the effortmust be made to integrate their interpretation without hidingthe difficulties and to resolve contradictions, especially re-garding paleontological metapopulations such as the Nean-derthals, where the mists of time become increasingly im-penetrable with chronological depth. In Neanderthalpaleodemographic death distributions by age, very few adultsare older than 40, while the promise of potential maximumlongevity implied by the quasi-biological continuum of mam-mals points to much more. One could even venture to assumethat the Neanderthal and modern human death distributionsshould be similar.

Increasingly detailed reconstitutions of climate, layeringmultiple sequences of variations that range from very long toshort periods (Sanchez Goni et al. 2008), and their connec-tions to the Neanderthal population through geographicallydistributed primary and secondary biomass raise new ques-tions. It is as if the last 10,000 yr of the Holocene, duringwhich the modern human metapopulation will reach 9 billionpeople, were a temperate niche of stability, as the latest similarniches date back to OIS 5 (Eemian: 114–130 kyr) and OIS11 (core: 400–420 kyr; see also Richerson, Boyd, and Bettinger2009). Except in these three temperate niches outside theMediterranean zone, the vegetation was mainly cold steppetundra and was regularly devastated by what would now beakin to catastrophic DO and H climate events. It can behypothesized that the demography of the Neanderthal meta-population, living under conditions where extreme environ-mental instability with short periods was the norm, was pri-marily stagnant, with frequent bottlenecks and episodes ofdecline.

Finally, the Neanderthals may have suffered the additionalhandicap of a system of specialized hunters, described as top-level carnivores (Richards and Trinkaus 2009), even thoughplants were also consumed (Henry, Brooks, and Piperno2011). Relative to a modern human, the estimated individual

Figu

re3.

Ne/

Na

and

Na

and

Nc

vari

atio

nde

pen

din

gon

T,

Ai

valu

e,an

d20

L 0pr

opor

tion

s.a,

Tge

ner

atio

nti

me

p17

.5.

b,T

p20

.c,

Tp

25.

Bocquet-Appel and Degioanni Neanderthal Demographic Estimates S211

metabolic cost of an adult Neanderthal is very high (3,500–5,000 kcal/d vs. 2,150–2,400 kcal/d for a male modern human;Churchill 2007; Sorensen and Leonard 2001; Steegman, Cerny,and Holliday 2002). Assuming a similar distribution of nu-trients between Neanderthals and modern humans, food in-take per capita for Neanderthals would have required almosttwice the mass of ungulate meat as that consumed by modernhumans. Mechanically, given similar natural hunting condi-tions, this level of ungulate consumption implies that theNeanderthal population density was half that of modern hu-mans. The demographic instability of this metapopulation ofspecialized hunters, which was small on average, and the var-iation of its geographical area of expansion and fragmentation,should help to understand why it stagnated technologically(Bocquet-Appel and Tuffreau 2009) and probably also socially(see the array of Hayden 2012), as it spent most of its evo-lutionary time moldering in the depths of a Boserupian de-mographic trap.

Measuring the acquisition or lack of acquisition by Ne-anderthals of a modern form of human behavior—that is,the ability, usually shown in modern humans, to express avery wide range of cognitive responses to contrasting so-cionatural situations—by simply comparing lists of archae-ological cultural items between the two groups is not anappropriate approach (see Shea 2011). If one of the meta-populations (the Neanderthals) remained largely within theconfines of the experience of the Boserupian traps men-tioned above while the other (modern human), numberingseveral million, was able to experience the African and Eur-asian population expansion, then comparing their lists ofcultural items at a point of fortuitous spatiotemporal contact(say in the Perigord at 40–35 kyr) does not inform us abouttheir different biological cognitive potential. The differencesbetween these lists will mirror those of the amplitudes ofthe socionatural experiences of these contemporaneousmetapopulations whose demographic numbers varied by afactor of 100 and perhaps even more. The cause of thecultural limitation of the Neanderthal metapopulation com-pared with that of modern humans may well have residedin its small numbers alone.

Acknowledgments

We thank the organizers and the Wenner-Gren Foundationfor taking the opportunity to participate in this very stimu-lating conference, Christine Verna for her comments, and twoinsightful anonymous reviewers. The figures were producedby Daniele Fouchier, Nathan Bocquet-Appel, and StephaneRenault.

References CitedAiello, Leslie C., and Peter Wheeler. 2003. Neanderthal thermoregulation and

the glacial climate. In Neanderthals and modern humans in the Europeanlandscape during the last glaciation: archaeological results of the Stage 3 project.

Tjeerd H. van Andel and William Davies, eds. Pp. 147–166. Cambridge:McDonald Institute.

Belle, Elise M. S., Uma Ramakrishnan, Joanna L. Mountain, and Guido Bar-bujani. 2006. Serial coalescent simulations suggest a weak genealogical re-lationship between Etruscans and modern Tuscans. Proceedings of the Na-tional Academy of Sciences of the USA 103:8012–8017.

Bermudez de Castro, Jose Maria, Maria Martinon-Torres, Marina Lozano,Susana Sarmiento, and Ana Muela. 2004. Paleodemography of the Ata-puerca-Sima de los Huesos hominin sample: a revision and new approachesto the paleodemography of the European Middle Pleistocene population.Journal of Anthropological Research 60(1):5–26.

Bermudez de Castro, Jose Maria, and Marıa Elena Nicolas. 1997. Palaeode-mography of the Atapuerca-SH Middle Pleistocene hominid sample. Journalof Human Evolution 33(2/3):333–355.

Bocquet, Jean-Pierre, and Claude Masset. 1977. Estimateurs en paleodemo-graphie. L’Homme 17:65–90.

———. 1982. Farewell to paleodemography. Journal of Human Evolution 11:321–333.

———. 1985. Paleodemography: resurrection or ghost? matters of moment.Journal of Human Evolution 14:107–111.

Bocquet-Appel, Jean-Pierre. 1982. Une esquisse de quelques mecanismes de-mographiques au paleolithique. In 1er Congres international de paleontologiehumaine. Jean Piveteau and Henry de Lumley, eds. Pp. 971–992. Paris:Institute de paleontologie humaine.

———. 1986. Small populations: demography and paleoanthropological in-ferences. Journal of Human Evolution 14:683–691.

———. 2002. Paleoanthropological traces of Neolithic demographic transi-tion. Current Anthropology 43:638–650.

———. 2008. Explaining the Neolithic demographic transition. In The Ne-olithic demographic transition and its consequences. Jean-Pierre Bocquet-Appel and Ofer Bar-Yosef, eds. Pp. 35–56. Dordrecht: Springer.

Bocquet-Appel, Jean-Pierre, and Juan Luis Arsuaga. 1999. Age distributionsof hominid samples at Atapuerca (SH) and Krapina indicate accumulationby catastrophe. Journal of Archaeological Science 26(3):327–338.

Bocquet-Appel, Jean-Pierre, and Jean-Noel Bacro. 2008. Estimation of an agedistribution with its confidence intervals using an iterative Bayesian pro-cedure and a bootstrap sampling approach. In Recent advances in paleo-demography: data, techniques, patterns. Jean-Pierre Bocquet-Appel, ed. Pp.63–82. New York: Springer.

Bocquet-Appel, Jean-Pierre, Pierre-Yves Demars, Lorette Noiret, and DmitryDobrowsky. 2005. Estimates of Upper Palaeolithic meta-population size inEurope from archaeological data. Journal of Archaeological Science 32:1656–1668.

Bocquet-Appel, Jean-Pierre, and Alain Tuffreau. 2009. Technological responsesof Neanderthals to macroclimatic variations (240,000–40,000 BP). HumanBiology 81:287–307.

Boserup, Ester. 1965. The conditions of agricultural growth. London: Allen &Unwin.

Briggs, Adrian W., Jeffrey M. Good, Richard E. Green, Johannes Krause,Tomislav Maricic, Udo Stenzel, Carles Lalueza-Fox, et al. 2009. Targetedretrieval and analysis of five Neandertal mtDNA genomes. Science325(5938):318–321.

Bruner, Emiliano, and Giorgio Manzi. 2006. Saccopastore 1: the earliest Ne-anderthal? a new look at an old cranium. In Neanderthals revisited: newapproaches and perspectives. Harvati Katerina and Terry Harrison, eds. Pp.23–36. Dordrecht: Springer.

Caspari, Rachel, and Sang-Hee Lee. 2004. Older age becomes common latein human evolution. Proceedings of the National Academy of Sciences of theUSA 101(30):10895–10900.

———. 2005a. Are OY ratios invariant? a reply to Hawkes and O’Connell.Journal of Human Evolution 49:654–659.

———. 2005b. Taxonomy and longevity: a reply to Minichillo. Journal ofHuman Evolution 49:646–649.

———. 2006. Is human longevity a consequence of cultural change or modernbiology? American Journal of Physical Anthropology 129(4):512–517.

Caussinus, Henri, and Daniel Courgeau. 2010. Estimer l’age sans le mesurer.Population 65(1):117–146.

Churchill, Steven E. 2007. Bioenergetic perspectives on Neandertal thermo-regulatory and activity budgets. In Neanderthals revisited: new approachesand perspectives. Harvati Katerina and Terry Harrison, eds. Pp. 113–133.Dordrecht: Springer.

Coale, Ansley J., Paul George Demeny, and Barbara Vaughan. 1983. Regional

S212 Current Anthropology Volume 54, Supplement 8, December 2013

model life tables and stable populations. Princeton, NJ: Princeton UniversityPress.

Conarda, Nicholas J., Michael Bolusc, and Susanne C. Munzeld. 2012. MiddlePaleolithic land use, spatial organization and settlement intensity in theSwabian Jura, southwestern Germany. Quaternary International 247:236–245.

Courtenay, Jackie, and Gigi Santow. 1989. Mortality of wild and captive chim-panzees. Folia Primatologia 52:67–177.

Crow, James F., and Motoo Kimura. 1970. An introduction to population ge-netics theory. New York: Harper & Row.

Cutler, Richard G. 1975. Evolution of human longevity and the genetic com-plexity governing aging rate. Proceedings of the National Academy of Sciencesof the USA 72(11):4664–4668.

Degioanni, Anna, Virginie Fabre, and Silvana Condemi. 2011. Genetique etpaleoanthropologie: deux approches pour un dialogue autour des Nean-dertaliens. Bulletins et Memoires de la Societe d’anthropologie de Paris 23(1):1–18.

d’Errico, Francesco, Maria Fernanda Sanchez-Goni, and Marian Vanhaeren2006. L’impact de la variabilite climatique rapide des OIS 3–2 sur le peu-plement de l’Europe. In L’homme face au climat. Paris: Odile Jacob.

Fabre, Virginie, Silvana Condemi, and Anna Degioanni. 2009. Genetic evi-dence of geographical groups among Neanderthals. PLoS ONE 4(4):e5151.

Felsenstein, Joseph. 1971. Inbreeding and variance effective numbers in pop-ulations with over-lapping generations. Genetics 168:581–597.

Finlayson, Clive, Darren A. Fa, Geraldine Finlayson, Francisco Giles Pacheco,Jose Marıa Gutierrez-Lopez, and Antonio Santiago Perez. 2001. Use of thelandscape by humans from the Middle Palaeolithic to the Neolithic: thecase of the northern shore of the Strait of Gibraltar. Almoraima 25:65–71.

Fisher, Ronald Aylmer. 1930. The genetic theory of natural selection. Oxford:Oxford University Press.

Frenzel, Burkhard. 1985. Die Umwelt des Menschen im Eiszeitalter. Quartar35/36:7–33.

Gamble, Clive. 1987. Man the shoveler: alternative models for Middle Pleis-tocene colonization in northern latitudes. In The Pleistocene old world: re-gional perspectives. Olga Soffer, ed. Pp. 81–98. New York: Plenum.

———. 1999. The Palaeolithic societies of Europe. Cambridge: Cambridge Uni-versity Press.

Ghirotto, Silvia, Francesca Tassi, Andrea Benazzo, and Guido Barbujani. 2011.No evidence of Neandertal admixture in the mitochondrial genomes ofearly European modern humans and contemporary Europeans. AmericanJournal of Physical Anthropology 146(2):242–252.

Gomez-Uchida, Daniel, Friso P. Palstra, Thomas W. Knight, and Daniel E.Ruzzante. 2013. Contemporary effective population and metapopulationsize (Ne and meta-Ne): comparison among three salmonids inhabiting afragmented system and differing in gene flow and its asymmetries. Ecologyand Evolution 3(3):569–580.

Goux, Jean-Michel. 1982. L’estimation du rapport de masculinite et de savariance en fonction de l’effectif. In 1er Congres international de paleontologiehumaine. Jean Piveteau and Henry de Lumley, eds. Pp. 993–996. Paris:Institute de paleontologie humaine.

Green, Richard E., Johannes Krause, Adrian W. Briggs, Tolmislav Maricic,Udo Stenzel, Martin Kircher, Nick Patterson, et al. 2010. A draft sequenceof the Neandertal genome. Science 328:710–722.

Green, Richard E., Johannes Krause, Susan E. Ptak, Adrian W. Briggs, MichaelT. Ronan, Jan F. Simons, Lei Du, et al. 2006. Analysis of one million basepairs of Neanderthal DNA. Nature 444(7117):330–336.

Green, Richard E., Anna-Sapfo Malaspinas, Johannes Krause, Adrian W.Briggs, Philip L. F. Johnson, Caroline Uhler, Matthias Meyer, et al. 2008.A complete Neandertal mitochondrial genome sequence determined byhigh-throughput sequencing. Cell 134:416–426.

Guatelli-Steinberg, Debbie. 2009. Recent studies of dental development inNeandertals: implications for Neandertal life histories. Evolutionary An-thropology 18:9–20.

Guatelli-Steinberg, Debbie, Donald J. Reid, Thomas A. Bishop, and ClarkSpencer Larsen. 2005. Anterior tooth growth periods in Neandertals werecomparable to those of modern humans. Proceedings of the National Acad-emy of Sciences of the USA 102(40):14197–14202.

Hammer, Michael F., and Stephen L. Zegura. 1996. The role of the Y chro-mosome in human evolutionary studies. Evolutionary Anthropology 5:116–134.

Harding, Rosalind M., Stephanie Malia Fullerton, Robert C. Griffiths, Jac-quelyn Bond, Martin J. Cox, Julie A. Schneider, Danielle S. Moulin, and J.

B. Clegg. 1997. Archaic African and Asian lineages in the genetic ancestryof modern humans. American Journal of Human Genetics 60:772–789.

Hawk, John. 2008. From genes to numbers: effective population sizes in hu-man evolution. In Recent advances in palaeodemography. Jean-Pierre Boc-quet-Appel, ed. Pp. 9–30. Dordrecht: Springer.

Hawkes, Kristen. 2006. Life history theory and human evolution: a chronicleof ideas and findings. In The evolution of human life history. Kristen Hawkesand Richard R. Paine, eds. Pp. 45–93. Oxford: Currey.

Hayden, Brian. 2012. Neandertal social structure? Oxford Journal of Archae-ology 31(1):1–26.

Henry, Amanda G., Alison S. Brooks, and Dolores R. Piperno. 2011. Micro-fossils in calculus demonstrate consumption of plant and cooked foods inNeanderthal diets (Shanidar III, Iraq; Spy I and II, Belgium). Proceedingsof the National Academy of Sciences of the USA 108(2):486–491.

Henry, Louis, and Yves Blayo. 1975. La population de la France de 1740 a1860. Population 30:71–122.

Hill, Kim, Christophe Boesch, Jane Goodall, Anne Pusey, Jennifer Williams,and Richard Wrangham. 2001. Mortality rates among wild chimpanzees.Journal of Human Evolution 40:437–450.

Hill, Kim Ronald, and A. Magdalena Hurtado. 1996. Ache life history: theecology and demography of a foraging people. New York: Aldine.

Hill, William G. 1972. Effective size of populations with overlapping gener-ations. Theoretical Population Biology 3:278–289.

Hofman, Michel A. 1993. Encephalization and the evolution of longevity inmammals. Journal of Evolutionary Biology 6:209–227.

Holliday, Trenton W. 1997. Body proportions in Late Pleistocene Europe andmodern human origins. Journal of Human Evolution 32:423–447.

Hoppa, Robert D., and James W. Vaupel. 2002. Paleodemography: age distri-butions from skeletal samples. Cambridge: Cambridge University Press.

Howell, Nancy. 1979. Demography of the Dobe !Kung. New York: AcademicPress.

Hublin, Jean Jacques. 1998. Climatic changes, paleogeography, and the evo-lution of the Neandertals. In Neandertals and modern humans in westernAsia. Akazawa Takeru, Kenichi Aoki, and Ofer Bar-Yosef, eds. Pp. 295–310.New York: Plenum.

Hublin, Jean Jacques, and Wil Roebroeks. 2009. Ebb and flow or regionalextinctions? on the character of Neandertal occupation of northern envi-ronments. Comptes rendus Palevol 8(5):503–509.

Isler, Karin, Christopher E. Kirk, Joseph M. A. Miller, Gene A. Albrecht, BruceR. Gelvin, and Robert D. Martin. 2008. Endocranial volumes of primatespecies: scaling analyses using a comprehensive and reliable data set. Journalof Human Evolution 55:967–978.

Jannetta, Ann Bowman, and Samuel H. Preston. 1991. Two centuries of mor-tality change in central Japan: the evidence from a temple death register.Population Studies 45:417–436.

Kendal, Jeremy, Jamshid J. Tehrani, and John Odling-Smee. 2011. Humanniche construction in interdisciplinary focus. Philosophical Transaction ofthe Royal Society B 366:785–792.

Krause, Johannes, Ludovic Orlando, David Serre, Bence Viola, Kay Prufer,Michael P. Richards, Jean-Jacques Hublin, Catherine Hanni, Anatoly P.Derevianko, and Svante Paabo. 2007. Neanderthals in central Asia andSiberia. Nature 449(7164):902–904.

Kremer, Michael. 1993. Population growth and technological change: onemillion BC to 1990. Quarterly Journal of Economics 108(3):681–716.

Krings, Matthias, Cristian Capelli, Frank Tschentscher, Helga Geisert, SonjaMeyer, Arndt von Haeseler, Karl Grossschmidt, Goran Possnert, Maja Pau-novic, and Svante Paabo. 2000. A view of Neandertal genetic diversity.Nature Genetics 26:144–146.

Krings, Matthias, Anne Stone, Ralf W. Schmitz, Heike Krainitzki, Mark Stone-king, and Svante Paabo. 1997. Neandertal DNA sequences and the originof modern humans. Cell 90:19–30.

Kuhn, Steven L., and Mary C. Stiner. 2006. What’s a mother to do? a hy-pothesis about the division of labor among Neandertals and modern hu-mans in Eurasia. Current Anthropology 47(6):953–980.

Kukla, George J., Michael L. Bender, Jacques-Louis de Beaulieu, Gerard Bond,Wallace S. Broecker, Piet Cleveringa, Joyce E. Gavin, et al. 2002. Last in-terglacial climates. Quaternary Research 58(1):2–13.

Kuznets, Simon. 1973. Modern economic growth: findings and reflections.American Economic Review 63(3):247–258.

Lalueza-Fox, Carles, Maria L. Sampietro, David Caramelli, Yvonne Puder,Martina Lari, Francesc Calafell, Cayetana Martinez-Maza, et al. 2005. Nean-dertal evolutionary genetics: mitochondrial DNA data from the Iberianpeninsula. Molecular Biology and Evolution 22:1077–1081.

Bocquet-Appel and Degioanni Neanderthal Demographic Estimates S213

Lucy, Dave, Robert. G. Ackroyd, and A. Mark. Pollard. 2002. Non-parametriccalibration for age estimation. Applied Statistics 52(2):185–196.

Lucy, Dave, Robert G. Ackroyd, A. Mark. Pollard, and Tore Solheim. 1996.A Bayesian approach to adult human age estimation from dental obser-vations by Johanson’s age changes. Journal of Forensic Sciences 41:189–194.

Mann, Alan. 1968. The paleodemography of Australopithecus. PhD disserta-tion, University of California, Berkeley.

Masset, Claude. 1973. La demographie des populations inhumees: essai depaleodemographie. L’Homme 13:95–131.

Mellars, Paul, and Jennifer C. French. 2011. Tenfold population increase inwestern Europe at the Neandertal-to-modern human transition. Science 333:623–627.

Miles, A. E. W. 1963. The dentition in the assessment of individual age inskeletal material. In Dental anthropology. Don Brothwell, ed. Pp. 191–209.Oxford: Pergamon.

———. 2001. The Miles method of assessing age from tooth wear revisited.Journal of Archeological Science 28:973–982.

Nei, Masatoshi, and Dan Graur. 1984. Extent of protein polymorphism andthe neutral mutation theory. Evolutionary Biology 27:73–118.

Neubauer, Simon, and Jean-Jacques Hublin. 2012. The evolution of humanbrain development. Evolutionary Biology 39(4):568–586.

Noonan, James P., Graham Coop, Sridhar Kudaravalli, Doug Smith, JohannesKrause, Joe Alessi, Feng Chen, et al. 2006. Sequencing and analysis ofNeanderthal genomic DNA. Science 314(5802):1113–1118.

Nunney, Leonard. 1993. The influence of mating system and overlappinggenerations on effective population size. Evolution 47(5):1329–1341.

Nunney, Leonard, and Diane. R. Elam. 1994. Estimating the effective popu-lation size of conserved populations. Conservation Biology (8):175–184.

Ovchinnikov, Iogor V., and Olga I. Kholina. 2010. Genome digging: insightinto the mitochondrial genome of Homo. PLoS ONE 5(12):e14278.

Patou-Mathis, Marie Helene. 2000. Neanderthal subsistence behaviours inEurope. International Journal of Osteoarchaeology 10(5):379–395.

Powell Adam, Stephen Shennan, and Mark G. Thomas. 2009. Late Pleistocenedemography and the appearance of modern human behavior. Science 324:1298–1301.

Ramirez-Rossi, Fernando V., and Jose Maria Bermudez de Castro. 2004. Sur-prisingly rapid growth in Neanderthals. Nature 428:936–939.

Rawlins, Richard G., and Matt J. Kessler. 1987. Demography of the free rangingCayo Santiago Macaques (1976–1983). In The Cayo Santiago Macaques.Richard G. Rawlins and Matt J. Kessler, eds. Pp. 48–72. Albany: StateUniversity of New York Press.

Ray, Nicolas. 2003. Modelisation de la demographie des populations humainesprehistoriques a l’aide de donnees environnementales et genetiques. PhDdissertation, Universite de Geneve.

Richard, Alison F. 1985. Primates in nature. New York: W. H. Freeman.Richards, Michael P., and Eric Trinkaus. 2009. Isotopic evidence for the diets

of European Neanderthals and early modern humans. Proceedings of theNational Academy of Science of the USA 1064(38):16034–16039.

Richerson, Peter J., Robert Boyd, and Robert L. Bettinger. 2009. Culturalinnovations and demographic changes. Human Biology 81(2):211–235.

Rogers, Alan R., and Lynn B. Jorde. 1995. Genetic evidence on modem humanorigins. Human Biology 67:1–36.

Ruff, Christopher B., Erik Trinkaus, and Trenton W. Holliday. 1997. Bodymass and encephalization in Pleistocene Homo. Nature 387:173–176.

Sacher, George A. 1959. Relation of lifespan to brain weight and body weightin mammals. In The lifespan of animals. Gordon Ethelbert Ward Wolsten-holme and Cecilia Mary O’Connor, eds. Pp. 115–133. CIBA FoundationColloquia on Ageing, vol. 5. Boston: Little-Brown.

———. 1975. Maturation and longevity in relation to cranial capacity inhominid evolution. In Primate functional morphology and evolution. RussellTuttle, ed. Pp. 417–441. The Hague: Mouton.

Sacher, George A., and Everett F. Staffeldt. 1974. Relation of gestation timeto brain weight for placental mammals: implications for the theory ofvertebrate growth. American Naturalist 108:593–615.

Sanchez Goni, Marıa Fernanda, Amaelle Landais, William J. Fletcher, FilipaNaughton, Stephanie Desprat, and Josette Duprat. 2008. Contrasting im-pacts of Dansgaard-Oeschger events over a western European latitudinal

transect modulated by orbital parameters. Quaternary Science Reviews27(11/12):1136–1151.

Serbezov, Dimitar, Per Erik Jorde, Louis Bernatchez, Esben Moland Olsen,and Leif Asbjørn Vøllestad. 2012. Life history and demographic determi-nants of effective/census size ratios as exemplified by brown trout (Salmotrutta). Evolutionary Applications 5:607–618.

Shea, John J. 2011. Neanderthal news: extinct species exhibits variability. Evo-lutionary Anthropology 20(5):198–200.

Shennan, Stephen. 2001. Demography and cultural innovation: a model andits implications for the emergence of modern human culture. CambridgeArchaeological Journal 11:5–16.

Simon, Julian. L. 1977. The economics of population growth. Princeton, NJ:Princeton University Press.

Smith, Tanya M., Michel Toussaint, Donald J. Reid, Anthony J. Olejniczak,and Jean-Jacques Hublin. 2007. Rapid dental development in a MiddlePaleolithic Belgian Neanderthal. Proceedings of the National Academy ofSciences of the USA 104(51):20220–20225.

Sorensen, M. V., and W. R. Leonard. 2001. Neandertal energetics and foragingefficiency. Journal of Human Evolution 40:483–495.

Steegmann, A. Theodore, Jr., Frank J. Cerny, and Trenton W. Holliday. 2002.Neandertal cold adaptation: Physiological and energetic factors. AmericanJournal of Human Biology 14:566–583.

Takahata, Naoyuki. 1993. Allelic genealogy and human evolution. MolecularBiology and Evolution 10:2–22.

Takahata, Naoyuki, Yoko Satta, and Jan Klein. 1995. Divergence time andpopulation size in the lineage leading to modem humans. Theoretical Pop-ulation Biology 48:198–221.

Teleki, Geza, Edward E. Hunt, and John-Henry Pfifferling. 1976. Demographicobservations (1963–1973) on the chimpanzees of Gombe National Park,Tanzania. Journal of Human Evolution 5:559–598.

Trinkaus, Erik. 1995. Neanderthal mortality patterns. Journal of ArchaeologicalScience 22:121–144.

———. 2011. Late Pleistocene adult mortality patterns and modern humanestablishment. Proceedings of the National Academy of Sciences of the USA108(4):1267–1271.

Trinkaus, Erik, and Robert L. Tompkins. 1990. The Neandertal life cycle: thepossibility, probability, and perceptibility of contrasts with recent humans.In Primate life history and Evolution. C. Jean De Rousseau, ed. Pp. 153–180. Monographs in Primatology 14. New York: Wiley-Liss.

van Andel, Tjeerd H., William Davies, and Bernard Weninger. 2003. Thehuman presence in Europe during the Last Glacial period. 1. Human mi-grations and the changing climate. In Neanderthals and modern humans inthe European landscape during the last glaciation: archaeological results of theStage 3 project. Tjeerd H. van Andel and William Davies, eds. Pp. 31–56.Cambridge: McDonald Institute for Archaeological Research.

van Andel, Tjeerd H., and P. Chronis Tzedakis. 1996. Palaeolithic landscapesof Europe and environs: 150,000–25,000 years ago: an overview. QuaternaryScience Reviews 15:481–500.

Whiteley, Andrew R., Jason A. Coombs, Mark Hudy, Zachary Robinson, KeithH. Nislow, and Benjamin H. Letcher. 2012. Sampling strategies for esti-mating brook trout effective population size. Conservation Genetics 13:625–627.

Wilson, Allan C., Rebecca L. Cann, Steven M. Carr, Matthew George, Ulf B.Gyllensten, Kathleen M. Helm-Bychowski, Russell G. Higuchi, et al. 1985.Mitochondrial DNA and two perspectives on evolutionary genetics. Bio-logical Journal of the Linnean Society 26:375–400.

Wood, James W. 1987. The genetic demography of the Gainj of Papua NewGuinea. 2. Determinates of effective population size. American Naturalist129(2):165–187.

———. 1998. A theory of pre-industrial population dynamics: demography,economy, and well-being in Malthusian systems. Current Anthropology 39:99–135.

Wright, Sewall. 1931. Evolution in Mendelian populations. Genetics 16:97–159.

———. 1938. Size of population and breeding structure in relation to evo-lution. Science 87:430–431.

S214 Current Anthropology Volume 54, Supplement 8, December 2013

� 2013 by The Wenner-Gren Foundation for Anthropological Research. All rights reserved. 0011-3204/2013/54S8-0005$10.00. DOI: 10.1086/673387

Agreements and Misunderstandingsamong Three Scientific Fields

Paleogenomics, Archaeology, and Human Paleontology

by Carles Lalueza-Fox

The emergence of paleogenomics (the study and analysis of ancient genomes) has provided a new, powerful sourceof information that can be used to test previous hypotheses regarding human evolution. However, various mis-understandings concerning the interpretation of genetic data in an archaeological and paleontological context andthe existence of different scientific goals tend to hinder the fluent and fruitful collaboration between these fields.Here we explore some of the subjects creating confusion, such as the problems associated with molecular clocks,the difference between sequence divergence and species divergence, and the limitations of the uniparental markers.Limited understanding of how the expression of a genome shapes the phenotype (including morphology andcognition) is the main obstacle to linking the genetic and the morphological evidence available. In the case ofNeanderthals (and probably Denisovans, too), it is obvious that the conspicuous morphological differences cannotbe explained by differences in a list of about 100 genes alone, thus suggesting that regulatory genomic elementsmust have been involved. A functional analysis of the genes involved as well as a study of the genomic architecture—a complexity level above the simple DNA message—could help us fill this gap. It is hoped that this future workwill lead to the emergence of an interrelated and multidisciplinary view of the study of the past based on realcollaborative efforts among disciplines.

Introduction

The interaction between archaeologists, paleontologists, andresearchers from the emerging field of paleogenomics hastraditionally been plagued by misunderstandings and a lackof collaborative efforts. Over the last three decades, molecularbiologists working on population analysis of human sampleshave usually tried to fit their results to hypotheses proposedpreviously on the basis of morphological or archaeologicalstudies. These hypotheses were often chosen at random fromthe available literature by the authors of these populationgenetics studies, who were clearly unfamiliar with the currentstate of the art in these other fields. Furthermore, the geneticresults themselves—especially with data, such as mitochon-drial DNA sequences, with limited phylogenetic resolvingpower—frequently did not allow the favoring of one hy-pothesis over another. For this reason, having possible supportfrom another field (paleontology, archaeology, even linguistics

Carles Lalueza-Fox is Director of the Paleogenomics Group atthe Institute of Evolutionary Biology (Consejo Superior deInvestigaciones Cientificas, Universitat Pompeu Fabra, Dr. Aiguader88, 08003 Barcelona, Spain [carles.lalueza@upf.edu]). This paper wassubmitted 3 VII 13, accepted 7 VIII 13, and electronically published8 XI 13.

or paleoclimatology) was seen as valuable and most-neededadditional evidence. This kind of subjective multidisciplinarymatch was very convenient for increasing the chances of pub-lication of those population genetic papers. However, manyof these analyses were in fact purely descriptive and providedvery limited insights into the genetic structure of past andextant human populations. As such, their real utility wouldbe in providing a context that with increasingly large numbersof samples could be used for testing competing explanationsin terms of past migrations and population affinities. Whenviewed from the other point of view, the behavior of archae-ologists and paleontologists has often been equally biased,sometimes ignoring widely accepted genetic results in a rathercondescending attitude of intellectual isolation. Many of theseresearchers seem to consider geneticists as newcomers to thestudy of the human past; this may be true, but they are hereto stay.

These disagreements between different fields can be par-tially explained by some limitations associated with the geneticmarkers typically employed as well as problems in our currentunderstanding of the relationship between genotype and phe-notype. Here, I would like to highlight some of these diffi-culties and suggest how they can be overcome. In the future,we can expect that an interrelated and multidisciplinary viewof our study of the past will be possible, and this can onlybe achieved with direct and real collaboration.

Lalueza-Fox Difficulties among Scientific Fields That Study the Past S215

Molecular Clocks, Sequence Divergence, andSpecies Divergence

We will start by discussing some of the problems associatedwith the evolutionary interpretation of the genetic data, asthey are usually the subject of misinterpretations by archae-ologists and paleontologists. The molecular clock hypothesisis based in the regularity of the mutation process in neutralgenetic regions along time, thus involving the possibility ofusing it as a time estimator for molecular evolution. Thereare, however, some problems with the accuracy of a molecularclock. First, the current genetic diversity (either a populationof study or a species) needs to be well characterized; second,the mutation rate needs to be known; third, we need to haveprecise dates to calibrate the clock (they are usually takenfrom the fossil record); and fourth, we need to work withselectively neutral genomic regions. All four factors can havetheir own limitations; for instance, there is conflicting evi-dence for estimating the mutation rates from family pedigreesand from evolutionary data (the former rate being usuallymuch faster than the later). Also, because of the existence ofubiquitous regulatory elements and undetected selectivesweeps, it is sometimes not so obvious that a particular ge-nomic region is neutrally evolving. Thus, it is not surprisingthat time estimates can always be subjected to refinementsand corrections. When we say, for instance, that the originof the Neanderthal mitochondrial DNA variation can be datedto about 110,000 years ago (Briggs et al. 2009), we are as-suming that the sampled Neanderthals are representative ofthe whole Neanderthal variation. If the next mitochondrialgenomes to be sequenced turn out to be more variable thanthe currently available ones, the “Neanderthal Eve” will bemoved dramatically back in time. Alternatively, if the futuremitochondrial genomes are quite similar to the previous ones,the date will not be significantly altered. Therefore, anotherobvious trait of the molecular clock date is its capacity ofbeing recalibrated depending on the sampling.

In the last decades, a new population genetics approachwith a strong mathematical base, know as coalescence, hasalso been developed. The coalescence theory allows us to gobackward in time from the existing genetic variation untilfinding its common ancestors, providing inferences on pop-ulation demography and genetic divergence. Another com-mon misunderstanding with other scientists dealing with thestudy of the past is the confusion between sequence diver-gence and species (or population) divergence. The coalescencetimes obtained always predate the real species divergence sim-ply because there is a certain genetic variation in any groupof individuals at any given time. The Italian geneticist GuidoBarbujani (Barbujani, Bertorelle, and Chikhi 1998:489) fa-mously illustrated this point with the following remark: “Sup-pose that some Europeans colonize Mars next year: if theysuccessfully establish a population, the common mitochon-drial ancestor of their descendants will be Paleolithic. But itwould not be wise for a population geneticist of the future

to infer from that a Paleolithic colonization of Mars.” There-fore, the smaller the ancestral population size, the closer se-quence divergence times and species divergence time wouldbe; but we have to keep in mind that both features do notneed to be coincident.

Mitochondrial DNA: Limitations ofUniparental Markers

Before the mass availability of genome-wide data, peopleworking on the genetics of human populations had to basetheir interpretations on single genetic loci, mainly uniparentalmarkers such as maternally inherited mitochondrial DNA(mtDNA) and the paternally inherited Y chromosome. Al-though it is frequently stated that mtDNA is just a single,uniparental genetic marker, the limitations of the mtDNA forinterpreting evolutionary processes are not fully recognizedin the population genetics literature. Because of the stochasticfactors associated with demography, some genetic markersmay reflect population or species history and some may not(Balloux 2010). This is related to lineage sorting—in otherwords, the process of gene-lineage fixation along an evolu-tionary process. Incomplete lineage sorting occurs when agene tree or genealogy differs from the species phylogeny, aphenomenon that produces conflicting phylogenies and non-monophyletic groups for a particular genetic marker. Theuniparental markers (mtDNA or Y chromosome) are greatlyaffected by these random processes. In this sense, a remarkablediscrepancy between mtDNA and nuclear DNA phylogenetictrees has been recently described for polar bears (Miller et al.2012). Of course, this variation in coalescence times alonggenetic markers would not be a problem if multiple nucleargenetic markers or even complete genomes could be gener-ated, as is increasingly the case even for extinct hominin spe-cies.

In this regard, the estimate divergence times for the sep-aration of Eurasian and African populations, generated fromthe observed mtDNA diversity, has yielded dates of !100,000years ago (almost always around 60,000–80,000 years ago),while the time depth for the African populations has neverbeen older than 200,000 years ago (the time of the so-calledmitochondrial Eve). Nevertheless, nuclear DNA divergencetimes obtained from complete genomes are estimated to bearound 600,000–800,000 years (Green et al. 2010; Reich et al.2010). It is probably oversimplistic to directly interpret thesemolecular clock dates from the point of view of a simplespeciation event because demographic events such as popu-lation fluctuations could have greatly affected the mitochon-drial as opposed to the nuclear genome diversity. In fact, nobottleneck is even needed. If the effective population size wasconstant at the time of the emergence of modern humans,the mtDNA would still coalesce at some point in the past(Weaver 2012). If this was the case, anterior demographicevents would have been “erased” and thus would be unde-tectable from the analysis of the mtDNA. The African thermal

S216 Current Anthropology Volume 54, Supplement 8, December 2013

conditions, unfavorable to DNA preservation, makes it likelythat no ancient DNA would ever be retrievable from speci-mens before the “mitochondrial Eve,” and thus we will needto rely more and more on the analysis of complete modernAfrican genomes.

Gene Flow from Archaic Hominins

One interesting example of a coalescent discrepancy betweenan mtDNA and the nuclear genome has been described inthe Denisova hominin (named after Denisova Cave in Siberia,Russia). The analysis of the complete mtDNA genome showedthat Neanderthals were the sister group of modern humansand that the Denisova lineage diverged from those of Ne-anderthals and modern humans about one million years ago(Krause et al. 2010). This tree was misinterpreted to representthe evolutionary relationships between these three species.However, subsequent analysis of the complete Denisova nu-clear genome produced a rather different picture in whichDenisova was now the sister group of Neanderthals, and bothlineages shared a common ancestor around 640,000 and withmodern humans around 804,000 years ago (Reich et al. 2010).The authors suggested that the Denisovan mtDNA lineagecould represent an archaic mtDNA that was introduced intoDenisovan ancestors by hybridization with some archaic hom-inin and subsequently preserved in the population by incom-plete lineage sorting. An additional individual from the samesite showed an almost identical mtDNA (Reich et al. 2010),thus demonstrating that this discordance between mtDNAand nuclear DNA is not restricted to a single Denisovan in-dividual. The authors suggested this kind of discordance isnot outside the range of what could be expected within apopulation. However, the subsequent Denisovan high-cov-erage genome (Meyer et al. 2012) showed this individual hada remarkably low heterozygosity (only ∼26%–33% of thatseen in modern Eurasians), which seems to indicate a verysmall population size.

While Denisovans could be descendants of a morpholog-ically unknown eastern form of hominin that inhabited largeareas of Asia while Homo neanderthalensis was mainly evolvingin Europe, the inferences drawn from the mtDNA alone, ifnuclear data were unavailable, would have been rather dif-ferent, pointing to a recent survival of more primitive hom-inin forms such as Homo erectus. A genomic comparison withmodern humans also found that modern Melanesians, butnot other non-African modern humans, share about 4.5% oftheir genomic regions with Denisova (Reich et al. 2010, 2011).If anything, the Denisova study shows that the emerging pic-ture of human evolution is one in which gene flow betweendifferent hominin populations (or species) was common. Asimilar result was found previously upon analysis of the Ne-anderthal genome in which non-African modern humansshare about 2.5% of their genomic regions with the former(Green et al. 2010). More recently, a genomic analysis in sub-Saharan populations has detected that about 2% of their ge-

nomes seems to be provided from yet another introgressionevent that took place around 35,000–50,000 years ago by con-tact with some archaic African lineage now extinct (Hammeret al. 2011). This archaic population, morphologically un-determined, would have split about 700,000 years ago fromthe lineage leading to the ancestors of modern humans (Ham-mer et al. 2011). Signals of admixture with Neanderthals havealso been detected in North African populations, probablyderiving from a back-to-Africa migration after the contact inthe Near East (Sanchez-Quinto et al. 2012). Thus, the complexevolutionary events that took place in Africa during the Mid-dle Stone Age are still being unraveled.

This is a crucial point, as some of the limitations of uni-parental markers chiefly arise when they are used to detectgene flow and hybridization events. We therefore need toredefine the use of mtDNA and the Y chromosome in humanevolution studies because they have clearly failed to detect thereal evolutionary processes that took place in the out-of-Africaexpansion of our species (see, e.g., Briggs et al. 2009; Kringset al. 1997; Serre et al. 2004). Moreover, thousands of modernhuman genomes from a large number of populations as wellas new ancient hominin genomes will be available in the nearfuture and will provide clearer answers to questions concern-ing the origin of our species than those obtained from uni-parental markers. The scientific time of the mtDNA and Ychromosome as the main tool to correlate with the fossilrecord is coming to an end.

Limitations of the First NeanderthalGenome Draft

Some people may think that not much was discovered aboutNeanderthals themselves after release of the first genome draft.In fact, it is easier to understand ourselves by comparing usto Neanderthals than to understand what makes a Neander-thal a Neanderthal (or a Denisovan a Denisovan; Lalueza-Foxand Gilbert 2011). This problem is derived from the lowgenomic coverage of the first draft (Green et al. 2010). Witha 1.3# coverage, if a particular read has an ancestral nucle-otide in a position where modern humans have a fixed, de-rived nucleotide, it is likely that this read is neither the productof an unknown, chimpanzee-like contamination nor post-mortem damage (fig. 1). However, those positions wheremodern humans have a fixed ancestral variant and Neander-thals a derived one are more difficult to validate (fig. 2). Inthis case, a Neanderthal read may harbor a novel variant, butit could also be simply due to damage or sequencing error.Of course, the damage tends to be template specific; therefore,increasing the coverage should make it possible to track thosegenes that have been modified in the Neanderthal evolution-ary lineage only (Lalueza-Fox and Gilbert 2011). The analysisof segregating loci in Neanderthals suffers from a similarshortcoming. Thus, with the current low coverage, it is im-possible to distinguish random damage (Briggs et al. 2007;Hofreiter et al. 2001) and/or background contamination (see,

Lalueza-Fox Difficulties among Scientific Fields That Study the Past S217

Figure 1. Only read of the Neanderthal genome draft (from Vi33.26) at the MC1R gene between positions 885 and 954. Thisillustrates the limitations of investigating Neanderthal-specific variants and also the heterozygosity using a low-coverage draft. TheVindija read does not have the Neanderthal-specific guanine substitution at nt 919 described in Monti Lessini and Sidron 1252.However, it could have a nondescribed, synonymous substitution at nt 942 because A to G substitutions are not known to beassociated with postmortem damage. However, only the genotyping of this position in additional Neanderthal samples or an increasedcoverage will allow us to confirm this new, Neanderthal-specific genetic variant.

e.g., Green et al. 2006; Wall and Kim 2007) from heterozy-gosity. Indeed, this can only be achieved with genome cov-erages of around 15–20#, something that is technically pos-sible but exceedingly expensive in most cases (Denisovan andNeanderthal specimens from Denisova cave, by now at 30–50# coverage, are a remarkable exception). Alternatively, tar-geted methods can be designed to retrieve a specific geneticmarker several times, as was the case for the ABO blood groupand bitter taste gene from two Neanderthal specimens (Lal-ueza-Fox et al. 2008, 2009, 2011). In any case, our under-standing of the genomic diversity of these extinct homininswill, it is hoped, be improved with a high-coverage genome.

Beyond the Genome

Amino acid positions in about 80 genes have been found todiffer between Neanderthals and modern humans by com-paring both genomes with those of chimpanzees (Burbano etal. 2010; Green et al. 2010). With an increased coverage thisfigure will likely increase to around 100 genes, as in the caseof Denisova (Meyer et al. 2012). These positions correspondto those where Neanderthals share the ancestral genetic var-iant with the chimpanzees but modern humans display afixed, derived variant. This is, of course, suggestive of func-tional differences in these genes (although it is not always thecase because the resulting proteins can have a similar efficiencyalbeit with some amino acid changes in the underlying genes).This list contains genes whose exact function remains un-known in most cases, and only when malfunction is somehowpresent (usually in the form of deleterious mutations) doesa particular disease emerge as a consequence. However, these80 preliminary genes include several associated with metab-olism, physiology, and cognition and some with more preciseroles, such as being involved in the movement of sperm, theexpression and development of follicle hairs in the skin, orsome olfactory receptors (Green et al. 2010). While this in-formation offers an exceptional opportunity to create a listof genes shaped by recent selection in modern humans andthus genes modeled by the common meaning of humankind,it would be a mistake to believe that phenotypical differencesbetween Neanderthals and modern humans can be explainedby this short genetic list alone (Lalueza-Fox and Gilbert 2011).

Some years ago, the publications of the human (2001) and

chimpanzee genomes (2006) failed to fulfill our expectationsin terms of being able to understand the genetic basis of theconspicuous morphological (and cognitive) differences thatexist between these two species. Indeed, those people whoassumed that a quick look at the genetic differences wouldprovide an easy answer to the evolutionary processes in-volved—for instance, in key hominin adaptations such asbipedalism or brain size and complexity—were certainly dis-appointed. The problem resides in both the difficulties inunderstanding gene function and also in the complexity ofthe genome operating above the simple DNA level. To startwith, what was once called “junk DNA” was found to befunctional even though these genetic regions do not code forany protein. This is partly due to the existence of many reg-ulatory elements that interact with networks of genes, thusshaping the final organism resulting from expression of thegenome (Carroll 2008). In other words, similar or even iden-tical genomes could produce different phenotypes as a resultof differences in the regulation of gene transcription (theprocess by which DNA makes RNA, the molecule from whichproteins are subsequently generated).

An Example of a Regulatory Element:microRNA

There are many types of genomic regulatory elements, in-cluding the so-called microRNAs (miRNA). The miRNAs aresmall, noncoding RNAs with a length of 19–25 nucleotidesin their mature form that act as posttranscriptional regulatorsof gene expression by acting on the DNA transcripts. It isestimated that miRNAs regulate more than 30% of all protein-coding genes, building complex regulatory networks that con-trol almost every cellular process. One set of such miRNAsis present only in present-day humans and is thus a goodcandidate for having contributed to human-specific pheno-types. The discovery of one miRNA, namely miR-1304, thatdiffers between two closely related species such as modernhumans and Neanderthals is of special interest: modern hu-mans carry what seems to be a fixed substitution, whereasNeanderthals present the ancestral allele in a nucleotide thatis located just in the seed region of miRNA-1304 and is there-fore likely to alter the spectrum of target genes for miR-1304(Green et al. 2010).

S218 Current Anthropology Volume 54, Supplement 8, December 2013

Figure 2. Single read of Vi33.25 at position 735 of the RPTN gene (chr1 position 150393996) showing the ancestral (e.g., sharedwith the chimpanzee) variant C instead of the T fixed in modern humans. Even in light of the problems associated with lowcoverage, we can be reasonably certain about reads showing ancestral status because no chimpanzee contamination can be expected.Thus, and maybe a bit paradoxically, the low-coverage Neanderthal genome draft is more useful for determining modern human-specific changes than Neanderthal-specific changes.

The genomic search for target genes for this ancestral miR-1304 has shown an increase of more than 15 times the numberof putative targets ( ) for the human miRNA, thusN p 515indicating an important functional evolution for miR-1304.The 36 predicted targets for Neanderthal miR-1304 includetwo important genes for teeth formation, namely enamelinand amelotin (Lopez-Valenzuela et al. 2012), and miRNAoverexpression experiments using a luciferase-based assayconfirmed that the ancestral version of miR-1304 greatly re-duces enamelin- and amelotin-associated reporter gene ex-pression by 50% (Lopez-Valenzuela et al. 2012). Interestingly,other genes in the Neanderthal miR-1304 list include cog-nitive genes such as TCF4 (associated with neuropsychiatricdisorders such as schizophrenia and impaired verbal learning)or CD24 (associated with multiple sclerosis).

Although it is difficult to determine how this down reg-ulation would affect the individual phenotype, it is knownthat the volume of coronal dentine in Neanderthal molars islarger than in modern humans. Because the absolute volumeis similar for both hominin groups, this results in significantlythinner cuspal enamel in Neanderthals than in recent humans(Macchiarelli et al. 2006). Thus, although the ameloblast se-cretion rates are similar, the enamel cusp forms faster in Ne-anderthals than in modern humans (Smith et al. 2010). An-other difference is found in the ameloblastic activity asreflected in the periodicity of long-period lines in the enamel(Retzius lines or perikymata; Aiello and Dean 1990). Severalstudies on dental growth have shown that the ontogeny inmost Neanderthal dentitions examined was more rapid thanthat of Homo sapiens individuals, either recent or fossil (Smithet al. 2010). As a result, current dental eruption tables sys-tematically overestimate the age of Neanderthal individualsat death while accurately predicting those of fossil H. sapiens(Smith et al. 2010).

However, the data generated from the 1,000 Genomes Proj-ect (released in 2011) have shown that the derived miR-1304,which was previously thought to be fixed in modern humans,is not; intriguingly, about 5%–7% of Asian individuals sharethe ancestral miR-1304 version with Neanderthals (Lopez-Valenzuela et al. 2012). This distribution schema fits themodel of genetic introgression from archaic to modern hu-mans as proposed in a recent study of certain alleles of HLAgenes (Abi-Rached et al. 2011), although it could also be theresult of selective sweeps within recent human populations,

which would be compatible with a beneficial role for the newderived miR-1304 allele. Because of the relatively recent di-vergence dates between Neanderthal and modern human ge-nomes (around 800,000 years), it is perhaps unrealistic toexpect to find many fixed differences between both humangroups, and even functionally important differences could beexpected to segregate to some extent in both lineages. It couldbe that the conspicuous phenotypic differences among ancienthuman lineages are due to the summatory effect of a particularcombination of genetic variants even if some of them seg-regate at low frequencies. If anything, the miRNA analysisagain shows the complexity involved in unraveling the humanevolutionary process.

Further functional studies could help our understandingof the link between regulation of the expression of genesassociated with enamel formation and the final teeth mor-phology. In any case, this is a nice example of what can beexpected in the future in terms of a view of the genomicarchitecture that goes beyond the simple reading of a DNAmessage.

Convergent Evolution

Another problem associated with our current lack of knowl-edge regarding the link between genotype and phenotype isthe analysis of possible convergent evolutionary traits in hom-inin species. This is related to what is known as the evolutionof “evolvability,” that is, the limited physical and even chem-ical possibilities of a body design to create restrictions on thepotential evolution of particular lineages. In the case of hom-inins, this could mean that only one set of adaptive traits canemerge with time, although it could also mean that similartraits are likely to appear independently in different homininlineages.

One interesting example of this was described in the MC1Rgene from Neanderthals (Lalueza-Fox et al. 2007). This genecodes for a protein in the membrane of melanocytes thatregulates the synthesis of two different pigments in the hairand skin: the dark, brownish eumelanine and the fair, reddishpheomelanine. A Neanderthal-specific variant was found toproduce a loss of function in the MC1R protein, thus resultingin fair skin and red hair in those Neanderthals carrying thisvariant. As in modern humans, it is likely that being hetero-zygous or homozygous for this particular mutation would

Lalueza-Fox Difficulties among Scientific Fields That Study the Past S219

produce phenotypes ranging from blond-reddish to “flame”red hair (Lalueza-Fox et al. 2007). However, it is worth em-phasizing that the Neanderthal mutation is not found in pres-ent modern humans; therefore, living red-haired peoplewould have a similar phenotype but for different genetic rea-sons (e.g., different mutations in the same gene). While thisis somehow an anecdotic phenotypic trait that is marginallyassociated to adaptation to high latitudes, it is possible thatother traits related to morphology and cognition could alsobe subjected to similar convergent processes. For instance, theincreased cranial capacity and brain organization that lead tocomplex human cognitive functions could, to some extent,have evolved in parallel along different hominin lineages. Thiscould explain the existence of aspects of modern symbolicbehavior in Neanderthals well before the arrival of modernhumans in Europe (Peresani et al. 2011; Zilhao et al. 2010).

It is likely that convergent evolution runs along particulargene networks that allow recurrent genetic modifications andthat thus trigger the repeated opportunity for natural selectionin the traits involved. This could explain underlying similar-ities in the phenotypical traits present in different homininlineages. These traits, known as “orthologous phenotypes” or“phenologs,” are defined as phenotypes related by the or-thology of the associated genes in two different species(McGary et al. 2010). Although MC1R is an example of con-vergent evolution at a protein-function level, it is likely thatthis phenomenon could be more prevalent in regulatory cir-cuits and could affect genes that tend to cluster together dur-ing the evolutionary process. The importance of convergentevolution in ancient hominins could be further explored whenmore is known about the precise genomic basis of specifichuman traits that can be observed in the fossil record.

Future Directions

The genetic basis of many of the traits observed in the fossilrecord is still unknown. This is partly due to the complexityof these traits but also to the problems associated with work-ing with functional genomics, the branch that studies thebiological function of the genes and their associated proteins.The study of most of the phenotypic traits will require theuse of animal models (for instance the creation of transgenicmice) and molecular techniques not used by paleogeneticists.An even higher level of multidisciplinary effort will thereforebe needed in the future.

Once this information becomes available, it will be possibleto check the genomic regions involved directly in the phe-notypic expression of extinct hominin genomes (speciallythose phenotypic traits that could be traced in the fossil rec-ord), and thus we will be able to finally understand some ofthe key issues of human evolution.

However, this complex enterprise can only be achieved withmultidisciplinary teams and real collaborations among ge-neticists, archaeologists, and paleontologists. A profound un-derstanding of the limitations and advantages of each disci-

pline through interdisciplinary trainings and meetings willallow different hypotheses to be tested from the availableevidence. More and more, all the disciplines studying the pastwill contribute from their own fields to the building of robustparadigms, and the current misunderstandings will, it is tobe hoped, fade away.

Acknowledgments

I am grateful to colleagues and to two anonymous reviewerswho provided constructive comments on the manuscript forthis article. My research is supported by a grant (BFU2012-34157) from the Ministerio de Economıa y Competitividadof Spain.

References CitedAbi-Rached, Laurent, Matthew J. Jobin, Subhash Kulkarni, Alasdair Mc-

Whinnie, Klara Dalva, Loren Gragert, Farbod Babrzadeh, et al. 2011. Theshaping of modern human immune systems by multiregional admixturewith archaic humans. Science 334(6052):89–94.

Aiello, Leslie C., and Christopher M. Dean. 1990. The microanatomy anddevelopment of teeth. In An introduction to human evolutionary anatomy.Pp. 106–132. London: Academic Press.

Balloux, Francois. 2010. Mitochondrial phylogeography: the worm in the fruitof the mitochondrial DNA tree. Heredity 104:419–420.

Barbujani, Guido, Giorgio Bertorelle, and Lounes Chikhi. 1998. Evidence forPaleolithic and Neolithic gene flow in Europe. American Journal of HumanGenetics 62(2):488–491.

Briggs, Adrian W., Jeffrey M. Good, Richard E. Green, Johannes Krause,Tomislav Maricic, Udo Stenzel, Carles Lalueza-Fox, et al. 2009. Targetedretrieval and analysis of five Neandertal mtDNA genomes. Science 325:318–321.

Briggs, Adrian W., Udo Stenzel, Philip L. F. Johnson, Richard E. Green, JanetKelso, Kay Prufer, Matthias Meyer, et al. 2007. Patterns of damage in ge-nomic DNA sequences from a Neandertal. Proceedings of the National Acad-emy of Sciences of the USA 104(37):14616–14621.

Burbano, Hernan A., Emily Hodges, Richard E. Green, Adrian W. Briggs,Johannes Krause, Matthias Meyer, Jeffrey M. Good, et al. 2010. Targetedinvestigation of the Neandertal Genome by array-based sequence capture.Science 328:723–725.

Carroll, Sean B. 2008. Evo-devo and an expanding evolutionary synthesis: agenetic theory of morphological evolution. Cell 134(1):25–36.

Green, Richard E., Johannes Krause, Adrian W. Briggs, Tomislav Maricic, UdoStenzel, Martin Kircher, Nick Patterson, et al. 2010. A draft sequence ofthe Neandertal genome. Science 328:710–722.

Green, Richard E., Johannes Krause, Susan E. Ptak, Adrian W. Briggs, MichaelT. Ronan, Jan F. Simons, Lei Du, et al. 2006. Analysis of one million basepairs of Neanderthal DNA. Nature 444:330–336.

Hammer, Michael F., August E. Woerner, Fernando L. Mendez, Joseph C.Watkins, and Jeffrey D. Wall. 2011. Genetic evidence for archaic admixturein Africa. Proceedings of the National Academy of Sciences of the USA 108(37):15123–15128.

Hofreiter, Michael, Viviane Jaenicke, David Serre, Arndt von Haeseler, andSvante Paabo. 2001. DNA sequences from multiple amplifications revealartifacts induced by cytosine deamination in ancient DNA. Nucleic AcidsResearch 29:4793–4799.

Krause, J., Q. Fu, J. M. Good, B. Viola, M. V. Shunkov, A. P. Derevianko, andS. Paabo. 2010. The complete mitochondrial DNA genome of an unknownhominin from southern Siberia. Nature 464(7290):894–897.

Krings, Matthias, Anne Stone, Ralf W. Schmitz, Heike Krainitzki, Mark Stone-king, and Svante Paabo. 1997. Neandertal DNA sequences and the originof modern humans. Cell 90:19–30.

Lalueza-Fox, Carles, Elena Gigli, Marco de la Rasilla, Javier Fortea, and An-tonio Rosas. 2009. Bitter-taste perception in Neandertals through the anal-ysis of TAS2R38 gene. Biology Letters 5:809–811.

Lalueza-Fox, Carles, Elena Gigli, Marco de la Rasilla, Javier Fortea, Antonio

S220 Current Anthropology Volume 54, Supplement 8, December 2013

Rosas, Jaume Bertranpetit, and Johannes Krause. 2008. Neandertal paleo-genomics in the ABO blood group gene. BMC Evolutionary Biology 8:342.

Lalueza-Fox, Carles, and M. Thomas P. Gilbert. 2011. Paleogenomics of archaichominins. Current Biology 21(24):R1002–R1009.

Lalueza-Fox, Carles, Holger Rompler, David Caramelli, Claudia Staubert, Giu-lio Catalano, David Hughes, Nadine Rohland, et al. 2007. A melanocortin1 receptor allele suggests varying pigmentation among Neanderthals. Science318:1453–1455.

Lalueza-Fox, Carles, Antonio Rosas, Almudena Estalrrich, Elena Gigli, PaulaF. Campos, Antonio Garcıa-Tabernero, Samuel Garcıa-Vargas, et al. 2011.Genetic evidence for patrilocal mating behaviour among Neandertal groups.Proceedings of the National Academy of Sciences of the USA 108(1):250–253.

Lopez-Valenzuela, Marıa, Oscar Ramırez, Antonio Rosas, Samuel Garcıa-Var-gas, Marco de la Rasilla, Carles Lalueza-Fox, and Yolanda Espinosa-Parrilla.2012. An ancestral allele of miR-1304 present in Neanderthals regulatesgenes involved in enamel formation and could explain dental differenceswith modern humans. Molecular Biology and Evolution 29(7):1797–1806.

Macchiarelli, Roberto, Luca Bondioli, Andre Debenath, Arnaud Mazurier,Jean-Francois Tournepiche, Wendy Birch, and M. Christopher Dean. 2006.How Neanderthal molar teeth grew. Nature 444:748–751.

McGary, Kriston L., Tae Joo Park, John O. Woods, Hye Ji Cha, John B.Wallingford, and Edward M. Marcotte. 2010. Systematic discovery of non-obvious human disease models through orthologous phenotypes. Proceed-ings of the National Academy of Sciences of the USA 107(14):6544–6549.

Meyer, Matthias, Martin Kircher, Marie-Theres Gansauge, Heng Li, FernandoRacimo, Swapan Mallick, Joshua G. Schraiber, et al. 2012. A high-coveragegenome sequence from an archaic Denisovan individual. Science 338:222–226.

Miller, Webb, Stephan C. Schuster, Andreanna J. Welch, Aakrosh Ratan, OscarC. Bedoya-Reina, Fangqing Zhao, Hie Lim Kim, et al. 2012. Polar andbrown bear genomes reveal ancient admixture and demographic footprintsof past climate change. Proceedings of the National Academy of Sciences ofthe USA 109(36):E2382–E2390.

Peresani, Marco, Ivana Fiore, Monica Gala, Matteo Romandini, and Antonio

Tagliacozzo. 2011. Late Neandertals and the intentional removal of feathersas evidenced from bird bone taphonomy at Fumane Cave 44 ky B.P., Italy.Proceedings of the National Academy of Sciences of the USA 108(10):3888–3893.

Reich, David, Richard E. Green, Martin Kircher, Johannes Krause, Nick Pat-terson, Eric Y. Durand, Bence Viola, et al. 2010. Genetic history of anarchaic hominin group from Denisova Cave in Siberia. Nature 468(7327):1053–1060.

Reich, David, Nick Patterson, Martin Kircher, Frederick Delfin, MadhusudanR. Nandineni, Irina Pugach, Albert Min-Shan Ko, et al. 2011. Denisovaadmixture and the first modern human dispersals into southeast Asia andOceania. American Journal of Human Genetics 89(4):516–528.

Sanchez-Quinto, Federico, Laura R. Botigue, Sergi Civit, Conchita Arenas,Marıa Carmen Avila-Arcos, Carlos D. Bustamante, David Comas, and CarlesLalueza-Fox. 2012. North African populations carry the signature of ad-mixture with Neandertals. PLoS ONE 7(10):e47765.

Serre, David, Andre Langaney, Mario Chech, Maria Teschler-Nicola, MajaPaunovic, Philippe Mennecier, Michael Hofreiter, Goran Possnert, andSvante Paabo. 2004. No evidence of Neandertal mtDNA contribution toearly modern humans. PLoS Biology 2(3):1–5.

Smith, Tanya M., Paul Tafforeau, Donald J. Reid, Joane Pouech, VincentLazzari, John P. Zermeno, Debbie Guatelli-Steinberg, et al. 2010. Dentalevidence for ontogenetic differences between modern humans and Nean-derthals. Proceedings of the National Academy of Sciences of the USA 107:20923–20928.

Wall, Jeffrey D., and Sung K. Kim. 2007. Inconsistencies in the Neanderthalgenomic DNA sequences. PLoS Genetics 3(10):e175.

Weaver, Timothy D. 2012. Did a discrete event 200,000–100,000 years agoproduce modern humans? Journal of Human Evolution 63:121–126.

Zilhao, Joao, Diego E. Angelucci, Ernestina Badal-Garcıa, Francesco d’Errico,Floreal Daniel, Laure Dayet, Katerina Douka, et al. 2010. Symbolic use ofmarine shells and mineral pigments by Iberian Neandertals. Proceedings ofthe National Academy of Sciences of the USA 107(3):1023–1028.

Current Anthropology Volume 54, Supplement 8, December 2013 S221

� 2013 by The Wenner-Gren Foundation for Anthropological Research. All rights reserved. 0011-3204/2013/54S8-0006$10.00. DOI: 10.1086/673503

Hominin Evolution in theMiddle-Late Pleistocene

Fossils, Adaptive Scenarios,and Alternatives

by Osbjorn M. Pearson

Hominins from Europe and Africa shed light on functional adaptations and other aspects of lifeways during theMiddle Paleolithic. By the end of that time span, Neanderthals and modern humans clearly differed physically andperhaps behaviorally. Explanations of the anatomical differences have largely focused on adaptation (directionalselection) to climate and habitual activity, but it is hard to rule out the alternative of genetic drift. Drift would haveaccelerated during periods of low population numbers, while selection operates best when populations are largeand expanding. Demographic changes almost certainly tracked climatic conditions in both continents. Environmentaland genetic data suggest that European hominins were primarily shaped by drift, while both factors operated inAfrica.

The period of time between 250 and 35 ka witnessed theemergence of Neanderthals in western Eurasia, modern hu-mans in Africa, and, at around 60 ka, the spread of modernhumans into Eurasia, where they replaced archaic humans,albeit with a small amount of interbreeding. Interpretationsof these events have tended to focus on different anatomicaland cultural adaptations as the key underlying forces respon-sible for producing the differences between modern humansand Neanderthals. The alternative, that genetic drift drovesome or perhaps many of the anatomical changes, has longbeen recognized (Howell 1957) but has received less emphasis.

The time is ripe for a reconsideration of scenarios for adap-tive change because of the accumulation of a critical mass ofnew evidence from paleoecology, genetics, anatomy, and chro-nology. Paleoclimatic records provide insights into why atleast some of the morphological and genetic evolution mayhave occurred. In this paper, I argue that climate and pop-ulation genetics are linked. Climate affects ecological pro-ductivity and biomass, which in turn affects human popu-lation numbers. Changes in population size have predictableconsequences for the expected rate of neutral genetic change.

The general outline of the evolution of modern humansand Neanderthals is well known (Arsuaga 2010; Arsuaga etal. 1997; Hublin 2009; Martınon-Torres et al. 2012; Stringer

Osbjorn M. Pearson is Associate Professor in the Department ofAnthropology at the University of New Mexico (MSC 01-1040,Albuquerque, New Mexico 87131, U.S.A. [ompear@unm.edu]). Thispaper was submitted 3 VII 13, accepted 20 VIII 13, and electronicallypublished 22 XI 13.

2007, 2011). Both populations diverged from a common an-cestor around 350,000 years ago as gauged by both geneticdifferences (Green et al. 2010) and divergence in cranial di-mensions modeled as the result of neutral evolution (Weaver,Roseman, and Stringer 2008).1 There is less agreement aboutthe deeper phylogeny of these lineages and related forms fromthe late Lower through early Middle Pleistocene, but thatperiod predates the central focus of this paper. Key anatomicaldifferences between Neanderthals and modern humans in-clude both the differential retention of primitive features ineach lineage as well as new features (apomorphies) in each.

In Europe, the Neanderthal lineage evolved a series of apo-morphies, including midfacial prognathism, a posterior po-sition of the mental foramen, a retromolar gap in the man-dible, a broad suprainiac fossa that is oval in form, a largejuxtamastoid process coupled with a small mastoid process,an occipital bun, double-arched browridges that are reducedin absolute volume and vertical thickness compared withthose of Middle Pleistocene hominins, and a substantiallylarger brain than those of most Middle Pleistocene hominins

1. Recent papers have produced a range of estimates for when theancestors of Neanderthals and modern humans split, ranging from ∼835ka for the average divergence for autosomal sequences (Green et al. 2010)to estimates of 270–400 ka for the population (rather than DNA se-quence) divergence time (Green et al. 2010) based on the same data.Other authors have calculated additional estimates for the divergencetimes between Neanderthals, Denisovans, and modern humans (e.g., Har-ris and Nielsen 2013; Li, Mulliken, and Reich 2010; Meyer et al. 2012;Reich et al. 2010). It is important to note that estimates of DNA diver-gence dates generally precede (often substantially) estimates of populationdivergence.

S222 Current Anthropology Volume 54, Supplement 8, December 2013

(Hublin 2009; Stringer 2007). Hublin (2009) considers thatthe full set of Neanderthal features were present by oxygenisotope stage (OIS) 7,2 and 3-D morphometric analyses of theface, temporal bone, and posterior cranial vault corroboratethis view (Harvati, Hublin, and Gunz 2010). This suite ofNeanderthal features had become common in European hom-inins by OIS 5, including the specimens from Krapina andSaccopatore, and they became even more frequent in OIS 4–3. This gradual increase in similarity to late Neanderthals hasbeen dubbed the “accretion model” and may have unfoldedas a single, long trend or perhaps in two pulses (Hublin 2009).A similar pattern applies to the evolution of Neanderthalpostcranial morphology (Trinkaus 1983, 2006). Scenarios forthe evolution of Neanderthal postcranial morphology havetended to emphasize adaptation to either the need to producelarge amounts of physical force or to preserve heat in a coldclimate.

The problem of whether Sima de los Huesos is young (ca.350 ka) or old (500–600 ka) complicates scenarios for thepace of evolutionary change in Europe (Stringer 2012). TheSima de los Huesos sample shows mosaics of Neanderthaland non-Neanderthal morphology in virtually all aspects ofits morphology (Arsuaga et al. 1997). Recently Martinon-Torres and colleagues (2012) have shown that the sample hasvery Neanderthal-like teeth; some of the nonmetrical featuresare even more common in the Sima de los Huesos samplethan in late, “classic” Neanderthals from OIS 4 to OIS 3,which casts doubt on simple models of a steady increase inNeanderthal features over time. Recent alternative hypothesesthat accept the greater antiquity for Sima de los Huesos haveproposed the presence of two lineages in Europe until 300–400 ka (Garcıa and Arsuaga 2011) or complicated scenariosof local extinction, recolonization, and admixture of two ormore populations (Dennell, Martinon-Torres, and Bermudezde Castro 2011).

While Neanderthals evolved in Europe, hominins in Africaevolved gradually toward a modern form (Brauer 2008; Pear-son 2008; Rightmire 2009). Modern human apomorphies in-clude a larger brain than generally observed in African craniadating to 300 ka or before; a more globular cranium (Lie-berman 2011; Lieberman, McBratney, and Krovitz 2002) withmore bossed parietals and an enlarged temporal lobe (Lie-berman 2011); an altered trajectory of endocranial growthrelative to Neanderthals (Gunz et al. 2012); a vertically shortface tucked beneath the frontal lobe (Lieberman 2011); re-tention of a canine fossa into adulthood; and the presence ofa projecting chin on the mandible (Stringer 2002, 2007). Theearliest crania that demonstrate the full suite of modern fea-

2. Both ice cores from Greenland and Antarctica as well as deep-seacores (and other data) help to reconstruct long-term patterns of climatechange. Many authorities prefer the term “marine isotope stage” for thissequence because the marine sequence is the longest and most complete,but in light of the importance of ice cores in illuminating the last 300kyr, I have used the older and more inclusive oxygen isotope stage (OIS)throughout this paper.

tures are Omo I, dated to 195 ka (Brown, McDougall, andFleagle 2012; Day and Stringer 1991; McDougall, Brown, andFleagle 2005), and the Herto crania at 150–160 ka (Clark etal. 2003; White et al. 2003).

Advances in imaging, especially synchrotron x-rays, whichallow researchers to peer inside teeth and count daily incre-ments of enamel accretion (Smith and Tafforeau 2008), haverevealed that Neanderthal children matured more rapidly thanmodern children (Smith et al. 2010). Neanderthals thus beara closer resemblance to the ancestral condition of even morerapid dental skeletal maturation of Homo erectus (Dean andSmith 2009; Dean et al. 2001; Graves et al. 2010). In contrast,the mandibular dentition of Jebel Irhoud 3, a juvenile latearchaic hominin from Morocco dating to 160 ka with affinitiesto modern humans (Hublin 2001; Hublin and Tillier 1981),preserves evidence of a slower, modern pace of dental de-velopment (Smith et al. 2007).

Both African and European Middle Pleistocene homininstended to be medium to tall in stature (Carretero et al. 2012)and very heavy for height relative to modern hunters andgatherers (Churchill et al. 2012; Kappelman 1996). A moreslender physique typified Omo I from Africa (Pearson et al.2008) and the early modern humans from Skhul and Qafzehin Israel (Carretero et al. 2012; Ruff, Trinkaus, and Holliday1997). Neanderthals remained at least 20% heavier relative tomodern human foragers of similar height (Kappelman 1996;Ruff, Trinkaus, and Holliday 1997). Carretero et al. (2012)proposed that this reduction in body mass may have been anevolutionary adaptation to a lifestyle that favored energy ef-ficiency. The stature of early modern humans from theLevalloiso-Mousterian of the Levant and the Gravettian ofEurope is particularly striking relative to Neanderthals andalmost all other samples from Europe before the twentiethcentury (Carretero et al. 2012).

Adaptation or Drift?

Over the last 50 years, the dominant view of the differencesbetween Neanderthals and modern humans has been that thedissimilarities in anatomy reflected adaptive differencesshaped by natural selection to meet specific challenges. Coon(1962), for example, argued that the enormous nose of Ne-anderthals had evolved to warm glacial air. Trinkaus’s (1983)influential analysis of the Shanidar Neanderthals emphasizedthat Neanderthal morphologies met adaptive needs for greaterstrength or leverage relative to modern humans. Trinkausargued that many distinctive facial features of Neanderthalsand their relatively large canines and incisors were adaptationsfor increased amounts of anterior biting. Some of the adaptivehypotheses have not received experimental support. For ex-ample, building on previous observations by Hylander (1977)about Neanderthal and Inuit noses, Rae, Koppe, and Stringer(2011) found no evidence that the Neanderthal face is coldadapted. Similarly, Clement, Hillson, and Aiello (2012) found

Pearson Hominin Evolution S223

no, or at best ambiguous, support for the hypothesis thatNeanderthal faces were specially shaped to resist anterior den-tal loading.

The evidence that Neanderthal bodies were adapted to acold climate lies in their wide hips, shortened distal limbsegments, short limbs relative to trunk length, and large ar-ticular surfaces and thick long bone shafts, all of which char-acterize recent humans whose ancestors have lived in coldclimates for thousands of years (Holliday 1997; Pearson 2000;Ruff 1994). Wide hips and robust long bones were alreadypresent in the Sima de los Huesos sample (Arsuaga et al.1999; Bonmatı et al. 2010) and may have been the primitivecondition for Middle Pleistocene Homo (Arsuaga et al. 1999).Wide hips may have also been inherited from Homo erectus(Simpson et al. 2008) rather than appearing as an evolutionarynovelty in Middle Pleistocene hominins.

Recently, Betti, von Cramon-Taubadel, and Lycett (2012)demonstrated that variance within pelvic dimensions of livinghumans tracked population history (distance from Africa)rather than climate while variance in the dimensions of thefemur and tibia correlated with minimum annual temperaturerather than population history. The implications of these find-ings are that contrary to previous conclusions (Ruff 1994),pelvic form appeared to have followed a pattern of largelyneutral evolution like most human cranial dimensions (Bettiet al. 2009, 2010; Roseman 2004; Weaver, Roseman, andStringer 2007, 2008). Given congruent estimates for staturebased on femurs and tibiae, the Sima de los Huesos sampleappears to have less shortened distal limb segments than Ne-anderthals (Carretero et al. 2012), which provides some evi-dence that European hominins evolved more cold adaptedproportions over time. In a review of 75 distinctive cranial,dental, and postcranial features of early modern humans andNeanderthals, Trinkaus (2006) concluded that only one quar-ter were unique to Neanderthals while twice that many wereunique to modern humans, a finding that means that Ne-anderthal morphology had remained fairly primitive whileearly moderns were much more derived. This could provideevidence that early modern humans had shifted to differentniches than archaic humans and had experienced a substantialpulse of selection that tailored them for their new habits.

Genetic drift provides an alternative explanation for mor-phological divergence (Howell 1957). Although this hypoth-esis has been marginalized historically, recent reviews haveemphasized its potential importance (e.g., Hublin 2009). Wea-ver, Roseman, and Stringer (2007, 2008) demonstrated thatif one applies a model of neutral evolution to expected di-vergence in cranial dimensions, the observed morphologicaldivergence between humans and Neanderthals could be ex-plained solely as the result of genetic drift over the last 350kyr. In addition, some recent approaches to cultural inno-vations also emphasize the role of chance, especially if changeis dependent on population size and density (e.g., Powell,Shennan, and Thomas 2009; Shennan 2001). These results

are exciting and motivate one to take a closer look at someof the recent genetic advances.

Genetics

Views of the origin of modern humans and our divergencefrom Neanderthals have been profoundly and perhaps deci-sively influenced by genetic data from living humans as wellas ancient DNA (aDNA) from Neanderthals. The completionof a draft of the Neanderthal nuclear genome (Green et al.2010) and recovery and analysis of nuclear and mitochondrialDNA from “Denisovans,” a third lineage that separated frommodern humans slightly before Neanderthals (Meyer et al.2012; Reich et al. 2010), stand out as signal achievements.These discoveries have decisively answered the question ofwhether interbreeding occurred between modern and archaichumans (it did in both cases) and opened new windows onwhich genes may have been involved in producing evolu-tionary novelties in both modern humans and Neanderthals.

Analysis of autosomal DNA indicates a divergence timebetween modern human and Neanderthal populations of270–440 ka (Reich et al. 2010). Work on aDNA has also shedmore light on Neanderthal population history, suggesting amarked bottleneck among their ancestors sometime beforethe time of the Mezmaiskaya neonate, 60–70 ka (Reich et al.2010), and another bottleneck after 48 ka (Dalen et al. 2012).Furthermore, all Neanderthal mtDNAs share a common an-cestor approximately 100 ka and a common ancestor withmodern humans ∼500 ka (Reich et al. 2010).

Modes of Genetic Evolution

At the genetic level, two of the fundamental means by whichevolution can occur are natural selection (referred to subse-quently simply as “selection”) and genetic drift. The two pro-cesses are not mutually exclusive, and both often act on apopulation at the same time. Selection generally works on agiven gene only if different alleles exist and one confers higherfitness than another, although epistasis (the interdependenceof genes to produce a phenotype) may produce a shiftingtarget for selection. The ultimate source of new alleles is mu-tation, which occurs rarely. Most mutations are either neutral(and have no effect on natural selection) or harmful (by in-terfering with gene function and thus causing deleterious ef-fects to the organism); only a small number of mutationsprove to be beneficial. Most selection pressures that have ac-tually been observed in nature are weak in strength; allelesunder strong positive selection rapidly move to fixation whilealleles under strong negative selection are rapidly removedfrom a population (Futuyma 1986).

In a larger population, one expects more of the rare, fa-vorable mutations to arise simply because the number of newmutations varies with population size (Cochran and Harp-ending 2009; Hawks et al. 2007). Large populations also tendto moderate, often to a great degree, the effects of drift. Thus,

S224 Current Anthropology Volume 54, Supplement 8, December 2013

large populations provide favorable conditions for the pro-duction of new, beneficial mutations; large and growing pop-ulations provide the most fertile ground for new mutationsto arise and increase in frequency.

Under a neutral model of evolution, most new mutationsare lost to drift (especially in small or numerically stable pop-ulations). In growing populations, new mutations are morelikely to be preserved, while in shrinking populations they aremore likely to be lost because of drift (Harpending et al. 1993,1998). Drift slows in large populations but accelerates in smallpopulations and can override the signal of all but the strongestselective pressures. As a result, population size emerges as akey variable in both selection and drift. Estimating populationsize in the past is difficult and invariably requires one to makea series of assumptions that are open to criticism. The prob-lem may not be intractable, however, because during the Mid-dle and Upper Pleistocene, recurrent 100,000-year-long glacialcycles drove climate change and almost certainly affectedhominin populations.

Geology and Paleoclimate

The climate in Europe in the Middle-Upper Pleistocene wasdominated by a high-amplitude 100,000-year cycle that ap-pears to have been determined by the eccentricity cycle in theearth’s orbit around the sun (deMenocal 2004). The glacialcycles show up very clearly in oxygen isotope values fromdeep-sea cores and ice cores from Greenland and Antarctica(deMenocal 2004). The paleoclimate of Africa presents a morecomplicated picture but one that is ultimately related to or-bital dynamics because of changes in air circulation and rain-fall that arose as consequences of the amount of solar radi-ation (insolation) that reached the earth (Siddall et al. 2010;Trauth, Larrasoana, and Mudelsee 2009). In Africa, oscilla-tions in precipitation were more crucial than temperature,and paleoclimatic records show that precipitation fluctuateddramatically in Africa during the Pleistocene.

Records of dust flux from deep-sea cores such as ODP 721/722 in the Arabian Sea and ODP 659 off of the coast ofMauritania can serve as proxies for precipitation in East Africaand the western Sahel and southern Sahara (deMenocal 2004;Trauth, Larrasoana, and Mudelsee 2009). The record stemsfrom long-standing patterns of atmospheric circulation. InJune, July, and August, clockwise-circulation monsoonalwinds blow moisture onshore in Somalia from the IndianOcean and carry dust from Somalia into the Arabian Sea.During the same months the southerly Shamai winds scourdust off of the Arabian Peninsula and deposit it in the ArabianSea. The dust-flux record from the Arabian Sea, ODP 721/722, therefore records both of these influences. Site ODP 659,off of the coast of Mauritania and Western Sahara, receivesa substantial amount of its dust from the southern Saharaand northern Sahel during these months. During December,January, and February, the direction of air circulation reversesover East Africa, and the Trade Winds blow air onshore over

East Africa and across the Sahel and much of the Sahara froma northeasterly counterclockwise direction. Site ODP 659 alsoreceives dust from Western Sahara and portions of the north-ern Sahara during these months.

Ample precipitation over East Africa and the southern Sa-hara and Sahel promote the growth of vegetation, which de-creases the amount of dust that winds scour off of the land.Periods of decreased precipitation diminish the amount ofvegetation and dependent biomass (including humans) andproduce more dust. The Arabian Sea dust core shows a100,000-year oscillation between wet and dry with the mostintense and long-lasting dry periods corresponding to themajor glacial advances in the Northern Hemisphere (fig. 1).Major dry phases would have been guaranteed to producegreatly expanded Sahara and Kalahari deserts and unfavorableconditions for human habitation. This seems to have hap-pened many times in the past, with conditions in OIS 2 serv-ing as a case in point (Brooks and Robertshaw 1990; Deaconand Lancaster 1988). In addition, cores and seismology ofseveral of the oldest East African great lakes, especially LakesMalawi and Tanganyika, have shown that substantial portionsof tropical Africa south of the equator experienced severedroughts over the last 200 kyr that would not have beeninferred from the oxygen isotope curve (Burnett et al. 2010;Cohen et al. 2007; Scholz et al. 2011), although the effects ofthese droughts appear to have been mitigated or absent atthe equator and cannot be generalized to the whole of Africa(Blome et al. 2012). Some of the reconstructed lake levels(e.g., for Lake Malawi) do not closely follow the dust curves(fig. 2), suggesting yet another layer of complexity in theclimatic record.

In Europe, major glaciations appear to have pushed hom-inins out of the northern European plain and Britain andinto southern refugia along the Mediterranean Sea (Dennell,Martinon-Torres, and Bermudez de Castro 2011; Stringer2006). Very wet and warm periods in Europe produced denseforests that may have also been unfavorable habitat (Roe-broeks, Conard, and Van Kolfschoten 1992), although inter-stadial periods seem to have been far more favorable for hom-inin populations than the coldest periods of glaciations.

Predictions

The direct influence of climatic conditions on population sizesin Europe and Africa allows a series of predictions about therelative ability of selection and drift to produce changes inhominin populations. Periods of large-scale glacial advancein Europe should produce periods of stress and low popu-lation numbers and rapid genetic drift in Neanderthals. Incontrast, periods of warmer but not yet heavily forested con-ditions would have supported a higher biomass of large her-bivores and the humans who preyed on them (Roebroeks,Conard, and Van Kolfschoten 1992), thus producing an in-crease in hominin population numbers and a decelerated rateof drift.

Figu

re1.

Key

even

tsin

gen

etic

evol

uti

on,

ages

offo

ssil

spec

imen

s,ox

ygen

isot

ope

stag

e(O

IS)

curv

es,

and

dust

-flu

xda

tafr

omth

eA

rabi

anSe

a.O

IScu

rves

adap

ted

from

Kle

in(2

009)

;du

st-fl

ux

curv

esfr

omD

onge

set

al.

(201

2).

S226 Current Anthropology Volume 54, Supplement 8, December 2013

Figure 2. Water levels in Lake Malawi over the last 145,000 years and 300,000 years of dust flux from different sea cores. Waterlevels adapted from Scholz et al. (2011); dust-flux data after Donges et al. (2012).

Mellars and French (2011) have argued that Neanderthalsin southwestern France had population numbers during theWurm Glaciation (OIS 4–3) that totaled approximately one-tenth (actually 1/9) as many individuals as the later Aurig-nacian occupation, although many of the assumptions thatled to this conclusion have been challenged (Dogandzic andMcPherron 2013) and defended (Mellars and French 2013).Mellars and French (2011) refrained from proposing an es-timate of the number of individuals this would have involved,but Bocquet-Appel et al.’s (2005) model for human popu-lation in Upper Paleolithic Europe produced an estimate ofthe population of Europe during the Aurignacian of 4,424people (95% confidence interval [CI]: 1,738–28,359). If oneaccepts Bocquet-Appel et al.’s (2005) estimates for the Au-rignacian and extends Mellars and French’s (2011) conclu-sions to the whole of Europe, it would imply that the Ne-anderthal population of Europe only totaled 492 individuals(95% CI: 193–3,151). Reviewing previous estimates of Ne-anderthal population numbers, Dennell, Martinon-Torres,and Bermudez de Castro (2011) proposed the Neanderthalpopulation of Europe totaled 3,000–5,000 during interstadialsand 1,500–2,500 during the depths of glacial advances, whenNeanderthal populations were confined to refugia in Iberia,

Italy, and the Balkans. At the higher end of estimates, Sørensen(2011) proposed a population of less than 10,000 individualsduring the Eemian interglacial (OIS 5e), when Neanderthalnumbers and inhabited territory may have been at a peak.Such low numbers would make sense given the large terri-tories required for the Neanderthal specialization on large-bodied prey (Stiner 2013). If these remarkably low populationnumbers are accurate, Neanderthals may never have beennumerous enough to experience conditions in which therewere enough individuals for favorable mutations to arise ata brisk pace. Rather, they may have been skirting the edge ofextinction for most of their existence, generally losing geneticdiversity as they did so. Neanderthals may have only rarelyexperienced periods of population growth and range expan-sion.

New data from the Denisovan genome from Siberia (Meyeret al. 2012) suggest that this population also had a strikinglylow long-term effective population size of approximately

individuals for the period between 400 and 100N p 1,667e

ka (Li, Patterson, and Reich 2012). It is important to bear inmind that effective population size can be different from (andsometimes lower by an order of magnitude or more) censussize (the actual number of individuals) and that Ne approx-

Pearson Hominin Evolution S227

imates the harmonic mean of the number of breeding indi-viduals over time. Nevertheless, for an effective populationsize to shrink from 16,667 before 400 ka to 1,667 after 400ka as the Denisovans did and apparently remained (Li, Pat-terson, and Reich 2012), the population must have crashedto 1,667 individuals (or fewer) one or more times.

Lest one think that Neanderthals and Denisovans were fun-damentally different from modern humans in the face ofclimatic instability, it is important to realize that some recentresearch to model effective population size in modern humanpopulations based on genomic data suggests that both theancestors of living Europeans and Chinese experienced oneor more severe bottlenecks between 40 and 20 ka such thatthe effective population size of each of these populationsshrank to a size of approximately during thisN p 1,200e

interval before rebounding to a higher size (to Ne between11,000 and 50,000) during the Holocene (Li and Durbin2011). This result implies that similar dynamics, likely at-tributable to climatic cycles, affected archaic and modern pop-ulations in Eurasia in very similar ways.

Recently, a number of authors have stressed that climaticdeterioration in Europe and the Near East could have led tothe local extinction of populations (Dennell, Martinon-Torres,and Bermudez de Castro 2011; Hublin and Roebroeks 2009;Shea 2011; Stewart and Stringer 2012; Stringer 2006). Britain,in particular, seems to have been abandoned with each majorglacial advance and then reoccupied, at least as long as a landbridge connected it to the continent (Stringer 2006). Heinrichevents, short periods of extreme cold followed by rapid warm-ing, during glaciations may have posed especially difficultchallenges for hominins in Europe (Stewart and Stringer2012) and perhaps contributed to a contraction in the rangeof Neanderthals in southern Iberia and the spread of modernhumans bearing Aurignacian technology into France andnorthern Spain (d’Errico and Sanchez Goni 2003).

In Africa, especially in East Africa, biomass and humanpopulation size were much more dependent on the availability(and predictability) of precipitation, and the dust-flux datafrom deep-sea cores provide an accessible gauge of precipi-tation (deMenocal 2004; Rohling et al. 2013). Periods of lowdust flux indicate more precipitation, more vegetation, moreanimal biomass, and more people. During these times selec-tion would logically have more power to create phenotypicchange, and genetic drift would be less influential. Periods ofhigh dust flux correspond to less precipitation, less vegetation,fewer people, and thus rapid genetic drift. The scale of theeffect of climatic changes on human populations is clearlyapparent in the dramatic decrease in the number of sites inthe Last Glacial Maximum in East Africa (Brooks and Rob-ertshaw 1990); difficulties for human populations likely con-tinued even after that, including evidence of the desiccationof Lake Victoria until ca. 14.5 ka (Williams et al. 2006) andthe desiccation of Lake Tana around the same time (Lamb etal. 2007; Marshall et al. 2011).

If selection was the crucial factor driving change in the

lineages of Neanderthals or modern humans, then majorchanges in anatomy in each lineage should emerge duringperiods that favor large population numbers. On the otherhand, if drift was the key force in driving the divergence ofNeanderthals and early modern humans, then key evolution-ary events and appearance of new morphologies should ap-pear during or immediately after periods of low populationnumbers. As new morphologies may be effectively invisiblein the fossil record during periods of contracted populationsize, they may, in fact, appear in the record only slightly later,once population sizes had rebounded.

Any test of these hypotheses faces practical limitations, in-cluding an incomplete fossil record, poor dating of some fos-sils, and inadequate resolution of current methods in pin-pointing morphological or genetic changes to exact spots inthe 100,000-year glacial and faster insolation cycles. Marginsof error for dates for fossils or genetic events may overlapboth favorable and unfavorable periods of climatic cycles. Afurther difficulty particular to Africa lies in the variability ofdust-flux records: different patterns occur in different coresaround Africa (fig. 2). For the sake of argument, I assume inthis paper that the record of dust flux from the Arabian Seais the most relevant to the origin of modern humans, butthis issue is certainly open to debate. As a case in point, Blomeet al. (2012) synthesized paleoclimatic records for the wholeof Africa using multiple proxies, including terrestrial, lacus-trine, and oceanic data (fig. 3). The resulting synthesis depictsa mosaic of wet and dry periods that are frequently asyn-chronous between regions and do not correspond in a con-sistent way to the OISs. Their results for East Africa are per-haps the most useful for inferences regarding the origin anddispersal of modern humans. Likewise, Rohling et al. (2013)present multiproxy data for the Mediterranean and Red Searegions, two areas that were crucial for hominin dispersalsfrom (and perhaps into) Africa during the last 500 kyr. Thearrows in figure 3 extend the analyses of Blome et al. (2012)by indicating possible population expansions within Africaand possible expansions into Arabia during the windows ofopportunity described by Rohling et al. (2013). In each case,populations can be inferred to have spread from regions withfavorable climate and thus presumably comparatively highhuman population density into regions previously nearly de-void of people but with newly favorable climatic conditions.

Results

Comparison of the time lines of paleoclimate, the fossil rec-ord, and genetic divergences and bottlenecks provide a roughcheck of whether key events occur in periods favorable forlarge population numbers or in periods unfavorable for largepopulations (fig. 1). In many cases, the relationships are highlysuggestive, but problems remain.

S228 Current Anthropology Volume 54, Supplement 8, December 2013

Figure 3. A, Geographical regions in Africa adapted from Blome et al. (2012, with permission from Elsevier). B, Climate in Africaadapted from Blome et al. (2012, with permission from Elsevier) with two periods of a wet Sahara coinciding with periods of lowsea level (following Rohling et al. 2013) indicated. The arrows show possible directions of colonization from regions of higherpopulation density into adjacent areas. Question marks symbolize that the geographic source of the colonizing population is uncertain(and thus, one might assume, largely of local origin).

Europe

The age of the fossils from Sima de los Huesos is a keyproblem for making sense of the tempo and mode of homininevolution in Europe over the last 500 kyr (Stringer 2012). Ifthe Sima de los Huesos dates to 500–600 ka (Bischoff et al.2007), then one can conclude the features characteristic ofthe later, “classic” Neanderthals dating to OIS 4–3 increasedin frequency very slowly within the Neanderthal lineage.Hominins with the full suite of Neanderthal cranial traitsappear by OIS 7 (Hublin 2009), but a further pulse of changemade western European Neanderthals from OIS 4–3 especiallydistinctive. It is difficult to tell whether this apparently lateinflection in the rate of “Neanderthalization” was the resultof selection within a large population during OIS 5 or ofrapid drift in a small population during OIS 4–3.

Neanderthal mtDNA sequences provide support for a latebottleneck in their population. A recent analysis of the ancient

mtDNA of a Neanderthal from Valdegoba, Spain, dating to48.5 ka, shows that all Neanderthal mtDNA sequences post-dating this time formed a compact, monophyletic groupwithin the known Neanderthal sequences (Dalen et al. 2012).These late, western Neanderthal mtDNA sequences have acoalescent age of 58 ka (the end of OIS 4) with a 95% CI54–77 ka. A severe population bottleneck (or selection, whichmay be less likely) could produce such a pattern. Dalen andcolleagues (2012) conducted an Approximate Bayesian Com-putation Analysis to test demographic models of neutral evo-lution that were the most likely to produce this pattern ofgreatly reduced diversity. The results showed that the mostprobable scenario (isolation and drift in the western and east-ern populations of Neanderthals at 48 ka) would have in-volved an effective population size (Ne) of western Neander-thals of only 300 females, a marked reduction from theestimate for the eastern subpopulation ( females).N p 2,000e

Pearson Hominin Evolution S229

The effective population size for the autosomal genes in theentire population is expected to be four times that of mtDNA(i.e., and 8,000 in the western and eastern sub-N p 1,200e

populations, respectively). The usual cautions about the dif-ference between census size and Ne apply, but one is left withthe strong impression that western European Neanderthalsexperienced a major population crash during OIS 4.

A second possibility for Europe is that the fossils from theSima de los Huesos date to only around 350 ka. Stringer(2012) argues that the younger age is supported by the mosaicpresence of many distinctive Neanderthal cranial, dental, andpostcranial features in the sample. These features are similarlycommon in European fossils dating to 300–200 ka but muchmore rare (or absent) in earlier fossils from other sites inEurope. Stringer (2012) also notes that an age of 600–500 kafor the Sima de los Huesos fossils would place them earlierthan the estimated population divergence times for the an-cestors of modern humans and Neanderthals and that thedated spelothems may, in fact, have been breached by a flowof younger sediments within the cave so that younger stratacontaining the hominins now underlie an only partially com-plete but older spelothem. However, Spanish researchers pre-fer the older date, noting that the younger age is contradictedby fossil fauna from the same deposit as the hominins, in-cluding relatively primitive fossils of Ursus deningeri and thevole Clethrionomys acrorhiza (Garcıa and Arsuaga 2011). Nev-ertheless, if the hominins from Sima de los Huesos date toaround 350 ka, the time span for drift would be cut in half,implying a more rapid pace of Neanderthalization later in thesequence. If this was the case, much of the genetic and mor-phological change may have been concentrated in bursts ofdrift that corresponded to major contractions in Neanderthalnumbers during OIS 8, 6, and 4. A shortened time span wouldalso create a stronger association between expansion in brainsize and adoption of Middle Paleolithic technologies between300 and 200 ka.

Africa

The Arabian Sea dust core shows a long relatively wet andstable period between 640 and 427 ka. This period is asso-ciated with the first appearance of Homo heidelbergensis (orHomo rhodesiensis, if this name is to be preferred) in Africa(i.e., the Bodo cranium, dated to 600 ka; Clark et al. 1994;Rightmire 1996) and, intriguingly, marked technological ad-vances represented by precociously early blade production andcore technology in the Kapthurin Formation at Lake Baringo(Johnson and McBrearty 2010; Tryon and McBrearty 2006).It is possible that large population sizes in Africa during muchof the Middle Pleistocene drove both cultural innovations andanatomical evolution via positive selection on beneficial newmutations.

The dust core also indicates marked dry periods in EastAfrica during OIS 8 (301–242 ka), OIS 6 (186–127 ka), andOIS 4–2 (71–12 ka), although Blome et al. (2012) suggest

that the last interval in East Africa was interrupted by a wetperiod around 55–50 ka. The origin of modern humans datesto OIS 6. The first fossils of recognizably modern form dateto 195–160 ka (i.e., the end of OIS 7 and into OIS 6). Allextant mtDNA sequences coalesce to a common ancestor at140–210 ka (Behar et al. 2008), and Y chromosomes coalesceat ka (Cruciani et al. 2011), although an ex-141.5 � 15.6tremely rare Y-chromosome haplotype from an African Amer-ican man was recently reported that coalesces with other Ychromosomes at 338 ka (Mendez et al. 2013). OIS 6 has beenlikened to the hyperarid conditions of the Last Glacial Max-imum in OIS 2 (Deacon and Lancaster 1988), which featuredgreatly decreased archaeological visibility of human popula-tions in much of Africa (Brooks and Robertshaw 1990). Ge-netic drift would be expected to be the dominant factor duringsuch a period, but it is worth reiterating that African climatewas a complex and regionally variable mosaic (Blome et al.2012).

Indirect evidence from autosomal genes also supports thehypothesis that the African ancestors of modern humans ex-perienced a major population bottleneck during this period.Fagundes et al. (2007) simulated several scenarios for theorigin of modern humans with a sample of 50 autosomal locithat were subsequently compared with observed patterns ofvariation in human nuclear loci. They found the best cor-respondence to observed patterns of human genetic variationin a model that features an origin of modern humans in Africafollowed by exponential population growth, expansion fromAfrica and replacement of archaic hominins outside of Africa(specifically in Asia in their model) followed by exponentialpopulation growth in Asia, and finally a migration from Asiato the Americas followed by a final burst of exponential pop-ulation growth in the New World. The best-fitting modelproduced a series of posterior estimates for demographic andhistorical parameters, including the age of the speciation eventthat produced modern humans (median: 141,455; 95% CI:103,535–185,642), the age of the migration from Africa (me-dian: 51,102; 95% CI: 40,135–70,937), the age of the colo-nization of the Americas (median: 10,280; 95% CI: 7,647–15,945), the size of the archaic African population (median:12,772; 95% CI: 6,604–20,211), the population size duringthe bottleneck during speciation (median: 600; 95% CI: 76–1,620), the size of the bottleneck when leaving Africa (median:462; 95% CI: 64–1,224), and the size of the bottleneck whenleaving Asia to settle the Americas (median: 452; 95% CI: 71–1,280).

Not all of the genetic data supports the conclusion that apopulation bottleneck produced modern humans, and someof the data strongly contradict that hypothesis. Using datafrom complete genomes of several modern men comprisingtwo Yoruba, three Europeans, one Chinese, and one Korean,Li and Durbin (2011) applied population genetics models toinfer changes in human effective population size over the lastmillion years. Intriguingly, their data showed no evidence ofa bottleneck between 200 and 100 ka. In fact, their results

S230 Current Anthropology Volume 54, Supplement 8, December 2013

show growth in effective population size from ∼450 ka until120–150 ka and very similar histories (and likely shared his-tories in an ancestral source population) of Yoruban, easternAsian, and European population size before 60 ka. After that,all three populations experienced bottlenecks, although theone that affected the ancestors of the Yoruba appears to havebeen less severe and allowed an earlier recovery.

Likewise, by applying a population genetics model to ex-pectations for (neutral) change in cranial dimensions, Weaver(2012) showed that crania that had dimensions that differedby one standard deviation from modern crania could be ex-pected by around 165 ka, which corresponds reasonably wellto when most researchers agree that modern (or nearly mod-ern) humans appear in the East African fossil record. Weaver’smodel, however, assumes a constant effective population. Apotential explanation for the apparent lack of a contractionin the effective population size of the ancestors of modernhumans in Africa is that if there were in fact bottlenecks withina subdivided population in Africa, following the bottlenecks,members of dissimilar populations mixed extensively, restor-ing to the resulting population much of the genetic variationthat existed before each bottleneck. The complex mosaic offavorable climates over time in different parts of Africa re-ported by Blome et al. (2012) may have provided the rightconditions for this sort of mechanism (fig. 3).

Given that both climatic and genomic data suggest a bot-tleneck in East Africa and Arabia after 60 ka, it is highly likelythat a substantial amount of genetic drift occurred in thepopulation of modern humans as they left Africa or for aperiod of time immediately afterward. As a result, outbreedingwould have been highly favorable if heterozygosity was greatlyincreased by these events, especially for loci such as the majorhistocompatibility complex, in which alleles from archaic Eur-asian populations are far more frequent in populations out-side of Africa than they are in other loci (Abi-Rached et al.2011).

Low lake levels in Lake Malawi and Lake Tanganyika andhigh levels of dust flux suggest generally unfavorable condi-tions for human population growth in tropical Africa duringmuch of OIS 5 (Blome et al. 2012; Scholz et al. 2011). Theresult was likely relatively rapid genetic drift and populationdifferentiation among modern humans in Africa. Anotherpopulation contraction in most of Africa in OIS 2 probablyaccounts for late (Holocene or terminal Pleistocene) appear-ance of crania that have cranial metrics that cannot be dis-tinguished from one or more extant populations (De Villiersand Fatti 1982; Habgood 1989). Early Holocene (and likelyfrom the end of OIS 2) human remains show evidence ofstrong morphological differentiation among African popu-lations. The spread of pastoralism and agricultural popula-tions in Africa has blurred or erased these stark distinctions(Tishkoff et al. 2009). For Africa, then, the dust-core andgenetic data suggest that selection may have been importantfrom 600–400 ka, but periods of drift had more potential tobe the dominant influence thereafter.

Conclusions

As a result of this comparison of records of paleoclimate,morphological change, and genetic change, it seems apparentthat many of the observed changes leading to Neanderthalswere more likely to have been the products of drift thanselection, whereas both drift and selection may have beenimportant in the emergence of modern humans. Many of thekey events appear to date to periods in which population sizeswere greatly reduced and genetic drift would have been rapid.The picture that emerges is one of human population historythat was highly (although almost certainly not exclusively)contingent on climatic changes.

One prominent example of this dependence on climatecomes from mtDNA intramatch distributions that show rapidpopulation growth in Africa at ca. 80 ka (Harpending et al.1993, 1998; Sherry et al. 1994); estimates of lake levels fromLakes Malawi and Tangyanika show a return of wet conditionsat the same time (Burnett et al. 2010; Scholz et al. 2011),although the pattern of wet and dry periods for Africa as awhole forms a complex mosaic that frequently departs fromthe pattern observed in Lakes Malawi and Tangyanika (Blomeet al. 2012). Rohling et al. (2013) demonstrate that favorableconditions combining low sea levels with elevated levels ofprecipitation to support terrestrial biomass (including hu-mans) would have facilitated movements across the Bab-el-Mendab strait between East Africa and Arabia only in narrowwindows of time, the latest of which dates to 70–65 ka. Thesepatterns help to illuminate a particularly irksome issue inresearch on the origin of modern humans: the question ofwhy modern humans only expanded out of Africa at 50–60ka if “anatomically modern” morphology arose between 200and 150 ka (e.g., Klein 2009). The answer seems to be thatclimatic conditions did not favor a large, interconnected pop-ulation in Africa between 125–ca. 80 ka because each regionexperienced one or more dry periods during this interval(Blome et al. 2012). Many of the famous cultural advancesassociated with the Upper Paleolithic and Late Stone Age arealso likely to have depended on population size and density(Powell, Shennan, and Thomas 2009; Premo and Kuhn 2010).

Acknowledgments

Many thanks to Erella Hovers and Steve Kuhn for the invi-tation to participate in this stimulating conference and to theWenner-Gren Foundation for organizing and funding it. In-sightful comments by conference participants, two anony-mous reviewers, and Steve Kuhn helped to refine the ideaspresented here; any remaining faults are my own.

References CitedAbi-Rached, Laurent, Matthew J. Jobin, Subhash Kulkarni, Alasdair Mc-

Whinnie, Klara Dalva, Loren Gragert, Farbod Babrzadeh, et al. 2011. Theshaping of modern human immune systems by multiregional admixturewith archaic humans. Science 334:89–94.

Pearson Hominin Evolution S231

Arsuaga, J.-L., I. Martınez, A. Gracia, and C. Lorenzo. 1997. The Sima de losHuesos crania (Sierra de Atapuerca, Spain): a comparative study. Journalof Human Evolution 33:219–281.

Arsuaga, Juan-Luis. 2010. Terrestrial apes and phylogenetic trees. Proceedingsof the National Academy of Sciences of the USA 107:8910–8917.

Arsuaga, Juan-Luis, Carlos Lorenzo, Jose-Miguel Carretero, Ana Gracia, Ig-nacio Martınez, Nuria Garcıa, Jose-Marıa Bermudez de Castro, and EudaldCarbonell. 1999. A complete human pelvis from the Middle Pleistocene ofSpain. Nature 399:255–258.

Behar, Doran M., Richard Villems, Himla Soodyall, Jason Blue-Smith, LuısaPereira, Ene Metspalu, Rosaria Scozzari, et al. 2008. The dawn of humanmatrilineal diversity. American Journal of Human Genetics 82:1130–1140.

Betti, Lia, Francois Balloux, William Amos, Tsunehiko Hanihara, and AndreaManica. 2009. Distance from Africa, not climate, explains within-populationphenotypic diversity in humans. Proceedings of the Royal Society B 276:809–814.

Betti, Lia, Francois Balloux, Tsunehiko Hanihara, and Andrea Manica. 2010.The relative role of drift and selection in shaping the human skull. AmericanJournal of Physical Anthropology 141:76–82.

Betti, Lia, Noreen von Cramon-Taubadel, and Stephen J. Lycett. 2012. Humanpelvis and long bones reveal differential preservation of ancient populationhistory and migration out of Africa. Human Biology 84:139–152.

Bischoff, James L., Ross W. Williams, Robert J. Rosenbauer, Arantza Aram-buru, Juan Luis Arsuaga, Nuria Garcıa, and Gloria Cuenca-Bescos. 2007.High-resolution U-series dates from the Sima de los Huesos hominids yields

kyrs: implications for the evolution of the early Neanderthal lineage.��600�66

Journal of Archaeological Science 34:763–770.Blome, Margaret Whiting, Andrew S. Cohen, Christian A. Tryon, Alison S.

Brooks, and Joellen Russell. 2012. The environmental context for the originsof modern human diversity: a synthesis of regional variability in Africanclimate 150,000–30,000 years ago. Journal of Human Evolution 62:563–592.

Bocquet-Appel, Jean-Pierre, Pierre-Yves Demars, Lorette Noiret, and DmitryDobrowsky. 2005. Estimates of Upper Paleolithic meta-population size inEurope from archaeological data. Journal of Archaeological Science 32:1656–1668.

Bonmatı, Alejandro, Asier Gomez-Olivencia, Juan-Luis Arsuaga, Jose MiguelCarretero, Ana Gracia, Ignacio Martınez, Carlos Lorenzo, Jose Marıa Ber-mudez de Castro, and Eudald Carbonell. 2010. Middle Pleistocene lowerback and pelvis from an aged human individual from the Sima de losHuesos site, Spain. Proceedings of the National Academy of Sciences of theUSA 107:18386–18391.

Brauer, Gunter. 2008. The origin of modern anatomy: by speciation or in-traspecific evolution? Evolutionary Anthropology 17:22–37.

Brooks, Alison S., and Peter Robertshaw. 1990. The Glacial Maximum intropical Africa: 22,000–12,000 BP. In Low latitudes, vol. 2 of The world at18,000 BP. Clive Gamble and Oolda Soffer, eds. Pp. 121–169. London:Unwin Hyman.

Brown, Francis H., Ian McDougall, and John G. Fleagle. 2012. Correlation ofthe KHS Tuff of the Kibish Formation to volcanic ash layers at other sites,and the age of early Homo sapiens (Omo I and Omo II). Journal of HumanEvolution 63:577–585.

Burnett, Allison P., Michael J. Soreghan, Christopher A. Scholz, and Erik T.Brown. 2010. Tropical East African climate change and its relation to globalclimate: a record from Lake Tanganyika, tropical East Africa, over the past90 kyr. Palaeogeography, Palaeoclimatology, Palaeoecology 303:155–167.

Carretero, Jose-Miguel, Laura Rodrıguez, Rebeca Garcıa-Gonalez, Juan-LuisArsuaga, Asier Gomez-Olivencia, Carlos Lorenzo, Alejandro Bonmatı, AnaGracia, Ignacio Martınez, and Rolf Quam. 2012. Stature estimation fromcomplete long bones in the Middle Pleistocene humans from the Sima delos Huesos, Sierra de Atapuerca (Spain). Journal of Human Evolution 62:242–255.

Churchill, Steven E., Lee R. Berger, Adam Hartstone-Rose, and B. HeadmanZondo. 2012. Body size in African Middle Pleistocene Homo. In Africangenesis: perspectives on hominin evolution. S. C. Reynolds and A. Gallagher,eds. Pp. 319–346. Cambridge: Cambridge University Press.

Clark, J. D., J. de Heinzelin, K. D. Schick, W. K. Hart, T. D. White, G.WoldeGabriel, R. C. Walter, et al. 1994. African Homo erectus: old radio-metric ages and young Oldowan assemblages in the Middle Awash Valley,Ethiopia. Science 264:1907–1910.

Clark, J. Desmond, Yonas Beyene, Giday WoldeGabriel, William K. Hart, PaulR. Renne, Henry Gilbert, Alban Defleur, et al. 2003. Stratigraphic, chro-nological and behavioral contexts of Pleistocene Homo sapiens from MiddleAwash, Ethiopia. Nature 423:747–752.

Clement, Anna F., Simon W. Hillson, and Leslie C. Aiello. 2012. Tooth wear,Neanderthal facial morphology and the anterior dental loading hypothesis.Journal of Human Evolution 62:367–376.

Cochran, Gregory, and Henry Harpending. 2009. The 10,000 year explosion:how civilization accelerated human evolution. New York: Basic.

Cohen, Andrew S., Jeffrey R. Stone, Kristina R. M. Beuning, Lisa E. Park,Peter N. Reinthal, David Dettman, Christopher A. Scholz, et al. 2007. Eco-logical consequences of early Late Pleistocene megadroughts in tropicalAfrica. Proceedings of the National Academy of Sciences of the USA 104:16422–16427.

Coon, C. S. 1962. The origin of races. New York: Knopf.Cruciani, Fulvio, Beniamino Trombetta, Andrea Massaia, Giovanni Destro-

Bisol, Daniele Sellitto, and Rosaria Scozzari. 2011. A revised root for thehuman Y chromosomal phylogenetic tree: the origin of patrilineal diversityin Africa. American Journal of Human Genetics 88:814–818.

Dalen, Love, Ludovic Orlando, Beth Shapiro, Mikael Brandstrom Durling,Rolf Quam, M. Thomas P. Gilbert, J. Carlos Dıez Fernandez-Lomana, EskeWillerslev, Juan Luis Arsuaga, and Anders Gotherstrom. 2012. Partial ge-netic turnover in Neandertals: continuity in the East and population re-placement in the West. Molecular Biology and Evolution 29:1893–1897.

Day, M. H., and C. B. Stringer. 1991. Les restes craniens d’Omo-Kibish etleur classification a l’interieur du genre Homo. L’Anthropologie 95:573–594.

Deacon, Janette, and N. Lancaster. 1988. Late Quaternary palaeoenvironmentsof southern Africa. Oxford: Oxford University Press.

Dean, Christopher, Meave G. Leakey, Donald Reid, Friedemann Schrenk, GaryT. Schwartz, Christopher Stringer, and Alan Walker. 2001. Growth processesin teeth distinguish modern humans from Homo erectus and earlier hom-inins. Nature 414:628–631.

Dean, M. Christopher, and B. Holly Smith. 2009. Growth and developmentof the Nariokotome youth, KNM-WT 15000. In The first humans: originand early evolution of the genus Homo. Frederick E. Grine, John G. Fleagle,and Richard E. Leakey, eds. Pp. 101–120. New York: Springer.

deMenocal, Peter B. 2004. African climate change and faunal evolution duringthe Pliocene-Pleistocene. Earth and Planetary Science Letters 220:3–24.

Dennell, Robin W., Marıa Martinon-Torres, and Jose M. Bermudez de Castro.2011. Hominin variability, climatic instability and population demographyin Middle Pleistocene Europe. Quaternary Science Reviews 30:1511–1524.

d’Errico, Francesco, and Maria Fernanda Sanchez Goni. 2003. Neandertalextinction and the millennial scale climatic variability of OIS 3. QuaternaryScience Reviews 22:769–788.

De Villiers, H., and L. P. Fatti. 1982. The antiquity of the Negro. South AfricanJournal of Science 78:321–332.

Dogandzic, Tamara, and Shannon P. McPherron. 2013. Demography and thedemise of Neandertals: a comment on “Tenfold population increase inWestern Europe at the Neandertal-to-modern human transition.” Journalof Human Evolution 64:311–313.

Donges, Jonathan F., Reik V. Donner, Martin H. Trauth, Norbert Marwan,Hans-Joachim Schellnhuber, and Jurgen Kurths. 2012. Nonlinear detectionof paleoclimate-variability transitions possibly related to human evolution.Proceedings of the National Academy of Sciences of the USA 108:20422–20427.

Fagundes, Nelson J. R., Nicholas Ray, Mark Beaumont, Samuel Neuen-schwander, Francisco M. Salzano, Sandro L. Bonatto, and Laurent Excoffier.2007. Statistical evaluation of alternative models of human evolution. Pro-ceedings of the National Academy of Sciences of the USA 104:17614–17619.

Futuyma, Douglas J. 1986. Evolutionary biology. 2nd edition. Sunderland, MA:Sinauer.

Garcıa, Nuria, and Juan Luis Arsuaga. 2011. The Sima de los Huesos (Burgos,northern Spain): palaeoenvironment and habitats of Homo heidelbergensisduring the Middle Pleistocene. Quaternary Science Reviews 30:1413–1419.

Graves, Ronda R., Amy C. Lupo, Robert C. McCarthy, Daniel J. Wescott, andDeborah L. Cunningham. 2010. Just how strapping was KNM-WT 15000?Journal of Human Evolution 59:542–554.

Green, Richard E., Johannes Krause, Adria W. Briggs, Tomislav Maricic, UdoStenzel, Martin Kircher, Nick Patterson, et al. 2010. A draft of the Nean-dertal genome. Science 328:710–722.

Gunz, Philipp, Simon Neubauer, Lubov Golovanova, Vladimir Doronichev,Bruno Maureille, and Jean-Jacques Hublin. 2012. A uniquely modern hu-man pattern of endocranial development: insights from a new cranial re-construction of the Neandertal newborn from Mezmaiskaya. Journal ofHuman Evolution 62:300–313.

Habgood, P. J. 1989. An examination of regional features on Middle and earlyLate Pleistocene sub-Saharan African hominids. South African ArchaeologicalBulletin 44:17–22.

S232 Current Anthropology Volume 54, Supplement 8, December 2013

Harpending, H. C., S. T. Sherry, A. R. Rogers, and M. Stoneking. 1993. Thegenetic structure of ancient human populations. Current Anthropology 34:483–496.

Harpending, Henry C., Mark A. Batzer, Michael Gurven, Lynn B. Jorde, AlanR. Rogers, and Stephen T. Sherry. 1998. Genetics and traces of ancientdemography. Proceedings of the National Academy of Sciences of the USA 95:1961–1967.

Harris, Kelley, and Rasmus Nielsen. 2013. Inferring demographic history froma spectrum of shared haplotype lengths. PLoS Genetics 9:e1003521.

Harvati, Katerina, Jean-Jacques Hublin, and Philipp Gunz. 2010. Evolutionof Middle-Late Pleistocene human cranio-facial morphology: a 3-D ap-proach. Journal of Human Evolution 59:445–464.

Hawks, John, Eric T. Wang, Gregory M. Cochran, Henry C. Harpending, andRobert K. Moyzis. 2007. Recent acceleration of human adaptive evolution.Proceedings of the National Academy of Sciences of the USA 104:20753–20758.

Holliday, Trenton W. 1997. Body proportions in Late Pleistocene Europe andmodern human origins. Journal of Human Evolution 32:423–447.

Howell, F. C. 1957. The evolutionary significance of variation and varietiesof “Neanderthal” man. Quarterly Review of Biology 32:330–347.

Hublin, Jean-Jacques. 2001. Northwestern African Middle Pleistocene hom-inids and their bearing on the emergence of Homo sapiens. In Human roots:Africa and Asia in the Middle Pleistocene. L. Barham and K. Robson-Brown,eds. Pp. 99–121. Bristol, UK: Western Academic & Specialist Press.

———. 2009. The origin of Neandertals. Proceedings of the National Academyof Sciences of the USA 106:16022–16027.

Hublin, Jean-Jacques, and Wil Roebroeks. 2009. Ebb and flow or regionalextinctions? on the character of Neandertal occupation of northern envi-ronments Comptes Rendus Palevol 8:503–509.

Hublin, J.-J., and A. M. Tillier. 1981. The Mousterian juvenile mandible fromIrhoud (Morocco): a phylogenetic interpretation. In Aspects of human evo-lution. Chris Stringer, ed. Pp. 167–185. London: Taylor & Francis.

Hylander, William L. 1977. The adaptive significance of Eskimo craniofacialmorphology. In Orofacial growth and development. A. A. Dahlberg and T.M. Graber, eds. Pp. 129–169. Paris: The Hague.

Johnson, Cara Roure, and Sally McBrearty. 2010. 500,000 year old blades fromthe Kapthurin Formation, Kenya. Journal of Human Evolution 58:193–200.

Kappelman, John. 1996. The evolution of body mass and relative brain sizein fossil hominids. Journal of Human Evolution 30:243–276.

Klein, Richard G. 2009. The human career. 3rd edition. Chicago: Universityof Chicago Press.

Lamb, Henry F., C. Richard Bates, Paul V. Coombes, Michael H. Marshall,Mohammed Umer, Sarah J. Davies, and Eshete Dejen. 2007. Late Pleistocenedesiccation of Lake Tana, source of the Blue Nile. Quaternary Science Reviews26:287–299.

Li, Heng, and Richard Durbin. 2011. Inference of human population historyfrom individual whole-genome sequences. Nature 475:493–496.

Li, Heng, James Mullikin, and David Reich. 2010. Date of population diver-gence between Neandertals and modern humans. Supplementary onlinematerials 14 to Green, Richard E., Johannes Krause, Adria W. Briggs, Tom-islav Maricic, Udo Stenzel, Martin Kircher, Nick Patterson, et al. 2010. Adraft of the Neandertal genome. Science 328:710–722.

Li, Heng, Nick Patterson, and David Reich. 2012. Inferred population sizechanges in the history of Denisovans. Supplementary note 17 to Meyer,Matthias, Martin Kircher, Marie-Theres Gansauge, Heng Li, Fernando Ra-cimo, Swanpan Mallick, Joshua G. Schraiber, et al. 2012. A high-coveragegenome sequence from an archaic Denisovan individual. Science 338:222–226.

Lieberman, Daniel E. 2011. The evolution of the human head. Cambridge, MA:Belknap.

Lieberman, Daniel E., Brandeis M. McBratney, and Gail Krovitz. 2002. Theevolution and development of cranial form in Homo sapiens. Proceedingsof the National Academy of Sciences of the USA 99:1134–1139.

Marshall, Michael H., Henry F. Lamb, Dei Huws, Sarah J. Davies, RichardBates, Jan Bloemendal, John Boyle, Melanie J. Leng, Mohammed Umer,and Charlotte Bryant. 2011. Late Pleistocene and Holocene drought eventsat Lake Tana, the source of the Blue Nile. Global and Planetary Change 78:147–161.

Martinon-Torres, Marıa, Jose Marıa Bermudez de Castro, Aida Gomez-Robles,Leyre Prado-Simon, and Juan Luis Arsuaga. 2012. Morphological descrip-tion and comparison of the dental remains from Atapuerca-Sima de losHuesos site (Spain). Journal of Human Evolution 62:7–58.

McDougall, Ian, Francis H. Brown, and John G. Fleagle. 2005. Stratigraphic

placement and age of modern humans from Kibish, Ethiopia. Nature 433:733–736.

Mellars, Paul, and Jennifer C. French. 2011. Tenfold population increase inwestern Europe at the Neandertal-to-modern human transition. Science 333:623–627.

———. 2013. Population changes across the Neanderthal-to-modern-humantransition in western France: a reply to Dogandzic and McPherron (2013).Journal of Human Evolution 65:330–333.

Mendez, Fernando L., Thomas Krahn, Bonnie Schrack, Astrid-Maria Krahn,Krishna R. Veeramah, August E. Woerner, Forka Leypey Mathew Fomine,et al. 2013. An African American paternal lineage adds an extremely ancientroot to the human Y chromosome phylogenetic tree. American Journal ofHuman Genetics 92:454–459.

Meyer, Matthias, Martin Kircher, Marie-Theres Gansauge, Heng Li, FernandoRacimo, Swanpan Mallick, Joshua G. Schraiber, et al. 2012. A high-coveragegenome sequence from an archaic Denisovan individual. Science 338:222–226.

Pearson, Osbjorn M. 2000. Activity, climate, and postcranial robusticity: im-plications for modern human origins and scenarios of adaptive change.Current Anthropology 41:569–607.

———. 2008. Statistical and biological definitions of “anatomically modern”humans: suggestions for an integrated approach to modern morphology.Evolutionary Anthropology 17:38–48.

Pearson, Osbjorn M., Danielle F. Royer, Frederick E. Grine, and John G.Fleagle. 2008. A description of the Omo I postcranial skeleton, includingnewly discovered fossils. Journal of Human Evolution 55:421–437.

Powell, Adam, Stephen Shennan, and Mark G. Thomas. 2009. Late Pleistocenedemography and the appearance of modern human behavior. Science 324:1298–1301.

Premo, L. S., and Steven L. Kuhn. 2010. Modeling effects of local extinctionson culture change and diversity in the Paleolithic. PLoS ONE 5:e15582.

Rae, Todd C., Thomas Koppe, and Chris B. Stringer. 2011. The Neanderthalface is not cold adapted. Journal of Human Evolution 60:234–239.

Reich, David, Richard E. Green, Martin Kircher, Johannes Krause, Nick Pat-terson, Eric Y. Durand, Bence Viola, et al. 2010. Genetic history of anarchaic hominin group from Denisova Cave in Siberia. Nature 468:1053–1060.

Rightmire, G. Philip. 1996. The human cranium from Bodo, Ethiopia: evidencefor speciation in the Middle Pleistocene? Journal of Human Evolution 31:21–39.

———. 2009. Middle and Later Pleistocene hominins in Africa and southwestAsia. Proceedings of the National Academy of Sciences of the USA 106:16046–16050.

Roebroeks, Wil, Nicholas J. Conard, and Thijs Van Kolfschoten. 1992. Denseforests, cold steppes, and the Palaeolithic settlement of Europe. CurrentAnthropology 33:551–586.

Rohling, Eelco J., Katherine M. Grant, Andrew P. Roberts, and Juan-CruzLarrasoana. 2013. Palaeoclimate variability in the Mediterranean and RedSea regions during the last 500,000 years: implications for hominin migra-tions. Current Anthropology 54(suppl. 8):S183–S201.

Roseman, Charles C. 2004. Detecting interregionally diversifying natural se-lection on modern human cranial form by using matched molecular andmorphometric data. Proceedings of the National Academy of Sciences of theUSA 101:12824–12829.

Ruff, C. B. 1994. Morphological adaptation to climate in modern and fossilhominids. Yearbook of Physical Anthropology 37:65–107.

Ruff, Christopher B., Erik Trinkaus, and Trenton W. Holliday. 1997. Bodymass and encephalization in Pleistocene Homo. Nature 387:173–176.

Scholz, C. A., A. S. Cohen, T. C. Johnson, J. King, M. R. Talbot, and E. T.Brown. 2011. Scientific drilling in the Great Rift Valley: the 2005 LakeMalawi Scientific Drilling Project: an overview of the past 145,000 years ofclimate variability in Southern Hemisphere East Africa. Palaeogeography,Palaeoclimatology, Palaeoecology 303:3–19.

Shea, John J. 2011. Transitions or turnovers? climatically-forced extinctionsof Homo sapiens and Neanderthals in the east Mediterranean Levant. Qua-ternary Science Reviews 27:2253–2270.

Shennan, Stephen. 2001. Demography and cultural innovation: a model andits implications for the emergence of modern human culture. CambridgeArchaeological Journal 11:5–16.

Sherry, S. T., A. R. Rogers, H. Harpending, H. Soodyall, T. Jenkins, and M.Stoneking. 1994. Mismatch distributions of mtDNA reveal recent humanpopulation expansions. Human Biology 66:761–775.

Pearson Hominin Evolution S233

Siddall, M., E. J. Rohling, T. Blunier, and R. Spahni. 2010. Patterns of millennialvariability over the last 500 ka. Climate of the Past 6:295–303.

Simpson, Scott W., Jay Quade, Naomi E. Levin, Robert Butler, GuillaumeDupont-Niven, Melanie Everett, and Sileshi Semaw. 2008. A female Homoerectus pelvis from Gona, Ethiopia. Science 322:1089–1092.

Smith, Tanya M., and Paul Tafforeau. 2008. New visions of dental tissueresearch: tooth development, chemistry, and structure. Evolutionary An-thropology 17:213–226.

Smith, Tanya M., Paul Tafforeau, Donald J. Reid, Rainer Grun, Stephen Eggins,Mohamed Boutakiout, and Jean-Jacques Hublin. 2007. Earliest evidence ofmodern life history in North American early Homo sapiens. Proceedings ofthe National Academy of Sciences of the USA 104:6128–6133.

Smith, Tanya M., Paul Tafforeau, Donald J. Reid, Joane Pouech, VincentLazzari, John P. Zermeno, Debbie Guatelli-Steinberg, et al. 2010. Dentalevidence for ontogenetic differences between modern humans and Nean-derthals. Proceedings of the National Academy of Sciences of the USA 107:20923–20928.

Sørensen, Bent. 2011. Demography and the extinction of European Nean-derthals. Journal of Anthropological Archaeology 30:17–29.

Stewart, J. R., and C. B. Stringer. 2012. Human evolution out of Africa: therole of refugia and climate change. Science 335:1317–1320.

Stiner, Mary C. 2013. An unshakable Middle Paleolithic? trends versus con-servatism in the predatory niche and their social ramifications. CurrentAnthropology 54(suppl. 8):S288–S304.

Stringer, Chris. 2002. Modern human origins: progress and problems. Phil-osophical Transactions of the Royal Society of London B 357:563–579.

———. 2006. Homo britannicus: the incredible story of human life in Britain.London: Penguin.

———. 2007. The origin and dispersal of Homo sapiens: our current state ofknowledge. In Rethinking the human revolution: new behavioural and bio-logical perspectives on the origin and dispersal of modern humans. Paul Mel-lars, Katie Boyle, Ofer Bar-Yosef, and Chris Stringer, eds. Pp. 15–20. Cam-bridge: McDonald Institute for Archaeological Research.

———. 2011. The origin of our species. London: Allen Lane.———. 2012. The status of Homo heidelbergensis (Schoetensack 1908). Evo-

lutionary Anthropology 21:101–107.Tishkoff, Sarah A., Floyd A. Reed, Francoise R. Friedlaender, Christopher

Ehret, Alessia Ranciaro, Alain Froment, Jibril B. Hirbo, et al. 2009. Thegenetic structure and history of Africans and African Americans. Science324:1035–1044.

Trauth, Martin H., Juan C. Larrasoana, and Manfred Mudelsee. 2009. Trends,rhythms and events in Plio-Pleistocene African climate. Quaternary ScienceReviews 28:399–411.

Trinkaus, Erik. 1983. The Shanidar Neandertals. New York: Academic Press.———. 2006. Modern human versus Neandertal evolutionary distinctiveness.

Current Anthropology 47:597–620.Tryon, Christian A., and Sally McBrearty. 2006. Tephrostratigraphy of the

Bedded Tuff Member (Kapthurin Formation, Kenya) and the nature ofarchaeological change in the later Middle Pleistocene. Journal of HumanEvolution 65:492–507.

Weaver, Timothy D. 2012. Did a discrete event 200,000–100,000 years agoproduce modern humans? Journal of Human Evolution 63:121–126.

Weaver, Timothy D., Charles C. Roseman, and Chris B. Stringer. 2007. WereNeandertal and modern human cranial differences produced by naturalselection or genetic drift? Journal of Human Evolution 53:135–145.

———. 2008. Close correspondence between quantitative- and molecular-genetic divergence times for Neandertals and modern humans. Proceedingsof the National Academy of Sciences of the USA 105:4645–4649.

White, Tim D., Berhane Asfaw, David DeGusta, Henry Gilbert, Gary D. Rich-ards, Gen Suwa, and F. Clark Howell. 2003. Pleistocene Homo sapiens fromMiddle Awash, Ethiopia. Nature 423:742–747.

Williams, Martin, Michael Talbot, Paul Aharon, Yassin Abdi Salaam, FrancesWilliams, and Knut Inge Brendeland. 2006. Abrupt return of the summermonsoon 15,000 years ago: new supporting evidence from the lower WhiteNile Valley and Lake Albert. Quaternary Science Reviews 25:2651–2665.

S234 Current Anthropology Volume 54, Supplement 8, December 2013

� 2013 by The Wenner-Gren Foundation for Anthropological Research. All rights reserved. 0011-3204/2013/54S8-0007$10.00. DOI: 10.1086/673752

Variability in the Middle Stone Ageof Eastern Africa

by Christian A. Tryon and J. Tyler Faith

CA� Online-Only Material: Supplement A

Eastern Africa is an important area to study early populations of Homo sapiens because subsets of those populationslikely dispersed to Eurasia and subsequently throughout the globe during the Upper Pleistocene. The Middle StoneAge (MSA) archaeology of this region, particularly aspects of stone-tool technology and typology, is highly variablewith only rare cases of geographic and temporal patterning. Although there are differences in timing and perhapsfrequency of occurrence, those elements that make up the MSA lithic tool kit are also found at contemporaneoussites elsewhere in Africa and Eurasia, making it difficult to identify a unique archaeological signal for hominindispersals out of eastern Africa. Rather, regional variation appears to be the outcome of possibly long-term interactionsbetween particular physical and social environments experienced by hominin populations.

The archaeological record of eastern Africa has the potentialto play a central role in our understanding of the behavioralevolution of modern human populations. The fossil recordfrom this region includes the earliest specimens attributed toHomo sapiens ∼195–154 ka (Clark et al. 2003; McDougall,Brown, and Fleagle 2005; White et al. 2003). Both fossil andgenetic evidence are consistent with this region providing thesource population(s) for subsequent dispersals out of Africaand feature prominently in models favoring the “southernroute” from Africa to Arabia (reviewed in Beyin 2011), witheach dispersing group sampling a portion of the biologicaland behavioral variability present in the parent population(Gunz et al. 2009; Prugnolle, Manica, and Balloux 2005; cf.Lycett and von Cramon-Taubadel 2008). Whereas the fossiland genetic data provide the best insights into past biologicalvariation, it is the archaeological record that provides therichest source of information on the behavioral variability offossil hominins.

All of the early fossils of H. sapiens from eastern Africa areassociated with Middle Stone Age (MSA) artifacts. These in-clude those from the Kibish Formation (Brown and Fuller2008; McDougall, Brown, and Fleagle 2005; Shea 2008), the

Christian A. Tryon is Assistant Professor in the Department ofAnthropology at Harvard University (Peabody Museum ofArchaeology and Ethnology, 11 Divinity Avenue, Cambridge,Massachusetts 02138, U.S.A. [christiantryon@fas.harvard.edu]). J.Tyler Faith is Postdoctoral Fellow in the Archaeology Program atthe School of Social Science of the University of Queensland(Brisbane, Queensland 4072, Australia [j.faith@uq.edu.au]). Thispaper was submitted 3 VII 13, accepted 20 VIII 13, and electronicallypublished 13 XII 13.

Herto Member of the Bouri Formation (Clark et al. 2003;White et al. 2003), Aduma (Haile-Selassie, Asfaw, and White2004; Yellen et al. 2005), and Porc Epic (Assefa 2006; Clarket al. 1984; Pleurdeau 2004; Vallois 1951) in Ethiopia andfrom the Upper Ngaloba Beds at Laetoli (Day, Leakey, andMagori 1980), Mumba Rockshelter (Brauer and Mehlman1988; Mehlman 1989), and perhaps the Lake Eyasi Beds (cf.Domınguez-Rodrigo et al. 2007, 2008; Mehlman 1987) inTanzania. We use the MSA archaeological record of easternAfrica to characterize the behaviors of hominin populationsthat included H. sapiens in addition to other archaic forms.This allows comparison with neighboring regions to betterunderstand the pattern and timing of archaeological diversityand innovation during the Middle and Upper Pleistocene.Put simply, is the archaeological record of eastern Africa con-sistent with the biological evidence that this region is thesource area for H. sapiens?

A few points of terminology require clarification. Followingthe International Commission on Stratigraphy, we refer to thetime interval bounded by the Brunhes-Matuyama paleomag-netic reversal dated to ∼781 ka and the onset of the LastInterglacial at ∼126 ka as the Middle Pleistocene. The UpperPleistocene lasts from the Last Interglacial until the Holoceneat ∼11.7 ka. By the MSA, we follow the common usage (e.g.,Clark 1988; Goodwin and Van Riet Lowe 1929; McBreartyand Brooks 2000) to refer to sites with lithic assemblages thatare characterized by stone or bone points and the frequentuse of Levallois methods for flake production. MSA sites lackthe large cutting tools such as cleavers and handaxes foundin Acheulian (Early Stone Age [ESA]) sites. Backed piecesmay be present but are less common than at Later Stone Age

Tryon and Faith Middle Stone Age Variability S235

(LSA) sites. Following McBrearty and Tryon (2006), the term“early” MSA (EMSA) refers to sites that predate the LastInterglacial. “Later” MSA (LMSA) sites date to within or afterthe Last Interglacial. Finally, “eastern Africa” refers to themodern-day countries of Tanzania, Kenya, Uganda, Ethiopia,Somalia, Djibouti, and Eritrea. Geographically, this region isbounded by the Indian Ocean, the Red Sea and a low elevationcoastal plain to the east, central highlands (11,000 m asl)dissected by the eastern arm of the Rift Valley, and the westernarm of the Rift to the west, with vegetation ranging fromdesert scrub to Afro-alpine forest. The area encompasses ∼3.6million km2; for comparison this is similar in size to westernEurope or India and is a third the size of the Sahara desertor the United States.

Eastern Africa as used here is defined by modern politicalboundaries as a matter of convenience, but as shown in figure1 and discussed below, most sites discussed in this paper alsoshare a common environmental context, occurring within ornear the boundaries of White’s (1983) Somali-Masai centerof regional endemism. We recognize that eastern Africa thusprovides a useful but imperfect geographic unit. Alternativeapproaches could emphasize different boundaries, for ex-ample, comparing Red Sea coastal sites such as Abdur inEritrea (Walter et al. 2000) with sites farther north and notincluded in this review such as Sodmein Cave in Egypt (Mer-cier et al. 1999; Vermeersch et al. 1994).

Sites are irregularly distributed throughout eastern Africa,largely concentrated within the East African Rift Valley system(fig. 1), as a result of geological exposure. We do not attemptto comprehensively review every known eastern African MSAarchaeological assemblage, but we emphasize those sites thatare sufficiently well published to determine the presence andabsence of archaeological attributes and have some degree ofchronological or stratigraphic control (see also Basell 2008).Summaries of these sites and their age estimates are providedin table 1, with their locations shown in figure 1. Stratigraphicand chronological control is particularly important becauseoverall site density is very low, with a complete lack of cov-erage for many areas; radiometric dates are few, and there isa rarity of caves or other deeply stratified sequences that allowready observation of change through time. In this, easternAfrica is different from western Europe, China, the Levant,or southern and northern Africa, and it relates largely tobedrock geology (i.e., abundant lavas and few limestone orquartzite deposits). As a result, much of the eastern Africanrecord consists of open-air sites used as living sites, huntinglocalities, areas for obtaining stone raw material, and otherfunctions (see, e.g., Tryon et al., forthcoming). Our inabilityto demonstrate contemporaneity among these sites in mostcases diminishes our ability to distinguish variation due toage as opposed to site function or environmental setting.

The irregular distribution of eastern African MSA sites isalso reflected in the discontinuous nature of their investiga-tion (Gabel 1984; Robertshaw 1990). There have been fewlong-term projects in the region since the seminal work con-

ducted in Uganda and Kenya in the 1920s and 1930s (e.g.,Leakey 1931, 1936; O’Brien 1939; Wayland 1934; Waylandand Burkitt 1932). The Central Rift Valley of Kenya is animportant exception, where multiple MSA sites have been thefocus of long-term study by Leakey (1931, 1936) and Isaacand his students (Isaac, Merrick, and Nelson 1972; Merrick1975), including Ambrose (1986, 2001, 2010) and others (e.g.,Anthony 1972; Waweru 2007). Surprisingly, despite investi-gation since the 1930s and the presence of relevant archae-ological material (Leakey et al. 1972; Mabulla 1990), OlduvaiGorge in Tanzania has played little role in our understandingof MSA sites in eastern Africa largely because of apathy onthe part of Mary Leakey (1984:213), the principal excavator:“In Africa, the hand axe culture did eventually give place toa surprisingly uninspiring group of industries lumped to-gether under the term Middle Stone Age; a stage in prehistoricarchaeology for which I have never been able to feel anyenthusiasm.” We hope that our paper, together with the othersreported in this volume, will serve to inspire new ideas andstimulate discussion concerning a time period that we believeis both interesting and important to understanding our evo-lutionary past.

Fossil Associations: The MSA Is NotExclusive to Homo sapiens

Although all of the early fossils of Homo sapiens are foundwith MSA artifacts, it is unlikely that our species was theexclusive author of MSA lithic technology. On present evi-dence, the oldest MSA sites in eastern Africa, at 1276 ka(Morgan and Renne 2008), are at least 70 kyr older than theoldest known H. sapiens fossil. The early fossils of H. sapiensand the populations they represent are highly variable, andthere is as yet no consensus on how to partition this variability(Gunz et al. 2009; Pearson 2008; Trinkaus 2005). Given thepossible presence of ancestral and sister taxa in the region(e.g., Hammer et al. 2011; Lachance et al. 2012), a morecautious reading of the available evidence would be that thevariability among MSA sites likely encapsulates the behavioraloutcomes of multiple hominin populations of varying taxo-nomic affinities. Direct linkage between particular homininsand specific archaeological entities is beyond the resolutionof our data, a problem that arises in other regions such asthe Levant (cf. Hovers 2009; Shea 2006a) and western Europe(cf. Slimak et al. 2011, 2012; Zwyns et al. 2012).

The MSA: Origins and Endings

The appearance and disappearance of MSA technologies canboth be defined as processes rather than events, typified bygradual, intermittent, and often complex patterns of changewith the loss of diagnostic ESA (Acheulian) or the additionof LSA elements over time. As reviewed below, this patternis consistent with technological change from existing, local

Figure 1. Sketch map showing eastern African Middle Stone Age sites discussed in the text and major biogeographic zones of White(1983).

Tryon and Faith Middle Stone Age Variability S237

antecedents and is in many ways comparable with the patternseen in western Europe. There, late Acheulian sites containLevallois technology with the number of handaxes decliningover time (Monnier 2006), and some Upper Pleistocene in-dustries may show the complex, perhaps nonlinear appear-ance of backed elements (Bordes and Teyssandier 2012) orretouched points (Slimak 2008).

The MSA Developed from Local Acheulian Antecedents

The overlap between age estimates for the earliest MSA andthe latest Acheulian sites supports the hypothesis of a pro-longed shift to MSA technologies. Gademotta, Ethiopia, is theoldest securely dated MSA site at 1276 ka (Morgan and Renne2008; Wendorf and Schild 1974). The youngest reportedAcheulian artifacts are surface collected from the Herto Mem-ber of the Bouri Formation of Ethiopia, dated to ∼154–160ka (Clark et al. 2003), and in situ material perhaps as recentas ∼125 ka from Abdur, Eritrea (Bruggeman et al. 2004; Walteret al. 2000). These young Acheulian sites are not without theirproblems. There remains the possibility that the Acheulianartifacts from the Herto Member are older than the datedsediments by an unknown but possibly large interval. TheAcheulian attribution of the material from Abdur is unfor-tunately not supported by detailed artifact descriptions orillustrations. However, if accurate, these results suggest a∼100–150 kyr overlap between Acheulian and MSA technol-ogies in eastern Africa.

The Kapthurin Formation, Kenya, currently provides thebest stratigraphic sequence showing the nature of the ap-pearance of MSA technologies within a single depositionalbasin (McBrearty and Tryon 2006; Tryon 2006; Tryon andMcBrearty 2006; Tryon, McBrearty, and Texier 2005) com-plemented by recent and ongoing work at Olorgesailie, Kenya(A. S. Brooks and J. E. Yellen, personal communication). Inthe Kapthurin Formation, sites with points and small Levalloiscores are interstratified with those with cleavers, suggestingthat Acheulian and MSA technologies overlapped temporallywithin the same geographic region (∼150 km2). Further, sev-eral elements of lithic technology found at MSA sites findtheir first expression in the Acheulian. These include the pro-duction of blades from cylindrical cores and particularly Le-vallois methods of flake production from assemblages withhandaxes and cleavers (Johnson and McBrearty 2010; Leakeyet al. 1969; McBrearty 1999; Tryon 2006). In the KapthurinFormation and elsewhere in eastern Africa (reviewed inSharon 2007; Tryon, McBrearty, and Texier 2005), Levalloistechnology formed one of several methods of producingAcheulian large flake blanks that could be transformed intoother tools. In each case, large (110 cm) Levallois flakes, oftenwith laterally retouched edges, were produced using the pref-erential method from centripetally prepared cores (fig. 2a).Levallois cores and flakes at younger sites in the KapthurinFormation are smaller and show a greater diversity of Levalloisapproaches (detailed below), perhaps linked to size reduction

of the desired Levallois flake blanks (Tryon, McBrearty, andTexier 2005). Finally, there is an apparent size gradient be-tween small (Acheulian) handaxes and large (MSA) points(McBrearty and Tryon 2006), consistent with a gradual shiftin artifact types (and perhaps functions) over time. Whetherthis size gradient masks different methods of production re-mains uninvestigated.

The End of the MSA

The end of the MSA was apparently a gradual but complexprocess rather than an event, with the emergence of the sub-sequent LSA developing from local MSA roots. At Enkapuneya Muto, Kenya, the sequence from ∼40 to 55 ka shows abasal MSA horizon with Levallois and discoidal methods offlake production and rare backed pieces. It is overlain by anindustry attributed to the LSA dominated by the productionof large (∼7 cm) backed blades and microliths, which is inturn overlain by an industry with abundant microliths (∼2–5 cm), MSA-like core reduction strategies, and ostrich eggshellbeads (Ambrose 1998).

In contrast, at Mumba Rockshelter, Tanzania, the strati-graphic sequence suggests a gradual change in the frequencyof typological and technologically important artifacts. Backedelements persist in low numbers across multiple strata, co-incident with a reduction in the frequency of Levallois coresand points and an increased use of bipolar percussion forflake production from ∼30 to 68 ka (Eren, Diez-Martin, andDomınguez-Rodrigo 2013; Gliganic et al. 2011; Marks andConard 2008; Mehlman 1989, 1991). The nature of the changeis such that the MSA or LSA attribution of a number ofindustries at Mumba is uncertain (Diez-Martın et al. 2009).Similar combinations of typically MSA (e.g., points) and LSA(e.g., backed pieces) artifacts are found at the Mochena Bo-rago sequence from Ethiopia (Brandt et al. 2012). As thesesites show, the apparent continuity of backed pieces amongstrata attributed to the MSA and LSA from eastern Africasites is distinct from the discontinuous appearance of backedpieces in southern Africa (cf. Howiesons Poort and Wiltonassemblages) or the late appearance of backed pieces in north-ern Africa (e.g., Close 2002; Deacon and Deacon 1999; Villaet al. 2010; Wurz 2013).

MSA Lithic Technological Variability

Stone tools and their manufacturing debris make up the bulkof our evidence for studying hominin behavioral variability.Table 1 summarizes this variability for a number of key datedsites in the region. As detailed below and recently emphasizedby Shea (2011b), the MSA record of lithic technology is highlyvariable from its first appearance. We first examine the var-iation within each of the major artifact classes summarizedin table 1. Moving from particular artifact types to artifactaggregates, we then conduct more formal analyses of inter-assemblage variation to more rigorously test hypotheses of

S238

Tabl

e1.

Art

ifac

tva

riab

ility

atea

ster

nA

fric

anM

iddl

eSt

one

Age

site

s

Site

/evi

den

ceLM

SA/

EM

SALa

t.Lo

ng.

Age

(ka)

Och

reB

eads

An

vil

Gri

nds

ton

eP

latf

orm

core

Dis

coid

alco

re

Leva

llois

pref

eren

tial

core

Leva

llois

recu

rren

tco

re

Leva

llois

poin

t/co

reB

ipol

arco

reB

lade

sB

acke

dpi

eces

Ret

ouch

edpo

int

Oth

erla

rge

bifa

ceH

eavy

-du

tyto

olM

isc.

reto

uch

ed%

reto

uch

edpi

eces

Mu

mba

Bed

III

low

er(N

aser

aIn

dust

ry)a

LMSA

�3.

5335

.32

30–4

0�

��

��

��

??

��

��

��

�5.

5M

um

baB

edV

(Mu

mba

In-

dust

ry)

LMSA

�3.

5335

.32

50–6

0�

��

��

��

??

��

��

��

�4.

9M

um

baB

edV

I-A

(Kis

ele

In-

dust

ry)a

LMSA

�3.

5335

.32

60–6

8�

��

��

��

?�

��

��

��

�13

.8M

um

baB

edV

I-B

(San

zako

In-

dust

ry)a

LMSA

�3.

5335

.32

168

��

��

��

�?

��

?�

��

��

10.5

Eya

siB

eds

EM

SA?

�3.

5235

.17

≥88–

132

��

��

��

�?

��

��

��

��

4.7

Nas

era

leve

ls6–

7(N

aser

aIn

dus-

try)

LMSA

�2.

7435

.36

118

–26a

��

��

��

�?

?�

��

��

��

3.5

Nas

era

leve

ls8/

9–11

(Mu

mba

Indu

stry

)LM

SA�

2.74

35.3

61

18–2

6a�

��

��

��

??

��

��

��

�10

Nas

era

leve

ls12

�25

(Kis

ele

Indu

stry

)LM

SA�

2.74

35.3

61

26–5

5a�

��

��

��

��

��

��

��

�5.

6W

asir

yaB

eds,

Ru

sin

gaIs

lan

dLM

SA�

0.42

34.1

738

�10

0�

��

��

��

��

��

��

��

�4.

1M

oche

na

Bor

ago

S-gr

oup

LMSA

6.90

37.7

643

��

�?

��

��

��

��

��

��

2.1

Moc

hen

aB

orag

oU

pper

T-g

rou

pLM

SA6.

9037

.76

43–4

5�

��

?�

��

��

��

��

��

�2.

7M

oche

na

Bor

ago

Low

erT

-gro

up

LMSA

6.90

37.7

645

–53

��

�?

��

�?

��

��

��

��

2.0

Mid

hish

i2

LSU

III

LMSA

10.9

547

.37

19–1

40?

?�

��

��

?�

��

��

��

�7.

3M

idhi

shi

2LS

UIV

LMSA

10.9

547

.37

40–9

0?

?�

��

��

?�

�?

��

��

�6.

2M

idhi

shi

2LS

UV

LMSA

10.9

547

.37

40–9

0?

?�

��

��

?�

�?

��

��

�3

Mid

hish

i2

LSU

VI

LMSA

10.9

547

.37

40–9

0?

?�

��

��

?�

�?

��

��

�7.

8P

orc

Epi

cLM

SA9.

6341

.87

33–7

9�

��

��

��

��

��

��

��

�8

Pro

lon

ged

Dri

ft(G

rJi1

1)LM

SA�

0.48

36.1

30–7

2�

��

��

�?

�?

��

��

��

�6.

1M

ugu

ruk

Form

a-ti

on,

Mem

ber

4,P

un

doM

ak-

war

Indu

stry

LMSA

?0.

0834

.63

110

–170

��

��

��

�?

��

��

��

��

1M

ugu

ruk

Form

a-ti

on,

Mem

ber

6,P

un

doM

ak-

war

Indu

stry

LMSA

?0.

0834

.63

110

–170

��

��

��

��

��

�?

��

�2.

4M

ugu

ruk

Form

a-ti

on,

Mem

ber

2,O

jalla

In-

dust

ryE

MSA

?0.

0834

.63

110

–170

��

��

��

�?

��

��

��

��

2.2

Mto

ngw

e,Lo

cal-

ity

KY-

55E

ast,

Upp

erG

rou

pLM

SA?

�4.

0839

.13

!12

5?�

��

��

�?

��

��

��

��

�7.

3

S239

Mto

ngw

e,Lo

cal-

ity

KY-

55E

ast,

Mid

dle

Gro

up

LMSA

?�

4.08

39.1

3!

125?

��

��

��

??

��

��

��

��

3.4

Mto

ngw

e,Lo

cal-

ity

KY-

55E

ast,

Low

erG

rou

pLM

SA?

�4.

0839

.13

!12

5?�

��

��

��

��

�?

��

��

�4.

1A

rdu

Bed

sB

/Cco

nta

ct,

Ad-

um

asi

teA

5LM

SA10

.42

40.5

280

–100

��

?�

��

�?

��

��

��

��

1A

rdu

Bed

sB

,A

d-u

ma

site

sA

4,A

8,A

8A,

VP

1/1,

VP

1/3

LMSA

10.4

240

.52

80–1

00�

��

��

��

?�

��

��

��

�2

Ard

uB

eds

B(b

ase)

,A

dum

asi

teA

1LM

SA10

.42

40.5

280

–100

��

��

��

�?

��

��

��

��

1.8

Kib

ish

Form

a-ti

on,

Mem

ber

II/I

II,

site

BN

SLM

SA5.

4135

.998

–123

��

��

��

��

�?

�?

��

��

1.6

Kib

ish

Form

a-ti

on,

Mem

ber

I,si

teA

HS

EM

SA5.

4135

.919

5–20

4?

��

��

��

?�

��

��

��

�1

Kib

ish

Form

a-ti

on,

Mem

ber

I,si

teK

HS

EM

SA5.

4135

.919

5–20

4�

��

��

��

?�

��

��

��

�5.

8K

apth

uri

nFo

r-m

atio

n,

site

Koi

milo

tLo

cus

2E

MSA

0.52

35.9

200–

250

��

��

��

��

��

��

��

��

0K

apth

uri

nFo

r-m

atio

n,

site

Koi

milo

tLo

cus

1E

MSA

0.52

35.9

200–

250

��

��

��

��

��

��

��

��

0G

adem

otta

For-

mat

ion

(ET

H72

-5)

EM

SA7.

9438

.6!

183

��

��

��

�?

��

��

��

��

2.2

Gad

emot

taFo

r-m

atio

n(E

TH

72-6

,E

TH

72-9

)E

MSA

7.94

38.6

!18

3�

��

��

��

��

��

?�

��

�1.

8G

adem

otta

For-

mat

ion

(ET

H72

-7B

,E

TH

72-1

)E

MSA

7.94

38.6

∼183

��

��

��

��

��

��

��

��

1.6

Gad

emot

taFo

r-m

atio

n(E

TH

72-8

B)

EM

SA7.

9438

.61

276

��

��

��

�?

��

��

��

��

3.8

Kap

edo

Tuff

sE

MSA

?1.

0736

.08

123–

135

��

��

��

��

��

��

��

��

3.1

Not

e.A

rtif

act

vari

abili

tyat

east

ern

Afr

ican

site

s:�

ppr

esen

t;�

pab

sen

t;?

pu

nce

rtai

n.

See

text

for

defi

nit

ion

san

dre

fere

nce

s.M

ult

iple

asse

mbl

ages

from

asi

ngl

esi

tesh

own

inst

rati

grap

hic

orde

r.Si

tes

are

attr

ibu

ted

toE

arly

MSA

(EM

SA)

orLa

ter

MSA

(LM

SA)

base

don

radi

omet

ric

date

s;u

nce

rtai

nti

esar

ise

for

un

date

dsi

tes,

thos

eth

atsp

anth

e∼1

26ka

bou

nda

rybe

twee

nth

eE

MSA

/LM

SA,

orw

her

ebi

ostr

atig

raph

icda

tadi

ffer

from

radi

omet

ric

age

esti

mat

es.

Lati

tude

and

lon

gitu

deex

pres

sed

asde

cim

alde

gree

s.a

Dat

afr

om19

38K

ohl-

Lars

enco

llect

ion

sra

ther

than

thos

eof

Meh

lman

(198

9);

thes

ear

elik

ely

over

esti

mat

es.

Figure 2. Selected cores and flakes from eastern African Middle Stone Age sites. a, Preferential Levallois core from the AcheulianFactory Site, Kapthurin Formation, Kenya. b, Preferential Levallois core from Gademotta, Ethiopia. c, Recurrent centripetal Levalloiscore from the Kapedo Tuffs, Kenya. d, Nubian core from the Wasiriya Beds, Rusinga Island, Kenya. e, Recurrent bipolar Levalloiscore from Nyogonyek, Kapthurin Formation, Kenya. f, Levallois flake from K’one, Ethiopia. g, h, Levallois points from Koimilot,Kapthurin Formation, Kenya, and Midhishi 2, Somalia. i, Discoidal core from the Wasiriya Beds, Rusinga Island, Kenya. j, Bladecore from the Kapedo Tuffs, Kenya. k, Single-platform core from Nasera Rockshelter, Tanzania. l, Bipolar core from NaseraRockshelter, Tanzania. Artifact illustrations after Gresham (1984); Kurashina (1978); Mehlman (1989); Tryon (2003); Tryon,McBrearty, and Texier (2005); Tryon, Roach, and Logan (2008); Tryon et al. (2012); and Wendorf and Schild (1974). Note that kand l use the lower scale bar; all others use the upper scale bar.

Tryon and Faith Middle Stone Age Variability S241

temporal and geographic variation within the eastern AfricanMSA.

Levallois Technology Is Highly Variable

Levallois technology in eastern Africa is as highly variable asit is in other well-studied regions such as western Europe orthe Levant (e.g., Delagnes and Meignen 2006; Hovers 2009).Although Levallois cores and flakes have been reported fromeastern African sites since at least the 1930s (e.g., Leakey1936), few sites in the region have been analyzed using thereconfigured understanding of the Levallois concept moststrongly associated with the work of Eric Boeda (1994, 1995).Briefly, the Levallois concept is an approach to flake produc-tion that targets the preparation and subsequent reduction ofa single core surface for the removal of relatively large andthick flake blanks (Eren and Lycett 2012). From this Levalloisflake removal surface, a single Levallois flake, blade, or pointis removed before repreparation of the convexities of the core(the preferential method), or multiple flakes, blades, or pointsare removed before repreparation (the recurrent method).The Levallois flake removal surface is shaped by the removalof “preparatory” flakes that alter core convexities. These con-vexities control the fracture pattern that in part dictates theform of the Levallois flake(s) removed from that surface (VanPeer 1992). Levallois flakes and preparatory flakes may beremoved using a number of different patterns, including re-movals from one direction (unidirectional), from oppositeends (bidirectional), about the circumference of the core (cen-tripetal), or subtle variations on these major themes (e.g.,unidirectional convergent flaking).

The combination of the particular Levallois method (pref-erential or recurrent) and the flake removal patterns (e.g.,unidirectional, bidirectional, centripetal) combine to producesubstantial variability within the Levallois approach to flakeproduction. Many of these variants are expressed at multipleMSA sites in eastern Africa (fig. 2b–2h). Importantly, variableapproaches to Levallois flake production are present at EMSAsites such as the Kapthurin Formation, Kenya (Tryon 2003,2006) and Gademotta/Kulkuletti, Ethiopia (Douze 2008;Wendorf and Schild 1974), as well as at LMSA sites, includingAduma (Yellen et al. 2005) and Porc Epic (Pleurdeau 2004)in Ethiopia and Rusinga Island in Kenya (Tryon et al. 2012).

Levallois points (fig. 2g, 2h) are present at several easternAfrican sites where they make up from ∼8% to 62% of re-covered Levallois flakes, such as Koimilot in the KapthurinFormation, Kenya (Tryon 2006); the Bird’s Nest Site (BNS)and Awoke’s Hominid Site (AHS) from the Kibish Formation,Ethiopia (Shea 2008); and Midhishi 2 in Somalia (Brandt andGresham 1989; Gresham 1984). This frequency is within therange of but is often greater than that found at some Levantinesites (cf. Hauck 2011; Hovers 2009:217). The Nubian Type 1method is a Levallois point variant distinguished by two elon-gated preparatory flakes removed from the distal end of thecore. Nubian Type 1 cores are prevalent to the north in the

Nile Valley and to the east in parts of the Arabian Peninsula(see Rose et al. 2011; Van Peer 1992, 1998) and provide someof the strongest archaeological evidence for connections be-tween Africa and Arabia in the Upper Pleistocene. NubianType 1 cores are found in eastern Africa at sites at or nearthe margins of the Nile drainage basin (fig. 2d), includingK’one and Aduma in Ethiopia (Kurashina 1978; Yellen et al.2005) and Rusinga Island in Kenya (Tryon et al. 2012), doc-umenting an extensive range for this Levallois variant. SomeLevallois approaches, such as that used for blade productionat some South African sites (Wurz 2002, 2013), are not foundin eastern Africa.

Beyond Levallois: Other Flake Production Methods

Although Levallois technology is a critical part of our un-derstanding of MSA sites, other forms of flake productionpersist, and additional, non-Levallois methods were intro-duced (table 1). Discoidal and single- and multiple-platformcores are widespread at MSA sites (fig. 2i, 2k). Discoidal coresresult from the alternate flaking of both sides of the peripheryof (typically) an oval cobble, resulting in a bifacially flaked,biconical core. Platform cores (including the “migrating planecores” of White and Pettitt 1995) result from the use of oneor more edges as striking platforms. Discoidal and platformcores are found at ESA (both Oldowan and Acheulian) sitesand form a technological substrate for the production ofsharp-edged flakes in some LSA assemblages (e.g., Mehlman1989). Bipolar cores (fig. 2l) resulting from the productionof small flakes using an anvil are also known from ESA (Ol-dowan) sites (e.g., de la Torre 2004). Bipolar cores occurirregularly at MSA sites in eastern Africa, including Naseraand Mumba rockshelters in Tanzania (Diez-Martın et al. 2009;Eren, Diez-Martin, and Domınguez-Rodrigo 2013; Mehlman1989) and Cartwright’s site in Kenya (Waweru 2007). Bladeor bladelet production (fig. 2j) occurs at EMSA assemblagesat Gademotta/Kulkuletti (Wendorf and Schild 1974) andLMSA assemblages at Aduma and Porc Epic in Ethiopia(Pleurdeau 2004; Yellen et al. 2005) and elsewhere. As notedpreviously, blades also occur in ESA (Acheulian) assemblagesin eastern Africa. Truncated-facetted pieces used for the pro-duction of small flakes are common at some Levantine andEuropean Middle Paleolithic sites (papers in McPherron2007) and some eastern African LSA sites (e.g., Newcomerand Hivernel-Guerre 1974). Although rare and probably un-derreported from eastern Africa, truncated-facetted pieces arereported from the MSA at Gademotta/Kulkuletti (Wendorfand Schild 1974:89); K’one Locality 5, Ethiopia (Kurashina1978); and perhaps Lukenya Hill (Clark 1988) and ProlongedDrift (Merrick 1975), Kenya.

Points: Functional, Spatial, and Temporal Variability

Along with the frequent use of Levallois technology, pointsare a defining element of the MSA. Point forms at eastern

S242 Current Anthropology Volume 54, Supplement 8, December 2013

African sites are highly variable in size and shape (fig. 3a–3d). “Point” refers to a broad category of artifacts made ofstone including unretouched, unifacial, and bifacial imple-ments made on Levallois and other flake blanks. The term isboth a morphological description (a pointed artifact) and anethnographically based functional inference (as the tip of aspear or other hunting implement). Studies of point shape,microwear patterns, and mastic traces from sites in easternAfrica and adjacent areas indicate that many points werehafted and probably used to tip spears, darts, or even arrows(e.g., Brooks et al. 2006; Donahue, Murphy, and Robbins2002–2004; Shea, 2006b; Van Peer, Rots, and Vermeersch2008; Waweru 2007).

We cannot assume that all points were used as armaturesor projectiles. Gademotta is the only eastern African MSA sitesubjected to two independent analyses of artifact function(Douze 2008; Wendorf and Schild 1993). Both analyses foundthat typologically defined points were used as cutting toolsrather than as spear tips, serving as an important reminderabout the potential dangers of inferring stone-tool functionfrom artifact form. Villa and Lenoir (2006; see also Villa,Delagnes, and Wadley 2005) have shown almost the oppositefor the European Middle Paleolithic record, where some “con-vergent scrapers” show impact damage consistent with theiruse as the tips of thrusting spears. Microwear analyses ofLevantine Levallois points emphasize the diversity of cuttingtasks served by these tools in addition to their possible useas spear tips (Beyries and Plisson 1998).

Whatever their function, for retouched pieces, EMSA av-erage point length ( mm, ) is significantly54.25 � 16.40 n p 50larger than LMSA points ( mm, ;42.22 � 14.30 n p 250Mann-Whitney U-test: , ; table A1 in CA�z p 5.222 P ! .001online supplement A). These size differences may reflectchanges in tool function, including the evolution of complexprojectile technology, which may appear at LMSA sites ∼40–100 ka (Brooks et al. 2006; Shea 2006b). Some stratigraphicsequences such as Aduma and Nasera show a monotonicdecrease in point size over time, whereas Mumba, Gademotta/Kulkuletti, and Gorgora rockshelters (Leakey 1943) do not(table A1).

Clark (1993) and McBrearty and Brooks (2000) have em-phasized geographic variation among MSA points at the sub-continental scale, although formal definitions or tests of theextent of many of these variants remain to be done. TheLupemban is one of the most distinct MSA regional variants,characterized by large (110 cm), thin, bifacially flaked lan-ceolate points (fig. 3a). Originally defined from sites in centralAfrica, Lupemban lanceolates are found as far east as the LakeVictoria region of Kenya. Although poorly dated in easternAfrica, the large size of Lupemban lanceolates suggests attri-bution to the EMSA, consistent with U-series age estimatesof 170–270 ka for Lupemban assemblages in Zambia (Barham2000) and the 110–170 ka age estimate from sedimentationrates published by McBrearty (1988) for western Kenya. Otherpoints from eastern African MSA sites are smaller but mor-

phologically highly variable and are not attributed to namedlarger archaeological entities equivalent to the Lupemban.While the small, subtriangular bifacially flaked forms (fig. 3b–3d) are distinct from lanceolate points or tanged pieces fromLupemban or Aterian sites, similar forms occur as far westas Mali (Soriano et al. 2010) and as far south as Botswana(Coulson, Staurset, and Walker 2011), reducing their utilityas a unique regional artifact form.

Other Shaped Tools

Although used to define MSA technology, points and indeedall forms of shaped or retouched tools are rare at MSA sites,typically making up !5%–7% of the flaked artifact total (table1). This is also true of many southern African MSA sites (e.g.,Thackeray 1989), and the rarity of retouch has made it dif-ficult to directly apply typologies such as that of Bordes (1961)that emphasize shaped or modified tools (see discussion inVilla, Delagnes, and Wadley 2005). In some areas, retouchfrequency is directly linked to the presence of fine-grainedraw materials, with retouch being rare on lava artifacts butmore common on those made of chert or similar rocks(Tryon, Roach, and Logan 2008). The Bordes system has beensuccessfully applied to sites in northern Africa where chert iswidespread (e.g., Hublin, Tillier, and Tixier 1987) and in east-ern Africa to sites such as Gademotta/Kulkuletti, Ethiopia(Wendorf and Schild 1974), where obsidian was used nearlyexclusively. Despite the relative rarity of retouched imple-ments, three tool classes are important to understanding MSAlithic technology: heavy-duty tools, scrapers, and backedpieces.

Several MSA sites have “heavy-duty tools” (sensu Clark2001b) such as picks (fig. 3h). These tools are also found inAcheulian or other earlier regional industries or industrialcomplexes such as the Sangoan and are likely a retention ofcharacteristic ESA technologies (table 1). These MSA sitesinclude Koimilot in the Kapthurin Formation (Tryon 2006)and the Kapedo Tuffs of Kenya (Tryon, Roach, and Logan2008), assemblages from Kibish Formation of Ethiopia (sur-face collected and not from the localities listed in table 1;Shea 2008), and the 168–130 ka basal Bed VI at MumbaRockshelter, Tanzania (Gliganic et al. 2011; Mehlman 1989:194). Similar tools also occur at the undated sites of Muguruk(McBrearty 1988), FxJi 61 near Koobi Fora (Kelly 1996:159),and the basal MSA levels at Mtongwe in Kenya (Omi 1986,1988).

As a tool class, scrapers have been reported from some ofthe earliest archaeological sites. However, in a qualitativesense, most scrapers from African Oldowan, Acheulian, andmany MSA sites are characterized by rare and irregular re-touch. These are very different, for example, from the classicscrapers defined by Bordes. It is only at MSA sites that somescraper forms appear that show continuously retouched edgesused to alter the shape of the tool (fig. 3f), a difference inform that may result from extending the use-life of hafted

Tryon and Faith Middle Stone Age Variability S243

Figure 3. Selected tools and beads from eastern African Middle Stone Age sites. a, Lupemban lanceolate point from Muguruk,Kenya. b, Point with basal thinning from Nasera Rockshelter, Tanzania. c, Point from Porc Epic, Ethiopia. d, Point from BNS,Kibish Formation, Ethiopia. e, Point with resharpening flake, Gademotta, Ethiopia. f, Scraper from Gademotta, Ethiopia. g, Grindstonefragment from Mumba Rockshelter, Ethiopia. h, Pick from Kapedo Tuffs, Kenya. i, Ostrich eggshell beads and production fragmentsfrom Mumba Rockshelter, Tanzania. j–m, Backed pieces from Mumba Rockshelter, Tanzania (j, k); Mtongwe, Kenya; and Enkapuneya Muto, Kenya. Artifact illustrations after Ambrose (1998); Clark et al. (1984); McBrearty (1986); Mehlman (1989); Omi (1986);Shea (2008); Wendorf and Schild (1974). Note that a through h use the upper scale bar; i through m use the lower scale bar.

implements through resharpening (cf. Keeley 1982), suggestedin particular by distinctive resharpening flakes found at Gad-emotta/Kulkuletti (Wendorf and Schild 1974; fig. 3e).

Backed pieces first appear at LMSA sites in eastern Africa.Backed pieces are flakes or blades with one lateral edge madesteep or blunted (“backed”) by abrasion or direct percussion.

Comparisons with historical examples (Clark, Phillips, andStaley 1974), experimental work (Clark and Prince 1978), andrare traces of ochre likely used as mastic (Ambrose 1998)suggest that backing is performed to facilitate hafting into aslotted wooden shaft consistent with findings from sites else-where (e.g., Villa and Soriano 2010). In eastern Africa, backed

S244 Current Anthropology Volume 54, Supplement 8, December 2013

pieces first appear ∼120–130 ka at the site of Deighton’s Cliff,Kenya (Ambrose and Deino 2010) but are a more regularfeature at LMSA sites ≤80 ka (fig. 3j–3m; table 1), includingPorc Epic (Clark et al. 1984; Pleurdeau 2004), Mochena Bo-rago (Brandt et al. 2012), and Mumba and Nasera rockshelters(Mehlman 1989, 1991). Backed pieces are also present in smallquantities from the !55 ka basal layers at Enkapune ya Muto(Ambrose 1998), with undated examples from probable MSAstrata at Mtongwe, Kenya (Omi 1984, 1986, 1988), and KiseseII Rockshelter, Tanzania (Inskeep 1962; Mehlman 1989:365),although published details from the latter site are scant.

Ochre and Grindstones

Ochre (or other mineral pigments) and grindstones (fig. 3g)are two key elements of MSA lithic technology. These twoartifact classes often co-occur (table 1), suggesting a functionalassociation. Ochre staining has been reported from somegrindstones, including those from the Kapthurin Formationand Enkapune ya Muto, in Kenya (Ambrose 1998; McBreartyand Brooks 2000). The Kapthurin Formation example (fromsite GnJh-15) is the earliest reported occurrence of grind-stones and ochre from eastern Africa, dated to ∼284–500 kaand associated with a lithic assemblage that cannot be con-fidently attributed to the Acheulian or MSA. Otherwise, grind-stones and ochre are found at the LMSA sites (table 1) ofAduma (Yellen et al. 2005) and Porc Epic (Clark et al. 1984)in Ethiopia; Mumba and Nasera rockshelters in Tanzania(Mehlman 1989); and Enkapune ya Muto in Kenya (Ambrose1998).

In addition to working ochre, grindstones may have alsobeen used to process seeds or other plant material, an activitywith considerably less archaeological visibility. Mercader(2009) reports starch grains from grindstones and other toolsfrom Ngalue, Mozambique, suggesting grass seed processing∼105 ka. The extent to which these results can be applied togrindstones at eastern African MSA sites is unknown butshould be a focus of future research. As emphasized by Kuhnand Stiner (2001), the appearance of grindstones for seedprocessing in MSA sites implies a shift toward lower-returnfoodstuffs that require substantial energy investment and thusa change in the foraging ecology of hominin populations.

Other Behaviors

We synthesize three other attributes of MSA hominin pop-ulations in eastern Africa: foraging behavior, territorial rangeinferred from raw material treatments, and symbolic behavior.Specifically, symbolic behavior concerns the treatment of thedead and the use of ornaments.

Foraging Behavior

As many stone tools served either directly or indirectly in thefood quest, we expect shifts in hominin diet to be reflectedin lithic assemblage composition. Although stone tools are

abundant, direct evidence of eastern African hominin diet issparse. Plants likely made up the bulk of the diet of anyhominin population living at or near the equator (Kelly 1995),but at present there is no direct evidence for plant con-sumption from eastern African MSA sites. As Marean (1997)notes, our models of reconstructing past foraging systems inthe region are limited by the lack of modern or historic for-agers (rather than pastoralists) from tropical grasslands, anenvironment that characterizes much of eastern Africa nowand in the Pleistocene.

Site location and faunal data provide two alternative meansof investigation. Sites such as Porc Epic, Ethiopia, and Nasera,Tanzania, have been interpreted as overlook sites situated neargame pathways (Clark 2001a; Mehlman 1989). Hunters ap-pear to have used natural features such as topographic lows,streams, or springs to acquire game at site GvJm46 at LukenyaHill (Marean 1990; 1997; Miller 1979) and Rusinga Island(Jenkins et al. 2012; Tryon et al. 2010) in Kenya. Ambrose(2001) has argued that MSA populations in the central partof the Rift Valley in Kenya positioned themselves at or nearthe ecotone between grassland and forest habitats in order tobest access resources from both environments. The use ofcoastal environments is demonstrated by MSA artifacts em-bedded within an ∼125 ka coastal reef off the coast of Eritrea(Walter et al. 2000) and an undated but well-stratified MSAartifact sequence in coastal dune sands at Mtongwe, Kenya(Omi 1984, 1986, 1988, 1991). Despite the importance ofcoastal environments for many out-of-Africa dispersal sce-narios (e.g., Bulbeck 2007), these two sites provide the only,albeit sparse, evidence for use of these environments fromeastern African MSA sites.

The faunal assemblages from GvJm46 at Lukenya Hill(Marean 1990) and Porc Epic (Assefa 2006) provide the onlylarge, well-studied, and published archaeofaunal MSA assem-blages from eastern Africa. Human exploitation of large mam-mals is also documented at other sites, including RusingaIsland, Kenya (Jenkins et al. 2012; Tryon et al. 2010), andLoiyangalani, Tanzania (Thompson 2005). Although the sam-ple is small, these studies suggest that at least by the laterparts of the Pleistocene, MSA foragers hunted a variety oflarge and small ungulates and selectively transported meat-rich elements to central places such as caves for further pro-cessing and consumption. Long-distance carcass transport tocentral places may distinguish MSA foraging strategies fromthose documented at other ESA sites in East Africa (e.g., Faith,Domınguez-Rodrigo, and Gordon 2009).

Territory and Movement Inferred from Raw MaterialTransport Data

Site-to-source distances for stone raw material provide thebest empirical estimate of the size of the physical and sociallandscapes familiar to early hominin populations. Comparedwith ESA hominins, groups making MSA artifacts used finer-grained rocks, particularly obsidian, more frequently (Feblot-

Tryon and Faith Middle Stone Age Variability S245

Augustins 1990; Merrick, Brown, and Nash, 1994). An ex-tensive, ongoing program to geochemically characterize RiftValley obsidian sources and artifacts provides the most de-tailed raw material transfer data (Ambrose et al. 2012; Mer-rick, Brown, and Nash 1994; Negash, Brown, and Nash 2011),summarized in table A2 in CA� online supplement A). ESAsite-to-source distances are !60 km, whereas eastern AfricanMSA site-to-source distances exceed 300 km. This increasesuggests expanded physical and social landscapes throughwhich stone artifacts were carried by highly mobile foragersand/or transferred through exchange.

MSA hominins apparently regularly transported obsidiancores, flakes, and finished tools ≤30 km, and in the case ofPorc Epic, 139 km (table A2). Beyond this and up to a distanceof 305 km, obsidian frequency declines, and only finishedtools and (resharpening?) flakes are found. These differencesmay reflect shifts in raw material procurement strategiesdriven by increased source distance relative to a group’s ter-ritorial range (from provisioning of places to provisioning ofindividuals; Kuhn 2004) or perhaps the trade/exchange offinished pieces rather than cores among different groups. Al-though sample size is small, sites at a similar distance (∼130–200 km) from the nearest source within the Rift Valley (PorcEpic) and outside of Rift Valley to the west in the Lake Victoriaregion (Songhor and Muguruk) show very different patterns(table A2). Obsidian is more rare (8% vs. !0.5%) and limitedto tools and flakes in the Lake Victoria region. This may resultfrom the relative difficulty of movement across the steep, oftendensely vegetated margins of the Rift rather than along thegrassy, open valley floor, suggesting a possible biogeographiccontrol to hominin movement within the region. Similarly,in central Europe, open habitats are associated with greaterstone transport distances (≤300 km) than in the topograph-ically complex region of western Europe (Feblot-Augustins1993).

Symbolic Behavior: Mortuary Practices and Ornaments

The 154–167 ka hominin skulls from the Herto Member ofthe Bouri Formation, Ethiopia, represent the only example ofperi- or postmortem treatment of the dead found at easternAfrican MSA sites. Here, cut and polish marks on the skullshave been interpreted as evidence of mortuary practice (Clarket al. 2003; White et al. 2003). The defleshed hominin skullfrom the ∼500 ka Acheulian site of Bodo (White 1986) liesonly 30 km away, possibly indicating significant time depthfor similar behaviors in the region.

Beads are the earliest direct evidence for personal orna-mentation from eastern African sites (fig. 3i). At Porc Epic,gastropod opercula were arguably worn as beads, co-occurwith MSA artifacts, and have been directly dated by the AMSradiocarbon method to between ∼33 and 143 ka (Assefa, Lam,and Mienis 2008). Ostrich eggshell beads and manufacturingdebris have been reported from Bed V and lower Bed III atMumba Rockshelter, Tanzania, now dated to 30–60 ka by

Gliganic et al. (2011). Mehlman (1989) considered Bed V andlower Bed III to contain industries intermediate between theMSA and LSA, whereas Diez-Martın et al. (2009) and Eren,Diez-Martin, and Domınguez-Rodrigo (2013) ascribe theselayers to the LSA, making the association of these beads withthe MSA uncertain. Conard (2004) reports directly dated∼29–33 ka ostrich eggshell beads from MSA/LSA strata atMumba. At Enkapune ya Muto, ostrich eggshell beads dateto ∼40 ka from an MSA/LSA stratum that overlies an industryattributed to the LSA (Ambrose 1998). It is unclear whetherexamples from 130 ka at Kisese II Rockshelter should beattributed to the MSA or LSA (Inskeep 1962; Leakey 1983:21).

In short, there are few demonstrable examples of bead useby MSA hominins in eastern Africa, and the behavior is arelatively late phenomenon. The appearance of beads marksone of the few apparent sharp breaks in the MSA record.Whether this is due to the appearance of a neural mutationthat led to the development of language and modern humanbehavior (Klein 2009), a shift to more durable forms of per-sonal expression (as suggested by Kuhn and Stiner 2007), orsampling bias due to a small sample of caves or rocksheltersis unclear. Rare, well-preserved Holocene burials such asNjoro River Cave, Kenya (Leakey and Leakey 1950), are alsopowerful reminders of the widespread use of seed beads orother perishable materials unlikely to preserve at MSA sites,potentially exaggerating the importance of the use of ostricheggshell as a medium for bead production.

Interassemblage Variability

To explore the nature of interassemblage variability amongeastern African MSA sites, we use presence/absence data forartifact classes (listed in table 1). Although variable artifacttypologies in use among researchers can reduce the utility ofsuch approaches (Vermeersch 2001), the categories used hereare sufficiently broad to minimize this problem. The data areused to examine (1) temporal variability among EMSA andLMSA sites (e.g., Shea 2011b), (2) geographic variability acrosseastern African sites (Clark 1988; McBrearty and Brooks2000), and (3) local, site-specific sources of variation. In thefollowing analyses, we conservatively treat all artifact classeswhose presence is uncertain (those with a question mark intable 1) as absent.

A correspondence analysis illustrating the association ofdifferent MSA assemblages with different artifact classes (fig.4, top) reveals temporal patterning among EMSA and LMSAassemblages. The EMSA artifact assemblages overlap in mul-tivariate space with many of the LMSA assemblages, but thereis a subset of LMSA assemblages that are distinct (Axis 1scores 1 0). These include the LMSA assemblages dated to!75 ka from Mumba, Nasera Rockshelter, Porc Epic Cave,Mochena Borago, and the undated middle and upper assem-blages from Mtongwe, which differ from EMSA sites by themore frequent presence of beads, ochre, backed pieces, bipolar

Figure 4. Correspondence analysis of artifact assemblage composition (top; table 1) across MSA localities. Site abbreviations:MuIII p Mumba Bed III; MuV p Mumba Bed V; MuVI-A p Mumba Bed VI-A; MuVI-B p Mumba Bed VI-B; Ey p EyasiBeds; NaN p Nasera (Nasera Industry); NaM p Nasera (Mumba Industry); NaKi p Nasera (Kisele Industry); Ru p WasiriyaBeds; MiIII p Midhishi 2 LSU III; MiIV p Midhishi 2 LSU IV; MiV p Midhishi 2 LSU V; MiVI p Midhishi 2 LSU VI; PE pPorc Epic; PD p Prolonged Drift; MtU p Mtongwe Upper Group; MtM p Mtongwe Middle Group; MtL p Mtongwe LowerGroup; Ar1 p Ardu Beds B/C; Ar2 p Ardu Beds B; Ar3 p Ardu Beds B (base); KiBNS p Kibish site BNS; KiAHS p Kibish siteAHS; KiKHS p Kibish site KHS; Ka1 p Kapthurin Formation, Koimilot Locus 1; Ka2 p Kapthurin Formation, Koimilot Locus2; Ga1 p Gademotta Formation (ETH72-5); Ga2 p Gademotta Formation (ETH72-6, ETH72-9); Ga3 p Gademotta Formation(ETH72-7b, ETH72-1); Ga4 p Gademotta Formation (ETH72-8B); KT p Kapedo Tuffs. Artifact abbreviations: Misc p miscel-laneous retouched piece; P core p platform core; LP core p Levallois preferential core; D core p discoidal core; HD tool pheavy-duty tool; LR core p Levallois recurrent core. The relationship between intersite distance (km) and the Dice-Sorensencoefficient calculated between all MSA assemblages (bottom). Solid line represents least squares regression.

Tryon and Faith Middle Stone Age Variability S247

cores, blades, grindstones, and anvils (fig. 4, top; table 1).Many of these artifact types are also found in the early LSAat Enkapune ya Muto (Ambrose 1998), and the LMSA sitesin which they are found have been characterized by some astransitional between the MSA and the LSA (e.g., Diez-Martınet al. 2009; Marks and Conard 2008). Our quantitative analysissupports their characterization as distinct from other “typical”EMSA and LMSA assemblages.

If geography is a meaningful correlate of assemblage com-position, we expect assemblages from sites that are closertogether to be more similar to one another than they are tomore distant sites. To quantify similarity between MSA as-semblages, we calculated Dice-Sørensen coefficients for allpairs of assemblages, measuring distance using latitude andlongitude coordinates (listed in table 1). For any given pair,the Dice-Sørensen coefficient is calculated as , where2j/(a � b)j is the number of artifact classes that co-occur at assemblagesA and B, a is the number of artifact classes at site A, and bis the number of artifact classes at site B. As illustrated infigure 4 (bottom), there is a weak but significant inverse re-lationship between pairwise site distance and the Dice-Sørensen coefficients ( , ), meaning thatr p �0.194 P ! .001pairs of assemblages that are nearby are more similar thanpairs of assemblages that are distant. This correlation is largelydriven by the tendency for assemblages from the same site(distance p 0) to be very similar to each other (fig. 4, top).Removing these from the analysis results in a much weaker,although still significant, correlation ( , ).r p �0.084 P p .035To the extent that the presence and absence of particularartifact classes can be interpreted as a measure of regionalvariation, our results suggest that geography plays a role indriving assemblage variability but that its effect is minimal.

Both the correspondence analysis (fig. 4, top) and the re-lationship between geographic distance and assemblage sim-ilarity (fig. 4, bottom) indicate that assemblages from the samelocality tend to be more similar to each other than they areto other sites. This is confirmed by a Mann-Whitney U-teston Dice-Sørensen coefficients calculated for assemblage fromsame locality versus assemblages from different localities( , ). This implies that local factors are az p �7.906 P ! .001dominant force driving interassemblage variability. These lo-cal factors might include site function, stone raw materialquality or abundance, or more speculatively small or restrictednetworks of information exchange within which traditions ofartifact manufacture and use were shared and maintained.

Environmental Controls

The environment structures aspects of the material record ofhuman foragers and environmental change may explain someof the variability among eastern African MSA sites. Archae-ological sites are unevenly distributed among four differentenvironmental or biogeographic zones defined by the distri-butions of endemic flora and fauna (fig. 1). Most sites occurwithin White’s (1983) Somali-Masai center of regional en-

demism (SMCRE). The SMCRE is the familiar dry savannaof eastern Africa characterized by habitats that range fromAcacia-Commiphora deciduous bushland/thicket to semides-ert grassland and shrubland with distinctive arid-adaptedfauna such as the oryx (Oryx beisa) and Grevy’s zebra (Equusgrevyi; Grubb et al. 1999). Faunal data from eastern AfricanMSA sites consistently show hominin occupation of grasslandhabitats broadly similar to those found in the SMCRE (Tryonet al. 2010, 2012). Other sites occur in the Eastern Foresttransitional zone along the Indian Ocean coast, the Lake Vic-toria regional mosaic (LVRM), or near the ecotone betweenthe SMCRE and scattered highland Afromontane areas (fig.1).

We have a poor understanding of the relationship betweenenvironmental and behavioral variability in eastern Africa be-cause of disparate spatial and temporal scales among paleoen-vironmental and archaeological data sets (Blome et al. 2012).The composition and boundaries of the biogeographic zonesdefined here likely shifted over time, providing different en-vironmental conditions for local hominin populations (seediscussion in Basell 2008), and it is these local conditions thatour analyses of interassemblage variability suggest are mostimportant in structuring archaeological variability. Coarse-grained analyses of hominin demography suggest that com-pared with other regions of Africa, hominin populations ineastern Africa responded to environmental change by minorshifts in settlement location (Blome et al. 2012). On a smallerscale, Ambrose (2001) suggested that MSA hominins, partic-ularly in the Lake Nakuru/Naivasha basin of Kenya, may havetracked ecotonal boundaries as they shifted in elevation withenvironmental change.

Evidence from the LVRM provides another example of thepossible relation between environmental and archaeologicalvariability in eastern Africa. Modern distributions of a varietyof plant and animal taxa consistently indicate eastward dis-persal of forest taxa into and across the LVRM from heavilyforested regions in central Africa during humid (i.e., inter-glacial) phases (e.g., Kingdon 1981; Rodgers, Owen, andHomewood 1982; Wronski and Hausdorf 2008). Conversely,LVRM MSA archaeological sites consistently include arid-adapted fauna such as oryx (O. beisa) and Grevy’s zebra (E.grevyi) that are characteristic of the SMCRE, suggesting west-ward dispersal during dry (i.e., glacial) conditions (Faith etal. 2013; Tryon et al. 2012). Hominin populations may wellhave followed a similar environmentally mediated pattern ofrange shifts. The LVRM conspicuously marks the easternmostlimit of Lupemban MSA sites that are most numerous in theforested regions of central Africa and appear to be associatedwith forested paleoenvironments (Barham 2000; Mercader2002). In contrast, LVRM MSA sites that co-occur with arid-adapted fauna have small points like those found to the eastin the Rift Valley (Tryon et al. 2012), a connection furtherdemonstrated by rare obsidian artifacts at LVRM sites withRift Valley sources (table A1).

The extent to which modern environments provide precise

S248 Current Anthropology Volume 54, Supplement 8, December 2013

analogues for Pleistocene eastern Africa is uncertain. Fossilfauna from MSA sites include a number of specialized grazersthat became extinct by the Holocene, implying importantdifferences in animal communities and grassland composition(Faith et al. 2011, 2012; Marean 1992, 1997). Five extinctmammals are reported from MSA sites, including an aardvark,Orycteropus crassidens (Lehmann 2009; MacInnes 1956), andfour bovids characterized by extreme hypsodonty and/or bodymass: a relative of the wildebeest, Rusingoryx atopocranion(Faith et al. 2011; Pickford and Thomas 1984); the giantwildebeest Megalotragus sp. (Kelly 1996; Tryon et al. 2012);an extinct blesbok, Damaliscus hypsodon (Faith et al. 2012;Marean and Gifford-Gonzalez 1991); and the giant longhornbuffalo Syncerus antiquus (Marean 1992; Tryon et al. 2012).Although extant species are present, extinct taxa are numer-ically dominant at MSA strata from Rusinga Island andGvJm46 at Lukenya Hill in Kenya (Marean 1992; Tryon et al.2012). Their dominance implies that dry grassland or scrubhabitats were more common than the seasonally moist shortgrasslands found today.

As Blome et al. (2012) stress, environmental change acrosseastern Africa is asynchronous, with different areas experi-encing variable changes in moisture availability and thus hab-itat change. The archaeological response to this variability seenamong Pleistocene faunas and eastern African MSA sites ap-pears to be small-scale movements (e.g., range expansion,topographic shifts) as an adaptation to changing environ-mental conditions. At present, there is no strong evidence forenvironmental change as a driver of behavioral innovationamong eastern African MSA sites, but this may reflect a lackof stratified MSA sequences associated with detailed paleoen-vironmental data.

Discussion

The high degree of variability characteristic of eastern AfricanMSA lithic technology limits our ability to identify an ar-chaeological signal linking eastern African MSA human pop-ulations to those that migrated out of Africa. This problemis exacerbated by the fact that many technical elements usedto manufacture MSA/MP artifacts, such as percussion, shap-ing (faconnage), retouch, biface manufacture, and wood-working are also present in older, Acheulian sites. This com-mon technological foundation makes likely the independentinvention of particular artifact forms even among dispersedpopulations of large-brained hominins (Shea 2006a; see alsoLycett 2007). For example, Levallois technology apparentlydeveloped at Acheulian sites in Africa and Eurasia from mul-tiple independent pathways (Tryon, McBrearty, and Texier2005; White, Ashton, and Scott 2011). Perhaps because ofthis, the MSA record from eastern Africa consists of lithictypes and technologies that are also found at similarly agedsites in other parts of Africa and Eurasia. Our comparativeanalyses of the eastern African data suggest that assemblagesfrom a single site are more similar to one another than they

are to those from other sites (regardless of geographic dis-tance), emphasizing the high diversity among these sites andthe difficulty of identifying a regional signature using artifacttypology.

While variability poses challenges to identifying a geo-graphic signature unique to the eastern African MSA, thereis some evidence for temporal change. The identification oftemporal patterning among some LMSA (many !75 ka) as-semblages on the basis of the more frequent presence of beads,ochre, bipolar cores, anvils, grindstones, and blades (fig. 4,top) suggests important behavioral changes during the laterPleistocene. All of these elements characterize sites of theregional LSA, imply a prolonged shift to LSA technologies,and suggest that the 70–35 ka interval of major populationdispersals within and out of Africa (Soares et al. 2012) wasone characterized by important technological changes. Someof the artifact forms underlying this technological shift havebeen suggested as markers of out-of-Africa population dis-persals to southern and eastern Asia (Mellars 2006b, 2006c),although Neanderthal populations in Europe apparently in-dependently invented similar elements during the same timeinterval (d’Errico and Stringer 2011).

Several authors have noted the absence of a clear archae-ological “out-of-Africa” signature, stressing that there is littlerecognizably (northern or eastern) “African” about the ar-chaeological record of early Homo sapiens in Asia, Europe, orAustralia (e.g., Shea, 2011b; Vermeersch 2001). For example,the oldest members of our species outside of Africa are foundat Qafzeh, Israel, in association with artifacts that fit com-fortably within the Levantine Mousterian (Hovers 2009). Evi-dence for symbolic behavior from Qafzeh, including beadsand ochre-stained burials, predate the oldest comparable evi-dence in Africa by at least 40 kyr (Bar-Yosef Mayer, Vander-meersch, and Bar-Yosef 2009; Hovers et al. 2003; Taborin2003; Vanhaeren et al. 2006). Similarly, although the earliestEuropean fossils of H. sapiens have tropical body proportionsassociated with a recent African origin (Pearson 2000), nofeatures of the Aurignacian (or later) Upper Paleolithic in-dustries suggest a technological link to Africa. The early recordfrom Australasia similarly lacks technological affinities withthe African MSA, consisting largely of simple forms of flakeproduction common to all Pleistocene archaeological sites(e.g., Mulvaney and Kamminga 1999). Mellars (2006b, 2006c)has suggested backed pieces as a candidate artifact form, butrepeated reinvention (e.g., during the Upper Pleistocene withthe Howiesons Poort and the mid-Holocene with the Wiltonindustries in South Africa) argues against its use as a markerof population dispersal. Shea (2011a) suggests that complexprojectile technology (i.e., bow and arrow) may have facili-tated the spread of H. sapiens out of Africa. However, in theabsence of similarities in the preserved (i.e., stone) elementsof this technology, the hypothesis remains difficult to test andmakes independent evolution impossible to rule out.

Should we expect to find an “out of Africa” signal in thearchaeological record at all? The Paleolithic archaeological

Tryon and Faith Middle Stone Age Variability S249

record is the outcome of behaviors mediated by particularsocial and physical environments. We expect the archaeolog-ical record to reflect changes in either the social or physicallandscape, and dispersals out of Africa were likely associatedwith both. Groups dispersing into the Levant and other partsof Eurasia likely encountered territories occupied by Nean-derthal (and other) hominin populations. While the precisenature of this interaction has long been debated, novel socialenvironments may have acted as a catalyst for behavioralchange, either through innovation or emulation. It may bethat the reason the earliest H. sapiens in the Levant haveartifact assemblages like those of the Neanderthals is that thisrepresents a successful behavioral strategy for that area. Homosapiens was the first hominin in Australia, and that continent’sdistinctive biota represent a dramatic change in physical en-vironment, perhaps explaining why the archaeological recordfrom that area is quite different from contemporaneous sitesfound in Africa (or Europe or the Levant). We do not findan “out-of-Africa” archaeological signature because the east-ern African record represents adaptations to that region’sunique setting; with new social and physical environments wefind new archaeological signatures. The outcome appears tobe different behaviors for different regions or environments,at least for hominins using MSA and MP technology, a featurethat may be distinct from those using LSA and UP technology(see Kuhn and Stiner 2001; Mercader and Brooks 2001).

From this perspective, evidence for hominin occupation ofthe Arabian Peninsula is particularly interesting. Rose (2004)and Armitage et al. (2011) used the presence of bifacial toolsto link the Arabian and eastern African records. However, thesample size is small, and bifaces have appeared independentlymultiple times in different areas (e.g., Rose 2007). In whatwe consider the only convincing archaeological evidence link-ing Africa and Arabia, Rose et al. (2011) demonstrated strongtechnological similarities in the specific details of core prep-aration and Levallois point production of the Nubian Type 1method. These cores are found largely at sites in the NileValley and its drainage basin in northeastern Africa and atsites in Oman dated to a relatively humid interval during theLast Interglacial ∼106 ka. What is most striking about this isthat multiple lines of evidence demonstrate a contemporarydispersal of eastern African flora and fauna (reviewed in Roseet al. 2011). Initial populations in Arabia may simply reflectan expansion of “Africa out of Africa.” African-like physicalenvironments in Arabia mitigated the need to adapt to novelenvironments, and with no (known) prior hominin popu-lations in the Arabian Peninsula, the social environment mayhave remained relatively stable. This hypothesis has clear par-allels with Dennell and Roebroek’s (2005) concept of “Sa-vannahstan,” with initial hominins dispersing into Asia re-maining within African-like environments. Later (∼55 ka)sites from the Arabian peninsula during arid intervals lackNubian Type 1 cores and suggest instead the development ofregionally distinct variants in Arabia and Africa with envi-ronmental change (Delagnes et al. 2012).

An emphasis on social and environmental factors shifts ourexpectations in searching for the origins of “modern humanbehavior.” If the archaeological signature of early H. sapiensvaries in relation to different social and physical environ-ments, then we should expect temporally and spatially variablepatterning in the expression of those elements linked to be-havioral modernity. This is consistent with the irregular tem-poral-spatial distribution of the archaeological signatures as-sociated with modernity (d’Errico 2003; d’Errico and Stringer2011; Habgood and Franklin 2008; McBrearty and Brooks,2000) and parallels d’Errico and Stringer’s (2011) “culturalmodel” of modern human behavioral origins and Conard’s(2008) model for Mosaic Polycentric Modernity. By reducingemphasis on the link between biological and behavioral mo-dernity (e.g., Hovers and Belfer-Cohen 2006; Kuhn and Hov-ers 2006; Lieberman and Bar-Yosef 2005) and emphasizingthe situational nature of the archaeological evidence (see alsoHenshilwood and Marean 2003), these models are consistentwith several lines of evidence, including an African biologicalorigin for our species but a Eurasian origin for the Aurig-nacian (Mellars 2006a), and the presence at Neanderthal sitesof some behaviors classically linked with modern humans(d’Errico and Stringer 2011). Such a perspective has the ad-vantage of shifting approaches to the eastern African MSArecord from those that scrutinize it for evidence of “modernhuman behavior” or archaeological signals of dispersal to onethat emphasizes it for what it is: the behavioral traces of earlypopulations of H. sapiens and closely related taxa (cf. Shea2011b; comments in Henshilwood and Marean 2003).

Conclusions

From an archaeological perspective, the MSA record of easternAfrica is highly variable and contains no typological or tech-nological elements that are uniquely derived relative to otherregions. Some change may be the result of subtle populationmovements or shifts in relation to environmental changerather than innovation. This is superimposed on generaltrends of point size decrease and a record beginning with theLast Interglacial that occasionally contains backed pieces andbeads and the more frequent presence of grindstones andochre, among other artifact classes. Some of the similaritieswith other regions likely represent analogous behaviors (suchas the origin and spread of Levallois technology), whereasothers may indeed be homologous, such as the very particularbehavioral “recipes” that define Nubian Type 1 technologyfound only in northeastern Africa and southeastern Arabia.Our ability to address these issues will certainly increase asmore African assemblages are studied in ways comparablewith those from other regions and as the number of studiedand well-published sites increases. The observed variability ofthe eastern African MSA record reduces its utility in identi-fying any sort of archaeological marker for dispersals “out ofAfrica.” Rather, it represents the long-term outcome of a series

S250 Current Anthropology Volume 54, Supplement 8, December 2013

of local adaptations made by Middle and Upper Pleistocenepopulations that included Homo sapiens.

Acknowledgments

C. A. Tryon would like to thank Erella Hovers, Steve Kuhn,Leslie Aiello, Laurie Obbink, and all the conference partici-pants for the wonderful week in Sweden. Any synthesis suchas this is the outcome of years of reflection and discussion,and we would like to acknowledge the following for directlyor indirectly contributing to the ideas presented here: SallyMcBrearty, Alison Brooks, Erella Hovers, Steve Kuhn, SusanAnton, Terry Harrison, Dan Peppe, Pierre-Jean Texier, JohnKingston, Kay Behrensmeyer, Rick Potts, Stanley Ambrose,Curtis Marean, and Nick Conard. We thank Sheila Nightingalefor the artifact illustrations. Portions of the research presentedhere were supported by grants from the National ScienceFoundation (BCS-0118345, BCS-0841530, BCS-1013199), theWenner-Gren Foundation, the Leakey Foundation, the Na-tional Geographic Society, New York University, and the Uni-versity of Queensland. Finally, none of this would have beenpossible without the love and support of Rhonda Kauffman,Violet Tryon, and Cornel Faith.

References CitedAmbrose, Stanley H. 1986. Hunter-gatherer adaptations to non-marginal en-

vironments: an ecological and archaeological assessment of the Dorobomodel. Sprache und Geschichte in Afrika 7(2):11–42.

———. 1998. Chronology of the Later Stone Age and food production inEast Africa. Journal of Archaeological Science 25:377–392.

———. 2001. Middle and Later Stone Age settlement patterns in the CentralRift Valley, Kenya: comparisons and contrasts. In Settlement dynamics inthe Middle Paleolithic and Middle Stone Age. N. J. Conard, ed. Pp. 21–44.Tubingen: Kerns.

———. 2010. Coevolution of composite tool technology, constructive mem-ory, and language: implications for the evolution of modern human be-havior. Current Anthropology 51(suppl.):S135–S147.

Ambrose, Stanley H., and Alan Deino. 2010. Technological diversity and geo-chronology of Middle Stone Age industries in the Central Rift Valley, Kenya:implications for the evolution of modern human behavior. Paleoanthro-pology 2010:A1.

Ambrose, Stanley H., Philip Slater, Jeffrey Ferguson, Michael Glascock, andIan Steele. 2012. Obsidian source survey and Late Quaternary artifact sourc-ing in Kenya: implications for early modern human behavior. Paleoan-thropology 2012:A1–A2.

Anthony, Barbara. 1972. The Stillbay question. In Congres panafricain de pre-histoire, Dakar 1967: actes de 6e session. H. J. Hugot, ed. Pp. 80–82. Paris:Les Imprimeries Reunies de Chambery.

Armitage, S. J., S. A. Jasim, A. E. Marks, A. G. Parker, V. I. Usik, and H.-P.Uerpmann. 2011. The southern route “out of Africa”: evidence for an earlyexpansion of modern humans into Arabia. Science 331:453–456.

Assefa, Zelalem. 2006. Faunal remains from Porc-Epic: paleoecological andzooarchaeological investigations from a Middle Stone Age site in south-eastern Ethiopia. Journal of Human Evolution 51:50–75.

Assefa, Zelalem, Y. M. Lam, and Henk K. Mienis. 2008. Symbolic use ofterrestrial gastropod opercula during the Middle Stone Age at Porc-EpicCave, Ethiopia. Current Anthropology 49:746–756.

Barham, Lawrence S. 2000. The Middle Stone Age of Zambia, south-centralAfrica. Bristol: Western Academic & Specialist Press.

Bar-Yosef Mayer, Daniella E., Bernard Vandermeersch, and Ofer Bar-Yosef.2009. Shells and ochre in Middle Paleolithic Qafzeh Cave, Israel: indicationsfor modern behavior. Journal of Human Evolution 56:307–314.

Basell, Laura S. 2007. An exploration of the Middle Stone Age of Eastern

Africa and Excavations at Rambogo Rock Shelter, Kenya. PhD thesis, Uni-versity of Cambridge.

———. 2008. Middle Stone Age (MSA) site distributions in eastern Africaand their relationship to Quaternary environmental change, refugia and theevolution of Homo sapiens. Quaternary Science Reviews 27:2484–2498.

Beyin, Ammanuel. 2011. Upper Pleistocene human dispersals out of Africa:a review of the current state of the debate. International Journal of Evolu-tionary Biology 2011: article ID 615094, doi:10.4061/2011/615094.

Beyries, Sylvie, and Hughes Plisson. 1998. Pointes ou outils triangulaires?donnees fontionelles dan le Mousterien levantin. Paleorient 24:5–24.

Blome, Margaret Whiting, Andrew S. Cohen, Christian A. Tryon, Alison S.Brooks, and Joellen Russel. 2012. The environmental context for the originsof modern human diversity: a synthesis of regional variability in Africanclimate 150,000–30,000 years ago. Journal of Human Evolution 62:563–592.

Boeda, Eric 1994. Le concept Levallois: variabilite des methodes. Paris: CentreNational de la Recherche Scientifique.

———. 1995. Levallois: a volumetric construction, methods, a technique. InThe definition and interpretation of Levallois technology. H. L. Dibble andO. Bar-Yosef, eds. Pp. 41–68. Madison, WI: Prehistory Press.

Bordes, Francois. 1961. Typologie du paleolithique ancien et moyen. Publicationsde l’Institut de prehistoire de l’Universite de Bordeaux, Memoire no. 1Bordeaux: Delmas.

Bordes, Jean-Guillaume, and Nicholas Teyssandier. 2012. The Upper Paleo-lithic nature of the Chatelperronian in south-western France: archeo-stratigraphic and lithic evidence. Quaternary International 259:95–101.

Brandt, Steven A., Erich C. Fisher, Elisabeth A. Hildebrand, Ralf Vogelsang,Stanley H. Ambrose, Josephine Lesur, and Hong Wang. 2012. Early MIS 3occupation of Mochena Borago Rockshelter, southwest Ethiopian highlands:implications for Late Pleistocene archaeology, paleoenvironments and mod-ern human dispersals. Quaternary International 274:38–54.

Brandt, Steven A., and Thomas H. Gresham. 1989. L’age de la pierre enSomalie. L’Anthropologie 94:459–482.

Brauer, Gunter, and Michael Mehlman. 1988. Hominid molars from a MiddleStone Age at the Mumba Rock Shelter, Tanzania. American Journal of Phys-ical Anthropology 75:69–76.

Brooks, Alison S., John E. Yellen, Lisa Nevell, and Gidon Hartman. 2006.Projectile technologies of the African MSA: implications for modern humanorigins. In Transitions before the transition: evolution and stability in theMiddle Paleolithic and Middle Stone Age. Erella Hovers and Steven L. Kuhn,eds. Pp. 233–256. New York: Springer.

Brown, Francis H., and Chad R. Fuller. 2008. Stratigraphy and tephra of theKibish Formation, southwestern Ethiopia. Journal of Human Evolution 55:366–403.

Bruggeman, J. Henrich, Richard T. Buffler, Mireille M. M. Guillaume, RobertC. Walter, Rudo von Cosel, Berhane N. Ghebretensae, and Seife M. Berhe.2004. Stratigraphy, palaeoenvironments and model for the deposition ofthe Abdur Reef Limestone: context for an important archaeological sitefrom the last interglacial on the Red Sea coast of Eritrea. Palaeogeography,Palaeoclimatology, Palaeoecology 203:179–206.

Bulbeck, David. 2007. Where river meets sea: a parsimonious model for Homosapiens colonization of the Indian Ocean rim and Sahul. Current Anthro-pology 48:315–321.

Clark, J. Desmond. 1988. The Middle Stone Age of East Africa and the be-ginnings of regional identity. Journal of World Prehistory 2:235–305.

———. 1993. African and Asian perspectives on the origins of modern hu-mans. In The origin of modern humans and the impact of chronometric dating.M. J. Aitken, C. B. Stringer, and P. A. Mellars, eds. Pp. 148–178. Princeton,NJ: Princeton University Press.

———. 2001a. Ecological and behavioral implications of the siting of MiddleStone Age rockshelter and cave settlements in Africa. In Settlement dynamicsof the Middle Paleolithic and Middle Stone Age. N. J. Conard, ed. Pp. 91–98. Tubingen: Kerns.

———. 2001b. The earlier cultures: Middle and Earlier Stone Age, vol. 3 ofKalambo Falls prehistoric site. Cambridge: Cambridge University Press.

Clark, J. Desmond, Yonas Beyene, Giday WoldeGabriel, William K. Hart, PaulR. Renne, Henry Gilbert, Alban Defleur, et al. 2003. Stratigraphic, chro-nological and behavioural contexts of Pleistocene Homo sapiens from Mid-dle Awash, Ethiopia. Nature 423:747–752.

Clark, J. Desmond, James L. Phillips, and Preston S. Staley. 1974. Interpre-tations of prehistoric technology from ancient Egyptian and other Sources.1. Ancient Egyptian bows and arrows and their relevance for African Pre-history. Paleorient 2:323–388.

Clark, J. Desmond, and G. R. Prince. 1978. Use-wear on Later Stone Age

Tryon and Faith Middle Stone Age Variability S251

microliths from Laga Oda, Haraghi, Ethiopia, and possible functional in-terpretations. Azania 13:101–110.

Clark, J. Desmond, Kenneth D. Williams, Joseph W. Michels, and Curtis A.Marean. 1984. A Middle Stone occupation site at Porc Epic Cave, DireDawa. African Archaeological Review 2:37–72.

Close, Angela E. 2002. Backed bladelets are a foreign country. In Thinkingsmall: global perspective on microlithization. S. L. Kuhn and R. G. Elston,eds. Pp. 31–44. Archeological Papers of the American Anthropological As-sociation, vol. 12. Washington, DC: American Anthropological Association.

Conard, Nicholas J. 2004. An overview of the patterns of behavioural changein Africa and Eurasia during the Middle and Late Pleistocene. In From toolsto symbols: from early hominids to modern humans. F. d’Errico and L. Back-well, eds. Pp. 294–332. Witswatersrand, South Africa: Witswatersrand Uni-versity Press.

———. 2008. A critical view of the evidence for a southern African originof behavioural modernity. South African Archaeological Society Goodwin Se-ries 10:175–179.

Coulson, Sheila, Sigrid Staurset, and Nick Walker. 2011. Ritualized behaviorin the Middle Stone Age: evidence from Rhino Cave, Tsodilo Hills, Bot-swana. Paleoanthropology 2011:18–61.

Day, Michael H., Mary D. Leakey, and Cassian Magori. 1980. A new hominidfossil skull (L.H. 18) from the Ngaloba Beds, Laetoli, northern Tanzania.Nature 284:55–56.

Deacon, Hilary J., and Janette Deacon. 1999. Human beginnings in SouthAfrica: uncovering the secrets of the Stone Age. Walnut Creek, CA: Altamira.

Delagnes, Anne, and Lillianne Meignen. 2006. Diversity of lithic productionsystems during the Middle Paleolithic in France: are there any chronologicaltrends? In Transitions before the transition: evolution and stability in theMiddle Paleolithic and Middle Stone Age. Erella Hovers and Steven L. Kuhn,eds. Pp. 85–108. New York: Springer.

Delagnes, Anne, Chantal Tribolo, Pascal Bertran, Michel Brenet, Remy Cras-sard, Jacques Jaubert, Lamya Khalidi, et al. 2012. Inland human settlementin southern Arabia 55,000 years ago: new evidence from the Wadi SurdudMiddle Paleolithic site complex, western Yemen. Journal of Human Evolution63:452–474.

de la Torre, Ignacio. 2004. Omo revisited: evaluating the technological skillsof Pliocene hominids. Current Anthropology 45:439–465.

Dennell, Robin W., and Wil Roebroeks. 2005. An Asian perspective on earlyhuman dispersal from Africa. Nature 438:1099–1104.

d’Errico, Francesco. 2003. The invisible frontier: a multiple species model forthe origin of behavioral modernity. Evolutionary Anthropology 12:188–202.

d’Errico, Francesco, and Chris B. Stringer. 2011. Evolution, revolution orsaltation scenario for the emergence of modern cultures? PhilosophicalTransactions of the Royal Society B 366:1060–1069.

Diez-Martın, Fernando, Manuel Domınguez-Rodrigo, Policarpo Sanchez, Au-dax Z. P. Mabulla, Antonio Tarrino, Rebeca Barba, Mary E. Prendergast,and Luis de Luque. 2009. The Middle to Later Stone Age technologicaltransition in East Africa: new data from Mumba Rockshelter Bed V (Tan-zania) and their implications for the origin of modern human behavior.Journal of African Archaeology 7:147–173.

Domınguez-Rodrigo, Manuel, Fernando Dıez-Martin, Audax Mabulla, LuisLuque, Luis Alcala, Antonio Tarrino, Jose A. Lopez-Saez, Rebecca Barba,and Pastory Bushozi. 2007. The archaeology of the Middle Pleistocenedeposits of Lake Eyasi, Tanzania. Journal of African Archaeology 5:47–78.

Domınguez-Rodrigo, Manuel, Audax Mabulla, Luis Luque, Jeroen W. Thomp-son, William J. Rink, Pastory Bushozi, Fernando Dıez-Martin, and LuisAlcala. 2008. A new archaic Homo sapiens fossil from Lake Eyasi, Tanzania.Journal of Human Evolution 54:899–903.

Donahue, Randolph E., Michael L. Murphy, and Lawrency H. Robbins. 2002–2004. Lithic microwear analysis of Middle Stone Age artifacts from WhitePaintings Rock Shelter, Botswana. Journal of Field Archaeology 29:155–163.

Douze, Katja. 2008. Le Early Middle Stone Age en Ethiopie: le contexte deGademotta: analyse technologique du site ETH-72-6. Memoire de Master2, Universite de Toulouse II, Le Mirail.

Eren, Metin I., Fernando Diez-Martin, and Manuel Domınguez-Rodrigo.2013. An empirical test of the relative frequency of bipolar reduction inBeds VI, V, and III at Mumba Rockshelter, Tanzania: implications for theEast African Middle to Late Stone Age transition. Journal of ArchaeologicalScience 40:248–256.

Eren, Metin I., and Stephen J. Lycett. 2012. Why Levallois? a morphometriccomparison of experimental “preferential” Levallois flakes versus debitageflakes. PLoS ONE 7:29723.

Faith, J. Tyler, Jonah N. Choiniere, Christian A. Tryon, Daniel J. Peppe, and

David L. Fox. 2011. Taxonomic status and paleoecology of Rusingoryx ato-pocranion (Mammalia, Artiodactyla), an extinct Pleistocene bovid from Ru-singa Island, Kenya. Quaternary Research 75:697–707.

Faith, J. Tyler, Manuel Domınguez-Rodrigo, Adam D. Gordon. 2009. Long-distance carcass transport at Olduvai Gorge? a quantitative examination ofBed I skeletal element abundances. Journal of Human Evolution 56:247–256.

Faith, J. Tyler, Richard Potts, Thomas W. Plummer, Laura C. Bishop, CurtisA. Marean, and Christian A. Tryon. 2012. New perspectives on middlePleistocene change in the large mammal faunas of East Africa: Damaliscushypsodon sp. nov. (Mammalia, Artiodactyla) from Lainyamok, Kenya. Pa-laeogeography, Palaeoclimatology, Palaeoecology 361/362:84–93.

Faith, J. Tyler, Christian A. Tryon, Daniel J. Peppe, and David L. Fox. 2013.The fossil history of Grevy’s zebra (Equus grevyi) in equatorial East Africa.Journal of Biogeography 40:359–369.

Feblot-Augustins, Jehanne. 1990. Exploitation des matieres premieres dansl’Acheuleen d’Afrique: perspectives comportementales. Paleo 2:27–42.

———. 1993. Mobility strategies in the Late Middle Paleolithic of centralEurope and western Europe: elements of stability and variability. Journalof Anthropological Archaeology 12:211–265.

Gabel, Creighton. 1984. Notes on the history of archaeology in sub-SaharanAfrica. Boston University African Studies Center Working Papers, no. 86.Boston: Boston University African Studies Center.

Gliganic, Luke A., Zenobia Jacobs, Richard G. Roberts, Manuel Domınguez-Rodrigo, and Audax Z. P. Mabulla. 2011. New ages for Middle and LaterStone Age deposits at Mumba Rockshelter, Tanzania: optically stimulatedluminescence dating of quartz and feldspar grains. Journal of Human Evo-lution 62:533–547.

Goodwin, Astley J. H., and Clarence van Riet Lowe. 1929. The Stone Agecultures of South Africa. Annals of the South Africa Museum 27:1–289.

Gossa, Tegenu, Yonatan Sahle, and Agazi Negash. 2012. A reassessment of theMiddle and Later Stone Age lithic assemblages from Aladi Springs, southernAfar Rift, Ethiopia. Azania 47:210–222.

Gresham, Thomas H. 1984. An investigation of an Upper Pleistocene ar-chaeological site in northern Somalia. MA thesis, University of Georgia,Athens.

Grubb, Peter, Oliver Sandrock, Ottmar Kullmer, Thomas M. Kaiser, and Fried-emann Schrenk. 1999. Relationships between eastern and southern Africanmammal faunas. In African biogeography, climate change, and early hominidevolution. T. Bromage and F. Schrenk, eds. Pp. 253–267. Oxford: OxfordUniversity Press.

Gunz, Phillipp, Fred L. Bookstein, Philipp Mitteroecker, Andrea Stadlmayr,Horst Seidler, and Gerhard W. Weber. 2009. Early modern human diversitysuggests subdivided population structure and a complex out-of-Africa sce-nario. Proceedings of the National Academy of Sciences of the USA 106:6094–6098.

Habgood, Phillip J., and Natalie R. Franklin. 2008. The revolution that didn’tarrive: a review of Pleistocene Sahul. Journal of Human Evolution 55:187–222.

Haile-Selassie, Yohannes, Berhane Asfaw, and Tim D. White. 2004. Hominidcranial remains from Upper Pleistocene deposits at Aduma, Middle Awash,Ethiopia. American Journal of Physical Anthropology 123:1–10.

Hammer, Michael F., August E. Woerner, Fernando L. Mendez, Joseph C.Watkins, and Jeffrey D. Wall. 2011. Genetic evidence for archaic admixturein Africa. Proceedings of the National Academy of Sciences of the USA 108:15123–15128.

Hauck, Thomas C. 2011. The Mousterian sequence of Hummal and its ten-tative placement in the Levantine Middle Paleolithic. In The Lower andMiddle Paleolithic in the Middle East and neighboring regions. J.-M. Le Ten-sorer, R. Jagher, and M. Otte, eds. Pp. 309–323. Etudes et recherches ar-cheologiques de l’Universite de Liege 126. Liege: Centre de recherches ar-cheologiques de l’Universite de Liege.

Henshilwood, Christopher S., and Curtis W. Marean. 2003. The origin ofmodern human behavior: critique of the models and their test implications.Current Anthropology 44:627–651.

Hovers, Erella. 2009. The lithic assemblages of Qafzeh Cave. Oxford: OxfordUniversity Press.

Hovers, Erella, and Anna Belfer-Cohen. 2006. “Now you see it, now youdon’t”: modern human behavior in the Middle Paleolithic. In Transitionsbefore the transition: evolution and stability in the Middle Paleolithic andMiddle Stone Age. Erella Hovers and Steven L. Kuhn, eds. Pp. 295–304.New York: Springer.

Hovers, Erella, Shimon Ilani, Ofer Bar-Yosef, and Bernard Vandermeersch.

S252 Current Anthropology Volume 54, Supplement 8, December 2013

2003. An early case of color symbolism: ochre use by early modern humansat Qafzeh Cave. Current Anthropology 44:491–522.

Hublin, Jean-Jacques, Anee-Marie Tillier, and Jacques Tixier. 1987. L’humerusd’enfant Mousterien (Homo 4) du Jeble Irhoud (Maroc) dans son contextearcheologique. Bulletin et Memoire de la Societe d’Anthropologie de Paris 4:115–142.

Inskeep, Raymond R. 1962. The age of the Kondoa rock paintings in the lightof recent excavations at Kisese II Rock Shelter. In Actes du IVe Congresspanafricain de prehistoire et l’etude du quaternaire. G. Mortelmans and J.Nenquin, eds. Pp. 249–256. Tervuren: Musee Royal de l’Afrique Centrale.

Isaac, Glynn L., Harry V. Merrick, and Charles M. Nelson. 1972. Stratigraphicand archaeological studies in the Lake Nakuru basin, Kenya. Palaeoecologyof Africa and the Surrounding Islands 6:225–234.

Jenkins, Kirsten, J. Tyler Faith, Christian A. Tryon, Daniel J. Peppe, SheilaNightingale, Julian A. Ogondo, Cara Roure Johnson, and Steve Driese. 2012.New excavations of a Late Pleistocene bonebed and associated MSA artifacts,Rusinga Island, Kenya. Paleoanthropology 2012:A17.

Johnson, Cara Roure, and Sally McBrearty. 2010. 500,000 year old blades fromthe Kapthurin Formation, Kenya. Journal of Human Evolution 58:193–200.

Keeley, Lawrence H. 1982. Hafting and retooling: effects on the archaeologicalrecord. American Antiquity 47:798–809.

Kelly, Alison J. 1996. Intraregional and interregional variability in the EastTurkana (Kenya) and Kenyan Middle Stone Age. PhD thesis, Rutgers Uni-versity, New Brunswick, NJ.

Kelly, Robert L. 1995. The foraging spectrum. Washington, DC: SmithsonianInstitution Press.

Kingdon, Jonathon. 1981. Where have the colonists come from? a zoogeo-graphical examination of some mammalian isolates in eastern Africa. Af-rican Journal of Ecology 19:115–124.

Klein, Richard G. 2009. The human career. 3rd edition. Chicago: Universityof Chicago Press.

Kuhn, Steven L. 2004. Upper Paleolithic raw material economies at Ucagızlıcave, Turkey. Journal of Anthropological Archaeology 23:431–448.

Kuhn, Steven L., and Erella Hovers. 2006. General introduction. In Transitionsbefore the transition: evolution and stability in the Middle Paleolithic andMiddle Stone Age. Erella Hovers and Steven L. Kuhn, eds. Pp. 1–12. NewYork: Springer.

Kuhn, Steven L., and Mary C. Stiner. 2001. The antiquity of hunter-gatherers.In Hunter-gatherers: an interdisciplinary perspective. C. Panter-Brick, R. H.Layton, and P. Rowley-Conwy, eds. Pp. 99–142. Cambridge: CambridgeUniversity Press.

———. 2007. Body ornamentation as information technology: towards andunderstanding of the significance of early beads. In Rethinking the humanrevolution: new behavioural and biological perspectives on the origin and dis-persal of modern humans. Paul Mellars, Katie Boyle, Ofer Bar-Yosef, andChris Stringer, eds. Pp. 45–54. Cambridge: McDonald Institute.

Kurashina, Hiro. 1978. An examination of prehistoric lithic technology ineast-central Ethiopia. PhD dissertation, University of California.

Lachance, Joseph, Benjamin Vernot, Clara C. Elbers, Bart Ferwerda, AlainFroment, Jean-Marie Bodo, Godfrey Lema, et al. 2012. Evolutionary historyand adaptation from high-coverage whole-genome sequences of diverseAfrican hunter-gatherers. Cell 150:457–469.

Leakey, Louis S. B. 1931. The Stone Age cultures of Kenya Colony. Cambridge:Cambridge University Press.

———. 1936. Stone Age Africa: an outline of prehistory in Africa. Oxford:Oxford University Press.

———. 1943. The industries of the Gorgora Rock Shelter, Lake Tana. Journalof the East Africa Natural History Society 17:199–203.

Leakey, Margaret, Philip V. Tobias, John E. Martyn, and Richard E. F. Leakey.1969. An Acheulean industry with prepared core technique and the dis-covery of a contemporary hominid mandible at Lake Baringo, Kenya. Pro-ceedings of the Prehistoric Society 3:48–76.

Leakey, Mary D. 1983. Africa’s vanishing art: the rock paintings of Tanzania.New York: Doubleday.

———. 1984. Disclosing the past: an autobiography. New York: McGraw-Hill.Leakey, Mary D., Richard L. Hay, D. L. Thurber, R. Protsch, and R. Berger.

1972. Stratigraphy, archaeology, and age of the Ndutu and Naisiusiu Beds,Olduvai Gorge, Tanzania. World Archaeology 3:328–341.

Leakey, Mary D., and Louis S. B. Leakey. 1950. Excavations at the Njoro RiverCave. Oxford: Clarendon.

Lehmann, Thomas. 2009. Phylogeny and systematics of the Orycteropodidae(Mammalia, Tubulidentata). Zoological Journal of the Linnean Society 155:649–702.

Lieberman, Daniel E., and Ofer Bar-Yosef. 2005. Apples and oranges: mor-phological versus behavioral transitions in the Pleistocene. In Interpretingthe past: essays on human, primate and mammal evolution. D. E. Lieberman,R. J. Smith, and J. Kelley, eds. Pp. 275–296. Boston: Brill Academic.

Lycett, Stephen J. 2007. Why is there a lack of Mode 3 Levallois technologiesin East Asia? a phylogenetic test of the Movius-Schick hypothesis. Journalof Anthropological Archaeology 26:541–575.

Lycett, Stephen J., and N. von Cramon-Taubadel. 2008. Acheulean variabilityand hominin dispersals: a model-bound approach. Journal of ArchaeologicalScience 35:553–562.

Mabulla, Audax Z. P. 1990. Preliminary report on an archaeological surveyof the Ndutu Beds, Olduvai Gorge, Tanzania. Nyame Akuma 33:20–25.

MacInnes, Donald G. 1956. Fossil Tubulidentata from East Africa. Fossil Mam-mals of Africa 10:1–38.

Marean, Curtis W. 1990. Late Quaternary paleoenvironments and faunal ex-ploitation in East Africa. PhD dissertation, University of California.

———. 1992. Implications of late Quaternary mammalian fauna from Lu-kenya Hill (south-central Kenya) for paleoenvironmental change and faunalextinctions. Quaternary Research 37:239–255.

———. 1997. Hunter-gatherer foraging strategies in tropical grasslands:model-building and testing in the East African Middle and Later Stone Age.Journal of Anthropological Archaeology 16:189–225.

Marean, Curtis W., and Dianne Gifford-Gonzalez. 1991. Late Quaternaryextinct ungulates of East Africa and palaeoenvironmental implications. Na-ture 350:418–420.

Marks, Anthony E., and Nicholas J. Conard. 2008. Technology vs. typology:the case for and against a transition from the MSA to the LSA at MumbaCave, Tanzania. In Space and time: which diachronies, which synchronies,which scales? typology vs technology. T. Aubry, F. Almeida, A. C. Araujo, andM. Tiffagom, eds. Pp. 123–131. BAR International Series 1831. Oxford:Archaeopress.

McBrearty, Sally. 1981. Songhor: a Middle Stone Age site in western Kenya.Quaternaria 23:171–190.

———. 1986. The archaeology of the Muguruk Site, Western Kenya. PhDdissertation, University of Illinois, Urbana-Champaign.

———. 1988. The Sangoan-Lupemban and Middle Stone Age sequence atthe Muguruk site, western Kenya. World Archaeology 19:379–420.

———. 1999. Archaeology of the Kapthurin Formation. In Late Cenozoicenvironments and hominid evolution: a tribute to Bill Bishop. P. Andrews andP. Banham, eds. Pp. 143–156. London: Geological Society.

McBrearty, Sally, and Alison S. Brooks. 2000. The revolution that wasn’t: anew interpretation of the origin of modern human behavior. Journal ofHuman Evolution 39:453–563.

McBrearty, Sally, and Christian A. Tryon. 2006. From Acheulian to MiddleStone Age in the Kapthurin Formation, Kenya. In Transitions before thetransition: evolution and stability in the Middle Paleolithic and Middle StoneAge. E. Hovers and S. L. Kuhn, eds. Pp. 257–277. New York: Springer.

McDougall, Ian, Francis H. Brown, and John G. Fleagle. 2005. Stratigraphicplacement and age of modern humans from Kibish, Ethiopia. Nature 433:733–736.

McPherron, Shannon P., ed. 2007. Tools versus cores: alternative approaches tostone tool analysis. Newcastle: Cambridge Scholars.

Mehlman, Michael J. 1987. Provenience, age and associations of archaic Homosapiens crania from Lake Eyasi, Tanzania. Journal of Archaeological Science14:133–162.

———. 1989. Late Quaternary archaeological sequences in northern Tanzania.PhD dissertation, University of Illinois, Urbana-Champaign.

———. 1991. Context for the emergence of modern man in eastern Africa:some new Tanzanian evidence. In Cultural beginnings: approaches to un-derstanding early hominid lifeways in the African savanna. J. Desmond Clark,ed. Pp. 177–196. Monographien (Romisch-Germanisches Zentralmuseum,Forschunginstitut fur Vorund Fruhgeschichte), vol. 19. Bonn: Habelt.

Mellars, Paul. 2006a. Archaeology and the dispersal of modern humans inEurope: deconstructing the “Aurignacian.” Evolutionary Anthropology 15:167–182.

———. 2006b. Going east: new genetic and archaeological perspectives onthe modern human colonization of Eurasia. Science 313:796–800.

———. 2006c. Why did modern human populations disperse from Africaca. 60,000 years ago? a new model. Proceedings of the National Academy ofSciences of the USA 103:9381–9386.

Mercader, Julio. 2002. Forest people: the role of African rainforests in humanevolution and dispersal. Evolutionary Anthropology 11:117–124.

Tryon and Faith Middle Stone Age Variability S253

———. 2009. Mozambican grass seed consumption during the Middle StoneAge. Science 326:1680–1683.

Mercader, Julio, and Alison S. Brooks. 2001. Across forests and savannas: LaterStone Age assemblages from Ituri and Semliki, Democratic Republic ofCongo. Journal of Anthropological Research 57:197–217.

Mercier, Norbert, Helene Valladas, Laurence Froget, Jean-Louis Joron, PierreM. Vermeersch, Phillip Van Peer, and Jan Moeyersons. 1999. Thermolu-minescence dating of a Middle Paleolithic occupation at Sodmein Cave,Red Sea mountains (Egypt). Journal of Archaeological Science 26:1339–1345.

Merrick, Harry V. 1975. Change in Later Pleistocene lithic industries in easternAfrica. PhD dissertation, University of California.

Merrick, Harry V., Francis H. Brown, and William P. Nash. 1994. Use andmovement of obsidian in the Early and Middle Stone Ages of Kenya andnorthern Tanzania. In Society, culture, and technology in Africa, vol. 11(supplement). S. T. Childs, ed. Pp. 29–44. Philadelphia: MASCA.

Miller, Sheryl F. 1979. Lukenya Hill, GvJm 46, excavation report. NyameAkuma 14:31–34.

Monnier, Gilliane. 2006. The Lower/Middle Paleolithic periodization in west-ern Europe: an evaluation. Current Anthropology 47:709–744.

Morgan, Leah E., and Paul R. Renne. 2008. Diachronous dawn of Africa’sMiddle Stone Age: new 40Ar/39Ar ages from the Ethiopian Rift. Geology 36:967–970.

Mulvaney, Derek J., and Johan Kamminga. 1999. Prehistory of Australia. Wash-ington, DC: Smithsonian Institution Press.

Negash, Agazi, Francis H. Brown, and Barbara Nash. 2011. Varieties andsources of artefactual obsidian in the Middle Stone Age of the Middle Awash,Ethiopia. Archaeometry 53:661–673.

Newcomer, Michael H., and Francois Hivernel-Guerre. 1974. Nucleus sureclat: technologie et utilisation par differentes cultures prehistoriques. Bul-letin de la Societe Prehistorique Francaise 71:119–128.

O’Brien, Terence P. 1939. The prehistory of Uganda protectorate. Cambridge:Cambridge University Press.

Omi, Giichi, ed. 1984. Mtongwe: an interim report of the East and NortheastAfrican Prehistory Project 1982. Matsumoto, Japan: Shinshu University.

———, ed. 1986. Mtongwe: an interim report of the East and Northeast AfricanPrehistory Project 1984. Matsumoto, Japan: Shinshu University.

———, ed. 1988. Mtongwe and Mgonga: an interim report of the East andNortheast African Prehistory Project 1986. Matsumoto, Japan: Shinshu Uni-versity.

———, ed. 1991. An interim report of the East and Northeast African PrehistoryProject 1988/1989. Matsumoto, Japan: Shinshu University.

Pearson, Osbjorn M. 2000. Postcranial remains and the origin of modernhumans. Evolutionary Anthropology 9:229–247.

———. 2008. Statistical and biological definitions of “anatomically modern”humans: suggestions for a unified approach to modern morphology. Evo-lutionary Anthropology 17:38–48.

Pickford, Martin, and Herbert Thomas. 1984. An aberrant new bovid (Mam-malia) in subrecent deposits from Rusinga Island, Kenya. Proceedings of theKoninklijke Akademie des Wetenschappen B 87:441–452.

Pleurdeau, David. 2004. Gestion des matieres premieres et comportements tech-niques dans le Middle Stone Age africain: les assemblages lithiques de la grottedu Porc Epic (Dire Dawa, Ethiopie). BAR International Series 1317. Oxford:Archaeopress.

Prugnolle, Franck, Andrea Manica, and Francois Balloux. 2005. Geographypredicts neutral genetic diversity of human populations. Current Biology15:159–160.

Robertshaw, Peter, ed. 1990. A history of African archaeology. London: JamesCurrey.

Rodgers, William A., Charles F. Owen, and Katherine M. Homewood. 1982.Biogeography of East African forest mammals. Journal of Biogeography 9:41–54.

Rose, Jeffrey. 2004. The question of Upper Pleistocene connections betweenEast Africa and South Arabia. Current Anthropology 45:551–555.

———. 2007. The Arabian Corridor Migration Model: archaeological evi-dence for hominin dispersals into Oman during the Middle and UpperPleistocene. Proceedings of the Seminar for Arabian Studies 37:219–237.

Rose, Jeffrey I., Vitaly I. Usik, Anthony E. Marks, Yamandu H. Hilbert, Chris-topher S. Galletti, Ash Parton, Jean Marie Geiling, Victor Cerny, Mike W.Morley, and Richard G. Roberts. 2011. The Nubian Complex of Dhofar,Oman: an African Middle Stone Age industry in southern Arabia. PLoSONE 6:e28239.

Sharon, Gonen, 2007. Acheulian large flake industries: technology, chronology,and significance. Oxford: Archaeopress.

Shea, John J. 2006a. The Middle Paleolithic of the Levant: recursion andconvergence. In Transitions before the transition: evolution and stability inthe Middle Paleolithic and Middle Stone Age. E. Hovers and S. L. Kuhn, eds.Pp. 189–212. New York: Springer.

———. 2006b. The origins of lithic projectile point technology: evidence fromAfrica, the Levant, and Europe. Journal of Archaeological Science 33:823–846.

———. 2008. The Middle Stone Age archaeology of the Lower Omo ValleyKibish Formation: excavations, lithic assemblages, and inferred patterns ofearly Homo sapiens behavior. Journal of Human Evolution 55:448–485.

———. 2011a. The archaeology of an illusion: the Middle-Upper Paleolithictransition in the Levant. In The Lower and Middle Palaeolithic in the MiddleEast and neighboring regions. J.-M. Le Tensorer, R. Jagher, and M. Otte, eds.Pp. 169–182. Etudes et recherches archeologiques de l’Universite de Liege126. Liege: Centre de recherches archeologiques de l’Universite de Liege.

———. 2011b. Homo sapiens is as Homo sapiens was: behavioral variabilityvs. “behavioral modernity” in Paleolithic archaeology. Current Anthropology52:1–35.

Slimak, Ludovic. 2008. The Neronian and the historical structure of culturalshifts from Middle to Upper Paleolithic in Mediterranean France. Journalof Archaeological Science 35:2204–2214.

Slimak, Ludovic, John I. Svendsen, Jan Mangerud, Hugues Plisson, HerbjørnP. Heggen, Alexis Brugere, and Pavel Y. Pavlov. 2011. Late Mousterianpersistence near the arctic circle. Science 332:841–845.

———. 2012. Response to “Comment on late Mousterian persistence nearthe arctic circle.” Science 335:167.

Soares, Pedro, Farida Alshamali, Joana B. Pereira, Veronica Fernandes, NunoM. Silva, Carla Alfonso, Marta D. Costa, et al. 2012. The expansion ofmtDNA haplogroup L3 within and out of Africa. Molecular Biology andEvolution 29:915–927.

Soriano, Sylvain, Michel Rasse, Chantal Tribolo, and Eric Huysecom. 2010.Ounjougou: a long Middle Stone Age sequence in the Dogon country(Mali). In West African archaeology: new developments, new perspectives. P.Allsworth-Jones, ed. Pp. 1–14. BAR International Series 2164. Oxford: Ar-chaeopress.

Taborin, Yves. 2003. La mer et les premiers hommes modernes. In Echangeset diffusion dans la prehistoire mediterraneenne. B. Vandermeersch, ed. Pp.113–122. Paris: Editions du comite des travaux historiques et scientifiques.

Thackeray, Anne I. 1989. Changing fashions in the Middle Stone Age: thestone artefact sequence from Klasies River main site, South Africa. AfricanArchaeological Review 7:33–57.

Thompson, Jessica C. 2005. The impact of post-depositional processes onbone surface modification frequencies: a corrective strategy and its appli-cation to the Loiyangalani site, Serengeti Plain, Tanzania. Journal of Ta-phonomy 3:67–89.

Trinkaus, Erik. 2005. Early modern humans. Annual Review of Anthropology34:207–230.

Tryon, Christian A. 2003. The Acheulian to Middle Stone Age transition:tephrostratigraphic context for archaeological change in the Kapthurin For-mation, Kenya. PhD dissertation, University of Connecticut.

———. 2006. “Early” Middle Stone Age lithic technology of the KapthurinFormation (Kenya). Current Anthropology 47:367–375.

Tryon, Christian A., J. Tyler Faith, Daniel J. Peppe, David L. Fox, Kieran P.McNulty, Kirsten Jenkins, Holly Dunsworth, and William Harcourt-Smith.2010. The Pleistocene archaeology and environments of the Wasiriya Beds,Rusinga Island, Kenya. Journal of Human Evolution 59:657–671.

Tryon, Christian A., J. Tyler Faith, Daniel J. Peppe, William F. Keegan, KristenN. Keegan, Kirsten H. Jenkins, Sheila Nightingale, et al. Forthcoming. Siteson the landscape: paleoenvironmental context of late Pleistocene archae-ological sites from the Lake Victoria Basin, equatorial East Africa. Quater-nary International.

Tryon, Christian A., and Sally McBrearty. 2006. Tephrostratigraphy of theBedded Tuff Member (Kapthurin Formation, Kenya) and the nature ofarchaeological change in the later Middle Pleistocene. Quaternary Research65:492–507.

Tryon, Christian A., Sally McBrearty, and Pierre-Jean Texier. 2005. Levalloislithic technology from the Kapthurin Formation, Kenya: Acheulian originand Middle Stone Age diversity. African Archaeological Review 22:199–229.

Tryon, Christian A., Daniel J. Peppe, J. Tyler Faith, Alex Van Plantinga, SheilaNightingale, Julian A. Ogondo, and David L. Fox. 2012. Middle Stone Ageartifacts and associated fauna from Rusinga and Mfangano islands, LakeVictoria, Kenya. Azania 47:14–38.

Tryon, Christian A., Neil T. Roach, and M. Amelia V. Logan. 2008. The Middle

S254 Current Anthropology Volume 54, Supplement 8, December 2013

Stone Age of the northern Kenyan Rift: age and context of new archaeo-logical sites from the Kapedo Tuffs. Journal of Human Evolution 55:652–664.

Vallois, Henri. V. 1951. La mandibule humaine fossile de la Grotte du Porc-Epic pres Dire-Daoua (Abyssinie). L’Anthropologie 55:231–238.

Vanhaeren, Marian, Francesco d’Errico, Chris Stringer, Sarah L. James, Jon-athan A. Todd, and Henk K. Mienis. 2006. Middle Paleolithic shell beadsin Israel and Algeria. Science 1785–1788.

Van Peer, Philip. 1992. The Levallois reduction strategy. Madison, WI: Prehis-tory Press.

———. 1998. The Nile Corridor and the out-of-Africa model: an examinationof the archaeological record. Current Anthropology 39(suppl.):S115–S140.

Van Peer, Philip, Veerle Rots, and Pierre M. Vermeersch. 2008. A wasted effortat the quarry: wear analysis and interpretation of an MSA lanceolate pointfrom Taramsa-8, Egypt. Paleoanthropology 2008:234–250.

Vermeersch, Pierre M. 2001. “Out of Africa” from an Egyptian point of view.Quaternary International 75:103–112.

Vermeersch, Pierre M., Philip Van Peer, Jan Moeyersons, and Wim Van Neer.1994. Sodmein Cave site, Red Sea mountains (Egypt). Sahara 6:31–40.

Villa, Paola, Anne Delagnes, and Lyn Wadley. 2005. A Late Middle Stone Ageartifact assemblage from Sibudu (KwaZulu-Natal): comparisons with theEuropean Middle Paleolithic. Journal of Archaeological Science 32:399–422.

Villa, Paola, and Michel Lenoir. 2006. Hunting weapons of the Middle StoneAge and Middle Paleolithic: spear points from Sibudu, Rose Cottage, andBouheben. South African Humanities 18:89–122.

Villa, Paola, and Sylvain Soriano. 2010. Hunting weapons of Neanderthalsand early modern humans in South Africa: similarities and differences.Journal of Anthropological Research 66:5–38.

Villa, Paola, Sylvain Soriano, Nicholas Teyssandier, and Sarah Wurz. 2010.The Howiesons Poort and MSA III at Klasies River main site, Cave 1A.Journal of Archaeological Science 37:630–655.

Walter, Robert, Richard T. Buffler, J. Henrich Bruggemann, Mireille M. M.Guillaume, Seife M. Berhe, Berhane Negassi, Yoseph Libeskal, et al. 2000.Early human occupation of the Red Sea coast of Eritrea during the lastinterglacial. Nature 405:65–69.

Waweru, Veronica N. 2007. Middle Stone Age technology at Cartwright’s Site,Kenya. PhD dissertation, University of Connecticut.

Wayland, E. J. 1934. Rifts, rivers, rain and early man in Uganda. Journal ofthe Royal Anthropological Institute of Great Britain and Ireland 64:333–352.

Wayland, E. J., and M. C. Burkitt. 1932. The Magosian culture of Uganda.Journal of the Royal Anthropological Institute of Great Britain and Ireland62:369–390.

Wendorf, Fred, and Romauld Schild. 1974. A Middle Stone Age sequence fromthe Central Rift Valley, Ethiopia. Warsaw: Institute of the History of MaterialCulture, Polish Academy of Sciences.

———. 1993. Probable functions of Mousterian points and convergentsidescrapers in the Middle Stone Age of Ethiopia. Quaternaria Nova 3:39–51.

White, Frank. 1983. The vegetation of Africa, a descriptive memoir to accompanythe UNESCO/AETFAT/UNSO vegetation map of Africa. Paris: UNESCO.

White, Mark, and Paul Pettitt. 1995. Technology of early Palaeolithic westernEurope: innovation, variability, and a unified framework. Lithics 16:27–40.

White, Mark J., Nick M. Ashton, and Beccy Scott. 2011. The emergence,diversity, and significance of the Mode 3 (prepared core) technologies. InThe ancient human occupation of Britain. N. M. Ashton, S. G. Lewis, andC. B. Stringer, eds. Pp. 53–66. Amsterdam: Elsevier.

White, Tim D. 1986. Cut marks on the Bodo cranium: a case of prehistoricdefleshing. American Journal of Physical Anthropology 69:503–509.

White, Tim D., Berhane Asfaw, David DeGusta, Henry Gilbert, Gary D. Rich-ards, Gen Suwa, and F. Clark Howell. 2003. Pleistocene Homo sapiens fromMiddle Awash, Ethiopia. Nature 423:742–747.

Wronski, Torsten, and Bernhard Hausdorf. 2008. Distribution patterns of landsnails in Ugandan rain forests support the existence of Pleistocene forestrefugia. Journal of Biogeography 35:1759–1768.

Wurz, Sarah. 2002. Variability in the Middle Stone Age lithic sequence,115,000–60,000 years ago at Klasies River, South Africa. Journal of Archae-ological Science 29:1001–1015.

———. 2013. Technological trends in the Middle Stone Age of South Africabetween MIS 7 and MIS 3. Current Anthropology 54(suppl. 8):S305–S319.

Yellen, John E., Alison S. Brooks, David M. Helgren, Martha Tappen, StanleyH. Ambrose, Raymond Bonnefille, James K. Feathers, et al. 2005. Thearchaeology of Aduma Middle Stone Age sites in the Awash Valley, Ethiopia.Paleoanthropology 10:25–100.

Zwyns, Nicolas, Wil Roebroeks, Shannon P. McPherron, Adam Jagich, andJean-Jacques Hublin. 2012. Comment on “A Late Mousterian persistencenear the arctic circle.” Science 335:167.

Current Anthropology Volume 54, Supplement 8, December 2013 S255

� 2013 by The Wenner-Gren Foundation for Anthropological Research. All rights reserved. 0011-3204/2013/54S8-0008$10.00. DOI: 10.1086/673529

Roots of the Middle Paleolithic in Eurasia

by Steven L. Kuhn

In this paper I aim to define basic characteristics of the Eurasian Middle Paleolithic (MP) and trace their development.I first outline some generalized features of MP material culture and behavior. A key characteristic of the MP isdiversity in lithic technology, manifest at various temporal and spatial scales. Looking more closely, this diversity iscomposed of repeated recurrences of a restricted range of alternatives. Many of the varieties of lithic productioncommon in the MP first appeared during the early MP and some even earlier. Other typical behaviors includefrequent use of fire, ochre, and hafting. These are first seen in Eurasia during the later MP (marine isotope stage[MIS] 6 and 7), the period when the entire suite of MP behaviors seems to coalesce. Many shared patterns of MPbehavior reflect a set of adaptations shaped by Eurasian habitats and climates. Other aspects of spatial and temporalvariation are more likely to reflect demographic conditions or patterns of cultural transmission among and withingroups.

The Eurasian Middle Paleolithic (MP), like the hominins thatcreated it, was a highly evolved entity, the product of hundredsof thousands of years of cultural and biological evolution. TheMP was also very long lived and extraordinarily widespread.It is undeniably the record of a very successful set of adap-tations. In the end, the MP ceded place to other ways of lifeand other ways of doing things, but this is not an especiallynoteworthy characteristic: to date that has been the ultimatefate of all human cultural endeavors.

Within the context of the conference “Alternative Pathwaysto Complexity: Evolutionary Trajectories in the Middle Pa-leolithic and Middle Stone Age,” the goals of this paper areto define key features of the MP in Eurasia and to trace theirdevelopment in time and space. Below I describe some centralcharacters of MP material culture and behavior as evidencefor an evolved set of adaptations and then assess their timedepth. A salient feature of the MP record is technologicaldiversity, which is manifest at a variety of scales, spatial aswell as temporal. Even though technology varies, however,MP hominins seem to have drawn on a limited and consistentarray of alternatives. Much local variability appears to reflecttactical and strategic responses to pressures from mobility,foraging, and raw material availability. Variation within MPtechnology at a larger scale may stem from demographic con-ditions and their effects on cultural transmission. Other novelcharacteristics of the MP record include well-developed py-rotechnology, hearth-focused activities within sites, and lim-ited use of material culture (pigments) for signaling. Manycommon forms of MP lithic technology show considerable

Steven L. Kuhn is Professor at the School of Anthropology of theUniversity of Arizona (Building 30, Tucson, Arizona 85721-0030,U.S.A. [skuhn@email.arizona.edu]). This paper was submitted 3 VII13, accepted 14 VIII 13, and electronically published 21 XI 13.

time depth, first appearing during the early Middle Pleisto-cene, if not earlier. Other features, such as consistent use offire and ochre, are first evident in Eurasia during the laterMP (marine isotope stage [MIS] 6 and 7). It is during thisinterval at the end of the MP when the entire MP behavioral“package” seems to come together (see also Richter 2011).

What Are the Key Shared Features of the MPand How Old Are They?

In identifying the principal shared features of the MP, I focuson the later, “classic” Mousterian of the late Pleistocene. Thisperiod, from roughly 120 to 35 ka (MIS 5e through mid-MIS3), saw the most-evolved—and by virtue of their age, thebest-preserved—examples of MP behavior. Some compari-sons with other times and places are unavoidable: derivedfeatures of behavior only become apparent through compar-ison with earlier or later periods (i.e., the Lower and UpperPaleolithic) or with contemporaneous cultural developmentselsewhere (the African Middle Stone Age [MSA]). I begin bydiscussing simple features, ones that may be apparent froma single assemblage. I subsequently discuss complex featuresthat emerge only from larger-scale comparisons among sitesor data types.

We have a reasonable understanding of the origins and timedepth of some typical MP features. One can only speculateabout the history of other forms of behavior. Systematic con-sideration of the history of Paleolithic cultural behaviors, evenfor a phenomenon as well studied as lithic technology, islimited by several factors. The widespread application of theBordes typology as the principal instrument for analyzing MPlithic assemblages during the mid- and late twentieth centurygreatly constrained perceptions of variability within the MP.

S256 Current Anthropology Volume 54, Supplement 8, December 2013

Archaeologists’ tendency to describe what they consider themost salient or important characteristics of technologies andto minimize variability (Bar-Yosef and Van Peer 2009) alsomakes it difficult to compare the material findings. For ex-ample, we know a good deal about handaxes in the LowerPaleolithic and methods of flake production in the MP butcomparatively little about flake production in the Acheulean.A third issue concerns variation in the density of useful datapoints. Because of unavoidable taphonomic and geologic fac-tors (Surovell and Brantingham 2007; Surovell et al. 2009),the number of well-preserved localities inevitably declines asa function of age. Moreover, for ecological or geological rea-sons, entire intervals are missing or poorly represented in therecords of some regions. These gaps in knowledge, along withuneven application of various dating methods, make it dif-ficult to tell equally detailed stories about all features of be-havior and material culture.

Simple Characters

Diverse Flake-Production Systems

The MP is conventionally distinguished from the Lower Pa-leolithic based on the overwhelming numerical dominance ofsmall flake tools over large core tools. Over the past 25 yearsit has become clear that MP toolmakers used an extraordi-narily diverse set of methods to produce blanks for these flaketools. These include multiple variants of Levallois (Boeda1994, 1995) but also a variety of non-Levallois systems, in-cluding discoid, prismatic blade production, and the “Quina”method (Bar-Yosef and Kuhn 1999; Boeda 1991; Boeda,Geneste, and Meignen 1990; Bourguignon 1996, 1997; Del-agnes, Jaubert, and Meignen 2007; Hiscock et al. 2009; Per-esani 2003) and a range of less widely documented techniques(e.g., Delagnes 1993; Faivre 2012; Geneste and Plisson 1996;Slimak 1999). Although a single “modal” system of produc-tion may dominate a given assemblage, more than one systemis nearly always represented. In some instances these coex-isting systems yielded blanks with very different functionalproperties (Boeda 1986; Kuhn, Arsebuk, and Howell 1996;Meignen, Delagnes, and Bourguignon 2009, table 2.1; Shim-elmitz, Barkai, and Gopher 2011), whereas in other instancesthey seem to represent alternative means of obtaining similarforms (Meignen 2007a).

Core tools such as handaxes and cleavers, while generallyassociated with the Lower Paleolithic Acheulean, are not ab-sent from the MP. However, MP handaxes often differ incharacter from Acheulean artifacts with similar forms (Villa2009). Sometimes (as in the Micoquian and Mousterian ofAcheulean Tradition) bifaces are treated as large flake tools:the bifacial shaping is essentially preparation of a blank foradditional marginal retouch and resharpening (Boeda,Geneste, and Meignen 1990; Soressi 2002; Soriano 2000). Thesmall bifaces and leaf points of central and eastern EuropeanMousterian industries (Bosinski 1967; Conard and Fischer

2000; Ruebens 2006) are also fundamentally different fromAcheulean bifacial handaxes in size, function, and methodsof production.

Flake-tool-dominated assemblages without handaxes dateback to ca. 400 ka in many parts of Eurasia (e.g., Burdukiewiczand Ronen 2003; Fluck 2011; Olle et al. 2013; Peris 2007).1

In some early assemblages the absence of large core tools isan effect of raw materials. In areas such as central Italy andparts of central Europe, knappable stone most often occursin small pebble form, making it impractical to manufacturelarge handaxes and cleavers. However, absence of large coretools is not inevitably a reflection of raw material constraints.In northern Europe, an area known for its rich sources ofexcellent flint, assemblages with and without handaxes wereproduced during roughly the same intervals of the MiddlePleistocene (Ashton and McNabb 1992; White 2000). Thescarcity of large Lower Paleolithic handaxes east of Italy andwestern Germany (Santoja and Villa 2006:429) is also certainlynot a matter of there being no suitable stone available.

In Europe, intra-assemblage diversity of flake-tool-pro-duction systems seems to be most evident in late- or post-Acheulean assemblages, where it is associated with a declininguse of general purpose core tools and greater emphasis onthe production of more specialized flake tools. White andAshton (2003) characterize much of earlier Acheulean flakemanufacture in northern Europe as being somewhat unstan-dardized, with “moving” core platforms producing amor-phous or polyhedral cores. Evidence for intra-assemblage var-iation is limited, but this may be because the number ofwell-described assemblages is small. Olle et al. (2013) arguethat multiple knapping methods are typical of most of thelong Middle Pleistocene sequence at Sierra de Attapuerca. Thesite of Yarımburgaz in Turkey, dating to MIS 7, also shows avaried repertoire of methods for producing flakes (Kuhn, Ar-sebuk, and Howell 1996). In the Levant, assemblages knownas Mugharan or Acheulo-Yabrudian are characterized byQuina-like production, blade manufacture and bifacial shap-ing existing side by side in the same assemblages (Jelinek1990). Yabrudian and allied assemblages date back to at least250–400 ka (Gopher et al. 2010; Mercier and Valladas 2003)and possibly earlier. Although most investigators assign themas late Lower Paleolithic, these flake-tool-dominated assem-blages bear a strong resemblance to later Mousterian assem-blages from Europe (Dibble 1991; Jelinek 1982; Le Tensoreret al. 2007).

Individual technological variants have quite different his-tories. Some methods of blank production, such as discoid,extend back to the Oldowan (reviewed by Barsky 2009; seealso papers in Peresani 2003). Others have more recent be-ginnings. The origins of the Levallois method deserve special

1. Note that I am confining the discussion to late- and post-Acheuleanindustries. The very earliest assemblages in Eurasia are all dominated bysmall flake tools (e.g., Carbonell and Rodriquez 2006), but these arequalitatively different sorts of assemblages and technologies from the MP.

Kuhn Roots of the Middle Paleolithic in Eurasia S257

treatment because the method is so strongly identified withthe Mousterian (White and Ashton 2003; White, Ashton, andScott 2011). Very ancient beginnings are claimed for Levallois,though in some cases the putative ancestral technologies arevery different from better known Mousterian versions. Keyaspects of the Levallois method, including the strategy of “pre-determining” flake forms by shaping cores extensively, can befound in Acheulean industries dating to near the Lower/Mid-dle Pleistocene boundary (e.g., Goren-Inbar 2011; Sharon andBeaumont 2006; White, Ashton, and Scott 2011). Cores madeon handaxes, fitting most criteria of Boeda’s (1994, 1995)definition of Levallois, appear to be an integral part of late-Acheulean artifact production systems in both the Levant(DeBono and Goren-Inbar 2001) and parts of northern Eu-rope (Tuffreau 1995; White and Ashton 2003; White, Ashton,and Scott 2011). Researchers have also commented more gen-erally on potential ties between Levallois and handaxe man-ufacture, which seem to associate in time and space (Lycett2007; Schick 1994; Villa 2009).

It is very likely that the Levallois method has multiple or-igins in Eurasia and Africa (Otte 1995; Sharon 2007; Villa2001; White and Ashton 2003:605). It is present discontin-uously in European late-Acheulean and early MP assemblagesdating as far back as MIS 11 (Santoja and Perez Gonzalez2010; Villa 2009:266; White, Ashton, and Scott 2011), al-though its frequency varies geographically. “Proto-Levallois”artifacts are reported from many Middle Pleistocene sites (e.g.,Barroso Ruız et al. 2011; Copeland 1998; Slimak et al. 2008;White and Ashton 2003; White, Ashton, and Scott 2011).Local sequences, such as that at Orgnac 3, may show the localrefinement of Levallois production during the latter part ofthe Middle Pleistocene (Moncel, Moigne, and Combier 2005;Moncel et al. 2011), but these cases are unlikely to representthe sole origin of Levallois in Europe. Rather, the Levalloismethod probably crystalized gradually because of changingfunctional and strategic roles of flake production in homininadaptations, channeled by the fracture mechanics of isotropicstone (White, Ashton, and Scott 2011:57).

Even though Levallois did not originate in the MP, thetechnology did become more widespread, though not ubiq-uitous. It may also have become more diversified, as did othermethods of flake production (White, Ashton, and Scott 2011).Many new variants on the basic theme of Levallois appeared,or at least have been described, in later Mousterian assem-blages. When exactly this diversification began is uncertain.There is an impression that pre-Mousterian Levallois is com-paratively homogeneous—mainly of the centripetal recurrentor preferential variety (White, Ashton, and Scott 2011)—butthis could simply be a function of restricted numbers of earlycases.

Another less-well-studied but equally coherent technolog-ical variant is called here, for want of a better term, the Quina/Yabrudian “pattern.” Here we are dealing with a linked setof strategies for artifact production, retouch, and use. Meth-ods of blank production (which may be varied) emphasize

the manufacture of large, thick flakes, often with cortical orotherwise blunt margins opposing a sharp edge (Al Qadi 2011;Bourguignon 1996, 1997; Turq 2000). Shaped tools, mainlyscrapers, typically show very characteristic scaled “Quina” re-touch produced by soft-hammer percussion. The resultingtools have the potential for extensive resharpening, althoughthis potential is not always realized in discarded tools. TheYabrudian appears first in the Levant more than 400,000 yearsago. Shortly thereafter, similar strategies of artifact productionand use can be found in Anatolia (the “proto-Charentian” ofKarain Cave; Otte et al. 1995, 1998) and south-central Europe(e.g., Velika and Mala Balanica [Mihailovic and Mihailovic2009; D. Mihailovic, personal communication, 2012]). Even-tually the forerunners of Quina industries appear in westernEurope by MIS 7 (Ashton and McNabb 1992; Geneste andPlisson 1996; Moncel 2008). On one hand, the complexityand number of associated traits making up the Quina/Yabru-dian pattern suggest that it is unlikely to be a result of repeatedindependent invention. Dates are consistent with the notionthat it spread from east to west, either by diffusion or pop-ulation movement. At the same time it is difficult to reconcilethe notion of maintaining a coherent set of associated be-haviors over 200,000 years and thousands of kilometers withany known process of cultural transmission.

It is worth mentioning the time depth of blade production,given its importance to the definition of the Upper Paleolithic(e.g., Clark’s [1969] Mode 4). As has been widely docu-mented, prismatic blade production is not exclusive to theUpper Paleolithic but exhibits considerable time depth in theMP within Eurasia (Bar-Yosef and Kuhn 1999; Meignen2007a, 2007b; Revillion and Tuffreau 1994). A few fairly recentMP assemblages even show limited production of small,bladelet-sized blanks (Faivre 2012; Slimak 1999). Systematicproduction of blade blanks in Eurasia can be traced back tothe Amudian in the Levant, dated to as early as 350–400 ka(Gopher et al. 2010). However, prismatic or non-Levalloisblade production is also a broader category than Levallois,discoid, or Quina production. Because it encompasses a widerange of variation, many local manifestations of blade pro-duction probably represent independent developments.

Habitual Transport of Artifacts

MP hominins regularly carried artifacts with them as theymoved across the landscapes on which they foraged. This isshown most clearly by the presence of artifacts made of rawmaterials obtained from sources located at some distance fromthe eventual locus of discard. Artifacts moved more than 5–20 km almost always make up a minority of MP assemblages,but at least a few are normally present. Maximum distancesof transport are normally between 20 and 50 km, but theymay exceed 100 km in unusual cases (Feblot-Augustins 1997,2009; Gamble 1999; Geneste 1988, 1990; Roebroeks, Kolen,and Rensink 1988; Slimak and Giraud 2007). Flake blanksand retouched tools were most often carried long distances,

S258 Current Anthropology Volume 54, Supplement 8, December 2013

but occasionally core tools or even cores were moved as well.Habitual transport of artifacts demonstrates that MP homi-nins could and regularly did anticipate needs for certain kindsof artifacts in advance, and therefore (1) activities involvingproduction and use of tools were staged across the landscape,and (2) tool use was planned or at least regular enough todemand habitual accommodation. Artifact transport mayeven provide a rough index of scales of foraging ranges (Bran-tingham 2006).

The regular transport of tools is a very old pattern, ex-tending at least back to the early Acheulean or the Oldowan.However, the distances over which artifacts were transporteddo appear to have increased over time through the late LowerPaleolithic and MP (Feblot-Augustins 1997, 2009; Gamble1999; Roebroeks, Kolen, and Rensink 1988). At some point,hominins began to carry artifacts over distances longer thanthey could have covered in a single bout of foraging, showingthat they were planning beyond immediate needs. Certainly,the displacement of artifacts more than 50 km in some MPsites is evidence for thinking about needs for tools severaldays in advance. This expanded scale of technological plan-ning could reflect changes in cognition, for example, an in-crease in the capacity to project future possibilities. Alter-natively, it may be due to expansions in ranging patternsassociated with colonization of northern latitudes and gen-erally deteriorating climate conditions throughout MIS 4.

Hafting and (Simple) Composite Tools

Hafting is one of the few real novelties in Eurasian MP tech-nologies. Evidence for hafting comes from modifications toshaped tools, use-wear studies, and more rarely, from visualor chemical detection of mastics on stone artifacts. Becausehafting-related modification is not necessarily obvious, andbecause use-wear and residue evidence are fugitive, we do nothave a very clear idea of how often things were hafted. How-ever, even if they are not very common, what appear to behafting-related modifications or wear on tools have been doc-umented in many places (e.g., Rotts 2009). Natural bitumenwas used as a mastic by MP hominins in the northern Levant(Boeda et al. 1996, 2008) as well as in Europe (Carciumaruet al. 2011). Reports of birchbark pitch, which supposedlyrequires a more elaborate manufacture process (Grunberg2002; Mazza et al. 2006; Pawlik and Thissen 2011), are es-pecially impressive. The recognition that MP tool makerssometimes intentionally produced very small flakes (!1 cm;Dibble and McPherron 2006; Hovers 2007) could also beindirect evidence for hafting, although it may have other ex-planations.

Points or pointed pieces are one commonly hafted class ofartifact, but other forms also show evidence for having beenattached to handles. Hafted MP points tend to be large, andit is suggested that they would have been attached to largethrust or thrown spears or used as knives (Churchill 1993;Shea 2006). Direct evidence for use as spear tips is rather

irregularly distributed (see Boeda et al. 2008; Villa and Lenoir2009; Villa et al. 2009), probably because it is not alwayspreserved or recognized. More generally, while there werehafted spear points in the Mousterian, not all pointed objectswere tips of weapons (Plisson and Beyries 1998), and pointswere not the only things hafted (Pawlik and Thissen 2011).

The significance of composite tools has been discussed ina number of contexts. The manufacture of a simple spearwith a stone point involves several distinct chaınes operatoires(for shaping the wooden haft, knapping the stone point, andpreparing the binder). The integration of multiple elements,each formed by a different process, shows a degree of staged,hierarchical organization to activities. At the very least thisimplies more complex cognitive processes than needed toproduce simpler, one-part tools. Some have drawn connec-tions between the ordered assembly of composite tools andgrammatically structured speech (Ambrose 2010). The assem-bly of composite tools also implies greater investment of timeand energy in the production of some items of technology.

The earliest purported evidence of hafting in Europe comesfrom Schoningen (Germany), where simple wooden handlesreportedly slotted to accommodate flake inserts date to ca.400 ka (Thieme and Maier 1995). Use-wear traces, resin onstone tools, and impact fractures on stone points are morecommon in MP assemblages dating to the later Middle andearly Upper Pleistocene (Boeda et al. 1996, 1999; Mazza etal. 2006; Rotts 2009; Villa and Lenoir 2009; Villa et al. 2009).In the Near East the best direct evidence for hafting is bitumentraces on late Mousterian artifacts from Syria (Boeda et al.1996). Fractures consistent with impact damage are reportedon early Mousterian pointed tools dating back to 200–250 ka(Shea 1988; Villa and Lenoir 2009:71), although the sizes andshapes of these artifacts suggest that they were better suitedto use as knives than spearheads (Shea 2006). Much earlierassemblages generally contain few pointed objects or otherlikely candidates for hafting.

Pyrotechnology

Fire was an important part of the adaptations of MP homi-nins. Where conditions of preservation are right, MP sitesgenerally contain evidence of fire, whether as well-definedhearths, dispersed ash in sediments, or burned bones andartifacts (e.g., Berna and Goldberg 2007; Goldberg et al. 2012;Karkanas 2002; Vallverdu et al. 2012). Moreover, traces of firein sites are found throughout the Mousterian range, fromnorthern Europe to the southern Levant. It is difficult to saywhether or not there is a correlation between pyrotechnologyand climatic conditions, but even sites in the relatively warm,dry Levant contain abundant evidence for the use of fire. Inearly Mousterian deposits at Hayonim cave (ca. 220 ka), forexample, wood ash residue makes up a large part of the sed-iment even though the cultural deposits are comparativelylow density (Schiegel et al. 1996; Stiner et al. 2005).

There is convincing evidence for repeated burning in a

Kuhn Roots of the Middle Paleolithic in Eurasia S259

limited area at Gesher Benot Ya’acov (Israel) dating to ca. 800ka (e.g., Alperson-Afil 2008), but not all sites of similar agecontain such clear traces of fire. Burning is largely absent inthe earlier Middle Pleistocene deposits at Atapuerca, for ex-ample (Roebroeks and Villa 2011). Fire becomes a regularpart of the record starting about 350–400 ka and is virtuallyubiquitous after 200 ka in both the Levant and Europe. Thereis some evidence for an increase in the frequency of burningat Qesem cave (Karkanas et al. 2007) and Tabun (R. Shim-elmitz, personal communication 2012) around 300 ka. In theirsynthetic analysis of the European evidence, Roebroeks andVilla also record two stepwise increases in the frequency ofevidence for fire jumps, one during MIS 7 and the secondduring MIS 5 (Roebroeks and Villa 2011:5211). However,Sandgathe et al. (2011) argue that full control of fire, specif-ically the ability to kindle fires at will, may have come later,a contention that will be very difficult to disprove.

Use of Pigments

The use of pigments is another aspect of material culture thatin Eurasia first appears in the MP. It has long been knownthat mineral pigments in the form of ochre and other metallicoxides were present in MP sites (reviewed by d’Errico 2008;Soressi and d’Errico 2007). We have few direct clues as tohow these materials were used—whether they were employedto color bodies, clothing, and/or objects or whether theyplayed other roles as binders or preservatives. It is safe toassume that they were used to decorate and to otherwisechange the appearance of things. Pigments do not seem tobe strongly associated with burials in the MP.

As of this writing, the earliest well-dated use of pigmentsin Europe comes from the early Mousterian site of MaastrichtBelvedere, dated to 200–250 ka (Roebroeks et al. 2012). Pig-ments are a regular though not ubiquitous element of the MParchaeological record dating to the Upper Pleistocene(d’Errico 2007, 2008; Soressi and d’Errico 2007). Interestingly,there is little evidence for the use of pigments in the NearEast except during MIS 5, when they are associated with an-atomically modern Homo sapiens fossils (Bar-Yosef Mayer,Vandermeersch, and Bar-Yosef 2009; Hovers and Belfer-Cohen 2013).

Absent Characteristics

It is not especially useful to use negative evidence to definethe MP or anything else, but it is worth recalling that someclasses of technology and material culture that are commonlater on in Eurasia and that appear in some contemporaneousassemblages from southern Africa are very rare if not com-pletely absent from the Mousterian. Generally, there is verylittle obvious use of durable material culture as signaling me-dia in MP assemblages. Possible ornaments have been re-ported from a few late MP sites (e.g., d’Errico 2007; Zilhaoet al. 2010), but they never seem to have become regular and

widespread among MP hominins except possibly for a specificperiod in the Levant when early H. sapiens was present in theregion (see Hovers and Belfer-Cohen 2013). Perishable ma-terials may possibly have been used for body decoration insome instances (Morin and Laroulandie 2012; Peresani et al2011), but evidence is currently spotty, as would be expected.MP assemblages contain few signs that bone, antler, or ivorywere regularly used as raw material for shaped tools, andgrinding and polishing were seldom employed to modify thesematerials. In marked contrast to later periods, there is alsolittle apparent variation across space or environments in thelevel of investment in or elaboration of technology related tofood procurement (Kuhn and Stiner 2001).

It is important that forms of material culture such as or-naments or bone tools occasionally do appear in MP contexts.Although these are often included in the list of featuresthought to define “modern human behavior” (e.g., McBreartyand Brooks 2000; Shea 2011), they are behavioral variables,not fixed markers of the “modern human condition.” Beadsand bone artifacts are not ubiquitous within the Upper Pa-leolithic or later periods: the presence or absence of all ofthese traits are functions of a range of influences operatingat both individual and group levels (cf. Shea 2011). If MPhominins occasionally made ornaments or used antler as araw material, we can assume that they were capable of rec-ognizing and exploiting the potential of these materials. Thequestion thus becomes not one of capability but of why theydid not regularly take advantage of that potential.

Complex/Emergent Characters

Geographic and Temporal Diversity

Compared with earlier periods and with some later ones, too,the Mousterian as a whole shows a striking degree of variationin methods used for making stone artifacts. This diversity isdetectable both within and among assemblages. A large partof the interassemblage diversity is apparent within a geo-graphic frame of reference, but some of it is also chronolog-ical. It is notable that not all areas show comparable levels ofvariation. How much, if any, technical variation is linked toenvironmental factors remains uncertain.

From a macroscopic perspective, Eurasian MP material cul-ture, which we know mainly through lithic technology, formsdistinctive regional or geographic groupings (e.g., Clark 2009;Conard and Fischer 2000; Delagnes, Jaubert, and Meignen2007; Mellars 1996:334–335; Ruebens 2006). Although certaintechnological elements, such as Levallois or discoid technol-ogy, are widely diffused, there are regional differences in spe-cific aspects of production and resulting artifact forms. Geo-graphic and temporal variation in MP lithic technology donot involve rapid appearance and turnover of novel methods.Instead, MP toolmakers “cycled through” a number of com-mon themes (Levallois, discoid, laminar production, Quina,etc.). Technological procedures that appear as novel additions

S260 Current Anthropology Volume 54, Supplement 8, December 2013

to a local sequence were often used previously (and later) inother places. There are also notable interregional contrasts inthe amount and patterning of technological variation. Forexample, the late Mousterian of southern France shows anotable (perhaps extreme) level of variation in lithic tech-nology. This variability resolves into temporal trends as wellas functional/strategic associations (Delagnes, Jaubert, andMeignen 2007; Delagnes and Meignen 2005; Meignen, Del-agnes, and Bourguignon 2009). On the other side of the Pyr-enees (Casanova I Martı et al. 2009; de la Torre 2013), tech-nological variation is much more limited, and there are fewerobvious temporal trends. Some of the contrasts in diversityamong regions may be a function of differing sample sizes,but the differences in evidence for change over time amongregions are unlikely to be due simply to the numbers of well-characterized assemblages.

Regionalization in western Eurasia becomes apparent rel-atively early, ca. 400 ka. Around this time major differencesin lithic assemblages are apparent between northwest Europe(Acheulean, Clactonian), Iberia (Acheulean) and central Eu-rope (core and flake tool/small tool industries), and the Levant(Acheulean, Acheulo-Yabrudian). There may be more subtledifferences as well, such as contrasts in late-Acheulean flaketechnology or handaxe production, but these are difficult toisolate because of the low density of well-dated and well-reported sites/assemblages. Researchers have long recognizedvariation across Europe in the shapes of Acheulean handaxes,but strong arguments have been mounted that handaxe shapesare more representative of raw material differences (Santojaand Villa 2006) or artifact life histories (McPherron 2006).Regional diversity in technologies of lithic production andmanufacture continues to increase over the course of the MP.Impressionistically it appears that the highest level of geo-graphic differentiation in artifact production and shapingcomes with the late Mousterian (ca. 70–45 ka). Interestingly,however, regionally distinctive projectile point types are notreadily detectable until the terminal Mousterian or “transi-tional” industries (e.g., Flas 2008; Shea 2006; Slimak 1999),and even then such artifacts are not universal.

Trajectories of change between 400 and 200 ka, leading tothe appearance of the full MP “package,” also differ regionally.In Europe as a whole, a certain level of technological diversityis maintained though the late Middle and early Upper Pleis-tocene. In fact, total diversity appears to increase over time.The number of variants of Levallois method may have ex-panded (White, Ashton, and Scott 2011), and some distinctivemodes of blank production, such as systematic manufactureof small bladelets (Faivre 2012), also develop in the laterMousterian. In marked contrast, technological diversity ac-tually declines over time in the Levant. Diversity is very highduring the second part of the Middle Pleistocene. The flake-tool-dominated assemblages making up Jelinek’s (1990) Mug-haran complex (Yabrudian, Amudian, and Acheulean facies)contain a wide range of methods for flake production. Theseassemblages are confined largely to cave sites and springs.

Contemporaneous open-air kill/butchery sites contain moretypical Acheulean assemblages with handaxes and limitedflake production (Porat et al. 2002; a similar pattern of “co-existence” may obtain in western Europe; Santoja and Villa2006). However, this technological diversity is truncatedaround 200–250 ka with the appearance of the early LevantineMousterian. There is some intersite diversity in lithic tech-nology in the early Mousterian that incorporates both Le-vallois and non-Levallois laminar production (Meignen2007a, 2007b), but the basic blank and artifact forms are fairlysimilar throughout the region (Meignen 1994; Monigal 2002).This tendency toward broad technological homogeneity ap-pears to persist throughout the Mousterian (250–50 ka), al-though the nature of the shared technology changes over time.The greatest level of “intra-assemblage” technical diversitymay be found in the “middle” Levantine Mousterian asso-ciated with archaic Homo sapiens fossils (Hovers 2009).

Diversity of Artifact Life Histories

Paleolithic artifacts are seldom recovered in mint condition.The great majority have entered the archaeological record aftersurviving one or more, sometimes many, cycles of use andrejuvenation. Part of the diversity of technological procedureswithin and among Mousterian assemblages can be groupedinto a series of alternative “life-history strategies” for flaketools, combining principles in the production of blanks withsubsequent patterns of use and maintenance. Multiple strat-egies may be represented in the same assemblage. At leastthree distinct life-history strategies have been identified.

1. A strategy based on exerting control over blank forms(“predetermination”) with comparatively limited subsequentshaping (e.g., Levallois, blade production). Artifacts may betransported, resharpened, and reduced, but the main strategyfor producing fresh edges is the manufacture of new blanks(Meignen, Delagnes, and Bourguignon 2009).

2. A strategy involving minimal control over blank formbut much subsequent shaping though retouch (Quina, Mi-coquian, etc.; Meignen, Delagnes, and Bourguignon 2009).The main means for producing fresh edges is through re-sharpening, extending the useful lives of tools. Artifacts maybe transported, but this is not always the case.

3. A third strategy encompassing both low levels of in-vestment in predetermining blank form and limited subse-quent shaping of blanks (discoid/denticulate). Fresh edges areobtained by extending the useful lives of cores. This mightbe referred to as an opportunistic or expedient strategy.

Researchers have identified other, more specific organiza-tional characteristics in some Mousterian assemblages. Inwhat are described as “ramified production systems,” a singleartifact may be treated alternatively as a tool or core over thecourse of its use-life (Bourgouignon, Faivre, and Turq 2004).A strategically ramified production system, exemplified by theQuina Mousterian, is thought to be distinguished from simplelateral recycling or scavenging of previously discarded material

Kuhn Roots of the Middle Paleolithic in Eurasia S261

because the initial design of the blank facilitates the diversefunctional roles that the object might fill.

This diversity in artifact life histories testifies to expansionin artifact functions and to the further integration of tool useinto all aspects of hominin life. Most flake tools of the MPwere not directly involved in the food quest; instead, theywere used to make artifacts out of other materials such aswood or hide (e.g., Anderson-Gerfaud 1990; Lemorini 2000;Plisson and Beyries 1998; Rotts 2009). The need to cope witha multitude of functional requirements and to maintain aready supply of artifacts for both predictable and unforeseenexigencies resulted in more complex cycles of artifact pro-duction and use than characterized previous periods. Varia-tion in artifact life histories is also tied to the mobility ofhominin individuals and groups, at least during the laterMousterian (after MIS 5; Delagnes and Rendu 2011; Kuhn1995). Certain artifact forms were routinely selected for trans-port based either on high edge/mass ratios or potential forresharpening (Eren and Lycett 2011). Others were producedmost often for local (in situ) use and discard.

The current state of knowledge is inadequate for evaluatingthe time depth of these more abstract strategic features ofMousterian technological behavior. The Acheulean and earlierindustries do show a certain level of redundancy or rigidityin artifact production and use. Flake tools tended to be ex-pediently or opportunistically used, whereas large core toolshad longer and more diversified life histories. Moreover, tra-jectories of manufacture and use were frequently linked toparticular raw materials (Goren-Inbar and Belfer-Cohen1998). Acheulo-Yabrudian assemblages in the Near East doshow more diverse trajectories of flake-tool production anduse, containing, for example, extensively curated and reducedscrapers alongside blade tools used with little or no modifi-cation (e.g., Shimelmitz, Barkai, and Gopher 2011). Data aretoo few to tie technological strategies to land use and foragingbehavior in the earlier periods. However, what we do knowsuggests that whereas the roots of many typical MP techno-logical practices are quite ancient, it is the associations andcontexts of these behaviors that varied. For example, in thelate MP of southern France, the Quina Mousterian is con-sistently associated with cold conditions and exploitation ofmigratory game—reindeer in particular (Delagnes and Rendu2011). However, that is certainly not the case for the Yabru-dian of the Levant. The same is true of discoid production,Levallois, or other procedures: they appear again and againbut in widely varying contexts.

Density/Retouch Relationship

In many sites of the Middle and Upper Paleolithic, the fre-quency of retouched tools is negatively correlated with thedensity of lithic artifacts in sediments (pieces per cubic meter;Barton et al. 2012; Kuhn 2004; Riel-Salvatore and Barton2004). The correlation is seldom very strong; many otherfactors, most notably varying rates of sediment accumulation,

also affect find densities. However, the relationship still ap-pears across a wide range of cases and areas. The most par-simonious explanation for this association involves simplestrategies that mobile populations use in keeping suppliedwith usable artifacts. It is clear that Mousterian homininsregularly carried artifacts with them as they moved from placeto place. If they remained in a particular place only for a veryshort period of time, they might rely exclusively on thosetransported artifacts, a few of which might then be depositedin the site. However, as the length of an occupation increased,both needs and opportunities would encourage hominins tocollect local raw materials and produce new tools on the spot.A series of short occupations would thus produce relativelylow-density deposits and relatively high frequencies of trans-ported and modified tools. Longer occupations would gen-erate both higher densities of artifacts and more unmodifieddebris from in situ manufacture (see Surovell et al. 2008 fora more extensive and rigorous discussion). Contra some au-thors (Barton et al. 2012; Riel-Salvatore and Barton 2004),this variation does not necessarily describe radically differentsettlement strategies. Instead, it reflects continuous variationin the duration of occupations.

It is not always possible to evaluate the relationship betweenartifact density and retouch frequency across multiple layers.Nonetheless, while the relationship is common or ubiquitousin Mousterian sites dating back to MIS 7, it has not yet beendemonstrated in earlier periods. At Tabun cave, for example,the expected relationship obtains in the Mousterian but notin the Acheulo-Yabrudian layers (Clark 2008; Kuhn and Clark,forthcoming). Bolomor Cave, a deeply stratified Middle Pleis-tocene site in Spain (Peris, Calatayud, and Valle 1997), alsodoes not show the expected pattern despite excellent pres-ervation and meticulous excavation. It is possible that ta-phonomy plays a role here, that accumulated geological dis-turbances over time disguise the patterning. However, thecrystallization of this tendency in the MP may also reflect ashift in land use and in the organization of lithic raw materialeconomies to service mobile populations.

Hearth-Focused Activities

Where geological conditions are amenable to the preservationof thermal features and spatial patterning in artifacts andother debris, Mousterian sites often show evidence for a rangeof activities centered around hearths (e.g., Henry et al. 2004;Mellars 1996:269–314; Vallverdu et al. 2012). This evidenceconsists of more or less well-defined fireplaces surrounded bya range of debris from toolmaking and food preparation andconsumption. Different zones of deposition/disposal may beevident, including toss zones for the largest/most intrusiveartifacts (Speth et al. 2012). Ashes may be removed fromhearths (rake out), indicating use over an extended period.Our ability to detect recurrent patterns in the use of space isdirectly limited by the number of well-preserved and well-excavated sites, so the record inevitably becomes more rarefied

S262 Current Anthropology Volume 54, Supplement 8, December 2013

with age. Stiner and colleagues argue that Middle Pleistoceneinhabitants of Qesem cave, first occupied as early as 400 ka,carried food to the cave and processed and consumed itaround fires in the same spaces they made tools (Stiner,Barkai, and Gopher 2011; Stiner, Gopher, and Barkai 2009).At the same time, butchery patterns may indicate that meatwas partitioned and shared out in a different manner than itwas among Mousterian or Upper Paleolithic groups.

Discussion: When and Why Did the MPCome Together?

The common characterization of the MP as a period of tech-nological stasis and monotony is misleading. In fact, there isa striking level of diversity, especially in methods for flakeproduction, within and among assemblages as well as withinand across regions. On the other hand, while there is certainlyevidence for refinement and diversification in manufacturetechnology, few truly new ways of doing things appeared dur-ing the period between 400 and 40 ka. The novel technologicaldevelopments of the MP are hafting and use of mastics toproduce composite tools, expansion of pyrotechnology, andthe use of pigments. Evidence of these phenomena is some-what discontinuous because of both preservation and typicalanalytical approaches.

In addition to diverse methods of flake production, MPtechnologies embody a varied range of artifact life histories,from long and complicated to short and simple. MP homininsused a variety of strategies to maintain a steady supply oftechnological aids not just for foraging but for working othermaterials into tools and clothing. In some cases, artifact lifehistories have shown to be associated with specific patternsof foraging or mobility. Recurrent correlations between re-touch frequencies and artifact densities in deeply stratifiedsequences and also indicate widely shared strategies for keep-ing supplied with artifacts and raw materials, strategies thatcontinued into the Upper Paleolithic. Along with the presenceof multiple reduction sequences and life histories within as-semblages, these facts testify to a high level of technologicalflexibility at both short (behavioral) and long (evolutionary)timescales. This flexibility in the manufacture of artifacts andmanagement of raw materials is in fact a defining feature ofthe MP. Although hard data are difficult to come by, it islikely that the MP shows greater procedural and tactical di-versity in lithic technology than both earlier and later culturalperiods.

At a local scale, the variety in MP lithic technologies andraw material economies attests to a significant degree of stra-tegic complexity. MP hominins were heavily dependent onstone artifacts as well as products made using stone tools.They were prepared to invest more time and energy than theirforebears in preparing multipart composite tools. What maydistinguish the MP from the Upper Paleolithic is the fact thatthese strategies are expressed entirely in lithic technology. Inthe Upper Paleolithic, more tactical and strategic variation is

expressed in choices among various raw materials (stone, os-seous materials, etc.) as well as in the design and assemblyof composite artifacts encompassing multiple materials. Inother words, while technological strategies of Upper Paleo-lithic hominins were at least as complex as those of MP hom-inins, lithic technology tells a smaller part of the story. Thewidespread shift to prismatic blade technology with the UpperPaleolithic, which appears to mark a reduction in technolog-ical diversity, may be in part a response to a need for regu-larization of components of composite artifacts (Bar-Yosefand Kuhn 1999).

There also is emerging evidence for strategic flexibility inland use during the MP, though it is more difficult to teaseout. Associations between lithic technology and choices ofmigratory or territorial game strongly suggest varying mo-bility regimes organized around the distribution of importantfood resources (e.g., Delagnes and Rendu 2011). But variationin mobility is as much about the assembly and disassemblyof groups of people as it is about mapping onto food. Someresearchers (Chazan 2009; Foley and Gamble 2009; Rolland2004) believe that domestic spaces and true “campsites” firstbecome recognizable in Eurasia during the later Middle Pleis-tocene (but see Villa 2009 for a different view). Evidence citedincludes increased use of caves and rockshelters, more fre-quent presence of burning and combustion features in sheltersites, regular transport of hunted meat to shelters for pro-cessing and consumption, and evidence that other activities(e.g., toolmaking and use) were conducted in the same spaces.However, these various features almost certainly coalescedgradually, and early MP “campsites” may not be identical tomore recent and more familiar ones. The recurrent density/retouch relationship described for many MP and Upper Pa-leolithic sites may be another symptom of this development.The contrast between low- and high-density assemblages thatfirst becomes apparent sometime after 250 ka shows variableand complex use of landscape with changing mixes of long-and short-term occupations involving different subsets of so-cial groups conducting different activities.

Diversity expressed over larger spatial and temporal scalesrequires a different sort of explanation. The geographic varietyin European MP technologies is undoubtedly tied closely tovariation in topography and climatic conditions across theEuropean landmass. By the end of MIS 5, MP hominin pop-ulations were distributed from north-central Europe to thesouthern Levant and from the Atlantic to the Altai Mountains.The radically different habitats that hominins encountered indifferent places as well as the severe fluctuations in climate(summarized in Andel and Davies 2003; Rohling et al. 2013)would inevitably have led to divergent adaptations amonglocal populations. A part of the spatial variability in methodsof flake production also probably reflects neutral variation.Small populations dispersed across Europe and western Asia,isolated from one another by distance and environmentalbarriers, would have provided an ideal setting for drift-likeprocesses and random loss and retention of certain behaviors.

Kuhn Roots of the Middle Paleolithic in Eurasia S263

The periodic contraction of local populations into refugiaduring harsh climatic intervals and their subsequent reex-pansion during times when conditions were more favorable(a source/sink model; Dennel 2009) could help maintain oreven amplify regional variation. Contrasts among regions inpatterns of chronological change and diversity might reflectvariation in population sizes and connectivity as well as eco-logical factors (reviewed in Kuhn 2012).

The suite of traits identified here as typical of the MP beganto appear in the second half of the Middle Pleistocene butcontinued to develop and coalesce throughout the UpperPleistocene (see also d’Errico 2007 for a discussion of otherelements of behavior). Many methods of flake productiontypical of the Mousterian first appear around 400–500 ka ifnot long before that. Likewise, regular use of fire and perhapssomething recognizable as “domestic spaces” may appear be-tween 300 and 400 ka. Pigments and hafting only becomeevident after 250 ka. There may be an expansion and diver-sification of flake production around this same period orslightly later. It should not be surprising that the gradualcoalescence of the characters that define the MP in westernEurasia resembles what has been documented for the AfricanMSA (d’Errico 2007; McBrearty and Brooks 2000). Both pat-terns are best understood as the gradual evolution of a broadsuite of cultural adaptations in response to shifting regionalenvironmental and demographic conditions. As we might alsoexpect, the specific adaptive patterns and sets of behaviorsdiffer across the two regions. Nonetheless, the trajectories ofdevelopment are similar in structure. Resemblances extend tothe discontinuous histories of many elements. In both theMP and MSA, specific forms of behavior—whether methodsfor working stone or approaches to ornamenting the body—appear, disappear, and sometimes reappear over time (Hoversand Belfer-Cohen 2006).

Specific behaviors and archaeological evidence aside, oneobvious difference between the MSA record of southern Africaand the MP record of Eurasia concerns the spatial scale atwhich directional change can be detected. It currently appearsthat a single sequence of industrial succession and techno-logical change characterizes South Africa, Namibia, andneighboring countries between MIS 5 and 3 (Jacobs and Rob-erts 2008; Jacobs et al. 2008; Wurz 2013). This contrastssharply with Europe, which shows a diversified patchwork ofindependent regional sequences. In one area the Mousterianmay exhibit a definite directional trend, while in a neighboringarea there may be a very different trend, or none at all. TheLevant shows a general pattern similar to southern Africa,with a single, fairly consistent set of cultural developmentsbetween 225 and 40 ka (Hovers and Belfer-Cohen 2013).However, the Levant actually covers a much smaller area thansouthern Africa, similar in size to one of the subregions withinEurope.

A number of researches have observed that the progressionof late MSA assemblages in southern Africa—in particularthe sequence of MSA 2, Stillbay, and Howieson’s Poort—

resembles cultural developments in the Eurasian Upper Pa-leolithic. Moreover, is fits expectations for cumulative culturalevolution or “cultural ratcheting” (Tennie, Call, and Toma-sello 2009; Tomasello 1999). However, the apparent absenceof such patterns within the MP does not necessarily indicatethat MP hominins were somehow unable to sustain culturaldevelopment over long periods. We can question the evidencefor cumulative cultural evolution in the Upper Paleolithic andMSA. The Stillbay and Howieson’s Poort, like the Aurignacianand Gravettian, appear more like equivalent (though differ-ent) cultural developments than examples of stepwise increasein the complexity or variety of human achievement. What isreally remarkable about these particular developments is thegeographic scale over which specific cultural variants arefound. As such, they may speak more to qualities of con-nectivity among subpopulations and the potential for culturetraits to spread than to “cultural ratcheting.”

The appearance of pigments in the MP marks a new rolefor material culture, that of signaling. The use of ochre byMP hominins certainly shows evidence for body decorationas a component in social interaction. We might also expectsome of the variation across space in MP artifact forms tobe evidence for conscious identity formation, what is some-times termed “emblemic style” (see Hegmon 1992; Wiessner1983). As discussed above, Mousterian assemblages do showconsiderable variation in modes of production of stone tools.However, lithic technology is an inherently stationary practicethat requires close proximity to observe. For this reason, var-iation in the fine details of technological procedures is likelyto represent simply repetition of different learned patternsrather than an active signaling of identity (Tostevin 2007).On the other hand, there is a widespread impression that theMP record contains little expressive “stylistic” variation, par-ticularly in artifacts such as points. This impression restslargely on the observation that in later periods, spear, dart,and arrow points tend to show stronger geographic and tem-poral patterning than other stone-tool forms, which is furtherpresumed to represent intentional assertion of group identitystyle. To a certain extent, appreciation of fine-scale stylisticvariation in artifact forms may have been inhibited by thedecades-long hegemony of the Bordes typological system andlater by the recognition that a good deal of variation in toolforms reflects use-lives rather than design (e.g., Dibble 1987,1995; McPherron 2006). Even so, regional differences in pointforms or in the shapes of other retouched tools, for thatmatter, are not as easily identified as they are, for example,in the sub-Saharan African MSA (McBrearty and Brooks2000). The bifacial points found in some central, southern,and eastern European assemblages are a possible exception,but even these are too widespread to nominate as a regionalor local style. In combination with the absence of consistentsystems of body ornamentation in durable materials, this ob-servation leads to the conclusion that intergroup signaling ofidentity was not an important function of MP material culturein general (Kuhn and Stiner 2007).

S264 Current Anthropology Volume 54, Supplement 8, December 2013

In the end, two factors than can best account for many ofthe emergent features of the MP are population size andpopulation structure. There is an emerging view that MPhominins existed at very low population densities comparedwith later Paleolithic or recent hunter-gatherers and that theirpopulations were highly fragmented. In addition to geneticevidence for small effective population sizes (Bocquet-Appeland Degioanni 2013; Lalueza-Fox 2013), there are ecologicalreasons to expect thinly distributed populations. Althoughthey consumed a wide variety of foods (e.g., Henry, Brooks,and Piperno 2010), the isotopic and faunal evidence suggestthat MP hominins tended to feed at a high trophic level,focusing their efforts on large-game animals (Bocherens andDrucker 2003; Gaudzinski-Windheuser and Niven 2009;Richards and Trinkaus 2009; Stiner 2013; Stiner and Kuhn2009; Stiner, Munro, and Surovell 2000; Stiner et al. 1999;Villa and Lenoir 2009). Because the collective biomass of largeherbivores is lower than small game or plants in terrestrialenvironments, the combination of large-game-focused for-aging and the comparatively high individual energy require-ments of Neanderthal anatomy (Froelhe and Churchill 2009;Snodgrass and Leonard 2009; Sorenson and Leonard 2001)leads to the inference that populations were sparse and co-resident groups very small, even in comparison with recentforagers. In aggregate, these diverse lines of evidence point topartitioning of MP populations into many small, demograph-ically fragile local demes. Because it would be difficult forsuch groups to maintain contact over large areas, social net-works would have been limited in scale and highly com-partmentalized (Gamble 1999).

Small absolute population sizes along with frequent localextinction of local demes could account for a range of typicalfeatures of the Eurasian MP. As has been widely discussed,small populations both limit the rates of invention and in-crease the likelihood that beneficial but rare novelties will belost (Bentley, Hahn, and Shennan 2004; Henrich 2004; Hoversand Belfer-Cohen 2006; Powell, Shennan, and Thomas 2009).Frequent extinction of local demes could also constrain ratesof drift and lead to recycling of old technological themes(Premo and Kuhn 2010). The limited use of technology forsignaling in the MP is also likely a consequence of smallgroups that seldom came into contact with and hence seldomneeded to identify themselves to strangers (Kuhn and Stiner2007). Likewise, the existence of many independent regionaltrajectories in technological change is most parsimoniouslyunderstood as a consequence of poorly interconnected socialnetworks that tended to keep cultural information local ratherthan potentiating its rapid diffusion (Kuhn 2012).

This proposition that MP populations were thinly distrib-uted and partitioned into small, disconnected local demesmight account for many features of the MP record, but thishypothesis in turn begs important questions. The most fun-damental is why hominins might have maintained such hightrophic levels and low demographic potential for so long. Canthe demographic patterns be attributed to environmental or

climatic constraints in contrast with the African MSA, forexample (Powell, Shennan, and Thomas 2009)? If so, why didMousterian hominins seldom transcend these limitations, bydiversifying the subsistence base, for example? Were cognitivefactors responsible for this apparent lack of flexibility (e.g.,Wynn and Coolidge 2004), and if so, what features of cog-nition are implicated and how can this be reconciled with theobserved technological variability and ability to colonize di-verse habitats? An alternative view is that because of theirparticular behavioral and biological evolutionary histories,MP populations were “locked into” a particular set of re-sponses such that immediate changes in behavior would havehad the immediate consequence of lower fitness (Andersson,Tornberg, and Tornberg 2014). In other words, the particularfitness landscapes on which MP populations operated maysimply have made certain evolutionary changes more difficultthan others (Brantingham, Kuhn, and Kerry 2004) indepen-dent of the essential characteristics of the hominins them-selves. Channeled into certain trajectories by its evolutionaryhistory, the particular experiment that was the MP played outin a different way than contemporary experiments in otherparts of the world.

References CitedAlperson-Afil, Nira. 2008. Continual fire-making by hominins at Gesher Benot

Ya’aqov, Israel. Quaternary Science Reviews 27:1733–1739.Al Qadi, Amjad. 2011. Le Yabroudien en Syrie: etat de la question et enjeux

de la recherche. In The Lower and Middle Palaeolithic in the Middle Eastand neighbouring regions. Marcel Otte, Reto Jagher, and Jean-Marie Le Ten-sorer, eds. Pp. 77–84. Etudes et Recherches Archeologiques de l’Universitede Liege, no. 126. Liege: Etudes et recherches archeologiques de l’Universitede Liege.

Ambrose, Stanley. 2010. Coevolution of composite-tool technology, construc-tive memory, and language: implications for the evolution of modern hu-man behavior. Current Anthropology 51(suppl.):S135–S147.

Andel, Thies van, and William Davies, eds. 2003. Neanderthals and modernhumans in the European landscape during the last glaciation. Cambridge:McDonald Institute for Archaeological Research.

Anderson-Gerfaud, Patricia. 1990. Aspects of behaviour in the Middle Pa-laeolithic: functional analysis of stone tools from southwest France. In Theemergence of modern humans. Paul Mellars, ed. Pp. 389–418. Edinburgh:Edinbugh University Press.

Andersson, Claes, Anton Tornberg, and Petter Tornberg. 2014. An evolu-tionary developmental approach to cultural evolution. Current Anthropol-ogy, forthcoming.

Ashton, Nick, and John McNabb. 1992. The interpretation and context of theHigh Lodge industries. In High Lodge: excavations by G. de G. Sieveking1962–68 and J. Cook 1988. Nick Ashton, Jill Cook, S. G. Lewis, and J. Rose,eds. Pp. 164–168. London: British Museum Press.

Barroso Ruız, Cecelio, Daniel Botella Ortega, Miguel Caparros, Anne MarieMoigne, Vincenzo Celiberti, Agnes Testu, Deborah Barsky, et al. 2011. TheCueva del Angel (Lucena, Spain): an Acheulean hunters habitat in the southof the Iberian Peninsula. Quaternary International 243:105–126.

Barsky, Deborah. 2009. An overview of some African and Eurasian Oldowansites: evaluation of hominin cognition levels, technological advancementand adaptive skills. In Interdisciplinary approaches to the Oldowan. ErellaHovers and David Braun, eds. Pp. 39–47. Vertebrate Paleobiology andPaleoanthropology Series. New York: Springer.

Barton, C. Michael, Julien Riel-Salvatore, John M. Anderies, and GabrielPopescu. 2011. Modeling human ecodynamics and biocultural interactionsin the Late Pleistocene of western Eurasia. Human Ecology 39(6):705–725.

Bar-Yosef, Ofer, and Steven Kuhn. 1999. The big deal about blades: laminartechnologies and human evolution. American Anthropologist 101:322–338.

Kuhn Roots of the Middle Paleolithic in Eurasia S265

Bar-Yosef, Ofer, and Philip Van Peer. 2009. The chaıne operatoire approach inMiddle Paleolithic archaeology. Current Anthropology 50(1):103–131.

Bar-Yosef Mayer, Daniella, Bernard Vandermeersch, and Ofer Bar-Yosef. 2009.Shells and ochre in Middle Paleolithic Qafzeh Cave, Israel: indications formodern behavior. Journal of Human Evolution 56(3):307–314.

Bentley, R. Alexander, Matthew W. Hahn, and Stephen Shennan. 2004. Ran-dom drift and culture change. Proceedings of the Royal Society of London B271:1443–1450.

Berna, Francesco, and Paul Goldberg. 2007. Assessing Palaeolithic fire tech-nology and associated hominin behavior in Israel. Israel Journal of EarthSciences 56:107–121.

Bocherens, Herve, and D. E. Drucker. 2003. Reconstructing Neandertal dietfrom 120,000 to 30,000 BP using carbon and nitrogen stable isotopic abun-dances. In Le role de l’environnement dans le comportements des chasseurs-cueilleurs prehistoriques. Marilyn Patou-Mathis and Herve Bocherens, eds.Pp. 1–8. BAR International Series 1105. Oxford: British ArchaeologicalReports.

Bocquet-Appel, Jean-Pierre, and Anna Degioanni. 2013. Neanderthal demo-graphic estimates. Current Anthropology 54(suppl. 8):S202–S213.

Boeda, E. 1986. La debitage Levallois de Biache St-Vaast (Pas-de-Calais): pre-miere etude technologique. In Chronostratigraphie et facies culturels du pa-leolithique inferieur et moyen dans l’europe du nord-ouest. Alain Tuffreau andA. Somme, eds. Pp. 209–218. Bulletin de l’Association Francaise pourl’Etude du Quaternaire, no. 26, supplement. Paris: L’Association Francaisepour l’Etude du Quaternaire.

———. 1991. Approche de la variabilite des systemes de production lithiquedes industries du paleolithique inferieur et moyen: chronique d’une var-iabilite attendee. Techniques et Culture 17/18:37–79.

———. 1994. Le concept Levallois: variabilite des methodes. Paris: CNRS.———. 1995. Levallois: a volumetric construction, methods, a technique. In

The definition and interpretation of Levallois technology. Harold Dibble andOfer Bar-Yosef, eds. Pp. 41–69. Madison, WI: Prehistory Press.

Boeda, Eric, S. Bonilauri, J. Connan, D. Jarvie, Norbert Mercier, M. Tobey,Helene Valladas, H. al Sakhel, and Sultan Muhesen. 2008. Middle Palaeo-lithic bitumen use at Umm el Tlel around 70,000 BP. Antiquity 82:853–886.

Boeda, Eric, Jacques Connan, Daniel Dessort, Sultan Muhesen, Norbert Mer-cier, Helene Valladas, and Nadine Tisnerat. 1996. Bitumen as a haftingmaterial on Middle Palaeolithic artefacts. Nature 380(6572):336–338.

Boeda, Eric, Jean-Michel Geneste, C. Griggo, Norbert Mercier, Sultan Mu-hesen, J.-L. Reyss, A. Taha, and Helene Valladas. 1999. A Levallois pointembedded in the vertebra of a wild ass (Equus africanus): hafting, projectilesand Mousterian hunting weapons. Antiquity 73(280):394–402.

Boeda, Eric, Jean-Michel Geneste, and Liliane Meignen. 1990. Identificationde chaınes operatoires lithiques du paleolithique ancient et moyen. Paleo 2:43–80.

Bosinski, Gerhard. 1967. Die mittelpalaolithischen Funde im westlichen Mit-teleuropa. Cologne: Bohlau.

Bourguignon, Laurence. 1996. La conception de debitage Quina. QuaternariaNova 6:149–166.

———. 1997. Le Mousterien de type Quina: nouvelle definition d’une entitetechnique. Doctoral thesis, Universite Paris X Nanterre.

Bourguignon, Laurence, Jean-Philippe Faivre, and Alain Turq. 2004. Rami-fication des chaınes operatoires: une specificite du Mousterien? Paleo 16:37–48.

Brantingham, P. Jeffery. 2006. Measuring forager mobility. Current Anthro-pology 47(3):435–459.

Brantingham, P. Jeffery, Steven Kuhn, and Kristopher W. Kerry. 2004. On thedifficulty of the Middle-Upper Paleolithic transition. In The Early UpperPaleolithic beyond western Europe. P. Jeffery Brantingham, Steven Kuhn, andKristopher W. Kerry, eds. Pp. 1–13. Berkeley: University of California Press.

Burdukiewicz, Jan Michal, and Avraham Ronen, eds. 2003. Lower Palaeolithicsmall tools in Europe and the Levant. BAR International Series 1115. Oxford:British Archaeological Reports.

Carbonell, Eudald, and Xose Pedro Rodrıquez. 2006. The first human settle-ment of Mediterranean Europe. Comptes Rendus Palevol 5(1/2):291–298.

Carciumaru, Marin, Radica-Mariana Ion, Elena Cristina Nitu, and Radu Ste-fanescu. 2012. New evidence of adhesive as hafting material on Middle andUpper Palaeolithic artefacts from Gura Cheii-Rasnov Cave (Romania). Jour-nal of Archaeological Science 39(7):1942–1950.

Casanova I Martı, Joel, Jorge Martınez Moreno, Rafael Mora Turcal, andIgnacio de la Torre. 2009. Strategies techniques dans le paleolithique moyendu sud-est des Pyrenees. L’Anthropologie 113:313–340.

Chazan, Michael. 2009. Assessing the Lower to Middle Paleolithic transition.In Sourcebook of Paleolithic transitions. Marta Camps and Parth Chuahan,eds. Pp. 237–243. New York: Springer.

Churchill, Steven, E. 1993. Weapon technology, prey size selection, and hunt-ing methods in modern hunter-gatherers: implications for hunting in thePalaeolithic and Mesolithic. In Hunting and animal exploitation in the laterPaleolithic and Mesolithic of Eurasia. Gail Peterkin, Harvey Bricker, and PaulMellars, eds. Pp. 11–24. Archaeological Papers of the American Anthro-pological Association no. 4. Washington, DC: American AnthropologicalAssociation.

Clark, Amy E. 2008. Changes in occupation intensity during the Lower andMiddle Paleolithic at Tabun Cave, Israel. MA thesis, University of Arizona,Tucson.

Clark, Geoffery. 2009. Accidents of history: conceptual frameworks in pa-leoarchaeology. In Sourcebook of Paleolithic transitions. Marta Camps andParth Chuahan, eds. Pp. 19–43. New York: Springer.

Clark, Grahame. 1969. World prehistory: a new outline. 2nd edition. Cambridge:Cambridge University Press.

Conard, Nicholas J., and B. Fischer. 2000. Are there recognizable culturalentities in the German Middle Palaeolithic? In Toward modern humans:Yabrudian and Micoquian, 400–50k years ago. Avraham Ronen and Wein-stein-Evron, eds. Pp. 7–24. BAR International Series 850. Oxford: BritishArchaeological Reports.

Copeland, Lorraine. 1998. The Lower Paleolithic of Jordan. In The Prehistoricarcheology of Jordan. Donald Henry, ed. Pp. 5–22. BAR International Series705. Oxford: British Archaeological Reports.

DeBono, Hila, and Na’ama Goren-Inbar. 2001. Note on a link between Acheu-lian handaxes and the Levallois method. Journal of the Israel PrehistoricSociety 31:9–23.

Delagnes, Anne. 1993. Un mode de production inedit au paleolithique moyendans l’industrie du niveau 6e du Pucheuil (Seine-Maritime). Paleo 5:111–120.

Delagnes, Anne, Jacques Jaubert, and Liliane Meignen. 2007. Les technocom-plexes du paleolithique moyen en Europe occidental dans leur cadre dia-chronique et geographique. In Neandertaliens: biologie et cultures. BernardVandermeersch, Bruno Maureille, and Yves Coppens, eds. Pp. 213–229.Documents Prehistoriques 23. Paris: CTHS.

Delagnes, Anne, and Liliane Meignen. 2005. Diversity of lithic productionsystems in the Middle Paleolithic in France: are there any chronologicaltrends? In Transitions before the transition: evolution and stability in theMiddle Paleolithic and Middle Stone Age. Erella Hovers and Steven Kuhn,eds. Pp. 85–107. New York: Springer.

Delagnes, Anne, and William Rendu. 2011. Shifts in Neandertal mobility,technology and subsistence strategies in western France. Journal of Archae-ological Science 38(8):1771–1783.

de la Torre, Ignacio, Jorge Martınez-Moreno, and Rafael Mora. 2013. Changeand stasis in the Iberian Middle Paleolithic: considerations on the signifi-cance of Mousterian technological variability. Current Anthropology54(suppl. 8):S320–S336.

Dennell, Robin. 2009. The Paleolithic settlement of Asia. Cambridge: CambridgeUniversity Press.

d’Errico, Francesco. 2007. The origin of humanity and modern cultures: ar-chaeology’s view. Diogenes 54(2):122–133.

———. 2008. The rouge et le noir: implications of early pigment use inAfrica, the Near East and Europe for the origin of cultural modernity. InCurrent themes in Middle Stone Age research. Marlize Lombard, ChristineSievers, and V. Ward, eds. Pp. 168–174. South African Archaeological SocietyGoodwin Series, vol. 10. Vlaeberg: South African Archaeological Society.

Dibble, Harold. 1987. The interpretation of Middle Paleolithic scraper mor-phology. American Antiquity 52:109–117.

———. 1991. Mousterian assemblage variability on an interregional scale.Journal of Anthropological Research 47(2):239–257.

———. 1995. Middle Paleolithic scraper reduction: background, clarification,and review of the evidence to date. Journal of Archaeological Method andTheory 2:299–368.

Dibble, Harold, and Shannon McPherron. 2006. The missing Mousterian.Current Anthropology 47(5):777–803.

Eren, Metin I., and Stephen J. Lycett. 2011. Why Levallois? a morphometriccomparison of experimental “preferential” Levallois flakes versus debitageflakes. PLoS ONE 7(1):e29273, doi:10.1371/journal.pone.002927.

Faivre, Jean-Philippe. 2012. A material anecdote but technical reality: bladeletand small blade production during the recent Middle Paleolithic at theCombe-Grenal rockshelter. Lithic Technology 37:5–24.

S266 Current Anthropology Volume 54, Supplement 8, December 2013

Feblot-Augustins, Jehanne. 1997. La circulation des matieres premieres au pa-leolithique. Etudes et Recherches Archeologiques de l’Universite de Liege,no. 75. Liege: Service de prehistoire, Universite de Liege.

———. 2009. Revisiting European Upper Palaeolithic raw material transfers:the demise of the cultural ecological paradigm? In Lithic materials andPaleolithic societies. Brian Adams and Brooke S. Blades, eds. Pp. 25–46. NewYork: Wiley-Blackwell.

Flas, Damien. 2008. La transition du paleolithique moyen au superieur dans laplaine septentrionale de l’Europe. Anthropologica et Praehistorica 119. Brus-sels: Royal Belgian Society of Anthropology and Prehistory.

Fluck, H. 2011. Culture in the Lower Palaeolithic: technological variability inMiddle Pleistocene Europe. In Investigating archaeological cultures: materialculture, variability and transmission. Benjamin W. Roberts and Marc VanderLindern, eds. Pp. 77–96. New York: Springer.

Foley, Robert, and Clive Gamble. 2009. The ecology of social transitions inhuman evolution. Philosophical Transactions of the Royal Society B 364:3267–3279.

Froelhe, Andrew, and Steven Churchill. 2009. Energetic competition betweenNeandertals and anatomically modern humans. Paleoanthropology 2009:96–116.

Gamble, Clive. 1999. Palaeolithic societies of Europe. Cambridge: CambridgeUniversity Press.

Gaudzinski-Windheuser, Sabine, and Laura Niven. 2009. Hominid subsistencepatterns during the Middle and Late Paleolithic in northwestern Europe.In The evolution of hominin diets: integrating approaches to the study ofPaleolithic subsistence. Jean-Jacques Hublin and Michael Richards, eds. Pp.99–111. Dordrecht: Springer.

Geneste, Jean-Michel. 1988. Systemes d’approvisionnement en matieres pre-mieres au paleolithique moyen et au paleolithique superieur en Aquitaine.In La mutation, vol. 8 of L’homme de Neandertal. Marcel Otte and JanuszK. Kozlowski, eds. Pp. 61–70. Etudes et Recherches Archeologiques del’Universite de Liege, no. 35. Liege: Universite de Liege.

———. 1990. Development des systemes de production lithique au cours dupaleolithique moyen en Aquitaine septentrionale. In Paleolithique moyenrecent et paleolithique superieur ancien en Europe: ruptures et transitions.Catherine Farizy, ed. Pp. 203–123. Memoires du Musee de Prehistoire d’Ile-de-France, no. 3. Nemours: Association pour la promotion de la recherchearcheologique en Ile-de-France.

Geneste, Jean-Michel, and Hugues Plisson. 1996. Production et utilisation del’outillage lithique dans le Mousterien du sud-ouest de la France: les Taresa Sourzac, Vallee de l’Isle, Dordogne. Quaternaria Nova 6:343–368.

Goldberg, Paul, Harold Dibble, Francisco Berna, Dennis Sandgathe, ShannonMcPherron, and Alain Turq. 2012. New evidence on Neandertal use of fire:examples from Roc de Marsal and Pech de l’Aze IV. Quaternary International247:325–340.

Gopher, Avi, Avner Ayalon, Miryam Bar-Matthews, Ran Barkai, Amos Frum-kin, Panagiotis Karkanas, and Ruthi Shahack-Gross. 2010. The chronologyof the Late Lower Paleolithic in the Levant based on U-Th ages of speleo-thems from Qesem cave, Israel. Quaternary Geochronology 5(6):644–656.

Goren-Inbar, Na’ama, and Anna Belfer-Cohen. 1998. The technological abil-ities of the Levantine Mousterian: cultural and mental capacities. In Nean-dertals and modern humans in western Asia. Takeru Akazawa, Kenichi Akoi,and Ofer Bar-Yosef, eds. Pp. 205–221. New York: Plenum.

———. 2011. Culture and cognition in the Acheulian industry: a case studyfrom Gesher Benot Ya’aqov. Philosophical Transactions of the Royal SocietyB 366:1038–1049.

Grunberg, Judith. 2002. Middle Palaeolithic birch-bark pitch. Antiquity 76:15–16.

Hegmon, Michelle. 1992. Archaeological research on style. Annual Review ofAnthropology 21:517–536.

Henrich, Joe. 2004. Demography and cultural evolution: how adaptive culturalprocesses can produce maladaptive losses: the Tasmanian case. AmericanAntiquity 69:197–214.

Henry, Amanda, Alison Brooks, and Delores Piperno. 2010. Microfossils incalculus demonstrate consumption of plants and cooked foods in Nean-derthal diets (Shanidar III, Iraq; Spy I and II, Belgium). Proceedings of theNational Academy of Sciences of the USA 108(2):486–491.

Henry, Donald, Harold J. Hietala, Arlene Rosen, Yuri Demidenko, Vitali Usik,and T. L. Armagan. 2004. Human behavioral organization in the MiddlePaleolithic: were Neanderthals different? American Anthropologist 106:17–31.

Hiscock, Peter, Alain Turq, Jean-Philippe Faivre, and Laurence Bourguignon.2009. Quina procurement and tool production. In Lithic materials and

Paleolithic societies. Brian Adams and Brooke Blades, eds. Pp. 232–246. NewYork: Wiley-Blackwell.

Hovers, Erella. 2007. The many faces of cores-on-flakes: a perspective fromthe Levantine Mousterian. In Cores or tools? alternative approaches to stonetool analysis. Shannon McPherron, ed. Pp. 42–74. Cambridge: CambridgeScholars Press.

———. 2009. The lithic assemblages of Qafzeh cave. Oxford: Oxford UniversityPress.

Hovers, Erella, and Anna Belfer-Cohen. 2006. “Now you see it, now youdon’t”: modern human behavior in the Middle Paleolithic. In Transitionsbefore the transition: evolution and stability in the Middle Paleolithic andMiddle Stone Age. Erella Hovers and Steven Kuhn, eds. Pp. 295–304. NewYork: Springer.

———. 2013. On variability and complexity: lessons from the LevantineMiddle Paleolithic record. Current Anthropology 54(suppl. 8):S337–S357.

Jacobs, Zenobia, and Richard G. Roberts. 2008. Testing times: old and newchronologies for the Howiesons Poort and Still Bay Industries in environ-mental context. South African Archaeological Society Goodwin Series 10:9–34.

Jacobs, Zenobia, Richard G. Roberts, Rex F. Galbraith, Hillary Deacon, RainerGrun, Alex Mackay, Peter Mitchell, Ralf Vogelsang, and Lynn Wadley. 2008.Ages for the Middle Stone Age of Southern Africa: implications for humanbehavior and dispersal. Science 322:733–735.

Jelinek, Arthur J. 1982. The Middle Paleolithic in the Southern Levant, withcomments on the appearance of modern Homo sapiens. In The transitionfrom Lower to Middle Paleolithic and the origin of modern man. AvrahamRonen, ed. Pp. 57–101. BAR International Series 151. Oxford: British Ar-chaeological Reports.

———. 1990. The Amudian in the context of the Mugharan tradition at theTabun Cave (Mount Carmel), Israel. In The emergence of modern humans.Paul Mellars, ed. Pp. 81–90. Ithaca, NY: Cornell University Press.

Karkanas, Panagiotis. 2002. Ash, bones and guano: a study of the mineralsand phytoliths in the sediments of Grotte XVI, Dordogne, France. Journalof Archaeological Science 29(7):721–732.

Karkanas, Panagiotis, Ruthi Shahack-Gross, Avner Ayalon, Miryam Bar-Mat-thews, Ran Barkai, Aamos Frumkin, Avi Gopher, and Mary C. Stiner. 2007.Evidence for habitual use of fire at the end of the Lower Paleolithic: site-formation processes at Qesem Cave, Israel. Journal of Human Evolution53(2):197–212.

Kuhn, Steven. 1995. Mousterian lithic technology: an ecological approach.Princeton, NJ: Princeton University Press.

———. 2004. Middle Paleolithic assemblage formation at Riparo Mochi. InProcessual archaeology. A. Johnson, ed. Pp. 31–60. London: Praeger.

———. 2012. Emergent patterns of creativity and innovation in early tech-nologies. In Creativity, innovation and human evolution. S. Elias, ed. Pp.69–87. Amsterdam: Elsevier.

Kuhn, Steven, Guven Arsebuk, and F. Clark Howell. 1996. The Middle Pleis-tocene lithic assemblage from Yarimburgaz Cave, Turkey. Paleorient 22(1):31–49.

Kuhn, Steven, and Amy E. Clark. Forthcoming. Artifact densities and assem-blage formation in the Paleolithic: evidence from Tabun Cave. Journal ofAnthropological Archaeology.

Kuhn, Steven, and Mary C. Stiner. 2001. The antiquity of hunter-gatherers.In Hunter-gatherers, an interdisciplinary perspective. Catherine Panter-Brick,Robert H. Layton, and Peter Rowley-Conwy, eds. Pp 99–142. Cambridge:Cambridge University Press.

———. 2007. Body ornamentation as information technology: towards anunderstanding of the significance of early beads. In Rethinking the humanrevolution: new behavioural and biological perspectives on the origins anddispersal of modern humans. Paul Mellars, Katie Boyle, Ofer Bar-Yosef, andC. Stringer, eds. Pp. 45–54. Cambridge: McDonald Institute of Archaeo-logical Research.

Lalueza-Fox, Carles. 2013. Agreements and misunderstandings among threescientific fields: paleogenomics, archaeology, and human paleontology. Cur-rent Anthropology 54(suppl. 8):S214–S220.

Lemorini, Cristina. 2000. Reconnaitre des tactiques d’exploitation du milieu aupaleolithique moyen: etude fonctionnelle des industries lithiques de GrottaBreuil (Latium) et de La Combette (Bonnieux, Vaucluse, France). BAR In-ternational Series 858. Oxford: British Archaeological Reports.

Le Tensorer, Jean-Marie, Reto Jagher, Phillipe Rentzel, Thomas Hauck, KristinIsmail-Meyer, Christine Pumpin, and Dorota Wojtczak. 2007. Long-termformation processes at the natural springs Nadaouiyeh and Hummal in theEl-Kowm oasis, central Syria. Geoarchaeology 22(6):621–639.

Kuhn Roots of the Middle Paleolithic in Eurasia S267

Lycett, Steven. 2007. Why is there a lack of Mode 3 Levallois technologies inEast Asia? a phylogenetic test of the Movius-Schick hypothesis. Journal ofAnthropological Archaeology 26(4):541–575.

Mazza, Paul Peter A., Fabio Martini, Benedetto Sala, Maurizio Magi, MariaPerla Colombini, Gianna Giachi, Francesco Landucci, Cristina Lemorini,Francesca Modugno, and Erica Ribechini. 2006. A new Palaeolithic dis-covery: tar-hafted stone tools in a European Mid-Pleistocene bone-bearingbed. Journal of Archaeological Science 33:1310–1318.

McBrearty, Sally, and Alison Brooks. 2000. The revolution that wasn’t: a newinterpretation of the origin of modern humans. Journal of Human Evolution39:453–463.

McPherron, Shannon. 2006. What typology can tell us about Acheulian hand-axe production. In Axe age: Acheulian tool-making from quarry to discard.Na’ama Goren-Inbar and Gonen Sharon, eds. Pp. 267–285. London: Equi-nox.

Meignen, Liliane. 1994. Paleolithique moyen au proche-orient: le phenomenelaminaire. In Les industries laminaires au paleolithique moyen. StephaneRevillion and Alain Tuffreau, eds. Pp. 125–161. Paris: CNRS.

———. 2007a. Middle Paleolithic blady assemblages in the Near East: areassessment. In Caucasus and the initial dispersals in the Old World. Pp.133–148. Russian Academy of Sciences, Institute of the History of MaterialCulture, vol. 21. St. Petersburg: PV.

———. 2007b. Le phenomene laminaire au proche-orient, du paleolithiqueinferieur aux debuts du paleolithique superieur. In Un siecle de constructiondu discours scientifique en prehistoire: congres du centenaire de la Societeprehistorique francaise. J. Evin, ed. Pp. 79–94. Paris: Societe prehistoriquefrancaise.

Meignen, Liliane, Anne Delagnes, and Laurence Bourguignon. 2009. Patternsof lithic raw material procurement and transformation during the MiddlePaleolithic in western Europe. In Lithic materials and Paleolithic societies.Brian Adams and Brooke Blades, eds. Pp. 15–24. New York: Wiley-Blackwell.

Mellars, Paul. 1996. The Neanderthal legacy. Princeton, NJ: Princeton Uni-versity Press.

Mercier, Norbert, and Helene Valladas. 2003. Reassessment of TL age estimatesof burnt flints from the Paleolithic site of Tabun Cave, Israel. Journal ofHuman Evolution 45:401–409.

Mihailovic, Dusan, and Bojiana Mihailovic. 2009. Cultural regionalization inthe Palaeolithic of the Middle Danube basin and western Balkans. In Leconcept de territoires dans le paleolithique superieur europeen. Francois Djind-jian, Jansz Koizlowski, and Nuno Bicho, eds. Pp. 93–101. BAR InternationalSeries 1938. Oxford: Archaeopress.

Moncel, Marie-Helene, ed. 2008. Le site de Payre: occupations humaines dansla vallee du Rhone a la fin du pleistocene moyen et au debut du pleistocenesuperieur. Memoire de la Societe Prehistorique Francais 46. Paris: Societeprehistorique francais.

Moncel, Marie-Helene, Anne-Marie Moigne, and Jean Combier. 2005. Pre-Neandertal behaviour during isotopic stage 9 and the beginning of stage8: new data concerning fauna and lithics in the different occupation levelsof Orgnac 3 (Ardeche, southeast France): occupation types. Journal of Ar-chaeological Science 32:1283–1301.

Moncel, Marie-Helene, Anne-Marie Moigne, Youssef Sam, and Jean Combier.2011. The emergence of Neanderthal technical behavior: new evidence fromOrgnac 3 (Level 1, MIS 8), southeastern France. Current Anthropology 52(1):37–75.

Monigal, Katheryn. 2002. The Levantine leptolithic: blade production fromthe Lower Paleolithic to the dawn of the Upper Paleolithic. PhD dissertation,Southern Methodist University, Dallas, TX.

Morin, Eugene, and Veronique Laroulandie. 2012. Presumed symbolic use ofdiurnal raptors by Neanderthals. PLoS ONE 7(3):e32856, doi:10.1371/journal.pone.0032856.

Olle, Andreu, Marina Mosquera, Xose Pedro Rodrıguez, Arturo de Lombera-Hermida, Marıa Dolores Garcıa-Anton, Paula Garcıa-Medrano, Luna Pena,et al. 2013. The Early and Middle Pleistocene technological record fromSierra de Atapuerca (Burgos, Spain). Quaternary International 295:138–167.

Otte, Marcel. 1995. The nature of Levallois. In The definition and interpretationof Levallois technology. Harold Dibble and Ofer Bar-Yosef, eds. Pp. 117–243. Madison, WI: Prehistory Press.

Otte, Marcel, Isın Yalcinkaya, Janusz Kozlowski, Ofer Bar-Yosef, Ignacio LopezBayon, and Harun Taskıran. 1998. Long-term technical evolution and hu-man remains in the Anatolian Palaeolithic. Journal of Human Evolution 34:413–431.

Otte, Marcel, Isın Yalcinkaya, Janusz Kozlowski, Ofer Bar-Yosef, Harun Tas-

kiran, and Piere Noiret. 1995. Evolution technique au paleolithique anciende Karain. L’Anthropologie 90:529–561.

Pawlik, Alfred, and Jurgen Thissen. 2011. Hafted armatures and multi-com-ponent tool design at the Micoquian site of Inden-Altdorf, Germany. Journalof Archaeological Science 38:1699–1708.

Peresani, Marco, ed. 2003. Discoid lithic technology: advances and implications.BAR International Series 1120. Oxford: Archaeopress.

Peresani, Marco, Ivana Fiore, Monica Gala, Matteo Romandini, and AntonioTagliacozzo. 2011. Late Neandertals and the intentional removal of feathersas evidenced from bird bone taphonomy at Fumane Cave 44 ky B.P., Italy.Proceedings of the National Academy of Sciences of the USA, doi:10.1073/pnas.1016212108.

Peris, Josep Fernandez. 2007. La Cova del Bolomor: las industrias lıticas delPleistoceno Medio en el ambito del Mediterraneo peninsular. Valencia: J. Agui-lar.

Peris, Josep Fernandez, P. M. Calatayud, and R. M. Valle. 1997. Cova delBolomor: los primeros habitantes de las tierras valencianas. Valencia: Dipu-tacion de Valencia.

Plisson, Hugues, and Sylvie Beyries. 1998. Pointes ou outils triangulaires?donnees fonctionnelles dans le Mousterien Levantin. Paleorient 24(1):5–16.

Porat, Naomi, Michael Chazan, Henry P. Schwarcz, and Liora Kolska Horwitz.2002. Timing the Lower to Middle Paleolithic boundary: new dates fromthe Levant. Journal of Human Evolution 43:107–122.

Powell, Adam, Stephen Shennan, and Mark Thomas. 2009. Late Pleistocenedemography and the appearance of modern human behavior. Science 324:1298–1301.

Premo, Luke, and Steven Kuhn. 2010. Modeling effects of local extinctionson culture change and diversity in the Paleolithic. PLoS ONE 5(12):e15582.

Revillion, Stephane, and Alain Tuffreau, eds. 1994. Les industries laminairesau paleolithique moyen. Paris: CNRS.

Richards, Michael, and Erik Trinkaus. 2009. Isotopic evidence for the diets ofEuropean Neanderthals and early modern humans. Proceedings of the Na-tional Academy of Sciences of the USA 106(38):16034–16039.

Richter, Jurgen. 2011. When did the Middle Paleolithic begin? In Neanderthallifeways, subsistence and technology: 150 years of Neanderthal studies. NicholasConard and Jurgen Richter, eds. Pp. 99–110. New York, Springer.

Riel-Salvatore, Julien, and C. Michael Barton. 2004. Late Pleistocene tech-nology, economic behavior, and land-use dynamics in southern Italy. Amer-ican Antiquity 69(2):257–274.

Roebroeks, Wil, Jan Kolen, and Eelco Rensink. 1988. Planning depth, antic-ipation and the organization of Middle Palaeolithic technology: the archaicnatives meet Eve’s descendents. Helenium 28:17–34.

Roebroeks, Wil, Mark J. Siera, Trine Kellberg Nielsen, Dimitri De Loecker,Josep Maria Pares, Charles E. S. Arps, and Herman J. Muchere. 2012. Useof red ochre by early Neandertals. Proceedings of the National Academy ofSciences of the USA 109:1889–1894.

Roebroeks, Wil, and Paola Villa. 2011. On the earliest evidence for habitualuse of fire in Europe. Proceedings of the National Academy of Sciences of theUSA 108(13):5209–5214.

Rohling, Eelco J., Katherine M. Grant, Andrew P. Roberts, and Juan-CruzLarrasoana. 2013. Paleoclimate variability in the Mediterranean and RedSea regions during the last 500,000 years: implications for hominin migra-tions. Current Anthropology 54(suppl. 8):S183–S201.

Rolland, Nicholas. 2004. Was the emergence of homebases and domestic firea punctuated event? a review of the Middle Pleistocene record in Eurasia.Asian Perspectives 43(2):248–280.

Rotts, Veerle. 2009. The functional analysis of the Mousterian and Micoquianassemblages of Sesselfelsgrotte, Germany: aspects of tool use and haftingin the European Late Middle Palaeolithic. Quartar 56:37–66.

Ruebens, Karen. 2006. A typological dilemma: Micoquian elements in con-tinental northwestern Europe during the Last Glacial Cycle (MIS 5d–3).Lithics 27:58–73.

Sandgathe, Dennis, Harold Dibble, Paul Goldberg, Shannon McPherron, AlainTurq, Laura Niven, and Jamie Hodgkins. 2011. On the role of fire in Nean-dertal adaptations in western Europe: evidence from Pech de l’Aze IV andRoc de Marsal. Paleoanthropology 2011:216–242.

Santoja, Manuel, and Arturo Perez Gonzale. 2010. Mid-Pleistocene Acheuleanindustrial complex in the Iberian Peninsula. Quaternary International 223/224:154–161.

Santoja, Manuel, and Paola Villa. 2006. The Acheulean of western Europe.In Axe age: Acheulian tool-making from quarry to discard. N. Goren-Inbarand G. Sharon, eds. Pp. 429–478. London: Equinox.

Schick, Kathy. 1994. The Movius line reconsidered. In Integrative paths to the

S268 Current Anthropology Volume 54, Supplement 8, December 2013

past. R. Corruccini and R. Chiochon, eds. Pp. 569–596. Engelwood Cliffs,NJ: Prentice Hall.

Schiegel, Solveg, Paul Goldberg, Ofer Bar-Yosef, and Stephen Weiner. 1996.Ash deposits in Hayonim and Kebara Caves, Israel: macroscopic, micro-scopic and mineralogical observations, and their archaeological implication.Journal of Archaeological Science 23:763–781.

Sharon, Gonen. 2007. Acheulian large flake industries: technology, chronologyand significance. BAR International Series 1701. Oxford: Archaeopress.

Sharon, Gonen, and Peter Beaumont. 2006. Victoria West: a highly standard-ized prepared core technology. In Axe age: Acheulian tool-making fromquarry to discard. N. Goren-Inbar and G. Sharon, eds. Pp. 181–199. London:Equinox.

Shea, John J. 1988. Spear points from the Middle Paleolithic of the Levant.Journal of Field Archaeology 15:441–456.

———. 2006. The origins of lithic projectile point technology: evidence fromAfrica, the Levant and Europe. Journal of Archaeological Science 33:823–846.

———. 2011. Homo sapiens is as Homo sapiens was. Current Anthropology52(1):1–35.

Shimelmitz, Ron, Ran Barkai, and Avi Gopher. 2011. Systematic blade pro-duction at late Lower Paleolithic (400–200 ka) Qesem Cave, Israel. Journalof Human Evolution 61(4):458–479.

Slimak, Ludovic. 1999. Mise en evidence d’une composante laminaire et la-mellaire dans un complex Mousterien du sud de la France. Paleo 11:89–110.

Slimak, Ludovic, and Y. Giraud. 2007. Circulations sur plusieurs centaines dekilometres durant le paleolithique moyen: contribution a la connaissancedes societes Neandertaliennes. Comptes Rendus Palevol 6:359–368.

Slimak, Ludovic, Steven Kuhn, Helene Roche, Damase Mouralis, Hilke Bui-tenhuis, Nur Balkan-Atlı, Didier Binder, Catherine Kuzucuoglu, and HerveGuillou. 2008. Kaletepe Deresi 3 (Turkey): archaeological evidence for earlyhuman settlement in Central Anatolia. Journal of Human Evolution 54(1):99–111.

Snodgrass, J. Josh, and William Leonard. 2009. Neandertal energetics revisited:insights into population dynamics and life history evolution. Paleoanthro-pology 209:220–237, doi:10.4207/PA.2009.ART31.

Sorenson, Mark V., and William R. Leonard. 2001. Neandertal energetics andforaging efficiency. Journal of Human Evolution 40:483–495.

Soressi, Marie. 2002. Le Mousterien de tradition Acheuleenne du sud-ouestde la France. PhD dissertation, University of Bordeaux 1.

Soressi, Marie, and Francesco D’Errico. 2007. Pigments, gravures, parures: lescomportements symboliques controverses des Neandertaliens. In Neander-taliens: biologie et cultures. Bernard Vandermeersch, Bruno Maureille, andYves Coppens, eds. Pp. 297–309. Documents Prehistoriques 23. Paris,CTHS.

Soriano, Sylvain. 2000. Outillage bifacial et outillage sur eclat au paleolithiqueancien et moyen: coexistence et interaction. PhD dissertation, UniversiteParis X Nanterre.

Speth, John D., Liliane Meignen, Ofer Bar-Yosef, and Paul Goldberg. 2012.Spatial organization of Middle Paleolithic occupation X in Kebara Cave(Israel): concentrations of animal bones. Quaternary International 247:85–102.

Stiner, Mary C. 2013. An unshakable Middle Paleolithic? trends versus con-servatism in the predatory niche and their social ramifications. CurrentAnthropology 54(suppl. 8):S288–S304.

Stiner, Mary C., Ran Barkai, and Avi Gopher. 2011. Hearth-side socioeco-nomics, hunting and paleoecology during the late Lower Paleolithic atQesem Cave, Israel. Journal of Human Evolution 60(2):213–233.

Stiner, Mary C., Avi Gopher, and Ran Barkai. 2009. Cooperative hunting andmeat sharing 400–200 kya at Qesem Cave, Israel. Proceedings of the NationalAcademy of Sciences of the USA 106:13207–13212.

Stiner, Mary C., and Steven Kuhn. 2009. Paleolithic diet and the division oflabor in Mediterranean Eurasia. In The evolution of hominid diets: integratingapproaches to the study of Paleolithic subsistence. M. Richards and J. J. Hublin,eds. Pp. 155–167. New York: Springer.

Stiner, Mary C., Steven Kuhn, Todd A. Surovell, Paul Goldberg, Amy V.Margaris, Liliane Meignen, Stephen Weiner, and Ofer Bar-Yosef. 2005. Bone,ash, and shell preservation in Hayonim Cave. In The faunas of Hayonimcave, Israel. Mary C. Stiner, ed. Pp. 59–79. American School of PrehistoricResearch Bulletin 48. Cambridge, MA: Peabody Museum of Archaeologyand Ethnology, Harvard University.

Stiner, Mary C., Natalie D. Munro, and Todd A. Surovell. 2000. The tortoise

and the hare: small game use, the broad spectrum revolution, and Paleolithicdemography. Current Anthropology 41(1):39–73.

Stiner, Mary C., Natalie D. Munro, Todd A. Surovell, Eitan Tchernov, andOfer Bar-Yosef. 1999. Paleolithic population growth pulses evidenced bysmall animal exploitation. Science 283:190–194.

Surovell, Todd A. 2008. Toward a behavioral ecology of lithic technology: casesfrom Paleoindian archaeology. Tucson: University of Arizona Press

Surovell, Todd A., and P. Jeffery Brantingham. 2007. A note on the use oftemporal frequency distributions in studies of prehistoric demography. Jour-nal of Archaeological Science 34:1868–1877.

Surovell, Todd A., J. Finely, G. M. Smith, P. Jeffery Brantingham, and RobertL. Kelly. 2009. Correcting temporal frequency distributions for taphonomicbias. Journal of Archaeological Science 36:1715–1724.

Tennie, Claudio, Josep Call, and Michael Tomasello. 2009. Ratcheting up theratchet: on the evolution of cumulative culture. Philosophical Transitions ofthe Royal Society B 364(1528):2405–2415.

Thieme, Hartmut, and Reinhard Maier, eds. 1995. Archaologische ausgrabungenim braunkohlentagebau Schoningen. Hannover: Landkreis Helmstedt.

Tomasello, Michael. 1999. The cultural origins of human cognition. Cambridge,MA: Harvard University Press.

Tostevin, Gilbert. 2007. Social intimacy, artefact visibility, and acculturationmodels of Neanderthal-modern human interaction. In Rethinking the hu-man revolution: new behavioural and biological perspectives on the origins anddispersal of modern humans. Paul Mellars, Katie Boyle, Ofer Bar-Yosef, andChris Stringer, eds. Pp. 341–357. Cambridge: MacDonald Institute for Ar-chaeological Research.

Tuffreau, Alain. 1995. The variability of Levallois technology in northernFrance and neighboring areas. In The definition and interpretation of Levalloistechnology. Harold Dibble and Ofer Bar-Yosef, eds. Pp. 413–427. Madison,WI: Prehistory Press.

Turq, Alain. 2000. Le paleolithique inferieur et moyen entre les Valees de laDordogne et do Lot. Paleo (suppl. 2). Les Eyzies: SAMRA.

Vallverdu, Josep, Susana Alonso, Amelia Bargallo, Raul Bartrolı, Gerard Cam-peny, Angel Carrancho, Isabel Exposito, et al. 2012. Combustion structuresof archaeological level O and Mousterian activity areas with use of fire atthe Abric Romanı rockshelter (NE Iberian Peninsula). Quaternary Inter-national 247:313–324.

Villa, Paola. 2001. Early Italy and the colonization of western Europe. Qua-ternary International 75:113–130.

———. 2009. The Lower to Middle Paleolithic transition. In Sourcebook ofPaleolithic transitions. Marta Camps and Parth Chuahan, eds. Pp. 265–270.New York: Springer.

Villa, Paola, Paolo Boscato, Filomena Ranaldo, and Annamaria Ronchitelli.2009. Stone tools for the hunt: points with impact scars from a MiddlePaleolithic site in southern Italy. Journal of Archaeological Science 36(3):850–859.

Villa, Paola, and Michel Lenoir. 2009. Hunting and hunting weapons of theLower and Middle Paleolithic of Europe. In The evolution of hominin diets:integrating approaches to the study of Paleolithic subsistence. Jean-JacquesHublin and Michael Richards, eds. Pp. 59–87. Dordrecht: Springer.

White, Mark. 2000. The Clactonian question: on the interpretation of core-and-flake assemblages in the British Lower Paleolithic. Journal of WorldPrehistory 14(1):1–62.

White, Mark, and Nicholas Ashton. 2003. Lower Paleolithic core technologyand the origins of the Levallois method in northwestern Europe. CurrentAnthropology 44:598–608.

White, Mark, Nick Ashton, and Beccy Scott. 2011. The emergence, diversityand significance of the Mode 3 (prepared core) technologies. In The ancienthuman occupation of Britain. Nick Ashton, Simon G. Lewis, and Chris B.Stringer, eds. Pp. 53–66. Amsterdam: Elsevier.

Wiessner, Polly. 1983. Style and social information in Kalahari San projectilepoints. American Antiquity 48:253–276.

Wurz, Sarah. 2013. Technological trends in the Middle Stone Age of SouthAfrica between MIS 7 and MIS 3. Current Anthropology 54(suppl. 8):S305–S319.

Wynn, Thomas, and Frederick Coolidge. 2004. The expert Neandertal mind.Journal of Human Evolution 46:467–487.

Zilhao, Joao, Diego E. Angelucci, Ernestina Badal-Garcıa, Francesco d’Errico,Floreal Daniel, Laure Dayet, Katerina Douka, et al. 2010. Symbolic use ofmarine shells and mineral pigments by Iberian Neandertals. Proceedings ofthe National Academy of Sciences of the USA 107(3):1023–1028.

Current Anthropology Volume 54, Supplement 8, December 2013 S269

� 2013 by The Wenner-Gren Foundation for Anthropological Research. All rights reserved. 0011-3204/2013/54S8-0009$10.00. DOI: 10.1086/673386

The Evolutionary Implications of Variation inHuman Hunting Strategies and Diet Breadth

during the Middle Stone Ageof Southern Africa

by Jamie L. Clark and Andrew W. Kandel

CA� Online-Only Material: Supplement A

In this paper, we analyze faunal data from southern Africa in order to explore the nature and extent of variabilityin human hunting strategies and diet breadth during the Middle Stone Age (MSA). Our analysis incorporates datafrom eight sites that span marine isotope stages (MIS) 6–3 (∼170–40 ka). The sample includes both coastal andinland sites; while it primarily derives from cave and rock shelter contexts (Blombos, Die Kelders, Diepkloof, KlasiesRiver, Pinnacle Point, Sibudu, and Ysterfontein), we also include one open-air site (Florisbad). The data indicatemarked changes in subsistence over time—of particular note is a spike in the exploitation of small ungulates andother small mammals during MIS 4. Trends in shellfish utilization also suggest an increasing dietary breadth overtime, although shellfish consistently represent only a small portion of the overall diet. We close with a discussionof several causal mechanisms (environment, demography, technology, and cognition) that could account for thepatterning evidenced in the data. While environmental change appears to play an important role in shaping thecharacter of the data, we argue that no single causal factor can account for the full range of variability.

Introduction

The period known as the Middle Paleolithic (MP) in Eurasiaand the Middle Stone Age (MSA) in Africa (roughly spanning300–30 ka) was a critical one in the development of ourspecies. This period witnesses the rise and spread of anatom-ically modern humans (AMH) and also incorporates both theflorescence—and demise—of the Neanderthals. Yet ourknowledge about the nature and extent of variability in humanbehavior during this period remains limited despite a growingrecognition that a deeper understanding of this variability maybe critical to answering larger questions about the evolutionof human culture and the ultimate success of AMH overarchaic populations such as the Neanderthals (e.g., Clark2011; Kuhn and Hovers 2006 and references therein; Shea

Jamie L. Clark is Assistant Professor in the Department ofAnthropology at the University of Alaska Fairbanks (P.O. Box 757720,Fairbanks, Alaska 99775-7720, U.S.A. [jlclark7@alaska.edu]).Andrew W. Kandel is Project Archaeologist in the Role of Culturein Early Expansions of Humans (ROCEEH) at the HeidelbergAcademy of Sciences and Humanities (Rumelinstraße 23, 72070Tubingen, Germany). This paper was submitted 3 VII 13, accepted7 VIII 13, and electronically published 8 XI 13.

2011). On the Eurasian side, this is due in large part to thelong-held view that the MP was a period of relative stasis inhuman behavior (see Kuhn and Hovers 2006 for a detaileddiscussion). While somewhat greater attention has been paidto variation over both space and time within the MSA (e.g.,Clark 1993; McBrearty and Brooks 2000), particularly as ex-pressed in the lithic record, the view of MSA behavior aslargely static has also been fairly well entrenched (see Klein2009 for a more detailed discussion). This is particularly thecase when considering subsistence behavior; Klein (2009) ar-gued that “in most instances where controlled comparisonsare possible, differences in species abundance among Mous-terian/MSA sites or between Mousterian/MSA sites . . . appearto reflect differences in site environment rather than in oc-cupant behavior” (553).

Our goals for this paper are twofold. Our primary goal isto present the results of a large-scale attempt to explore var-iation in human hunting behavior and diet breadth in theMSA of southern Africa. To this end, we have compiled faunaldata from eight sites, including both coastal and inland lo-calities; while we hoped to incorporate material from theentire span of the MSA (∼300–30 ka; MIS 8–3), the natureof the available data forced us to restrict our study to materialsspanning ∼170–40 ka. Our second goal is to evaluate the

S270 Current Anthropology Volume 54, Supplement 8, December 2013

relative importance of environmental, demographic, tech-nological, and cognitive changes on the observed variabilityin subsistence behavior.

Previous Research

The Eurasian record. The majority of research on variationin hunting behavior within the Late Pleistocene (and its po-tential evolutionary implications) has focused on comparisonsbetween the MP and Upper Paleolithic (UP) in Eurasia. Themost influential work in this regard has been that of MaryStiner and colleagues (e.g., Stiner 2002, 2006; Stiner and Kuhn2006; Stiner and Munro 2002; Stiner, Munro, and Surovell2000), who utilized data from across the Mediterranean todemonstrate a marked shift in subsistence around the timeof the MP-UP transition, most notably as related to the ex-ploitation of small game. While MP populations primarilyexploited slow-moving, slow-growing small game such as tor-toises and marine mollusks, UP hunters focused on agile, fast-maturing animals such as hares and birds. This changemarked what Stiner considered to be a categorical shift inhuman predator-prey dynamics (Stiner 2006). However, thispattern is not necessarily a universal one. Work by Adler andcolleagues in the Caucasus (Adler and Bar-Oz 2009; Adler etal. 2006) demonstrated relatively little change in human sub-sistence behavior across the MP-UP transition. Similarly,when Gaudzinski-Windheuser and Niven (2009) compiledfaunal data from several Paleolithic sites in NW Europe, theyfound that subsistence behavior was relatively consistent be-tween the MP and earliest UP, with the most marked changesin subsistence occurring later in the UP.

Large-scale, specifically directed analyses of changes in hu-man subsistence behavior within the MP are rare (exceptionsinclude Delagnes and Rendu 2011; Gaudzinski 2006; Stiner2006, 2013; Stiner and Kuhn 1992). Outside the long-heldnotion of behavioral stasis during this period, there are other,more practical reasons why such studies are uncommon.Gaudzinski (2006) highlighted some of these reasons, includ-ing a dearth of faunal assemblages subjected to detailed taph-onomic analyses, which raises concerns about the compara-bility of data sets. Coarse chronological resolution has alsoprevented the development of “far-reaching conclusions”concerning evolutionary trends in subsistence behavior dur-ing the MP (Gaudzinski 2006).

Stiner (2006, 2013) looked at variation in MP subsistenceecology in the Mediterranean region. She found that very fewtrends were apparent within the MP with the possible excep-tion of mild harvesting pressure on slow-turnover prey pop-ulations after 50 ka. Although limiting their study to changeover time at a single site (Kebara Cave, Israel), work by Spethand Clark (2006) was consistent with these results; they arguedthat during the last ∼12,000 years of the MP, the inhabitantsof Kebara overhunted their large-game resources while alsointensifying their use of lower-ranked gazelle and fallow deer.

It is not yet clear whether the results obtained for the

Mediterranean are applicable across a wider geographic re-gion. However, a recent study by Delagnes and Rendu (2011)based on lithic and faunal materials recovered from 68 MPassemblages in southwestern France (spanning MIS 6–3) doc-umented clear (and seemingly interrelated) shifts in tech-nology and subsistence among late Neanderthals in the region,particularly after MIS 4. The authors were unable to identifya single factor that could account for the documented vari-ability, and ultimately they stressed the importance of mul-ticausal explanatory models for understanding behavioral di-versity in the archaeological record.

The African record. Just as research on variation in theEurasian record has primarily focused on comparisons of theMP and UP, so too has research on the African record largelycentered on comparisons of the MSA and Later Stone Age(LSA). Best known in this respect is the work of Klein andcolleagues (e.g., Klein 1979, 2001, 2009; Klein and Cruz-Uribe1996, 2000; Steele and Klein 2009), who have argued thatMSA populations were less effective hunters than their LSAcounterparts. Utilizing data from several MSA and LSA as-semblages, Klein demonstrated that LSA sites generally pre-serve higher frequencies of dangerous prey (particularly buf-falo; e.g., Klein and Cruz-Uribe 2000). He interprets thesedata as indicating that MSA foragers were limited in theirability to capture certain types of ungulate prey. However,this argument has been contested; most notably, Faith (2008)utilized a wider range of samples and a variety of statisticalmethods to argue that dangerous game (including buffaloesand wild pigs) are equally abundant in both the MSA andLSA. Although Weaver, Steele, and Klein (2011) critiquedFaith’s methodology, their reanalysis nonetheless corroboratesFaith’s primary result that MSA sites do not appear to bedominated by eland or be deficient in dangerous game com-pared with LSA sites (see Faith 2011b).

Steele and Klein (2009) presented the results of a muchmore comprehensive comparison of MSA and LSA subsis-tence. In this study, they applied the framework used forexploring variation in subsistence strategies in the EurasianMP and UP to the MSA/LSA record from coastal South Africa,utilizing data from nine archaeological sites. Although fo-cusing on the coast, Steele and Klein deliberately includedsites from different ecological zones, arguing that combiningthese zones made it more likely that any patterns identifiedin the data reflected changes in human behavior rather thanclimatic/environmental variation.

Steele and Klein (2009) looked for variation in several di-mensions: (1) large-game exploitation (using both prey mor-tality profiles and measures of processing intensity), (2) theabundance and diversity of small game (particularly small,slow game such as tortoises and shellfish vs. small, fast gamesuch as hares and birds), and (3) shellfish and tortoise size,which can serve as an indicator of collection intensity. Theyconcluded that LSA populations consumed a wider variety ofresources and exploited the available resources more inten-

Clark and Kandel Variation in Middle Stone Age Hunting Strategies S271

Figure 1. Map of MSA sites discussed in the text. Dots indicatesites that were included in the study; Xs indicate sites that wereexcluded. A11 p Apollo 11, BBC p Blombos Cave, BC pBorder Cave, BFA p Bundu Farm, BPA p Boomplaas Cave,BRS p Bushman Rock Shelter, DK1 p Die Kelders 1, DRS pDiepkloof Rock Shelter, EQC p Equus Cave, FLO p Florisbad,HB p Herold’s Bay, KB p Kalkbank, KK p Kudu Koppie,KRM p Klasies River Main Site, MLK p Melikane, PLL pPlover’s Lake, PP13B p Pinnacle Point Cave 13B, SCB p Strath-alan Cave B, SEH p Sehonghong, SIB p Sibudu, YFT1 pYsterfontein 1.

sively than their MSA counterparts, and they suggested thatthese differences related both to larger and denser humanpopulations and perhaps to the development of LSA tech-nologies.

As was the case for the Eurasian MP, larger-scale attemptsto explore variation in human hunting behavior within theMSA have been relatively limited (exceptions include Dus-seldorp 2012; Lombard and Clark 2008; Thompson 2010b).There are a number of reasons for this. First, as was the casefor the MP, this is due in part to notions of relative stasis inhuman behavior during the MSA (e.g., Klein 2009). Second,the number of relatively well-published MSA faunal assem-blages was quite limited until recently. Just as the scale ofresearch into the MSA has markedly increased in the past twodecades, so too has the number of sites with published faunaldata. Since 2000, details on the fauna from six key MSA sites(Blombos, Die Kelders, Diepkloof, Pinnacle Point 13B, Si-budu, and Ysterfontein) have been published, with data fromseveral other sites expected in the near future (e.g., Boom-plaas, Kudu Koppie, Melikane, and Sehonghong). Finally,broader-scale comparisons of MSA faunal assemblages havebeen hindered by the use of different analytical and quantifi-cation methods such that all data sets are not necessarilydirectly comparable (see Thompson 2010b for a detailed dis-cussion of some of the theoretical, empirical, and method-ological obstacles facing those attempting larger comparativestudies of MSA subsistence).

The view of MSA subsistence behavior as static was soentrenched that as recently as 2008, Lombard and Clark fo-cused their attention on documenting whether variation inhuman hunting behavior even existed over the course of theMSA. Combining lithic and faunal data sets, they summarizedcurrent interpretations about hunting behavior spanning fromthe lowermost pre–Still Bay at Blombos (∼100 ka) throughthe post–Howieson’s Poort at Sibudu (∼58 ka); however, theydid not attempt to identify any broader directional trends inthe data.

Thompson (2010b) explored variation in faunal exploita-tion at two MSA sites that preserve deposits spanning ∼170–73 ka. She personally analyzed the faunal material from Blom-bos and Pinnacle Point 13B, promoting comparability of thetwo data sets. Looking at variation in dietary composition,processing, and transport behavior, she concluded that sub-sistence decisions were not uniform over time or space. Shedid not identify any clear temporal trends but rather proposedthat the identified variation may be closely related to differ-ences in site context.

In light of the recent publication of faunal data from anumber of well-excavated MSA sites preserving extensive,well-stratified deposits, the time is ripe for a larger-scale con-sideration of variation in MSA subsistence behavior over time.While recognizing that this type of study is not entirelystraightforward (see below for a more detailed discussion ofthe caveats), we believe that the available data are sufficientlycomparable to produce meaningful results.

Materials

Criteria for Inclusion

Given that our goal is to explore variation in human huntingstrategies and dietary breadth over time, we limited our anal-ysis to well-dated sites that preserved both faunal and lithicremains. A further requirement was that the assemblages wereprimarily anthropogenic in origin (see fig. 1 for a map of sitesdiscussed in the text and table A1 in CA� online supplementA for a list of excluded sites). As such, we excluded remainsfrom well-known MSA sites such as Boomplaas (Faith 2011a),Equus Cave (Klein, Cruz-Uribe, and Beaumont 1991), the“old collections” at Florisbad (Brink 1988; Kuman and Clarke1986), Herolds Bay Cave (Brink and Deacon 1982), Kalkbank(Hutson and Cain 2008), and Plovers Lake (de Ruiter et al.2008), which appeared to be either natural death sites oraccumulated primarily by agents other than humans.

We also chose to limit our analysis to sites for which mam-malian faunal data were reported as number of identifiedspecimens (NISP) rather than minimum number of individ-uals (MNI). While a detailed discussion of the relative merits(and faults) of these quantification methods is beyond thescope of this paper, our choice was based more on practicalconsiderations. We wanted to limit our analysis to a singlequantification method, and a larger proportion of sites pre-sented mammalian faunal data in NISP form. As such, we

Table 1. Mammalian assemblages included in the analysis and grouped by marine isotope stage (MIS)

AssemblageAge range

(ka)NISP

(mammals) References

MIS 6:PP13B LB Silt 349–152 6 Marean et al. 2010; Rector and Reed 2010PP13B LC-MSA Lower 174–153 2,826 Marean et al. 2010; Thompson 2010aPP13B Lower DB Sand 164–159 147 Marean et al. 2010; Rector and Reed 2010; Thompson 2010a

MIS 5:YFT1 Lower 140–124 1,222 Avery et al. 2008YFT1 Middle 128–114 266 Avery et al. 2008YFT1 Upper 135–123 1,836 Avery et al. 2008PP13B LBG Sand 134–94 1,391 Marean et al. 2010; Rector and Reed 2010; Thompson 2010aPP13B LC-MSA Middle/Upper 130–120 314 Marean et al. 2010; Thompson 2010aFLO Unit F 127–115 268 Brink 1987; Kuman, Inbar, and Clarke 1999DRS Lower MSA 118–90 300 Steele and Klein 2013; Tribolo et al. 2013DRS MSA-Mike n.d. 63 Steele and Klein 2013; Tribolo et al. 2013DRS Pre-SB Lynn 110–90 37 Steele and Klein 2013; Tribolo et al. 2013DRS Still Bay 119–99 166 Steele and Klein 2013; Tribolo et al. 2013DRS Early HP 119–95 290 Steele and Klein 2013; Tribolo et al. 2013KRM LBS 115–110 419 Feathers 2002; Wurz 2002; van Pletzen 2000KRM Cave 1 (38/LBS) 115–110 853 Feathers 2002; R. G. Klein, unpublished dataPP13B Lower Roof Spall 114–106 1,510 Marean et al. 2010; Rector and Reed 2010; Thompson 2010aKRM Cave 1 (37/RBS) 110–95 1,173 Feathers 2002; Wurz 2002; R. G. Klein, unpublished dataKRM Cave 1 (17/SASL) 110–95 1,493 Feathers 2002; Wurz 2002; R. G. Klein, unpublished dataKRM Cave 1 (16/SASU) 110–95 1,794 Feathers 2002; Wurz 2002; R. G. Klein, unpublished dataKRM Cave 1 (15/SAS) 110–95 944 Feathers 2002; Wurz 2002; R. G. Klein, unpublished dataKRM SASb 110–95 1,938 Feathers 2002; van Pletzen 2000; Wurz 2002BBC M3 105–91 287 Henshilwood et al. 2011; Thompson and Henshilwood 2011BBC M3 105–91 1,212 Henshilwood et al. 2001, 2011PP13B Upper DB Sand 102–91 2,950 Marean et al. 2010; Rector and Reed 2010; Thompson 2010aPP13B LB Sand 2 102–91 4 Marean et al. 2010; Rector and Reed 2010PP13B SB Sand/ Upper Roof Spall 98–91 5,330 Marean et al. 2010; Rector and Reed 2010; Thompson 2010aDRS MSA-Jack 98–80 114 Steele and Klein 2013; Tribolo et al. 2013KRM Cave 1 (14/SASR) 95–85 3,357 Feathers 2002; Wurz 2002; R. G. Klein, unpublished dataKRM SASm 95–72 884 Jacobs et al. 2008a; van Pletzen 2000KRM SASt 95–72 205 Jacobs et al. 2008a; van Pletzen 2000PP13B LB Sand 1 94–91 763 Marean et al. 2010; Rector and Reed 2010; Thompson 2010aDRS Intermediate HPa 94–57 955 Steele and Klein 2013; Tribolo et al. 2013BBC M2 lower 91–77 121 Henshilwood et al. 2011; Thompson and Henshilwood 2011BBC M2 upper 80–74 468 Henshilwood et al. 2011; Thompson and Henshilwood 2011BBC M2 91–74 1,167 Henshilwood et al. 2001, 2011BBC M1 74–68 934 Henshilwood et al. 2011; Thompson and Henshilwood 2011BBC M1 74–68 1,727 Henshilwood et al. 2001, 2011

MIS 4:KRM Cave 1 (13/WS) 78–63 186 d’Errico, Garcia Moreno, and Rifkin 2012; R. G. Klein, un-

published dataKRM Upper 72–58 1,017 Jacobs et al. 2008a; van Pletzen 2000DK1 Layer 15 70–60 336 Feathers and Bush 2000; Klein and Cruz-Uribe 2000DK1 Layer 14 70–60 2,126 Feathers and Bush 2000; Klein and Cruz-Uribe 2000DK1 Layer 13 70–60 374 Feathers and Bush 2000; Klein and Cruz-Uribe 2000DK1 Layer 12 70–60 960 Feathers and Bush 2000; Klein and Cruz-Uribe 2000DK1 Layer 11 70–60 458 Feathers and Bush 2000; Klein and Cruz-Uribe 2000DK1 Layer 11 70–60 541 Feathers and Bush 2000; Marean et al. 2000DK1 Layer 10 70–60 1,996 Feathers and Bush 2000; Klein and Cruz-Uribe 2000DK1 Layer 10 70–60 2,424 Feathers and Bush 2000; Marean et al. 2000DK1 Layer 9 70–60 4,414 Feathers and Bush 2000; Klein and Cruz-Uribe 2000DK1 Layer 8 70–60 107,939 Feathers and Bush 2000; Klein and Cruz-Uribe 2000DK1 Layer 7 70–60 9,200 Feathers and Bush 2000; Klein and Cruz-Uribe 2000DK1 Layer 6 70–60 56,476 Feathers and Bush 2000; Klein and Cruz-Uribe 2000DK1 Layer 4/5 70–60 15,414 Feathers and Bush 2000; Klein and Cruz-Uribe 2000SIB HP 67–60 5,463 Jacobs et al. 2008a; J. L. Clark, unpublished data

MIS 3:SIB Post HP MSA 2 62–56 1,090 Jacobs et al. 2008b; J. L. Clark, unpublished dataSIB Post HP MSA 1 62–56 843 Jacobs et al. 2008b; J. L. Clark, unpublished data

Clark and Kandel Variation in Middle Stone Age Hunting Strategies S273

Table 1 (Continued)

AssemblageAge range

(ka)NISP

(mammals) References

DRS Late HP 57–47 1,023 Steele and Klein 2013; Tribolo et al. 2013DRS Post HP n.d. 481 Steele and Klein 2013; Tribolo et al. 2013SIB Late MSA 52–44 1,985 Jacobs et al. 2008b; Wadley, Plug, and Clark 2008SIB Final MSA 42–35 303 Jacobs et al. 2008b; Wadley, Plug, and Clark 2008

Note. Age ranges are best estimates based on the data found in the literature. Assemblages are ordered chronologically except in cases that wouldreverse their stratigraphic sequence. NISP p number of identified specimens; n.d. p no date.a The Intermediate HP at DRS contains layers dating to both MIS 5 and MIS 4. Because the faunal data were provided only for the unit as a wholeand because the majority of the dates fall within MIS 5, we included it in the MIS 5 sample. See CA� online supplement A.

excluded sites with only MNI data—for example, the Wendt1968–1972 excavations at Apollo 11 (Thackeray 1979), BorderCave (Klein 1977), Kudu Koppie (Pollarolo et al. 2010), andthe Opperman 1988–1991 excavations at Strathalan Cave B(Brink’s data in Opperman 1996). Although Klein (1976) onlypresented MNI data from Cave 1 at Klasies River, he kindlyprovided us with NISP counts for the same sample, allowingus to include these results in our analysis. However, becauseresearchers primarily presented MNI counts for the molluscanfauna, we report shellfish data using this quantificationmethod.

We excluded other sites because faunal remains were as-signed to taxa in a way that was inconsistent with the otherdata sets (e.g., while NISP data are available for the newexcavations at Apollo 11, Vogelsang et al. [2010] utilized cat-egories such as “small or medium antelope,” making it dif-ficult to compare data). Finally, the nature of the availabledates meant that some sites could not be assigned to a specificmarine isotope stage (MIS); this was the case for Bundu Farm(Brink’s data in Kiberd 2006; Hutson 2012) and BushmanRock Shelter (Badenhorst and Plug 2012; Louw 1969; Plug1981).

Sites/Assemblages Included in the Study

Based on these criteria, the total vertebrate sample comprises60 assemblages from eight sites (table 1): Blombos Cave(BBC), Die Kelders Cave 1 (DK1), Diepkloof Rock Shelter(DRS), Florisbad (FLO), Klasies River Main Site (KRM), Pin-nacle Point Cave 13B (PP13B), Sibudu Cave (SIB), and Yst-erfontein 1 (YFT1). Shellfish data were available for 28 as-semblages from five of these sites (table 2): BBC, DRS, KRM,PP13B, and YFT1. More detailed information about site con-text and the faunal remains can be found in supplement A.

While we recognize that these sites may have been excavatedusing different recovery strategies and that the zooarchaeol-ogists responsible for producing and publishing the faunaldata may have employed differing analytical strategies, wecontend that the assemblages included here are sufficientlycomparable so as to provide a reliable basis for a considerationof variation in human hunting behavior during the course ofthe MSA. With a single exception—the Singer and Wymer(1982) excavations at Klasies River—all of the material in-

cluded here comes from excavations that employed modernrecovery strategies, including the use of fine mesh sieves (!5mm), and thus recovery biases should be limited. By restrict-ing our analysis to sites providing NISP counts for mam-malian fauna, we avoid potential issues arising from the useof multiple quantification methods and also avoid any po-tential complications related to the nonstandardized calcu-lation of MNI values. Although more thorough taphonomicanalyses are clearly warranted, we are confident that humanswere the major contributors to the faunas under consider-ation. Finally, similar to Steele and Klein (2009), we feel thatthe inclusion of remains from a variety of ecological zones(including inland and coastal sites as well as sites from sum-mer and winter rainfall areas) makes it more likely that anytrends identified in the data signify real differences in humanbehavior rather than simply reflecting environmental change.

Methods

For all analyses reported here, we grouped the assemblagesby MIS; this made the resulting data easier to manage andhad the added benefit of making evaluating whether the po-tential variance in the data was driven primarily by climatechange more straightforward. The number of mammalianassemblages dated to MIS 5 and 4 ( and 16, respec-n p 35tively) are larger than those for MIS 6 and 3 ( and 6,n p 3respectively); in fact, the sample from MIS 6 is limited to onesite (PP13B), while that from MIS 3 is limited to two (DRSand SIB). For shellfish, MIS 5 also contained the largest num-ber of assemblages ( ), with MIS 6, 4, and 3 representedn p 22by two assemblages each. Given the nature of the sample andthe variation in sample size across MIS, our analyses are bynecessity more qualitative than quantitative in nature.

We considered grouping the assemblages by archaeologicalcultures in order to explore other potential dimensions ofvariation in hunting behavior, as some proportion of the stonetools likely served as weapon tips. However, this considerationproved to be untenable in large part because there are toofew conventions for assigning material to specific MSA cul-tures. Especially before and after the better-known Still Bay(SB) and Howieson’s Poort (HP) phases of the MSA, lithicassemblages are highly variable and trends are poorly under-stood (see Wurz 2013).

S274 Current Anthropology Volume 54, Supplement 8, December 2013

Table 2. Shellfish assemblages included in the analysis grouped by marine isotope stage (MIS)

Assemblage MNI Weight (g) kcal People days References

MIS 6:PP13B Lower LC-MSA 28 156 64 .0 Jerardino and Marean 2010PP13B West Lower Sand 1 2 1 .0 Jerardino and Marean 2010

MIS 5:YFT1 Lower 7,087 109,744 49,149 24.6 Avery et al. 2008YFT1 Middle 1,231 14,109 7,404 3.7 Avery et al. 2008YFT1 Upper 1,540 25,870 13,146 6.6 Avery et al. 2008PP13B West Middle Sand 11 30 18 .01 Jerardino and Marean 2010PP13B Middle LC-MSA 53 398 175 .1 Jerardino and Marean 2010PP13B Upper LC-MSA 35 156 64 .03 Jerardino and Marean 2010DRS Lower MSA . . . 22 17 .01 Steele and Klein 2013DRS MSA-Mike . . . 1 1 .001 Steele and Klein 2013DRS Pre-SB Lynn . . . 17 14 .01 Steele and Klein 2013DRS Still Bay . . . 16 10 .005 Steele and Klein 2013DRS Early HP . . . 69 35 .02 Steele and Klein 2013KRM LBS (MSA I) . . . 6,577 . . . . . . Langejans et al. 2012PP13B Lower Roof Spall 238 2,066 846 .4 Jerardino and Marean 2010KRM SAS (MSA II) . . . 4,656 . . . . . . Langejans et al. 2012BBC M3 2,692 98,218 75,456 37.7 Langejans et al. 2012PP13B West Upper Sand 49 369 206 .1 Jerardino and Marean 2010PP13B Upper Roof Spall 234 2,292 1,012 .5 Jerardino and Marean 2010PP13B SB Sand 68 823 362 .2 Jerardino and Marean 2010DRS MSA-Jack . . . 111 63 .03 Steele and Klein 2013DRS Intermediate HP . . . 92 45 .02 Steele and Klein 2013BBC M2 3,118 24,253 15,867 7.9 Langejans et al. 2012BBC M1 3,203 24,363 14,801 7.4 Langejans et al. 2012

MIS 4:KRM Upper (HP) 605 . . . . . . . . . Langejans et al. 2012KRM Upper (MSA III) 213 . . . . . . . . . Langejans et al. 2012

MIS 3:DRS Late HP . . . 996 630 .3 Steele and Klein 2013DRS Post-HP . . . 1065 783 .4 Steele and Klein 2013

Notes. Nutritional values for shellfish (kcal, people-days) calculated according to Buchanan (1988). See text for details on calculations. Assemblagesare ordered chronologically except in cases that would reverse their stratigraphic sequence.

Exploring Variation in Hunting Strategies and Diet Breadth

Prey choice models predict that top ranked prey (those witha higher energetic yield relative to search, pursuit, and han-dling costs) will always be taken on encounter; if encounterrates with preferred prey decline, hunters are expected tobroaden their diets by including more lower-yield resources(e.g., Bird and O’Connell 2006; Kaplan and Hill 1992; Lupo2007). Within this framework, there are a number of wayswe can utilize the available data. Ethnographic and empiricalstudies have shown that when acquired individually, prey rankscales closely with body size such that large prey are generallyhigher ranked than smaller animals (Broughton and Grayson1993; Hawkes, Hill, and O’Connell 1982; Kelly 1995; Lupo2007; Lupo and Schmitt 2005). As such, an increase in theexploitation of small (and presumably lower-ranked) animalsmay indicate an increase in dietary breadth.

Ungulate Size Classes

The first way we approached the data was to explore variationin the exploitation of the various ungulate size classes over

time with the assumption that the smallest ungulates wouldbe of the lowest rank. Bovids are the most common ungulatesin the region under consideration; given the number of dif-ferent bovid species present in southern Africa and the re-sulting difficulty in identifying fragmentary remains to spe-cies, the majority of the bovid remains in MSA assemblagesare assigned to size class (see table 3 for details); these classeshave also been utilized more generally (cf. Thompson 2010a;Thompson and Henshilwood 2011). For this analysis, we in-cluded all ungulate remains identifiable to species as well asthose identifiable to size class.

Large versus Small (!4.5-kg) Mammals

An increase in the exploitation of small game may also reflectan expansion of dietary breadth. Although work by Stinerand colleagues (e.g., Stiner, Munro, and Surovell 2000) fo-cused not just on the relative frequency of small game com-pared with larger game but more specifically on changes inthe types of small game present in any given assemblage (i.e.,slow-moving/slow-growth prey vs. agile/fast-maturing prey),

Clark and Kandel Variation in Middle Stone Age Hunting Strategies S275

Table 3. Size classes utilized for southern African faunal assemblages

Size classesWeight

(kg) Species (list not all inclusive)Brain (1974) Klein (1976)

Size 1 Small 4.5–23 Blue duiker, common duiker, klipspringer, steenbokSize 2 Small-medium 23–84 Mountain reedbuck, bushbuck, springbok, impalaSize 3 Large-medium 85–295 Red hartebeest, wildebeest, roan antelope, kudu, zebraSize 4 Large 296–900 African buffalo, elandSize 5 Very large 1900 Giant buffalo (extinct)

the available data limited us to a more general comparisonof small versus larger game. Furthermore, because data onnonmammalian game (such as tortoises, birds, and fish) werenot consistently available, we restricted our analysis to smallmammals.

Given that 4.5 kg is a common cutoff for small game amongzooarchaeologists working in southern Africa (cf. Marean etal. 2000; Thompson 2010a; Thompson and Henshilwood2011), we included species with an average weight of less than4.5 kg in our “small mammal” category but excluded micro-mammal remains, defined as less than 0.75 or 0.3 kg, de-pending on the analyst. We also included generic “small mam-mal” NISP counts if these were clearly defined as representinganimals of the appropriate size. For this particular analysis,we excluded those assemblages for which data on small mam-mals were not provided.

Dangerous versus Docile Prey

A third approach to exploring changes in diet breadth—andhunting strategies more broadly—involves a consideration ofvariation in the exploitation of dangerous or aggressive game.As previously discussed, Klein has proposed that the exploi-tation of dangerous game serves as an indicator of huntingability (and thus perhaps cognitive capacity; Klein 2001; Kleinand Cruz-Uribe 1996). On a broader level, though, one couldargue that dangerous/aggressive game would be comparativelylow-ranked resources, and thus an increase in the exploitationof such taxa could serve as an indicator of increased dietarybreadth (Clark 2009, 2011). This is because capturing thisgame could potentially pose a greater risk (and thus wouldbe higher cost) than the pursuit of more docile game.

In exploring this issue, we began by comparing the relativefrequencies of eland (Taurotragus oryx) and buffalo (Synceruscaffer and Pelorovis antiquus) not only because this compar-ison has been a popular one among those looking at variationin human subsistence behavior (e.g., Dusseldorp 2012; Faith2008; Klein 1979, 2001) but also because the species in ques-tion have broadly similar habitat requirements, and as such,variation in relative frequency of these taxa should not simplyreflect changes in environmental conditions over time.

We also compared the relative frequency of suids—warthog(Phacochoerus africanus) and bushpig (Potamochoerus larva-tus)—with the frequency of similar sized bovids (Size 2) be-

cause the wild pigs are known to be aggressive and wouldhave been more dangerous to hunt than the bovids. The twosuids occupy different habitat types, with warthogs occupyingmore open environments and bushpigs occupying closed en-vironments. Thus variation in the relative frequency of suidsshould not simply be a reflection of environmental change.

Use of Nonvertebrate Resources: Shellfish Exploitation

An increase in the exploitation of nonvertebrate resourcesmay also indicate an expansion of dietary breadth. Here, weare particularly interested in looking at variation in shellfishexploitation during the MSA. Shellfish represent a predictableresource that is easy to collect, although they have a highprocessing cost relative to the number of calories provided.Shellfish appear as early as ∼164 ka in the southern AfricanMSA (Marean et al. 2007). While people collected shellfishin what often appears to be large amounts, the calories thatshellfish provide are surprisingly low compared with verte-brate resources (Buchanan 1988). This suggests that shellfishrepresented just one part of a varied MSA subsistence strategy.

We explored variation in the diversity of shellfish over timeby reviewing the published shellfish data and combining theshells into groups (see table A2). Because we are interestedin examining human diet, we excluded incidental species thatwere not likely exploited as food. Of the five sites for whichshellfish data are available, DRS and YFT1 represent SouthAfrica’s cooler west coast, while BBC, KRM, and PP13B comefrom the warmer south coast. Because of differences in oceantemperature along each coast, some shell species are not pres-ent in both regions. Therefore, we examined the data fromeach coast separately.

We also explored variation in the relative contribution ofshellfish to the diet. To accomplish this, we converted shellweights into nutritional values to evaluate the dietary im-portance of shellfish, although weights were not available forKRM. The nutritional values stem from Buchanan’s (1988)investigation of the dietary contribution of marine and ter-restrial resources. He reports that two limpet species (Cymbulagranatina and Scutellastra granularis) yield an average of 350kJ of energy for each 100 g of shell present, whereas blackmussels (Choromytilus meridionalis) generate 150 kJ from thesame shell mass. Because nutritional data were not availablefor other shellfish, we assume that gastropods (including other

S276 Current Anthropology Volume 54, Supplement 8, December 2013

limpet species) are similar to limpets and that bivalves aresimilar to the black mussels analyzed by Buchanan. We alsoassume a modern nutritional requirement of 2,000 kcal (8,370kJ) per person per day for calculating person-days of food.

The third way we examined variation in subsistence wasto look at the relative frequency of ungulates to shellfish. Forconsistency in our comparison, we only used data from as-semblages that contained both ungulates and shellfish. Werecognize the potential problems inherent in comparing un-gulate NISP with shellfish MNI values. However, because weare interested in analyzing change over time, this index pro-vides a relative measure of the intensity with which peopleutilized shellfish.

Other Methods for Exploring Variation in Subsistence

Two lines of evidence that have been commonly utilized whenexploring variation in hunting strategies and/or diet breadthcould not be considered here. The first relates to changes inprey mortality profiles; Klein and colleagues (Klein 1979,2001; Klein and Cruz-Uribe 1996; Steele and Klein 2009) havelargely utilized this data in discussions about the huntingabilities of MSA versus LSA populations; for example, theyhave argued that when MSA populations do capture large,dangerous game, they tend to capture very young or old in-dividuals as opposed to prime-aged adults. Unfortunately, agedata were not consistently available, so we could not explorevariability along this dimension.

More intensive carcass use and processing has also beencited as evidence for subsistence intensification (and thuswider dietary breadth), for example, the exploitation of mar-row from skeletal parts with low marrow utility (or from smallgame) and the production of bone grease have relatively highprocessing costs relative to energetic yield (e.g., Binford 1978;Munro and Bar-Oz 2005). A consideration of variation incarcass use and processing requires not only skeletal elementfrequency data but also detailed taphonomic data. Unfortu-nately, these data are not yet available for most of the assem-blages under consideration.

Controlling for Environmental Change: Browsersversus Grazers

Finally, given that the data under consideration incorporatematerials from four MISs, including two glacial/stadial periods(MIS 6 and 4) and two interglacial/interstadial periods (MIS5 and 3), we were also interested in exploring what degree ofvariance in the data might be driven primarily by environ-mental change. To this end, we compared the relative fre-quency of browsers and grazers by MIS. Classification wasbased on Gagnon and Chew (2000) and Skinner and Chim-imba (2005). Note that there are a number of extinct speciesin the sample. Because we cannot be certain about their di-etary preferences, we exclude extinct taxa from the analysis.

Results

Variation in Ungulate Size Classes over Time

Figure 2a presents data on variation in the frequency of thevarious ungulate size classes by MIS in both graphical andtabular format. While these data appear to show a verymarked increase in the smallest ungulates in MIS 4, this pat-tern is potentially driven by a single assemblage, DK1 Layer6, which had an NISP of more than 18,000 Size 1 ungulates.Given the fact that raptors have been implicated as a majorcontributor to the small ungulate fauna in other layers at thesite (particularly DK1 Layer 10), we removed this assemblagefrom the analysis to see how the pattern would change. How-ever, we should note that new work by Armstrong (2013)indicates that humans contributed to the collection of bothsmall ungulates and other small game at the site. Figure 2bpresents the results of this analysis. The same general trendis apparent; we still see a marked increase in the smallestungulates from MIS 5 to MIS 4 with a subsequent (and evenmore marked) reduction in MIS 3. If we consider the smallestungulates to be the lowest-ranked prey, these data are con-sistent with an expansion of dietary breadth from MIS 6 toMIS 4 followed by a decline in diet breadth in MIS 3.

Another pattern in these data can be seen in the relativeproportion of large and very large ungulates (Size 3 andlarger), which, according to prey choice models, should bethe highest-ranked prey. Comparatively speaking, large andvery large ungulates are more frequent during the interglacial/interstadial periods (MIS 5 and 3; 150% in both cases) thanin MIS 6 and 4. Given that the larger ungulates in the samplealso tend to be those that occupy more open environments,this could imply that variation in the presence of the largestungulates is linked to environmentally driven changes in theavailability of this game.

Variation in Large versus Small Mammals over Time

Figure 3a presents data on the occurrence of large versus smallmammals. As was the case for the previous analysis, one sam-ple from DK1 was large enough to potentially skew the re-sults—DK1 Layer 8 contained more than 100,000 Cape dunemole rat (Bathyergus suillus) bones. Again, in the absence ofdetailed taphonomic analysis, it is unclear whether these wereaccumulated predominantly by humans. While Klein andCruz-Uribe (2000) point out that those layers at DK1 withthe most Cape dune mole rat bones are also those with evi-dence for less intensive human occupation, new work byArmstrong (2013) implicates humans in the collection of atleast some of these remains. Figure 3b shows the results ofthe analysis with the assemblage from DK1 Layer 8 removed.Again, the general pattern remained the same. In fact, thedata show the same trend as the ungulate size-class data—arelatively small increase in small mammals from MIS 6 toMIS 5, a very marked increase in MIS 4, and then an evenmore pronounced decline in MIS 3. If small mammals are

Clark and Kandel Variation in Middle Stone Age Hunting Strategies S277

Figure 2. a, Ungulate size-class data by MIS (in NISP). b, Ungulate size-class data by MIS (in NISP) with DK1 Layer 6 removedfrom analysis.

considered low-ranked prey, this indicates an increase in di-etary breadth from MIS 6 to MIS 4.

In considering the implications of these data, two pointsdeserve mention. First, as previously discussed, our samplefrom MIS 6 is limited to data from PP13B (Thompson 2008,2010a). However, Thompson conducted detailed taphonomicanalyses of the PP13B fauna and argued that the relativelylow frequency of small game was a real phenomenon and notsimply a reflection of taphonomic factors unique to this site.In fact, she drew particular attention to the relative dearth ofsmall game at PP13B especially compared with other MSAsites along the southwestern coast of South Africa (Thompson2010a).

There is also a more general question about what propor-tion of the small mammals present in MSA sites was intro-duced by human activity. The small mammals from many ofthe assemblages under consideration have not been subjectedto detailed taphonomic analyses; for example, Milo’s (1994,1998) taphonomic study of the KRM fauna was limited tothe larger mammal (14.5 kg) remains, as was Marean et al.’s(2000) taphonomic study of the material from DK1 layers 10and 11. Even when taphonomic analyses have been com-pleted, the primary accumulating agent for the small mam-mals is not always clear. For example, although Thompson’s(2010a) work at PP13B indicated some human involvementwith the small fauna, direct evidence for human interaction

S278 Current Anthropology Volume 54, Supplement 8, December 2013

Figure 3. a, Relative frequency of large versus small (!4.5-kg) mammals by MIS (in NISP). b, Relative frequency of large versussmall (!4.5-kg) mammals by MIS (in NISP) with DK1 Layer 8 removed from study.

with the small game component of the HP fauna from Sibuduremains elusive (J. L. Clark, unpublished data). All this beingsaid, the fact that the small ungulates, which do seem to beprimarily accumulated by human activity, and small mammalfauna show the same trends could indicate that humans alsoaccumulated the latter.

Prey Characteristics: Dangerous versus Docile Game

Figure 4 presents data on the relative frequencies of dangerousversus docile prey. Figure 4a shows the proportion of elandversus buffalo; note that the analysis was restricted to MIS

5–3 because the sample size for these taxa in MIS 6 was 1.The relative frequency of buffalo declines from MIS 5 to MIS4 and then subsequently increases in MIS 3. This could beinterpreted as an indication of a reduction in diet breadthduring MIS 4. However, the fact that MIS 5 and 3 have similarproportions of buffalo could also be a reflection of more openenvironments during the interglacial/interstadial periods.

Figure 4b presents the data on the relative proportion of

suids versus similar sized bovids (Size 2). While the sample

size from MIS 6 is quite small ( ), there is a steadyn p 67

increase in suids over time, although their relative frequencyis always low. This pattern could suggest a gradual increase

Clark and Kandel Variation in Middle Stone Age Hunting Strategies S279

Figure 4. Dangerous versus docile prey. a, NISP counts for buffalo (including the extinct Pelorovis antiquus) versus eland by MIS.b, NISP counts for suids versus similar sized bovids (Size 2) by MIS.

in intensification over time; however, these data are difficultto interpret for a variety of reasons. For example, becausesuids are typically less common in coastal settings, we expectthem to occur at different frequencies around the varioussites included in the analysis.

Shellfish Data

Figure 5 shows increasing diversification in shellfish exploi-tation from MIS 6 to MIS 4. This is consistent with an increasein dietary breadth over time. Figure 6a presents data on therelative frequency of ungulates to shellfish. Shellfish are rare

in MIS 6 but predominate in MIS 5 and then drop in MIS

4. As a means of controlling for the unequal sample size in

the data set, we plotted the data from a single site, KRM,

where we found a similar trend (fig. 6b). Given that MIS 5

is an interglacial period when sea level was high and com-

parable with the modern coastline, the prevalence of shellfish

in MIS 5 is to be expected.

The data presented in table 2 regarding the dietary con-

tribution of shellfish deserve particular note. These data

indicate that the amount of shellfish in BBC M3, the largest

assemblage for which shell weights were available, would

have provided subsistence for a maximum of 38 person-

S280 Current Anthropology Volume 54, Supplement 8, December 2013

Figure 5. Distribution of shell groups by MIS and region (MNI).

days (ca. 75,000 kcal), which is the caloric equivalent ofa single bontebok (Damaliscus dorcas). The next largestsample, from YFT1 Lower, would have provided 25 person-days (ca. 49,000 kcal), or the energy equivalent of onebushbuck (Tragelaphus scriptus), one steenbok (Raphiceruscampestris), and one hare (Lepus spp.). In contrast, thecontribution from all 10 layers at PP13B is 1.5 person-days (ca. 3,000 kcal), the equivalent of four Cape dunemole rats. Because the reported samples represent only aportion of the total excavated assemblage, the overall di-etary contribution would have been greater. Nonetheless,these data average hundreds, if not thousands, of years ofshell accumulation, belying the impression that shellfishplayed a major dietary role during this period.

Browser/Grazer Data

Figure 7a presents data on the relative proportion of browsersand grazers over time. Because a majority of the browsersfrom MIS 4 came from a single species from the HP depositsat Sibudu (Philantomba monticola, or blue duiker; n p

), we removed the Sibudu HP assemblage from the anal-1,962ysis; the resulting data are presented in figure 7b. While thetotal sample for MIS 6 is small ( ), in both cases, brows-n p 13

ers dominate the glacial/stadial periods (MIS 6 and 4), withgrazers dominating during the interglacial/interstadial periods(MIS 5 and 3). This suggests more trees and brush duringthe glacial periods and more open environments during theinterglacials. These data fit within our current understandingof paleoenvironmental conditions during the period underconsideration; for example, Chase (2010) utilized multiplelines of evidence to demonstrate the presence of cool andhumid conditions during MIS 4 followed by dramatic arid-ification in MIS 3 (see also Ziegler et al. 2013).

Discussion

When we consider the data presented above, some broadertrends emerge (see fig. 8 for a visual summary of our results).No matter which way we approach the data, MIS 6 preservesthe least evidence of subsistence intensification. Moreover,several lines of evidence are consistent with an expansion ofdietary breadth from MIS 6 to MIS 4. Most notable is theaforementioned similarity in the data relating to the exploi-tation of the smallest (Size 1) ungulates and the small mam-mals, both of which would presumably have been low-rankedgame. The data indicate a slight increase in the exploitationof these species from MIS 6 to MIS 5 with a marked inten-

Clark and Kandel Variation in Middle Stone Age Hunting Strategies S281

Figure 6. a, Frequency of shellfish versus ungulates (shellfish data expressed as MNI, ungulate as NISP) by MIS. b, Frequency ofshellfish versus ungulates (shellfish data expressed as MNI, ungulate as NISP) at KRM.

sification in their collection in MIS 4 and significant declinein the focus on these prey in MIS 3. The increasing diversityin shellfish exploitation also fits within a model of expandingdietary breadth from MIS 6 to MIS 4.

However, not all of the data are consistent with these re-sults. For example, if buffalo are considered to be lower rankedthan eland (given the higher pursuit costs for this aggressivespecies), one might expect to see an increase in the exploi-tation of buffalo from MIS 6 to MIS 4 and a decline in MIS3. While data from MIS 6 are unavailable, we actually foundthe opposite trend, with a decline in the relative frequency ofbuffalo from MIS 5 to MIS 4 and an increase in MIS 3. Along

similar lines, if the exploitation of shellfish were primarily areflection of subsistence intensification, we might expect tosee an increase in the relative frequency of shellfish from MIS5 to MIS 4, and yet, this is not the case. However, in bothcases, the results may ultimately have more to do with en-vironmental change than with subsistence intensification inthat shifts in climate could have affected availability or ac-cessibility of these prey.

Taken as a whole, the data indicate marked variance insubsistence over time. We now consider the potential role ofa number of different causal mechanisms: environment, de-mography, technology, and cognition.

S282 Current Anthropology Volume 54, Supplement 8, December 2013

Figure 7. a, Browser/grazer data by MIS (in NISP). b, Browser/grazer data by MIS (in NISP) with Sibudu HP removed from analysis.

Explaining Variance in the Data: Linkages toEnvironmental Change?

As already discussed, some of the variance in the data seemsto be best explained as a reflection of changes in the envi-ronment. In fact, even the large increase in both the smallestungulates and small mammals in MIS 4 could signify sub-sistence intensification resulting from changes in the land-scape during this glacial period. As such, the changes in thedata may largely reflect shifts in prey availability driven byclimate change. However, we do not believe that environ-mental change can explain the full range of variance; forinstance, if this were the case, one might expect a higherfrequency of small ungulates/mammals in MIS 6, and this isnot the pattern that we see.

Linkages to Demographic Change?

It is notoriously difficult to identify changes in demographyin the MSA. It is possible that the observed expansion ofdietary breadth from MIS 6 to MIS 4 is related to an increasein population pressure caused by either expanding populationsizes or the concentration of existing populations. However,as previously discussed, this pattern could also be related toenvironmental change, and distinguishing between these twoalternatives is not currently possible. One potential way ofaddressing this would be to explore the evidence for variationin shellfish and tortoise size across the MSA, because a declinein the size of these species has been linked to populationgrowth/pressure (e.g., Steele and Klein 2005/2006; Stiner etal. 1999).

Clark and Kandel Variation in Middle Stone Age Hunting Strategies S283

Figure 8. Schematic representation of overall trends in the data.

Linkages to Variation in Hunting Technology?

A lack of agreed-on conventions for assigning lithic materialto specific MSA cultures/techno-complexes has precluded usfrom exploring the relative impact of technological change onthe variance present in our data set in any detail. While newclassification systems for MSA lithics have been proposed(Conard, Porraz, and Wadley 2012; Lombard et al. 2012), itremains to be seen whether researchers will apply these guide-lines. However, it is worth noting that a number of scholarshave utilized data from the SB and HP in an attempt toexplore the potential relationships between lithic and faunalchange, although the results of these studies are often at oddswith each other (e.g., Dusseldorp 2012; Lombard and Clark2008; McCall 2007; Newlander and Clark 2010).

In terms of well-published faunal assemblages subjected todetailed taphonomic analyses, data from the SB are currentlylimited to those from Blombos. The HP is also limited to asingle sample—that from Sibudu. While other sites includedin this analysis do contain SB and HP deposits (i.e., the SBis also present at Sibudu and Diepkloof, while the HP hasbeen identified at Klasies and Diepkloof), full taxonomic and/or taphonomic data are thus far lacking. Given this, we feelthat more specific comparisons of subsistence between thesetwo phases should await the publication of a larger numberof assemblages.

This being said, we would like to make one importantpoint. The increase in small ungulates (and small game moregenerally) that we identified in MIS 4 has often been rec-ognized as a phenomenon specific to the HP. For example,Backwell, d’Errico, and Wadley (2008) argued that the bonepoints and backed tools associated with the HP “may signifypart of a hunting adaptation to small prey in a closed forestenvironment” (1577). However, our data make it apparentthat this pattern cannot be uniquely linked to the technologyof the HP. The DK1 fauna is similarly dominated by smallungulates and small mammals—and despite the recovery of

a large lithic assemblage, no diagnostic HP tools have beenrecovered (Thackeray 2000). This implies that we need tothink more carefully about the functional significance of thetechnology of the HP, at least as far as it can be related tosubsistence behavior. This is further underscored by new datafrom Diepkloof, where the HP spans MIS 5–3, periods inwhich we found a lesser focus on small game (Porraz et al.2013; Tribolo et al. 2013).

Linkages to Variation in Cognition?

We should begin here by admitting our own personal biaseson this point; we assume that all of the populations repre-sented in this analysis likely shared the same cognitive capacityand behavioral capabilities. That said, evaluating cognitiveabilities via hunting strategies is far from straightforward. ThatMSA peoples were capable hunters is no longer questioned(e.g., Clark and Plug 2008; Faith 2008, 2011b; Milo 1998;Thompson and Henshilwood 2011). Although we did see anincrease in the exploitation of one type of dangerous game—suids—over time, we are not convinced by the argument thata greater focus on dangerous game primarily signifies in-creased cognitive abilities (e.g., Klein and Cruz-Uribe 1996).This observation could indicate intensification caused by in-creased population size or reflect improvements in huntingtechnology that would reduce the danger associated with theprocurement of these species. Unfortunately, it is not cur-rently possible to distinguish between these alternatives.

Wadley (2010) argues that the use of remote capture tech-nologies indicates an essentially modern cognition. She pro-posed that the increased focus on taxa from closed environ-ments and the diverse small game assemblage evidenced inthe HP at Sibudu could reflect use of these technologies. Ifone agrees with this hypothesis, then our data suggest thatall MIS 4 populations in southern Africa—even those withoutthe HP—were utilizing remote capture technology and thuspossessed complex cognitive abilities.

Summary: The Implication of Variability in Subsistenceduring the MSA

Although we cannot fully evaluate the influence of techno-logical and demographic change on the observed patterns, itseems clear that no single causal mechanism will adequatelyaccount for the range of variation evidenced in the data. Ofcourse, we are not only interested in understanding the driv-ing forces behind this variability but also in its broader evo-lutionary implications. At present, it would appear that sub-sistence behavior in the MSA was more variable than it wasin the MP (Stiner 2013; but see Delagnes and Rendu 2011).This could be a reflection of a greater degree of behavioralflexibility among MSA populations. And yet, it is interestingto note that we could not identify clear directional trendsassociated with increasing intensification or behavioral com-plexity. In fact, we only identified one example of a continuous

S284 Current Anthropology Volume 54, Supplement 8, December 2013

directional trend, the case of the increasing frequency of suidsover time. The patterns in the faunal data are thus consistentwith those from other classes of material culture. During theMSA, certain innovative technologies and artifacts with sym-bolic significance also seem to come and go rather than show-ing an accretional trend toward increasing complexity, andthis is something the MSA record actually does seem to sharewith the MP (cf. Hovers and Belfer-Cohen 2006, 2013; Lom-bard 2012; Wurz 2013).

Conclusions

Large-scale, specifically directed studies aimed at character-izing and explaining the nature and extent of variability inhuman behavior within the MSA have generally been lackingdespite the fact that a deeper understanding of this variabilityis clearly relevant to larger questions about human behavioralevolution and the rise and spread of AMH. There are a num-ber of reasons for this, both practical and theoretical. How-ever, in light of the recent publication of faunal data from anumber of key MSA sites, we deemed it valuable to conducta large-scale exploration of variation in MSA subsistence be-havior incorporating material from eight sites that span ∼174–40 ka (MIS 6–3).

Because we were interested in exploring evidence for var-iation in diet breadth over time, we largely approached thedata from the perspective of prey choice models under whichsmaller game is considered lower-ranked prey. Thus, an in-crease in the exploitation of these taxa should indicate abroadening of the diet. Notably, multiple strands of evidencesuggest an increase in dietary breadth that peaked in MIS 4;not only was there a marked increase in the smallest ungulates(Size 1), but also there was also an equally large rise in theexploitation of small mammals (!4.5 kg). Both of these typesof prey declined noticeably in MIS 3. The diversity of shellfishalso peaked in MIS 4, which also indicates increasing dietarybreadth. Interestingly, although many scholars emphasize theimportance of shellfish exploitation in the MSA, the availabledata suggest that even the largest MSA shell accumulationsreflect a small contribution to the overall diet.

Explaining the variance in the data is not necessarily astraightforward task. Some dimensions of variability do seemto be driven primarily by environmental change; for example,the increased frequency in shellfish in MIS 5 (relative to MIS6 and 4) makes sense given that sea levels would have beenhighest during this period. Furthermore, MIS 5 and 3 showa similar (and high) frequency of grazers, particularly com-pared with MIS 6 and 4, which show a predominance ofbrowsers. This suggests greater shrub and tree cover duringthe glacial periods and more open environments in thewarmer phases, which is consistent with our broader under-standing of paleoenvironmental conditions in the region (e.g.,Chase 2010; Ziegler et al. 2013).

What remains unclear is the reason behind the spike in thesmallest ungulates and small mammals during MIS 4. This

may be related to environmental change, or it could reflectintensification driven by population growth, although at thecurrent time we cannot effectively reconstruct populationsizes during the different stages of the MSA. Interestingly, thisspike in smaller fauna has been linked to the particular tech-nologies of the HP. However, given that our sample includesdata from DK1, a site occupied at the same time as the HPbut lacking its characteristic technology (Thackeray 2000), weargue that technology is not the force driving this pattern.

While preparing this paper, we identified several factorsthat will be critical to the success of future studies of this sort.First, more detailed taphonomic analyses are needed. Thisparticularly applies to the smaller fauna, as the accumulatingagents of these taxa are often unclear. New work by Armstrong(2013) on the DK1 fauna is a step in the right direction.Second, data from more sites will be necessary to ascertainhow robust the identified patterns are. For example, it willbe interesting to see whether MIS 6 will consistently showthe least evidence for the exploitation of low-ranked prey.Third, finer grained analyses will require better dating of ex-isting collections. Finally, we think that a consideration of theeffect of technological change on human subsistence behaviorcould prove very insightful in understanding variation in thefaunal data; however, this type of analysis will not truly bepossible until the development of a more standardized systemfor defining archaeological cultures in the MSA. Conard, Por-raz, and Wadley (2012), Lombard et al. (2012), and Porrazet al. (2013) have made significant steps toward this goal.

Although we may not yet fully understand the causes driv-ing the variability identified in the faunal record, we feel thatwe have made an excellent start in documenting the natureand extent of variation in subsistence behavior during theMSA. In fact, the degree of variance in the data exceeded ourexpectations. Our results indicate that subsistence behaviorduring the MSA was more highly variable than that in theMP, although the broader evolutionary implications of thisare not immediately apparent.

Acknowledgments

J. L. Clark would like to thank Erella Hovers and Steve Kuhnfor their invitation to participate in the symposium on whichthis issue of Current Anthropology is based. Thanks to RichardKlein for providing unpublished data from Klasies River Cave1, James Brink for providing PDFs on Florisbad, Leslie Aielloand Laurie Obbink at Wenner-Gren for their advice and as-sistance, and Teresa Steele and an anonymous reviewer fortheir helpful comments. We also wish to thank Aaron Arm-strong, Shaw Badenhorst, Genevieve Dewar, Tyler Faith, andJarod Hutson for providing supplementary information. Re-search by J. L. Clark was supported by an Alexander vonHumboldt Foundation postdoctoral research fellowship, theLeakey Foundation, and the Palaeontological Scientific Trust.

Clark and Kandel Variation in Middle Stone Age Hunting Strategies S285

References CitedAdler, Daniel S., and Guy Bar-Oz. 2009. Seasonal patterns of prey acquisition

and inter-group competition during the Middle and Upper Paleolithic ofthe southern Caucasus. In The evolution of hominin diets: integrating ap-proaches to the study of Palaeolithic subsistence. Jean-Jacques Hublin andMichael P. Richards, eds. Pp. 127–140. Vertebrate Paleobiology and Pa-leoanthropology Series. Dordrecht: Springer.

Adler, Daniel. S., Guy Bar-Oz, Anna Belfer-Cohen, and Ofer Bar-Yosef. 2006.Ahead of the game: Middle and Upper Paleolithic hunting behaviors in thesouthern Caucasus. Current Anthropology 47:89–118.

Armstrong, Aaron. 2013. Small mammal utilization by Middle Stone Agehumans at Die Kelders Cave 1, Western Cape, South Africa: Paleoanthro-pology Society meeting abstracts, Honolulu, HI, 2–3 April 2012. Paleoan-thropology 2013:A2.

Avery, Graham, David Halkett, Jayson Orton, Teresa Steele, Madelon Tusenius,and Richard G. Klein. 2008. The Ysterfontein 1 Middle Stone Age rockshelter and the evolution of coastal foraging. South African ArchaeologicalSociety Goodwin Series 10:66–89.

Backwell, Lucinda, Franceso d’Errico, and Lyn Wadley. 2008. Middle StoneAge bone tools from the Howiesons Poort layers, Sibudu Cave, South Africa.Journal of Archaeological Science 35:1566–1580.

Badenhorst, Shaw, and Ina Plug. 2012. The faunal remains from the MiddleStone Age levels of Bushman Rock Shelter in South Africa. South AfricanArchaeological Bulletin 67:16–31.

Binford, Lewis R. 1978. Nunamiut ethnoarchaeology. New York: AcademicPress.

Bird, Douglas W., and James F. O’Connell. 2006. Behavioral ecology andarchaeology. Journal of Archaeological Research 14:143–188.

Brain, C. K. 1974. Some suggested procedures in the analysis of bone accu-mulations from southern African Quaternary sites. Annals of the TransvaalMuseum 29:1–7.

Brink, James S. 1987. The archaeozoology of Florisbad, Orange Free State. Mem-oirs van die Nasionale Museum (Bloemfontein) no. 24. Bloemfontein, SouthAfrica: National Museum.

———. 1988. The taphonomy and palaeoecology of the Florisbad springfauna. Paleoecology of Africa 19:169–179.

Brink, James S., and Hilary J. Deacon. 1982. A study of a last interglacial shellmidden and bone accumulation at Herolds Bay, Cape Province, SouthAfrica. In Palaeoecology of Africa: Southern African Society for QuaternaryResearch, Proceedings of the VIth Biennial Conference held at the TransvaalMuseum, Pretoria, 26–29 May 1981, vol. 15. J. C. Vogel, E. A. Voigt, T. C.Partridge, E. M. v. Z. Bakker, and J. A. K. Coetzee, eds. Pp. 31–39. Rot-terdam: Balkema.

Broughton, Jack M., and Donald K. Grayson. 1993. Diet breadth, adaptivechange and the White Mountain fauna. Journal of Archaeological Science20:331–336.

Buchanan, W. F. 1988. Shellfish in prehistoric diet, Elands Bay, S.W. Cape Coast,South Africa. British Archaeological Reports International Series 455. Ox-ford: BAR.

Chase, Brian M. 2010. South African paleoenvironments during marine ox-ygen isotope stage 4: a context for the Howiesons Poort and Still Bayindustries. Journal of Archaeological Science 37:1359–1366.

Clark, Jamie L. 2009. Testing models on the emergence and nature of modernhuman behavior: Middle Stone Age fauna from Sibudu Cave (South Africa).PhD dissertation, University of Michigan, Ann Arbor.

———. 2011. The evolution of human culture during the later Pleistocene:using fauna to test models on the emergence and nature of “modern”human behavior. Journal of Anthropological Archaeology 30:273–291.

Clark, Jamie L., and Ina Plug. 2008. Animal exploitation strategies during theSouth African Middle Stone Age: Howiesons Poort and post–HowiesonsPoort fauna from Sibudu Cave. Journal of Human Evolution 54:886–898.

Clark, J. D. 1993. African and Asian perspectives on the origins of modernhumans. In The origin of modern humans and the impact of chronometricdating. M. J. Aitken, Curtis B. Stringer, and Paul A. Mellars, eds. Pp. 148–178. Princeton, NJ: Princeton University Press.

Conard, Nicholas J., Guillaume Porraz, and Lyn Wadley. 2012 What is in aname? characterizing the “post-HP” at Sibudu. South African ArchaeologicalBulletin 67:180–199.

Delagnes, Anne, and William Rendu. 2011. Shifts in Neandertal mobility,technology and subsistence strategies in western France. Journal of Archae-ological Science 38:1771–1783.

d’Errico, Franceso, Renata Garcia Moreno, and Riaan F. Rifkin. 2012. Tech-

nological, elemental and colorimetric analysis of an engraved ochre frag-ment from the Middle Stone Age levels of Klasies River Cave 1, SouthAfrica. Journal of Archaeological Science 39:942–952.

de Ruiter, Darryl J., Juliet K. Brophy, Patrick J. Lewis, Steven E. Churchill,and Lee R. Berger. 2008. Faunal assemblage composition and paleoenvi-ronment of Plovers Lake, a Middle Stone Age locality in Gauteng Province,South Africa. Journal of Human Evolution 55:1102–1117.

Dusseldorp, Gerrit L. 2012. Tracking the influence of technological changeon Middle Stone Age hunting strategies in South Africa. Quaternary In-ternational 270:70–79.

Faith, J. Tyler. 2008. Eland, buffalo, and wild pigs: were Middle Stone Agehumans ineffective hunters? Journal of Human Evolution 55:24–36.

———. 2011a. Late Quaternary megafaunal extinctions in southern Africa’sCape Floral Region. PhD dissertation, George Washington University, Wash-ington, DC.

———. 2011b. Ungulate biogeography, statistical methods, and the profi-ciency of Middle Stone Age hunters. Journal of Human Evolution 60:315–317.

Feathers, James K. 2002. Luminescence dating in less than ideal conditions:case studies from Klasies River Main Site and Duinefontein, South Africa.Journal of Archaeological Science 29:177–194.

Feathers, James K., and Daniel A. Bush. 2000. Luminescence dating of MiddleStone Age deposits at Die Kelders. Journal of Human Evolution 38:91–119.

Gagnon, Mario, and Amy E. Chew. 2000. Dietary preferences in extant AfricanBovidae. Journal of Mammalogy 81:490–511.

Gaudzinski, Sabine. 2006. Monospecific or species-dominated faunal assem-blages during the Middle Paleolithic in Europe. In Transitions before thetransition: evolution and stability in the Middle Paleolithic and Middle StoneAge. Erella Hovers and Steven L. Kuhn, eds. Pp. 137–147. New York:Springer.

Gaudzinski-Windheuser, Sabine, and Laura Niven. 2009. Hominin subsistencepatterns during the Middle and Late Paleolithic in northwestern Europe.In The evolution of hominin diets: integrating approaches to the study ofPalaeolithic subsistence. Jean-Jacques Hublin and Michael P. Richards, eds.Pp. 99–111. Vertebrate Paleobiology and Paleoanthropology Series. Dor-drecht: Springer.

Hawkes, Kristen, Kim Hill, and James F. O’Connell. 1982. Why hunters gather:optimal foraging and the Ache of eastern Paraguay. American Ethnologist9:379–398.

Henshilwood, Curtis S., Francesco d’Errico, Karen L. van Niekerk, Yvan Co-quinot, Zenobia Jacobs, Stein-Erik Lauritzen, Michel Menu, and RenataGarcıa-Moreno. 2011. A 100,000-year-old ochre-processing workshop atBlombos Cave, South Africa. Science 334:219–222.

Henshilwood, Curtis S., Judith C. Sealy, Royden Yates, Kathryn Cruz-Uribe,Paul Goldberg, Frederick E. Grine, Richard G. Klein, Cedric Poggenpoel,Karen Van Niekerk, and Ian Watts. 2001. Blombos Cave, Southern Cape,South Africa: Preliminary report on the 1992–1999 excavations of the Mid-dle Stone Age levels. Journal of Archaeological Science 28:421–448.

Hovers, Erella, and Anna Belfer-Cohen. 2006. “Now you see it, now youdon’t”: modern human behavior in the Middle Paleolithic. In Transitionsbefore the transition: evolution and stability in the Middle Paleolithic andMiddle Stone Age. Erella Hovers and Steven L. Kuhn, eds. Pp. 295–304.New York: Springer.

———. 2013. On variability and complexity: lessons from the LevantineMiddle Paleolithic record. Current Anthropology 54(suppl. 8):S337–S357.

Hutson, Jarod M. 2012. A comparative study of life and death at Middle StoneAge open-air sites within the southern African interior. PhD dissertation,University of Nevada, Reno.

Hutson, Jarod M., and Chester R. Cain. 2008. Reanalysis and reinterpretationof the Kalkbank faunal accumulation, Limpopo Province, South Africa.Journal of Taphonomy 6:399–428.

Jacobs, Zenobia, Richard G. Roberts, Rex F. Galbraith, Hilary J. Deacon, RainerGrun, Alex Mackay, Peter Mitchell, Ralf Vogelsang, and Lyn Wadley. 2008a.Ages for the Middle Stone Age of southern Africa: implications for humanbehavior and dispersal. Science 322:733–735.

Jacobs, Zenobia, Ann G. Wintle, Geoffrey A. T. Duller, Richard G. Roberts,and Lyn Wadley. 2008b. New ages for the post–Howiesons Poort, late andfinal Middle Stone Age at Sibudu Cave, South Africa. Journal of Archaeo-logical Science 35:1790–1807.

Jerardino, Antonieta, and Curtis W. Marean. 2010. Shellfish gathering, marinepaleoecology and modern human behavior: perspectives from cave PP13B,Pinnacle Point, South Africa. Journal of Human Evolution 59:412–424.

Kaplan, Hillard, and Kim Hill. 1992. The evolutionary ecology of food ac-

S286 Current Anthropology Volume 54, Supplement 8, December 2013

quisition. In Evolutionary ecology and human behavior. E. A. Smith and B.Winterhalder, eds. Pp. 167–201. Hawthorne, NY: Aldine.

Kelly, Robert L. 1995. The foraging spectrum: diversity in hunter-gatherer life-ways. Washington, DC: Smithsonian Institution.

Kiberd, Philip. 2006. Bundu Farm: a report on archaeological and palaeoen-vironmental assemblages from a pan site in Bushmanland, Northern Cape,South Africa. South African Archaeological Bulletin 61:189–201.

Klein, Richard G. 1976. The mammalian fauna of the Klasies River Mouthsites, southern Cape Province, South Africa. South African ArchaeologicalBulletin 31:75–98.

———. 1977. The mammalian fauna from the Middle and Later Stone Age(Later Pleistocene) levels of Border Cave, Natal Province, South Africa.South African Archaeological Bulletin 32:14–27.

———. 1979. Environmental and ecological implications of large mammalsfrom Upper Pleistocene and Holocene sites in southern Africa. Annals ofthe South African Museum 81:223–283.

———. 2001. Southern Africa and modern human origins. Journal of An-thropological Research 57:1–16.

———. 2009. The human career: human biological and cultural origins. 3rdedition. Chicago: University of Chicago Press.

Klein, Richard G., and Kathryn Cruz-Uribe. 1996. Exploitation of large bovidsand seals at Middle and Later Stone Age sites in South Africa. Journal ofHuman Evolution 31:315–334.

———. 2000. Middle and Later Stone Age large mammal and tortoise remainsfrom Die Kelders, Cave 1, Western Cape Province, South Africa. Journal ofHuman Evolution 38:169–195.

Klein, Richard G., Kathryn Cruz-Uribe, and Peter B. Beaumont. 1991. En-vironmental, ecological, and paleoanthropological implications of the latePleistocene mammalian fauna from Equus Cave, Northern Cape Province,South Africa. Quaternary Research 36:94–119.

Kuhn, Steven L., and Erella Hovers. 2006. General introduction. In Transitionsbefore the transition: evolution and stability in the Middle Paleolithic andMiddle Stone Age. Erella Hovers and Steven L. Kuhn, eds. Pp. 1–11. NewYork: Springer.

Kuman, Kathleen, and R. J. Clarke. 1986. Florisbad: new investigations at aMiddle Stone Age hominid site in South Africa. Geoarchaeology 1:103–125.

Kuman, Kathleen, Moshe Inbar, and R. J. Clarke. 1999. Palaeoenvironmentsand cultural sequence of the Florisbad Middle Stone Age hominid site,South Africa. Journal of Archaeological Science 26:1409–1425.

Langejans, Geeske H. J., Karen L. van Niekerk, Gerrit L. Dusseldorp, and J.Francis Thackeray. 2012. Middle Stone Age shellfish exploitation: potentialindications for mass collecting and resource intensification at Blombos Caveand Klasies River, South Africa. Quaternary International 270:80–94.

Lombard, Marlize. 2012. Thinking through the Middle Stone Age of sub-Saharan Africa. Quaternary International 270:140–155.

Lombard, Marlize, and Jamie L. Clark. 2008. Variability and change in MiddleStone Age hunting behavior: aspects from the lithic and faunal records. InAnimals and people: archaeozoological papers in honour of Ina Plug. ShawBadenhorst, Peter Mitchell, and Jonathan C. Driver, eds. Pp. 46–56. BritishArchaeological Reports International Series 1849. Oxford: Archaeopress.

Lombard, Marlize, Lyn Wadley, Janette Deacon, Sarah Wurz, Isabelle Parsons,Moleboheng Mohapi, Joane Swart, and Peter Mitchell. 2012. South Africanand Lesotho Stone Age sequence updated (I). South African ArchaeologicalBulletin 67:123–144.

Louw, A. W. 1969. Bushman Rock Shelter, Ohrigstad, Eastern Transvaal: apreliminary investigation, 1965. South African Archaeological Bulletin 24:39–51.

Lupo, Karen D. 2007. Evolutionary foraging models in zooarchaeological anal-ysis: recent applications and future challenges. Journal of ArchaeologicalResearch 15:143–189.

Lupo, Karen D., and Dave N. Schmitt. 2005. Small prey hunting technologyand zooarchaeological measures of taxonomic diversity and abundance:ethnoarchaeological evidence from central African forest foragers. Journalof Anthropological Archaeology 24:335–353.

Marean, Curis W., Yoshiko Abe, Carol J. Frey, and Robert C. Randall. 2000.Zooarchaeological and taphonomic analysis of the Die Kelders Cave 1 layers10 and 11 Middle Stone Age larger mammal fauna. Journal of HumanEvolution 38:197–233.

Marean, Curtis W., Miryam Bar-Matthews, Jocelyn Bernatchez, Erich Fisher,Paul Goldberg, Andy I. R. Herries, Zenobia Jacobs, et al. 2007. Early humanuse of marine resources and pigment in South Africa during the MiddlePleistocene. Nature 449:905–908.

Marean, Curtis W., Miryam Bar-Matthews, Erich Fisher, Paul Goldberg, Andy

Herries, Panagiotis Karkanas, Peter J. Nilssen, and Erin Thompson. 2010.The stratigraphy of the Middle Stone Age sediments at Pinnacle Point Cave13B (Mossel Bay, Western Cape Province, South Africa). Journal of HumanEvolution 59:234–255.

McBrearty, Sally, and Alison S. Brooks. 2000. The revolution that wasn’t: anew interpretation of the origin of modern human behavior. Journal ofHuman Evolution 39:453–563.

McCall, Grant S. 2007. Behavioral ecological models of lithic technologicalchange during the later Middle Stone Age of South Africa. Journal of Ar-chaeological Science 34:1738–1751.

Milo, Richard G. 1994. Human-animal interactions in southern African pre-history: a microscopic study of bone damage signatures. PhD dissertation,University of Chicago.

———. 1998. Evidence for hominid predation at Klasies River Mouth, SouthAfrica, and its implication for the behavior of early modern humans. Journalof Archaeological Science 25:99–133.

Munro, Natalie D., and Guy Bar-Oz. 2005. Gazelle bone fat processing in theLevantine Epipaleolithic. Journal of Archaeological Science 32:223–239.

Newlander, Khori, and Jamie L. Clark. 2010. One size doesn’t fit all? modelingtechnological change in the Middle Stone Age of southern Africa: Paleo-anthropology Society meeting abstracts, St. Louis, MO, 13–14 April 2010.Paleoanthropology 2010:A23.

Opperman, H. 1996. Strathalan Cave B, Northeastern Cape Province, SouthAfrica: evidence for human behaviour 29,000–26,000 years ago. QuaternaryInternational 33:45–53.

Plug, Ina. 1981. Some research results on the Late Pleistocene and EarlyHolocene deposits of Bushman Rock Shelter, Eastern Transvaal. South Af-rican Archaeological Bulletin 36:14–21.

———. 2004. Resource exploitation: animal use during the Middle Stone Ageat Sibudu Cave, KwaZulu-Natal. South African Journal of Science 100:151–158.

Pollarolo, Luca, J. Wilkins, K. Kuman, and L. Galletti. 2010. Site formationat Kudu Koppie: a late Earlier and Middle Stone Age site in northernLimpopo Province, South Africa. Quaternary International 216:151–161.

Porraz, Guillaume, Pierre-Jean Texier, Will Archer, Michel Piboule, Jean-Phi-lippe Rigaud, and Chantal Tribolo. 2013. Technological successions in theMiddle Stone Age sequence of Diepkloof Rock Shelter, Western Cape, SouthAfrica. Journal of Archaeological Science 40:3376–3400.

Rector, Amy L., and Kaye E. Reed. 2010. Middle and late Pleistocene faunasof Pinnacle Point and their paleoecological implications. Journal of HumanEvolution 59:340–357.

Shea, John J. 2011. Homo sapiens is as Homo sapiens was: behavioral variabilityvs. “behavioral modernity” in Paleolithic archaeology. Current Anthropology52:1–35.

Singer, Ronald, and John Wymer. 1982. The Middle Stone Age at Klasies RiverMouth in South Africa. Chicago: University of Chicago Press.

Skinner, John D., and Christian T. Chimimba. 2005. The mammals of thesouthern African sub-region. 3rd edition. Singapore: Cambridge UniversityPress.

Speth, John D., and Jamie L. Clark. 2006. Hunting and overhunting in theLevantine late Middle Paleolithic. Before Farming 3:1–42.

Steele, Teresa E., and Richard G. Klein. 2005/2006. Mollusk and tortoise sizeas proxies for Stone Age population density in South Africa: implicationsfor the evolution of human cultural capacity. Munibe 57:221–237.

———. 2009. Late Pleistocene subsistence strategies and resource intensifi-cation in Africa. In The evolution of hominin diets: integrating approaches tothe study of Palaeolithic subsistence. Jean-Jacques Hublin and Michael P.Richards, eds. Pp. 113–126. Vertebrate Paleobiology and Paleoanthropology.Dordrecht: Springer.

———. 2013. The Middle and Later Stone Age faunal remains from DiepkloofRock Shelter, Western Cape, South Africa. Journal of Archaeological Science40:3453–3462.

Stiner, Mary C. 2002. Carnivory, coevolution, and the geographic spread ofthe genus Homo. Journal of Archaeological Research 10:1–63.

———. 2006. Middle Paleolithic subsistence economy in the Mediterraneanregion. In Transitions before the transition: evolution and stability in theMiddle Paleolithic and Middle Stone Age. Erella Hovers and Steven L. Kuhn,eds. Pp. 213–231. New York: Springer.

———. 2013. An unshakable Middle Paleolithic? trends versus conservatismin the predatory niche and their social ramifications. Current Anthropology54(suppl. 8):S288–S304.

Stiner, Mary C., and Steven. L. Kuhn. 1992. Subsistence, technology, and

Clark and Kandel Variation in Middle Stone Age Hunting Strategies S287

adaptive variation in Middle Paleolithic Italy. American Anthropologist 94:306–339.

———. 2006. Changes in the “connectedness” and resilience of Paleolithicsocieties in Mediterranean ecosystems. Human Ecology 34:693–712.

Stiner, Mary C., and Natalie D. Munro. 2002. Approaches to prehistoric dietbreadth, demography, and prey ranking systems in time and space. Journalof Archaeological Method and Theory 9:181–214.

Stiner, Mary C., Natalie D. Munro, and Todd A. Surovell. 2000. The tortoiseand the hare: small-game use, the broad spectrum revolution, and Paleo-lithic demography. Current Anthropology 41:39–73.

Stiner, Mary C., Natalie D. Munro, Todd A. Surovell, Eitan Tchernov, andOfer Bar-Yosef. 1999. Paleolithic population growth pulses evidenced bysmall animal exploitation. Science 283:190–194.

Thackeray, Anne I. 2000. Middle Stone Age artefacts from the 1993 and 1995excavations of Die Kelders Cave 1, South Africa. Journal of Human Evolution43:15–26.

Thackeray, J. Francis. 1979. An analysis of faunal remains from archaeologicalsites in southern southwest Africa (Namibia). South African ArchaeologicalBulletin 34:18–33.

Thompson, Jessica C. 2008. Zooarchaeological tests for modern human be-havior at Blombos Cave and Pinnacle Point Cave 13B, southwestern Cape,South Africa. PhD dissertation, Arizona State University, Tempe.

———. 2010a. Taphonomic analysis of the Middle Stone Age faunal assem-blage from Pinnacle Point Cave 13B, Western Cape, South Africa. Journalof Human Evolution 59:321–339.

———. 2010b. Variability in Middle Stone Age faunal exploitation and useof the physical and social landscapes in the southwestern Cape, South Africa.In Settlement dynamics of the Middle Paleolithic and Middle Stone Age, vol.3. Nicholas J. Conard and Anne Delagnes, eds. Pp. 11–38. Tubingen: Kerns.

Thompson, Jessica C., and Christopher S. Henshilwood. 2011. Taphonomicanalysis of the Middle Stone Age larger mammal faunal assemblage fromBlombos Cave, southern Cape, South Africa. Journal of Human Evolution60:746–767.

Tribolo, Chantal, Norbert Mercier, Eric Douville, Jean-Louis Joron, Jean-LouisReyss, Daniel Rufer, Nadia Cantin, et al. 2013. OSL and TL dating of theMiddle Stone Age sequence at Diepkloof Rock Shelter (South Africa): aclarification. Journal of Archaeological Science 40:3401–3411.

van Pletzen, Liezl. 2000. The large mammal fauna from Klasies River. MAdissertation, University of Stellenbosch.

Vogelsang, Ralf, Jurgen Richter, Zenobia Jacobs, Barbara Eichhorn, VeerleLinseele, and Richard G. Roberts. 2010. New excavations of Middle StoneAge deposits at Apollo 11 rockshelter, Namibia: stratigraphy, archaeology,chronology and past environments. Journal of African Archaeology 8:185–218.

Wadley, Lyn. 2010. Were snares and traps used in the Middle Stone Age anddoes it matter? a review and case study from Sibudu, South Africa. Journalof Human Evolution 58:179–192.

Wadley, Lyn, Ina Plug, and Jamie L. Clark. 2008. The contribution of Sibudu’sfauna to an understanding of KwaZulu-Natal environments at 60, 50 and37 ka. In Animals and people: archaeozoological papers in honour of Ina Plug.Shaw Badenhorst, Peter Mitchell, and Jonathan C. Driver, eds. Pp. 34–45.British Archaeological Reports International Series 1849. Oxford: Archaeo-press.

Weaver, Timothy D., Teresa E. Steele, and Richard G. Klein. 2011. The abun-dance of eland, buffalo, and wild pigs in Middle and Later Stone Age sites.Journal of Human Evolution 60:309–314.

Wurz, Sarah. 2002. Variability in the Middle Stone Age lithic sequence,115,000–60,000 years ago at Klasies River, South Africa. Journal of Archae-ological Science 29:1001–1015.

———. 2013. Technological trends in the Middle Stone Age of South Africabetween MIS 7 and MIS 3. Current Anthropology 54(suppl. 8):S305–S319.

Ziegler, Martin, Margit H. Simon, Ian R. Hall, Stephen Barker, Chris Stringer,and Rainer Zahn. 2013. Development of Middle Stone Age innovationlinked to rapid climate change. Nature Communications 4: doi:10.1038/ncomms2897.

S288 Current Anthropology Volume 54, Supplement 8, December 2013

� 2013 by The Wenner-Gren Foundation for Anthropological Research. All rights reserved. 0011-3204/2013/54S8-0010$10.00. DOI: 10.1086/673285

An Unshakable Middle Paleolithic?Trends versus Conservatism in the Predatory Niche

and Their Social Ramifications

by Mary C. Stiner

The great temporal and geographic span of the Middle Paleolithic (MP) raises many questions about behavioralvariation within this period and its evolutionary significance. This paper focuses on MP predator economics andits social ramifications by examining the data for possible trends in the size of the hominin ecological footprint,hunting practices, trophic level, food sharing, and the intensity with which sites were occupied. Middle Paleolithichominins were big game hunters, and they were rather specialized in their focus on ungulate prey. Low-cost gatherablesmall prey were a perennial if minor contribution to MP diets at lower latitudes, but the overall breadth of themeat diet remained narrow throughout the period. Discernible trends in the MP are few. Foraging innovations ofthe MP include marine shellfish exploitation by 120,000 years BP (possibly earlier), galvanization of the prime-ageungulate hunting niche, and hearth-centered domestic camps. The density of zooarchaeological material seems toincrease during the last 30,000 years of MP existence, implying mild increases in human populations. Importantaspects of carcass processing and meat sharing in the MP do not show much variation but do indicate closecooperation and habitual sharing among group members. Contrasts to late Lower Paleolithic butchery patterns mayilluminate more formal patterns of meat sharing in the MP and after. The seeming rigidity of MP hunting economicscould have been the secret to its widespread success for 200,000 years.

Introduction

Variation in Middle Paleolithic (MP) behavior and ecologycan be explored at many temporal and geographic scales, fromlocal to the quasi-global. Temporal variation is seen in MPsite densities, for example, and much geographic variationcan be found in the prey species hunted. One may argue,nonetheless, that the MP is more clearly defined by devel-opments with the outset and close of this long culture periodthan by strong trends within it. Of course all arguments aboutthe MP come down to how much we choose to make of thevariation we find, and usually this is accomplished by com-parisons with recent human behavior. Though useful to apoint, modern comparators are something of an Achilles’ heelin evolutionary interpretation because hindsight has nothingto do with the evolutionary processes that gave rise to or wereimportant during the MP. If compared with the Lower Pa-leolithic, the MP was a time of innovation. If compared withthe Upper Paleolithic (UP), the MP was arguably a time ofstability and inflexibility. The hallowed standard of early“modern human behavior” is only particularly relevant where

Mary C. Stiner is Professor in the School of Anthropology at theUniversity of Arizona (P.O. Box 210030, Tucson, Arizona 85721-0030,U.S.A. [mstiner@email.arizona.edu]). This paper was submitted 3VII 13, accepted 24 VII 13, and electronically published 29 X 13.

and when very late MP populations first came in contact withAfrican Middle Stone Age (MSA) or UP populations.

This review of the MP emphasizes the archaeological recordof animal exploitation. While predation is but one way ofexamining past hominin behavior, predation is one of themost powerful interactive forces in the evolution of life. Therecan be no question that predation habits define a number ofimportant themes in the human evolution story not simplywith respect to food energetics and ecology but also to thebases of social cooperation and demography. I have two goalsin writing this essay. The first is to document changes inhunting strategies and dietary breadth within the Mediter-ranean MP and between the MP and periods just before andafter it. One cannot hope to cover the full volume of workthat has been done on MP economics in Eurasia. I will insteadfocus mainly on the archaeological record of the Mediterra-nean Basin while bringing information from other areas tothe discussion as much as possible. The Mediterranean regionis a key testing ground for studies of variation in homininsubsistence because of the great variety of edible foods it offers(Blondel and Aronson 1999) and the high density of archae-ological research. I will consider whether the variation ob-served in the faunal record is better explained by environ-mental fluctuations or by social, demographic, or cognitivechanges. The second goal of this essay is to reconcile the scopeof the observed faunal trends with the temporal and geo-

Stiner Trends versus Conservatism in the Predatory Niche S289

graphic breadth of the MP culture period as a whole. Beforethe invasion of anatomically modern humans (AMH) intotheir range, MP populations experienced selective conditionsthat were somewhat distinct from those that confronted themwhen new, competing populations were present. The value ofpredatory specialization, for example, and the potential forpopulation growth would almost certainly have changed inthe face of interspecific competition.

I will also explore the position that the differences betweenthe MP and the UP or the MSA need not have been aboutdifferences in native cognitive ability. Cognitive differencesmight well have existed between MP and MSA or UP pop-ulations, but the evidence we have in hand does not reallypoint to this. Alternative explanations for the differing evo-lutionary trajectories of the MP, MSA, and UP may lie in thedemographic potential and stability of these Late Pleistocenepopulations and, intimately linked to these population char-acteristics, the longevity and stability of knowledge-sharinginstitutions (sensu Henrich 2004; Poteete and Ostrom 2004).Physiology does not predict cultural differences or intelligenceamong recent humans, and the same may have been trueamong Pleistocene MP and MSA populations. Lest we forget,MP artifacts, or artifacts very much like them, were madeand used by both Neanderthals and earliest AMHs in theLevant (Bar-Yosef et al. 1986; Vandermeersch 1989). Consid-erable planning capability is indicated for the MP from themany pathways by which flake or blade blank forms wereproduced (cf. Boeda 1994; Bourguignon, Faivre, and Turq2004; Delagnes and Meignen 2006; Kuhn 1995). In fact, MPlithic reduction systems tend to be a good deal more com-plex—and the outcomes more elegant—than most MSA sys-tems for working stone. Zooarchaeological data meanwhileindicate that MP hunting practices relied on deep environ-mental knowledge, forethought, and close cooperation amonggroup members.

Any attempt to review variation in MP subsistence is facedwith patchy evidence. More problematic, however, is the factthat possible cultural continuity or gradients between Eurasiaand North Africa are underexplored. We should in principleexpect evidence of behavioral variation to increase as the areaof a paleoculture’s distribution increases merely because ofthe likelihood of cultural drift. What we bump up againstinstead is an academic legacy of imposed geographic andpaleoculture boundaries that undermines attempts to look atreal variation. If we assume that the MP simply did not existin Africa, even in areas nearest to the Mediterranean Basin,then much of the potential variability is eclipsed by an ar-bitrary segregation of archaeological records at the scale ofcontinents. If, on the other hand, paleocultures such as theAterian of North Africa can be included in the MP, then thevariation embodied by the MP paleoculture will seem fargreater.

Let us assume for the sake of this discussion that the MPwas confined strictly to Europe and the whole of western Asia.This is still a huge area and one that encompasses tremendous

ecological variation because of the latitudinal gradient, diversetopographies, and complex maritime-continental interfaces.Undaunted by variable environments, MP hominins thrivedin Eurasia for about 200,000 years. They were expert huntersof large game animals wherever they lived, and they movedlarge quantities of food and tool stone around to suit theirneeds. The large game species that they hunted followed re-gional variation in animal community composition (cf.Grayson and Delpech 1998; Griggo 2004; Munzel and Conard2004; Patou-Mathis 2000; Speth and Tchernov 2001; Stark-ovich 2011; Stiner 1994, 2005; Valensi and Psathi 2004). Inthis regard, MP hunters appear to have been very flexible.They were much less flexible in their use of small animals,in strong contrast to recent hunter-gatherers, UP foragers,and possibly some MSA foragers. From a modernist per-spective, it seems odd that MP hominins did not expand theirmeat diet much at all in response to variation in biotic di-versity over the more than 200,000 years of their existence inEurasia.

One need not be a MP apologist, on the other hand, tofind that opinion currently is stacked against recognizing“modernity” of any sort in the MP. There is no basis forclaiming that MP populations were inherently less imaginativein solving problems. They imagined reasons to bury deceasedgroup members (e.g., Bar-Yosef et al. 1988, 1992; Haydal etal. 1995); they turned to the sea for new foods by at least120,000 years ago (Stiner 1994); they collected uniquely at-tractive objects such as rock crystals and pretty shells (e.g.,Zilhao et al. 2010); they used marine sponges (Stiner 1994)and mineral pigments (Rendu 2010; Roebroeks et al. 2012);they hafted stone artifacts to organic handles (e.g., Boeda etal. 1996; Koller, Baumer, and Mania 2001; Shea 1989); theymade scraping tools from marine shells (Darlas 2007:360–361; Stiner 1994:187–188; Vitagliano 1984); they used feathersfrom large birds (Peresani et al. 2011); and they invented orrefined many distinct techniques for regulating blank formin stone working. Important findings on the MSA from theAfrican continent, such as very early shell ornaments and bonetools (e.g., Bouzouggar et al. 2007; d’Errico et al. 2005; Hen-shilwood et al. 2004), have primed us to expect precociousbehavioral patterns in this part of the world more generally.In the interest of healthy debate, why do seemingly innovativeacts in the MP hardly register in the story of “revolutionary”human developments?

The elephant in the room is the apparently capacious brainsize of the Neanderthals. Scaled to body mass, AMH andrecent humans have nothing on MP Neanderthals in thisregard (Klein 1999; Rightmire 2003). Argue as we may fordifferent wiring within the brain, it could all be special plead-ing given the equally great metabolic burden of Eurasian MP,African MSA, and recent human brains (Snodgrass, Leonard,and Robertson 2009). Surprisingly less attention has beengiven to the conditions behind the final burst in homininbrain expansion during the Middle Pleistocene, which led intothe MP. Behaviors loosely associated with the tail end of this

S290 Current Anthropology Volume 54, Supplement 8, December 2013

neurological trend included more extensive use of caves whereavailable, ubiquitous fire technology, and residential campsthat supported diverse activities. Thereafter we see remarkableconsistency in how MP people reaped the benefits of largeprey through extensive meat transport to base camps, fire-centered food processing, formalized butchering, and meatsharing.

From my perspective, the lists of differences thought todistinguish the MP and late MSA (McBrearty and Brooks2000) or the MP from the Eurasian UP (Mellars 1989, 1996)are reducible to three phenomena: (1) diet breadth and as-sociated radiations weapons technology, (2) durable art (i.e.,beads and inscribed ochre and ostrich eggshells), and (3) thedegree of cultural volatility or “changeability.” In these re-gards, some phases of the MSA and the UP do seem to havemore in common with each other than either has with theMP.

Behavioral Variation Within and Bracketingthe MP Period

The most obvious component of MP subsistence is large gameexploitation—prey selection and capture, carcass processing,land use, and provisioning of sites with meat. Other importantthemes in MP subsistence are small game exploitation andforaging technology. Also important are the nature of sitesand their functions as evidenced from material contents, oc-cupation intensity, and the diversity of activities represented.Residential hub sites are very much in evidence in the MP,and their development is relevant to the nature of coopera-tion, delayed food returns, and domestic arenas of social in-teraction.

Living High in the Food Web

It is safe to say that MP people were dedicated large gamehunters. While they must also have eaten plant foods, espe-cially in lower latitude environments, MP hominins occupiedmany cold regions in which plant productivity was highlyseasonal. In the absence of storage and milling tools for mak-ing calorie-rich seeds or nuts digestible, meat was the onlyhigh-quality food available to northern MP populations formany months of the year. Important rare finds of seeds orcharred nut hulls in combustion features (Barton et al. 1999;Madella et al. 2002) and phytolith and starch grains preservedin Neanderthal dental calculus (Henry, Brooks, and Piperno2011) simply confirm what we all expect (Jones 2009), namely,that MP hominins ate plant foods and sometimes used fireto cook them.

What too many discussions of diet breadth fail to grasp,however, is the core importance of search and processing costsof different foods in relation to their nutritional yield. Thisis a problem of how differing proportions of high- and low-cost food sources in the diet reflect variable human invest-ments (sensu Stephens and Krebs 1986). Such a contrast po-

tentially exists within the spectra both of plant use and animaluse. Unfortunately, plant macro- and microfossil evidencecannot tell us much about the proportional contributions ofhigh- versus low-cost plants to early diets. Some archaeolo-gists turn to the technological record for answers, focusingparticularly on durable milling, scraping, and poundingequipment to learn about investment and processing costs.Plant-heavy diets among recent foragers tend to correlate withemphasis on staple plant seeds and/or nuts (Keeley 1995),and heavy, durable tools are needed to extract the carbohy-drates from them efficiently. Apart from fire, the MP generallylacks such tools. A distinct line of evidence from light isotopestends to support the idea of high levels of carnivory in MPhominins in Europe (e.g., Bocherens 2009), but data on MPpopulations at lower latitudes are still too few.

As for variable investments in the meat diet, the zooar-chaeological record is the richest source of information onthe range of animals eaten and the relative importance of eachtype provided that the prey species possess skeletons of somekind. Faunal records present their own set of challenges tointerpretation, but the inorganic components of bone andcarbonate shells are well preserved in many regions of theworld. Equally important, the preserved remains can be re-lated in a consistent if generalized way to consumable organicparts within and across prey species.

Large prey in the size range of gazelles/roe deer to aurochs/bison were the core meat sources during the MP (cf. Delagnesand Rendu 2011; Gaudzinski-Windheuser and Niven 2009;Hoffecker and Cleghorn 2000; Rabinovich and Tchernov1995; Speth and Tchernov 2007). In north-central Europe,MP hominins exploited large prey irrespective of globalwarming through the Eemian Interglacial (Gaudzinski 2004).A concerted focus on large prey was also typical in the bi-otically diverse Mediterranean Basin, where 95%–99% of theanimal foods procured by weight were large herbivores (fig.1; Stiner 2005). Within the “large prey” size window, MPhunters responded easily to variation in the species availableto them, hunting mainly red deer, ibex, or wild goat in someareas; aurochs, bison, horse, or reindeer in others; and camelor fallow deer and gazelle in yet others (cf. Adler and Bar-Oz 2009; Burke 2000; Conard, Bolus, and Munzel 2012; Davidand Farizy 1999; Gaudzinski and Roebroeks 2000; Griggo2004; Hoffecker 2009; Rendu et al. 2012; Reynaud Savioz andMorel 2005; Rosell et al. 2012; Stiner 2009).

Rhino and straight-tusked elephant or mammoth remainsalso occur at some MP sites, particularly in open settings.Whether these megafaunal species were hunted or scavengedremains an open question, but their meat clearly was exploitedin some cases (Bocherens 2009; Conard and Prindiville 2000;David and Poulain 1990; Griggo 2004; Hoffecker 2009; Patou-Mathis 2000; see also Villa 1990). Megafauna remains in cavesites are uncommon if present at all probably because of thehigh cost of transporting their heavy bones. Megafauna gen-erally drop out of MP records in accordance with the ex-tinction of these species in local ecosystems (e.g., Garrard

Stiner Trends versus Conservatism in the Predatory Niche S291

Figure 1. Percentage of total prey biomass represented by ungulate prey in three Mediterranean faunal series (adapted from Stinerand Kuhn 2006). Major climate cycles represented by the marine oxygen isotope record are adapted from Martinson et al. (1987).

1983; Hoffecker 2009; Hortola and Martınez Navarro 2013;Stuart 1991:488–502), but aurochs-sized and smaller ungu-lates never lost their importance to MP hunters.

A long-standing signature of human hunting is the ten-dency to focus on prime-age artiodactyl prey (Stiner 1990).In reality, human-generated ungulate mortality patterns varymuch more than this, but most patterns fall between non-selective and strongly biased to prime adults, averaging to amild bias to prime adult prey (figs. 2, 3; Stiner 2005). Thisbias in prey age selection is widespread in the MP and UP(Gaudzinski and Roebroeks 2000; Hoffecker and Cleghorn2000; Patou-Mathis 2000; Rendu 2010; Speth and Tchernov1998; Steele 2004; Stiner 1990, 2005; Yesherun, Bar-Oz, andWeinstein-Evron 2007). The bias has also been documentedin a few late Lower Paleolithic (LP) sites such as at Waller-theim in Germany (Gaudzinski 1995) and Qesem Cave inIsrael (Stiner, Barkai, and Gopher 2011). The ubiquity ofprime-focused hunting in the MP is striking, however, andmay suggest a galvanization of this dimension of the largegame hunting niche.

The ungulate body parts that MP hominins carried fromkill sites to camps varied with circumstance, but repetitivebiases are found in caves and rock shelters. Ideally we wouldlike to have equally reliable data on prey body-part profilesacross open and natural shelter contexts, but preservationoften favors samples from shelters. Shelter faunas nonethelessprovide us with some analytical benefits because food mustbe carried to them. Most group members were likely presentat these sites, making sharing difficult to avoid. Evidence formeat sharing in shelters can be explored by (1) consideringthe quality and quantities of meat carried into them, (2)attempting to follow the path of meat redistribution on site,and (3) comparing the scope of activities at camps that mayrelate to modes of economic cooperation.

Ungulate skeletal representation typically varies from wholebodies to a bias favoring meaty or marrow-rich limb partsand heads (e.g., Burke 2000; Conard and Prindiville 2000;Griggo 2004; Patou-Mathis 1993, 2000; Speth and Tchernov2001; Stiner 2005; Valensi 2000). Figure 4 exemplifies therange of variation typical of medium-sized species from con-secutive layers in Ucagızlı Cave II (MP) and Ucagızlı Cave I(early UP) in southern Turkey. Here and elsewhere, MP for-agers often left axial parts behind at kill sites or exploitedthem in the field. Differential preservation of dense versusspongy bones (following Lyman 1984) may explain the an-atomical biases in some cases (e.g., Hoffecker, Baryshnikov,and Potapova 1991; Lam and Pearson 2005). However, thepatterns are clearly anthropogenic in a great many other cases(e.g., Burke 2000; Conard and Prindiville 2000; Rendu 2010;Rendu et al. 2012; Speth and Tchernov 2001, 2007; Starkovich2011; Stiner 1994, 2002, 2005) because there is no correlationto bone density, or similarly dense skeletal features are quan-tified across the axial and appendicular anatomy to inferbody-part profiles (Stiner 2002).

The MP is rife with examples of delayed consumption ofhigh-quality animal parts. Carrying meat and marrow-richbones to a central place allowed for more thorough process-ing, but the large volumes of meat moved also testify to anethic of premeditated sharing. Evidence for delayed meat con-sumption is also found at late LP sites such as at Qesem Cave(Stiner, Barkai, and Gopher 2011) and probably the open siteof Gesher Benot Ya’akov (Rabinovich, Gaudzinski-Windheu-ser, and Goren-Inbar 2008), but the MP differs in the sheerabundance of clear-cut cases. In comparison with the stratifiedlate LP record of Qesem Cave (Gopher et al. 2005), wherebutchering activities involved mainly simple defleshing andmarrow extraction, MP meat sharing also appears more for-malized (Stiner, Barkai, and Gopher 2011). The types of tool

S292 Current Anthropology Volume 54, Supplement 8, December 2013

Figure 2. Models a and b averaged values for observed ungulatemortality patterns created by various human and nonhumanpredators. Shaded panels represent natural variations in the agestructures of living ungulate populations and thus also nonse-lective mortality patterns and mortality patterns caused by at-tritional factors such as disease, accidents, and malnutrition. Eachcorner of the diagram represents a strong bias toward the des-ignated prey age group. Observed means are for recent spottedhyena (1), wolf (2), Cape hunting dog (3), tiger (4), African lion(5), Holocene and recent human hunters (6), MediterraneanEpipaleolithic and Upper Paleolithic hunters (7), and Mediter-ranean Middle Paleolithic hunters (8). An asterisk indicates theaverage for the Acheulo-Yabrudian (late Lower Paleolithic) fallowdeer assemblages from Qesem Cave.

marks on the ungulate bones from Qesem Cave are redun-dant, and the cut-mark orientations are relatively chaotic. Thelatter observation indicates many procedural interruptions ordiverse positions while cutting flesh and perhaps a more in-dividualized way of consuming shared meat. MP and earlyUP butchering practices are more similar to one another, withmuch consistency in cut-mark orientations, suggesting a morecomplex, socially canalized way of sharing meat.

No discussion of the MP would be complete without somemention of the scavenging controversy. In the 1980s, LewisR. Binford challenged a long-standing assumption that earlyhominins (particularly LP hominins) were big game hunters(Binford 1981). He even suggested that MSA and MP hom-inins might have been obligate scavengers of other predators’kills (Binford 1984:244–249, 1988). The second hypothesiswas generally refuted by subsequent analyses (e.g., Chase 1986,1989; Grayson and Delpech 1994; Speth and Tchernov 2007;Stiner 1990, 1994). I, a student of Binford, concluded fromlong study of MP faunas in Italy that “Mousterian peoplewere competent hunters of small, medium and large ungulates

but at times turned to very different foraging agendas whileusing coastal caves in Latium, scavenging the very same kindsof ungulates they hunted elsewhere in addition to collectingshellfish and tortoises” (Stiner 1994:371). With the pendularreturn to the view of MP folk as big game hunters, scavengingbehavior seems to have sunk beneath our notice (but seeConard and Prindiville 2000). This is unfortunate, becausenonconfrontational scavenging is a form of gathering aboutwhich we still know little for the MP. At the site of Grottadei Moscerini in west-central Italy, probable evidence for scav-enging head parts, nearly always from old animals, coincideswith an unambiguous record of collecting shellfish and tor-toises (Stiner 1994).

Small Game Exploitation and Its Implications

Large game hunting practices generally fail to differentiate theeconomies of MP and UP societies in Eurasia (see Adler andBar-Oz 2009; Adler et al. 2006; Gaudzinski 2000, 2004;Grayson and Delpech 2003; Hoffecker 2009; Munzel and Con-ard 2004; Stiner 1994, 2005; Stiner, Barkai, and Gopher 2011).Economic differences are more apparent in how these paleo-cultures filled gaps in large game availability with small an-imals. In warm-temperate and subtropical environments, MPhominins favored small animals that were easy to collect—tortoises, marine shellfish, ostrich eggs, and large lizards ifavailable. This makes good sense from the viewpoint of a preychoice (optimality) model, as the low handling costs of slow-moving or immobile small animals compensate for their smallbody size. Small, quick animals—particularly birds, hares, andfish—generally have higher handling costs, and MP peopledid not pursue these prey most of the time (Aura et al. 2002;Costamagno and Laroulandie 2004; Laroulandie 2004; Stiner,Munro, and Surovell 2000). Whatever flexibility may haveexisted in MP foraging systems, it seldom extended to animalsor plants with high capture or processing costs.

Although habitual use of small, quick animals is rare inthe MP overall, fascinating exceptions exist in southwesternEurope, usually involving rabbits (Aura et al. 2002), a colonialburrower, and some birds. La Canelettes in France containssignificant numbers of cut-marked rabbit and bird bonesalongside large game remains (Cochard et al. 2012). Anotherexception is found in the early MP at Bolomor Cave in Va-lencia, Spain. Here cut-marked rabbit remains occur in rel-atively high percentages in multiple horizons along with somebirds and many tortoise and ungulate remains (Blasco andFernandez Peris 2009). The fact that these examples are lo-calized occurrences that repeat through multiple occupationssuggests that unique aspects of the locality made this unusualmanner of foraging feasible. Some instances of MP bird ex-ploitation, again very rare, associate with ochre use and mayhave been for obtaining large feathers (Peresani et al. 2011;Rendu 2010). Use of quick, small prey is not the rule in theMP, however, and the above examples stand out because oftheir rarity (Cochard et al. 2012).

Stiner Trends versus Conservatism in the Predatory Niche S293

Figure 3. Observed mortality pattern distribution in artiodactylprey from Mediterranean Epipaleolithic (EPI), Upper Paleolithic(UP), and Middle Paleolithic (MP) assemblages from Israel (Hay-onim Cave, Meged Rockshelter, Kebara Cave), Lebanon (Ksar’Akil), Turkey (Ucagızlı I Cave), and various cave sites in west-central and northern Italy. Acheulo-Yabrudian (late Lower Pa-leolithic) cases from Qesem Cave appear as diamonds.

A major shift in the overall breadth of the meat diet in theMediterranean region coincides with the MP-UP transition(fig. 5; Stiner 2001; Stiner et al. 1999). Differences in smallgame use between the two Paleolithic periods are made allthe more fascinating by the fact that diet diversification andintensified use of resources can support a larger populationbase. In the Levant, this transition in small game use is linkedto harvesting pressure on sensitive tortoise populations. Themean sizes of tortoises declined (Speth and Tchernov 2002;Stiner et al. 1999), and unnatural skewing is evidenced in theage (size) structure of the harvested animals, which points toheavy exploitation by humans in particular (Stiner 2005:139–147). These findings imply that human populations in theregion first exceeded the potential of high-ranked, high-returnresources to support them after about 50,000 years ago (seeSpeth and Clark 2006 for effects within the late MP). De-clining availability of high-yield resources, which raises searchcosts, is considered the main reason why foragers turn tofoods that give a lower return for the effort (Stephens andKrebs 1986:17–24). Middle Paleolithic foragers’ great depen-

dence on large game and on slow-moving and slow-growingsmall animals such as Mediterranean tortoises implies thatMP populations were consistently small and highly dispersed.Given that the first detectable human demographic pulse oc-curred more or less at the threshold of the MP-UP culturaltransition, some very late Mousterian populations may havebeen uniquely affected by expansions of early UP populationsin Eurasia.

The African MSA is not as extensively documented as theEurasian MP. Especially troubling is the data gap for the in-terval 60,000–20,000 years ago (within marine isotope stage[MIS] 3) in both South and North Africa (Steele and Klein2009) and spotty reporting in East Africa. There are hints ofa greater range of subsistence and technological behaviors insome areas and phases of the MSA (reviewed by McBreartyand Brooks 2000; Steele and Klein 2009). Scarce bird remainsare reported in some North African sites, small mammal re-mains (hare and hyrax) in the Ethiopian site of Porc-Epic(Assefa 2006), catfish bones in Aduma Middle Awash Valleysites (Yellen et al. 2005), and early bone harpoons (Yellen etal. 1995) and grinding tools that may have been suitable forprocessing seeds or nuts (McBrearty and Brooks 2000). Thesecases do not seem to constitute a single trend but ratheroscillate substantially with environmental changes (especiallyin MIS 4; Clark and Kandel 2013). The MSA data nonethelesssuggest a contrast to the MP, even to the MP in warmer climes.Better documented jumps in diet breadth occur in AfricanLate Stone Age sequences (Klein 1999; Steele and Klein 2009)and post–Large Glacial Maximum sequences in Eurasia(Munro 2004; Stiner 2001; Stiner et al. 1999).

Evidence of early shellfishing has attracted new attention.It is said by some to be evidence of precocious expansionsin dietary breadth and cognitive advances during the MSA inSouth Africa (Marean et al. 2007). The MP has not figuredmuch in this discussion, but it should (see Colonese et al.2011). Evidence for probable MP shellfishing in the Medi-terranean region—mainly limpets, large clams, mussels, oys-ters, and turbans—has been widely reported in archaeologicalworks of the 1950s onward at Gibraltar (Baden-Powell 1964;Garrod et al. 1928); on the Italian coasts of Liguria, Lazio,and Puglia (Blanc 1958–1961; Palma di Cesnola 1965, 1969);and at Haua Fteah in Cyrenaica, North Africa (Klein andScott 1986). Detailed taphonomic studies are available foronly some of these sites, but recent results confirm that hom-inins exploited shellfish at Cueva Perneras and Cueva de losAviones in Spain (Zilhao et al. 2010), Kalamakia Cave inGreece (Darlas 2007), Grotta dei Moscerini in Italy (Stiner1994), Ucagızlı Cave II in southern Turkey (Stiner 2009), andin Morocco (Steele and Alverez-Fernandez 2011). Several ofthe MP sites on Gibraltar contain shellfish remains, and al-though some are almost certainly archaeological, their attri-bution to hominin activities continues to be controversial (cf.Barton 2000; Erlandson and Moss 2001; Fernandez-Jalvo andAndrews 2000; Finlayson et al. 2008; Freeman 1981; Kleinand Steele 2008; Stringer et al. 2008). The earliest secure case

S294 Current Anthropology Volume 54, Supplement 8, December 2013

Figure 4. Comparison of standardized skeletal element frequencies (observed MNE/expected MNE) by layer and anatomical regionfor medium-sized ungulate prey at the Epipaleolithic and early Upper Paleolithic site of UcI (layers EPI though I, youngest tooldest) and the Middle Paleolithic site of UcII (layers A through C-D in bottom row, arranged youngest to oldest) in southernTurkey. Unevenness in the anatomical profile indicates biases in body-part representation relative to the complete animal anatomy.Dental elements were not used for the calculations of head-part abundance.

for MP shellfish exploitation comes from Grotta dei Mos-cerini, a deeply stratified cave site on the Gaetan coast of Italythat spans 115–65 ka (Stiner 1994:180–192).

The earliest record of MSA shellfish exploitation at PinnaclePoint (PP13B) in South Africa has been dated to about 160ka (Marean et al. 2007). Taken at face value, shellfishing onthe South African coast began roughly 40,000–50,000 yearsbefore it appeared on Mediterranean shores. However, geo-logical studies of shoreline changes indicate that much of theMediterranean archaeological record is now submerged or hasbeen lost to the erosion with marine transgression followingthe Last Interglacial (Lambeck 1996; Van Andel and Tzedakis1996). Given that all of the known shell-bearing MP sitesoccur at or just a few meters above modern sea level, onecan assume that earlier coastal occupations have been scouredclean by waves or inundated by rising seas (reviewed in Baileyand Flemming 2008; Colonese et al. 2011). For the moment,115–110 ka is merely the minimum age for shellfish collectingin the Mediterranean region. Suggesting significantly differentbehavioral thresholds between the MP and MSA for marineforaging therefore is premature.

Marean et al. (2007) argue that early shellfishing in SouthAfrica fueled a demographic expansion of MSA populationsalong coastlines northward into Eurasia. The energetic basisof this proposition is highly questionable, but even if such athing occurred, the expanding MSA populations would soon

have bumped into Mediterranean Neanderthals already in thehabit of gathering mussels, limpets, and turbans from shore-line rocks and digging for clams buried in the sand. The ideathat the coast was a main corridor for colonization is doubtfulin any case given that good shellfishing areas were virtuallyabsent along the coast from Egypt to northern Galilee. Im-portantly, Clark and Kandel (2013) show that the quantitiesof shells in coastal MSA are low as a rule and on par withthose in the MP sites.

Labor Allocation and Foraging Equipment

This is not a paper on Paleolithic technology, but a few re-marks about foraging implements are relevant to any consid-eration of hominin subsistence. First, it is clear that late LPand MP hominins routinely hunted large game animals longbefore the development of stone-tipped and bone-tippedweapons. In Eurasia, stone-tipped weapons appear in the laterMP as localized traditions. Hafted Levallois points are re-ported in the Levant (Shea 1989, 1993), for example, andbifacial Blattzspitzen in Germany (Muller-Beck 1988) alongwith rare pointed bone artifacts (Conard, Bolus, and Munzel2012). Modest weapons diversification therefore is suggestedby the late MP. Bifacial stone and bone points are also reportedin the African MSA (McBrearty and Brooks 2000; Shea 2009).Because stone-tipped weapons are not universal to the MP

Stiner Trends versus Conservatism in the Predatory Niche S295

Figure 5. Changes in evenness in the exploitation of three smallprey categories based on the inverse of Simpson’s index (1 pleast even and the narrowest diet; 3 p most even and thereforethe broadest meat diet). Prey are low-cost slow game and higher-cost quick-running terrestrial mammals and quick-flying birds;fish are very rare and therefore are excluded. Black-filled circlesrepresent the Ucagızlı II-I faunal series; unfilled and gray-filledcircles denote other Mediterranean series in Italy and Israel (fromStiner 2001). Sources: late MP from Kebara Cave ca. 65–50 ka(Speth and Tchernov 2001); Paleolithic Italy, Israel, and Turkey(Munro 2004; Stiner 1994, 2001, 2005, 2009; Stiner, Barkai andGopher 2011); Greece, Klissoura Cave 1 (Starkovich 2011);Franchthi Cave (Stiner and Munro 2011). (Adapted from Stiner2001, 2005.)

or the MSA, however, wooden spears were almost certainlythe default weapon of choice. In Eurasia, this represents acontinuation of late LP hunting equipment (cf. Jacob-Friesen1956; Thieme 1997).

The evidence forces us to decouple suppositions about Pa-leolithic hunting capabilities and complex hunting weapons.Improvements in weapons design are typical of the UP (e.g.,Knecht 1997), and increased weapons efficiency or reliabilitymay have reduced an individual’s procurement time, risk perforay, and the minimum hunting party size needed to capturelarge animals. Middle Paleolithic hunters nonetheless man-aged very well without complex weapons, relying instead onextensive cooperation among group members. The impor-tance of cooperation in MP big game hunting is underscoredby the apparently low population densities of the period. Closeteamwork would have been utterly essential to MP huntingsuccess and maintaining minimum party size a perennial con-cern.

While there is no necessary link between the ability to bringdown large game animals and weapons complexity, there isa predictable relation between the exploitation of small, quickanimals and technologies that reduce capture costs in Pa-leolithic Eurasia. This is because the quick prey are small andthe returns are too low relative to human metabolic needs inthe absence of tools that make harvesting efficient (sensuOswalt 1976). Making and maintaining the harvesting toolsalso incurs costs, which some later Paleolithic infrastructures

eventually came to support. We have seen that MP peoplecould obtain quick animals if they wanted to but seldom choseto do so. Early radiations in small game hunting equipmentare reported in some MSA records of Africa (e.g., fishing;McBrearty and Brooks 2000), apparently in connection withexpanding dietary breadth.

Expanding dietary breadth has labor and social connota-tions. The benefits of niche (labor) separation within humangroups should increase as the diet broadens over long periodsof time because of less overlap or symmetry in personal sched-ules and the locations where various foods can be obtained.Such within-population diversification would be most ad-vantageous in environments where key resources occur atdisparate locations or times or if distinct methods and toolsare required to obtain them efficiently. These conditions mayalso result in greater autonomy with respect to what individ-uals add to a pool of shareable foods.

To appreciate the social ramifications of expanding dietbreadth across the MP-UP, one must consider diversificationin technology and diet simultaneously. Upper Paleolithic pop-ulations readily diversified along either dimension or both inresponse to environmental circumstance. They produced, forexample, weatherproof clothing and related items to supportthe food quest in cold areas and diverse food harvesting equip-ment in warmer areas. Acknowledging that we know littleabout organic artifacts of any Paleolithic period, the MP none-theless shows less flexibility in both technology and the scopeof the meat diet. There is considerable evidence for hide work-ing in the MP (Meignen et al. 1989; Villa, Bon, and Castel2001), suggesting that European Neanderthals wore skin gar-ments. However, evidence of working dry hide (cured leather)in the form of microwear traces on stone artifacts is com-paratively scarce in the MP (e.g., Anderson-Gerfaud 1990;Beyries 1987; Lemorini 2000; Martınez-Molina 2005). It isonly with the UP that the types of artifacts commonly as-sociated ethnographically with tailored, weather-resistantclothing—bone needles and awls—also became a regular partof the Eurasian archaeological record. Upper Paleolithic so-ciety seems to have differed from MP society because of awider range of economic and social roles.

Kuhn and Stiner (2006) hypothesize that MP women, chil-dren, and men all participated more consistently and directlyin large game hunting than is generally seen among recentforagers or UP foragers. This is not to say that socioeconomicroles among MP individuals by age or gender were identical.Rather, the point is that a comparatively narrow economicfocus on large game and very low population densities con-strained labor organization and land use in historically uniqueways. Like other predators that hunt socially, humans cangain significant advantages over large prey if some membersof the hunting party act as artificial surrounds or funnels thatmove the quarry toward the killers. (Only substantive in-creases in weapon efficiency can alter this dynamic.) Individ-ual roles in MP hunting would have varied from direct phys-ical contact with prey to diverse, safer tasks for helpers. In

S296 Current Anthropology Volume 54, Supplement 8, December 2013

such small-scale societies that lacked sophisticated killingtools, group members had to be available on short noticemuch of the time. Beating the bushes, carrying meat andbones back to camps, processing carcasses, and tanning hideswere all essential to gaining high payoffs from a hunt. Morehelpers meant that more meat and bone could be carried backto camps, cutting losses to scavengers and allowing more tobe gained from the meat, marrow, and hides of large preyanimals. Given the perennial risk of groups being stretchedtoo thin and the apparently limited development of regionalsocial networks, within-group proximity would also have beenthe main safety net for MP people.

Site Function and Occupation Intensity

Two modes have been noted in MP site data: (1) ephemeraloccupations with high mobility and (2) relatively intensiveoccupations with high material inputs. The two conditionsdo not seem to represent a single continuum. The relativefrequencies of ungulate and lithic materials are positively andstrongly correlated in the richer cave sites (Moncel and Dau-jeard 2012; Riel-Salvatore and Barton 2004; Stiner and Kuhn1992), implying that tool manufacture and maintenance wereintimately connected to large game exploitation. The relationis weaker or nonexistent in sites with scant materials and highrates of tool retouch and exotic raw materials (e.g., GrottaGuattari, Grotta dei Moscerini [Kuhn 1995], and YarımburgazCave in Thrace [Kuhn and Stiner 2010]). Delagnes and Rendu(2011) propose that the duration or length of lithic reductionsequences may also have varied with mobility and huntingstrategies. The sparse hominin occupations in caves tend tointercalate or be intermixed with carnivore occupations (Bru-gal and Jaubert 1991; Gamble 1999; Patou-Mathis 2000; Spethand Tchernov 1998; Stiner 1994; Stiner, Arsebuk, and Howell1996; Straus 1982; Villa and Soressi 2000). Gamble (1986)was among the first to examine the distribution of this phe-nomenon across the Mediterranean area. Where hominidcomponents are thin, carnivore components often are thickand easily recognized (e.g., Aura et al. 2002; Brugal and Jau-bert 1991).

Other MP sites instead are dominated by hominid-gen-erated debris. Early MP examples in the Levant include Hay-onim Cave (Stiner 2005) and Misliya Cave (Yesherun, Bar-Oz, and Weinstein-Evron 2007). The quantities of faunalmaterial in these sites are not tremendous relative to sedimentvolume, but all of the high-quality ungulate body parts thatwe would expect to be valued by hunters are represented alongwith hearth traces and evidence of butchering and marrowprocessing. Later MP examples are both more common andoften richer in material, such as in Kebara Cave (Bar-Yosefet al. 1992; Speth and Tchernov 1998) and Qafzeh Cave (Ra-binovich and Tchernov 1995). Comparable situations are re-ported throughout Europe (e.g., Chase 1986; David and Pou-lain 1990; Farizy, David, and Jaubert 1994; Gaudzinski 1995;Jaubert et al. 1990; Stiner 1994). Some of the variation in the

density of material in MP sites must relate to the season ofoccupation (e.g., Boyle 2000; Hoffecker and Cleghorn 2000;Patou-Mathis 2000; Pike-Tay et al. 1999), and it seems likelythat cold-weather camps may have lasted longer (e.g., Conard,Bolus, and Munzel 2012). However, the generally opposingrelation between hominin and carnivore presence signals animportant divide in the function of sites within MP territories.This phenomenon continues through the early UP in someregions, suggesting that low population densities could be partof the explanation.

The MP site record is not entirely trendless. There is anincrease in the number of sites after about 70 ka (e.g., MirazonLahr and Foley 2003; Van Andel et al. 2003). The ravages oftime complicate interpretation of this trend but cannot ex-plain it away. Moreover, a higher proportion of the later MPsites were occupied more intensively. For example, humanoccupation intensity in the long consecutive sequences ofHayonim (early MP) and Kebara Caves (late MP) has beencompared through integrated analyses of faunal and lithicassemblages, hearth features, and wood ash deposits (Meig-nen, Speth, and Stiner 2006; Meignen et al. 2010). The MPoccupations in Hayonim Cave are repetitive in content butsmall in scale. Thermoluminescence dates suggest a sedimentaccumulation rate of about 1 m per 10,000–15,000 years. Bycontrast, each meter accumulated in roughly 3,000 years inKebara Cave (Goldberg and Bar-Yosef 1998), and the densitiesof lithic artifacts are also much higher.

While the intensity of human presence at Hayonim wasgenerally low, all stages of stone tool production are in evi-dence (Meignen et al. 2010), and raw material came mainlyfrom local sources (Delage, Meignen, and Bar-Yosef 2000).Ungulates and tortoises were exploited repeatedly, and evi-dence of fire-aided food preparation is widespread (Stiner2005; Weiner, Goldberg, and Bar-Yosef 2002). Also commonis evidence of postdepositional burning of bone and lithicmaterial, indicating considerable reuse of certain locationswithin the cave (Stiner et al. 2001). The variety of economicactivities represented tells us that Hayonim cave served mainlyas a residential camp. Complete on-site core reduction anda diversified tool kit rule out the possibility of a special usesite (Meignen et al. 2010). The pace of material buildup (den-sity) was significantly higher overall during the late MP inKebara Cave (Albert et al. 2000, 2003; Goldberg and Bar-Yosef 1998; Meignen et al. 2010; Weiner, Goldberg, and Bar-Yosef 2002). A concentrated bone midden occurs along thenorth wall in Units XI-IX (Speth and Tchernov 1998), andcarcass processing and consumption occurred at spatially dis-crete locations inside the cave (Speth and Clark 2006; Spethand Tchernov 2007). Hearth features are prevalent and deeplystacked with little or any sediment formation between lenses(Bar-Yosef et al. 1992; Meignen, Bar-Yosef, and Goldberg1989). A full suite of lithic production is represented in Kebara(Meignen et al. 2010), generally on local flint, and the toolswere used for diverse tasks, including butchery, woodworking,cutting, and scraping of hard and medium materials (Plisson

Stiner Trends versus Conservatism in the Predatory Niche S297

and Beyries 1998; Shea 1991). The absence of exotic rawmaterial and the complete reduction sequence of cores in thesite indicate large, well-stocked encampments of prolongedduration (Meignen et al. 2010). Some MP layers in KebaraCave display considerable heterogeneity in site structure, ap-parently in response to rapid debris buildup. The relegationof bone trash along the peripheries of habitation areas isexpected to increase with the duration of an occupation (Bin-ford 1978, 1991, 1998; Brooks 1984; Galanidou 2000; Gorecki1991). This condition is not seen in the early MP of HayonimCave even though this site and Kebara served mainly as res-idential encampments. There evidently was more impetus tomanage domestic space in the later, denser site.

Residential camps are places where all group members maycongregate and engage in a wide range of activities. By thisdefinition, residential camps are clearly in evidence through-out the MP and UP. If we try to count the number of distinctactivities in dense sites, however, there seem to be fewer rep-resented in MP camps than in UP camps (and fewer still inthe late LP camps). Trends in the complexity of site activityhistories may be undermined by differential preservation butnot where durable classes of technological material are in-volved.

There is at least one other aspect of site use—the centralityof fire—that may define the MP period as a whole. Fire tech-nology emerged well before the MP (Alperson-Afil andGoren-Inbar 2010; Gowlett et al. 2005; Karkanas et al. 2007;Preece et al. 2006), but the record of fire becomes ubiquitouswith the beginning of the MP (e.g., Albert, Berna, and Gold-berg 2012; Fernandez Peris et al. 2012; Goldberg et al. 2012;Roebroeks and Villa 2011; Rolland 2004; Vallverdu et al.2012). Hearths were magnets for carcass processing activitiesin MP residential sites (Rosell et al. 2012; Speth 2006; Stiner2005; Vallverdu et al. 2012; Vaquero and Pasto 2001), andthey probably were intensely social spaces as well (Foley andGamble 2009; Gamble 1999). Evidence of fire is not preservedin every MP site, but it is as common in MP sites as it is inUP sites if time-dependent preservation is taken into account(Roebroeks and Villa 2011). Sandgathe et al. (2011) take adifferent view, arguing that fire only became a regular partof life in the late MP, and that earlier use of fire was limitedto curation of live coals obtained from natural fires. It isdifficult to reconcile Sandgathe et al.’s interpretation againstthe rich, long-term fire records in the late LP site of QesemCave based on the distribution of burned bones (Stiner,Barkai, and Gopher 2011), the early MP of Hayonim Cave(Weiner, Goldberg, and Bar-Yosef 2002) and Bolomor Cave(Fernandez Peris et al. 2012), or geologist A. Segre’s obser-vations from the excavation of Grotta dei Moscerini (Stiner1994:46–52). Repeated use of fire in these early cultural seriesspans distinct climate cycles and environmental regimes. It isinteresting that MP fire records are much clearer in someregions than in others and not necessarily in the coldest en-vironments. Whether this anomaly reflects regional behavioraldifferences or taphonomic differences in preservation poten-

tial (Goldberg et al. 2012) is not fully resolved, but probablyit is due to the latter.

Evidence Summary

It is a good deal easier to find subsistence trends between theMP and the periods before and after it than within the MPitself. The transition from the late LP to early MP seemsgradual with respect to dependence on large game, basic hunt-ing equipment, aspects of blade technology, and prey ageselection. One possible contrast to the late LP is a more com-plex pattern of meat butchering and sharing based on cut-mark data, but more work is needed on this issue. Anothercontrast to the LP is the consistent and rather more complex“residential” nature of many MP sites, including the ubiquityof fire. The emergence of burial customs and pigment usealso set the MP apart from the late LP. These practices almostcertainly relate to important social developments in the MPeven if we do not know just what they were about. The shiftsbetween the MP and early UP tend to be more abrupt, al-though not with respect to big game hunting or the dense-ephemeral site contrast within territories. However, the UPdisplays novel solutions for managing foraging risk throughclear expansions in dietary breadth, the use of artistic media,and rapid technological diversification. A number of archae-ologists argue that these qualities of the UP contributed togreater population stability and environmental carrying ca-pacity.

These observations have led to an impression of MP pop-ulations as proverbially stuck in the mud, left behind by evo-lution. Sampling and preservation issues notwithstanding, thelate MP may hint at a few subtle changes within the period.Hominin populations may have grown, based on site structureevidence and numbers of sites, although there are few if anyindications of increased predator pressure on large or smallanimals resources (but see Speth and Tchernov 2007 on Lev-antine cases). Dated burials also cluster in the later part ofthe MP. And the late introduction of stone-tipped huntingweapons suggests mild increases in hunting efficiency in someregions. MP meat diets nonetheless remained narrow nearlyeverywhere.

Part of the explanation for greater diet breadth in the Af-rican MSA could be environmental, because so much of theMP range falls in higher latitude environments, where thechoice of animal and plant foods would have been less varied.The same cannot be argued for the contrast between theMediterranean MP (including the late MP) and the early UP,as both existed sequentially in the same parts of the worldand each weathered many shifts in climate, environment, andfood supplies. Dietary and technological evidence show usthat both MP and UP societies depended on systems of di-vided, collaborative labor, but MP activities indicate a nar-rower spectrum of on-site activities and probably greater uni-formity in ranging patterns of group members.

The most readily visible differences between the early and

S298 Current Anthropology Volume 54, Supplement 8, December 2013

late MP seem to be about numbers of people—rates of sitevisitation, population density at the regional scale, and pos-sibly social group sizes. The potential for population growthin the MP would have been severely constrained by frequentresidential moves and an economy based on a high-qualitybut unpredictable nutrient supply (Boone 2002; Munzel andConard 2004; Stiner, Munro, and Surovell 2000). Of course,MP females maintained viable levels of reproductive successthrough thick and thin. How did they pull it off, if the socialsafety nets of the MP did not extend very far beyond theboundaries of small groups? In the absence of reliable regionalnetworks, group coherence, vigilant concern for group mem-bers, and close cooperation would have been utterly criticalto MP survival. One would also expect males to remain innatal territories through adulthood if large game hunting werethe core of the economy. Recent findings by Lalueza-Fox etal. (2011) on the late MP individuals from El Sidron Cavesuggest that may indeed have been the case.

Why Is There Not More Variationwithin the MP?

One of the great challenges to studying the MP is explainingthe appearance of stasis in spite of the considerable intelli-gence of these hominins. Middle Paleolithic humans existedon narrow meat diets across a wide range of latitudes. Whywas there not more variation in small game use in the bi-otically diverse Mediterranean Basin? The prey animals werethere. The typically low population densities of the MP arepart of the answer. A heavy dependence on large game impliesa very high position in the food web, and it seems that con-sumer abundance seldom exceeded the supply of high-qualityfoods at the regional scale. While large game animals can yieldhigh average return rates, they associate with high variancesin supply (reviewed in Kelly 1995; Kuhn and Stiner 2006)and thus limit the reproductive potential of any highly car-nivorous population. These populations compensate withhigh mobility, which also facilitates recolonization where localextinctions occurred.

It is striking that even the late MP cultural record changedonly a little, because this is generally when AMH populationsexpanded into Eurasia. The conservative nature of MP culturemay have been the outcome of a very successful and long-standing adaptation, one that became fixed via the process ofspecialization. Paleolithic archaeologists, including me, havevacillated to and fro in attempts to classify hunter-gathererforaging as more specialized or generalized. Some of the con-fusion stems from how the terms may be applied at severaldistinct scales of behavior. All hominins were omnivores, forexample, and omnivorousness connotes a “generalist” feedingstrategy. Specialization is only informative if considered forone well-controlled aspect of econiche at a time. Middle Pa-leolithic hominins were unusually specialized in their focuson large game, with limited supplementation from small gameanimals. Streamlining and elegance are common signatures

of specialization along with a narrower tolerance range in oneor more dimensions of niche (Pianka 1978:253–256). Spe-cialization brings greater efficiency in the exploitation of cer-tain core resources that in turn may reduce the group’s abilityto be efficient in exploiting others. Because of this, it is dif-ficult for specialists to retreat from time-hardened success.Middle Paleolithic populations went far down the path oflarge game specialization, a path made possible in part byhigh ungulate biomass during the ice ages (e.g., Discamps2011).

The low population densities of the MP raise questionsabout the importance of cultural drift, because isolation canfoster regional variation and distinct trajectories of changeover the long term. Of course high mobility and the tele-graphic scent of fire meant that MP folk were not totally cutoff from one another. Considerable temporal and spatial var-iation nonetheless exists in the details of MP stone working.Rather than periodic, directional diffusion of novel ways ofdoing things, we see rather stochastic patterns or reappearanceof a set array of technological variants in different areas (e.g.,Delagnes and Meignen 2006; Kuhn 2013; Moncel and Dau-jeard 2012; Peresani 2012; Vaquero et al. 2012). This patternmay be a consequence of demographic factors in that low-density populations are subject to comparatively high ratesof local extinction and recolonization. These conditionswould reduce the probability of cumulative cultural evolutionvia complex learning traditions while promoting some degreeof nondirectional variation in material culture (Premo andKuhn 2010; Stiner and Kuhn 2006). The MP has the ap-pearance of a stable adaptation with plenty of flexibility builtinto it. But it was headed nowhere in particular—no evidenceof strong directional selection—consistent with the status ofa successful and well-tuned system. The MP was not a deadend in itself but rather the victim of the historically uniquecolonization by a competing population.

A Zooarchaeological Agenda

The evidence synthesized in this essay exposes some clear-cutopportunities for further research. Specifically, how differentwere MSA and MP population responses to food supplies if,for example, we compare them during climate intervals ofmore stability and less stability at a millennial scale (see alsoClark and Kandel 2013)? So-called Mediterranean-type en-vironments exist in southern Europe, the Levant, North Af-rica, and South Africa. The temporal and spatial records arenot yet well connected in climate science, but these regionscan serve as geographic laboratories. The null hypothesiswould be that climate is not the driver of human behavioralevolution. Siddall and colleagues (Siddall, Chappell, and Pot-ter 2006; Siddall et al. 2003) identify the climate intervals of40–80, 120–140, and 160–200 ka as particularly variable basedon Red Sea–Ocean proxies for global sea level changes. Theyidentify the climate intervals of 80–120, 140–160, and 200–220 ka as least variable. The most recent contrasting pair of

Stiner Trends versus Conservatism in the Predatory Niche S299

intervals, 40–80 and 80–120 ka, seem particularly suitable forcomparisons of MP and MSA strategies because faunal andother data are fairly abundant.

In order to undertake such a comparison, two other thingsmust be done. First, zooarchaeologists must get away frommaking lists of anecdotal evidence and focus on systematic,quantitative analysis of the full scope of the animal diet (big,small, vertebrate, and shelled invertebrate). Richard Klein andcolleagues have set a remarkable precedent for a holistic modeof data presentation in South Africa, a practice that sadly isout of vogue. The same limitations once applied to zooar-chaeological work in Eurasia, but this has changed in the pastfew decades, and now there are complete faunal accounts formuch of southern Eurasia. Second, although we cannot hopeto calibrate quantitative faunal data to plant consumption forthe Paleolithic in any region, what we can do is compare thefaunal record in quantity and character with the record ofdurable plant processing equipment in a more controlledmanner. The idea would be to look for evidence of heightenedprocessing investments in plant use. Scattered reports, mainlyanecdotal, suggest that ground-stone artifacts are fairly com-mon in some MSA records. Given that these artifacts are nota notable part of MP records, here lies a potential contrastwaiting for systematic consideration. Of course small MSAmilling equipment suitable for preparing mineral and organicpigments must be distinguished from large implements suit-able for seed processing, but experimental evidence providessome expectations. Such a strategy might provide an alter-native way of looking at the problem of how and why theMP and MSA seem different from one another and assessingthe independence of their evolutionary trajectories.

Acknowledgments

This paper is an outcome of the 2012 Wenner-Gren Sym-posium held in Sweden and titled “Alternative Paths to Com-plexity: Evolutionary Trajectories in the Middle Paleolithicand Middle Stone Age.” I am deeply grateful to the organizersof this conference, Erella Hovers and Steve Kuhn, and toWenner-Gren colleagues Leslie Aiello and Laurie Obbink fortheir many efforts in making this one of the most stimulatingand enriching symposia in which I have had the honor toparticipate. I also thank Jamie Clark and two anonymousCurrent Anthropology reviewers for their thoughtful and crit-ical remarks on an earlier version of this paper that certainlyhelped improve it.

References CitedAdler, D. S., and G. Bar-Oz. 2009. Seasonal patterns of prey acquisition and

inter-group competition during the Middle and Upper Palaeolithic of thesouthern Caucasus. In Evolution of human diets: integrating approaches tothe study of Palaeolithic subsistence. J.-J. Hublin and M. Richards, eds. Pp.127–140. Dordrecht: Springer.

Adler, D. S., G. Bar-Oz, A. Belfer-Cohen, and O. Bar-Yosef. 2006. Ahead of

the game: Middle and Upper Paleolithic hunting behaviors in the southernCaucasus. Current Anthropology 47:89–118.

Albert, R. M., O. Bar-Yosef, L. Meignen, and S. Weiner. 2003. Quantitativephytolith study of hearths from the Natufian and Middle Palaeolithic levelsof Hayonim Cave (Galilee, Israel). Journal of Archaeological Science 30:461–480.

Albert, R. M., F. Berna, and P. Goldberg. 2012. Insights on Neanderthal fireuse at Kebara Cave (Israel) through high-resolution study of prehistoriccombustion features: evidence from phytoliths and thin sections. Quater-nary International 247:278–293.

Albert, R. M., S. Weiner, O. Bar-Yosef, and L. Meignen. 2000. Phytoliths inthe Middle Palaeolithic deposits of Kebara Cave, Mt. Carmel, Israel: studyof the plant materials used for fuel and other purposes. Journal of Archae-ological Science 27:931–947.

Alperson-Afil, N., and N. Goren-Inbar. 2010. Ancient flames and controlleduse of fire, vol. 2 of The Acheulian site of Gesher Benot Ya’aqov. Dordrecht:Springer.

Anderson-Gerfaud, P. 1990. Aspects of behavior in the Middle Paleolithic:functional analysis of stone tools from southwest France. In The emergenceof modern humans: an archaeological perspective. P. Mellars, ed. Pp. 389–418. Edinburgh: Edinburgh University Press.

Assefa, Z. 2006. Faunal remains from Porc-Epic: paleoecological and zooar-chaeological investigations from a Middle Stone Age sites in southeasternEthiopia. Journal of Human Evolution 51(1):50–75.

Aura, J. E., V. B. Villaverde, M. Perez Ripoll, R. Martinez Valle, and P. Cal-atayud Guillem. 2002. Big game and small prey: Paleolithic and Epipaleo-lithic economy from Valencia (Spain). Journal of Archaeological Method andTheory 9(3):215–268.

Baden-Powell, D. F. W. 1964. Gorham’s Cave, Gibraltar: report on the climaticequivalent of the marine Mollusca. In The excavation of Gorham’s Cave:Gibraltar, 1951–1954. J. Waechter, ed. Pp. 216–218. Bulletin of the Instituteof Archaeology 4. London: University of London.

Bailey, G. N., and N. C. Flemming. 2008. Archaeology of the continental shelf:marine resources, submerged landscapes and underwater archaeology. Qua-ternary Science Review 27:2153–2165.

Barton, R. N. E. 2000. Mousterian hearths and shellfish: late Neanderthalactivities on Gibraltar. In Neanderthals on the edge: papers presented at theconference marking the 150th anniversary of the Forbes’ Quarry discovery,Gibraltar. C. B. Stringer, R. N. E. Barton, and J. C. Finlayson, eds. Pp. 211–220. Oxford: Oxbow.

Barton, R. N. E., A. P. Currant, Y. Fernandez-Jalvo, J. C. Finlayson, P. Goldberg,R. Macphail, P. B. Pettitt, and C. B. Stringer. 1999. Gibraltar Neanderthalsand results of recent excavations in Gorham’s, Vanguard and Ibex Caves.Antiquity 73:13–23.

Bar-Yosef, O., H. Laville, L. Meignen, A. M. Tillier, B. Vandermeersch, B.Arensburg, A. Belfer-Cohen, P. Goldberg, Y. Rak, and E. Tchernov. 1988.La sepulture neanderthalienne de Kebara (unite XII). In L’homme de Ne-anderthal, vol. 5. M. Otte, ed. Pp. 17–24. Liege: ERAUL.

Bar-Yosef, O., B. Vandermeersch, B. Arensburg, A. Belfer-Cohen, P. Goldberg,H. Laville, L. Meignen, et al. 1992. The excavations in Kebara Cave, Mt.Carmel. Current Anthropology 33(5):497–550.

Bar-Yosef, O., B. Vandermeersch, B. Arensburg, P. Goldberg, H. Laville,L. Meignen, Y. Rak, E. Tchernov, and A.-M. Tillier. 1986. New dataon the origin of modern man in the Levant. Current Anthropology 27:63–64.

Beyries, S. 1987. Variabilite de l’industrie lithique asu Mousterien: approchefunctionelle sur quelques gisments francaises. BAR International Series 328.Oxford: British Archaeological Reports.

Binford, L. R. 1978. Nunamiut ethnoarchaeology. New York: Academic Press.———. 1981. Bones: ancient men and modern myths. New York: Academic

Press.———. 1984. Faunal remains from Klasies River Mouth. New York: Academic

Press.———. 1988. Etude taphonomique des restes fauniques de la grotte Vaufrey.

In La Grotte Vaufrey a Cenac et Saint-Julien (Dordogne): paleoenvironnements,chronologie et activites humaines. J.-P. Rigaud, ed. Pp. 535–563. Memoiresde la Societe Prehistorique Francaise, vol. 19. Paris: CNRS.

———. 1991. When the going gets tough, the tough get going: Nunamiutlocal groups, camping patterns, and economic organization. In Ethnoar-chaeological approaches to mobile campsites: hunter-gatherer and pastoralistcase studies. C. S. Gamble and W. A. Boismer, eds. Pp. 25–137. Ann Arbor,MI: International Monographs in Prehistory.

———. 1998. Hearth and home: the spatial analysis of ethnographically doc-

S300 Current Anthropology Volume 54, Supplement 8, December 2013

umented rock shelter occupations as a template for distinguishing betweenhuman and hominid use of sheltered space. In Thirteenth UISPP CongressProceedings: Forli, 8–14 September 1996. N. J. Conard and F. Wendorf, eds.Pp. 229–239. Forli, Italy: ABACO.

Blanc, A. C. 1958–1961. Industria musteriana su calcare e su valve di Meretrixchione associata con fossili di elefante e rinocerante: nuovi giacimenti costieridel Capo de Leuca. Quaternaria 5:308–313.

Blasco, R., and J. Fernandez Peris. 2009. Middle Pleistocene bird consumptionat level XI of Bolomor Cave (Valencia, Spain). Journal of ArchaeologicalScience 36:2213–2223.

———. 2012. A uniquely broad spectrum diet during the Middle Pleis-tocene at Bolomor Cave (Valencia, Spain). Quaternary International252:16–31.

Blondel, J., and J. Aronson. 1999. Biology and wildlife of the Mediterraneanregion. Oxford: Oxford University Press.

Bocherens, H. 2009. Neanderthal dietary habits: review of the isotopic evi-dence. In Evolution of human diets: integrating approaches to the study ofPalaeolithic subsistence. J.-J. Hublin and M. Richards, eds. Pp. 214–250.Dordrecht: Springer.

Boeda, E. 1994. Le concept Levallois: variabilite des methodes. Paris: CNRS.Boeda, E., J. Connan, D. Dessort, S. Muhesen, N. Mercier, H. Valladas, and

N. Tisnerat. 1996. Bitumen as a hafting material on Middle Palaeolithicartefacts. Nature 380:336–338.

Boone, J. 2002. Subsistence strategies and early human population history:an evolutionary ecological perspective. World Archaeology 34:6–25.

Bourguignon, L., J. P. Faivre, and A. Turq. 2004. Ramification des chaınesoperatoires: une specificite du Mousterien? Paleo 16:37–48.

Bouzouggar A., R. N. E. Barton, M. Vanhaeren, F. D’errico, S. Collcutt,T. Higham, E. Hodge, et al. 2007. 82,000-year-old shell beads fromNorth Africa and implications for the origins of modern human be-havior. Proceedings of the National Academy of Sciences of the USA104(24):9964–9969.

Boyle, K. V. 2000. Reconstructing Middle Palaeolithic subsistence strategiesin the south of France. International Journal of Osteoarchaeology 10:336–356.

Brooks, A. 1984. San land-use patterns, past and present: implications forsouthern African prehistory. In Frontiers: southern African archaeology today.M. Hall, G. Avery, D. M. Avery, M. L. Wilson, and A. J. B. Humphreys,eds. Pp. 40–52. BAR International Series 207. Oxford: British ArchaeologicalReports.

Brugal, J.-P., and J. Jaubert. 1991. Les gisements paleontologiques Pleistocenesa indices de frequentation humaine: un nouveau type de comportementde predation? Paleo 3:15–41.

Burke, A. 2000. The view from Starosele: faunal exploitation at a MiddlePalaeolithic site in western Crimea. International Journal of Osteoarchaeology10:325–335.

Chase, P. G. 1986. The hunters of Combe Grenal: approaches to Middle Paleolithicsubsistence in Europe. BAR International Series 286. Oxford: British Ar-chaeological Reports.

———. 1989. How different was Middle Palaeolithic subsistence? a zooar-chaeological perspective on the Middle to Upper Palaeolithic transition. InThe human revolution: behavioural and biological perspectives on the originsof modern humans. P. Mellars and C. Stringer, eds. Pp. 321–337. Princeton,NJ: Princeton University Press.

Clark, J. L., and A. W. Kandel. 2013. The evolutionary implications of variationin human hunting strategies and diet breadth during the Middle Stone Ageof southern Africa. Current Anthropology 54(suppl. 8):S269–S287.

Cochard, D., J.-P. Brugal, E. Morin, and L. Meignen. 2012. Evidence of smallfast game exploitation in the Middle Paleolithic of Les Canalettes Aveyron,France. Quaternary International 264:32–51.

Colonese, A., M. A. Mannino, D. E. Bar-Yosef Mayer, D. A. Fa, C. J.Finlayson, D. Lubell, and M. C. Stiner. 2011. Marine mollusc exploi-tation in Mediterranean prehistory: an overview. Quaternary Interna-tional 239:86–103.

Conard, N. J., M. Bolus, and S. C. Munzel. 2012. Middle Paleolithic land use,spatial organization and settlement intensity in the Swabian Jura, south-western Germany. Quaternary International 247:236–245.

Conard, N. J., and T. J. Prindiville. 2000. Middle Palaeolithic hunting econ-omies in the Rhineland. International Journal of Osteoarchaeology 10:286–309.

Costamagno, S., and V. Laroulandie. 2004. L’exploitation des petits vertebresdans les Pyrenees francaises du Paleolithique au Mesolithique: un inventairetaphonomique et archeozoologique. In Petits animaux et societes humaines:

du complement alimentaire aux ressources utilitaires. J.-P. Brugal and J. Desse,eds. Pp. 369–382. Antibes: APCDA.

Darlas, A. 2007. Le Mousterien de Grece a la lumiere des recentes recherches.L’anthropologie 111:346–366.

David, F., and C. Farizy. 1999. Mauran (Haute-Garonne, France). In The roleof early humans in the accumulation of European Lower and Middle Palaeo-lithic bone assemblages (Ergebisse eines Kolloquiums). Pp. 267–277. Mono-graphien des Romisch-Germanischen Zentralmuseums 42. Mainz: Rom-isch-Germanischen Zentralmuseums.

David, F., and T. Poulain. 1990. La faune de grands mammiferes des niveauxXI et XC de la Grotte du Renne a Arcy-sur-Cure (Yonne): etude prelimi-naire. In Paleolithique Moyen Recent et Paleolithique Superieur Ancien enEurope. C. Farizy, ed. Pp. 319–323. Memoires du Musee de Prehistoire d’Ilede France 3. Ile de France: Editions de l’Association pour la Promotion dela Recherche Archeologique.

Delage, C., L. Meignen, and O. Bar-Yosef. 2000. Chert procurement and theorganization of lithic production in the Mousterian of Hayonim Cave (Is-rael). Journal of Human Evolution 38:A10–A11.

Delagnes, A., and L. Meignen. 2006. Diversity of lithic production systemsduring the Middle Paleolithic in France: are there any chronological trends?In Transitions before the transition: evolution and stability in the MiddlePaleolithic and Middle Stone Age. E. Hovers and S. L. Kuhn, eds. Pp. 85–107. Santa Barbara, CA: Springer.

Delagnes, A., and W. Rendu. 2011. Shifts in Neandertal mobility, technologyand subsistence strategies in western France. Journal of Archaeological Science38:1771–1783.

d’Errico, F., C. Henshilwood, M. Vanhaeren, and K. Van Niekerke. 2005.Nassarius kraussianus shell beads from Blombos Cave: evidence forsymbolic behaviour in the Middle Stone Age. Journal of Human Evo-lution 48:3–24.

Discamps, E. 2011. Hommes et hyenes face aux recompositions des com-munautes d’Ongules (MIS 5–3): elements pour un cadre paleoecologiquedes societes du Paleolithique moyen et superieur ancien d’Europe del’Ouest. PhD dissertation, Universite Bordeaux 1.

Erlandson, J. M., and M. L. Moss. 2001. Shellfish feeders, carrion eaters, andthe archaeology of aquatic adaptations. American Antiquity 66(3):413–432.

Farizy, C., F. David, and J. Jaubert, eds. 1994. Hommes et bisons du PaleolithiqueMoyen a Mauran (Haute Garonne). Gallia prehistoire, suppl. 30. Paris:CNRS.

Fernandez-Jalvo, Y., and P. Andrews. 2000. The taphonomy of Pleistocenecaves, with particular reference to Gibraltar. In Neanderthals on the edge:papers presented at the conference marking the 150th anniversary of the Forbes’Quarry discovery, Gibraltar. C. B. Stringer, R. N. E. Barton, and J. C. Fin-layson, eds. Pp. 171–182. Oxford: Oxbow.

Fernandez Peris, J., V. Barciela Gonzalez, R. Blasco, F. Cuartero, H. Fluck, P.Sanudo Die, and C. Verdasco. 2012. The earliest evidence of hearths insouthern Europe: the case of Bolomor Cave (Valencia, Spain). QuaternaryInternational 247:267–277.

Finlayson, C., A. F. Darren, F. J. Jimenez-Espejo, J. S. Carrion, G. Finlayson,F. Giles Pacheco, J. Rodrıguez Vidal, C. B. Stringer, and F. Martınez. 2008.Gorham’s cave, Gibraltar: the persistence of a Neanderthal population.Quaternary International 181:64–71.

Foley, R., and C. Gamble. 2009. The ecology of social transitions in humanevolution. Philosophical Transactions of the Royal Society B 364:3267–3279.

Freeman, L. G. 1981. The fat of the land: notes on Paleolithic diet in Iberia.In Omnivorous primates: gathering and hunting in human evolution. R. S.O. Harding and G. Teleki, eds. Pp. 104–165. New York: Columbia UniversityPress.

Galanidou, N. 2000. Patterns in caves: foragers, horticulturalists, and the useof space. Journal of Anthropological Archaeology 19:243–275.

Gamble, C. 1986. The Palaeolithic settlement of Europe. Cambridge: CambridgeUniversity Press.

———. 1999. The Hohlenstein-Stadel revisited: the role of early humans inthe accumulation of European Lower and Middle Palaeolithic bone assem-blages. In The role of early humans in the accumulation of European Lowerand Middle Palaeolithic bone assemblages (Ergebisse eines Kolloquiums). Pp.305–324. Monographien des Romisch-Germanischen Zentralmuseums 42.Mainz: Romisch-Germanischen Zentralmuseums.

Garrard, A. N. 1983. The Palaeolithic faunal remains from Adlun andtheir ecological context. In Adlun in the Stone Age: the excavations ofD. A. E. Garrod in the Lebanon, 1958–1963. D. A. Roe, ed. Pp. 397–

Stiner Trends versus Conservatism in the Predatory Niche S301

413. BAR International Series 159(ii). Cambridge: British Archaeolog-ical Reports.

Garrod, D. A. E., L. H. D. Buxton, G. E. Smith, D. M. A. Bate, R. C. Spiller,M. A. C. Hinton, and P. Fischer. 1928. Excavation of a Mousterian rock-shelter at Devil’s Tower, Gibraltar. Journal of the Royal Anthropological In-stitute of Great Britain and Ireland 58:33–113.

Gaudzinski, S. 1995. Wallertheim revisited: a reanalysis of the fauna from theMiddle Palaeolithic site of Wallertheim (Rheinhessen/Germany). Journal ofArchaeological Science 22:51–66.

———. 2000. On the variability of Middle Palaeolithic procurement tactics:the case of Salzgitter Lebenstedt, Northern Germany. International Journalof Osteoarchaeology 10:396–406.

———. 2004. A matter of high resolution? the Eemian Interglacial (OIS 5e)in north-central Europe and Middle Palaeolithic subsistence. InternationalJournal of Osteoarchaeology 14(3/4):201–211.

Gaudzinski, S., and W. Roebroeks. 2000. Adults only: reindeer hunting at theMiddle Palaeolithic site Salzgitter Lebenstedt, northern Germany. Journalof Human Evolution 38:497–521.

Gaudzinski-Windheuser, S., and L. Niven. 2009. Hominin subsistencepatterns during the Middle and Late Paleolithic in northwestern Eu-rope. In Evolution of human diets: integrating approaches to the studyof Palaeolithic subsistence. J.-J. Hublin and M. Richards, eds. Pp. 99–111. Dordrecht: Springer.

Goldberg, P., and O. Bar-Yosef. 1998. Site formation processes in Kebara andHayonim caves and their significance in Levantine prehistoric caves. InNeanderthals and modern humans in western Asia. T. Akazawa, K. Aoki, andO. Bar-Yosef, eds. Pp. 107–125. New York: Plenum.

Goldberg, P., H. Dibble, F. Berna, D. Sandgathe, S. J. P. McPherron, andA. Turq. 2012. New evidence on Neandertal use of fire: examples fromRoc de Marsal and Pech de l’Aze IV. Quaternary International 247:325–340.

Gopher, A., R. Barkai, R. Shimelmitz, M. Khalaily, C. Lemorini, I. Heshkovitz,and M. C. Stiner. 2005. Qesem Cave: an Amudian site in central Israel.Journal of the Israel Prehistory Society 35:69–92.

Gorecki, P. P. 1991. Horticulturalists as hunter-gatherers: rock shelterusage in Papua New Guinea. In Ethnoarchaeological approaches to mo-bile campsites: hunter-gatherer and pastoralist case studies. C. S. Gambleand W. A. Boismer, eds. Pp. 237–262. Ann Arbor, MI: InternationalMonographs in Prehistory.

Gowlett, J. A. J., J. Hallos, S. Hounsell, V. Brant, and N. C. Debenham. 2005.Beeches Pit: archaeology, assemblage dynamics and early fire history of aMiddle Pleistocene site in East Anglia, UK. Journal of Eurasian Prehistory3:3–38.

Grayson, D. K. and F. Delpech. 1994. The evidence for Middle Palaeolithicscavenging from Couche VIII, Grotte Vaufrey (Dordogne, France). Journalof Archaeological Science 21:359–375.

———. 1998. Changing diet breadth in the early Upper Palaeolithic of south-western France. Journal of Archaeological Science 25:1119–1129.

———. 2003. Ungulates and the Middle-to-Upper Paleolithic transition atGrotte XVI (Dordogne, France). Journal of Archaeological Science 30:1633–1648.

Griggo, C. 2004. Mousterian fauna from Dederiyeh Cave and comparisonswith fauna from Umm El Tlel and Douara Cave. Paleorient 30(1):149–162.

Haydal, J., T. Oguchi, T. Akazawa, S. Muhesen, Y. Dodo, Y. Mizoguchi, Y.Abe, Y. Nishiaki, and S. Ohta. 1995. Neanderthal infant burial from theDederiyeh cave in Syria. Paleorient 21:77–86.

Henrich, J. 2004. Demography and cultural evolution: how adaptive culturalprocesses can produce maladaptive losses: the Tasmanian case. AmericanAntiquity 69(2):197–214.

Henry, A. G., A. S. Brooks, and D. R. Piperno. 2011. Microfossils in calculusdemonstrate consumption of plants and cooked foods in Neanderthal diets(Shanidar III, Iraq; Spy I and II, Belgium). Proceedings of the NationalAcademy of Sciences of the USA 108(2):486–491.

Henshilwood, C., F. d’Errico, M. Vanhaeren, K. Van Niekerk, and Z.Jacobs. 2004. Middle Stone Age shell beads from South Africa. Science304:404.

Hoffecker, J. F., G. Baryshnikov, and O. Potapova. 1991. Vertebrate re-mains from the Mousterian site of Il’skaya I (northern Caucasus,USSR): new analysis and interpretation. Journal of Archaeological Sci-ence 18:113–147.

Hoffecker, J. F., and N. Cleghorn. 2000. Mousterian hunting patterns in the

northwestern Caucasus and the ecology of the Neanderthals. InternationalJournal of Osteoarchaeology 10:368–378.

Hoffecker, J. R. 2009. Neanderthal and modern human diet in eastern Europe.In Evolution of human diets: integrating approaches to the study of Palaeolithicsubsistence. J.-J. Hublin and M. Richards, eds. Pp. 87–98. Dordrecht:Springer.

Hortola, P., and B. Martınez-Navarro. 2013. The Quaternary megafaunal ex-tinction and the fate of Neanderthals: an integrative working hypothesis.Quaternary International 295:69–72.

Jacob-Friesen, K. H. 1956. Eiszeitliche Elephantenjager in der LuneburgerHeide. Jarbuch des Romisch-Germanischen Zentralmuseums, Mainz 3:1–22.

Jaubert, J., M. Lorblanchet, H. Laville, R. Slott-Moller, A. Turq, and J.-P.Brugal. 1990. Les chasseurs d’aurochs de la Borde. Documents d’ArcheologieFrancaise, no. 27. Paris: Maison des Sciences de l’Homme.

Jones, M. 2009. Moving north: archaeobotanical evidence for plant diet inMiddle and Upper Paleolithic Europe. In Evolution of human diets: inte-grating approaches to the study of Palaeolithic subsistence. J.-J. Hublin andM. Richards, eds. Pp. 171–180. Dordrecht: Springer.

Karkanas, P., R. Shahack-Gross, A. Ayalon, M. Bar-Matthews, R. Barkai, A.Frumkin, A. Gopher, and M. C. Stiner. 2007. Evidence for habitual use offire at the end of the Lower Paleolithic: site formation processes at QesemCave, Israel. Journal of Human Evolution 53(2):197–212.

Keeley, L. H. 1995. Protoagricultural practices among hunter-gatherers: across-cultural survey. In Last hunters–first farmers: new perspectives on theprehistoric transition to agriculture. T. D. Price and A. B. Gebauer, eds. Pp.243–272. Santa Fe, NM: School of American Research.

Kelly, R. 1995. The foraging spectrum: diversity in hunter-gatherer lifeways.Washington, DC: Smithsonian Institution.

Klein, R. G. 1999. The human career: human biological and cultural origins.2nd edition. Chicago: University of Chicago Press.

Klein, R. G., and K. Scott. 1986. Reanalysis of faunal assemblages from theHaua Fteah and other Late Quaternary archaeological sites in CyrenaicanLibya. Journal of Archaeological Science 13:515–542.

Klein, R. G., and T. Steele. 2008. Gibraltar data are too sparse to inform onNeanderthal exploitation of coastal resources. Proceedings of the NationalAcademy of Sciences of the USA 105(51):E115.

Knecht, H. 1997. Projectile technology. New York: Plenum.Koller, J., U. Baumer, and D. Mania. 2001. High-tech in the Middle Palaeo-

lithic: Neandertal-manufactured pitch identified. European Journal of Ar-chaeology 4:385–397.

Kuhn, S. L. 1995. Mousterian lithic technology: an ecological perspective. Prince-ton, NJ: Princeton University Press.

———. 2013. Roots of the Middle Paleolithic in Eurasia. Current Anthropology54(suppl. 8):S255–S268.

Kuhn, S. L., and M. C. Stiner. 2006. What’s a mother to do? a hypothesisabout the division of labor among Neandertals and modern humans inEurasia. Current Anthropology 47:953–980.

———. 2010. Summary of findings and conclusion. In Culture and biologyat a crossroads: the Middle Pleistocene record of Yarimburgaz Cave (Thrace,Turkey). F. C. Howell, G. Arsebuk, S. L. Kuhn, M. Ozbasaran, and M. C.Stiner, eds. Pp. 301–319. Istanbul: Zero Books/Ege.

Lalueza-Fox, C., A. Rosas, A. Estalrrich, E. Gigli, P. F. Campos, A. Garcıa-Tabernero, S. Garcıa-Vargas, et al. 2011. Genetic evidence for partrilocalmating behavior among Neandertal groups. Proceedings of the NationalAcademy of Sciences of the USA 108(1):250–253.

Lam, Y. M., and O. M. Pearson. 2005. Bone density studies and the inter-pretation of the faunal record. Evolutionary Anthropology 14(3):99–108.

Lambeck, K. 1996. Sea-level change and shore-line evolution in Aegean Greecesince Upper Paleolithic time. Antiquity 70:588–611.

Laroulandie, V. 2004. Exploitation des ressources aviaires durant le Paleo-lithique en France: bilan critique et perspectives. In Petits animaux et societeshumaines: du complement alimentaire aux ressources utilitaires. J.-P. Brugaland J. Desse, eds. Pp. 163–172. Antibes: APCDA.

Lemorini, C. 2000. Reconnaıtre de tactiques d’exploitation du milieu au paleo-lithique moyen: la contribution de l’analyse fonctionelle: etude fonctionelle desindustries lithiques au Grotta Breuil (Latium, Italie) et de La Colombette(Bonnieux, Valcluse, France). BAR International Series 858. Oxford: BritishArchaeological Reports.

Lyman, R. L. 1984. Bone density and differential survivorship of fossil classes.Journal of Anthropological Archaeology 3:259–299.

Madella, M., E. Hovers, M. K. Jones, P. Goldberg, and Y. Goren. 2002. Theexploitation of plant resources by Neanderthals in Amud Cave (Israel): the

S302 Current Anthropology Volume 54, Supplement 8, December 2013

evidence from phytolith studies. Journal of Archaeological Science 29:703–719.

Marean, C. W., M. Bar-Matthews, J. Bernatchez, E. Fisher, P. Goldberg, A. I.R. Herries, Z. Jacobs, et al. 2007. Early human use of marine resources andpigment in South Africa during the Middle Pleistocene. Nature 449:905–909.

Martınez-Molina, K. 2005. Analisis functional de industrias lıticas del Plies-toceno superior: el Paleolıtico Medio del Abric Romanı (Capellades, Bar-celona) y del Paleolıtico Superior de Ucagızlı (Hatay, Turquıa) y del Molıdel Salt (Vimbodı, Tarragona): cambios en los patrones funcionales entreel Paleolıtico Medio y el Superior. PhD dissertation, Universitat Rovira IVirgili, Tarragona.

Martinson, D. G., N. G. Pisias, J. D. Hays, J. Imbrie, T. C. Moore, and N. J.Shackleton. 1987. Age dating and the orbital theory of the ice ages: devel-opment of a high-resolution 0 to 300,000-year chronostratigraphy. Qua-ternary Research 27:1–29.

McBrearty, S., and A. Brooks. 2000. The revolution that wasn’t: a new inter-pretation of the origin of modern human behavior. Journal of HumanEvolution 39:456–463.

Meignen, L., O. Bar-Yosef, and P. Goldberg. 1989. Les structures de com-bustion Mousteriennes de la grotte de Kebara (Mont Carmel, Israel). InNature et fonctions des foyers prehistoriques. M. Olive and Y. Taborin, eds.Pp. 141–146. Memoires du Musee de Prehistoire d’Ile de France, vol. 2.Nemours: APRAIF.

Meignen, L., O. Bar-Yosef, M. Stiner, S. Kuhn, P. Goldberg, and S. Weiner.2010. Apport des analyses mineralogiques (en spectrometrie infrarouge atransformee de fourier) a l’interpretation des structures anthropiques: lesconcentrations osseuses dans les niveaux Mousteriens des Grottes de Kebaraet Hayonim (Israel). PALEO 3(suppl.):93–107.

Meignen, L., J. D. Speth, and M. C. Stiner. 2006. Middle Paleolithic settlementpatterns in the Levant. In Transitions before the transition: evolution andstability in the Middle Paleolithic and Middle Stone Age. E. Hovers and S.L. Kuhn, eds. Pp. 149–169. New York: Kluwer.

Mellars, P. 1989. Major issues in the emergence of modern humans. CurrentAnthropology 30:349–385.

———. 1996. The Neanderthal legacy: an archaeological perspective from west-ern Europe. Princeton, NJ: Princeton University Press.

Mirazon Lahr, M., and R. Foley. 2003. Demography, dispersal and humanevolution in the Last Glacial period. In Neandertals and modern humans inthe European landscape during the Last Glaciation. T. H. Van Andel and W.Davies, eds. Pp. 241–256. Cambridge: McDonald Institute for Archaeolog-ical Research, University of Cambridge.

Moncel, M.-H., and C. Daujeard. 2012. The variability of the Middle Palaeo-lithic on the right bank of the Middle Rhone Valley (southeast France):technical traditions or functional choices? Quaternary International 247:103–124.

Muller-Beck, H. 1988. The ecosystem of the “Middle Paleolithic” (late LowerPaleolithic) in the Upper Danube region: a stepping stone to the UpperPaleolithic. In Upper Pleistocene prehistory in western Eurasia. H. Dibbleand A. Montet-White, eds. Pp. 233–254. Monograph 54. Philadelphia: Uni-versity Museum, University of Pennsylvania.

Munro, N. 2004. Zooarchaeological measures of hunting pressure and oc-cupation intensity in the Natufian. Current Anthropology 45(suppl.):S5–S33.

Munzel, S., and N. J. Conard. 2004. Change and continuity in subsistenceduring the Middle and Upper Palaeolithic in the Ach Valley of Swabia(south-west Germany). International Journal of Osteoarchaeology 14:225–243.

Oswalt, W. H. 1976. An anthropological analysis of food-getting technology. NewYork: Wiley.

Palma di Cesnola, A. 1965. Notizie preliminari sulla terza campagna di scavinella Grotta del Cavallo (Lecce). Rivista di Scienze Preistoriche 25:3–87.

———. 1969. Il musteriano della Grotta del Poggio a Marina di Camerota(Salerno). In Estratto dagli scritti sul Quaternario in onore di Angelo Pasa.Pp. 95–135. Verona: Museo Civico di Storia Naturale.

Patou-Mathis, M. 1993. Etude taphonomique et paleoethnologique de la faunede l’Abri des Canalettes. In L’Abri des Canalettes: un habitat Mousterien surles Grands Causses (Nant, Aveyron). L. Meignen, ed. Pp. 199–237. Paris:CNRS.

———. 2000. Neanderthal subsistence behaviours in Europe. InternationalJournal of Osteoarchaeology 10:379–395.

Peresani, M. 2012. Fifty thousand years of flint knapping and tool shapingacross the Mousterian and Uluzzian sequence of Fumane Cave. QuaternaryInternational 247:125–150.

Peresani, M., I. Fiore, M. Gala, M. Romandini, and A. Tagliacozzo. 2011. LateNeandertals and the intentional removal of feathers as evidenced from birdbone taphonomy at Fumane Cave 44 ky B.P., Italy. Proceedings of the Na-tional Academy of Sciences of the USA 108:3888–3893.

Pianka, E. R. 1978. Evolutionary ecology. 2nd edition. New York: Harper &Row.

Pike-Tay, A., V. Cabrera Valdes, and F. Bernaldo de Quiros. 1999. Seasonalvariations of the Middle–Upper Paleolithic transition at El Castillo, CuevaMorin and El Pendo (Cantabria, Spain). Journal of Human Evolution 36:283–317.

Plisson, H., and S. Beyries. 1998. Pointes ou outils triangulaires? donneesfonctionnelles dans le Mousterien Levantin. Paleorient 24(1):5–24.

Poteete, A. R., and E. Ostrom. 2004. Heterogeneity, group size and collectiveaction: the role of institutions in forest management. Development andChange 35(3):435–461.

Preece, R. C., J. A. J. Gowlett, S. A. Parfitt, D. R. Bridgland, and S. G. Lewis.2006. Humans in the Hoxnian: habitat, context and fire use at Beeches Pit,West Stow, Suffolk, UK. Journal of Quaternary Science 21:485–496.

Premo, L. S., and S. L. Kuhn. 2010. Modeling effects of local extinctions onculture change and diversity in the Paleolithic. PLoS ONE 5:15582.

Rabinovich, R., S. Gaudzinski-Windheuser, and N. Goren-Inbar. 2008. Sys-tematic butchering of fallow deer (Dama) at the early middle PleistoceneAcheulian site of Gesher Benot Ya‘aqov (Israel). Journal of Human Evolution54(1):134–149.

Rabinovich, R., and E. Tchernov. 1995. Chronological, paleoecological andtaphonomical aspects of the Middle Paleolithic site of Qafzeh, Israel. InArchaeozoology of the Near East, vol. 2. H. Buitenhuis and H.-P. Uerpmann,eds. Pp. 45–71. Leiden: Backhuys.

Rendu, W. 2010. Hunting behavior and Neanderthal adaptability in the LatePleistocene site of Pech-de-l’Aze I. Journal of Archaeological Science 37:1798–1810.

Rendu, W., S. Costamagno, L. Meignen, and M. C. Soulier. 2012. Monospecificfaunal spectra in Mousterian contexts: implications for social behavior.Quaternary International 247:50–58.

Reynaud Savioz, N., and P. Morel. 2005. La faune de Nadaouiyeh Aın Askar(Syrie centrale, pleistocene moyen): apercu et perspectives. Revue Paleo-biologie Geneve 10:31–35.

Riel-Salvatore, J., and C. M. Barton. 2004. Late Pleistocene technology, eco-nomic behavior, and land-use dynamics in southern Italy. American Antiq-uity 69(2):257–274.

Rightmire, G. P. 2003. Brain size and encephalization in early to mid-Pleistocene Homo. American Journal of Physical Anthropology 124(2):109–123.

Roebroeks, W., M. J. Sier, T. Kellberg Nielsen, D. de Loecker, J. M. Pares, C.E. S. Arps, and H. J. Mucher. 2012. Use of red ochre by early Neandertals.Proceedings of the National Academy of Sciences of the USA 109(6):1889–1894.

Roebroeks, W., and P. Villa. 2011. On the earliest evidence for habitual useof fire in Europe. Proceedings of the National Academy of Sciences of theUSA 108:5209–5214.

Rolland, N. 2004. Was the emergence of home bases and domestic fire apunctuated event? a review of the Middle Pleistocene record in Eurasia.Asian Perspectives 43:248–280.

Rosell, J., I. Caceres, R. Blasco, M. Bennasar, P. Bravo, G. Campeny, M.Esteban-Nadal, et al. 2012. A zooarchaeological contribution to establishoccupational patterns at Level J of Abric Romanı (Barcelona, Spain). Qua-ternary International 247:69–84.

Sandgathe, D. M., H. L. Dibble, P. Goldberg, S. P. Mcpherron, A. Turq, L.Niven, and J. Hodgkins. 2011. On the role of fire in Neandertal adaptationsin western Europe: evidence from Pech de l’Aze IV and Roc de Marsal,France. Paleoanthropology 2011:216�242.

Shea, J. 1989. A functional study of the lithic industries associated with hom-inid fossils in the Kebara and Qafzeh caves, Israel. In The human revolution.P. Mellars and C. Stringer, eds. Pp. 611–625. Princeton, NJ: PrincetonUniversity Press.

———. 1991. The behavioral significance of Levantine Mousterian industrialvariability. PhD dissertation, Harvard University, Cambridge, MA.

———. 1993. Lithic use-wear evidence for hunting by Neandertals and earlymodern humans from the Levantine Mousterian. In Hunting and animalexploitation in the later Palaeolithic and Mesolithic of Eurasia. G. L. Peterkin,H. Bricker, and P. Mellars, eds. Pp. 189–197. Archaeological Papers of theAmerican Anthropological Association, vol. 4. Washington, DC: AmericanAnthropological Association.

Stiner Trends versus Conservatism in the Predatory Niche S303

———. 2009. The impact of projectile weaponry on Late Pleistocene homininevolution. In Evolution of human diets: integrating approaches to the studyof Palaeolithic subsistence. J.-J. Hublin and M. Richards, eds. Pp. 189–199.Dordrecht: Springer.

Siddall, M., J. Chappell, and E.-K. Potter. 2006. Eustatic sea level during pastinterglacials. In The climate of past interglacials. F. Sirocko, T. Litt, M. Claus-sen, and M.-F. Sanchez-Goni, eds. Pp. 75–92. Developments in QuaternaryScience 7. Amsterdam: Elsevier.

Siddall, M., E. Rohling, A. Almogi-Labin, C. Hemleben, D. Meischner, I.Schmelzer, and D. A. Smeed. 2003. Sea-level fluctuations during the lastglacial cycle. Nature 423(6942):853–858.

Snodgrass, J. J., W. R. Leonard, and M. L. Robertson. 2009. The energeticsof encephalization in early hominids. In Evolution of human diets: integratingapproaches to the study of Palaeolithic subsistence. J.-J. Hublin and M. Rich-ards, eds. Pp. 15–29. Dordrecht: Springer.

Speth, J. D. 2006. Housekeeping, Neandertal-style: hearth placement and mid-den formation in Kebara Cave (Israel). In Transitions before the transition:evolution and stability in the Middle Paleolithic and Middle Stone Age. E.Hovers and S. L. Kuhn, eds. Pp. 171–188. New York: Springer.

Speth J. D., and J. L. Clark. 2006. Hunting and overhunting in the Levantinelate Middle Palaeolithic. Before Farming 3:1–42.

Speth, J. D., and E. Tchernov. 1998. The role of hunting and scavenging inNeanderthal procurement strategies: new evidence from Kebara Cave (Is-rael). In Neanderthals and modern humans in west Asia. T. Akazawa, K.Aoki, and O. Bar-Yosef, eds. Pp. 223–239. New York: Plenum.

———. 2001. Neandertal hunting and meat-processing in the Near East:evidence from Kebara Cave (Israel). In The early human diet: the role ofmeat. C. Stanford and H. Bunn, eds. Pp. 52–72. Oxford: Oxford UniversityPress.

———. 2002. Middle Paleolithic tortoise use at Kebara Cave (Israel). Journalof Archaeological Science 29(5):471–483.

———. 2007. The Middle Paleolithic occupations at Kebara Cave: a faunalperspective. In Kebara Cave Mt. Carmel, Israel: the Middle Paleolithic ar-chaeology, part 1. O. Bar-Yosef and L. Meignen, eds. Pp. 165–260. AmericanSchool of Prehistoric Research Bulletin 49. Cambridge, MA: Peabody Mu-seum.

Starkovich, B. 2011. Subsistence change during the Middle to Upper Paleo-lithic at Klissoura Cave 1 (Peloponnese, Greece). PhD dissertation, Uni-versity of Arizona, Tucson.

Steele, T. 2004. Variation in mortality profiles of red deer (Cervus elaphus) inMiddle Palaeolithic assemblages from western Europe. International Journalof Osteoarchaeology 14:307–320.

Steele, T., and E. Alverez-Fernandez. 2011. Initial investigations into the ex-ploitation of coastal resources in North Africa during the Late Pleistoceneat Grottes des Contrebandiers, Morocco. In Trekking the shore: changingcoastlines and the antiquity of coastal settlement. N. F. Bicho, J. A. Haws,and L. G. Davis, eds. Pp. 383–403. New York: Springer.

Steele, T. E., and R. G. Klein. 2009. Late Pleistocene subsistence strategies andresource intensification in Africa. In Evolution of human diets: integratingapproaches to the study of Palaeolithic subsistence. J.-J. Hublin and M. Rich-ards, eds. Pp. 113–126. Dordrecht: Springer.

Stephens, D. W., and J. R. Krebs. 1986. Foraging theory. Princeton, NJ: Prince-ton University Press.

Stiner, M. C. 1990. The use of mortality patterns in archaeological studies ofhominid predatory adaptations. Journal of Anthropological Archaeology 9:305–351.

———. 1994. Honor among thieves: a zooarchaeological study of Neandertalecology. Princeton, NJ: Princeton University Press.

———. 2001. Thirty years on the “broad spectrum revolution” and Palaeo-lithic demography. Proceedings of the National Academy of Sciences of theUSA 98(13):6993–6996.

———. 2002. On in situ attrition and vertebrate body part profiles. Journalof Archaeological Science 29:979–991.

———. 2005. The faunas of Hayonim Cave (Israel): a 200,000-year record ofPaleolithic diet, demography and society. American School of PrehistoricResearch Bulletin 48. Cambridge, MA: Peabody Museum.

———. 2009. Prey choice, site occupation intensity and economic diversityacross the Middle to early Upper Palaeolithic at Ucagızlı Caves I and II(Hatay, Turkey). Before Farming. http://www.waspress.co.uk/journals/beforefarming/journal_20093/abstracts/index.php#200933.

Stiner, M. C., G. Arsebuk, and F. C. Howell. 1996. Cave bears and Paleolithicartifacts in Yarımburgaz Cave, Turkey: dissecting a palimpsest. Geoar-chaeology 11(4):279–327.

Stiner, M. C., R. Barkai, and A. Gopher. 2011. Hearth-side socioeconomics,hunting and paleoecology during the late Lower Paleolithic at Qesem Cave,Israel. Journal of Human Evolution 60(2):213–233.

Stiner, M. C., and S. L. Kuhn. 1992. Subsistence, technology, and adaptivevariation in Middle Paleolithic Italy. American Anthropologist 94:12–46.

———. 2006. Changes in the “connectedness” and resilience of Paleo-lithic societies in Mediterranean ecosystems. Human Ecology 34(5):693–712.

Stiner, M. C., S. L. Kuhn, T. A. Surovell, P. Goldberg, L. Meignen, S. Weiner,and O. Bar-Yosef. 2001. Bone preservation in Hayonim Cave (Israel): amacroscopic and mineralogical study. Journal of Archaeological Science 28:643–659.

Stiner, M. C., and N. D. Munro. 2011. On the evolution of Paleolithic dietand landscape at Franchthi Cave (Peloponnese, Greece). Journal of HumanEvolution 60:618–636.

Stiner, M. C., N. D. Munro, and T. A. Surovell. 2000. The tortoise and thehare: small game use, the broad spectrum revolution, and Palaeolithic de-mography. Current Anthropology 41:39–73.

Stiner, M. C., N. D. Munro, T. A. Surovell, E. Tchernov, and O. Bar-Yosef.1999. Paleolithic population growth pulses evidenced by small animal ex-ploitation. Science 283:190–194.

Straus, L. G. 1982. Carnivores and cave sites in Cantabrian Spain. Journal ofAnthropological Research 38:75–96.

Stringer, C. B., J. C. Finlayson, R. N. E. Barton, Y. Fernandez-Jalvo, I. Caceres,R. C. Sabin, E. J. Rhodes, et al. 2008. Neanderthal exploitation of marinemammals in Gibraltar. Proceedings of the National Academy of Sciences ofthe USA 105(38):14319–14324.

Stuart, A. J. 1991. Mammalian extinctions in the Late Pleistocene of northernEurasia and North America. Biological Reviews 66:453–562.

Thieme, H. 1997. Lower Palaeolithic hunting spears from Germany. Nature385:807–810.

Valensi, P. 2000. The archaeozoology of Lazaret Cave (Nice, France). Inter-national Journal of Osteoarchaeology 10:357–367.

Valensi, P., and E. Psathi. 2004. Faunal exploitation during the Middle Pa-laeolithic in south-eastern France and north-western Italy. InternationalJournal of Osteoarchaeology 14:256–272.

Vallverdu, J., S. Alonso, A. Bargallo, R. Bartrolı, G. Campeny, A. Car-rancho, I. Exposito, et al. 2012. Combustion structures of archaeolog-ical level O and Mousterian activity areas with use of fire at the AbricRomanı rockshelter (NE Iberian Peninsula). Quaternary International247:313–324.

Van Andel, T. H., W. Davies, B. Weninger, and O. Joris. 2003. Archaeologicaldates as proxies for the spatial and temporal human presence in Europe:a discourse on the method. In Neandertals and modern humans in theEuropean landscape during the Last Glaciation. T. H. Van Andel and W.Davies, eds. Pp. 21–29. Cambridge: McDonald Institute for ArchaeologicalResearch, University of Cambridge.

Van Andel T. H., and P. C. Tzedakis. 1996. Palaeolithic landscape of Europeand environs, 150,000–25,000 years ago: an overview. Quaternary ResearchReview 15:481–500.

Vandermeersch, B. 1989. The evolution of modern humans: recent evidencefrom southwest Asia. In The human revolution: behavioural and biologicalperspectives on the origins of modern humans. P. Mellars and C. B. Stringer,eds. Pp. 155–164. Edinburgh: Edinburgh University Press.

Vaquero, M., M. G. Chacon, M. D. Garcia-Anton, B. Gomez de Soler, K.Martinez, and F. Cuartero. 2012. Time and space in the formation of lithicassemblages: the example of Abric Romanı Level J. Quaternary International247:162–181.

Vaquero, M., and I. Pasto. 2001. The definition of spatial units in MiddlePalaeolithic sites: the hearth-related assemblages. Journal of ArchaeologicalScience 28:1209–1220.

Villa, P. 1990. Torralba and Aridos: elephant exploitation in Middle PleistoceneSpain. Journal of Human Evolution 19:299–309.

Villa, P., F. Bon, and J.-C. Castel. 2001. Fuel, fire and fireplaces in the Paleolithicof western Europe: review of Economie des combustibles au Paleolithique, I,by Thery-Parisot, ed. Review of Archaeology 23(1):33–42.

Villa, P., and M. Soressi. 2000. Stone tools in carnivore sites: the case of BoisRoche. Journal of Anthropological Research 56:187–215.

Vitagliano, S. 1984. Nota sul Pontiniano della Grotta dei Moscerini, Gaeta(Latina). Il paleolitico e il mesolitico nel Lazio: atti della XXIV RiunioneScientifica. Pp. 155–164. Florence: Istituto Italiano di Preistoria e Proto-storia.

S304 Current Anthropology Volume 54, Supplement 8, December 2013

Weiner, S., P. Goldberg, and O. Bar-Yosef. 2002. Three-dimensional distri-bution of minerals in the sediments of Hayonim Cave, Israel: diageneticprocesses and archaeological implications. Journal of Archaeological Science29(11):1289–1308.

Yellen, J. E., A. S. Brooks, E. Cornelissen, R. G. Klein, M. Mehlman, and K.Stewart. 1995. A Middle Stone Age worked bone industry from Katanda,upper Semliki River valley (Kivu), Zaire. Science 268:553–556.

Yellen, J., A. S. Brooks, D. M. Helgren, M. Tappen, S. H. Ambrose, R. Bonne-

fille, J. Feathers, et al. 2005. The archaeology of Aduma Middle Stone Agesites in the Awash Valley, Ethiopia. Paleoanthropology 10:25–100.

Yeshurun, R., G. Bar-Oz, and M. Weinstein-Evron. 2007. Modern huntingbehavior in the early Middle Paleolithic: faunal remains from Misliya Cave,Mount Carmel, Israel. Journal of Human Evolution 53(6):656–677.

Zilhao, J., D. E. Angelucci, E. Badal-Garcıa, F. D’errico, F. Daniel, L. Dayet,K. Douka, et al. 2010. Symbolic use of marine shells and mineral pigmentsby Iberian Neandertals. Proceedings of the National Academy of Sciences ofthe USA 107(3):1023–1028.

Current Anthropology Volume 54, Supplement 8, December 2013 S305

� 2013 by The Wenner-Gren Foundation for Anthropological Research. All rights reserved. 0011-3204/2013/54S8-0011$10.00. DOI: 10.1086/673283

Technological Trends in the Middle Stone Ageof South Africa between MIS 7 and MIS 3

by Sarah Wurz

The range of technological elements that marks the Middle Stone Age originated more than 300,000 years ago andformed the basic tool kit for an extended period of time. No spatial and chronological patterns can be identifiedfrom the Early Middle Stone Age until marine isotope stage (MIS) 5, and there is no cumulative trend of increasingcomplexity and diversity; instead, periods of complexity come and go. The Howiesons Poort and Still Bay techno-complexes, broadly associated with MIS 4, are recognized across various ecological zones of South Africa. Thesetechno-complexes contain relatively more retouched tools and exhibit heightened levels of what is described asinnovative practices. The Howiesons Poort is the best understood industry in the Middle Stone Age of South Africa.Its unifying technological characteristic is the almost exclusive use of a blade and bladelet production system, butsubtle changes in types of backed artefacts, other retouched tools, and raw material exploitation patterns occurthrough time. Technological preferences change in the ensuing MIS 3 period, and other strategies and implementtypes become popular, particularly unifacial points, but trends once again are less clear. This historical review ofthe technological diversity in the Middle Stone Age of South Africa emphasizes that the roots of some innovationsmay lie in the earlier Middle Stone Age and that innovation is best understood within the context of local historicaltrajectories of technological change in South Africa.

The Middle Stone Age period of South Africa is a key milieuwithin which Homo sapiens developed, diverged, and migrated(e.g., Behar et al. 2008; Dusseldorp, Lombard, and Wurz 2013;Lombard, Schlebusch, and Soodyall, forthcoming; Pickrell etal. 2012; Schlebusch et al. 2012). Numerous Middle StoneAge cave and open-air sites occur all over South Africa andbear testimony to the successful adaptation of hunter-gath-erers over the past 300,000 years. Stratified sequences fromcaves acted as culture-stratigraphic anchors since the “MiddleStone Age” (MSA) concept was established (Goodwin 1928).A recent exercise in ordering Stone Age sites with relativelyrecent chronometric dates shows that some progress has beenmade in expanding the MSA spatiotemporal framework ofSouth Africa and Lesotho (Lombard and Haidle 2012). Theaim here is to describe MSA technological developments fromSouth Africa, a large area covering more than 1.2 millionsquare kilometers. The considerable size of the area precludesan exhaustive discussion, and only fairly general trends areaddressed. The noticeable growth in investigations into theMSA in the past decade involve some cave sites from the Capecoastal area, which necessarily biases this discussion towardthese locations, but sites from other regions—for example,

Sarah Wurz is Senior Researcher in the Institute for HumanEvolution of the University of the Witwatersrand (Private Bag 3, POWITS, 2050, Johannesburg, South Africa [sarah.wurz@wits.ac.za]).This paper was submitted 3 VII 13, accepted 18 VII 13, andelectronically published 29 X 13.

Sibudu Cave, Border Cave, and Kathu Pan—also contributesignificantly to the latest insights.

The dominant narrative for the MSA of South Africaemerging from a number of recent publications (e.g., Comp-ton 2011; McCall 2007; McCall and Thomas 2012; Mellars2006; Mourre, Villa, and Henshilwood 2010; Soriano, Villa,and Wadley 2007; Villa, Delagnes, and Wadley 2005; Villa etal. 2010; Ziegler et al. 2013) is that “Middle Paleolithic–like”variability occurred from the earliest MSA until around 77,000years ago. The first time that “Upper Paleolithic–like” vari-ability developed was during marine isotope stage (MIS) 4,with a spurt of innovation and complexity associated withthe Still Bay and Howiesons Poort techno-complexes. Theinnovations include formal bone tools, beads, processing ofochre, engraved ochre and ostrich eggshell, and compoundadhesive manufacture. The technological sophistication of theHowiesons Poort and the Still Bay is evident in the manu-facture of geometric-backed tools and bifacial points, specifictechniques of blade and bladelet reduction, pressure flaking,and heat treatment of stone. These levels of innovation andcomplexity are thought to signify symbolically mediated be-havior (Henshilwood and Marean 2003) or cognitive com-plexity (Wadley, Hodgkiss, and Grant 2009), perhaps evenconferring evolutionary advantage to the groups of Homosapiens who migrated out of Africa after ∼60,000 years ago(Mellars 2006; Mourre, Villa, and Henshilwood 2010). After∼58,000 years there is a return to Middle Paleolithic–liketechnological variability and a dearth of complex stone tools.

S306 Current Anthropology Volume 54, Supplement 8, December 2013

Figure 1. Map of sites mentioned in the text.

The lack of perceived complexity and sophistication of thepost–Howiesons Poort technologies indicates a return to“nonmodern” levels of behavior and cognition (but see Lom-bard and Parsons 2010; Marean 2010; Wurz 2008). This nar-rative exaggerates, but it does capture the dominant percep-tion of the development of technological complexity in theMSA of South Africa. A historical perspective on the spa-tiotemporal variability of stone tool assemblages dating tobetween ca. 300,000 and 40,000 years ago provides the op-portunity to evaluate whether this narrative is the best-fitmodel for the MSA of South Africa. The MISs are used as anorganizational framework only as it is clear that terrestrialtemperature and precipitation changes do not coincide withthem (Blome et al. 2012).

Early Middle Stone Age

The transition from the Early Stone Age to the MSA is am-biguous, and handaxes and other large cutting tools co-occurwith typical MSA flakes, blades, and points for an extended

period, from 500,000 to 125,000 years ago, mainly in open-air sites all over South Africa. The Fauresmith industry, con-sidered transitional between the Early Stone Age and the MSA,contains small handaxes, blades, and points. Its temporal des-ignation is unclear, and although relatively young dates(around 150,000 years ago) at, for example, Bundu Farm andRooidam (Herries 2011) are noted, dates of between 280,000and 500,000 years ago may be more realistic (Beaumont andVogel 2006; Herries 2011; Porat et al. 2010). The Sangoanfrom the northern parts of South Africa, with its core axesand MSA debitage, is as yet undated but is in all probabilityalso a regionally distinct transitional industry (Kuman 2007).At Duinefontein in the Western Cape (fig. 1), large cuttingtools occur as late as 125,000 years ago (Feathers 2002; Kleinet al. 1999). The chronological ambiguities often associatedwith Middle Pleistocene sites (cf. Barham 2012) impede aclear understanding of when handaxes and other core toolsceased being part of the technological repertoire of Pleistocenehunter-gatherers. The earliest MSA assemblage recorded with-out Acheulean elements is an a ca. 280,000-year-old assem-

Wurz Technological Trends in the Middle Stone Age of South Africa S307

blage from Florisbad (Grun et al. 1996; Kuman, Inbar, andClark 1999). It is not particularly diagnostic, and platformcores and a small number of triangular flakes and blades aredescribed.

The Early MSA dates to between 300,000 and 130,000 yearsago (Lombard et al. 2012; Volman 1984) and is characterizedby prepared-core technologies, including Levallois, discoid,and blade methods. These methods had been invented morethan half a million years ago. A recurring issue is whetherthe origin of the Levallois in South Africa is in the VictoriaWest industry (Van Riet Lowe 1945). In the Victoria Westindustry, a Levallois-like technique is employed to prepare adomed surface for the removal of a large preferential flake,resulting in different shapes of discarded cores, including thefamous hoenderbek (hen beak) cores (Lycett 2011; McNabband Beaumont 2012). Subsequent investigations of these coresdescribe them as “proto-Levallois” (White, Ashton, and Scott2011) or “para-Levallois” to avoid assumptions of phyloge-netic relationships (Lycett 2009). At Canteen Kopje, VictoriaWest technology dating to more than 500,000 years ago occursalongside prepared-core or Levallois technology (McNabband Beaumont 2011, 2012). Thus, a unilinear evolution fromthe Acheulean to Levallois technology is highly unlikely. Theearly presence of Levallois at this site confirms that it hasdeep roots in South Africa and co-occurs with typical EarlyStone Age methods. In the South African MSA, “Levallois”encompasses several types of prepared cores, including radialcores (Volman 1984; Wilkins, Pollarolo, and Kuman 2010).Whether these cores conform strictly to the “Levallois” con-cept or not is less important than appreciating that prepared-core technology provided a new level of technological flexi-bility (White, Ashton, and Scott 2011).

In South Africa, as in East Africa (Tryon and Faith 2013)and the Levant (Hovers and Belfer-Cohen 2013), laminartechnology occurs in the earlier part of the Middle Pleistocene.Large blades together with small Fauresmith handaxes, Le-vallois points and flakes are associated with sites from theNorthern Cape (Beaumont and Morris 1990; Sampson 1974).For one of these assemblages—Kathu Pan 1 stratum 4a, re-cently redated to ∼500,000 years ago (Porat et al. 2010)—asystematic blade production system is described (Wilkins andChazan 2012), confirming Sampson’s (1968) impression thatlarge blade production has an ancient origin in South Africa.The cores are centripetally prepared before the removal ofblades with hard-hammer percussion. The blade cores are“Levallois-like” (Wilkins and Chazan 2012:15), but they arenot Levallois in concept because the flaking surfaces are muchmore convex than the lower platform surfaces, and the in-tersection of the surfaces is not a plane. A number of bladelets(!12 mm in width) also occur in this assemblage. Some lam-inar products are retouched into unifacial points, and theseand nonretouched triangular flakes show evidence for haftingand use as spear tips (Wilkins et al. 2012). It is significantthat this level of complexity occurs at such an early date, andit would be more so if it could be demonstrated more com-

prehensively that the museum collections analyzed originatefrom the recently dated “stratum 4a” at Kathu Pan (Wadley2013).

There has been very little opportunity to integrate paleoen-vironmental and technological data of the Early MSA becauseregional environmental proxies and technological trajectoriesstill need to be developed. An exception is the “Cape FloralRegion—South Coast Model” for the origins of modern hu-mans (Marean 2010:432, 2011). In this model MIS 6, one ofthe coldest and perhaps most arid stages of the Quaternary(Marean 2010; Petit et al. 1999), is related to a genetic bot-tleneck (e.g., Behar et al. 2008; Blum and Jakobsson 2011;Fagundes et al. 2007; Gonder et al. 2007) when populationscrashed and survived only in favorable refugia. The 162,000-year-old (Jacobs 2010) assemblage from Cave 13B at PinnaclePoint on the southern Cape coast is considered such a re-fugium (Compton 2011; Marean 2010). The southern Capecoast, being somewhat buffered from extreme aridity effects,would have provided predictable faunal, geophyte, and shell-fish food resources. The continued occupation of sites suchas Klasies River and Blombos Cave in MIS 5 on the southernCape coast is seen as a fluorescence of populations that sur-vived in the southern Cape refugium during MIS 6. The Cave13B assemblage is associated with the earliest use of shellfish(Marean et al. 2007) and with Levallois technology, points,blades, and bladelets (!10 mm in width; Marean et al. 2007:906). The bladelets have been reported as the first of theirkind in the South African MSA (Thompson, Williams, andMinichillo 2010; but see Wilkins and Chazan 2012).

Certain genetic, paleoenvironmental, and archaeologicaldata imply that the southern Cape MIS 6 refugium scenariomay need reconsideration. African resequencing data (Sjodinet al. 2012) indicate, for example, that a demographic modelwithout an ancestral bottleneck in MIS 6 is a possibility (seealso Lombard, Schlebusch, and Soodyall, forthcoming). It isfurther questionable whether the MIS 6 glacial resulted inextreme aridity in all areas of southern Africa. At Florisbad,an inland site, occupation events are related to wetter phases(Scott and Brink 1992), and one of these occurs in MIS 6, at

years ago (Kuman, Inbar, and Clark 1999).157,000 � 21,000At Florisbad and the surrounding areas, highly productiveecosystems occurred periodically between 400,000 and100,000 years ago. The nine extant and six extinct species ofgrazers identified in various units reflect an open grasslandecosystem similar to that of the Serengeti (Brink 2005; Brinkand Lee-Thorpe 1992). Geological indicators, wetland species,bird remains, and palynological and phytolith data signal thepresence of pans or lakes in the central interior at these times(Loock and Grobler 1988; Manengold and Brink 2011; Scottand Rossouw 2005; Van Zinderen Bakker 1989; Visser andJoubert 1991). It is not surprising that this area was occupiedby humans during productive times, as evidenced by the∼160,000-year-old “Florisbad industry” (Sampson 1974), ahighly retouched industry with many side-scrapers. Elongatedflakes and a variety of core types, including prepared cores,

S308 Current Anthropology Volume 54, Supplement 8, December 2013

occur (Kuman 1989). Florisbad and other inland sites suchas Border Cave, Bundu Farm, Lincoln Cave, and WonderwerkCave show that hunter-gatherer groups lived beyond thesouthern Cape during MIS 6 (Lombard et al. 2012).

MIS 5 Middle Stone Age

Only a limited number of assemblages from MIS 5 (ca. 128–75 ka; Carto et al. 2009) have both reliable dates and recentlypublished descriptions. Spatial and chronological patterningremains vague, but this may have more to do with the ana-lytical approaches followed and intensity of research focusthan a true absence of cultural patterning. For a considerableperiod of time lithic analysts were guided by a typologicalparadigm to describe MSA retouched artefacts, blades, con-vergent flakes (or points), and cores in terms of typometricalattributes (e.g., Kaplan 1990; Singer and Wymer 1982; Thack-eray 1989, 2000; Thackeray and Kelly 1988; Volman 1981,1984; Wadley and Harper 1989), but more technological de-scriptions are emerging (e.g., Porraz et al. 2013b; Will et al.2013; Wurz 2002). The MIS 5 occurrence from Klasies Riverdates to between ca. 115,000 and 80,000 years ago (Wurz2002) and consists of over 10 m of deposits. It constitutesone of the largest collection of artefacts from this time rangein South Africa, sourced from both older and modern ex-cavation techniques, and this provides an opportunity to in-vestigate technological trends through time. It is subdividedinto the Light Brown Sand and Shell and Sand Members(Deacon and Geleijnse 1988) and two corresponding culturalunits, the MSA I and the MSA II (Singer and Wymer 1982),or MSA 2a and 2b (Volman 1981). The morphometric sim-ilarity of the MSA I and II quartzite-dominated assemblagesat Klasies River has been interpreted as evidence for tech-nological continuity (Singer and Wymer 1982; Thackeray1989), but end product profile, platform details, debitagetypes, core geometry, and scar patterning (after Chazan 1997;Van Peer 1992) indicate observable technological change(Wurz 2000, 2002). Multivariate statistical analyses supportthis contention and reveal significant differences between theend products of the two taxons, especially in terms of platformthickness and length (Wurz et al. 2003). For this reason dif-ferent terminology for the MSA I and MSA II at Klasies Riveris suggested. “Klasies River” is the name given for the MSAI because it is the first more detailed description for this typeof occurrence (Wurz 2000), and “Mossel Bay” is the namegiven for the MSA II because this name has historical pre-cedence for assemblages from the southern Cape that fit thisdescription (Sampson 1974).

The Klasies River techno-complex is characterized in partby a recurrent blade reduction strategy. Some of the end prod-ucts have the same characteristics as those of the HowiesonsPoort—curved blades and points with diffused bulbs of per-cussion and platforms showing rubbing and removal of smallflakes from the dorsal edge (Wurz 2000, 2002). Blade andpoint production similar to that of the MSA II (Mossel Bay)

also occur (Wurz 2012). Assemblages from Florisbad, Yster-fontein 1, and Hoedjies Punt are broadly contemporaneouswith the Klasies River techno-complex. It is uncertain howthe ∼120,000-year-old hornfels-dominated assemblage fromFlorisbad with blades, triangular flakes, bladelets, and fewformal tools (Kuman, Inbar, and Clark 1999) compares. TheYsterfontein (YFT 1) techno-complex dating to either 1120ka or MIS 5c-a (Avery et al. 2008) is a flake industry in mainlysilcrete and calcrete with a few blades. The formal tool com-ponent consists of denticulates (Wurz 2012). The nearbyHoedjies Punt industry of the last interglacial age resemblesYFT 1 technology closely (Will et al. 2013) and that from SeaHarvest (Volman 1978). The three Western Cape techno-com-plexes differ from the Klasies River occurrence, and there isa strong possibility that they are younger (Will et al. 2013).

The Mossel Bay from Klasies River dating to between∼100,000 and 80,000 years ago is characterized by a unipolarrecurrent Levallois reduction process for points and blades.The majority of the end products have large, faceted platformsassociated with prominent bulbs of percussion and straightprofiles (Wurz 2000, 2002). A similar assemblage comes fromDiepkloof (Porraz et al. 2013b) and Cape St. Blaize (Sampson1974; Thompson and Marean 2008). The Pinnacle Point Cave13B assemblage is typologically but not necessarily techno-logically similar (Thompson, Williams, and Minichillo 2010).Commonalities between the stratified sequences from Caveof Hearths (not dated) from the north of South Africa andKlasies River have been noted (Sampson 1974; Sinclair andMcNabb 2005; Volman 1984). In Bed 4, Cave of Hearths,long flake-blades and convergent points occur, while the over-lying levels comprise a Levallois industry with blades, radialflakes, and, especially noticeable, convergent points. The Caveof Hearth assemblage served as a comparative basis for themany surface occurrences found in the Free State (Sampson1968). The South African interior contains a multitude ofsurface sites (Beaumont and Morris 1990; Sampson 1968,1974). For example, in an area of 216 square miles in themiddle reaches of the floodplains of the Orange River, 290MSA well-preserved occurrences, though undated, are iden-tified (Sampson 1968). Sampson comments on their similarityto the pre–Howiesons Poort Klasies River and Cave of Hearthsassemblages. One important difference, however, is a muchlarger extent of bifacial and unifacial flaking at the Cave ofHearths, Orange River, and also Border Cave assemblages inwhat has been termed the “Pietersburg” Industry (Mason1962; Sampson 1974). It would be interesting to discover howthese bifacials and unifacials compare with those of other timeperiods and areas. This will provide the context for under-standing technological innovations involving invasive uni-and bifacial retouch. The blades and points are probably mul-tifunctional tools. Some triangular pointed flakes and bladeshave damage consistent with use in a longitudinal cuttingaction (Brink and Henderson 2001; Henderson 2001; Kuman1989). The unretouched MIS 5 points from Cave 13B at Pin-nacle Point (Bird, Minichillo, and Marean 2007; Schoville

Wurz Technological Trends in the Middle Stone Age of South Africa S309

2010) were used in a similar way. It is likely that both theretouched and unretouched points of this time period servedas spear tips (Lombard and Phillipson 2010; McBrearty andBrooks 2000; Wilkins et al. 2012). The populations of MIS 5were efficient hunters (see Clark and Kandel 2013) and gath-erers, and integrating technological and foraging behaviormore extensively is a priority for future research (d’Errico andBanks 2013).

The Still Bay and Howiesons Poort

Still Bay. The Still Bay techno-complex has fascinated re-searchers and collectors since the first discovery of bifaciallanceolate points, the type artefact of the Still Bay, on theCape Flats in 1870 (Henshilwood 2012; Minichillo 2005). Anumber of Still Bay sites occur in stratified contexts—forexample, Dale Rose Parlour (Trappieskop), Peers Cave, Diep-kloof, Hollow Rock Shelter, Blombos Cave in the WesternCape, Sibudu Cave and Umhlatuzana in KwaZulu Natal, andApollo 11 in Namibia—in addition to several surface occur-rences (Minichillo 2005; Steele et al. 2012; Wadley 2007). Mostof the current dating evidence suggests that the Still Bay spansthe end of MIS 5 and the beginning of MIS 4 (72–60 ka;Blome et al. 2012). There are relatively few dates, but opticallystimulated luminescence (OSL) dating assays from BlombosCave, Sibudu, Apollo 11, and Diepkloof indicate a durationfor the Still Bay period of around 7,700 years, from 75,500to 67,800 years ago (Henshilwood 2012; Jacobs et al. 2008,2012), while Hogberg and Larsson report preliminary OSLdates of ca. 72,000 and 80,000 BP. Thermoluminescence datesplace the Still Bay industry from Diepkloof much earlier, witha mean age of 109,000 years ago (Tribolo et al. 2009, 2013).This implies that the Still Bay may have a much longer du-ration in South Africa than previously thought (Porraz et al.2013a, 2013b; Tribolo et al. 2013).

The Still Bay is characterized by bifacially retouched foliatepoints with lenticular cross-sections. Their shapes vary fromnarrowly elliptic to lanceolate, and they either have a wide-angled pointed butt or two pointed apices (Henshilwood2012; Minichillo 2005; Villa et al. 2009; Wadley 2007). Theshape of Still Bay points may be unique in the MSA (cf. Porrazet al. 2013b). An intersite comparative study involving bifacialpoints from Dale Rose Parlour, Hollow Rock Shelter, PeersCave, Blombos Sands, Kleinjongensfontein, and Cape Hang-klip (fig. 1) demonstrates that they occur in a wide varietyof sizes: they vary from 34 mm to more than 120 mm inlength (Minichillo 2005). The technology of blank productionin the Still Bay is relatively unknown and consists of threerather short production sequences—a unifacial flake reduc-tion sequence and two versions of a bifacial block chaıneoperatoire—at Hollow Rock shelter (Hogberg and Larsson2011). At Diepkloof a reduction sequence for the productionof laminar blanks and flakes is present but awaits furtherdescription (Porraz et al. 2013b).

An extensive attribute analysis of bifacial points is presented

by Villa et al. (2009) on the largest available collection (352bifacial elements) in South Africa, from Blombos Cave (Hen-shilwood 2012). Direct internal percussion (with a hard ham-mer) and shaping by marginal percussion (with a soft or softstone hammer) were used in the production process of sil-crete, quartzite, and quartz bifacials. Initially, it was felt thatpressure flaking and heat treatment (Brown 2011; Brown etal. 2009) were not part of the bifacial production process, butsubsequent experimentation (Mourre, Villa, and Henshil-wood 2010) finds evidence for both pressure flaking and heattreatment on 4% of the silcrete bifacial points from BlombosCave. Mourre replicated bifacial points on heat treated silcreteusing a bone tool for pressure flaking. New diagnostic traitsfor this technique are described, and heat treatment is re-garded as a prerequisite for pressure flaking on silcrete.Mourre, Villa, and Henshilwood (2010) regard pressure flak-ing used for the Still Bay bifacials as the first occurrence ofits kind anywhere in the world, thus an invention or inno-vation. Intentionally heated silcrete artefacts also occur in theDiepkloof SB (Porraz et al. 2013b; Schmidt et al. 2013). Somebifacials at Hollow Rock Shelter appear to have been pressureflaked with a fine-tipped small tool without heat treatment(Hogberg and Larsson 2011), a technique purportedly dif-ferent from the more advanced pressure-flaking techniquesfound in Europe and North America. The Diepkloof Still Baybifacial elements ( ) are not pressure flaked (Porraz etn p 74al. 2013b), and this is also true for the bifacials from Sibudu(Soriano et al. 2007; Wadley 2013), a feature that is partlyrelated to the raw materials used.

The majority of bifacial points in the Still Bay are in localor near-local materials. At Blombos Cave, for example, 72%of the bifacials are in silcrete, which probably originated 20–30 km from the site. At Diepkloof most bifacials are in localquartzite (80%), and at Hollow Rock Shelter and Sibudu Caveabout half are in local quartzite and dolerite, respectively. Itis likely that the bifacials are multifunctional tools (Henshil-wood 2012)—residue patterns on complete bifacials suggestthat they functioned as knives as well as projectile points(Lombard 2006; Wadley 2007). One proposition is that theywere used exclusively as parts of hand-delivered spears be-cause their morphometric attributes and macrofractures aresimilar to Paleo-Indian points (Villa and Soriano 2010; Villaet al. 2009). Another is that they may have functioned as darts(Shea 2009). Because the bifacials from Diepkloof display rareimpact-like fractures and frequent resharpening of the lateraledges, they may have primarily functioned as cutting imple-ments (see also Minichillo 2005).

The onset of the Still Bay is correlated with climatic changesthat occurred at the end of MIS 5 (Thackeray 2009; Ziegleret al. 2013). This may have pressured populations to adoptbifacial reduction as a strategy to economize lithic raw ma-terial and maintain tools more intensively in the context oflong-distance residential moves, a pattern inferred for bifacialindustries from other parts of the world as well (McCall andThomas 2012). A period of severe climatic instability did

S310 Current Anthropology Volume 54, Supplement 8, December 2013

occur ∼72,000 years ago as indicated by, for example, theCrevice Cave carbon isotopic record (Bar-Matthews et al.2010), the EPICA EDML curve (EPICA Community Members2006), and core CD154-17-17K (Ziegler et al. 2013). For Bar-Matthews et al. (2010) and Ziegler et al. (2013), the Still Baysignals a technological innovation in response to this punc-tuated environmental event, a scenario that becomes unlikelyin the face of a longer duration of the Still Bay.

Howiesons Poort. The Howiesons Poort (HP) is the mostwidespread MSA industry recorded in southern Africa and isknown from more than 20 sites south of the Zambezi (Hen-shilwood 2012; Lombard 2005, 2009), and new occurrenceswere reported recently (e.g., Kandel and Conard 2012; Steeleet al. 2012). Thanks to the distinctiveness of the geometric-backed artefacts, the industry has served as horizon markerfor the MSA of South Africa (Deacon 1989, 1992) for decades.Howiesons Poort sites occur at, for example, the HP namesite, Boomplaas, Border Cave, Diepkloof, Klasies River, KleinKliphuis, Rose Cottage Cave, Sibudu, and Um-hlatuzana in South Africa; at Melikane and Ntloana Tsoanain Lesotho; and at Apollo 11 in Namibia and a number ofsurface sites. The HP is sometimes considered to be a relativelyshort-lived entity occurring within well-understood time lim-its because eight sites have single-grain OSL dates of betweenca. 64,800 and 59,500 years ago (Cochrane, Doelman, andWadley 2013; Jacobs et al. 2008).

There are a number of younger and older dates for the HP.At Klasies River, for example, the HP is associated with elec-tron spin resonance and thermoluminescence dates of be-tween 50,000 and 60,000 years ago (Eggins et al. 2005; Feathers2002; Tribolo, Mercier, and Valladas 2005). The HP fromBorder Cave dates to around 75,000 years ago (Grun andBeaumont 2001; Grun, Beaumont, and Stringer 1990), andthis is somewhat similar to dates from Pinnacle Point andDiepkloof. Unit SADBS at Pinnacle Point 5–6 (Brown 2011;Brown et al. 2012) contains small blades, notched pieces, andbacked artefacts as well as quartzite blades and points. Thisindustry is interpreted as transitional to the HP (Brown etal. 2012), but in most respects it falls within the techno-typological range of variability of the HP (see also Porraz etal. 2013b). The Diepkloof TL dates are between ∼109,000 and∼52,000 years ago (Porraz et al. 2013b; Tribolo et al. 2013).It is suggested that the single-grain OSL dates for the earlyHP (previously interpreted as Still Bay) are erroneous becauseof methodological problems (Guerin et al. 2013; Tribolo etal. 2013). Whether the TL dating from Diepkloof demon-strates that the HP is not a horizon marker and is of muchlonger duration than previously thought (Porraz et al. 2013a,2013b; Tribolo et al. 2013) is still open to debate.

Despite the increased research focus on the HP, its primaryelements remained unchanged. It is a small-blade industrywith backed and notched artefacts and lesser proportions ofpieces esquillees, scrapers, unifacial points, and partly bifacialpoints (Deacon 1995; Goodwin and Van Riet Lowe 1929;

Harper 1997; Kaplan 1990; Minichillo 2005; Sampson 1974;Singer and Wymer 1982; Thackeray 1992; Volman 1984; Wurz1999). The blades originate from a recurrent blade productionsystem using a soft hammer, with preparation of the lateraland distal convexities by elongated debordantes or sometimesflat centripetal removals. The small blades—with high-angled,sometimes extremely small, off-center platforms with rubbeddorsal edges—are typical of the HP (Wurz 2000). Recent de-scriptions from a strict French chaıne operatoire perspective(Porraz et al. 2013b; Soriano, Villa, and Wadley 2007; Villaet al. 2010) provide more details on the blade reduction strat-egy followed at Klasies River, Rose Cottage Cave, and Diep-kloof. It is described as non-Levallois (Porraz et al. 2013b;Villa et al. 2010) because the intersection of the debitagesurface with the platform and nonproduction surface is nota plane, and convexities are more pronounced at the begin-ning of the reduction. Technological continuity may exist be-tween this system and the strategies followed in Middle Pleis-tocene Kathu Pan and the 115,000-year-old Klasies River(MSA I) industries. In both cases the cores are described as“Levallois-like,” with very convex upper surfaces (Wilkins andChazan 2012; Wurz 2002, 2010). The backed artefacts aremostly on blade blanks, but at some sites, such as Diepkloof,they may be on flakes (Mackay 2008b; Porraz et al. 2013b).Experimental knapping by S. Soriano and G. Porraz indicatemarginal percussive gestures with a soft stone hammer (Porrazet al. 2008; Soriano, Villa, and Wadley 2007; Villa et al. 2010).It is not clear whether local raw materials were used, but thisis essential in replicative experiments. Villa et al. (2010) sug-gest that the HP is the first known occurrence of small bladesproduced by marginal percussion and used as blanks for themanufacture of backed pieces. A noteworthy departure fromthe chaıne operatoire–like methodologies to investigate HPvariability is the materialist approach in which categories areconstructed without notions of “ideal forms, the mental pre-disposition of the maker, or the presumed goals of the artisan”(Hiscock 2007:201). From this perspective the HP from Diep-kloof and Klein Kliphuis is interpreted as a time period inwhich core mass was transformed into flake length more ef-ficiently than in preceding or subsequent technological sys-tems (Mackay 2008a, 2009). The ratio of edge length to massof complete flakes is used as a proxy for flaking efficiency.

The HP is not a static entity, and changes through time inthe proportions of backed tools and notched artefacts occur(e.g., Harper 1994; Mackay 2011; Porraz et al. 2013b; Singerand Wymer 1982; Soriano, Villa, and Wadley 2007; Volman1984; Wadley and Mohapi 2008). Bifacially flaked pieces dooccur in some HP assemblages (e.g., Deacon 1995; Wurz 2000:91), and their presence does not necessarily indicate an earlyphase of the HP (Porraz et al. 2013a, 2013b). Changes in rawmaterial proportions, especially quartz and silcrete, are alsonoted (Mackay 2011; Porraz et al. 2013b; Wurz 2000).

The HP consists of different phases at sites such as Um-hlatuzana, Rose Cottage Cave, Sibudu Cave, Diepkloof, andKlasies River (Porraz et al. 2013b; Soriano, Villa, and Wadley

Wurz Technological Trends in the Middle Stone Age of South Africa S311

2007; Villa et al. 2010; Wadley and Mohapi 2008). At KlasiesRiver, for example, a flake production strategy occurs in themiddle of the sequence continuing into the upper levels, andthis is associated with changes in the blade parameters andpercussion technique. Despite these more subtle changes andmorphometric differences in the cores (Clarkson 2010), theHP technological signal is similar at sites as far apart as RoseCottage Cave, Klasies River, and Border Cave (Villa et al.2010). The perception that the selection of fine-grained “highquality, exotic” (Singer and Wymer 1982) raw material oc-curred in HP assemblages (e.g., McCall and Thomas 2012)needs to be qualified. There is in fact considerable variabilityin the raw material composition of different sites, and thesources are often local (Lombard 2005; Minichillo 2006; Wad-ley 2008; Wadley and Mohapi 2008; Wurz 2000). What ismore relevant is that fine-grained rocks were preferentiallyselected.

Experimental, use-wear, and trace analyses of the HPbacked artefacts indicate that they were inserts in compositeprojectile weapons and that hafting arrangements and the rawmaterial of choice for hafts changed through time (Lombard2008; Wadley and Mohapi 2008). A complex chain of actionswas followed to create compound adhesives, with red ochreand plant gum among the ingredients for hafting of thebacked artefacts (Wadley 2013; Wadley, Hodgkiss, and Grant2009). The impact fractures on backed artefacts imply thatthey functioned as spear points (Villa and Soriano 2010).Small quartz segments from Sibudu Cave may represent thefirst bow-and-arrow technology, a crucial invention (Hen-shilwood 2012; Lombard 2011; Lombard and Phillipson2010). Several lines of evidence suggest that the quartz seg-ments tipped arrows in a transverse position: their morpho-metric dimensions (Wadley and Mohapi 2008), experimentsby Lombard and Pargeter (2008), impact fractures (Lombard2011), microresidues (Lombard 2007, 2008), and use-traceevidence (Lombard 2011). It is likely that different types ofbacked artefacts were used in different kinds of hunting ac-tivities. They were used to perform a variety of functions,including cutting (Igreja and Porraz 2013; McBrearty andBrooks 2000; Wadley and Mohapi 2008). A bone point fromthe Sibudu HP layers at Sibudu Cave, similar to unpoisonedbone arrow points from much later contexts, may also havebeen used in bow-and-arrow technology (Backwell, d’Errico,and Wadley 2008; d’Errico, Backwell, and Wadley 2012). Itseems that bow-and-arrow technology disappeared by the endof the HP (Wadley and Mohapi 2008).

The relationship between climatic change and HP culturalexpressions is a much-discussed topic (Cochrane, Doelman,and Wadley 2013). The colder conditions during this glacialperiod may be reflected by the angular spall in HP levels atBorder Cave, Rose Cottage Cave, and Boomplaas Cave (asreviewed in Chase 2010). The colder conditions are relatedto more open environments (Bar-Matthews et al. 2010; Dea-con and Lancaster 1988; Deacon et al. 1984), an inferencesupported by the presence of grazers in southern Cape ar-

chaeological sites (e.g., Klein 1976; Van Pletzen 2000). A num-ber of proxies (e.g., EPICA Community Members 2004; NorthGreenland Ice Core Project Members 2004; Pahnke et al. 2003;Ziegler et al. 2013) indicate the colder conditions of MIS 4.The speleothem carbon and oxygen isotope ratios from Crev-ice Cave, close to Pinnacle Point, are interpreted as evidencefor more summer rain and more grass cover in this timeperiod (Bar-Matthews et al. 2010). The grassier habitats inthe southern Cape in MIS 4 may be linked to wetter con-ditions due to the shift of the position of the winter rainfallbelt during glacial advances (Chase and Meadows 2007). Mi-crofaunal evidence (e.g., Avery 1982, 1992) from the southernCape points to colder and moister conditions during this timeperiod. Similar conditions occurred at Sibudu Cave, wherethe fauna (Clark 2011; Clark and Plug 2008) and charcoal(Allott 2006; Hall, Woodborne, and Scholes 2008) indicate acool, moist environment with evergreen forest but with awoodland/savannah habitat close by. At Sibudu and a fewother HP sites, a marked increase in the exploitation of smallmammals and the smallest ungulates in MIS 4 occur, indi-cating closed environments and an expansion of diet breadth(Clark and Kandel 2013). It could be expected that a mosaicof open and closed environments occurred in MIS 4.

The HP has been interpreted as an adaptation to fluctuatingMIS 4 environments in the context of residential mobility andrisk reduction through investment in technology (e.g., Am-brose 2002; McCall 2007) and time-dependent foraging (Min-ichillo 2006). Relatively more time and energy had been ex-pended to produce the geometric-backed artefacts forcomposite tools (Bird and O’Connell 2006; Elston and Bran-tingham 2002). Another interpretation is that the HP and StillBay reflect a response to the expansion and isolation of mod-ern human populations between 80,000 and 60,000 years ago(Jacobs and Roberts 2009), as may be implied by some geneticstudies (Atkinson, Gray, and Drummond 2009; Behar et al.2008; Quintana-Murci et. al. 2008; Scheinfeldt, Sameer, andTishkoff 2010; Tishkoff et al. 2007; see also Lombard, Schle-busch, and Soodyall, forthcoming). The interplay betweenclimate, demography, mobility, and technological responsesis a complex topic (d’Errico and Banks 2013) only touchedon here because the focus is on technological change in theMSA (for recent reviews see Brown 2011; McCall and Thomas2012). To investigate more productively whether climate hasbeen a motor of behavioral change in the HP, less coarse toolsneed to be developed to integrate climatic data with culturalchange (d’Errico and Henshilwood 2007).

After the Howiesons Poort

In his 2008 review of MIS 3 MSA, Mitchell (2008) notes thatthis period is relatively unexplored in comparison with MIS4, and this remains true. The transition to the post-HP∼58,000 years ago preceded a large-scale change in climate∼57,000 years ago (Cochrane 2008). Occupational hiatusesoccur in some southern Cape sites in MIS 3 (∼50–20 ka),

S312 Current Anthropology Volume 54, Supplement 8, December 2013

perhaps due to hyperarid conditions (Deacon and Thackeray1984; Steele and Klein 2009), but otherwise MIS 3 was markedby fluctuating climatic conditions that included very wet pe-riods (Deacon and Lancaster 1988; Mitchell 2008). Proxiesfrom Tswaing Crater, Wonderkrater, Elands Bay, and the CapeFlats show that rainfall was sometimes less but still compa-rable to levels experienced by the South African Highveldtoday. The occurrence of more than 25 MSA sites dating tothis time period in southern Africa (Mitchell 2008) is testi-mony to successful adaptation in the post-HP time period.

MIS 3 assemblages are associated with a number of labels,including post-HP, MSA 3, MSA III, or informal late MSA(Lombard et al. 2012). In most assemblages there is a returnto the larger blades and points that characterize the MSA, butcontinuities between the HP and post-HP at Klasies River andRose Cottage Cave do occur (Clark 1999; Lombard and Par-sons 2010; Soriano, Villa, and Wadley 2007; Villa et al. 2010;Wadley and Harper 1989; Wurz 2000). Technological trendsare much less apparent than in MIS 4, even though the samevariety of analytical approaches has been applied for the post-HP. Many sites dating to between ∼58,000 and 45,000 yearsago contain points, often unifacially retouched, for example,at Border Cave, Klein Kliphuis, Sibudu Cave, Diepkloof, Kla-sies River, Umhlatuzana, and Rose Cottage Cave in SouthAfrica and Melikane, Ntloana Tsoana, and Sehonghong inLesotho (Lombard et al. 2012). The most extensive analysisof the post-HP is from Sibudu Cave. The technology is notvery elaborate, with a minimal degree of predetermination ofshapes (Villa, Delagnes, and Wadley 2005), and the retouchedcomponent, mainly consisting of unifacial points, is well de-scribed (Cochrane 2006; Conard, Porraz, and Wadley 2012;Mohapi 2012; Villa, Delagnes, and Wadley 2005; Wadley2005). These points exhibit unique attributes: they have fac-eted platforms and are relatively broader, thicker, and longerthan the other point types in the Sibudu sequence (Mohapi2012). It is suggested that post-HP assemblages in South Af-rica dating to between 58,000 and 45,000 years ago be termedthe “Sibudu” (Lombard et al. 2012). It has already stimulateddebate on whether this label is appropriate for all post-HPsites in this time range (e.g., Conard, Porraz, and Wadley2012; Porraz et al. 2013b).

Final MSA assemblages, dating to between 45,000 and20,000 years ago (Lombard et al. 2012), are associated witha wide variety of point types, such as hollow-based pointsand bifacial and unifacial points. Beaumont (1978) suggeststhat a ∼40 ka industry at Border Cave with small, irregularmicrolithic flakes in quartz—frequently produced by bipolarreduction and a few retouched tools, including outils ecail-lees—represents an “Early Later Stone Age” (ELSA). In sup-port of this hypothesis, Villa et al. (2012) describe the tech-nology of ELSA and preceding industries and also microliths(here referring to bladelet fragments and small flakes) haftedwith pitch from Podocarpus elongates bark. Like Beaumont,they propose that the ∼43,000-year-old assemblage at BorderCave marks the beginning of the Later Stone Age (LSA) in

South Africa. Organic artefacts—such as bone arrowheads,notched bones for notational purposes, and digging sticks andbone awls similar to those used by ethnographically knownSan—are identified (d’Errico et al. 2012) and interpreted asthe emergence of adaptations that are typical of the San. Thesenew results do not clarify the considerable ambiguity sur-rounding the MSA-LSA transition appreciably as the contin-ued presence of typical MSA stone tool assemblages until22,000 years ago (Deacon and Deacon 1999) still preventsunderstanding the technological difference between the MSAand LSA in terms of a clear-cut Rubicon (Clark 1999; Mitchell2012; Wadley 1993).

Concluding Discussion

In South Africa, as in other areas of Africa, the transitionbetween the ESA and MSA is not well understood. The tran-sition was between ∼500,000 and 285,000 years ago, and dur-ing this time period MSA and Acheulean elements occur to-gether. The technological antecedents of the MSA are inprepared-core technology associated with Acheulean assem-blages and Fauresmith blade and point technologies. In EarlyMSA assemblages (300,000–130,000 years ago), blade tech-nology, Levallois flake and point technologies, bladelets, andretouched types such as scrapers, denticulates, and retouchedpoints form part of the stone tool inventory. In MIS 5 thesame diversity of technical elements continues. There is muchpotential in further investigating the spatiochronological ex-tent and details of the technological trends documented byearlier syntheses (e.g., Sampson 1974; Volman 1984). Cur-rently it is only in MIS 4 that the technological trends areclear, but this may simply be the result of the more intensiveanalytical scrutiny of this period. Components of the Still Baybifacials and their production processes are thoroughly de-scribed, but the debitage processes await further consider-ation. The Howiesons Poort techno-complex is described toa fuller extent than that for the Still Bay. It encompasses acharacteristic blade and bladelet associated with different pro-portions of flakes, backed geometric shapes, notched pieces,and other retouched classes. The Still Bay and HowiesonsPoort are widely considered as distinct technological pulsesof a relatively short duration each, but the ambiguity in datingresults may mean that this idea needs revisiting. Post–How-iesons Poort technologies occurring between 58,000 and45,000 years ago are united by the presence of unifacial pointsin what has been termed the Sibudu techno-complex. TheLSA may be heralded by an ∼42,000-year-old assemblage fromBorder Cave with bipolar technology, small quartz flakes andbladelets, and San-like bone tools. This assemblage is the onlyoccurrence of its kind in South Africa at this time. It may bean indication that regionally distinct technological trajectoriesexisted around 40,000 years ago. Hypotheses that are con-structed to investigate how the MSA and LSA contrast hand-icap understanding of cultural trends (Mitchell 2002) as it

Wurz Technological Trends in the Middle Stone Age of South Africa S313

masks continuities and the complex mosaic of culturalchange.

Several claims of “firsts” or innovations are associated withthe Still Bay and Howiesons Poort. These are pressure flakingin combination with heat treatment for the production of StillBay silcrete bifacial points, marginal percussion with a softstone hammer and the transformation of blades into geo-metrics in the Howiesons Poort, and the use of small How-iesons Poort quartz-backed artefacts as transverse arrowheadsin what may represent the earliest known bow-and-arrowtechnology. Noticeably more objects—such as beads (d’Errico,Vanhaeren, and Wadley 2008; d’Errico et al. 2005; Henshil-wood et al. 2004), engraved ochre (Henshilwood and d’Errico2011; Henshilwood, d’Errico, and Watts 2009; Henshilwoodet al. 2002; Mackay and Welz 2008), engraved ostrich eggshell(Texier et al. 2010, 2013), and formal bone tools (d’Erricoand Henshilwood 2007; d’Errico, Moreno, and Rifkin 2012;Henshilwood et al. 2001)—occur within the MIS 4 techno-complexes. These innovations are interpreted as evidence forsymbolically mediated behavior (Henshilwood 2012; Hen-shilwood and Dubreuil 2011 [but see comments on the latter];Henshilwood and Marean 2003). The innovative aspects ofbehavior of interest in developing explanatory models of MSAbehavior are listed by d’Errico and Banks (2013). Among theseare the ochre tool kits from Blombos Cave dating to ∼100,000years ago (Henshilwood et al. 2011) and pre–MIS 4 engravedochres (d’Errico, Moreno, and Rifkin 2012; Henshilwood,d’Errico, and Watts 2009; Watts 2010). This places incidencesof “symbolic behavior” (Henshilwood and Dubreuil 2011)and complex cognition (Wadley 2010, 2013) earlier than MIS4. Insights into the cognitive capabilities of MSA people arefurther explored by other frameworks (e.g., Lombard andHaidle 2012; Wynn and Coolidge 2011), including those thatcast the theoretical net wider by incorporating evolutionarytheory (Dubreuil 2011; Ellis 2011; Gontier 2012). Furtherresearch on how MSA technological trends covary with trendsin foraging (Clark and Kandel 2013) and other kinds of be-havior (d’Errico and Banks 2013; Lombard 2012) is neededto develop integrated models of cultural development in theMSA.

The search for innovation drives research on the MSA,explicitly so during the past decade. What is considered “in-novative” technologically is often sourced from the UpperPaleolithic from Europe (but see Wadley 2013). Combiningapproaches developed for the European Paleolithic with localones to analyze MSA assemblages has led to more detaileddescriptions of technologies, but it is pointless to describeassemblages as “Middle” or “Upper Paleolithic” in character(e.g., Soriano, Villa, and Wadley 2007; Villa, Delagnes, andWadley 2005; Villa et al. 2010). Eurocentric interpretation ofthe MSA is not new, and already in 1928 Goodwin counseledagainst seeing the South African Stone Age through “Euro-pean spectacles” (Goodwin 1928:29). Technological change inthe MSA is best interpreted within the context of local his-torical trajectories. Inferring the origins of any technological

choice requires identifying local antecedents. Historical factorsstructure the adaptive configuration of populations as ex-plained by, for example, Kuhn (2006) in the “rugged fitnesslandscape” model. Populations would have drawn on partic-ular technical expressions that they were already familiar with.When seen from this perspective, the “innovativeness” of theHowiesons Poort and Still Bay is not revolutionary. Artefactproduction systems in MIS 4 are different in detail and notin kind from those of other MSA industries. Pre–Still BayMSA populations were able to use most of the stone tooltechniques that flourished in MIS 4. For example, the bladetechnology of the Howiesons Poort was present in the∼110,000-year-old Klasies River assemblage; bladelet produc-tion is ancient; heat treatment of stone occurs already 162,000years ago; and bifacial flaking occurred in MIS 5. Severalresearchers have noted that fluctuating backed artefact pro-duction characterizes the MSA from at least 230,000 yearsago (Barham 2002; Hiscock and O’Connor 2006; Wadley andMohapi 2008). The same argument can be made for organicartefacts. The enigmatic remains of what was possibly part ofa 121,000-year-old throwing stick is described from Florisbad(Bamforth and Henderson 2003), and bone tools occur inthe ∼100,000-year-old levels from Klasies River (Wurz 2000).

Climatic and paleoenvironmental reconstruction plays animportant role in developing explanatory theory for the MSA,but the relationship between climatic and technologicalchange is indirect and multifaceted (Cochrane, Doelman, andWadley 2013; Jacobs and Roberts 2009). In the MSA, as inthe LSA, climate and environment structured the availablesubsistence choice but had no deterministic effect on tech-nology. LSA technological changes are described as filteringthrough the social web of interactions in a complex way (Dea-con 1984:285). Farther afield, in Europe, for example, Ne-anderthal technological changes in MIS 9–7 occur indepen-dent of climatic change (Moncel et al. 2011).

Although this review focuses on the MSA of South Africa,the area south of Zambezi and the Kunene forms an ecolog-ical, cultural, and archaeological unit (Mitchell 2002). It isthus relevant to compare technological trends in the SouthAfrican MSA with those of countries such as Botswana, Zim-babwe, Lesotho, southern Mozambique, and Swaziland. Ona superficial level the technological patterning of South Africais different from what is described for these countries, al-though the same technical elements, such as retouched pointsand backed artefacts, occur (Barham and Mitchell 2008;McBrearty and Brooks 2000). Nevertheless, the absence ofadequate chronometric control and the lack of a unified de-scriptive taxonomy to compare assemblages impede a thor-ough understanding of how the areas relate. The spatio-chronological integrity of the South African MSA is mostlyinsufficient to identify clear trends, and it is far from “sortedfully” (Goodwin and Van Riet Lowe 1929:7; Thackeray 1992:399). There are adequate data to postulate that the MSA inSouth Africa was a period characterized by fluctuating tech-nological “fashions” (cf. Thackeray 1989), but the chrono-

S314 Current Anthropology Volume 54, Supplement 8, December 2013

logical and regional details remain largely obscure. One hy-pothesis is that similar cycles of technological change occurover the subcontinent (Volman 1984; Wurz 2012). Volman(1984:194) notes “a similar pattern of assemblage changethrough time occurs in MSA sequences south of the Lim-popo,” and Sampson (1974) suggests notable chronologicaland regional patterning for the South African MSA. This levelof change can be most clearly observed for the Still Bay andHowiesons Poort techno-complexes. The lack of recentlydemonstrated patterning on a similar scale for the rest of theSouth African MSA may be an artefact of research intensityor may indicate an actual absence of subcontinental scalepatterning. Technological patterning in MIS 3, for example,reveals that regionally variable technological trajectories ex-isted at times in the prehistory of South Africa (Mitchell2012). Progress in this regard would be optimal if new in-vestigations identify trends and construct hypotheses in thecontext of data generated by previous generations.

Acknowledgments

My thanks go to Erella Hovers and Steve Kuhn for the in-vitation to take part in this conference and to them and LaurieObbink and Leslie Aiello for organizing the conference. Tworeviewers are thanked for their helpful comments on the man-uscript. I also thank the African Origins Platform of the Na-tional Research Foundation of South Africa for funding myresearch.

References CitedAllott, Lucy. 2006. Archaeological charcoal as a window on the palaeovege-

tation and wood use during the Middle Stone Age at Sibudu Cave. SouthAfrican Humanities 18:173–201.

Ambrose, Stanley H. 2002. Early microlithic industries in sub-Saharan Africa.In Thinking small: global perspectives on microlithization. Robert G. Elstonand Steven L. Kuhn, eds. Pp. 9–29. Archeological Papers of the AmericanAnthropological Association, no. 12. Arlington, VA: American Anthropo-logical Association.

Atkinson, Quentin D., Russel D. Gray, and Alexei J. Drummond. 2009. Bayes-ian coalescent inference of major human mitochondrial DNA haplogroupexpansions in Africa. Proceedings of the Royal Society B 276:367–373.

Avery, Graham, David Halkett, Jayson Orton, Teresa Steele, Madelon Tusenius,and Richard Klein. 2008. The Ysterfontein 1 Middle Stone Age rock shelterand the evolution of coastal foraging. South African Archaeological SocietyGoodwin Series 10:66–89.

Avery, Margaret D. 1982. Micromammals as palaeoenvironmental indicatorsand an interpretation of the late Quaternary in the southern Cape Province,South Africa. Annals of the South African Museum 85:183–377.

———. 1992. The environment of early modern humans at Border Cave,South Africa: micromammalian evidence. Palaeogeography, Palaeoclimatol-ogy, Palaeoecology 91:71–87.

Backwell, Lucinda, Francesco d’Errico, and Lyn Wadley. 2008. Middle StoneAge bone tools from the Howiesons Poort layers, Sibudu Cave, South Africa.Journal of Archaeological Science 35:1566–1580.

Bamforth, Marion K., and Zoe L. Henderson. 2003. A reassessment of thewooden fragment from Florisbad, South Africa. Journal of ArchaeologicalScience 30:637–651.

Barham, Lawrence S. 2002. Backed tools in the Middle Pleistocene centralAfrica and their evolutionary significance. Journal of Human Evolution 43:585–603.

———. 2012. Letter to the editor: clarifying some fundamental errors in

Herries’ “A chronological perspective on the Acheulian and its transitionto the Middle Stone Age in Southern Africa: the question of the Fauresmith”(2011). International Journal of Evolutionary Biology (2012):1–5.

Barham, Lawrence, and Peter Mitchell. 2008. The first Africans: African ar-chaeology from the earliest tool makers to most recent foragers. Cambridge:Cambridge University Press.

Bar-Matthews, Miryam, Curtis W. Marean, Zenobia Jacobs, Panagiotis Kar-kanas, Erich C. Fisher, Andy I. R. Herries, Kyle Brown, et al. 2010. A highresolution and continuous isotopic speleothem record of paleoclimate andpaleoenvironment from 90 to 53 ka from Pinnacle Point on the south coastof South Africa. Quaternary Science Reviews 29:2131–2145.

Beaumont, Peter B. 1978. Border Cave. PhD thesis, University of Cape Town.Beaumont, Peter B., and David Morris. 1990. Guide to archaeological sites in

the Northern Cape. Kimberly, South Africa: McGregor Museum.Beaumont, Peter B., and John C. Vogel. 2006. On a timescale for the past

million years of human history in central South Africa. South African Journalof Science 102:217–228.

Behar, Doron M., Richard Villems, Himla Soodyall, Jason Blue-Smith, LuisaPereira, Ene Metspalu, Rosaria Scozzari, et al. 2008. The dawn of humanmatrilineal diversity. American Journal of Human Genetics 82:1130–1140.

Bird, Catherine, Tom Minichillo, and Curtis W. Marean. 2007. Edge damagedistribution at the assemblage level on Middle Stone Age lithics: an image-based GIS approach. Journal of Archaeological Science 34:771–780.

Bird, Douglas W., and James F. O’Connell. 2006. Behavioral ecology andarchaeology. Journal of Archaeological Research 14:143–188.

Blome, Margaret W., Andrew S. Cohen, Christian A. Tryon, Alison S. Brooks,and Joellen Russell. 2012. The environmental context for the origins ofmodern human diversity: a synthesis of regional variability in African cli-mate 150,000–30,000 years ago. Journal of Human Evolution 62:563–592.

Blum, Michael G. B., and Mattias Jakobsson. 2011. Deep divergences of humangene trees and models of human origins. Molecular Biology and Evolution28:889–898.

Brink, James S. 2005. The evolution of the black wildebeest (Connochaetesgnou) and modern large mammal faunas of central southern Africa. PhDthesis, University of Stellenbosch.

Brink, James S., and Zoe L. Henderson. 2001. A high-resolution last interglacialMSA horizon at Florisbad in the context of other open-air occurrences inthe central interior of southern Africa: an interim statement. In MiddlePalaeolithic and Middle Stone Age settlement systems. Nicholas J. Conard,ed. Pp. 1–20. Tubingen: Kerns.

Brink, James S., and Julia Lee-Thorp. 1992. The feeding niche of an extinctspringbok, Antidorcas bondi (Antilopini, Bovidae), and its palaeoenviron-mental meaning. South African Journal of Science 88:227–229.

Brown, Kyle S. 2011. The sword in the stone: lithic raw material exploitationin the Middle Stone Age at Pinnacle Point Site 5–6, southern Cape, SouthAfrica. PhD dissertation, University of Cape Town.

Brown, Kyle S., Curtis W. Marean, Andy I. R. Herries, Zenobia Jacobs, ChantalTribolo, David Braun, David L. Roberts, et al. 2009. Fire as an engineeringtool of early modern humans. Science 325:859–862.

Brown, Kyle S., Curtis W. Marean, Zenobia Jacobs, Benjamin J. Schoville,Simen Oestmo, Erich C. Fisher, Jocelyn Bernatchez, et al. 2012. An earlyand enduring advanced technology originating 71,000 years ago in SouthAfrica. Nature 491:590–593.

Carto, Shannon L., Andrew J. Weaver, Renee Hetherington, Yin Lam, andEdward C. Wiebe. 2009. Out of Africa and into an ice age: on the role ofglobal climate change in the late Pleistocene migration of early modernsOut of Africa. Journal of Human Evolution 56:139–151.

Chase, Brian M. 2010. South African palaeoenvironments during marine ox-ygen isotope stage 4: a context for the Howiesons Poort and Still Bayindustries. Journal of Archaeological Science 37:1359–1366.

Chase, Brian M., and Michael E. Meadows. 2007. Late Quaternary dynamicsof southern Africa’s winter rainfall zone. Earth Science Reviews 84:103–138.

Chazan, Michael. 1997. Lead review: redefining Levallois. Journal of HumanEvolution 33:719–735.

Clark, Amelia M. B. 1999. Late Pleistocene technology at Rose Cottage Cave:a search for modern behavior in a Middle Stone Age context. AfricanArchaeological Review 16:93–119.

Clark, Jamie L. 2011. The evolution of human culture during the later Pleis-tocene: using fauna to test models on the emergence and nature of ‘‘mod-ern” human behavior. Journal of Anthropological Archaeology 30:273–291.

Clark, Jamie L., and Andrew W. Kandel. 2013. The evolutionary implicationsof variation in human hunting strategies and diet breadth during the MiddleStone Age of southern Africa. Current Anthropology 54(suppl. 8):S269–S287.

Wurz Technological Trends in the Middle Stone Age of South Africa S315

Clark, Jamie L., and Ina Plug. 2008. Animal exploitation strategies during theSouth African Middle Stone Age: Howiesons Poort and post–HowiesonsPoort fauna from Sibudu Cave. Journal of Human Evolution 54:886–898.

Clarkson, Chris. 2010. Regional diversity within the core technology of theHowiesons Poort techno-complex. In New perspectives on old stones: ana-lytical approaches to Palaeolithic technologies. Stephen Lycett and Parth Chau-han, eds. Pp. 43–60. New York: Springer.

Cochrane, Grant W. G. 2006. An analysis of lithic artefacts from the ∼60 kalayers of Sibudu Cave. Southern African Humanities 18:69–88.

———. 2008. The transition from Howiesons Poort to post–Howiesons Poortindustries in southern Africa. South African Archaeological Society GoodwinSeries 10:157–167.

Cochrane, Grant W. G., Trudy Doelman, and Lyn Wadley. 2013. Anotherdating revolution for prehistoric archaeology? Journal of ArchaeologicalMethod and Theory 20:42–60.

Compton, John S. 2011. Pleistocene sea-level fluctuations and human evo-lution on the southern coastal plain of South Africa. Quaternary ScienceReviews 30:506–527.

Conard, Nicholas J., Guillaume Porraz, and Lyn Wadley. 2012. What is in aname? characterizing the “post–Howieson’s Poort” at Sibudu. South AfricanArchaeological Bulletin 67 (196):180–199.

Deacon, Hilary J. 1989. Late Pleistocene palaeoecology and archaeology inthe southern Cape, South Africa. In The human revolution: behavioural andbiological perspectives on the origins of modern humans. Paul Mellars andChris Stringer, eds. Pp. 547–564. Edinburgh: Edinburgh University Press.

———. 1992. Southern Africa and modern human origins. Transactions ofthe Royal Society B 337:177–183.

Deacon, Hilary J., and Janette Deacon. 1999. Human beginnings in SouthAfrica: uncovering the secrets of the Stone Age. Cape Town: David Philip.

Deacon, Hilary J., Janette Deacon, Anton Scholtz, J. Francis Thackeray, JamesS. Brink, and John C. Vogel. 1984. Correlation of palaeoenvironmental datafrom the Late Pleistocene and Holocene deposits at Boomplaas Cave, south-ern Cape. In Late Cenozoic palaeoclimates of the Southern Hemisphere. JohnC. Vogel, ed. Pp. 339–352. Rotterdam: Balkema.

Deacon, Hilary J., and Vera B. Geleijnse. 1988. The stratigraphy and sedi-mentology of the main site sequence, Klasies River, South Africa. SouthAfrican Archaeological Bulletin 43:5–14.

Deacon, Hilary J., and Francis Thackeray. 1984. Late Pleistocene environ-mental changes and implications for the archaeological record in southernAfrica. In Late Cenozoic palaeoclimates of the Southern Hemisphere. John C.Vogel, ed. Pp. 375–390. Rotterdam: Balkema.

Deacon, Janette. 1984. The Later Stone Age of southernmost Africa. CambridgeMonographs in African Archaeology 12, BAR International Series 213. Ox-ford: BAR.

———. 1995. An unsolved mystery at the Howiesons Poort name site. SouthAfrican Archaeological Bulletin 50:110–120.

Deacon, Janette, and Nick Lancaster. 1988. Late Quaternary Palaeoenviron-ments of southern Africa. Oxford: Clarendon.

d’Errico, Francesco, Lucinda Backwell, Paola Villa, Ilaria Degano, JeannetteJ. Lucejko, Marion K. Bamford, Thomas F. G. Higham, Maria Perla Col-ombini, and Peter B. Beaumont. 2012. Early evidence of San material culturerepresented by organic artifacts from Border Cave, South Africa. Proceedingsof the National Academy of Sciences of the USA 109(33):13214–13219.

d’Errico, Francesco, Lucinda Backwell, and Lyn Wadley. 2012. Identifyingregional variability in Middle Stone Age bone technology: the case of SibuduCave. Journal of Archaeological Science 39(7):2479–2495.

d’Errico, Francesco, and William E. Banks. 2013. Identifying mechanismsbehind Middle Paleolithic and Middle Stone Age cultural trajectories. Cur-rent Anthropology 54(suppl. 8):S371–S387.

d’Errico, Francesco, and Christopher S. Henshilwood. 2007. Additional evi-dence for bone technology in the southern African Middle Stone Age.Journal of Human Evolution 52:142–163.

d’Errico, Francesco, Christopher S. Henshilwood, Marian Vanhaeren, andKaren van Niekerk. 2005. Nassarius kraussianus shell beads from BlombosCave: evidence for symbolic behaviour in the Middle Stone Age. Journal ofHuman Evolution 48:3–24.

d’Errico, Francesco, Renata G. Moreno, and Riaan F. Rifkin. 2012. Techno-logical, elemental and colorimetric analysis of an engraved ochre fragmentfrom the Middle Stone Age levels of Klasies River Cave 1, South Africa.Journal of Archaeological Science 39:942–952.

d’Errico, Francesco, Marian Vanhaeren, and Lyn Wadley. 2008. Possible shellbeads from the Middle Stone Age layers of Sibudu Cave, South Africa.Journal of Archaeological Science 35(10):2675–2685.

Dubreuil, Benoıt. 2011. The other middle-range theories: mapping behaviourand the evolution of the mind. In Homo symbolicus: the dawn of language,imagination and spirituality. Christopher S. Henshilwood and Francescod’Errico, eds. Pp. 185–204. Amsterdam: Benjamins.

Dusseldorp, Gerrit, Marlize Lombard, and Sarah Wurz. 2013. PleistoceneHomo and the updated Stone Age sequence of South Africa. South AfricanJournal of Science 109(5/6):1–7.

Eggins, Stephen M., Rainer Grun, Malcolm T. McCulloch, Alistair W. G. Pike,John Chappell, Les Kinsley, Graham Mortimer, et al. 2005. In situ U-seriesdating by laser-ablation multi-collector ICPMS: new prospects for Quater-nary geochronology. Quaternary Science Reviews 24:2523–2538.

Ellis, George F. R. 2011. Biology and mechanisms related to the dawn oflanguage. In Homo symbolicus: the dawn of language, imagination and spir-ituality. Christopher S. Henshilwood and Francesco d’Errico, eds. Pp. 163–184. Amsterdam: Benjamins.

Elston, Robert G., and P. Jeffrey Brantingham. 2002. Microlithic technologyin northern Asia: a risk-minimizing strategy of the Late Paleolithic andEarly Holocene. In Thinking small: global perspectives on microlithization.Robert G. Elston and Steven L. Kuhn, eds. Pp. 104–117. Washington: Amer-ican Anthropological Association.

EPICA Community Members. 2004. Eight glacial cycles from an Antarctic icecore. Nature 429:623–628.

———. 2006. One-to-one coupling of glacial climate variability in Greenlandand Antarctica. Nature 444:195–198.

Fagundes, Nelson J. R., Nicolas Ray, Mark Beaumont, Samuel Neuenschwan-der, Francisco M. Salzano, Sandro L. Bonatto, and Laurent Excoffier. 2007.Statistical evaluation of alternative models of human evolution. Proceedingsof the National Academy of Sciences of the USA 104:17614–17619.

Feathers, James K. 2002. Luminescence dating in less than ideal conditions:case studies from Klasies River main site and Duinefontein, South Africa.Journal of Archaeological Science 29(2):177–194.

Gonder, Mary K., Holly M. Mortensen, Floyd A. Reed, Alexandra de Sousa,and Sarah A. Tishkoff. 2007. Whole-mtDNA genome sequence analysis ofancient African lineages. Molecular Biology and Evolution 24:757–768.

Gontier, Nathalie. 2012. Selectionist approaches in evolutionary linguistics:an epistemological analysis. International Studies in the Philosophy of Science26(1):67–95.

Goodwin, A. John H. 1928. An introduction to the Middle Stone Age of SouthAfrica. South African Journal of Science 25:410–418.

Goodwin, A. John H., and Clarence Van Riet Lowe. 1929. The Stone Agecultures of South Africa. Annals of the South African Museum 27:9–51.

Grun, Rainer, and Peter B. Beaumont. 2001. Border Cave revisited: a revisedESR chronology. Journal of Human Evolution 40:467–482.

Grun, Rainer, Peter B. Beaumont, and Chris B. Stringer. 1990. Electron spinresonance dating evidence for early modern humans at Border Cave inSouth Africa. Nature 344(6266):537–539.

Grun, Rainer, James S. Brink, Nigel A. Spooner, Lois Taylor, Chris B. Stringer,Robert G. Franciscus, and Andrew S. Murray. 1996. Direct dating of Flor-isbad hominid. Nature 382:500–501.

Guerin, Guillaume, Andrew S. Murray, Mayank Jain, Kristina J. Thomsen,and Norbert Mercier. 2013. How confident are we in the chronology ofthe transition between Howieson’s Poort and Still Bay? Journal of HumanEvolution 64:314–317.

Hall, Grant, Stephan Woodborne, and M. Scholes. 2008. Stable carbon isotoperatios from archaeological charcoal as palaeoenvironmental indicators.Chemical Geology 247:384–400.

Harper, Philip T. N. 1994. The Middle Stone Age sequence at Rose CottageCave: a search for continuity and discontinuity. Master’s thesis, Universityof the Witwatersrand, Johannesburg.

———. 1997. The Middle Stone Age sequence at Rose Cottage Cave: a searchfor continuity and discontinuity. South African Journal of Science 93(10):470–475.

Henderson, Zoe. 2001. Spatial patterning at southern African Middle Pleis-tocene open-air sites: Florisbad, Duinefontein 2/2 and Mwanganda’s Village.PhD thesis, University of Cambridge.

Henshilwood, Christopher S. 2012. Late Pleistocene techno-traditions insouthern Africa: a review of the Still Bay and Howiesons Poort, c. 75–59ka. Journal of World Prehistory 25:205–237.

Henshilwood, Christopher S., and Francesco d’Errico. 2011. Middle Stone Ageengravings and their significance to the debate on the emergence of symbolicmaterial culture. In Homo symbolicus: the dawn of language, imaginationand spirituality. Christopher S. Henshilwood and Francesco d’Errico, eds.Pp. 75–96. Amsterdam: Benjamins.

S316 Current Anthropology Volume 54, Supplement 8, December 2013

Henshilwood, Christopher S., Francesco d’Errico, Curtis W. Marean, RichardG. Milo, and Royden Yates. 2001. An early bone tool industry from theMiddle Stone Age at Blombos Cave, South Africa: implications for theorigins of modern human behaviour, symbolism and language. Journal ofHuman Evolution 41:631–678.

Henshilwood, Christopher S., Francesco d’Errico, Marian Vanhaeren, Karenvan Niekerk, and Zenobia Jacobs. 2004. Middle Stone Age shell beads fromSouth Africa. Science 304:404.

Henshilwood, Christopher S., Francesco d’Errico, Karen L. van Niekerk, YvanCoquinot, Zenobia Jacobs, Stein-Erik Lauritzen, Michel Menu, and RenataGarcıa-Moreno. 2011. A 100,000-year-old ochre-processing workshop atBlombos Cave, South Africa. Science 334(6053):219–222.

Henshilwood, Christopher S., Francesco d’Errico, and Ian Watts. 2009. En-graved ochres from the Middle Stone Age levels at Blombos, South Africa.Journal of Human Evolution 57:27–47.

Henshilwood, Christopher S., Francesco d’Errico, Royden Yates, Zenobia Ja-cobs, Chantal Tribolo, Geoff A. T. Duller, Norbert Mercier, et al. 2002.Emergence of modern human behavior: Middle Stone Age engravings fromSouth Africa. Science 295:1278–1280.

Henshilwood, Christopher S., and Benoıt Dubreuil. 2011. The Still Bay andHowiesons Poort, 77–59 ka: symbolic material culture and the evolutionof the mind during the African Middle Stone Age. Current Anthropology52(3):361–400.

Henshilwood, Christopher S., and Curtis W. Marean. 2003. The origin ofmodern human behavior: a critique of the models and their test implica-tions. Current Anthropology 44:627–651.

Herries, Andy I. R. 2011. A chronological perspective on the Acheulian andits transition to the Middle Stone Age in southern Africa: the question ofthe Fauresmith. International Journal of Evolutionary Biology 2011:1–25.

Hiscock, Peter. 2007. Looking the other way: a materialist/technological ap-proach to classifying tools and implements, cores and retouched flakes. InTools versus cores? alternative approaches to stone tool analysis. Shannon P.McPherron, ed. Pp. 98–222. Newcastle: Cambridge Scholars.

Hiscock, Peter, and Sue O’Connor. 2006. An Australian perspective on modernbehaviour and artefact assemblages. Before Farming 2:1–10.

Hogberg, Anders, and Lars Larsson. 2011. Lithic technology and behaviouralmodernity: new results from the Still Bay site, Hollow Rock Shelter, WesternCape Province, South Africa. Journal of Human Evolution 61:133–155.

Hovers, Erella, and Anna Belfer-Cohen. 2013. On variability and complexity:lessons from the Levantine Middle Paleolithic record. Current Anthropology54(suppl. 8):S337–S357.

Igreja, Marina, and Guillaume Porraz. 2013. Functional insights into the in-novative Early Howiesons Poort technology at Diepkloof Rock Shelter(Western Cape, South Africa). Journal of Archaeological Science 40:3475–3491.

Jacobs, Zenobia. 2010. An OSL chronology for the sedimentary deposits fromPinnacle Point Cave 13B: a punctuated presence. Journal of Human Evo-lution 59(3/4):289–305.

Jacobs, Zenobia, Elspeth H. Hayes, Richard G. Roberts, Rex F. Galbraith, andChristopher S. Henshilwood. 2012. An improved OSL chronology for theStill Bay layers at Blombos Cave, South Africa: further tests of single-graindating procedures and a re-evaluation of the timing of the Still Bay industryacross southern Africa. Journal of Archaeological Science 40:579–594.

Jacobs, Zenobia, and Richard G. Roberts. 2009. Were environmental or de-mographic factors the driving force behind Middle Stone Age innovationsin southern Africa? South African Journal of Science 105:333–334.

Jacobs, Zenobia, Richard G. Roberts, Rex F. Galbraith, Hilary J. Deacon, RainerGrun, Alex Mackay, Peter Mitchell, Ralf Vogelsang, and Lyn Wadley. 2008.Ages for the Middle Stone Age of southern Africa: implications for humanbehavior and dispersal. Science 322:733–735.

Kandel, Andrew W., and Nicholas J. Conard. 2012. Settlement patterns duringthe Earlier and Middle Stone Age around Langebaan Lagoon, Western Cape(South Africa). Quaternary International 270:15–29.

Kaplan, Jonathan. 1990. The Umhlatuzana rock shelter sequence: 100,000years of Stone Age history. Natal Museum Journal of Humanities 2:1–94.

Klein, Richard G. 1976. The mammalian fauna of the Klasies River Mouthsites, southern Cape Province, South Africa. South African ArchaeologicalBulletin 31:75–98.

Klein, Richard G., Graham Avery, Kathryn Cruz-Uribe, David Halkett, TimHart, Richard G. Milo, and Thomas P. Volman. 1999. Duinefontein 2: anAcheulean site in the Western Cape Province of South Africa. Journal ofHuman Evolution 37:153–190.

Kuhn, Steven L. 2006. Trajectories of change in the Middle Paleolithic in Italy.

In Transitions before the transition: evolution and stability in the MiddlePaleolithic and Middle Stone Age. Erella Hovers and Steven L. Kuhn, eds.Pp. 109–120. New York: Springer.

Kuman, Kathleen. 1989. Florisbad and (Gi: the contribution of open airsites to study of the Middle Stone Age in southern Africa. PhD dissertation,University of Pennsylvania, Philadelphia.

———. 2007. The Earlier Stone Age in South Africa: site context and theinfluence of cave studies. In Breathing life fossils: taphonomic studies in honorof C.K. (Bob) Brain. Travis Rayne Pickering, Kathy Schick, and NicholasToth, eds. Pp. 181–198. Bloomington IN: Stone Age Institute Press.

Kuman, Kathleen, Moshe Inbar, and Ronald J. Clark. 1999. Palaeoenviron-ments and cultural sequence of the Florisbad Middle Stone Age hominidsite, South Africa. Journal of Archaeological Science 26:1409–1425.

Lombard, Marlize. 2005. The Howiesons Poort of South Africa: what we know,what we think we know, and what we need to know. South African Hu-manities 17:33–55.

———. 2006. First impressions on the functions and hafting technology ofStill Bay pointed artefacts from Sibudu Cave. South African Humanities 18:27–41.

———. 2007. The gripping nature of ochre: the association of ochre withHowiesons Poort adhesives and Later Stone Age mastics from South Africa.Journal of Human Evolution 53:406–419.

———. 2008. Finding resolution for the Howiesons Poort through the mi-croscope: micro-residue analysis of segments from Sibudu Cave, SouthAfrica. Journal of Archaeological Science 35:26–41.

———. 2009. The Howiesons Poort of South Africa amplified. South AfricanArchaeological Bulletin 189:4–12.

———. 2011. Quartz-tipped arrows older than 60 ka: further use-trace evi-dence from Sibudu, KwaZulu-Natal, South Africa. Journal of ArchaeologicalScience 38:1918–1930.

———. 2012. Thinking through the Middle Stone Age of sub-Saharan Africa.Quaternary International 270:140–155.

Lombard, Marlize, and Miriam N. Haidle. 2012. Thinking a bow-and-arrow:cognitive implications of Middle Stone Age bow and stone-tipped arrowtechnology. Cambridge Archaeological Journal 22(2):237–264.

Lombard, Marlize, and Justin Pargeter. 2008. Hunting with Howiesons Poortsegments: pilot experimental study and the functional interpretation ofarchaeological tools. Journal of Archaeological Science 35:26–41.

Lombard, Marlize, and Isabelle Parsons. 2010. Fact or fiction? behaviouraland technological reversal after 60 ka in southern Africa. South AfricanArchaeological Bulletin 65:224–228.

Lombard, Marlize, and Laurel Phillipson. 2010. Indications of bow and stone-tipped arrow use 64,000 years ago in KwaZulu-Natal. Antiquity 84:635–648.

Lombard, Marlize, Carina Schlebusch, and Himla Soodyall. Forthcoming.Early Homo sapiens in southern Africa (88–160 ka): an overview of themost recent DNA, fossil and archaeological evidence. South African Journalof Science.

Lombard, Marlize, Lyn Wadley, Janette Deacon, Sarah Wurz, Isabelle Parsons,Moleboheng Mohapi, Joane Swart, and Peter Mitchell. 2012. South Africanand Lesotho Stone Age sequence updated. South African ArchaeologicalBulletin 67(195):123–144.

Loock, Johan C., and N. Johann Grobler. 1988. The regional geology of Flor-isbad. Navorsinge van die Nasionale Museum, Bloemfontein 5:489–497.

Lycett, Stephen J. 2009. Are Victoria West cores ‘‘proto-Levallois”? a phylo-genetic assessment. Journal of Human Evolution 56:175–191.

———. 2011. “Most beautiful and most wonderful”: those endless stone toolforms. Journal of Evolutionary Psychology 9(2):143–171.

Mackay, Alex. 2008a. A method for estimating edge length from flake di-mensions: use and implications for technological change in the southernAfrican MSA. Journal of Archaeological Science 35:614–622.

———. 2008b. On the production of blades and its relationship to backedartefacts in the Howiesons Poort at Diepkloof, South Africa. Lithic Tech-nology 33(1):87–99.

———. 2009. History and selection in the late Pleistocene archaeology of theWestern Cape, South Africa. PhD thesis, Australian National University,Canberra.

———. 2011. Potentially stylistic differences between backed artefacts fromtwo nearby sites occupied 60,000 years before present in South Africa.Journal of Anthropological Archaeology 30(2):235–245.

Mackay, Alex, and Aara Welz. 2008. Engraved ochre from a Middle Stone Agecontext at Klein Kliphuis in the Western Cape of South Africa. Journal ofArchaeological Science 35:1521–1532.

Wurz Technological Trends in the Middle Stone Age of South Africa S317

Manengold, Albrecht, and James S. Brink. 2011. Descriptions and palaeo-ecological implications of bird remains from the Middle Pleistocene ofFlorisbad, South Africa. Palaontologische Zeitschrift 85:19–32.

Marean, Curtis W. 2010. Pinnacle Point Cave 13B (Western Cape Province,South Africa) in context: the Cape floral kingdom, shellfish, and modernhuman origins. Journal of Human Evolution 59:425–443.

———. 2011. Coastal South Africa and the co-evolution of the modern hu-man lineage and the coastal adaptation. In Trekking the shore: changingcoastlines and the antiquity of coastal settlement. Nuno F. Bicho, JonathanA. Haws, and Loren G. Davis, eds. Pp. 421–440. New York: Springer.

Marean, Curtis W., Miryam Bar-Matthews, Jocelyn Bernatchez, Erich Fisher,Paul Goldberg, Andy I. R. Herries, Zenobia Jacobs, et al. 2007. Early humanuse of marine resources and pigment in South Africa during the MiddlePleistocene. Nature 449:905–908.

Mason, Revil J. 1962. Prehistory of the Transvaal: a record of human activity.Johannesburg: Witwatersrand University Press.

McBrearty, Sally, and Alison S. Brooks. 2000. The revolution that wasn’t: anew interpretation of the origin of modern human behavior. Journal ofHuman Evolution 39:453–563.

McCall, Grant S. 2007. Behavioral ecological models of lithic technologicalchange during the later Middle Stone Age of South Africa. Journal of Ar-chaeological Science 34:1738–1751.

McCall, Grant S., and Jonathan T. Thomas. 2012. Still Bay and HowiesonsPoort foraging strategies: recent research and models of culture change.African Archaeological Review 29:7–50.

McNabb, John, and Peter Beaumont. 2011. A report on the archaeologicalassemblages from excavations by Peter Beaumont at Canteen Koppie, NorthernCape, South Africa. University of Southampton Series in Archaeology 4.Oxford: Archaeopress.

———. 2012. Excavations in the Acheulean levels at the Earlier Stone Agesite of Canteen Koppie, Northern Province, South Africa. Proceedings of thePrehistoric Society 78:51–71.

Mellars, Paul. 2006. Why did modern human populations disperse from Africaca. 60,000 years ago? a new model. Proceedings of the National Academy ofSciences of the USA 103:9381–9386.

Minichillo, Tom. 2005. Middle Stone Age lithic study, South Africa: an ex-amination of modern human origin. PhD dissertation, University of Wash-ington, Seattle.

———. 2006. Raw material use and behavioural modernity: Howiesons Poortlithic foraging strategies. Journal of Human Evolution 50:359–364.

Mitchell, Peter. 2002. Hunter-gatherer archaeology in southern Africa: recentresearch, future trends. Before Farming 1(3):1–36.

———. 2008. Developing the archaeology of marine isotope stage 3. SouthAfrican Archaeological Society Goodwin Series 10:52–65.

———. 2012. San origins and transition to the Later Stone Age: new researchfrom Border Cave, South Africa. South African Journal of Science 108(11/12):1–2.

Mohapi, M. 2012. Point morphology and the Middle Stone Age culturalsequence of Sibudu Cave, KwaZulu-Natal, South Africa. South African Ar-chaeological Bulletin 67(195):5–15.

Moncel, Marie-Helene, Anne-Marie Moigne, Youssef Sam, and Jean Combier.2011. The emergence of Neanderthal technical behavior: new evidence fromOrgnac 3 (Level 1, MIS 8), southeastern France. Current Anthropology 52(1):37–75.

Mourre, Vincent, Paola Villa, and Christopher S. Henshilwood. 2010. Earlyuse of pressure flaking on lithic artifacts at Blombos Cave, South Africa.Science 330:659–662.

North Greenland Ice Core Project Members. 2004. High-resolution record ofNorthern Hemisphere climate extending into the last interglacial period.Nature 431:147–151.

Pahnke, Katharina, Rainer Zahn, Harry Elderfield, and Michael Schulz. 2003.340,000-year centennial-scale marine record of Southern Hemisphere cli-matic oscillation. Science 301:948–952.

Petit, Jean-Robert, Jean Jouzel, Dominique Raynaud, Narcisse I. Barkov, Jean-Marc Barnola, Isabelle Basile, Michael Bender, et al. 1999. Climate andatmospheric history of the past 420,000 years from the Vostok ice core,Antarctica. Nature 399:429–436.

Pickrell, Joseph K., Nick Patterson, Chiara Barbieri, Falko Berthold, LindaGerlach, Tom Guldemann, Blesswell Cure, et al. 2012. The genetic pre-history of southern Africa. Nature Communications 3:1143, doi:10.1038/ncomms2140.

Porat, Naomi, Michael Chazan, Rainer Grun, Maxime Aubert, Vera Eisen-mann, and Liora K. Horwitz. 2010. New radiometric ages for the Fauresmith

industry from Kathu Pan, southern Africa: implications for the Earlier toMiddle Stone Age transition. Journal of Archaeological Science 37(2):269–283.

Porraz, Guillaume, John E. Parkington, Jean-Philippe Rigaud, Christopher M.Miller, Cedric Poggenpoel, Chantal Tribolo, Will Archer, et al. 2013a. TheMSA sequence of Diepkloof and the history of southern African Late Pleis-tocene populations. Journal of Archaeological Science 40:3542–3552.

Porraz, Guillaume, Pierre-Jean Texier, Will Archer, Michel Piboule, Jean-Phi-lippe Rigaud, and Chantal Tribolo. 2013b. Technological successions in theMiddle Stone Age sequence of Diepkloof Rock Shelter, Western Cape, SouthAfrica. Journal of Archaeological Science 40:3376–3400.

Porraz, Guillaume, Pierre-Jean Texier, Jean-Philippe Rigaud, John Parkington,Cedric Poggenpoel, and Dave Roberts. 2008. Preliminary characterizationof a Middle Stone Age lithic assemblage preceding the “classic” HowiesonsPoort complex at Diepkloof Rock Shelter, Western Cape Province, SouthAfrica). South African Archaeological Society Goodwin Series 10:105–121.

Quintana-Murci, Lluis, Helene Quach, Christine Harmant, Francesca Luca,Blandine Massonnet, Etienne Patin, Lucas Sica, et al. 2008. Maternal tracesof deep common ancestry and asymmetric gene flow between Pygmyhunter-gatherers and Bantu-speaking farmers. Proceedings of the NationalAcademy of Sciences of the USA 105:1596–1601.

Sampson, C. Garth. 1968. The Stone Age industries of the Orange River schemeand South Africa. National Museum, Bloemfontein Memoir, no. 6. Bloem-fontein, South Africa: National Museum.

———. 1974. The Stone Age archaeology of southern Africa. New York: Aca-demic Press.

Scheinfeldt, Laura B., Soi Sameer, and Sarah A. Tishkoff. 2010. Working towarda synthesis of archaeological, linguistic, and genetic data for inferring Af-rican population history. Proceedings of the National Academy of Sciences ofthe USA 107(suppl. 2):8931–8938.

Schlebusch, Carina M., Pontus Skoglund, Per Sjodin, Lucie M. Gattepaille,Dena Hernandez, Flora Jay, Sen Li, et al. 2012. Genomic variation in sevenKhoe-San groups reveals adaptation and complex African history. Science338:374–379.

Schmidt, Patrick, Guillaume Porraz, Aneta Slodczyk, Ludovic Bellot-Gurlet,William Archer, and Christopher E. Miller. 2013. Heat treatment in theSouth African Middle Stone Age: temperature induced transformations ofsilcrete and their technological implications. Journal of Archaeological Science40:3519–3531.

Schoville, Benjamin J. 2010. Frequency and distribution of edge damage onMiddle Stone Age lithic points, Pinnacle Point 13B, South Africa. Journalof Human Evolution 59:378–391.

Scott, Louis, and James S. Brink. 1992. Quaternary palaeoenvironments ofpans in central South Africa: palynological and palaeontological evidence.Suid-Afrikaanse Geograaf 19:22–34.

Scott, Louis, and Lloyd Rossouw. 2005. Reassessment of botanical evidencefor palaeoenvironments at Florisbad, South Africa. South African Archae-ological Bulletin 60:96–102.

Shea, John J. 2009. The impact of projectile weaponry on Late Pleistocenehominin evolution. In The evolution of hominin diets: integrating approachesto the study of Palaeolithic subsistence. Jean-Jacques Hublin and MichaelRichards, eds. Pp. 89–199. Dordrecht: Springer.

Sinclair, Anthony, and John McNabb. 2005. All in a day’s work: Middle Pleis-tocene individuals, materiality and the lifespace at Makapansgat, SouthAfrica. In The individual hominid in context: archaeological investigations ofLower and Middle Palaeolithic landscapes, locales and artefacts. Clive Gambleand Martin Porr, eds. Pp. 176–196. London: Routledge.

Singer, Ronald, and John Wymer. 1982. The Middle Stone Age at Klasies RiverMouth in South Africa. Chicago: Chicago University Press.

Sjodin, Per, Agnes E. Sjostrand, Mattias Jakobsson, and Michael G. B. Blum.2012. Resequencing data provide no evidence for a human bottleneck inAfrica during the penultimate glacial period. Molecular Biology and Evolution2012:1–30.

Soriano, Sylvain, Paola Villa, and Lyn Wadley. 2007. Blade technology andtool forms in the Middle Stone Age of South Africa: the Howiesons Poortand post–Howiesons Poort at Rose Cottage Cave. Journal of ArchaeologicalScience 34:681–703.

Steele, Teresa E., and Richard G. Klein. 2009. Late Pleistocene subsistencestrategies and resource intensification in Africa. In The evolution of hominindiets: integrating approaches to the study of Palaeolithic subsistence. Jean-Jacques Hublin and Michael P. Richards, eds. Pp. 113–126. New York:Springer.

Steele, Teresa E., Alex Mackay, Jayson Orton, and Steve Schwortz. 2012. Var-

S318 Current Anthropology Volume 54, Supplement 8, December 2013

sche Rivier 003, a new Middle Stone Age site in southern Namaqualand,South Africa. South African Archaeological Bulletin 67(195):108–119.

Texier, Pierre-Jean, Guillaume Porraz, John Parkington, Jean-Philippe Rigaud,Cedric Poggenpoel, Christopher Miller, Chantal Tribolo, et al. 2010. A How-iesons Poort tradition of engraved ostrich eggshell containers dated to60,000 years ago at Diepkloof Rock Shelter, South Africa. Proceedings of theNational Academy of Sciences of the USA 107:6180–6185.

Texier, Pierre-Jean, Guillaume Porraz, John Parkington, Jean-Philippe Rigaud,Cedric Poggenpoel, and Chantal Tribolo. 2013. The context, form andsignificance of the MSA engraved ostrich eggshell collection from DiepkloofRock Shelter, Western Cape, South Africa. Journal of Archaeological Science40:3412–3431.

Thackeray, Anne I. 1989. Changing fashions in the Middle Stone Age: thestone artefact sequence from Klasies River main site, South Africa. AfricanArchaeological Review 7:33–57.

———. 1992. The Middle Stone Age south of the Limpopo River. Journal ofWorld Prehistory 6(4):385–431.

———. 2000. Middle Stone Age artefacts from the 1993 and 1995 excavationsof Die Kelders Cave 1, South Africa. Journal of Human Evolution 38:147–168.

Thackeray, Anne I., and Alison J. Kelly. 1988. A technological and typologicalanalysis of Middle Stone Age assemblages antecedent to the HowiesonsPoort at Klasies River main site. South African Archaeological Bulletin 43:15–26.

Thackeray, J. Francis. 2009. Chronology, climate and technological innovationassociated with the Howiesons Poort and Still Bay industries in South Africa.South African Journal of Science 105:90.

Thompson, Erin, and Curtis Marean. 2008. The Mossel Bay lithic variant:120 years of Middle Stone Age research from Cape St Blaize Cave to PinnaclePoint. South African Archaeological Society Goodwin Series 10:90–104.

Thompson, Erin, Hope M. Williams, and Tom Minichillo. 2010. Middle andlate Pleistocene Middle Stone Age lithic technology from Pinnacle Point13B (Mossel Bay, Western Cape Province, South Africa). Journal of HumanEvolution 59:358–377.

Tishkoff, Sarah A., Mary K. Gonder, Brenna M. Henn, Holly Mortensen, AlecKnight, Christopher Gignoux, Neil Fernandopulle, et al. 2007. History ofclick-speaking populations of Africa inferred from mtDNA and Y chro-mosome genetic variation. Molecular Biology and Evolution 24:2180–2195.

Tribolo, Chantal, Norbert Mercier, E. Douville, J.-L. Joron, D. Reyss, N. Rufer,N. Cantin, et al. 2013. OSL and TL dating of the Middle Stone Age sequenceat Diepkloof Rock Shelter (South Africa): a clarification. Journal of Ar-chaeological Science 40:3401–3411.

Tribolo, Chantal, Norbert Mercier, and Helene Valladas. 2005. Chronology ofthe Howiesons Poort and Still Bay technocomplexes: assessment and newdata from luminescence. In From tools to symbols: from early hominids tomodern humans. Francesco d’Errico and Lucinda Backwell, eds. Pp. 493–512. Johannesburg: Witwatersrand University Press.

Tribolo, Chantal, Norbert Mercier, Helene Valladas, Jean-Louis Joron, PierreGuibert, Yannick Lefrais, Madeleine Selo, et al. 2009. Thermoluminescencedating of a Stillbay–Howiesons Poort sequence at Diepkloof Rock Shelter(Western Cape, South Africa). Journal of Archaeological Science 36:730–739.

Tryon, Christian A., and J. Tyler Faith. 2013. Variability in the Middle StoneAge of eastern Africa. Current Anthropology 54(suppl. 8):S234–S254.

Van Peer, P. 1992. The Levallois reduction strategy. Monographs in WorldArchaeology, no. 13. Madison, WI: Prehistory Press.

Van Pletzen, Vos L. 2000. The large mammal fauna from Klasies River. Master’sthesis, University of Stellenbosch.

Van Riet Lowe, Clarence. 1945. The evolution of the Levallois technique inSouth Africa. Man 45:49–59.

Van Zinderen Bakker, Eduard M. 1989. Middle Stone Age palaeoenvironmentsat Florisbad (South Africa). Palaeoecology of Africa 20:133–154.

Villa, Paola, Anne Delagnes, and Lyn Wadley. 2005. A Late Middle Stone Ageartifact assemblage from Sibudu (KwaZulu-Natal): comparisons with theEuropean Middle Paleolithic. Journal of Archaeological Science 32:399–422.

Villa, Paola, Marie Soressi, Christopher S. Henshilwood, and Vincent Mourre.2009. The Still Bay points of Blombos Cave (South Africa). Journal ofArchaeological Science 36:441–460.

Villa, Paola, and Sylvain Soriano. 2010. Hunting weapons of Neanderthalsand early modern humans in South Africa. Journal of Anthropological Re-search 66:5–38.

Villa, Paola, Sylvain Soriano, Nicolas Teyssandier, and Sarah Wurz. 2010. TheHowiesons Poort and MSA III at Klasies River Mouth, Cave 1A. Journalof Archaeological Science 37:630–655.

Villa, Paola, Sylvain Soriano, Tsenka Tsanova, Ilaria Degano, Thomas F. G.Higham, Francesco d’Errico, Lucinda Backwell, Jeannette J. Lucejko, MariaPerla Colombini, and Peter B. Beaumont. 2012. Border Cave and the be-ginning of the Later Stone Age in South Africa. Proceedings of the NationalAcademy of Sciences of the USA 109(33):13208–13213.

Visser, Johan N. J., and Alta Joubert. 1991. Cyclicity in the Late Pleistoceneto Holocene spring and lacustrine deposits at Florisbad, Orange Free State.South African Journal of Geology 94:123–131.

Volman, Tom. 1978. Early archaeological evidence for shellfish collecting. Sci-ence 201:911–913.

———. 1981. The Middle Stone Age in the southern Cape. PhD dissertation,University of Chicago.

———. 1984. Early prehistory of Southern Africa. In Southern African pre-history and paleoenvironments. Richard G. Klein ed. Pp. 169–220. Rotter-dam: Balkema.

Wadley, Lyn. 1993. The Pleistocene Later Stone Age south of the LimpopoRiver. Journal of World Prehistory 7:243–296.

———. 2005. A typological study of the final Middle Stone Age stone toolsfrom Sibudu Cave, KwaZulu-Natal. South African Archaeological Bulletin60:51–63.

———. 2007. Announcing a Still Bay Industry at Sibudu Cave, South Africa.Journal of Human Evolution 52:681–689.

———. 2008. The Howiesons Poort Industry of Sibudu Cave. South AfricanArchaeological Society Goodwin Series 10:122–132.

———. 2010. Compound-adhesive manufacture as a behavioral proxy forcomplex cognition in the Middle Stone Age. Current Anthropology 51(suppl.1):S111–S119.

———. 2013. Recognizing complex cognition through innovative technologyin Stone Age and Palaeolithic sites. Cambridge Archeological Journal 23(2):163–183.

Wadley, Lyn, and Philip T. N. Harper. 1989. Rose Cottage revisited: Malan’sMiddle Stone Age collection. South African Archaeological Bulletin 44:23–32.

Wadley, Lyn, Tamaryn Hodgskiss, and Michael Grant. 2009. Implications forcomplex cognition from the hafting of tools with compound adhesives inthe Middle Stone Age, South Africa. Proceedings of the National Academyof Sciences of the USA 106(24):9590–9594.

Wadley, Lyn, and Moleboheng Mohapi. 2008. A segment is not a monolith:evidence from the Howiesons Poort of Sibudu, South Africa. Journal ofArchaeological Science 35:2594–2605.

Watts, Ian. 2010. The pigments from Pinnacle Point Cave 13B, Western Cape,South Africa. Journal of Human Evolution 59:392–411.

White, Mark, Nick Ashton, and Beccy Scott. 2011. The emergence, diversityand significance of Mode 3 (prepared core) technologies. In The ancienthuman occupation of Britain. Nick Ashton, Simon Lewis, and Chris Stringer,eds. Pp. 53–65. Amsterdam: Elsevier.

Wilkins, Jayne, and Michael Chazan. 2012. Blade production ∼500 thousandyears ago at Kathu Pan 1, South Africa: support for a multiple originshypothesis for Early Middle Pleistocene blade technologies. Journal of Ar-chaeological Science 39:1883–1900.

Wilkins, Jayne, Luca Pollarolo, and Kathleen Kuman. 2010. Prepared corereduction at the site of Kudu Koppie in northern South Africa: temporalpatterns across the Earlier and Middle Stone Age boundary. Journal ofArchaeological Science 37:1279–1292.

Wilkins, Jayne, Benjamin J. Schoville, Kyle S. Brown, and Michael Chazan.2012. Evidence for early hafted hunting technology. Science 338:942–946.

Will, Manuel, John E. Parkington, Andrew W. Kandel, and Nicholas J. Conard.2013. Coastal adaptations and the Middle Stone Age lithic assemblages fromHoedjiespunt 1 in the Western Cape, South Africa. Journal of Human Evo-lution 64:518–537.

Wurz, Sarah. 1999. The Howiesons Poort at Klasies River: an argument forsymbolic behaviour. South African Archaeological Bulletin 54(69):38–50.

———. 2000. The Middle Stone Age sequence at Klasies River, South Africa.PhD dissertation, University of Stellenbosch.

———. 2002. Variability in the Middle Stone Age lithic sequence, 115,000–60,000 years ago at Klasies River, South Africa. Journal of ArchaeologicalScience 29:1001–1015.

———. 2008. Modern human behaviour at Klasies River. South African Ar-chaeological Society Goodwin Series 10:150–156.

———. 2010. Middle Stone Age tools from Klasies River main site: conven-tions and symbolic cognition. In The cutting edge: stone tools and the evo-lution of cognition. April Nowell and Iain Davidson, eds. Pp. 135–158.Boulder: Colorado University Press.

Wurz Technological Trends in the Middle Stone Age of South Africa S319

———. 2012. The significance of MIS 5 shell middens on the Cape coast:lithic perspective from Klasies River and Ysterfontein 1. Quaternary Inter-national 270:61–69.

Wurz, Sarah, Niel J. le Roux, Sugnet Gardner, and Hilary J. Deacon. 2003.Discriminating between the end products of the Earlier Middle Stone Agesub-stages at Klasies River using biplot methodology. Journal of Archaeo-logical Science 30:1107–1126.

Wynn, Thomas, and Frederick L. Coolidge. 2011. The implications of theworking memory model for the evolution of modern cognition. Interna-tional Journal of Evolutionary Biology 2011:1–12.

Ziegler, Martin, Margit H. Simon, Ian R. Hall, Stephen Barker, Chris Stringer,and Rainer Zahn. 2013. Development of Middle Stone Age innovationlinked to rapid climate change. Nature Communications 4:1905, doi:10.1038/ncomms2897.

S320 Current Anthropology Volume 54, Supplement 8, December 2013

� 2013 by The Wenner-Gren Foundation for Anthropological Research. All rights reserved. 0011-3204/2013/54S8-0012$10.00. DOI: 10.1086/673861

Change and Stasis in the IberianMiddle Paleolithic

Considerations on the Significance of MousterianTechnological Variability

by Ignacio de la Torre, Jorge Martınez-Moreno, and Rafael Mora

CA� Online-Only Material: Supplements A and B

The European Mousterian has traditionally been portrayed as a long period of technological stasis as opposed tothe technotypological dynamism of Upper Paleolithic cultures. The classic debate on Mousterian variability explainedinterassemblage differences either by ethnic, cultural, functional, and chronological or by paleoenvironmental causes,but variability was based on typological considerations. Recently, technological factors have been introduced indiscussions over time trends and geographic differences in the Mousterian. This paper will address the topic byreviewing technological strategies in the Iberian Middle Paleolithic. Three sites from northeastern Spain are chosenas a case study to address the existence of directional patterns in the Iberian Mousterian. We conclude that albeitdiachronic variability exists, it does not show patterning, which suggests stochastic variation rather than directionalchange in the technological strategies of Iberian Neanderthals.

The notion of variability is at the heart of the study of theMousterian. Spanning ca. 300 kyr and present in hundredsof sites across western Europe, the search for explanations ofinterassemblage variability is consubstantial to Middle Pale-olithic research. The classic debate on the interpretation ofthe Mousterian facies privileged cultural (Bordes 1972), func-tional (Binford 1973), diachronic (Mellars 1969), or climatic(Laville 1973) explanations, but it was strictly typological; thatis, it was based on the assumption that Bordes’s tool typeswere meaningful categories. Since the 1980s, the advent oftechnological approaches in Mousterian research (e.g., Boeda1986; Geneste 1985) and the consideration of tool types asthe result of different stages of reduction (Dibble 1987) ledto explanations of most variability by raw material constraintsand functional or settlement dynamics (e.g., Dibble and Rol-land 1992). New attempts to understand the Mousterian faciesfrom a functional perspective (e.g., Beyries 1988) and an in-

Ignacio de la Torre is Reader in Palaeolithic Archaeology at theInstitute of Archaeology of University College London (31–34Gordon Square, WC1H 0PY London, United Kingdom [i.torre@ucl.ac.uk]). Jorge Martınez-Moreno is Research Associate and RafaelMora is Professor of Prehistory at the Centre d’Estudis del PatrimoniArqueologic de la Prehistoria of the Universitat Autonoma deBarcelona (Facultat de Lletres, 08193 Bellaterra, Spain). This paperwas submitted 3 VII 13, accepted 27 VIII 13, and electronicallypublished 20 XII 13.

novative emphasis on the role of ecological factors in tech-nological behavior (Kuhn 1995) added a novel dimension tothe study of Middle Paleolithic variability.

In recent years, generalization of technological perspectivesin most of western European Middle Paleolithic studies andthe dramatic improvement of dating techniques for chro-nologies beyond the accelerator mass spectrometry (AMS) 14Cmethod have provided new tools to address the question ofvariability and time depth, now with an emphasis on theunderstanding of knapping techniques rather than tool types.Interest in evolution, stasis, and rhythms of change duringthe Middle Paleolithic (e.g., Kuhn and Hovers 2006) has shednew light in Neanderthal cultural adaptations, with pioneeringstudies focused on the technological characterization ofMousterian diachronic variability (Delagnes and Meignen2006) and its functional/ecological correlation (Delagnes andRendu 2011).

The aim of this paper is to apply some of these perspectivesto the Mousterian of the Iberian Peninsula with the specificpurpose of addressing the potential significance of diachronicpatterns in lithic technology. A long history of research, abun-dance of karstic systems susceptible of preserving archaeo-logical deposits, and relatively milder conditions than in mostof Europe during the Pleistocene explains the substantialnumber of Mousterian sites known in Spain and Portugal.However, problems on the chronostratigraphic correlation ofassemblages plus the variety of nomenclatures and analytical

de la Torre, Martınez-Moreno, and Mora Change and Stasis in the Iberian Middle Paleolithic S321

perspectives employed make it difficult to produce overallassessments such as those currently available for adjacentregions (Delagnes and Meignen 2006; Jaubert 2011) and theNear East (Goren-Inbar and Belfer-Cohen 1998; Hovers andBelfer-Cohen 2013).

In order to overcome the sample disparity derived fromsuch methodological and empirical pitfalls, three Iberian sitesexcavated and studied within the same research program (Ca-sanova et al. 2009; Martınez-Moreno et al. 2010; Mora et al.2011) in the area of La Noguera (northeast Spain) are usedin this paper to search for temporal trends in the MiddlePaleolithic. Our results suggest that although diachronic var-iation is detected, it does not follow any particular pattern.This case study from La Noguera will then be discussed withinthe context of the Iberian Mousterian, in which a lack of timetrends also prevails. Instead, the observed technological var-iability could be associated with intrasite specifics and, po-tentially, regional idiosyncrasies. This apparent lack of tem-poral patterns in the Iberian Middle Paleolithic contrasts withneighboring areas such as France, where chronological trendsseem to exist (e.g., Delagnes and Meignen 2006). By com-bining different units of analysis (from particular case studiesin northeast Spain to the regional scale of Iberia and to itscontextualization within the sequence of southwestern Eu-rope), the purpose of this paper is to contribute to the debateon technological change and stasis during the Middle Pale-olithic and the nature and causes of Mousterian technologicalvariability.

The Middle Paleolithic of La Noguera:A Case Study on the Analysis ofTechnological Diachronic Patterns

Trago, Roca dels Bous, and Cova Gran are three rocksheltersin the region of La Noguera (Lleida, Catalunya) located withina radius of less than 20 km at the Marginal Exterior Sierrasof the Eastern Pre-Pyrenees in the northeast of the IberianPeninsula (CA� Online Supplement A: fig. A1). Trago hasyielded eight archaeological levels, all attributed to the Mous-terian. Thermoluminescence dating (CA� Online Supple-ment B: table B1) situates the bottom layer (UA3) at ap-proximately 126 kyr and the uppermost (UA1) level of themain excavation area at ca. 42 kyr (Casanova et al. 2009).Ongoing excavations in Roca dels Bous have yet to reach thebedrock, and so far four main levels (R3, N10, N12, and N14)have been unearthed (Martınez-Moreno et al. 2010; Mora,de la Torre, and Martınez-Moreno 2004), all correspondingto the Mousterian. R3 was dated to kyr, while38.8 � 1.2underlying levels seem to be placed beyond the range of AMS14C (see table B1). Cova Gran is a huge (12,500 m2) rockshelterwith Upper Pleistocene and early Holocene deposits (Moraet al. 2011). Four Mousterian levels (S1B, S1C, S1D, and S1E)have been documented so far, situating Cova Gran among

the most recent Middle Paleolithic sites in northern Iberia,perhaps as late as 33–32 kyr (Martınez-Moreno, Mora, andde la Torre 2010).

The three sites share a number of features enabling inter-assemblage comparisons. Both Roca dels Bous and Trago aresituated over two main river valleys that are natural corridorsbetween the Ebro plain and the first ranges of the Pyrenees(fig. A1), while Cova Gran is located in a small subsidiaryvalley about 10 km north from Roca dels Bous. Pleistocenebiotopes should have been similar in the three sites, withvariations in the diversity of game according to climatic pulsesbut comparable density and availability of biotic resources.Availability of raw material was also similar; flint of poorquality is present in the immediate surroundings of Cova Granand Trago but not in the proximity of Roca dels Bous. Goodquality flint was not locally available in any of the three sites.Quartzite and other metamorphic cobbles were readily avail-able from river beds in the three sites, particularly in Rocadels Bous and Trago.

In order to evaluate the possible existence of temporaltrends, in this paper the Roca dels Bous (Martınez-Moreno,Mora, and de la Torre 2010; Mora, de la Torre, and Martınez-Moreno 2004), Cova Gran (Martınez-Moreno, Mora, and dela Torre 2010), and Trago (Casanova 2009) assemblages willbe organized diachronically; Trago levels cover oxygen isotopestage (OIS) 5e (UA3), OIS 5/4 (UA2), and OIS 3 (UA1). Rocadels Bous levels are probably within the same chronologicalrange as UA1, but for the sake of comparison, N12 and N10have been placed after the sequence of Trago. Cova Gran levelsS1D, S1C, and S1B have yielded consistent dates post 40 kyrand therefore are considered here to belong to the latter partof OIS 3.

The artefacts ( ) analyzed from the three sitesN p 59,001are distributed unevenly, with 23,557 lithics from Trago,19,569 from Roca dels Bous, and 15,875 from Cova Gran(CA� Supplement B: table B2). Relative frequencies of mainstone tool types are shown in figure 1. The percentage of coresdoes not indicate any clear patterning; while cores are con-sistently less abundant in Cova Gran than in the two oldersites, N10 at Roca dels Bous yields the highest proportion ofcores of all levels. Therefore, the comparative paucity of coresin the Cova Gran levels is better explained by site-specificeconomic variables rather than by temporal trends. In con-trast, frequency of flakes seems to decrease steadily through-out the sequence, with the exception of S1C in Cova Gran(fig. 1B). The meaning of such consistent reduction on thefrequency of flakes is obscure and could tentatively be relatedto recurrent export of flakes off site. A divide seems to existin the sequence with regard to the frequency of retouchedtools; while the four older assemblages yield very low fre-quencies, all levels from N10 to S1B contain percentages ofretouched tools over 7% (fig. 1C). Even though no cumulativeincrease is observed across the four younger levels, the factthat a higher frequency of retouched tools is present in twosites (Roca dels Bous and Cova Gran) rules out the possibility

Figure 1. Frequency and size of main categories in Trago (Ua3, Ua2, Ua1), Roca dels Bous (N12, N10), and Cova Gran (S1D, S1C,S1B). A, Percentage of cores. B, Percentage of whole flakes. C, Percentage of retouched tools. D, Average length of cores. E, Elongation(length divided by width) of whole flakes. F, Mean length of retouched tools. G, Mean length of whole flakes. H, Average lengthof cores, whole flakes, and retouched tools. All data from CA� Online Supplement B: table B2. A color version of this figure isavailable in the online edition of Current Anthropology.

de la Torre, Martınez-Moreno, and Mora Change and Stasis in the Iberian Middle Paleolithic S323

that this pattern was site-specific and could suggest someintensification in the shaping of blanks among latest Pre-Pyrenean Neanderthals.

Metric analysis of the main categories (fig. 1; CA� Sup-plement B: table B3) reflects consistency on lithic dimensionsthroughout the sequence. Average length of cores and re-touched tools is remarkably similar from UA3 to S1B (fig.1D, 1F), and even though flakes are larger in the oldest as-semblages (fig. 1G), the length/width ratio (fig. 1E) of Tragoflakes resembles that of Cova Gran, which yields the mostrecent layers. The widely documented pattern of larger blankselection for retouching during the Mousterian (e.g., Dibbleand Rolland 1992; Geneste 1985) is also verified in each andevery one of the assemblages analyzed here, which show con-sistent larger size of retouched tools over unmodified blanks.

In general, data from figure 1 (see also table B3) do notindicate any kind of temporal trend but rather remarkablyhomogeneous size range for the main lithic categories acrossall assemblages. The only possible divergence is shown byN10 at Roca dels Bous, whose cores and flakes are noticeablysmaller than those in the rest of the levels. However, the short-term nature of this occupation (Martınez-Moreno, Mora, andde la Torre 2004; Mora, de la Torre, and Martınez-Moreno2004) rather than any chronological pattern explains partic-ularities of N10 more satisfactorily.

In order to explore the possible presence/existence ofdiachronic patterns on knapping methods, nearly 500 coreswere considered in our analysis (CA� Supplement B: tableB4). Levallois cores are poorly represented, with only N12and S1C yielding relevant percentages. Bifacial hierarchicalcentripetal (sensu Casanova et al. 2009; de la Torre and Mora2004) and unifacial methods are the most common in severalassemblages, whereas discoid cores sensu stricto are rare. Fig-ure 2A suggests that neither Levallois nor any other flakingmethods show temporal trends. To avoid subjectivities derivedfrom flaking methods classification, a comparison (fig. 2B)was made between expedient cores—those that bear only afew scars (e.g., unifacial) and which underwent limited and/or unstandardized (e.g., multifacial) reduction—versus struc-tured cores—those including several stages of preparationand/or production (e.g., Levallois, bifacial hierarchical cen-tripetal and discoid). Albeit expedient methods are rarelymentioned in the literature, they are common in the Pre-Pyrenean Mousterian and indicate opportunistic reduction ofpart of the raw material stocks. However, once again no tem-poral pattern is discerned in the variation between expedientand structured methods: while structured methods tend toprevail (as expected from any Mousterian assemblage), ex-pedient flaking is common at the beginning (UA3), middle(N10), and end (S1B) of the sequence.

Low variability of retouched tool types is well attested acrossthe northeastern Iberian Mousterian sites (Mora 1988) andis confirmed in the Pre-Pyrenees sequence. Figure 2C showsthat sidescrapers and denticulates predominate, while points,endscrapers, and other retouched tools are represented resid-

ually (see also CA� Supplement B: table B5). There seem tobe some differences between Trago and the other two sites;while in Trago nearly all shaped tools are either denticulatesor sidescrapers, both in Roca dels Bous and Cova Gran, highertypological variability is attested, with points, endscrapers, andother tool types present in all levels. Trago also consistentlyyields higher proportions of sidescrapers than Roca dels Bousand Cova Gran. However, it is unclear whether such a patternresponds to temporal trends or whether it could be explainedinstead by settlement dynamics specific to the earlier site ofTrago.

It remains to assess the role of raw materials in discerningpotential time trends. Figure 2D suggests an overall preferencefor flint across the sequence in all levels apart from N12.Interestingly, the four more recent layers show clear preferencefor flint, whereas the earlier assemblages of Trago yield a moremixed procurement of flint and quartzite (see also CA� Sup-plement B: table B6). However, the exception of N12 in Rocadels Bous, where the raw material procurement strategy iscompletely reversed to that of the later levels, precludes es-tablishing a clear diachronic pattern throughout the assem-blages. The same applies to lithic categories with particulartechnological relevance; figure 2E indicates that with the ex-ception of N12, flint cores prevail in all assemblages. Similarly,flint was more often selected for structured knapping thanquartzite. No temporal trend can be observed, either, in theselection of raw material for retouched tools; while the youn-ger levels of Cova Gran reflect a strong preference for flintover quartzite, such preferential selection is exacerbated inUA1 at Trago, one of the earlier assemblages. Althoughquartzite retouched tools in N12 are more abundant thanflint in absolute terms, flint was preferentially used for re-touching, which is consistent with the pattern observed in therest of the sequence (fig. 2F).

Chronological Patterns of the IberianMiddle Paleolithic

Although better known for putatively yielding the last Ne-anderthal traces anywhere in Europe (e.g., Finlayson et al2006; Zilhao 2008), the Iberian record also contains a numberof Middle and early Upper Pleistocene sequences. In principle,this should provide the time depth required to evaluatediachronic trends in the Iberian Mousterian. Unfortunately,the wealth of Middle Paleolithic sites in Iberia (fig. 3) is notaccompanied by a solid chronostratigraphic framework; inrecent years, research agendas have targeted dating of the lastMousterian and the transition to the Upper Paleolithic(d’Errico and Sanchez-Goni 2003; Maroto et al 2012; Mar-tınez-Moreno, Mora, and de la Torre 2010; Zilhao 2006),whereas older sequences are poorly constrained or simply lackany radiometric dating. Thus, figure 4 must be seen as a verypreliminary attempt to order chronologically some of the rel-evant sites in the Iberian Mousterian.

S324 Current Anthropology Volume 54, Supplement 8, December 2013

Figure 2. Temporal trends in the technology and raw materials of Trago, Roca dels Bous, and Cova Gran. LEV p Levallois;DIS p discoid; BHC p bifacial hierarchical centripetal; UF p unifacial; MF p multifacial. A, Relative frequency of main corereduction methods. B, Relative frequency of structured (LEV, BHC, DIS) versus expedient (bipolar, MF, UF) reduction methods.C, Relative frequency of main retouched types. All data from CA� Online Supplement B: tables B4, B5. D, Breakdown of rawmaterials per assemblage. E, Raw material breakdown of quartzite and flint cores. F, Raw material breakdown of quartzite and flintretouched tools. All data from CA� Online Supplement B: table B6. A color version of this figure is available in the online editionof Current Anthropology.

The level TD 10.1 at Atapuerca, dated to kyr,337 � 29could represent the earliest evidence of Middle Paleolithictechnology in Iberia (Rodrıguez 2004). Its large lithic andfossil assemblage shares some elements of continuity with theAcheulean (e.g., handaxes), but cores and retouched flakesindicate more standardized systems typical of the Mousterian(Olle et al 2013). Cueva Hora and Cueva del Angel, both insouthern Spain, may potentially contain very old Mousterian

assemblages, but a robust radiometric frame has yet to bedeveloped. At present, the best documented cultural succes-sion for the latter part of the Middle Pleistocene is that fromBolomor, in eastern Spain. Here, radiometric dates bracketbetween OIS 9 and OIS 5e more than a dozen archaeologicalunits in which denticulate and sidescraper-rich layers alternateand no handaxes are recorded (Fernandez-Peris 2007). Theearliest levels of Bolomor (XVII–XV) are positioned between

de la Torre, Martınez-Moreno, and Mora Change and Stasis in the Iberian Middle Paleolithic S325

Figure 3. Geographic locations of the main Middle Paleolithic sites in Iberia. A color version of this figure is available in the onlineedition of Current Anthropology.

347 and 242 kyr and are considered as early Middle Paleolithic(Fernandez-Peris et al. 2008), which is in agreement with datafrom Atapuerca TD 10.1.

Both the Bolomor sequence and Atapuerca TD 10.1 pointto an emergence of the Middle Paleolithic earlier than 300,000BP. This would be in agreement with Mousterian-like featuresseen in late Acheulean open-air sites such as Ambrona AS6,dated over 350 kyr (Santonja and Villa 2006), but it is at oddswith the chronology of other late Acheulean sites in the Man-zanares (e.g., Arriaga, Oxıgeno), Guadiana, and Duero valleys,for which terminal Middle Pleistocene ages have been pro-posed. This calls for either a long coexistence of Acheuleanand Mousterian technologies between 300 and 100 kyr, agradual disappearance of handaxes in early Middle Paleolithicsequences as recorded elsewhere (e.g., Moncel et al 2011),and/or for a need to reconsider the chronology of some ofthe central Iberia terraces, which is in reality poorly con-strained.

Other pre-OIS 5e sequences in caves/rockshelters presentsimilar dating problems. Lezetxiki in northern Spain hasyielded dates 1240 kyr for level VII (Falgueres, Yokoyama,and Arrizabalaga 2005), but most of the other radiometricages are inconsistent. Cueva de los Aviones, Cueva de lasGrajas, and Cueva Hora, all in the southern half of Spain,contain Mousterian (and in the case of Cueva Hora alsoAcheulean) layers attributed to the Middle Pleistocene, butno reliable radiometric dates have been published. It is likelythat several other caves from figure 4 contain pre-OIS 5eMousterian deposits, but at present, apart from Bolomor, onlya few yield radiometric dates. Two of those caves are Carihuelaand Trago (southern and northeastern Spain, respectively);lower deposits from Carihuela range between 146 and 117kyr (Vega et al. 1997) and show clear Mousterian features,while the bottom of the Middle Paleolithic sequence in Tragois radiometrically dated to the beginning of the last interglacial(Casanova et al. 2009).

Figu

re4.

Tim

esp

anof

rele

van

tM

oust

eria

nsi

tes

inIb

eria

(iso

topi

ccu

rve

adap

ted

from

Gam

ble

1986

).G

iven

the

inco

nsi

sten

cies

ofra

diom

etri

cre

sult

sin

an

um

ber

ofas

sem

blag

esan

dth

ela

ckof

date

sin

seve

ral

oth

ers,

this

grap

hic

shou

ldbe

con

side

red

only

asa

very

prel

imin

ary

atte

mpt

toor

gan

ize

tem

pora

llyth

eIb

eria

nM

iddl

eP

aleo

lithi

c.

de la Torre, Martınez-Moreno, and Mora Change and Stasis in the Iberian Middle Paleolithic S327

Figure 4 could lead us to believe that the majority of Iberiancave sites have no Mousterian occupation for most of OIS 5.However, that is likely to be an artefact of dating methods(in many sites no techniques other that 14C have been at-tempted), and research program contingencies (often the bot-tom of stratigraphy has not yet been reached, and lower,unreported levels might be present). Focusing on the few sitesfor which OIS 5 levels are published, some of those alreadymentioned contain Early Glacial deposits, and several otherscan also be considered (see fig. 4). Among the latter, El Castilloand Cova Negra have yielded relatively consistent radiometricdates across thick sequences; in El Castillo, the classic Mous-terian is capped between 70 kyr (the age of level 22) and 39kyr (dating proposed for the Early Aurignacian of level 18;Rink et al. 1997). Cova Negra contains a thick Mousterianstratigraphy spanning the whole of the Early Glacial (OIS5d-a), OIS 4, and part of OIS 3. Although radiometrically notas robust as Cova Negra, the Mousterian sequence of CuevaBajondillo is also constrained between OIS 5 and OIS 3 (Cor-tes Sanchez et al. 2007), and so is the thick stratigraphy ofCovalejos, dated between 100 and 40 kyr (Montes and San-guino 2005), and Oliveira, dated between 170 and 35 kyr(Angelucci and Zilhao 2009).

According to the available dates, most of Mousterian sitesin Iberia would correspond to the latest part of the MiddlePaleolithic. While the possibility of a sudden proliferation ofNeanderthal sites after 45 kyr cannot be excluded, it is alsovery likely that the radiocarbon dating threshold rather thanan actual gap in the archaeological record explains the absenceof longer spans in the OIS 3 assemblages plotted in figure 4.Similarly, the very recent radiocarbon dates obtained in anumber of Iberian sites may be due to problems of the limitsof AMS 14C; therefore, many of the late Mousterian datesfrom Iberian sites could be seen as minimum ages (e.g., Mar-tınez-Moreno, Mora, and de la Torre 2010).

Leaving aside chronometric problems, the Iberian recordcontains some firmly dated sequences that account for con-tinuous occupation throughout the first half of OIS 3. Prob-ably the most outstanding is Abric Romanı, which yields well-constrained Mousterian levels between 60 and 40 kyr(Bischoff, Julia, and Mora 1988). Apart from several of thesites mentioned above, Esquilleu, El Salt, Vanguard Cave, anda few others (see fig. 4) also yield levels 145 kyr overlaid bylate Mousterian assemblages. They all attest to thick sequencesof Middle Paleolithic and, in principle, should enable assessingdiachronic patterns in Mousterian technology, as discussed inthe following section.

Technological Trends in theMousterian of Iberia

If chronostratigraphic correlations are weak because of a lackof radiometric dates (especially for levels beyond the rangeof AMS 14C), interassemblage comparisons of the Iberian

Mousterian become even more difficult for several other rea-sons. First, many of the classic studies of the Iberian MiddlePaleolithic were conducted within the typological approach,and therefore technological data are only available from recentpublications. Even when technological studies are available,there is a focus on the late Mousterian and its comparisonwith the early Upper Paleolithic. In consequence, availableinformation is biased to the detriment of older, pre-OIS 3Mousterian assemblages, for which technological data are of-ten very scarce. There is also a tendency toward qualitativedescriptions against quantitative data that sometimes makesstatistical comparisons unfeasible. More importantly, there isa generalized lack of consistency on the methodology em-ployed to classify knapping methods and technological pat-terns; different conceptions exist on the meaning of discoid,Levallois, other centripetal cores, expedient flaking, and soforth, and therefore similar reduction methods are namedwith different terms. The opposite also occurs, with termssuch as “discoid” used to classify cores that could be consid-ered as recurrent Levallois or alike according to recent tech-nological conceptions. All of this makes it difficult to con-textualize the Pre-Pyrenean diachronic patterns within theIberian Middle Paleolithic sequence and constrains the num-ber of case studies that provide comparable data. Further-more, available data from other case studies elsewhere in Ibe-ria do not necessarily comprehend every technological aspect,which precludes a unified assessment of all analytical criteria(e.g., raw material selection, tool types, core reduction tech-niques, etc.) across the Iberian chronostratigraphic sequence.Hence, the patchy character of the available data determinesthe comparisons below, which are built on a composite viewof a number of Iberian case studies.

At present, typological classifications of the Iberian MiddlePaleolithic are unusual, and so the Bordesian taxonomy oftable 1 is limited to data from some 1960s–1980s studies. Asexpected, no diachronic patterns are observed in the variationof Mousterian facies; in Cova Negra, one of the few datedMiddle Paleolithic sequences, Quina-type Charentian was rec-ognized at the bottom (level XIV: OIS 5d-b) and middle(XIII–XII, X–IX, VII–VI: OIS 4) of the stratigraphy, and thesame occurs with the para-Charentian/typical Mousterian,identified both in OIS 4 (level XI, VIII) and at the beginningof OIS 3 (Villaverde 1984). While Cova Negra can be usedas an instance of nondirectional temporal variability of Mous-terian facies, Axlor provides an example of homogeneity oftool types through time; here, all levels are classified as typicalor Quina Charentian (Baldeon 1999) despite the thick stra-tigraphy of the rockshelter. Despite reservations derived fromthe absence of radiometric dates, Axlor and several othersequences from table 1 can be placed at the onset of OIS 3(see fig. 4). Given the variety of Mousterian facies representedin this time interval, it may be concluded that no temporalpatterning exists.

The relative typological homogeneity of the Iberian Mous-terian in comparison with the neighboring sequence of France

S328 Current Anthropology Volume 54, Supplement 8, December 2013

Table 1. Mousterian facies in some classic Spanish assemblages

Denticulate Mousterian Charentian Typical Mousterian MTA Mousterian with cleavers

Conde 6 Castillo Beta (level 22) Casares Cova Negra V Morın 13/14, 15–17Morın 17b2, 11–5 Hornos de la Pena Cova Negra XI Abauntz Castillo Alfa (level 20)La Flecha Zajara Cova Negra IV–I Pendo XIIIPendo 4 Cova Negra XIV–XII Lezetxiki VII–V AmaldaRomanı Cova Negra X–IX Amalda VII Lezetxiki V–VI

Cova Negra VII–VI Morın 17a–17b1, 16–13 GatzarriaLezetxiki IV Pendo 12–6Pena Miel MolletCueva MillanLa ErmitaMorın 12Pendo 5AxlorErmitaEudoviges

Sources. Baldeon 1993, 1999; Butzer 1981; Cabrera 1984; Freeman 1966; Ripoll and de Lumley 1965; and Villaverde 1984.Note. MTA p Mousterian of Acheulean tradition.

has sometimes been explained by limited availability of goodraw materials in Spain and Portugal. While flint is ubiquitousaround a number of sites in Andalusia and in some parts ofthe Mediterranean, quartzite and other metamorphic rockspredominate in large parts of western, central, and northernIberia. Despite this unequal distribution of raw materials, wecan discuss whether any diachronic patterns exist in the Mid-dle Paleolithic procurement strategies. Again, data are notavailable for all assemblages, so here we selected case studiesin which both raw materials and their availability in the land-scape have been investigated (CA� Supplement B: table B7).In the early Middle Paleolithic of TD 10.1, local flint is thepredominant raw material, but a particular type of exotic flintis also documented for the first time in the Atapuerca se-quence (Olle et al. 2013), providing yet another element ofdifferentiation between the previous Acheulean and TD 10.1.The thick sequence of Bolomor shows variations in raw ma-terial procurement (see fig. 5), but there seems to be no tem-poral patterning, and changes are related to climatic pulsesduring the Middle Pleistocene; small marine flint pebbles werepreferentially selected when locally available during sea trans-gressions in detriment of other raw materials such as quartzite(accessible in the early part of the sequence) and especiallylimestone, an immediately available rock (Fernandez Peris etal. 2008).

In the case of Abric Romanı, the location of raw materialsources was more fixed, with flint available in a radius of 5–10 km and limestone and quartz in the immediate (1 km)surrounding of the rockshelter. Although from farther dis-tances, flint was consistently preferred in most of the Romanılevels (Vaquero 1999). Raw material variation is documentedthroughout the sequence, but figure 5 shows no directionalitytoward any particular rock type. The Mousterian sequencefrom El Esquilleu probably spans more than 20 kyr, but sub-stantial uniformity in raw material procurement is foundacross the levels (Manzano et al. 2005); percentage of quartz-

ite, the most commonly used rock (table B7), varies through-out the sequence, but again no particular trend is observed(fig. 5). Most of the rocks were procured from a stream bed200 m from the cave, and 99% of raw materials were foundwithin a 5-km radius (Manzano et al. 2005), although somestone tools are reported to come from more distant sources(Carrion et al. 2008).

Cova Gran and Trago present an analogous pattern to thesites mentioned above; flint was readily available in Cova Granand Trago and consistently preferred in both sequences. Thecase of Trago is particularly enlightening: UA3 and UA1 areca. 80 kyr apart, and yet raw material procurement for eachis remarkably similar (see table B7). Among the case studiesselected for figure 5, only Roca dels Bous and Axlor suggestsome divergence; in the case of Roca dels Bous, N10 presentsopposite trends to N12 in the use of local (quartzite) versusimported (flint) rocks. However, the current lack of data forunderlying levels makes it difficult to ascertain whether suchreversal in the use of raw material was episodic or insteadcorresponds to a temporal trend in the sequence. With regardto Axlor, Rıos (2008) situates most of the flint sources between15 and 30 km from the site, whereas other raw material—such as quartz, lutite, and so forth—were local (!10 km).Based on data from Baldeon (1999), it could then be statedthat the sequence of Axlor sees a steady increase of exoticraw materials (fig. 5). This diachronic pattern could, poten-tially, be linked to increasingly larger foraging ranges and/ormore demanding manufacture processes that required higherquality raw materials. Time trends have also been detected insome Portuguese Mousterian sites, but they point in the op-posite direction; according to Zilhao (2001), in the sequenceof Gruta da Oliveira, there is a steady decrease in the use offlint, and Gruta do Caldeirao shows a similar trend, which isrelated to mobility patterns focused on the exploitation oflocal quartzite sources.

With some exceptions (e.g., Rıos 2008; Vaquero et al. 2012),

Figure 5. Raw material percentages in selected Iberian Middle Paleolithic assemblages. All data from CA� Online Supplement B:table B7. A color version of this figure is available in the online edition of Current Anthropology.

S330 Current Anthropology Volume 54, Supplement 8, December 2013

relationships between raw material procurement and otheraspects of Neanderthal ecology (e.g., Geneste 1985; Kuhn1995) are yet to be explored in the Iberian Middle Paleolithic,so it is not easy to assess the role of other subsistence activitiesin raw material acquisition and how specific foraging strat-egies may affect temporal patterns. Nonetheless, with thiscaution in mind, time trends seem to be an exception ratherthan the norm; most of the case studies discussed here showconservative patterns in the acquisition of raw materials, inwhich fluctuations do exist (not unexpectedly, for in severalinstances we are dealing with sequences spanning manythousands of years), but such variations may be explained byenvironmental constraints (e.g., Bolomor) or contingent set-tlement dynamics (Roca dels Bous, Romanı) rather than bydirectional trends toward local or exotic raw materials.

After excluding the existence of diachronic trends in ty-pological and raw material patterns, we shall now discusswhether patterns in the variability of flaking methods can betraced. As mentioned above, this issue is particularly difficultto address; one reason is that disparity in the conceptions ofLevallois, discoid, and other centripetal methods obscure po-tential interassemblage comparison of structured flakingmethods. On the other hand, a fairly large number of availablestudies focus on such structured techniques, and simpler,more expedient flaking solutions are frequently excluded fromthe analysis. While consideration of Levallois and discoid be-comes on occasion sophistic and might not help to discerngeneral patterns (de la Torre 2009), differentiation betweenstructured/long reduction sequences and expedient/short se-ries cores (Casanova et al. 2009) may provide a more effectiveground for time-depth comparisons. Such a comparativeframe, nonetheless, is only available for a few sequences, sodiscussion will be restrained here to some particular case stud-ies in which comparable technological data exist.

Discoid-like methods are known in Iberia since the LowerPleistocene; Vaquero and Carbonell (2003) consider the re-current bifacial centripetal method of Atapuerca TD 6 to befairly similar to discoid flaking and report an increase of thistechnique throughout the sequence. According to Olle et al.(2013), centripetal strategies progressively become more stan-dardized, leading to the Levallois-like technique from UpperTD 10.1 (OIS 9), in which morphometrical predeterminationis observed. The roughly contemporary level XVII of Bolomorcontains poor evidence of Levallois, while in upper levels suchas XII–VII (OIS 6), hierarchization of cores (recurrent cen-tripetal Levallois and/or hierarchical discoid) is attested (Fer-nandez Peris et al. 2008). In OIS 5e, Bolomor levels VI–Icontain discoid and Levallois flaking accompanied by trifacialand Kombewa cores (Fernandez Peris et al. 2008). Trago UA3,with a similar chronology, shows predominance of expedientmethods followed by bifacial hierarchical centripetal (BHC)reduction, the latter being potentially similar to the recurrentbifacial centripetal method of Atapuerca (Vaquero and Car-bonell 2003) and to many of the cores described as discoidand Levallois in Bolomor.

Most of the data on knapping systems correspond to OIS4 and especially OIS 3. In the central Mediterranean region,Fernandez-Peris et al. (2008) state that Mousterian assem-blages of this chronology show predominance of Levallois asopposed to earlier periods. An example is the long sequenceof El Salt, dated between 60 and 40 kyr, which shows prev-alence of recurrent centripetal Levallois methods (Galvan etal. 2006). A different pattern is reported in Cueva Bajondillo(southern Spain), where discoid flaking is better representedin the upper levels, albeit Levallois is common across thesequence (Cortes Sanchez 2008).

Detailed reports of flaking techniques are now available fora number of north Iberian OIS 4/3 assemblages. In Catalunya,discoid and discoid-like (e.g., BHC) methods are omnipresentin the Middle Paleolithic as opposed to lower frequencies ofLevallois (Mora 1988). Although Cantabria follows a similarpattern, this region shows higher variability than Catalunya,which is probably explained by the larger number of sitesdocumented. Carrion et al. (2008) state that although Leval-lois is not abundant, it is present in most of Cantabrian sites,usually in the recurrent centripetal modality. Discoid corespredominate and are often made on flakes (Carrion et al.2008), and Kombewa flaking is also attested (e.g., Rıos 2008).Quina flaking is now identified in El Esquilleu (Carrion et al.2008), Axlor, Gatzarri, Lezetxiki, and Amalda (Rıos 2008),but bearing in mind the recentness of the definition of thiscore reduction technique (Bourguignon 1997), it will be un-surprising if over the next few years Quina flaking is docu-mented elsewhere. Bladelet production is suggested at CuevaMorın, El Castillo, and Covalejos (Bernaldo de Quiros, San-chez-Fernandez, and Maıllo 2010) within Mousterian levelswhere no traces of Upper Paleolithic are reported.

Indistinctive Quina and discoid flaking are documented inquartzite and flint, while Levallois is consistently made in thehighest quality raw materials available (Carrion et al. 2008).Beyond recurrence of this pattern, and despite the wealth ofMousterian assemblages in the Cantabrian region, no clearcorrelations are yet available between knapping methods andother contextual elements. A particularly telling recent ex-ample illustrates the impenetrability of the meaning of flakingmethods variability; at El Sidron, the lithic assemblage wasaimed at one very specific activity, that is, butchering otherNeanderthals. Yet both discoid and Levallois cores were flakedin this single and task-specific event (Santamarıa et al. 2010).

While acknowledging pitfalls derived from the disparity ofcore classification systems and chronometric problems of as-semblages, we have compiled a few of the available case studiesin order to assess diachronically north Iberia Mousterianknapping methods (CA� Supplement B: table B8). Somesequences such as Gabasa indicate no change in primary flak-ing techniques. Abric Romanı does show diachronic pattern-ing toward higher frequency of Levallois in the upper levels,whereas variations in El Esquilleu are nondirectional: discoiddominates the bottom and the top of the sequence, whileQuina and Levallois methods are the commonest in some of

de la Torre, Martınez-Moreno, and Mora Change and Stasis in the Iberian Middle Paleolithic S331

Figure 6. Chronostratigraphic position of flaking techniques in the OIS 4/3 assemblages from CA� Online Supplement B: tableB8.

the middle levels. Therefore, no clear patterns emerge fromthe diachronic assessment of intrasite variation.

With regard to intersite variability, figure 6 attempts toorder chronologically the main flaking systems in the OIS 4–OIS 3 case studies (see also table B8). Late OIS 4/early OIS3 assemblages such as the lower levels of El Esquilleu, AbricRomanı, and (perhaps) Gabasa show predominance of discoidmethods. If we were to assume that dates at the limit of theradiocarbon method are correct, then the trend would be

more variable in the 45–39 kyr interval; Levallois predomi-

nates in El Castillo level 20e, top levels of Abric Romanı,

Axlor N, and Arrillor level emj. Nonetheless, in a similar

chronological span, discoid and discoid-like (e.g., BHC) tech-

niques are the primarily flaking methods in Gabasa, Trago

UA1, Roca dels Bous level N12, and Morın levels 11 and 13,

while Quina predominates in El Esquilleu levels XIII–XVI and

Axlor B–D. The pattern is not clearer in post-39-kyr assem-

blages; Quina, Levallois, and discoid follow one another in

the El Esquilleu top levels, while BHC flaking predominates

in Cova Gran S1D–S1C and unifacial methods in Cova Gran

S1B and Roca dels Bous N10. In short, figure 6 certifies for

technical systems the same pattern (or rather, the lack of any)

discussed above for typological facies and raw material pro-

curement: an absence of directional changes or temporal

trends.

Technological Variability of the IberianMiddle Paleolithic in the Context ofSouthwestern Europe

The case study of the Middle Paleolithic at La Noguera regionhas been presented above to examine possible time trends inthe technology of the Iberian Mousterian. Our results suggestthat albeit changes occur, they do not show directional pat-terns. Framing the three Pre-Pyrenean case studies within thewider context of the Iberian Mousterian, although hinderedby methodological and empirical constraints, seems to suggestthat the lack of diachronic patterning in technological strat-egies is applicable to most of the Portuguese and SpanishMiddle Paleolithic. In truth, this is hardly surprising andcomes to confirm on technological grounds what typologicalapproaches (e.g., Bordes 1972) had long reported; that is, asubstantial part of western European Middle Paleolithic as-semblages show no defined time trends.

However, it is important to stress that absence of temporaldirectionality does not mean randomness of variation. In-trasite variability in Roca dels Bous (Martınez-Moreno, Mora,and de la Torre 2004; Mora, de la Torre, and Martınez-Moreno2004), Bolomor (Fernandez Peris 2007), Romanı (Vaquero1999), Axlor (Rıos 2008), El Esquilleu (Baena et al. 2005),and others may respond to alternative settlement dynamicspotentially related to distinct foraging strategies by Neander-

S332 Current Anthropology Volume 54, Supplement 8, December 2013

thal groups. A challenge in current studies of intrasite vari-ability is to decipher why changing technological behaviorsare documented in contexts where raw material sources werefixed and biotic resources would have been similar throughtime; that will surely help explain the apparently randomchanges detected in some archaeological sequences.

Once methodological disparities are overcome and the se-rious chronometric problems are addressed, interassemblagecomparisons may also play a more important role in modernanalysis of Iberian Middle Paleolithic technological strategiesand in the recognition of temporal trends. The Iberian recordhas the required time depth to explore more than 250 kyr ofMousterian history, and it is unfortunate that substantial partsof this research focus only on the Neanderthal extinctioninterval. Also, the absence of time trends and the attributionof intrasite variability to contingent settlement strategies donot preclude the existence of other variables of differentiationin the Mousterian of Iberia with respect to the adjacent regionof France and within Iberia itself. Traditionally attributed toraw material constraints, the scarcity of preferential flake Le-vallois cores in Iberia is noticeable; they certainly appear insome Middle Paleolithic sites, but they are much less commonthan Levallois recurrent centripetal methods. Predominanceof recurrent centripetal techniques is a feature shared withthe Mousterian of the Aquitaine basin, but while in the Frenchsequence some evolution within Levallois methods is docu-mented (Turq 2000), such variation is not identified in Iberia.Turq (2000) also reports numerous Kombewa flakes both formaking handaxes and cleavers and as core blanks. WhileKombewa is not yet much reported in Iberian assemblages,the role of flakes as cores is increasingly recognized; as detailedby Bourguignon, Faivre, and Turq (2004) across French sites,such “ramification” of chaınes operatoires may explain theremarkably small size of stone tools in some Mousterian as-semblages from both the north (Rıos 2008) and south (CortesSanchez 2007) ends of the Iberian Peninsula. It is unlikely,however, that this so-called ramification explains all instancesof small-sized assemblages, for in some cases, such as Cuestade la Bajada (Santonja et al 2000) and Bolomor (FernandezPeris 2007), raw material constraints may be responsible,while in others (e.g., Roca dels Bous) the small size of artefactsis due to intentional extreme exhaustion of nodular cores(Mora, de la Torre, and Martınez-Moreno 2004).

Regional variability of tool types should also be investi-gated. In general, the diversity of retouched tools seems tobe lower in Iberia than in southwestern France. Levallois andMousterian points, backed knives, and others are present ina number of Iberian Mousterian sites but show even lowerfrequencies than in southwestern France. Sidescrapers dom-inate most Iberian sites apart from Catalunya and some Can-tabrian sites, and such sidescraper-rich assemblages are oftenconsidered as “typical” Mousterian. Cantabria and Catalunyaseem to present some idiosyncrasies. The presence of cleaversis well documented in some Cantabrian coast (El Castillo,Morın, Lezetxiki, Pendo, Amalda, and Gatzarria) and French

Pyrenees (Isturitz, Abri Olha, Calavante, and Noisetier)Mousterian assemblages. Although subsequent studies (e.g.,Cabrera, Pike-Tay, and Bernaldo de Quiros 2004; Freeman1966) excluded Bordes’s original proposal of the Vasconianas a distinct facies, new hypotheses (e.g., Thiebault et al. 2012)somehow rescue the original idea and propose a specific tech-nological entity for the group of Mousterian sites with cleav-ers. Also, denticulates predominate in some Cantabrian andmany Catalonian assemblages, which is not a common patternelsewhere in Iberia. This prevalence of denticulate assem-blages, especially in Catalonian sites, has been known for along time (Mora 1988; Ripoll and de Lumley 1965) and hasalso recently been proposed as sharing traits with French Py-renean assemblages (Thiebault et al. 2012). Albeit further cor-relations and interassemblage comparative work is required,these examples seem to suggest that regional variability inIberia could exist.

For the moment, however, it is difficult to compare datadirectly on temporal and regional variation from Iberia withneighboring regions such as France. In the latter, the highdensity of sites, development of a reliable radiometric frame-work, and consistency of technological study and publicationof site reports has enabled researchers to construct a solidchronostratigraphic sequence where time trends in technologycan be evaluated more precisely. Technological reviews beganin the wake of the 1960s–1980s typological debate on themeaning of Middle Paleolithic variability, and even in recentyears (e.g., Delagnes, Jaubert, and Meignen 2007) attemptshave been made to adequate the Mousterian facies to thecurrently prevalent technological reading of assemblages.Thus, a meaningful association is proposed to exist betweenthe typological facies of La Ferrassie Mousterian and the Le-vallois method, the Denticulate Mousterian and discoid flak-ing, the Quina Mousterian facies and Quina reduction, andthe typical Mousterian of Levallois facies and Levallois re-current flaking (Delagnes, Jaubert, and Meignen 2007).

Since the beginning of the chaıne operatoire approach inthe 1980s, certain diachronic patterns in technology were dis-cerned, particularly with regard to the evolution of the Le-vallois technique (Geneste 1990). In the Aquitaine basin, Turq(2000) reports a change from unidirectional and bidirectionalrecurrent Levallois in pre-OIS 5 assemblages to centripetalrecurrent Levallois in Last Glacial sites. He also states thatfrom OIS 5, centripetal methods dominated alongside Kom-bewa and Quina flaking. According to Delagnes and Meignen(2006), preferential Levallois is consistently older and com-moner in the north than in southern France, and centripetalrecurrent Levallois became dominant after OIS 5. The RhoneValley (Moncel and Daujeard 2012) offers a different picture,however, in which Levallois is mostly centripetal during OIS8 while unidirectional and bidirectional Levallois cores pre-dominate in OIS 4 and 3. Actually, the OIS 8–OIS 3 sequenceat the Rhone Valley (Moncel and Daujeard 2012:113) some-what resembles the random variation of flaking systems andtool-type frequencies discussed above for Iberia, showing sto-

de la Torre, Martınez-Moreno, and Mora Change and Stasis in the Iberian Middle Paleolithic S333

chastic predominance of Levallois and discoid methods andnondirectional changes in the percentage of scrapers andpoints.

Other technological indicators may potentially bear clearerchronological meaning. In northern France, blade productionis limited to the early OIS 5 (Delagnes and Meignen 2006)and does not appear in the Rhone Valley before the last in-terglacial, although in southeastern France, blade assemblagescontinues after the early OIS 5 (Moncel and Daujeard 2012).Blade and/or bladelet production seem also to be character-istic of some late Mousterian assemblages; elongated blanksin the southwestern France Mousterian of Acheulean tradition(MTA) could be a precursor for the Chatelperronian bladetechnology (Pelegrin and Soressi 2007), while in the southeastthe Charentian would evolve into blade and bladelet industrieswith a marked regional character (Slimak 2008).

In summary, evidence seems to suggest that some timetrends exist in the Middle Paleolithic technology of France.Recent overviews of the Aquitaine record report a successionof the predominance of flaking systems, with prevalence ofLevallois and laminar reduction in the OIS 7–OIS 5 sitesfollowed by Quina, MTA, and discoid/denticulate assemblagesfrom OIS 4 until the transitional industries to the UpperPaleolithic (Delagnes and Rendu 2011). Changes within theMTA are also given a chronological significance, with MTAtype A assemblages being consistently older than the MTAtype B (Soressi 2004), and !50 kyr sites from southeast Franceare reported to undergo a process of microlithization andtypological specialization (Slimak 2008).

Although the transitional industries to the Upper Paleo-lithic are beyond the scope of this paper, it is worth notingthat most indicators of Mousterian variability in France areclustered in the later Middle Paleolithic. Delagnes and Meig-nen (2006) report higher diversification of flaking systems inlater stages of the Mousterian in which production of blankswith a low degree of predetermination predominates. Idio-syncratic industries such as the MTA seem to be mostly con-strained to post-50-kyr sequences (Soressi 2004), and a similartime trend applies to the appearance of new tool types andblade/bladelet production in the Rhone Valley (Slimak 2008).Therefore, there seems to be some kind of “shift of gear” thataccelerates technological change in the !50-kyr MousterianFrench sequence, resulting in a more conspicuous interas-semblage variability. Such variability certainly bears chrono-logical connotations, potentially regional differentiation, andputatively also cultural evolutionary implications, given thetemporal proximity to the early Upper Paleolithic.

Causes of these general temporal trends have also beeninvestigated. For example, recent studies have aimed to es-tablish links between particular knapping systems and sub-sistence strategies, providing an explanatory cause for chro-nological patterns. But so far results are contradictory;assuming an association between flaking methods and mo-bility patterns, Delagnes and Rendu (2011) relate predomi-nance of Levallois and blade production with low transport-

ability of blanks and low mobility patterns, link the MTA withhigh mobility Neanderthal groups, and see discoid/denticu-late-dominated assemblages as the result of multipurpose andhighly versatile tool kits. Scott and Ashton (2011), however,propose for early Levallois assemblages exactly the oppositeand consider that Levallois represents increased transport ofblanks and extended curation. Likewise, whereas some pro-pose that MTA bifaces were multipurpose tools (e.g., Soressi2004), others see these as single-tasked artefacts related tobutchery made by highly mobile Neanderthals (Delagnes andRendu 2011).

Beyond interpretive problems of interassemblage variabil-ity, it remains clear that the French record contains a solidand reliable chronostratigraphic record and that such a recordcould indicate some kind of time patterning in Mousteriantechnology (Delagnes, Jaubert, and Meignen 2007; Delagnesand Meignen 2006; Delagnes and Rendu 2011). For the mo-ment, such trends are particularly conspicuous in the laterpart of the Mousterian and could be related to a process ofregionalization, particularly in the south of France (Slimak2008). Regionalization phenomena are seen in adjacentregions of Europe at the end of the Middle Paleolithic notonly on typological grounds but also according to techno-logical indicators, such as in Italy (Kuhn 2006). However, thistemporal trend has yet to be discerned in the Iberian Pen-insula, where a number of challenges discussed in this paperconfound any potential patterns. When compared with theneighboring area of France, the Iberian Middle Paleolithicrecord shows more technotypological homogeneity, both re-gionally and diachronically. In order to ascertain whether thislack of variability in the Iberian Mousterian is real or anartefact of empirical and methodological problems, furtherefforts are required to develop a chronostratigraphic and tech-nological data set framework comparable with that availablein other parts of western Europe and the Near East.

Conclusions

The debate on change versus stasis is inseparable from theassessment of Mousterian technology. Two questions can beasked here (Kuhn 2006). First, did major technological in-novations occur during the Middle Paleolithic? Many wouldagree with Kuhn (2006) that the whole package of techno-logical features documented in the latest Mousterian is alreadypresent during the early Middle Paleolithic. Second, and adifferent issue, does Mousterian variability shows directionaltrends? With some exceptions (e.g., Mellars 1969), the ty-pological approach failed to detect such temporal patterns inthe Middle Paleolithic repertoire.

In recent years, technological perspectives have also ex-plored the existence of diachronic trends but from the viewof the knapping systems. For example, Scott and Ashton(2011) argue that most of early Middle Paleolithic sites inNorthern Europe show predominance of preferential Levalloismethods that are not so conspicuous in later assemblages. In

S334 Current Anthropology Volume 54, Supplement 8, December 2013

the same vein, laminar technologies are reported to be morecommon in earlier Middle Paleolithic assemblages (Bar-Yosefand Kuhn 1999). In general, the French sequence seems topresent at the end of the Mousterian greater diversity of flak-ing methods coexisting at the same time and larger frequencyof lower predetermination systems (Delagnes and Meignen2006).

Can we apply such observations to the Iberian record? Thispaper has argued that according to the evidence currentlyavailable, time trends either do not exist or are not discerniblebecause of a lack of comparable data. We acknowledge thatthe La Noguera Middle Paleolithic does not provide enoughtime depth and sufficient sites to build up a solid sample. Itshould also be remembered that chronostratigraphic andmethodological problems are too acute to allow drawing firmconclusions on interassemblage correlations for the whole ofIberia. Therefore, this paper should be seen as an attempt tospeculate about the diachronic potential of the Iberian MiddlePaleolithic rather than a categorical exclusion of the existenceof the time patterns documented elsewhere in western Europe.As stated in an earlier section, it would be unsurprising ifsuch patterns are detected once the Iberian record is organizedin a more reliable chronological framework and technologicalstudies become readily available.

For now, however, directionality of assemblage variabilityis not visible in the Iberian Middle Paleolithic. Quoting Kuhn(2006:110), “While the Mousterian may have changed overtime it was not going anywhere in particular,” Iberia included.This does not mean that it is unworthy to keep trying todisentangle the long historical trajectory of the Mousterian.As pointed out above, most of current efforts in Mousterianresearch in Spain and Portugal are privileging only the tran-sition to the Upper Paleolithic. This is definitely not uniqueto Iberia, and it has been argued that also elsewhere “what isalmost never addressed is what was going on earlier in theMiddle Paleolithic, before modern humans and the UpperPaleolithic came on the scene” (Kuhn and Hovers 2006:3).Only further research can help to address this problem.

Acknowledgments

We thank Erella Hovers, Steven Kuhn, and Leslie Aiello forthe invitation to participate in the “Alternative Pathways toComplexity” Wenner-Gren symposium and to contribute tothis special issue of Current Anthropology. Comments by theguest editors and two anonymous reviewers on an earlier draftof this paper are gratefully acknowledged. Excavations in thePre-Pyrenees Middle Paleolithic are funded by the SpanishMinisterio de Ciencia e Innovacion (HAR2010-15002). R.Mora is grateful for the support of the Institucio Catalana deRecerca i Estudis Avancats.

References CitedAngelucci, Diego E., and Joao Zilhao. 2009. Stratigraphy and formation pro-

cesses of the Upper Pleistocene deposit at Gruta da Oliveira, AlmondaKarstic System, Torres Novas, Portugal. Geoarchaeology 24:277–310.

Baena, Javier, E. Carrion, B. Ruiz, B. Ellwood, C. Sese, J. Yravedra, J. Jorda,et al. 2005. Paleoecologıa y comportamiento humano durante el PleistocenoSuperior en la comarca de Liebana: la secuencia de la Cueva del Esquilleu,Occidente de Cantabria, Espana. In Neandertales cantabricos: estado de lacuestion. Jose A. Lasheras and Ramon Montes Barquın, eds. Pp. 461–487.Monografıas Museo de Altamira 20. Madrid: Ministerio de Cultura, Sub-direccion General de Publicaciones, Informacion y Documentacion.

Baldeon, Amelia. 1993. El yacimiento de Lezetxiki (Gipuzkoa, Paıs Vasco):los niveles musterienses. Munibe 45:3–97.

———. 1999. El abrigo de Axlor (Bizkaia, Paıs Vasco): las industrias lıticasde sus niveles Musterienses. Munibe 51:9–121.

Bar-Yosef, Ofer, and Steven L. Kuhn. 1999. The big deal about blades: laminartechnologies and human evolution. American Anthropologist 10:322–338.

Bernaldo de Quiros, Federico, Granada Sanchez-Fernandez, and Jose M.Maıllo. 2010. Technological characteristics at the end of the Mousterian inCantabria: the El Cıastillo and Cueva Morın (Spain). In The Upper Pa-laeolithic revolution in global perspective: papers in honour of Sir Paul Mellars.Katherine V. Boyle, Clive Gamble, and Ofer Bar-Yosef, eds. Pp. 153–160.Cambridge: McDonald Institute.

Beyries, Sylvia. 1988. Functional variability of lithics sets in the Middle Pa-leolithic. In Upper Pleistocene prehistory of western Eurasia. Harold L. Dibbleand Anta Montet-White, eds. Pp. 213–223. Philadelphia: University Mu-seum, University of Philadelphia.

Binford, Lewis R. 1973. Interassemblage variability: the Mousterian and thefunctional argument. In The explanation of culture change: models in pre-history. Colin Renfrew, ed. Pp. 227–254. London: Duckworth.

Bischoff, James L., Ramon Julia, and Rafael Mora. 1988. Uranium series datingof the Mousterian occupation at Abric Romanı, Spain. Nature 332:68–70.

Boeda, Eric. 1986. Approche technologique du concept Levallois et evaluationde son champ d’application. PhD dissertation, Universite Paris X Nanterre.

Bordes, Francois. 1972. A tale of two caves. New York: Harper & Row.Bourguignon, Laurence. 1997. Le Mousterien de type Quina: nouvelle defin-

ition d’une entite technique. PhD dissertation, Universite Paris X Nanterre.Bourguignon, Laurence, Jean-Philippe Faivre, and A. Turq. 2004. Ramification

des chaınes operatoires: une specificite du Mousterien? Paleo 16:37–48.Butzer, Karl W. 1981. Cave Sediments, Upper Pleistocene stratigraphy and

Mousterian facies in Cantabrian Spain. Journal of Archaeological Science 8:133–183.

Cabrera, Victoria. 1984. El yacimiento de la cueva de “El Castillo” (PuenteViesgo, Santander). Bibliotheca Praehistorica Hispana, vol. 22. Madrid:Consejo Superior de Investigaciones Cientıficas, Instituto Espanol de Pre-historia.

Cabrera, Victoria, A. Pike-Tay, and F. Bernaldo de Quiros. 2004. Trends inMiddle Paleolithic settlement in Cantabrian Spain: the Late Mousterian atCastillo Cave. In Settlement dynamics of the Middle Paleolithic and MiddleStone Age, vol. 2. Nicholas J. Conard, ed. Pp. 437–460. Tubingen: Kerns.

Carrion, Elena, J. Baena, C. Conde, F. Cuartero, and M. Roca. 2008. Varia-bilidad tecnologica en el musteriense de Cantabria. In Variabilidad tecnicadel Paleolıtico Medio en el sudoeste de Europa, vol. 14. Rafael Mora, JorgeMartınez-Moreno, Ignacio de la Torre, and Joel Casanova, eds. Pp. 279–318. Barcelona: Treballs d’Arqueologia.

Casanova, Joel. 2009. Estrategias tecnologicas de los neandertales en la ver-tiente sur del Pirineo oriental. PhD dissertation, Universidad Autonomade Barcelona.

Casanova, Joel, R. Mora, J. Martınez-Moreno, and I. de la Torre. 2009. Stra-tegies techniques dans le Paleolithique Moyen du sud-est des Pyrenees.L’Anthropologie 113:313–340.

Cortes Sanchez, Miguel, ed. 2007. Cueva Bajondillo (Torremolinos): secuenciacronocultural y paleoambiental del Cuaternario reciente en la Bahıa de Malaga.Malaga: Diputacion de Malaga.

———. 2008. Variabilidad tecnologica en el Paleolıtico Medio meridionaliberico: un punto de partida. In Variabilidad tecnica del Paleolıtico Medioen el sudoeste de Europa, vol. 14. Rafael Mora, Jorge Martınez-Moreno,Ignacio de la Torre, and Joel Casanova, eds. Pp. 105–119. Barcelona: Treballsd’Arqueologia.

Delagnes, Anne, Jacques Jaubert, and Liliane Meignen. 2007. Les technocom-plexes du Paleolithique moyen en Europe occidentale dans leur cadre diach-ronique et geographique. In Les neandertaliens, biologie et cultures. Bernard

de la Torre, Martınez-Moreno, and Mora Change and Stasis in the Iberian Middle Paleolithic S335

Vandermeersch and Bruno Maureille, eds. Pp. 213–229. Paris: Editions duComite des Travaux Historiques et Scientifiques.

Delagnes, Anne, and Liliane Meignen. 2006. Diversity of lithic productionsystems during the Middle Paleolithic in France: are there any chronologicaltrends? In Transitions before the transition: evolution and stability in theMiddle Paleolithic and Middle Stone Age. Erella Hovers and Steven. L. Kuhn,eds. Pp. 85–107. New York: Springer.

Delagnes, Anne, and Will Rendu. 2011. Shifts in Neandertal mobility, tech-nology and subsistence strategies in western France. Journal of ArchaeologicalScience 38:1771–1783.

de la Torre, Ignacio. 2009. Technological Strategies in the Lower Pleistoceneat Peninj (west of Lake Natron, Tanzania). In The cutting edge: new ap-proaches to the archaeology of human origins. Kathy Schick and NicholasToth, eds. Pp. 93–113. Bloomington, IN: Stone Age Institute.

de la Torre, Ignacio, and Rafael Mora. 2004. El Olduvayense de la seccion tipode Peninj (Lago Natron, Tanzania). Monografies del CEPAP 1. Barcelona:Centre d’estudis del patrimoni arqueologic de la prehistoria.

d’Errico, Francesco, and Maria F. Sanchez-Goni. 2003. Neandertal extinctionand the millennial scale climatic variability of OIS 3. Quaternary ScienceReviews 22:769–788.

Dibble, Harold L. 1987. The interpretation of Middle Paleolithic scraper mor-phology. American Antiquity 52:109–117.

Dibble, Harold L., and Nicholas Rolland. 1992. On assemblage variability inthe Middle Paleolithic of western Europe: history, perspectives, and a newsynthesis. In The Middle Paleolithic: adaptation, behaviour and variability.Harold L. Dibble and Paul Mellars, eds. Pp. 1–28. Philadelphia: UniversityMuseum, University of Philadelphia.

Falgueres, Christophe, Y. Yokoyama, and A. Arrizabalaga. 2005. La geocron-ologıa del yacimiento pleistocenico de Lezetxiki (Arrasate, Paıs Vasco): crı-tica de las dataciones existentes y algunas nuevas aportaciones. Munibe 57:93–106.

Fernandez Peris, Josep. 2007. La Cova del Bolomor (Tavernes de la Valldigna,Valencia): las industrias del Pleistoceno medio en el ambito del Mediterraneopeninsular. Serie de Trabajos Varios, no. 108. Valencia: Servicio de Inves-tigacion Prehistorica, Diputacion Provincial de Valencia.

Fernandez Peris, Josep, V. Barciela, R. Blasco, F. Cuartero, and P. Sanudo.2008. El Paleolıtico Medio en el territorio valenciano y la variabilidad tecno-economica de la Cova del Bolomor. In Variabilidad tecnica del PaleolıticoMedio en el sudoeste de Europa, vol. 14. Rafael Mora, Jorge Martınez-Moreno, Ignacio de la Torre, and J. Casanova, eds. Pp. 141–169. Barcelona:Treballs d’Arqueologia.

Finlayson, Clive, F. Giles, J. Rodrıguez-Vidal, D. A. Fa, J. M. Gutierrez, A. S.Perez, G. Finlayson, et al. 2006. Late survival of Neanderthals at the south-ernmost extreme of Europe. Nature 443:850–853.

Freeman, Leslie G. 1966. The nature of Mousterian facies in Cantabrian Spain.American Anthropologist 68:230–237.

Galvan, Bertila, C. M. Hernandez, M. I. Francisco, and A. Rodrıguez. 2006.Datos para la caracterizacion del final del musteriense en los valles de Alcoi.In En el centenario de la cueva de El Castillo: el ocaso de los Neandertales.Victoria Cabrera, Federico Bernaldo de Quiros, and Jose M. Maıllo, eds.Pp. 127–141. Madrid: Centro Asociado a la Universidad Nacional de Ed-ucacion a Distancia en Cantabria.

Gamble, C. 1986. The Palaeolithic settlement of Europe. Cambridge: CambridgeUniversity Press.

Geneste, Jean-Michel. 1985. Analyse lithique d’industries mousteriennes duPerigord: une approche technologique du comportement des groupes hu-mains au Paleolitique Moyen. Bordeaux: PhD dissertation, Universite deBordeaux I.

———. 1990. Developpement des sytemes de production lithique au coursdu Paleolithique moyen en Aquitaine septentrionale. In Paleolithique moyenet recent et Paleolithique superieur ancien en Europe: ruptures et transitions:examen critique des documents archeologiques: actes du colloque internationalde Nemours, 9–10 mai 1988. Catherine Farizy, ed. Pp. 203–213. Nemours:Memoires du Musee de Prehistoire d’Ile-de-France.

Gonzalez Urquijo, Jesus E., J. J. Ibanez, J. Rıos, and L. Bourguignon. 2006.Aportes de las nuevas excavaciones en Axlor sobre el final del PaleolıticoMedio. In En el centenario de la cueva de El Castillo: el ocaso de los Nean-dertales. Victoria Cabrera, Federico Bernaldo de Quiros, and Jose M. Maıllo,eds. Pp. 269–289. Madrid: Centro Asociado a la Universidad Nacional deEducacion a Distancia en Cantabria.

Goren-Inbar, Naama, and Anna Belfer-Cohen. 1998. The technological abil-ities of the Levantine Mousterians: cultural and mental capacities. In Nean-

dertals and modern humans in western Asia. Takeru Akazawa, Kenichi Aoki,and Ofer Bar-Yosef, eds. Pp. 205–221. New York: Plenum.

Hovers, Erella, and Anna Belfer-Cohen. 2013. On variability and complexity:lessons from the Levantine Middle Paleolithic record. Current Anthropology54(suppl. 8):S337–S357.

Jaubert, Jacques. 2011. Archeosequences du Paleolithique moyen du sud-ouestde la France: quel bilan un quart de siecle apres Francois Bordes? In FrancoisBordes et la prehistoire. Francois Delpech and Jacques Jaubert, eds. Pp. 235–253. Paris: Editions du Comite des Travaux Historiques et Scientifiques.

Kuhn, Steven L. 1995. Mousterian lithic technology: an ecological perspective.Princeton, NJ: Princeton University Press.

———. 2006. Trajectories of change in the Middle Paleolithic of Italy. InTransitions before the transition: evolution and stability in the Middle Pale-olithic and Middle Stone Age. Erella Hovers and Steven L. Kuhn, eds. Pp.109–120. New York: Springer.

Kuhn, Steven L., and Erella Hovers. 2006. General introduction. In Transitionsbefore the transition: evolution and stability in the Middle Paleolithic andMiddle Stone Age. Erella Hovers and Steven L. Kuhn, eds. Pp. 1–11. NewYork: Springer.

Laville, Henri. 1973. The relative position of Mousterian industries in theclimatic chronology of the Early Wurm in the Perigord. World Archaeology4:323–329.

Manzano, Ivan, J. Baena, A. Lazaro, D. Martın, L. Dapena, M. Roca, and E.Moreno. 2005. Analisis de los recursos lıticos en la Cueva del Esquilleu:gestion y comportamiento durante el Musteriense (Comarca de la Liebana,Occidente de Cantabria). In Neandertales cantabricos: estado de la cuestion.Jose A. Lasheras and Ramon Montes, eds. Pp. 285–300. Monografıas Na-cional y Centro de Investigacion de Altamira 20. Madrid: Ministerio deCultura, Subdireccion General de Publicaciones, Informacion y Documen-tacion.

Maroto, Julia, M. Vaquero, A. Arrizabalaga, J. Baena, E. Baquedano, J. Jorda,R. Julia, et al. 2012. Current issues in Late Middle Palaeolithic chronology:new assessments from northern Iberia. Quaternary International 247:15–25.

Martınez-Moreno, Jorge, Ignacio de la Torre, Rafael Mora, and Joel Casanova.2010. Technical variability and changes in the pattern of settlement at Rocadels Bous (southeastern Pre-Pyrenees, Spain). In Settlement dynamics of theMiddle Paleolithic and Middle Stone Age, vol. 3. Nicholas J. Conard andAnne Delagnes, eds. Pp. 485–507. Tubingen: Kerns.

Martınez-Moreno, Jorge, Rafael Mora, and Ignacio de la Torre. 2004. Meth-odological approach for understanding Middle Palaeolithic settlement dy-namics at La Roca dels Bous (Noguera, Catalunya, Northeast Spain). InSettlement dynamics of the Middle Paleolithic and Middle Stone Age, vol. 2.Nicholas J. Conard, ed. Pp. 393–413. Tubingen: Kerns.

———. 2010. The Middle-to-Upper Palaeolithic transition in Cova Gran(Catalonia, Spain) and the extinction of Neanderthals in the Iberian Pen-insula. Journal of Human Evolution 58:211–226.

Mellars, Paul. 1969. The chronology of Mousterian industries in the Perigordregion of south-west France. Proceedings of the Prehistoric Society 35:134–171.

Moncel, Marie-Helene, and Camille Daujeard. 2012. The variability of theMiddle Palaeolithic on the right bank of the Middle Rhone Valley (southeastFrance): technical traditions or functional choices? Quaternary International247:103–124.

Moncel, Marie-Helene, Anne-Marie Moigne, Youssef Sam, and Jean Combier.2011. The emergence of Neanderthal technical behavior: new evidence fromOrgnac 3 (level 1, MIS 8), southeastern France. Current Anthropology 52:37–75.

Montes, Ramon, and Juan Sanguino. 2005. Nuevos datos para el conocimientodel Paleolıtico Medio en el centro de la Region Cantabrica: la Cueva deCovalejos. In Neandertales cantabricos: estado de la cuestion. J. A. Lasherasand R. Montes, eds. Pp. 489–538. Monografıas Nacional y Centro de In-vestigacion de Altamira 20. Madrid: Ministerio de Cultura, SubdireccionGeneral de Publicaciones, Informacion y Documentacion.

Mora, Rafael. 1988. El Paleolıtico Medio en Catalunya. PhD dissertation,Universidad de Barcelona.

Mora, Rafael, A. Benito-Calvo, Jorge Martınez-Moreno, P. Gonzalez, and I.de la Torre. 2011. Chrono-stratigraphy of the Upper Pleistocene and Ho-locene archaeological sequence in Cova Gran (south-eastern Pre-Pyrenees,Iberian Peninsula). Journal of Quaternary Science 26:635–644.

Mora, Rafael, Ignacio de la Torre, and Jorge Martınez-Moreno. 2004. MiddlePalaeolithic mobility and land use in the southwestern Pyrenees: the ex-ample of level 10 in La Roca dels Bous (Noguera, Catalunya, northeast

S336 Current Anthropology Volume 54, Supplement 8, December 2013

Spain). In Settlement dynamics of the Middle Paleolithic and Middle StoneAge, vol. 2. Nicholas J. Conard, ed. Pp. 415–435. Tubingen: Kerns.

Olle, A., M. Mosquera, X. P. Rodrıguez, A. de Lombera-Hermida, M. D.Garcıa-Anton, P. Garcıa-Medrano, L. Pena, et al. 2013. The early and MiddlePleistocene technological record from Sierra de Atapuerca (Burgos, Spain).Quaternary International 295:138–167.

Pelegrin, Jacques, and Marie Soressi. 2007. Le Chatelperronien et ses rapportsavec le Mousterien. In Les Neandertaliens, biologie et cultures. Bernard Van-dermeersch and Bruno Maureille, eds. Pp. 283–296. Documents prehis-toriques 23. Paris: Editions du Comite des Travaux Historiques et Scien-tifiques.

Rink, W. J., H. P. Schwarcz, H. K. Lee, V. Cabrera Valdes, F. Bernaldo deQuiros, and M. Hoyos. 1997. ESR dating of Mousterian levels at El CastilloCave, Cantabria, Spain. Journal of Archaeological Science 24:593–600.

Rıos, Joseba. 2008. Variabilidad tecnologica en el Paleolıtico Medio de losPirineos Occidentales: una expresion de las dinamicas historicas de lassociedades neandertales. In Variabilidad tecnica del Paleolıtico Medio en elsudoeste de Europa, vol. 14. Rafael Mora, Jorge Martınez-Moreno, Ignaciode la Torre, and Joel Casanova, eds. Pp. 171–194. Barcelona: Treballsd’Arqueologia.

Ripoll, Eduardo, and Henry de Lumley. 1965. El Paleolıtico medio en Cataluna.Ampurias 26/27:1–67.

Rodrıguez, Xose P. 2004. Atapuerca y el inicio del Paleolıtico medio en Europa.In Arqueologıa, vol. 4 of Miscelanea en homenaje a Emiliano Aguirre. Pp.417–431. Zona Arqueologıa, no. 4. Madrid: Museo Arqueologico Regional.

Santamarıa, David, J. Fortea, M. d. l. Rasilla, L. Martınez, E. Martınez, J. C.Canaveras, S. Sanchez-Moral, et al. 2010. The technological and typologicalbehaviour of a Neanderthal group from El Sidron Cave (Asturias, Spain).Oxford Journal of Archaeology 29:119–148.

Santamarıa, David, L. Montes, and P. Utrilla. 2008. Variabilidad tecnica delPaleolıtico Medio en el valle del Ebro: la Cueva de los Moros I de Gabasa(Peralta de Calasanz, Huesca). In Variabilidad tecnica del Paleolıtico Medioen el sudoeste de Europa, vol. 14. Rafael Mora, Jorge Martınez-Moreno,Ignacio de la Torre, and Joel Casanova, eds. Pp. 319–339. Barcelona: Treballsd’Arqueologia.

Santonja, Manuel, A. Perez-Gonzalez, P. Villa, C. Sese, E. Soto, R. Mora, V.Eisenmann, and M. Dupre. 2000. El yacimiento paleolıtico de Cuesta dela Bajada (Teruel). In 3º Congresso de Arqueologia Peninsular: Paleolıtico dapenınsula iberica, vol. 2. Rodrigo de Balbın, N. Bicho, E. Carbonell, B.Hockett, A. Moure, L. Raposo, M. Santonja, and G. Vega, eds. Pp. 169–186. Porto: ADECAP.

Santonja, Manuel, and Paola Villa. 2006. The Acheulean of western Europe.In Axe age: Acheulian toolmaking from quarry to discard. Naama Goren-Inbar and Gonen Sharon, eds. Pp. 429–478. London: Equinox.

Scott, Becky, and Nick Ashton. 2011. The Early Middle Palaeolithic: the Eu-

ropean context. In The ancient human occupation of Britain. Nick Ashton,S. G. Lewis, and C. B. Stringer, eds. Pp. 91–112. Amsterdam: Elsevier.

Slimak, Ludovic. 2008. Faire des outils entre le 50eme et le 35eme millenaireen France mediterraneene: systemes techniques et evolutions anthropolo-giques. In Variabilidad tecnica del Paleolıtico Medio en el sudoeste de Europa,vol. 14. Rafael Mora, Jorge Martınez-Moreno, Ignacio de la Torre, and JoelCasanova, eds. Pp. 65–86. Barcelona: Treballs d’Arqueologia.

Soressi, Marie. 2004. From the Mousterian of Acheulian tradition type A totype B: a change in technical tradition, raw material, task, or settlementdynamic? In Settlement dynamics of the Middle Paleolithic and Middle StoneAge, vol. 2. Nicholas J. Conard, ed. Pp. 343–366. Tubingen: Kerns.

Thiebaut, Celine, V. Mourre, P. Chalard, D. Colonge, A. Coudenneau, M.Deschamps, and A. Sacco-Sonador. 2012. Lithic technology of the finalMousterian on both sides of the Pyrenees. Quaternary International 247:182–198.

Turq, Alain. 2000. Le Mousterien de tradition Acheuleenne. In Paleolithiqueinferieur et moyen entre Dordogne et Lot. Special issue, Paleo (suppl. 2):244–273.

Vaquero, Manuel. 1999. Intrasite spatial organization of lithic production inthe Middle Palaeolithic: the evidence of the Abric Romanı (Capellades,Spain). Antiquity 73:493–504.

Vaquero, Manuel, and Eudald Carbonell. 2003. A temporal perspective onthe variability of the discoid method in the Iberian Peninsula. In Discoidlithic technology: advances and implications. Marco Peresani, ed. Pp. 67–81.BAR International Series 1120. Oxford: Archaeopress.

Vaquero, Manuel, M. G. Chacon, M. D. Garcıa-Anton, B. Gomez de Soler,K. Martınez, and F. Cuartero. 2012. Time and space in the formation oflithic assemblages: the example of Abric Romanı level J. Quaternary Inter-national 247:162–181.

Vega, Luis G., P. Cosano, A. Villar, O. Escarpa, and T. Rojas. 1997. Las in-dustrias de la interfase Pleistoceno Medio-Superior en la cueva de la Ca-rihuela (Pinar, Granada). In II Congreso de Arqueologıa Peninsular, vol. 1.Rodrigo de Balbın and Primitiva Bueno, eds. Pp. 105–118. Zamora, Spain:Fundacion Rei Afonso Henriques.

Villaverde, Valentin. 1984. La Cova Negra de Xativa y el Musteriense de laregion central del Mediterraneo espanol. Valencia: Diputacion Provincial deValencia.

Zilhao, Joao. 2001. Middle Paleolithic settlement patterns in Portugal. InSettlement dynamics of the Middle Paleolithic and Middle Stone Age, vol. 1.Nicholas J. Conard, ed. Pp. 597–608. Tubingen: Kerns.

———. 2006. Chronostratigraphy of the Middle-to-Upper Paleolithic Tran-sition in the Iberian Peninsula. Pyrenae 37:7–84.

———. 2008. The Ebro frontier revisited. In The Mediterranean from 50,000to 25,000 BP: turning points and new directions. Marta Camps and CarolynSzmidt, eds. Pp. 293–311. Oxford: Oxbow.

Current Anthropology Volume 54, Supplement 8, December 2013 S337

� 2013 by The Wenner-Gren Foundation for Anthropological Research. All rights reserved. 0011-3204/2013/54S8-0013$10.00. DOI: 10.1086/673880

On Variability and ComplexityLessons from the Levantine Middle Paleolithic Record

by Erella Hovers and Anna Belfer-Cohen

CA� Online-Only Material: Supplements A, B

A century of research has led to the recognition of multiple levels of technological variability in the Levantine MiddlePaleolithic (MP) that cannot be resolved through single-cause explanatory models. Recent ecological models arguefor continual occupation of the region and competitive coexistence of Neanderthal and modern human populations.Current paleogenetic studies underline the feasibility of the latter scenario. The Levantine MP offers a perspectiveon the interface of historical circumstances and long-term evolutionary mechanisms that structured in-tandemtrajectories of technological and behavioral changes as well as insights into the dynamics of nondirectional behavioralcomplexities in the archaeological record.

Paleoanthropologists dichotomize the global post-AcheulianPaleolithic record of 400,000–50,000 yr ago between the Mid-dle Paleolithic (MP) of Eurasia/northern Africa and the Mid-dle Stone Age (MSA) of sub-Saharan Africa (e.g., Marks2008). While the MP record (on large temporal and geo-graphic scales) has been perceived as relatively monotonousand poor in innovations (Kuhn and Stiner 1998), the MSArecord appears studded with technological and symbolic ad-vances (but see Tryon and Faith 2013). This distinction hasraised questions about underlying biological, cognitive, de-mographic, and social differences between populations in thetwo metaregions and how those affected the trajectories ofcultural evolution in both. Importantly, the record of eachregion averages hominin interactions with local social andphysical environments that had occurred on variable intra-and interregional geographic and temporal scales. Thesehigher-resolution processes should be identified and explainedif broader evolutionary trajectories are to be understood(Hovers 2009:246). Accordingly, we narrow down our dis-cussion to the Levantine record of 250,000–50,000 yr ago,which constitutes a discrete MP entity (Bar-Yosef 2006). Ouraim here is to focus on major traits of the Levantine MPrecord as a behavioral system.

The onset of the MP in the Levant is recognized throughthe disappearance of the bifacial tools that characterized thepreceding Lower Paleolithic (LP) complexes, a shift toward

Erella Hovers and Anna Belfer-Cohen are Professors at the Instituteof Archaeology of the Hebrew University of Jerusalem (Mt. Scopus,Jerusalem 91905, Israel [hovers@mscc.huji.ac.il; belferac@mscc.huji.ac.il]). This paper was submitted 3 VII 13, accepted 30 VIII 13, andelectronically published 20 XII 13.

production of flake-based assemblages (Goren-Inbar 1994),and the proliferation of Levallois flaking systems. At the endof the MP period, the Levallois technology gives way to UpperPaleolithic (UP) blade-oriented lithic reduction systems cou-pled with an assortment of less formal flake-oriented tech-nologies. When examined in detail, the MP record presentsa mosaic pattern of diversification, loss, and reemergence ofcultural traits throughout its duration (Belfer-Cohen andGoren-Inbar 1994; Goren-Inbar and Belfer-Cohen 1998; Hov-ers 2009; Hovers and Belfer-Cohen 2006).

A number of reasons make the Levant, a small (ca. 1,000km north–south by up to 400 km east–west) geographicallycircumscribed region (Hovers 2009; Shea 2003), suitable to ahigh-resolution “dissection” of the MP record (fig. 1).1 Arecurrent problem in environmental reconstructions is thatglobal models, often derived from marine data sources, areincompatible in scale with the size of terrestrial tracts inves-tigated archaeologically and do not take into account theeffects of local physiography (Rohling et al. 2013). The richhistory of research on Levantine prehistoric and climatic rec-ords provides a dense matrix of data and allows for a morerealistic approach to the question of the interaction betweenbehavior and environmental variability (Goring-Morris, Hov-ers, and Belfer-Cohen 2009; see Hovers 2009 for references).Moreover, while Levantine climate corresponded to globalclimate patterns throughout the Quaternary, the amplitudeof changes has been less dramatic than in Europe, under-mining simple models of climate-driven extinctions or bio-geographical movements of hominins and related cultural

1. For site-by-site details and dating, see Ahmad (2009), Bisson et al.(forthcoming), Conard et al. (2010), Hauck (2011), Hovers (2009, apps.1, 5), and Sharon et al. (2010).

S338 Current Anthropology Volume 54, Supplement 8, December 2013

Figure 1. Climatic variation shown against site chronologies. The δ18O records from Peqi’in and Soreq caves serve as proxy for therange and magnitude of climatic changes throughout the MP period between 230,000 and 40,000 yr ago (see text). Dated MP sitesshown here are placed within their chronological ranges and paleoclimatic contexts as inferred by various dating methods. Wherestratigraphic sequences are dated, the arrows show the chronological intervals involved. For the sites of Amud, Kebara, Tabun,Hayonim, and Ein Difla, only the thermoluminescence chronology is shown. Note the millennial scale variations clearly observedin the paleoclimatic record. Paleoclimatic data from Bar-Matthews et al. (2003). See Hovers (2009, app. 6) for detailed chronologicalinformation on radiometrically dated MP sites. Reproduced from Hovers (2009, fig. 8.2) with permission from Oxford UniversityPress.

changes.2 Furthermore, the arrival of exogenous populationsinto a region is a potential source for cultural variation andchange. At least two such events (the respective arrivals ofmodern humans and of Neanderthals) are documented in theLevantine MP, providing an opportunity to investigate howsuch historical events are expressed and how they influencedthe local trajectories of cultural evolution.

In order to understand how a complex system works, adetailed analysis of the parts and their interactions, followedby an integrated view of why the particular arrangement ofparts evolved, is essential (Corning 2002). Accordingly, wediscuss the features that define the Levantine MP, relyingmostly on stone tool characteristics. We then identify potentialenvironmental, demographic, and social sources of the be-havioral variability associated with the organization of thelithic technological features. Finally, we turn to questionsabout cultural evolution in the context of the Middle andUpper Pleistocene in the Levant: can one identify the range

2. Almogi-Labin, Bar-Matthews, and Ayalon (2004), Almogi-Labin etal. (2009), Enzel et al. (2008), Frumkin, Bar-Yosef, and Schwarcz (2011),and Hallin, Schoninger, and Schwarcz (2012) all present detailed researchon modern and ancient environments; see summary and discussions inHovers (2009).

of transmission processes, and are they unique to the Levant

within the MP world?

The possibility of behavioral differences between MP mod-

erns and Neanderthals is a major issue in elucidating the

Levantine MP, yet we opted to first explore the properties of

the technological system(s) and only then examine their pos-

sible correlation with biological distinctions (Hovers 2006;

Lieberman and Bar-Yosef 2005). We premise that the cognitive

abilities of MP hominins (and possibly earlier ones; Goren-

Inbar 2011) were comparable, although they may have been

expressed and executed differently because of differences in

life histories, reproductive success, social organization, or dif-

ferent modes and rates of social transmission (e.g., Belfer-

Cohen and Hovers 2010; Hovers and Belfer-Cohen 2006;

Langbroek 2012; Rossano 2010; Wynn and Coolidge 2012).

A second premise is that the variety of human responses to

shifting and unstable natural and social environments is un-

likely to reflect strictly genetically determined responses (Ehr-

lich and Feldman 2003; Pulliam and Dunford 1980). By MP

times, social choices are to be considered as one of the major

factors in the shaping of a cultural record (Hovers 2009:6–7;

Sterelny 2012).

Hovers and Belfer-Cohen Variability and Complexity S339

The View from the Ground: Characteristics ofthe Levantine MP

Lithic Production Systems

Levantine MP lithic assemblages conform to definitions ofthe Mousterian technocomplex and are often referred to as“Levantine Mousterian.” Those are flake industries charac-terized by (1) absence of “core tools” (a generic name forbifaces, cleavers, spheroids), and (2) a quantitative dominanceof Levallois flaking methods (as defined by Boeda 1993;Boeda, Geneste, and Meignen 1990) used over other formalflaking systems to produce all blank morphotypes (flakes,blades, and points). That said, Levallois flaking systems mayco-occur within any assemblage with variable frequencies ofblanks and cores from laminar (as defined by Meignen 1998)and discoidal production systems. “True” (UP-style) bladecores occur in variable frequencies in all published assem-blages. Cores on flakes represent a discrete production systemencountered variably in all assemblages, yet the size and shaperanges of its products often fall within those resulting fromflaking nodules (Goren-Inbar 1988; Hovers 2007, 2009).Quina flaking and bifacial shaping are absent, although theyare known from the earlier Acheulo-Yabrudian (Garrod andBate, 1937; Gopher et al. 2005; Jelinek 1982:3). Flaking sys-tems are modular, with preferential and recurrent methodsof Levallois flaking combined variably with modes of corepreparation (i.e., centripetal, unidirectional, unidirectionalconvergent, and bidirectional). Most Levallois morphotypeswere produced through more than one combination of flakingmodes and methods (Bar-Yosef and Van Peer 2009; Hovers2009; Inizan, Roche, and Tixier 1992; Van Peer 1992).Changes of flaking modes during the course of reduction areoften documented, attempting to extend the core’s use-lifebefore its discard (Bar-Yosef et al. 1992; Hovers 1998b, 2009;Nishiaki and Copeland 1992). Only in a few assemblages doesa single Levallois flaking variant account for over half of theLevallois blank population (e.g., Bar-Yosef et al. 1992; Goren-Inbar 1990; Hauck 2011; Hovers 2009).

Flint was the nearly exclusive raw material used for allflaking systems. Systematic use of limestone or basalt wasreported from only two open-air sites (Gilead 1980, 1988;Goren-Inbar 1990). Flint is readily available throughoutmost areas of the Levant as part of the geological substrate(Barkai and Gopher 2009; Delage 2007; Druck 2004; Eksh-tain et al. 2012; Henry 1995b). Procurement was mostlyfrom primary sources, which were exploited selectively ac-cording to quality, size, and shape suitability for the formalflaking systems (Levallois or Laminar). Transport distancesof partially prepared cores from sources to sites rangedbetween !5 and 20 km (typically !10 km), whereas finishedproducts were rarely transported over distances of 30–40km (Delage 2007).

The production of elongated morphotypes may havepromoted selection of nodules of specific size or propor-

tions conducive to cost-effective core preparation proce-dures and successful removals (Gordon 1993; Wojtzak2011). Levallois flakes were typically detached from broadcores that enabled exploitation of a large surface, and theywere relatively unconstrained by the proportions of theoriginal raw material. Throughout the MP there was atendency to select larger blanks (usually Levallois blanks)for making side scrapers (Hovers 2009 and referencestherein). There are no indications for using fire as an en-gineering tool for lithic raw material manipulation (unlikein the southern African MSA; Brown et al. 2009; Mourre,Villa, and Henshilwood 2010).

Extractive Technologies

Evidence as to the dietary resources used by MP hominins isrestricted in most cases to animal remains because of theirbetter preservation. Direct evidence for dietary plant exploi-tation is rare (Henry, Brooks, and Piperno 2011; Henry et al.2004; Lev, Kislev, and Bar-Yosef 2005; Madella et al. 2002;Matsutani 1973, 1987), yet it demonstrates that Levantine MPpopulations engaged in the collection and processing (thelatter rarely attested to, but see Lev, Kislev, and Bar-Yosef2005) of plant foods, including cereal seeds and lentils. Thepreservation bias has led to an unbalanced focus on tech-nologies of meat acquisition (Hovers 1998a), although on thebasis of nutritional reasoning (Speth 2010 and referencestherein), plants would constitute an important part of thediet of prehistoric hunter-gatherers in most regions, pendingavailability. Even the high metabolic costs of European Ne-anderthals (Aiello 2003; Churchill 2006; MacDonald, Roe-broeks, and Verpoort 2009; Steegman, Cherny, and Holliday2002) may have been met by exploitation of a diverse resourcebase rather than heavy reliance on large game (Hockett 2012).Given the mid-low latitude of the Levant, prehistoric hunter-gatherers would rely extensively on vegetal resources (Binford2001; Cordain et al. 2000; el Zaatari et al. 2011; Hayden, 1981;Kelly 1995).

A similar preservation bias against tools made of organicmaterials renders stone tools the main source for understand-ing MP extractive technologies. It is assumed that stone ar-tifacts were used on animal and plant materials, yet directevidence to that effect is uncommon. Analyses of microscopicedge damage and residues of adhesive and mastics on stonetools provide clues as to the construction of extractive tools.Use-wear studies indicate that both unretouched and re-touched blanks were used as multipurpose implements forcutting and scraping of various animal tissues and plant ma-terials (Beyries 1988; Bonilauri et al. 2007; Shea 1991). Thisis similar to the known technological practices among extanthunter-gatherers (Oswalt 1976). Modification signs on thebase of MP triangular points are consistent with experimentalhafting damage (Shea 1988, 1998), although evidence for thin-ning or notching of the proximal ends in preparation for

S340 Current Anthropology Volume 54, Supplement 8, December 2013

hafting is scant. Pointed blanks bearing such marks, withputative impact fractures at the tip and edge damage,3 wereinterpreted as hafted hunting weapons—used either by thrust-ing (Shea 2006) or thrown over short distances (Boeda et al.1999)—often doubling as cutting tools.

The most compelling evidence for hafting comes fromUmm el-Tlel, where ca. 70,000 yr ago bitumen was procuredfrom a distance of 40 km, processed (minimal heating wasinvolved), and used systematically as an adhesive for a largenumber of retouched and unretouched blanks, including butnot exclusively Levallois points (Boeda et al. 1998, 1999, 2008;see also Friedman et al. 1994–1995). This practice stoppedbetween 70,000 and 40,000 yr ago even though the productionand use of Levallois points continued. Combined with thelow frequencies of MP points with impact fractures (e.g., Shea2006; Villa et al. 2009), these data imply that pointed formswere not necessarily used as composite hunting weapons(Boeda et al. 1999; Plisson and Beyries 1998; Shea 2006).There is also sporadic evidence for the use of bitumen in thenearby site of Hummal (Hauck et al. 2013).

Whether hafting occurred and whether it was restricted tospecific-function tools are questions of interest because a shiftfrom handheld tools (e.g., the Schoningen wooden spears;Thieme 1997) to hafted (composite) tools was a significanttechnological innovation and possibly bears on higher cog-nitive capacities (e.g., Greenfield 1991; Wadley, Hodgkiss, andGrant 2009).

Hearths and Combustion Features

Combustion features have been documented in long cavesequences from the late LP onward (Henry et al. 2004; Kar-kanas et al. 2007; Meignen, Goldberg, and Bar-Yosef 2007;Shahack-Gross et al. 2008). The presence of burned lithicdebris and artifacts suggests that the lack of clear hearths instratigraphically shallow open-air sites is potentially a pres-ervation bias (see Gilead 1988 for an exception).

Fireplaces in Levantine MP sites were either flat surface(“simple”) or pit hearths (March et al. 2012) constructedwith wood without evidence for use of auxiliary combus-tibles (Albert, Berna, and Goldberg 2012; Albert et al. 1999,2003; Madella et al. 2002; Shahack-Gross et al. 2008). Be-ginning from the Early MP (EMP, ca. 250,000–170,000 yrago), combustion features occurred as small, thin lensesrepresenting flat fires directly laid on the soil substrate; aslarge hearths, sometimes in basin-like shallow structures;and as massive (primary or secondary) accumulations ofashes (Goldberg and Bar-Yosef 1998; Meignen, Goldberg,and Bar-Yosef 2007; Rabinovich and Hovers 2004; Sha-

3. Attempts to identify other diagnostic wear—e.g., Wallner lines,which provide evidence about projectile, thrusting, or other manners ofusing hunting armature (Hutchings 2011)—on pointed items in assem-blages from the whole MP time range have been unsuccessful, possiblybecause of the use of flint (K. Hutchings, personal communication, 2009.)

hack-Gross et al. 2008; Yeshurun, Bar-Oz, and Weinstein-Evron 2007). The different hearth types may have serveddifferent social or economic roles, though empirical datato this effect are still missing.

Pigments, Shells, and Burials

Several sporadic instances of processing and use of “non-utilitarian” materials are known from Levantine MP sitesdated to marine isotope stage (MIS) 5e-b. Ochre has beenreported from Skhul and Qafzeh caves (120,000–90,000 yrago). It was collected locally (10–30 km from the sites, contraSalomon et al. 2012). In both sites pyrotechnology was usedto transform yellow goethite to red hematite. The operationalsequences of preparing ochre powder involved in some in-stances heating as well as scraping of ochre lumps with sharpstone tools, abrasion, and grinding (Godfrey-Smith and Ilani2004; Hovers et al. 2003; Salomon et al. 2012). In both in-stances the pigment use was inferred to have been of symbolicimportance.

No dietary expansion into marine resources has been doc-umented in the Levantine MP, contrary to its European andAfrican counterparts (Avery et al. 2008; Cortes-Sanchez et al.2011; Finlayson, Barton, and Stringer 2001; Klein and Cruz-Uribe 1996; Marean et al. 2007; Stiner 1994; Stringer et al.2008). Nutritionally insignificant, naturally perforated shellsoccurring sporadically in Qafzeh and Skhul are interpretedas beads (Bar-Yosef Mayer, Vandermeersch, and Bar-Yosef2009; Vanhaeran et al. 2006; Walter 2003). The shells aresimilar to those found in Africa during oxygen isotope stage5–4 (Nassarius gibbosulus in Skhul) and in European sites(Glycemerys asurbica in Qafzeh; Bar-Yosef Mayer, Vander-meersch, and Bar-Yosef 2009; Zilhao et al. 2010). The distanceof Qafzeh from the Mediterranean shore (∼40 km) suggestsselective collection and inland transport (or exchange).

Incised cortices on Levallois cores are known from Qaf-zeh Cave, where some of the ochre lumps are also incised,and Quneitra (∼54,000 yr ago; Goren-Inbar 1990; Hovers,Vandermeersch, and Bar-Yosef 1997; Hovers et al. 2003;Marshack 1996). The presence of ochre residues withinsome of the incisions on the Qafzeh item (Nowell, d’Errico,and Hovers 2001) suggests that the two types of finds arelinked causally.

Several instances of modern (Skhul and Qafzeh) and Ne-anderthal (Tabun, Kebara, Amud, Dederiyeh) skeletal remainsare identified as burials based on their specific interment con-texts (Belfer-Cohen and Hovers 1992; Hovers, Kimbel, andRak 2000; Hovers et al. 1995; Nilsson 1998; Tillier 1990).Burials are found in thick undifferentiated deposits (Skhul),or within several stratigraphic horizons, interment took placein a given locality over a period of time. When adjusted forsediment volume per time (calculated for Qafzeh, Amud, andKebara), the behavior seems more common among modernhumans.

Hovers and Belfer-Cohen Variability and Complexity S341

From Data to Behavior

Decisions made by hunter-gatherers regarding “the selectionand integration of strategies for making, using, transporting,and discarding tools and the material needed for their man-ufacture and maintenance” (Nelson 1991:57) culminate intechnological organization: a repertoire of behavioral strate-gies aiming to reduce survival risks by providing technologicalmeans when needed (Bamforth and Bleed 1997; Bousman2005; Johnson 1978; Moore 1981; Reynolds 1978; Ugan,Bright, and Rogers 2003). It entails variable degrees of mo-bility and raw material curation (recycling and maintenance;Andrefsky 1994; Binford 1977, 1979, 1989; Bleed 1986; Kuhn1993, 1994; Marks 1988; Shott 1989) in response to fluctu-ations in the temporal and spatial distributions of resources.Table A1 in CA� online supplement A shows archaeologicalproxies for organizational strategies.

In the rich Mediterranean ecological zone, the clumpeddistributions and geographic seasonality gradients of vegetalresources combined with the small and stable territories ofthe main prey animals created nearly yearlong patchy mo-saics.4 Availability of peak quality resources was segregatedspatially and temporally. The importance of annuals andgrasses would increase at the expense of trees as one movedeast and/or south, where plants and animals that are obligatedrinkers are found near water bodies and could be exploitedthroughout the year, albeit intermittently. Beyond such focallocalities (e.g., the sites of Hummal or Umm el Tlel), plantsin the semiarid zone and in hyperarid areas would be clusteredin the winter along ephemeral drainages. Gregarious migra-tory species or resident species frequenting water sourceswould be a comparatively abundant and predictable food sup-ply even if only for a short term.

Ecological considerations (e.g., Dyson-Hudson and Smith1978; Kelly 1983, 1995) suggest that in low and medium-lowlatitudes, plant resources play a major role in determining theintervals, timing, and distances of group mobility because theyrequire continuous resource monitoring to ensure accuratescheduling of their harvest, and because long-distance trans-port of abundant yet short-lived resources is economicallywasteful in the absence of storage technology. Such constraintsare optimally mitigated by moving consumers to the resourcepatch (residential mobility; Bar-Yosef and Meadow 1995; Met-calfe and Barlow 1992; O’Shea 1981). Given a patchy resourcedistribution over short distances, MP hominins in the Levantmight have prioritized high residential mobility levels andopted to curate raw material by artifact recycling (“remakingan implement into a different kind of tool”; Odell 1996:95)and blank maintenance by reshaping (e.g., resharpening;

4. Hovers (2009:200–204) presents details and references for the eco-logical conditions inferred from paleoenvironmental data and reconstruc-tions.

Bamforth 1986; Eren et al. 2005; Geneste 1985; Thiebaut etal. 2010; see Nelson 1991).5

Moving Around: Artifact Use-Lives

Throughout the MP, the characteristic pattern of raw materialprocurement involved transport of raw material as partly pre-pared cores or as blanks over distances typically ranging from5 to 20 km to caves or open-air sites, transforming them intotemporary raw material sources (Kuhn’s [1995] strategy of“provisioning of places”).6 Assemblages usually contain prod-ucts from all postdecortication stages of core reduction—corepreparation, blank removal, core reshaping, blank retouch,and discard. Blank recycling and reshaping in order to extendthe artifacts’ use-lives are more common in EMP assemblageswhere retouched tools are relatively abundant. In late MPassemblages, the final shape of a tool was dictated mainlyduring flaking rather than through retouch. Retouch intensity(invasiveness on the blank’s surface) and extent (length alongthe edge) tend to be low regardless of the production systememployed; the distance from raw material source is correlatedwith the amount or intensity of retouch only occasionally(Hovers 2009:209–211).

Raw material was redistributed from workshop/habitationsites across the landscape as tools intended for activities else-where. Levallois blanks constituted the mobile component ofthe knapped stone technology because they were bigger andrelatively thinner than other flakes, providing higher utilityratios per unit mass (Eren and Lycett 2012; Geneste 1985:526;Hovers 2009:77–80; Kuhn 1994). In the absence of long re-fitted sequences (Far’a II and Tor Faraj being exceptions; De-midenko and Usik 2004; Gilead 1988), the best indication fortransport of finished blanks to and from sites (possibly asindividual tool kits; Binford’s [1979] strategy of “personalgear”) are blank/core ratios. Elevated frequencies of pointedand elongated blanks in assemblages where suitable cores andcore-trimming elements are rare suggest import of preparedblanks (Ekshtain et al. 2013; Henry 1995a; Jelinek 1982; Meig-nen 1998; Sharon et al. 2010), while the absence of such

5. It has been argued that in Europe Neanderthals were also highlymobile because of their heavy reliance on faunal resources, which areless dense on the landscape (see Kuhn 2013). One of the first adaptationsof Neanderthals in the Levant would have been reduction in costs ofmaintaining their metabolically expensive anatomy, which would selectfor a fast organizational response by adjusting their extractive behaviorand associated technological organization to the different local conditions.Moreover, if European Neanderthals relied on plants more heavily thanpreviously thought (El Zaatari et al. 2011; Henry, Brooks, and Piperno2011; Hockett 2012), changes from “European” to “Levantine” organi-zational strategies may not have been so drastic. A biological responseto relaxed environmental pressures, e.g., in the form of changes in theproportions of Levantine Neanderthal long bones (Arensburg and Belfer-Cohen 1998 and references therein) may have lagged behind.

6. Transport distances to Negev sites are argued to be on the scale oftens or hundreds of meters (Gilead 1980; Munday 1976), and raw materialexploitation in them is more expedient (Hovers 2009:280).

S342 Current Anthropology Volume 54, Supplement 8, December 2013

elements from assemblages where appropriate cores do existimplies their removal from the assemblages (e.g., Qafzeh VII,sites D44, D45, and D2 in the Negev, Kebara VIII, and TabunC; Hovers 2009; Meignen and Bar-Yosef 1992; Munday 1976;Shea 1991).

The patterns of raw material choice and transport to “cen-tral places” observed regionwide are significant in showingthat (1) raw material was evaluated on the basis of techno-logical knowledge rather than exploited opportunistically onthe basis of its availability; (2) in many instances, raw materialprocurement was conducted as a focused activity at distancesthat may have exceeded daily foraging ranges; (3) throughoutthe MP, lithic technological activities were organized acrossfamiliar territories; and (4) the differences in core designs(e.g., Wallace and Shea 2006) and in frequencies of heavilycurated items between the EMP and the later MP suggestchanges in organizational strategies (table A1). While resi-dential mobility seems to have been prioritized in both, toolavailability was negotiated differently in the EMP comparedwith the later MP, with different emphases on raw materialcuration/recycling and tool maintainability (in the sense ofBleed 1986), respectively.

Places of Activity: Caves and Open-Air Sites

Some open-air sites are located next to discrete resources (e.g.,quarry sites on raw material sources; hunting stations in prox-imity to water bodies where fauna could be anticipated). Ac-tivities may have been focused on specific tasks (e.g., Marksand Freidel 1977; Sharon and Oron 2013), leading to differ-ences in technology-assisted activities and therefore in lithicassemblage compositions (e.g., Marks and Freidel 1977). Dif-ferences in the compositions of lithic assemblages can be ex-pected between open-air sites, which might have served moreoften as task-specific places as compared with cave sites, whichwere used more often as habitation sites.

The evidence from Levantine MP caves is generally con-sistent with their being central places/habitation sites. Thelithic assemblages represent variable mixtures of provisioningstrategies as well as production and maintenance of stonetools on-site. Faunal remains indicate selective transport basedon body size, meat utility, and marrow content: large fauna(rhinos, aurochs, equids) are scarce, whereas gazelle- and fal-low deer-size prey were brought into the caves as nearly com-plete carcasses (Rabinovich and Hovers 2004; Rabinovich andTchernov 1995; Speth 2012; Speth and Clark 2006; Speth andTchernov 2007; Stiner 2005; Yeshurun, Bar-Oz, and Wein-stein-Evron 2007). The faunal assemblages from open-air sitesinclude the remains of large-bodied mammals in addition tothe smaller species known from caves (Davis, Rabinovich, andGoren-Inbar 1988; Gilead and Grigson 1984; Griggo 1998; E.Hovers et al., unpublished manuscript, 2013; Rabinovich andHovers 2004; Sharon et al. 2010; Zaidner et al., forthcoming).Thus open-air sites complement subsistence patterns gleaned

from cave occupations. Beyond the trivial observation thatonly animals that were available could be hunted, the com-positions of faunal assemblages reflect complex decision-mak-ing processes rather than mere availability (Speth 2012). Thus,lithic raw material management in most open-air localitiesresembles that of caves: provisioning, on-site knapping, lowinvestment in blank recycling, and high typological diversity.An important difference that sets open-air sites in the Med-iterranean zone apart from caves is the low frequencies ofLevallois blanks and the elevated frequencies of expedienttools such as notches/denticulates (fig. 2; Hovers 2009, table8.2). The exceptions observed at Umm el Tlel and Hummal(Boeda, Griggo, and Noel-Soriano 2001; Hauck 2011) maybe related to their locations in oases, often being used ashabitation sites.

Lithic and faunal data converge to show that Levantine MPsites in the Mediterranean zone, in various geographic loca-tions and in cave and open-air contexts, were usually a “mixedbag” of organizational practices. Outside of the Mediterraneanecological zone, assemblages are more skewed toward mobiletechnological elements (Hauck 2011; Hovers 1997; Shea1998).

Places of Activity: Features of Levantine MP Occupations

Caves and a few open-air sites (Hummal, Umm el Tlel;Nesher-Ramleh; Zaidner et al., forthcoming) were repeatedlyoccupied; accumulated stratigraphic sequences span ca.200,000–20,000 yr, yet with stratigraphic and occupation hi-atuses. Other open-air sites represent shorter temporal ranges.Complex depositional processes and high densities obfuscatedistinctions between isolated occupations of either short orlong duration versus some long and intensive occupations aswell as the estimates of the occupying group sizes.

Explicit as well as latent spatial patterns show that LevantineMP sites are not randomly cluttered debris concentrations.Mainly late MP sites show indications for differential use ofspace. Some spatial patterns are site specific while others arerecognized repeatedly in various sites. The latter includehearth-related knapping activities, ash cleaning into desig-nated areas, and evidence for well-defined activity areas ex-clusively related to animal carcass treatment/discard to theexecution of specific lithic reduction sequences or the disposalof human remains. Separation of lithic reduction stages acrossoccupation space—for example, core preparation in one areaand advanced knapping carried out near hearths in anotherarea as well as areas designated for stone stocking—were alsoobserved (Alperson-Afil and Hovers 2005; Garrod and Bate1937:62–63; Gilead 1988; Henry et al. 2004; Hietala 2003;Hovers et al. 1995, 2011; Meignen, Goldberg, and Bar-Yosef2007; Oron and Goren-Inbar 2013; Rabinovich and Hovers2004; Schick and Stekelis 1977; Shahack-Gross et al. 2008;Speth 2006; Speth et al. 2012). Particular activities were some-times conducted repeatedly in the same area within a site,

Figure 2. Values of IL, ILty, and percentages of retouched blades in various MP sites. Reprinted from Hovers (2009, fig. 8.4) withpermission from Oxford University Press.

S344 Current Anthropology Volume 54, Supplement 8, December 2013

suggesting that the specific functional assignments of areaswere retained over time.

The functions of sites within their respective settlementsystems changed through time. Taphonomy, body part andage distribution of ungulate remains, and frequency changeswithin the pertinent lithic assemblages suggest that the rolesof sites shifted between hunting to habitation or butcheringlocations (Boeda, Griggo, and Noel-Soriano 2001; Hauck2011; Meignen et al. 2006; Speth and Tchernov 2007). Faunaldata have also been used to infer occupation duration/inten-sity. At EMP Hayonim Cave (ca. 200,000 yr ago), the largesizes of slow-growing tortoises that were harvested by thehominins implied infrequent and ephemeral occupations (Sti-ner, Munro, and Surovell 2000). In contrast, throughout thelate MP sequence of Kebara Cave (60,000–48,000 yr ago), thesite’s occupants gradually shifted to exploitation of smaller-bodied animals and younger individuals of those species thatoffered lower meat or fat returns. This has been interpretedas hunting at ever-increasing distances from the site to com-pensate for locally diminishing returns resulting from over-hunting (Speth 2004; Speth and Clark 2006). Hunting pres-sure throughout the Upper Pleistocene may have had along-term negative effect on the abundance of large ungulates(Davis, Rabinovich, and Goren-Inbar 1988).

An emphasis on “simple” (bidirectional or unidirectional)core reduction modes (Kuhn 1995; Parry and Kelly 1987;Wallace and Shea 2006) increases gradually during the laterpart of the MP. This may correspond to either reduction inmobility levels or an increase in site use intensity. However,changes in site function(s) are not necessarily reflected in allfeatures of the occupation. At Qafzeh Cave, hearths, burials,and micromammals occur only in the early layers, yet changesin lithic technology do not correspond to this dichotomy.Shifts in retouched tool-type frequencies crosscut the dis-tinction between the lower and upper sections, and the pro-portions of the large ungulates remain similar through time(Bar-Yosef and Vandermeersch 1993; Hovers 2009; Rabino-vich and Tchernov 1995).

What Is “The Levantine MP Variability”?

Because of the intrinsic properties of the prehistoric record(Goring-Morris, Hovers, and Belfer-Cohen 2009) but alsobecause of taphonomic and research biases, the hominin evo-lutionary record is constructed from unequal distributions ofdata points in time and space. Characterizations of LevantineMP lithic technology sometimes selectively focus on isolatedtechnological aspects rather than whole-assemblage proper-ties. Still, the Levantine MP can be characterized as a discretearchaeological phenomenon. Behavioral features that contin-ued from earlier periods survived throughout the record, oth-ers that were lost early on were not regained, and a few in-novations held on throughout the period.

The Beginning of the Levantine MP: Gains, Losses, andVariations on Themes

The onset of the MP in Eurasia and of the MSA in Africa ischaracterized by the disappearance of LP bifaces from thearchaeological record,7 which appears as a rupture from earlierindustries, concurrent with the proliferation of Levallois flak-ing as a system for flake production. The origins of Levalloisflaking concepts in the Levant are traced to the late Acheulian,with two pathways identified for this change: recycling hand-axes into (centripetal) Levallois cores for preferential flakes,and flake removals from the center of a core surface butwithout rigid hierarchical relationship between core surfacesand platforms (e.g., DeBono and Goren-Inbar 2001; Malin-sky-Buller, Grosman, and Marder 2011).8 However, in the lateLP Acheulo-Yabrudian technocomplex (400,000–220,000 yrago) that immediately precedes the MP, evidence for Levalloisflaking (e.g., in Adlun, Tabun, Masloukh, Qesem Caves, Ya-brud I Rockshelter) is nonexistent (Gopher et al. 2005) orscanty at best (Copeland 1983; Garrod and Bate 1937:79–89;Jelinek et al. 1973; Skinner 1970; E. Hovers, personal obser-vations). This is unlike the long “transitional” phases in Africaor Europe at more or less the same time.

The systematic production of elongated blanks is well doc-umented in the Acheulo-Yabrudian, yet these flaking tech-nologies (Shimelmitz, Barkai, and Gopher 2011) do not seemto continue into the EMP. The increased control over cores(and thus over the shapes of the products), which is a tech-nological advantage of the Levallois and Laminar systems, ledto a range of recurrent flake morphologies that are the hall-mark of the Levantine Mousterian throughout its course. Sys-tematic production of pointed elongated blanks as desiredproducts, obtained through such enhanced lithic technolog-ical control (sometimes with further modification by re-touch), is a novel component of EMP flaking systems andappears to be a local innovation (see below).

Shea, Davis, and Brown (2001) concluded from experi-mental work that the early elongated points were more suitedfor use as knives than as tips of heavy thrusting weapons, andtheir emergence in the Levantine MP record may representthe inception of hafting techniques. Importantly, elongatedpoints co-occur in the earliest MP assemblages with short,subtriangular Levallois points (Ashkenazi 2005; Copeland1975; Meignen 1998, 2011; Monigal 2002; Nishiaki 1989;Weinstein-Evron et al. 2003; Wojtzak 2011). The two mor-photypes undergo a frequency shift over the course of the

7. MP bifaces in Europe are a late phenomenon, different technolog-ically from the Acheulian ones (e.g., Soressi 2005, and references therein).Therefore, they should not be considered as a continuous trait but ratheras technological innovations within the European MP.

8. Similar phenomena during early MIS 8 in Europe (Moncel et al.2011; Tuffreau 1995; White, Ashton, and Scott 2011) are identified asimmediate precursors of Levallois methods, whereas in eastern Africathey are regarded as fully Levallois, yet they occur in a mosaic patternwith ESA bifacial elements (McBrearty and Tryon 2006; Tryon, McBrearty,and Texier 2005).

Hovers and Belfer-Cohen Variability and Complexity S345

MP, with the frequencies of the long points declining abruptlyfrom ca. 170,000 yr ago, while short (unretouched or lightlyretouched) Levallois points become more numerous (thoughrarely the major component) through time (Hovers 2009:216,table 8.3; but see Hauck 2011).

It is doubtful that the rise of broad-based points is relatedto changing extractive strategies or to their greater efficacy ashunting weapon components. By current archaeozoologicalevidence, species compositions remained similar from theLate LP throughout the MP.9 This is especially true for theMediterranean ecological zone because of its higher sustain-ability in the face of climate variations (Belmaker and Hovers2011; Hovers 2009 and references therein). Hence, changesin faunal resource structure need not have driven the changein point frequencies (Shea 2006).

Habitual fire making and fire control are the main behav-ioral features that continued uninterrupted from the late LPinto and through the MP, with a range of variation in sizeand shape of the fire structures encountered from the EMPonward (e.g., Weiner, Goldberg, and Bar-Yosef 2002; Yeshu-run, Bar-Oz, and Weinstein-Evron 2007). In contrast, ochre(and the evidence of its heating to change its color) as wellas marine shells occur briefly, during MIS 5e-b, only in Skhuland Qafzeh. Another MP innovation is hafting with adhesives.While its earliest known occurrence is in late Middle Pleis-tocene Europe (Mazza et al. 2006), it appeared in the Levantonly during MIS 4 and involved different components fromthose used either in Europe (where birch tar was used inaddition to bitumen; Carciumaru et al. 2012; Grunberg 2002;Pawlik and Thissen 2011) or in Africa (ochre; Lombard 2007;Rots and Van Peer 2006; Rots, Van Peer, and Vermeersch 2011;Wadley 2005), suggesting that it may be another local in-novation, again involving the use of fire as an engineeringtool. Because of the functional advantages of hafting withadhesive, Boeda et al. (2008) ascribe the gap in bitumen useat Umm el Tlel to changes in territorial behavior, implicitlyassuming its continuous use.

The evolution of mortuary behavior in the Levantine MPdiffers from either Africa or Europe, where, respectively, bonemodifications and polishes unrelated to cannibalism (Bodo,Herto) and body caching and cannibalism (Sierra de Ata-puerca) appear already in the Middle Pleistocene. In Europethis behavior developed into interment in the Upper Pleis-tocene (Hovers and Belfer-Cohen 2013; Pettitt 2010). In theLevantine MP, burial appears without recognized precursorsduring MIS 5–3. The stratigraphic associations of “nonutili-tarian” objects (see above) strictly with burials of modernhumans provides contextual support for the suggestion thatthese traits may constitute a novel symbolic/behavioral suite

9. Large-game species usually reflect local availability (Bar-Yosef 2004),though there is a possibility of choice-related bias (e.g., Marder et al.2011).

of this population (Hovers, Vandermeersch, and Bar-Yosef,1997; Hovers et al. 2003).10

Changes within the Levantine MP Record

A now-classical model of linear development of LevantineMP lithic industries based on the sequence of Tabun Cave(Bar-Yosef 1998; Copeland 1975; but see Ronen 1979; tableB1 in CA� online supplement B) emphasizes the presenceor absence of a few selected technological traits (e.g., Laminarproducts, specific modes of Levallois flaking or of Levalloismorphotypes) as characteristics that differentiate three chron-ocultural phases of the Levantine MP. However, as the di-versity of MP flaking systems (i.e., the number of flakingsystems in the technological repertoire) had already becomeestablished in the EMP and persisted throughout the period(Goren-Inbar and Belfer-Cohen 1998), assemblages portrayshifts in the frequencies of technological variants rather thanany single dominant variant or morphotype. This is a sig-nificant distinction, because an “assemblage” is a basic ana-lytical unit that nonetheless averages events of group-leveldecision making. Following this, interassemblage variationsand how they are patterned chronologically and geographi-cally offer interesting insights as regards the behaviors theymight reflect.

Interassemblage variability is identified along two mainaxes. Along the temporal axis, a major change occurred be-tween the EMP and the later MP in relative frequencies oftechnological variants, with more obvious raw material andartifact curation strategies observed in the earlier MP. Geo-graphic variations are less conspicuous in the EMP.11 In con-trast, interassemblage variability in the later MP patternslargely geographically (the Mediterranean vs. the semiarid andarid zones) even though the strategies of raw material curationin both zones emphasize the knapping of maintainable, ver-satile tools (Bleed 1986). Within the Mediterranean zone, laterMP interassemblage variability decreases with geographic dis-tance; that is, assemblages from sites that are geographicallycloser to one another (or occur at the same locality through-out its stratigraphy) are more similar than those that arefarther away.12 With the exception of Tabun, assemblages ina given sequence tend to be more similar to one another(Hovers 1997).

10. The claim is stronger in Qafzeh given the more accurate excavationtechniques applied during fieldwork. The Skhul evidence is not as per-suasive. Note that ochre, shells, or incised items do not occur in theburials yet are associated with them stratigraphically and are absent fromcontexts without burials.

11. Although this may be partly related to the lower resolution of theEMP record.

12. This is probably the reason why the temporal Tabun model failsto encompass the full range of variability of the Levantine Mousterian(Goren-Inbar and Belfer-Cohen 1998; Hovers 1998b; Meignen 2011). Themodel holds better for assemblages within the Mediterranean region(Hauck 2011; Hovers 1997, 2009) but is less successful in chronologicalattribution of assemblages outside the Mediterranean zone.

S346 Current Anthropology Volume 54, Supplement 8, December 2013

Figure 3. Lithic densities/accumulation time, in four MP sites.Data for Hayonim and Kebara, Bar-Yosef (1998); for Qafzeh,Hovers (1997); for Amud, E. Hovers, personal observation. Acolor version of this figure is available in the online edition ofCurrent Anthropology.

Changes in Settlement Patterns

On the basis of paleoecological reconstructions, the distri-bution and structure of resources did not shift significantlythroughout the MP, and human groups would tend to besmall and highly mobile. The change in curation behaviorthrough time therefore suggests a shift in associated organi-zational strategies—specifically, the geographic scale of groupmovements. It occurs in tandem with changes in the markersof occupation intensity within the Mediterranean zone: den-sities of the lithic assemblages (fig. 3) and faunal remains pertime unit as well as investment in site features seem to increasefrom ca. 130,000 yr ago, setting EMP occurrences apart fromlater ones.13 The archaeological record suggests that at leastfrom that time, larger populations were concentrated withinthe Mediterranean zone, ecologically the most attractive partof the Levant.

Such changes in modes of occupation and the geographicpatterning of interassemblage technological variability areconsistent with a scenario of increased group territorialitywithin the Mediterranean zone. MP groups that convergedinto the Mediterranean zone would exploit more efficientlythe seasonal availability of resources within their contractedterritories.14 Repeated burials in specific locations may be aresult of this contraction process, perhaps denoting the emer-gence of some degree of territoriality and corporateness (Hov-ers 2001, 2009). Ecological constraints (see above) may havekept group size at low numbers, but sites within a group’sterritory would be visited more frequently, leading to a morerapid and extensive accumulation of archaeological remainsand to recurring spatial relationships between domestic fea-tures within the sites (Hovers 1997, 2001; Meignen et al.2006). Outside of the Mediterranean zone, groups probablystill exploited larger territories (Boeda et al. 2008; Hauck2011).

The changes in settlement patterns between the EMP and

13. Misliya Cave, with a large hearth (Yeshurun, Bar-Oz, and Wein-stein-Evron 2007) and high lithic and faunal densities (Y. Zaidner, per-sonal communication), is exceptional in the EMP. One possible expla-nation is that it was the locale for social interactions of groups thatotherwise used sites ephemerally throughout their large exploitation ter-ritories. Initial observations suggest an increase in artifact and faunaldensities at Nesher-Ramleh at ca. 170,000 yr ago (Zaidner et al., forth-coming).

14. An area of 120,000 km2 in the Mediterranean zone (roughly thesize of the area in which MP sites are known) could carry between 8,640and 14,992 persons (see Hovers 2009:206 for discussion; Shea 2004:170–173). Note that the estimates here are based on values for extant groupsin temperate forests, where primary production is lower than in theMediterranean forests; also, effects of technological extractive aids wasnot factored into the equations. Thus carrying capacity may have beenhigher during MP times. Based on these numbers alone, a 25-personband of hunter-gatherers could survive in the Mediterranean zone in anannual territory as small as 350–400 km2 (realistically, it would have beenslightly larger); a viable population of 500 individuals (Wobst 1974, 1976)would require a territory of at least 800–1,000 km2. In terms of carryingcapacity, viable groups could survive in small territories as postulatedhere.

later MP are not tantamount to an overall increase in pop-ulation size. If anything, settlement patterns in the very lateMP and in the UP show more ephemeral than intensive oc-cupations (Goring-Morris, Hovers, and Belfer-Cohen 2009;Rabinovich 2003; Speth and Tchernov 2007). There are noindications for social, ritual, and cultural innovations thatmight be associated with robust demographic networks ofhigh-density (or growing) populations (e.g., Hovers and Bel-fer-Cohen 2006; Jochim 1983; Powell, Shennan, and Thomas2009; Shennan 2001). To the contrary, indications for suchcultural intensification (e.g., shell beads or pigment use) in-creased gradually throughout the UP from sporadic occur-rences in the earliest phases (e.g., Bar-Yosef Mayer 2005; Kuhnet al. 2001), while burials occur only later in the UP. Someconceptual approaches to the volumetric management ofcores were retained from the MP to the UP, and minute yetsubstantial changes in core-shaping techniques (e.g., abrasion)as a way to facilitate hafting occurred only in later stages ofthe UP (Belfer-Cohen and Goring-Morris 2007, 2009; Belfer-Cohen and Hovers 2010; see Meignen and Bar-Yosef 2002).

Given these observations, a parsimonious explanation ofthe trends in land-use patterns within the later MP is thatthey depict a historical occurrence rather than a long-termevolutionary trajectory of gradual population growth duringthe MP. The posited human density in the Mediterraneanzone is attributed to the effects of changing landscapes andenvironmental constraints (e.g., high sea levels, high standsof Lake Lisan in the Dead Sea Rift that reduced the area ofhabitable land, greater aridity outside of the Mediterraneanzone), which may have resulted in an influx of people fromthe arid areas. Possibly this is also related to dispersal events

Hovers and Belfer-Cohen Variability and Complexity S347

into the Levant during environmental windows of opportu-nity in the later MP (Bar-Yosef 2000; Drake et al. 2011; Lahrand Foley 1998; Vaks et al. 2007).

Discussion

From the broadscale, low-resolution cultural evolutionaryperspective, the Eurasian MP record has been perceived as aperiod of cultural stasis over some 200,000 yr without cu-mulative changes that led to cultural evolution. This has beenattractively conceptualized by “rugged fitness landscapes”(Boyd and Richerson 1996; Dobzhansky 1951), whereby sig-nificant, costly changes that enhance fitness only occurredwhen physical and social conditions on the adaptive landscapewere disrupted dramatically (e.g., the shift to sedentism; Bel-fer-Cohen and Bar-Yosef 2000). This broad perspective, how-ever, does not tell us much about the dynamics that createdand preserved such stasis. If we are to address questions aboutalternative pathways to a known end result of an evolutionaryprocess, a higher-resolution record of the historical details isrequired. Otherwise explanatory hypotheses become tauto-logical.

Potential Causes of Variability

The Levantine MP constitutes a distinct entity within the MPworld (Bar-Yosef 2006). This overview has underlined theproperties that distinguish it. The repertoire of material cul-ture behaviors defining the Levantine MP falls within therange known from Europe and Africa, yet the abruptness ofboth loss and emergence of the defining lithic characteristicsdiffers from the gradual mosaic-like processes observed in thelate Middle Pleistocene of Europe and Africa. Some nonlithicbehaviors—cave occupation, the use and control of fire, anda prime-adult, large-game focused hunting—continued overfrom the LP. The LP/MP turnover in lithic practices was fol-lowed by 200,000-yr-long variations on a restricted repertoirewith few and relatively short-lived innovations.

A detailed terrestrial paleoclimatic record of the UpperPleistocene Levant enables an understanding of the effects ofclimate variations on a regional scale, which is more relevantthan globally recognized climate patterns for understandingthe behavior of prehistoric foragers. Given the mild amplitudeof Levantine climate changes, this leads to the insight thataspects of material culture variability, even if they are envi-ronment related (e.g., patterns of raw material use, tool func-tions), do not respond directly to climatic shifts. Similarly,predictions that lithic variability or organizational strategieswill be clearly dichotomized according to the two homininspecies present in the Levantine MP are questionable boththeoretically and empirically (Hovers 2006, 2009; Liebermanand Bar-Yosef 2005).

Microevolutionary Processes: Cultural Transmission

An alternative hypothesis for explaining the variability in thematerial culture record of the Levantine MP invokes micro-evolutionary mechanisms of cultural transmission (e.g., Bet-tinger, Boyd, and Richerson 2009). The discernible changebetween the EMP and later MP reflects reorganization of land-use patterns. Interassemblage variability observed in the laterMP stems from flexible combinations of provisioning andcuration strategies along ecological gradients. We have sug-gested that these changes are related to demographic shifts.

Indeed, loss and retention of cultural diversity are associ-ated with demographic properties; for example, loss of di-versity is greater, and will occur faster, in small groups. Somedemographic events (local extinctions, demic diffusion) act(respectively) as direct causes of eradication/reduction of cul-tural traits (Henrich 2004; Mesoudi 2011b; Premo and Kuhn2010).15 Otherwise, changes in cultural diversity may accu-mulate because of random drift (Mesoudi 2011a; Neiman1995) or socially mediated (“biased”) cultural transmissionby individuals, which leads to group-level changes (Mesoudi2011a; Richerson and Boyd 2005). The latter takes on variableforms of “conformist” (frequency-dependent) and “prestige”(agent-dependent) biases. Both transmission mechanisms re-quire that social institutions of learning (e.g., socially rec-ognized learning environments where individuals can copyprevalent variants or select mentors from whom they chooseto learn) be in place (Henrich and Boyd 1998; Sterelny 2012).

Over time, both mechanisms lead to loss of diversity atvariable paces depending on the strength of model selectionby individual decision makers or the number of mentors(Boyd and Richerson 1985; Mesoudi 2011a, 2011b; Mesoudiand Lycett 2009). A population size threshold of N p ∼500differentiates between marginal and substantial loss of culturalcomplexity (Vaesen 2012:e40989).16 On the other hand, gainsin technological diversity (i.e., incorporation of inventions associety-wide innovations; Renfrew 1978) are not dependentonly on population size. Even if a new behavior is beneficialand population size is above 250 individuals, it will not spreadand cultural diversity will not increase if levels of intercon-nectedness (e.g., the number of linkages between individualsin networks of different mating behaviors) are low (Vaesen2012). Social learning can occur in kin-based learning envi-ronments (e.g., extended families) with a restricted trans-mission network or within broader social contexts that maypromote the spread of behaviors selected for learning becauseof their economic payoff or social rewards (Vaesen 2012;Young 2012). Even if the same behavior is transmitted within

15. Vital information may be lost when an influential or knowledge-able individual passes away without having passed on the information(Whallon, Lovis, and Hitchcock 2011).

16. Under certain conditions of the mathematical models, this thresh-old may be as high as 1,290 individuals (Vaesen 2012 and referencestherein).

S348 Current Anthropology Volume 54, Supplement 8, December 2013

each context, the end result in terms of cultural diversificationand accruing novel behaviors will differ significantly.

Given the idea that demographic aspects can spur or slowdown technological change (e.g., Henrich 2004, 2010; Powell,Shennan, and Thomas 2009; Shennan and Steele 2009), wefind it useful to consider the possibility of understanding theLevantine MP stasis from this microevolutionary perspective.Right now, this notion can be explored only informally, giventhe incompatibility of the evolutionary timescale of the MPrecord and the temporal scale of decision-making and socialtransmission processes (e.g., Bettinger, Boyd, and Richerson2009; Jordan 2007) as well as the incipient stage of the relevantformal modeling.

Recognizing Cultural Transmission Processesin the Levantine MP

To address processes of biased cultural transmission, two is-sues need to be tackled. The first is whether there existedenvironments of social transmission, and the second iswhether demographic continuity is a feasible scenario in theLevantine MP.

1. Although their archaeological signatures may be the mostversatile and most difficult to decipher, habitation sites werelikely the locations of significant intragroup social interactions(e.g., Rolland 2004). The use and control over fire generateda domestic and social context for knowledge transmission. Inaddition to the cooperation implied by hunting, transport ofhigh-utility body parts into caves and their butchery awayfrom hunting locations suggest that delayed consumption andfood sharing were part of the subsistence-cum-social activitiesthat took place within habitation sites. The evidence of do-mestic space allocation and differentiation, raw material con-centration and redistribution, and delayed returns in foodsharing, which occur variably in MP open-air and cave sites,suggests that the social relationship of Levantine MP hominins(moderns and Neanderthals) extended beyond immediateeconomic returns and was formally constructed. Such a socialorganization would provide opportunities for biased trans-mission of cultural traits, for example, highly specific ways ofstone tool making or a social perception of a site’s space (e.g.,David and Kramer 2001). Appearance of mortuary practicesas a local phenomenon in the later Levantine MP may besignificant in view of suggestions that the human practice ofburial evidences “new forms of social organization” (Binford1968). This contrasts with the Acheulo-Yabrudian popula-tions, whose social behaviors around fireplaces may have haddifferent contexts for resource redistribution and lower levelsof intergroup cooperation than did MP groups (Stiner, Go-pher, and Barkai 2011). Notably, the Acheulo-Yabrudian pat-tern itself may represent a shift from the social structure ofearlier Acheulian groups, hypothesized to have operated aslarge, loosely, and pragmatically united corporate groupsbased on dyadic relationships between mothers and their off-spring (Marx 2004).

At present there is no concrete evidence that Neanderthalsand moderns overlapped chronologically in the Levant. Thishas led to scenarios of competitive exclusion, recursion, oralternate extinctions due to climate (Shea 2003, 2006, 2008).However, the absence of evidence for temporal overlap maybe an artifact of the small samples of identifiable humanremains that archaeologists have at their disposal (Hovers2006). If all dated MP assemblages are considered, the tem-poral distribution is less discontinuous (fig. 4), and somedegree of continuity (and coexistence) cannot be rejected(Hovers 2006, 2009). Paleogenetic analyses (albeit notoriouslytransient) have implications that are potentially consistentwith this hypothesis (Green et al. 2010; though see Erikssonand Manica 2012).

The persistent occurrence of the same repertoire of flakingtechnologies throughout the MP is noteworthy. Long-termoccurrence of any single flaking system need not imply con-tinuity of information transmission (and thus demographic)systems. However, the consistent use of the same technologicalrepertoire is less likely to represent repeated independent rein-ventions because complex knapping processes do necessitateteaching (or, at least, direct observation).17 Accordingly, weassume some degree of demographic continuity in the Med-iterranean zone during the later MP. Still, we cannot—anddo not—argue that it prevailed at all times regionwide.

2. The archaeological record is spotty, and discontinuityscenarios are the null hypothesis in many studies of culturalevolution, rendering drift, biological extinctions, dispersals,and collapses of demographic systems as the main mecha-nisms implicated in shaping the pace and trajectory of Pa-leolithic cultural evolution (e.g., Boyd and Richerson 1985,2009; Henrich 2004; Mesoudi 2011a:8; Premo and Kuhn2010). Notably, the Levantine MP archaeological record ischaracterized by retention rather than gain or loss. The suiteof nonutilitarian innovations in Skhul and Qafzeh during MIS5, which disappeared altogether by MIS 4 time, represents anexception. Such traits appear slightly earlier and less sporad-ically in Africa (although there, too, they are not continuousthroughout the whole MSA time span; Barham 2002; Cain2006; Clark 2011; d’Errico, Garcıa Moreno, and Rifkin 2011;Marean et al. 2007; Texier et al. 2010; although see Marshack1981 and Roebroeks et al. 2012 for early occurrences in Eu-rope). These traits may have been introduced into the Lev-antine behavioral repertoire by dispersing populations ofmodern humans from Africa. Yet if there was indeed a degreeof demographic continuity in the Levant, it may be difficultto decide whether these temporally constrained phenomenarepresented demic or cultural diffusion processes (Collard etal. 2010). The loss of these features may have entailed physicalextinction but could alternatively be a reflection of collapse

17. Modern-day knappers have indicated that mastering any form ofLevallois knapping is training, skill, and cognitively expensive comparedwith earlier or later formal flaking systems (M. Eren, G. Sharon, and J.Shea, personal communication).

Hovers and Belfer-Cohen Variability and Complexity S349

Figure 4. Distribution of dated MP assemblages. Darker vertical bars indicate stratigraphic horizons with remains of anatomicallymodern humans; lighter vertical bars indicate stratigraphic horizons with Neanderthal remains. For Tabun, both the thermolu-minescence (TL) and electron spin resonance (ESR) date ranges are shown. The question mark and the arrow reflect the controversyabout the stratigraphic context of the Tabun C1 fossil and its corrected placement in layer B rather than C (base graphic, courtesyOfer Bar-Yosef). A color version of this figure is available in the online edition of Current Anthropology.

of the social networks of small groups (Hovers and Belfer-Cohen 2006; Hovers et al. 2003), where even beneficial traitsmay not survive (Vaesen 2012).

Conversely, there are surprisingly few indications, especiallyfor a region with documented population inflows, for diver-sification through diffusion (cultural or demic) from neigh-boring areas. Not a single piece that could be related to theAfrican MSA or European MP lithic systems is published fromLevantine MP archaeological contexts despite the presence ofhighly diagnostic morphotypes and practices in these areas(e.g., Aterian or bifacial points, Quina flaking, Micoquianbifaces). That elongated EMP points represented “a momentof contact” with African populations (Brooks et al. 2006:249)is questionable on chronological and technological grounds.

The Levantine MP emerges as a period when small humangroups had a robust enough social transmission environmentbut were not interconnected sufficiently to generate techno-logical innovations (Henrich 2010; Young 2012) leading tocultural evolutionary change that is observed archaeologically.The spatial parameters of the settlement systems identified inthe Levantine MP (e.g., small, local clusters) could have beenconducive to creating short-term diversity that was nonethe-less at high risk of being lost because of biased transmissionbefore becoming habitual innovations (conformity bias es-pecially weakens the relationship between diversity and in-novations; Kandler and Laland 2009). This explains why thelater MP, when groups were more densely packed in the Med-

iterranean zone, was not a period of accelerated rates of majorinnovations (the persistence of mortuary behavior being anotable exception). In this scenario, such groups had to keepoccasional physical contact to maintain viable populations(e.g., a site such as Misliya Cave could serve as an “aggregationlocality” where otherwise widely dispersed EMP homininscould have interacted occasionally). The minimal number ofmembers in such networks is ∼500 (Wobst 1974, 1976). Thisoverlaps with the suggested critical population size thresholdbeyond which major losses of cultural diversity occur whenbiased transmission takes place (Vaesen 2012).18

Concluding Comments

The Levantine case study, as perceived in this paper, indicatesthat environmental conditions influence human niche con-struction (Laland and O’Brien 2010) but in ways more in-tricate and subtle than has previously been suggested. Forexample, it has been suggested recently that habitual use offire emerged in Europe only post-LP (Roeberoeks and Villa2011) and that European Neanderthals were unable to controlfire (Sandgathe et al. 2011). It was also suggested that in

18. This scenario can be tested archaeologically because it predicts thatgiven similar ecological and economic conditions, assemblages in differentsites will differ in the fine nuances of lithic (or other, if preserved)technologies (e.g., Hovers 1997, 2009; Meignen 1995; and see MacKay2011). Research to that effect is currently under way.

S350 Current Anthropology Volume 54, Supplement 8, December 2013

Europe, Neanderthal social bonding was dyadic (as in theLevantine Acheulian) while exploiting very large territories(Kuhn and Stiner 2006; Stiner 2013). If those claims are val-idated, they may suggest that MP hominins in Europe asopposed to those in the Levant constructed different envi-ronments of social interaction as well as microevolutionarymechanisms of cultural transmission.

That said it is important to bear in mind that the geographicscale of the Levant differs significantly from those of Africaor Europe. Still, if we are to compare demographic patternsand cultural trajectories of the Levant with those largerregions, the geographic “clumping” that typifies culturaltransmission of small-scale social units requires detailed anal-yses on comparable geographic scales (e.g., Bocquet-Appeland Tuffreau 2009; de la Torre 2013; Wurz 2013).

Humble gains in technological diversity and slow changesin the pace of cultural accumulation characterize also theLevantine UP. Despite the diversification of lithic technologies(Goring-Morris and Belfer-Cohen 2003), it was not until wellinto the UP that beneficial technological innovations (e.g.,microlithization and highly composite tools: Belfer-Cohenand Goring-Morris 2002; the use of shell beads as informationtransmission systems: Kuhn et al. 2001) spread, indicating apossible change in the levels of interconnectedness of socialgroups (as modeled by Kandler and Laland 2009; Young2012). If a cultural and demographic “Upper Paleolithic Rev-olution” occurred, it happened during, rather than at thebeginning, of the UP. The pace of diversification and inno-vativeness remained relatively slow till the close of the Pleis-tocene. Thus most of the UP resembles the “state of affairs”observed in the local MP during which Levantine homininsoccupied suboptimal social fitness peaks, getting by with“good enough.” It is this particular slow pace of culturalchange that characterizes the Levantine prehistoric record ofthe late Middle and Upper Pleistocene and that opens upinteresting research questions about the mechanisms of cul-tural transformation and stability and the historical and evo-lutionary questions underlying them.

Acknowledgments

Steve Kuhn, John Speth, Christian Tryon, Yossi Zaidner, andtwo anonymous reviewers provided insightful critical com-ments on earlier drafts of this paper.

References CitedAhmad, G. S. 2009. The Middle Paleolithic stone tool assemblage from Ar

Rasfa: reconstructing Late Pleistocene human behavior in the Jordan RiftValley. MA thesis, State University of New York, Stony Brook.

Aiello, Leslie C. 2003. Neanderthal thermo-regulation and the glacial climate.In Neanderthals and modern humans in Europe during the last glaciation.Tjeerd H. van Andel and William Davies, eds. Pp. 147–156. Cambridge:McDonald Institute for Archaeological Research.

Albert, Rosa-Maria, Ofer Bar-Yosef, Liliane Meignen, and Steve Weiner. 2003.Quantitative phytolith study of hearths from the Natufian and Middle

Palaeolithic levels of Hayonim Cave (Galilee, Israel). Journal of Archaeo-logical Science 30:461–480.

Albert, Rosa-Maria, Francesco Berna, and Paul Goldberg. 2012. Insights onNeanderthal fire use at Kebara Cave (Israel) through high resolution studyof prehistoric combustion features: evidence from phytoliths and thin sec-tions. Quaternary International 247:278–293.

Albert, Rosa-Maria, Ofer Lavi, Lara Estroff, Steve Weiner, Alexander Tsatskin,Avraham Ronen, and Simcha Lev-Yadun. 1999. Mode of occupation ofTabun Cave, Mt Carmel, Israel, during the Mousterian period: a study ofthe sediments and phytoliths. Journal of Archaeological Science 26:1249–1260.

Almogi-Labin, Ahuva, Miryam Bar-Matthews, and Avner Ayalon. 2004. Cli-mate variability in the Levant and northeast Africa during the Late Qua-ternary based on marine and land records. In Human paleoecology in theLevantine corridor. Naama Goren-Inbar and John D. Speth, eds. Pp. 117–134. Oxford: Oxbow.

Almogi-Labin, Ahuva, Miryam Bar-Matthews, Dan Shriki, Elina Kolosovsky,Martine Paterne, Bettina Schilman, Avner Ayalon, et al. 2009. Climaticvariability during the last 90 ka of the southern and northern LevantineBasin as evident from marine records and speleothems. Quaternary ScienceReviews 28:2882–2896.

Alperson-Afil, Nira, and Erella Hovers. 2005. Differential use of space at theNeandertal site of Amud Cave, Israel. Eurasian Prehistory 3:3–22.

Andrefsky, William. 1994. The geological occurrence of lithic material andstone tool production strategies. Geoarchaeology 9:375–391.

Arensburg, Baruch, and Anna Belfer-Cohen. 1998. Neandertals and moderns:re-thinking the Levantine Middle Paleolithic hominids. In Neandertals andmodern humans in Western Asia. Takeru Akazawa, Kenichi Aoki, and OferBar-Yosef, eds. Pp. 311–322. New York: Plenum.

Ashkenazi, Hila. 2005. Standardization of blade production in the LevantineMP. MA dissertation, Hebrew University of Jerusalem.

Avery, D. Margaret, David Halkett, Jayson Orton, Teresa Steel, Madelon Tus-enius, and Richard Klein. 2008. The Ysterfontein 1 Middle Stone Age rockshelter and the evolution of coastal foraging. South African ArchaeologicalSociety Goodwin Series 10:66–89.

Bamforth, Douglas B. 1986. Technological efficiency and tool curation. Amer-ican Antiquity 51:38–50.

Bamforth, Douglas B., and Peter Bleed. 1997. Technology, flaked stone, andrisk. In Rediscovering Darwin: evolutionary theory and archaeological expla-nation. C. Michael Barton and Geoffrey A. Clark, eds. Pp. 109–139. Ar-chaeological Papers of the American Anthropological Association, no. 7.Arlington, VA: American Anthropological Association.

Barham, Lawrence S. 2002. Systemic pigment use in the Middle Pleistoceneof South-Central Africa. Current Anthropology 43:181–189.

Barkai, Ran, and Avi Gopher. 2009. Changing the face of the earth: humanbehavior at Sede Ilan, an extensive Lower-Middle Paleolithic quarry site inIsrael. In Lithic materials and Palaeolithic societies. Brian Adams and BrookeS. Blades, eds. Pp. 174–185. New York: Blackwell.

Bar-Matthews, Miryam, Avner Ayalon, Mabs Gilmour, Alan Matthews, andChris J. Hawkesworth. 2003. Sea-land oxygen isotopic relationships fromplanktonic froaminifera and speleothems in the eastern Mediterranean re-gion and their implication for paleorainfall during interglacial intervals.Geochimica et Cosmochimica Acta 67:3181–3199.

Bar-Yosef, Ofer. 1998. The chronology of the Middle Paleolithic of the Levant.In Neandertals and modern humans in western Asia. Takeru Akazawa, Ken-ichi Aoki, and Ofer Bar-Yosef, eds. Pp. 39–56. New York: Plenum.

———. 2000. The Middle and Early Upper Paleolithic in southwest Asia andneighboring regions. In The geography of Neandertals and modern humansin Europe and the greater Mediterranean. Ofer Bar-Yosef and David Pilbeam,eds. Pp. 107–156. Cambridge, MA: Peabody Museum of Archaeology andEthnography.

———. 2004. Eat what is there: hunting and gathering in the world of Ne-anderthals and their neighbours. International Journal of Osteoarchaeology14:333–342.

———. 2006. Between observations and models: an eclectic view of MiddlePaleolithic archaeology. In Transitions before the transition: evolution andstability in the Middle Paleolithic and Middle Stone Age. Erella Hovers andSteven L. Kuhn, eds. Pp. 305–325. New York: Springer.

Bar-Yosef, Ofer, and Richard H. Meadow. 1995. The origins of agriculture inthe Near East. In Last hunters, first farmers. T. Douglas Price and A.-B.Gebauer, eds. Pp. 39–94. Santa Fe, NM: School of American Research.

Bar-Yosef, Ofer, and Bernard Vandermeersch. 1993. Modern humans in theLevant. Scientific American 268:94–99.

Hovers and Belfer-Cohen Variability and Complexity S351

Bar-Yosef, Ofer, Bernard Vandermeersch, Baruch Arensburg, Anna Belfer-Cohen, Paul Goldberg, Henri Laville, Liliane Meignen, et al. 1992. Theexcavations in Kebara Cave, Mt. Carmel. Current Anthropology 33:497–550.

Bar-Yosef, Ofer, and Philip Van Peer. 2009. The chaıne operatoire approach inMiddle Paleolithic archaeology. Current Anthropology 50:103–131.

Bar-Yosef Mayer, Daniella E. 2005. The exploitation of shells as beads in thePalaeolithic and Neolithic of the Levant. Paleorient 31:176–185.

Bar-Yosef Mayer, Daniella E., Bernard Vandermeersch, and Ofer Bar-Yosef.2009. Shells and ochre in Middle Paleolithic Qafzeh Cave, Israel: indicationsfor modern behavior. Journal of Human Evolution 56:307–314.

Belfer-Cohen, Anna, and Ofer Bar-Yosef. 2000. Early sedentism in the NearEast: a bumpy ride to village life. In Life in Neolithic farming communities:social organization, identity, and differentiation. Ian Kuijt, ed. Pp. 19–38.New York: Kluwer Academic/Plenum.

Belfer-Cohen, Anna, and Naama Goren-Inbar. 1994. Cognition and com-munication in the Levantine Lower Palaeolithic. World Archaeology 26:144–157.

Belfer-Cohen, Anna, and A. Nigel Goring-Morris. 2002. Why microliths? mi-crolithization in the Levant. In Thinking small: global perspectives on mi-crolithization. Robert G. Elston and Steven L. Kuhn, eds. Pp. 57–68. Ar-chaeological Papers of the American Anthropological Association, no. 12.Santa Fe, NM: American Anthropological Association.

———. 2007. From the beginning: Levantine Upper Palaeolithic culturalchange and continuity. In Rethinking the human revolution: new behaviouraland biological perspectives on the origin and dispersal of modern humans.Paul Mellars, Katie Boyle, Ofer Bar-Yosef, and Chris Stringer, eds. Pp. 199–205. Cambridge: McDonald Institute for Archaeological Research.

———. 2009. The shift from the Middle Palaeolithic to the Upper Palaeolithic:Levantine perspectives. In The Mediterranean from 50,000 to 25,000 BP:turning points and new directions. Marta Camps and Carolyn C. Szmidt,eds. Pp. 89–100. Oxford: Oxbow.

Belfer-Cohen, Anna, and Erella Hovers. 1992. In the eye of the beholder:Mousterian and Natufian burials in the Levant. Current Anthropology 33:463–471.

———. 2010. Modernity, enhanced working memory, and the Middle toUpper Paleolithic record in the Levant. Current Anthropology 51(suppl. 1):S167–S175.

Belmaker, Miriam, and Erella Hovers. 2011. Ecological change and the ex-tinction of the Levantine Neanderthals: implications from a diachronicstudy of micromammals from Amud Cave, Israel. Quaternary Science Re-views 30:3196–3209.

Bettinger, Robert L., Robert Boyd, and Peter J. Richerson. 2009. Culturalinnovations and demographic change. Human Biology 81:211–235.

Beyries, Silvie. 1988. Functional variability of lithic sets in the Middle Pa-laeolithic. In Upper Pleistocene prehistory of western Euroasia. Harold L.Dibble and Anta Montet-White, eds. Pp. 213–225. University MuseumMonograph 54. Philadelphia: University Museum, University of Pennsyl-vania.

Binford, Lewis R. 1977. Forty-seven trips: a case study in the character ofarchaeological site formation process. In Stone tools as cultural markers. R.V. S. Wright, ed. Pp. 24–36. Canberra: Australian Institute of AboriginalStudies.

———. 1979. Organization and formation processes: looking at curated tech-nologies. Journal of Anthropological Research 35:255–273.

———. 1989. Isolating the transition to cultural adaptations: an organiza-tional approach. In The emergence of modern humans: biocultural adaptationsin the Late Pleistocene. Erik Trinkaus, ed. Pp. 18–41. Cambridge: CambridgeUniversity Press.

———. 2001. Constructing frames of reference: an analytical method for ar-chaeological theory building using hunter-gatherer and environmental datasets. Berkeley: University of California Press.

Binford, Sally R. 1968. Early Upper Pleistocene adaptations in the Levant.American Anthropologist 70:707–717.

Bisson, Michael, April Nowell, M. Poupart, C. Cordova, T. J. Pokines, and C.Ames. Forthcoming. WE-2: the Middle Paleolithic of the Wadi Enoqqiyarevisited. In Jordan’s prehistory: past and future revisited. Gary Rollefson andBill Finlayson, eds. Amman: Department of Antiquities of the HashemiteKingdom of Jordan.

Bleed, Peter. 1986. The optimal design of hunting weapons: maintainabilityor reliability. American Antiquity 51:737–747.

Bocquet-Appel, Jean-Pierre, and Alan Tuffreau. 2009. Technological responses

of Neanderthals to macroclimatic variations (240,000–40,000 BP). HumanBiology 81:287–307.

Boeda, Eric. 1993. Le debitage discoıde et le debitage Levallois recurrent cen-tripete. Bulletin de la Societe Prehistorique Francaise 90:392–404.

Boeda, Eric, Stephanie Bonilauri, Jacque Connan, Dan Jarvie, Norbert Mercier,Mark Tobey, Helene Valladas, Heba al Sakhel, and Sultan Muhesen. 2008.Middle Paleolithic bitumen use at Umm el Tlel around 70,000 BP. Antiquity82:853–861.

Boeda, Eric, Jacques Connan, and Sultan Muhesen. 1998. Bitumen as haftingmaterial on Middle Paleolithic artifacts from the el-Kowm Basin, Syria. InNeandertals and modern humans in western Asia. Takeru Akazawa, KenichAoki, and Ofer Bar-Yosef, eds. Pp. 181–204. New York: Plenum.

Boeda, Eric, Jean-Michel Geneste, Christophe Griggo, Norbert Mercier, SultanMuhesen, J. L. Reyss, A. Taha, and Helene Valladas. 1999. A Levallois pointembedded in the vertebra of a wild ass (Equus africanus): hafting, projectilesand Mousterian hunting weapons. Antiquity 73:394–402.

Boeda, Eric, Jean-Michel Geneste, and Liliane Meignen. 1990. Identificationde chaıne operatoires lithiques du paleolithique ancient et moyen. Paleo 2:43–80.

Boeda, Eric, Christophe Griggo, and Sandrine Noel-Soriano. 2001. Differentsmodes d’occupation du site d’Umm el Tlel au cours du paleolithique moyen(El Kowm, Syrie centrale). Paleorient 27:13–28.

Bonilauri, Stephanie, Eric Boeda, Christophe Griggo, Heba Al-Sakhel, andSultan Muhesen. 2007. Un eclat de silex mousterien coince dans un bassind’autruche (Struthio camelus) a Umm el Tlel (Syrie centrale). Paleorient 33:39–46.

Bousman, C. Britt. 2005. Coping with risk: Later Stone Age technologicalstrategies at Blydefontein Rock Shelter, South Africa. Journal of Anthro-pological Archaeology 24:193–226.

Boyd, Robert, and Peter J. Richerson. 1985. Culture and the evolutionaryprocess. Chicago: University of Chicago Press.

———. 1996. Why culture is common, but cultural evolution is rare. Pro-ceedings of the British Academy 88:77–93.

———. 2009. Voting with your feet: payoff biased migration and theevolution of group beneficial behavior. Journal of Theoretical Biology257:331–339.

Brooks, Alison, John E. Yellen, Lisa Nevell, and Gideon Hartman. 2006. Pro-jectile technologies of the African MSA: implications for modern humansorigins. In Transitions before the transition: evolution and stability in theMiddle Paleolithic and Middle Stone Age. Erella Hovers and Steven L. Kuhn,eds. Pp. 233–255. New York: Springer.

Brown, Kyle S., Curtis W. Marean, Andy I. R. Herries, Zenobia Jacobs, ChantalTribolo, David Braun, David L. Roberts, Michael C. Meyer, and JocelynBernatchez. 2009. Fire as an engineering tool of early modern humans.Science 325:859–862.

Cain, Chester R. 2006. Implications of the marked artifacts of the MiddleStone Age in Africa. Current Anthropology 47:675–681.

Carciumaru, Marin, Rodica-Mariana Ion, Elenna-Christina Nitu, and RaduStefanescu. 2012. New evidence of adhesive as hafting material on Middleand Upper Palaeolithic artefacts from Gura Cheii-Rasnov Cave (Romania).Journal of Archaeological Science 39:1942–1950.

Churchill, Steven E. 2006. Bioenergetic perspectives on Neanderthal ther-moregulatory and activity budgets. In Neanderthals revisited: new ap-proaches and perspectives. Katrina Harvati and Terry Harrison, eds. Pp.113–133. Vertebrate Paleobiology and Paloeanthroplogy Series. Dor-drecht: Springer.

Clark, Jamie L. 2011. The evolution of human culture during the Later Pleis-tocene: using fauna to test models on the emergence and nature of“modern” human behavior. Journal of Anthropological Archaeology 30:273–291.

Collard, Mark, Briggs Buchanan, Marcus J. Hamilton, and Michael J. O’Brien.2010. Spatiotemporal dynamics of the Clovis-Folsom transition. Journal ofArchaeological Science 37:2513–2519.

Conard, Nicholas, M. Masri, K. Bretzkei, H. A. Naipieralai, B. Weltel, andAndrew W. Kandel. 2010. The 2008 excavation at the Middle Paleolithicsite of Wadi Mushkuna Rockshelter, Damascus Province, Syria. In Docu-menting the Annual Excavation Reports concerning the archaeological activitiesin Syria. Special issue, Chronique archeologique en Syrie 6:13–22.

Copeland, Lorraine. 1975. The Middle and Upper Palaeolithic in Lebanonand Syria in the light of recent research. In Problems in prehistory: NorthAfrica and the Levant. Fred Wendorf and Angela Close, eds. Pp. 317–350.Dallas, TX: Southern Methodist University Press.

S352 Current Anthropology Volume 54, Supplement 8, December 2013

———. 1983. Levallois/non-Levallois determinations in the Early LevantineMousterian: problems and questions for 1983. Paleorient 9:15–27.

Cordain, Loren, Janette B. Miller, S. Boyd Eaton, Neil Mann, Susanne H.Holt, and John D. Speth. 2000. Plant-animal subsistence ratios and mac-ronutrient energy estimations in worldwide hunter-gatherer diets. AmericanJournal of Clinical Nutrition 71:682–692.

Corning, Peter A. 2002. The re-emergence of “emergence”: a venerable conceptin search of a theory. Complexity 7:18–30.

Cortes-Sanchez, Miguel, Arturo Morales-Muniz, Marıa D. Simon-Vallejo,Marıa C. Lozano-Francisco, Jose L. Vera-Pelaez, Clive Finlayson, JoaquınRodrıgues-Vidal, et al. 2011. Earliest known use of marine resources byNeanderthals. PLoS ONE 6:e24026.

David, Nick, and Carole Kramer. 2001. Ethnoarchaeology in action. Cambridge:Cambridge University Press.

Davis, Simon J. M., Rivka Rabinovich, and Naama Goren-Inbar. 1988. Qua-ternary extinctions and population increase in western Asia: the animalremains from Biq’at Quneitra. Paleorient 14:95–105.

DeBono, Hila, and Naama Goren-Inbar. 2001. Note on a link between Acheu-lian handaxes and the Levallois method. Journal of the Israel PrehistoricSociety 31:9–24.

Delage, Christophe, ed. 2007. Chert availability and prehistoric exploitation inthe Near East. BAR International Series 1615. Oxford: Archaeopress.

de la Torre, Ignacio, Jorge Martınez-Moreno, and Rafael Mora. 2013. Changeand stasis in the Iberian Middle Paleolithic: considerations on the signifi-cance of Mousterian technological variability. Current Anthropology54(suppl. 8):S320–S336.

Demidenko, Y. E., and V. I. Usik. 2004. Into the mind of the maker: refittingstudy and technological reconstructions. In Neandertals in the Levant: be-havioral organization and the beginnings of human modernity. D. O. Henry,ed. Pp. 107–155. London: Continuum.

d’Errico, Francesco, Renata Garcıa Moreno, and Riaan F. Rifkin. 2011. Tech-nological, elemental and colorimetric analysis of an engraved ochre frag-ment from the Middle Stone Age levels of Klasies River Cave 1, SouthAfrica. Journal of Archaeological Science 39:942–952.

Dobzhansky, Theodosius G. 1951. Genetics and human origins. New York:Columbia University Press.

Drake, Nick A., Roger M. Blench, Simon J. Armitage, Charlie S. Bristow, andKevin H. White. 2011. Ancient watercourses and biogeography of the Saharaexplain the peopling of the desert. Proceedings of the National Academy ofSciences of the USA 108:458–462.

Druck, Dotan. 2004. The exploitation of flint resources throughout the se-quence of prehistoric cultures in Nahal Mea’rot, Mt. Carmel. MA thesis,University of Haifa.

Dyson-Hudson, Rada, and Eric Alden Smith. 1978. Human territoriality: anecological reassessment. American Anthropologist 80:21–41.

Ehrlich, Paul, and Marcus W. Feldman. 2003. Genes and cultures: what createsour behavioral phenome? Current Anthropology 44:87–107.

Ekshtain, Ravid, Moshe Inbar, Ianir Milevski, Micka Ullman, and Omry Bar-izlai. 2012. Giv’at Rabi East: a new Middle Paleolithic knapping site in theLower Galilee, Israel. Paleorient 37:107–122.

Ekshtain, Ravid, Ariel Malinsky-Buller, Shimon Ilani, Irena Segal, andErella Hovers. 2013. Raw material exploitation around the Middle Pa-leolithic site of ’Ein Qashish. Quaternary International, forthcoming,doi:10.1016/j.quaint.3013.07.025.

El Zaatari, Sireen, Frederick E. Grine, Peter S. Ungar, and Jean-Jaque Hublin.2011. Ecogeographic variation in Neandertal dietary habits: evidence fromocclusal molar microwear texture analysis. Journal of Human Evolution 61:411–424.

Enzel, Yehouda, Rivka Amit, Uri Dayan, Onn Crouvi, Ron Kahana, BaruchZiv, and David Sharon. 2008. The climatic and physiographic controls ofthe eastern Mediterranean over the late Pleistocene climates in the southernLevant and its neighboring deserts. Global and Planetary Change 60:165–192.

Eren, Metin I., Manuel Domınguez-Rodrigo, Steven L. Kuhn, Dan S. Adler,Ian Le, and Ofer Bar-Yosef. 2005. Defining and measuring reduction inunifacial stone tools. Journal of Archaeological Science 32:1190–1201.

Eren, Metin I., and Stephen J. Lycett. 2012. Why Levallois? a morphometriccomparison of experimental “preferential” Levallois flakes versus debitageflakes. PLoS ONE 7:e29273.

Eriksson, Anders, and Andrea Manica. 2012. Effect of ancient populationstructure on the degree of polymorphism shared between modern humanpopulations and ancient hominins. Proceedings of the National Academy ofSciences of the USA 109:13956–13960.

Finlayson, Geraldine, R. Barton, and Chris Stringer. 2001. The Gibraltar Ne-anderthals and their extinction. In Les premiers hommes modernes de lapeninsule Iberique. Joao Zilhao, T. Aubry, and A. Carvalho, eds. Pp. 117–122. Trabalhos de Arqueologia 17. Lisbon: Instituto Portugues de Arqueo-logia.

Friedman, Erich, Naama Goren-Inbar, Amnon Rosenfeld, Ofer Marder, andFelix Burian. 1994–1995. Hafting during Mousterian times: further evi-dence. Journal of the Israel Prehistoric Society 26:8–31.

Frumkin, Amos, Ofer Bar-Yosef, and Henry P. Schwarcz. 2011. Possible pa-leohydrologic and paleoclimatic effects on hominin migration and occu-pation of the Levantine Middle Paleolithic. Journal of Human Evolution 60:437–451.

Garrod, Dorothy A. E., and Dorothea M. A. Bate. 1937. The Stone Age ofMount Carmel, vol. 1. Oxford: Clarendon.

Geneste, Jean-Michel. 1985. Analyse lithique d’industries mousteriennes duPerigord: une approche technologique des comportements des groupeshumains au paleolithique moyen. These Sc., Universite de Bordeaux 1.

Gilead, Isaac. 1980. A Middle Palaeolithic open-air site near Tell Far’ah, west-ern Negev: preliminary report. Israel Exploration Journal 30:52–62.

———. 1988. Le site mousterien de Fara II (Neguev septentrional, Israel) etle remontage de son industrie. L’Anthropologie 92:797–807.

Gilead, Isaac, and Caroline Grigson. 1984. Far’ah II: a Middle Palaeolithicopen-air site in the Northern Negev, Israel. Proceedings of the PrehistoricSociety 50:71–91.

Godfrey-Smith, Dorothy I., and Shimon Ilani. 2004. Past thermal history ofgoethite and hematite fragments from Qafzeh Cave deduced from thermalactivation characteristics of the 110 centigrade TL peak of enclosed quartzgrains. Revue d’Archeometrie 28:185–190.

Goldberg, Paul, and Ofer Bar-Yosef. 1998. Site formation processes at Keabaraand Hayonim caves and their significance in Levantine prehistoric caves.In Neandertals and modern humans in western Asia. Takeru Akazawa, Ken-ichi Aoki, and Ofer Bar-Yosef, eds. Pp. 107–126. New York: Plenum.

Gopher, Avi, Ran Barkai, Ron Shimelmitz, Hamoudi Khalaily, Christina Le-morini, Israel Hershkovitz, and Mary C. Stiner. 2005. Qesem Cave: anAmudian site in central Israel. Journal of the Israel Prehistoric Society 35:69–92.

Gordon, David. 1993. Mousterian tool selection, reduction, and discard atGhar, Israel. Journal of Field Archaeology 20:205–218.

Goren-Inbar, Naama. 1988. Too small to be true? reevaluation of coreson flakes in Levantine Mousterian assemblages. Lithic Technology 17:37–44.

———. 1990. The lithic assemblages. In Quneitra: a Mousterian site on theGolan Heights. Qedem Naama Goren-Inbar, ed. Pp. 61–167. Jerusalem:Institute of Archaeology.

———. 1994. The Lower Palaeolithic of Israel. In The archaeology of societyin the Holy Land. Thomas. E. Levy, ed. Pp. 93–109. London: LeicesterUniversity Press.

———. 2011. Behavioral and cultural origins of Neanderthals: a Levantineperspective. Continuity and discontinuity in the peopling of Europe: one hun-dred fifty years of Neanderthal study. Silvana Condemi and Gerd-ChristianWeninger, eds. Pp. 89–100. Dordrecht: Springer.

Goren-Inbar, Naama, and Anna Belfer-Cohen. 1998. The technological abil-ities of the Levantine Mousterians: cultural and mental capacities. In Nean-dertals and modern humans in western Asia. Takeru Akazawa, Kenichi Aoki,and Ofer Bar-Yosef, eds. Pp. 205–222. New York: Plenum.

Goring-Morris, Adrian N., and Anna Belfer-Cohen, eds. 2003. More than meetsthe eye: studies on Upper Palaeolithic diversity in the Near East. Oxford:Oxbow.

Goring-Morris, A. Nigel, Erella Hovers, and Anna Belfer-Cohen. 2009. Thedynamics of Pleistocene and early Holocene settlement patterns and humanadaptations in the Levant: an overview. In Transitions in prehistory: essaysin honor of Ofer Bar-Yosef. John J. Shea and Daniel E. Lieberman, eds. Pp.185–252. Oxford: Oxbow.

Green, Richard E., Johannes Krause, Adrian W. Briggs, Tolmislav Maricic,Udo Stenzel, Martin Kircher, Nick Patterson, et al. 2010. A draft sequenceof the Neandertal genome. Science 328:710–722.

Greenfield, Patricia M. 1991. Language, tools, and brain: the ontogeny andphylogeny of hierarchically organized sequential behavior. Behavioral andBrain Sciences 14:531–595.

Griggo, Christophe. 1998. La faune mousterienne du site d’umm el Tlel (Syrie):etude preliminarie: Fouilles 1991–1993. Cahiers el l’Euphrates 8:11–25.

Grunberg, J.-M. 2002. Middle Palaeolithic birch-bark pitch. Antiquity 76:15–16.

Hovers and Belfer-Cohen Variability and Complexity S353

Hallin, Kristin A., Margaret J. Schoeninger, and Henry P. Schwarcz. 2012.Paleoclimate during Neandertal and anatomically modern human occu-pation at Amud and Qafzeh, Israel: the stable isotope data. Journal of HumanEvolution 62:59–73.

Hauck, T. C. 2011. Mousterian technology and settlement dynamics in thesite of Hummal (Syria). Journal of Human Evolution 61:519–537.

Hauck, Thomas C., Jacques Connan, Armelle Charrie-Duhaut, Jean-Marie LeTensorerd, and Heba al Sakhel. 2013. Molecular evidence of bitumen inthe Mousterian lithic assemblage of Hummal (Central Syria). Journal ofArchaeological Science 40(8):3252–3262. http://dx.doi.org/10.1016/j.jas.2013.03.022.

Hayden, Brian. 1981. Subsistence and ecological adaptations of modernhunter-gatherers. In Omnivorous primates: gathering and hunting in humanevolution. Robert S. O. Harding and Geza Teleki, eds. Pp. 344–421. NewYork: Columbia University Press.

Henrich, Joseph. 2004. Demography and cultural evolution: how adaptivecultural processes can produce maladaptive losses: the Tasmania case. Amer-ican Antiquity 69:197–214.

———. 2010. The evolution of innovation-enhancing institutions. Innovationsin cultural systems. Michael J. O’Brien and Stephen J. Shennan, eds. Pp.99–120. Cambridge MA, London England. Cambridge: MIT Press.

Henrich, Joseph, and Robert Boyd. 1998. The evolution of conformist trans-mission and the emergence of between-group differences. Evolution andHuman Behavior 19:215–241.

Henry, Amanda G., Alison S. Brooks, and Dolores R. Piperno. 2011. Micro-fossils in calculus demonstrate consumption of plants and cooked foods inNeanderthal diets (Shanidar III, Iraq; Spy I and II, Belgium). Proceedingsof the National Academy of Sciences of the USA, doi:10.1073/pnas.1016868108.

Henry, Donald O. 1995a. The influence of mobility levels on Levallois pointproduction, Late Levantine Mousterian, southern Jordan. In The definitionand interpretation of Levallois technology. Harold L. Dibble and Ofer Bar-Yosef, eds. Pp. 185–200. Madison, WI: Prehistory Press.

———. 1995b. The Middle Paleolithic sites. In Prehistoric cultural ecology anddiversity: insights from southern Jordan. Donald O. Henry, ed. Pp. 49–84.New York: Plenum.

Henry, D. O., H. J. Hietala, A. M. Rosen, Y. E. Demidenko, V. I. Usik,and T. L. Armagan. 2004. Human behavioral organization in the MiddlePaleolithic: were Neanderthals different? American Anthropologist 106:17–31.

Hietala, H. J. 2003. Site structure and material culture in space on the TorFaraj living floors. In Neandertals in the Levant: behavioral organization andthe beginnings of human modernity. Donald O. Henry, ed. Pp. 198–233.London: Continuum.

Hockett, Bryan. 2012. The consequences of Middle Paleolithic diets on preg-nant Neanderthal women. Quaternary International 264:78–82.

Hovers, Erella. 1997. Variability of lithic assemblages and settlement patternsin the Levantine Middle Paleolithic: implications for the development ofhuman behavior. PhD dissertation, Hebrew University of Jerusalem.

———. 1998a. Comment on “Neandertal and early modern human behav-ioral variability: a regional-scale approach to lithic evidence for hunting inthe Levantine Mousterian” by J. J. Shea. Current Anthropology 39(suppl.):S65–S66.

———. 1998b. The lithic assemblages of Amud Cave: implications for theend of the Mousterian in the Levant. In Neandertals and modern humansin western Asia. Takeru Akazawa, Kenichi Aoki, and Ofer Bar-Yosef, eds.Pp. 143–163. New York: Plenum.

———. 2001. Territorial behavior in the Middle Paleolithic of the SouthernLevant. In Settlement dynamics of the Middle Paleolithic and Middle StoneAge. Nicholas Conard, ed. Pp. 123–152. Tubingen Publications in Prehis-tory. Tubingen: Kerns.

———. 2006. Neandertals and modern humans in the Middle Paleolithic ofthe Levant: what kind of interaction? In When Neandertals and modernsmet. Nicholas Conard, ed. Pp. 65–86. Tubingen: Kerns.

———. 2007. The many faces of cores-on-flakes: a perspective from theLevantine Mousterian. In Cores or tools? alternative approaches to stone toolanalysis. Shannon P. McPherron, ed. Pp. 42–74. Cambridge: CambridgeScholars Press.

———. 2009. The lithic assemblages of Qafzeh Cave. New York: Oxford Uni-versity Press.

Hovers, Erella, and Anna Belfer-Cohen. 2006. “Now you see it, now youdon’t”: modern human behavior in the Middle Paleolithic. In Transitionsbefore the transition: evolution and stability in the Middle Paleolithic and

Middle Stone Age. Erella Hovers and Steven L. Kuhn, eds. Pp. 295–304.New York: Springer.

———. 2013. Insights into early mortuary practices of Homo. In The Oxfordhandbook of the archaeology of death and burial. Sarah Tarlow and LivNilsson-Stutz, eds. Oxford: Oxford University Press.

Hovers, Erella, Shimon Ilani, Ofer Bar-Yosef, and Bernard Vandermeersch.2003. An early case of color symbolism: ochre use by early modern humansin Qafzeh Cave. Current Anthropology 44:491–522.

Hovers, Erella, William H. Kimbel, and Yoel Rak. 2000. Amud 7: still, a burial:response to Gargett. Journal of Human Evolution 39:253–260.

Hovers, Erella, Ariel Malinsky-Buller, Mae Goder-Goldberger, and Ravid Eksh-tain. 2011. Capturing a moment: identifying short-lived activity locationsin Amud Cave, Israel. In The Lower and Middle Palaeolithic in the MiddleEast and neighbouring regions: Basel symposium, mai 8–10 2008. Jean-MarieLe Tensorer, Rheto Jagher, and Marcel Otte, eds. Pp. 101–114. Etudes etrecherches archeologiques de l’Universite de Liege 126. Liege: University ofLiege Press.

Hovers, Erella, Yoel Rak, Ron Lavi, and William H. Kimbel. 1995. Hominidremains from Amud Cave in the context of the Levantine Middle Paleolithic.Paleorient 21:47–61.

Hovers, Erella, and Adi Raveh. 2000. The use of a multivariate display tech-nique in the analysis of inter-assemblage lithic variability: a case study fromQafzeh Cave, Israel. Journal of Archaeological Science 27:1023–1038.

Hovers, Erella, Bernard Vandermeersch, and Ofer Bar-Yosef. 1997. A MiddlePalaeolithic engraved artefact from Qafzeh Cave, Israel. Rock Art Research14:79–87.

Hutchings, W. Karl. 2011. Measuring use-related fracture velocity in lithicarmatures to identify spears, javelins, darts, and arrows. Journal of Archae-ological Science 38:1737–1746.

Inizan, Marie-Louise, Helene Roche, and Jacques Tixier. 1992. Technologyof knapping stone: prehistoire de la Pierre Taillee 3. Meudon, France:CREP.

Jelinek, Arthur. 1982. The Middle Paleolithic in the southern Levant, withcomments on the appearance of modern Homo sapiens. In The transitionfrom Lower to Middle Palaeolithic and the origin of modern man. AvrahamRonen, ed. Pp. 57–104. BAR International Series 151. Oxford: BAR.

Jelinek, Arthur J., William Farrand, George Haas, Aharon Horowitz, and PaulGoldberg. 1973. New excavations at the Tabun Cave, Mount Carmel, Israel,1967–1972: a preliminary report. Paleorient 1:151–183.

Jochim, Michael B. 1983. Palaeolithic cave art in ecological perspective. Inhunter-gatherer economy in prehistory: a European perspective. Geoff Bailey,ed. Pp. 212–219. Cambridge: Cambridge University Press.

Johnson, Gregory A. 1978. Information sources and the development ofdecision-making organization. In Social archaeology: beyond subsistence anddating. Charles L. Redman, M. J. Berman, E. V. Curtin, W. J. LonghorneJr., N. M. Versaggi, and C. Wansen, eds. Pp. 87–112. New York: AcademicPress.

Jordan, P. 2007. Continuity and change in different domains of culture: anemerging approach to understanding diversity in technological traditions.In The model-based archaeology of socionatural ecosystems. T. A. Kohler andS. E. van der Leeuw, eds. Pp. 13–39. Santa Fe, NM: School for AdvancedResearch.

Kandler, Anne, and Kevin N. Laland. 2009. An investigation of the relationshipbetween innovation and cultural diversity. Theoretical Population Biology76:59–67.

Karkanas, Panagiotis, Ruth Shahack-Gross, Avner Ayalon, Miryam Bar-Matthews, Ran Barkai, Amos Frumkin, Avi Gopher, and Mary C. Stiner.2007. Evidence for habitual use of fire at the end of the Lower Paleolithic:site-formation processes at Qesem Cave, Israel. Journal of Human Evolution53:197–212.

Kelly, Robert L. 1983. Hunter-gatherer mobility strategies. Journal of Anthro-pological Research 39:277–306.

———. 1995. The foraging spectrum: diversity in hunter-gatherer lifeways.Washington, DC: Smithsonian Institution.

Klein, Richard G., and K. Cruz-Uribe. 1996. Exploitation of large bovids andseals at Middle and Later Stone Age sites in South Africa. Journal of HumanEvolution 31:315–334.

Kuhn, Steven L. 1993. Mousterian technology as adaptive response: a casestudy. In Hunting and animal exploitation in the late Palaeolithic andMesolithic of Europe. Gail H. Peterkin, Harvey M. Bricker, and PaulMellars, eds. Pp. 20–25. Washington, DC: American AnthropologicalAssociation.

S354 Current Anthropology Volume 54, Supplement 8, December 2013

———. 1994. A formal approach to the design and assembly of transportedtoolkits. American Antiquity 59:426–442.

———. 1995. Mousterian lithic technology: an ecological perspective. Princeton,NJ: Princeton University Press.

———. 2013. Roots of the Middle Paleolithic in Eurasia. Current Anthropology54(suppl. 8):S255–S268.

Kuhn, Steven L., and Mary C. Stiner. 1998. Middle Paleolithic “creativity”:reflections on an oxymoron? In Creativity in human evolution and prehistory.Steven Mithen, ed. Pp. 143–164. London: Routledge.

———. 2006. What’s a mother to do? the division of labor among Ne-anderthals and modern humans in Eurasia. Current Anthropology 47:953–980.

Kuhn, Steven L., Mary C. Stiner, David S. Reese, and Erskin Gulec. 2001.Ornaments of the earliest Upper Paleolithic: new insights from theLevant. Proceedings of the National Academy of Sciences of the USA 98:7641–7646.

Lahr, Martha M., and Robert Foley. 1998. Towards a theory of modern humanorigins: geography, demography, and diversity in recent human evolution.Yearbook of Physical Anthropology 41:137–176.

Laland, Kevin, and Michael O’Brien. 2010. Niche construction theory andarchaeology. Journal of Archaeological Method and Theory 17:303–322.

Langbroek, Marco. 2012. Trees and ladders: a critique of the theory of humancognitive and behavioural evolution in Palaeolithic archaeology. QuaternaryInternational 270:4–14.

Lev, Ephraim, Mordechai E. Kislev, and Ofer Bar-Yosef. 2005. Mousterianvegetal food in Kebara Cave, Mt. Carmel. Journal of Archaeological Science32:475–484.

Lieberman, Daniel E., and Ofer Bar-Yosef. 2005. Apples and oranges: mor-phological versus behavioral transitions in the Pleistocene. In Interpretingthe past: essays in human, primate, and mammal evolution in honor of DavidPilbeam. Daniel E. Lieberman, Richard J. Smith, and Jay Kelley, eds. Pp.275–296. Boston: Bril Academic.

Lombard, Marlize. 2007. The gripping nature of ochre: the association ofochre with Howiesons Poort adhesives and Later Stone Age mastics fromSouth Africa. Journal of Human Evolution 53:406–419.

MacDonald, Katharine, Will Roebroeks, and Alexander Verpoort. 2009. Anenergetics perspective on the Neandertal record. In The evolution of hominindiets: integrating approaches to the study of palaeolithic subsistence. Jean-Jacque Hublin and Michael P. Richards, eds. Pp. 211–220. Vertebrate Pale-obiology and Paleoanthropology. Dordrecht: Springer.

Mackay, Alex. 2011. Potentially stylistic differences between backed artefactsfrom two nearby sites occupied 60,000 years before present in South Africa.Journal of Anthropological Archaeology 30:235–245.

Madella, Marco, Martin K. Jones, Paul Goldberg, Yuval Goren, and ErellaHovers. 2002. The exploitation of plant resources by Neanderthals in AmudCave (Israel): the evidence from phytolith studies. Journal of ArchaeologicalScience 29:703–719.

Malinsky-Buller, Ariel, Leore Grosman, and Ofer Marder. 2011. A case oftechno-typological lithic variability and continuity in the late Lower Pa-laeolithic. Before Farming 2011/1(art. 3):1–32.

March, Ramiro, Alexandre Lucquin, Delphine Joly, Juan Ferreri, and Mo-hamad Muhieddine. 2012. Processes of formation and alteration of ar-chaeological fire structures: complexity viewed in the light of experimentalapproaches. Journal of Archaeological Method and Theory, doi:10.1007/s10816-012-9134-7.

Marder, Ofer, Reuven Yeshurun, Ronit Lupu, Guy Bar-Oz, Miriam Bel-maker, Naomi Porat, Hagai Ron, and Amos Frumkin. 2011. Mammalremains at Rantis Cave, Israel, and Middle-Late Pleistocene humansubsistence and ecology in the southern Levant. Journal of QuaternaryScience 26:769–780.

Marean, Curtis W., Miryam Bar-Matthews, J. Bernatchez, Erich Fisher, PaulGoldberg, Andy I. R. Herries, Zenobia Jacobs, et al. 2007. Early human useof marine resources and pigment in South Africa during the Middle Pleis-tocene. Nature 449:905–908.

Marks, Anthony E. 1988. The curation of stone tools during the Upper Pleis-tocene: a view from the central Negev, Israel. In Upper Pleistocene prehistoryof western Eurasia. Harold L. Dibble and Anta Montet-White, eds. Pp. 275–285. Philadelphia: University Museum, University of Pennsylvania.

———. 2008. Into Arabia, perhaps, but if so, from where? Proceedings of theSeminar for Arabian Studies 38:15–24.

Marks, Anthony E., and David A. Freidel. 1977. Prehistoric settlement patternsin the Avdat/Aqev area. In The Avdat/Aqev area, part 2, vol. 2 of Prehistory

and paleoenvironments in the central Negev, Israel. Anthony E. Marks, ed.Pp. 131–158. Dallas, TX: Southern Methodist University Press.

Marshack, Alexander. 1981. On Paleolithic ochre and the early uses of colorand symbol. Current Anthropology 22:188–191.

———. 1996. A Middle Paleolithic symbolic composition from the GolanHeights: the earliest known depictive image. Current Anthropology 37:357–365.

Marx, Emmanuel. 2004. Bands and other corporate societies in Acheulianculture. In Human paleoecology in the Levantine corridor. Naama Goren-Inbar and John D. Speth, eds. Pp. 89–104. Oxford: Oxbow.

Matsutani, A. 1973. Microscopic study on the “amorphous silica” in sedimentsfrom the Douara Cave. In The Paleolithic site of Doura Cave in Syria, partI. Hisashi Suzuki and Fuyuji Takai, eds. Pp. 17–21. Tokyo: University ofTokyo Press.

———. 1987. Plant remains from the 1984 excavations at Douara Cave. InPaleolithic site of Douara Cave and paleogeography of Palmyra Basin, partIV: 1984 excavations. Takeru Akazawa and Y. Sakaguchi, eds. Pp. 117–122.Tokyo: University of Tokyo Press.

Mazza, Paul P. A., Fabio Martini, Benedeto Sala, Maurizio Magi, Maria P.Colombini, Gianna Giachi, Francesco Landucci, Christina Lemorini, Fran-cesco Modugno, and Erika Ribechini. 2006. A new Palaeolithic discovery:tar-hafted stone tools in a European Mid-Pleistocene bone-bearing bed.Journal of Archaeological Science 33:1310–1318.

McBrearty, Sally, and Christian A. Tryon. 2006. From Acheulean to MiddleStone Age in the Kapthurin Formation, Kenya. In Transitions before thetransition: evolution and stability in the Middle Paleolithic and Middle StoneAge. Erella Hovers and Steven L. Kuhn, eds. Pp. 257–277. New York:Springer.

Meignen, Liliane. 1995. Levallois lithic production systems in the MiddlePalaeolithic of the Near East: the case of the unidirectional method. In Thedefinition and interpretation of Levallois technology. Harold L. Dibble andOfer Bar-Yosef, eds. Pp. 361–380. Madison, WI: Prehistory Press.

———. 1998. A preliminary report on Hayonim cave lithic assemblages inthe context of the Near Eastern Middle Palaeolithic. In Neandertals andmodern humans in western Asia. Takeru Akazawa, Kenichi Aoki, and OferBar-Yosef, eds. Pp. 165–180. New York: Plenum.

———. 2011. The contribution of Hayonim Cave assemblage to the under-standing of the so-called Early Levantine Mousterian. In The Lower andMiddle Palaeolithic in the Middle East and neighbouring regions: Basel sym-posium, mai 8–10 2008. Jean-Marie Le Tensorer, Rheto Jagher, and MarcelOtte, eds. Pp. 85–100. Etudes et recherches archeologiques de l’Universitede Liege 126. Liege: University of Liege Press.

Meignen, Liliane, and Ofer Bar-Yosef. 1992. Middle Paleolithic variability inKebara Cave, Israel. In The evolution and dispersal of modern humans inAsia. Takeru Akazawa, Kenichi Aoki, and Tatsuku Kimura, eds. Pp. 129–148. Tokyo: Hakusen-Sha.

———. 2002. The lithic industries of the Middle and Upper Paleolithic ofthe Levant: continuity or break? Archaeology, Ethnology and Anthropologyof Eurasia 3:12–21.

Meignen, Liliane, Ofer Bar-Yosef, John D. Speth, and Mary C. Stiner. 2006.Middle Paleolithic settlement patterns in the Levant. In Transitions beforethe transition: evolution and stability in the Middle Paleolithic and MiddleStone Age. Erella Hovers and Steven L. Kuhn, eds. Pp. 149–169. New York:Springer.

Meignen, Liliane, Paul Goldberg, and Ofer Bar-Yosef. 2007. The hearths atKebara Cave and their role in site formation processes. In Kebara Cave,Mt. Carmel, Israel: the Middle and Upper Paleolithic archaeology, part I. OferBar-Yosef and Liliane Meignen, eds. Pp. 91–122. Cambridge, MA: AmericanSchool of Prehistoric Research, Peabody Museum of Anthropology andEthnology, Harvard University.

Mesoudi, Alex. 2011a. Cultural evolution: how Darwinian theory can explainhuman culture and synthesize the social sciences. Chicago: University of Chi-cago Press.

———. 2011b. Variable cultural acquisition costs constrain cumulative cul-tural evolution. PLoS ONE 6:e18239.

Mesoudi, Alex, and Stephen J. Lycett. 2009. Random copying, frequency-dependent copying and culture change. Evolution and Human Behavior 30:41–48.

Metcalfe, D., and K. R. Barlow. 1992. A model for exploring the optimal trade-off between field processing and transport. American Anthropologist 94:340–356.

Moncel, Marie-Helen, Anne-Marie Moigne, Youssef Sam, and Jean Combier.2011. The emergence of Neanderthal technical behavior: new evidence from

Hovers and Belfer-Cohen Variability and Complexity S355

Orgnac 3 (Level 1, MIS 8), southeastern France. Current Anthropology 52:37–75.

Monigal, Katherine. 2002. The Levantine leptolithic: blade production fromthe Lower Paleolithic to the dawn of the Upper Paleolithic. PhD dissertation,Southern Methodist University, Dallas, TX.

Moore, James A. 1981. The effects of information networks in hunter-gatherersocieties. In Hunter-gatherer foraging strategies: ethnographic and archaeo-logical case studies. Bruce Winterhalder and Eric A. Smith, eds. Pp. 194–217. Prehistoric Archeology and Ecology. Chicago: University of ChicagoPress.

Mourre, Vincent, Paola Villa, and Christopher S. Henshilwood. 2010. Earlyuse of pressure flaking on lithic artifacts at Blombos Cave, South Africa.Science 330:659–662.

Munday, Frederick. C. 1976. Intersite variability in the Mousterian of theCentral Negev. In The Avdat/Aqev area, part 1, vol. 1 of Prehistory andpaleoenvironments in the central Negev, Israel. Anthony E. Marks, ed. Pp.113–140. Dallas, TX: Southern Methodist University Press.

Neiman, Frederick D. 1995. Stylistic variation in evolutionary perspective:inferences from decorative diversity and interassemblage distance in Illinoiswoodland ceramic assemblages. American Antiquity 60:7–36.

Nelson, Margaret C. 1991. The study of technological organization. In Ar-chaeological method and theory, vol. 3. Michael B. Schiffer, ed. Pp. 57–100.Tucson: University of Arizona Press.

Nilsson, Liv. 1998. Dynamic cadavers: a “field-anthropological” analysis of theSkateholm II burials. Lund Archaeological Review 4:5–17.

Nishiaki, Yoshihiro. 1989. Early blade industries in the Levant: the placementof Douara IV industry in the context of the Levantine Early Middle Pa-leolithic. Paleorient 15:215–229.

Nishiaki, Yoshihiro, and Lorraine Copeland. 1992. Keoue Cave, northern Leb-anon, and its place in the context of the Levantine Mousterian. In Theevolution and dispersal of modern humans in Asia. Takeru Akazawa, KenichiAoki, and Tatsuku Kimura, eds. Pp. 107–127. Tokyo: Hakusen-sha.

Nowell, April, Francesco d’Errico, and Erella Hovers. 2001. The origin ofsymbolism in the Near East. Abstracts of the 66th annual meeting, April 18-22, 2001, New Orleans, Louisiana. Washington, DC: Society for AmericanArchaeology.

Odell, George H. 1996. Economizing behavior and the concept of curation.In Stone tools: theoretical insights into human prehistory. George H. Odell,eds. Pp. 51–80. New York: Plenum.

Oron, Maya, and Naama Goren-Inbar. 2013. Mousterian intra-site spatialpatterning at Quneitra, Golan Heights. Quaternary International. http://dx.doi.org/10.1016/j.quaint.2013.04.013.

O’Shea, John. 1981. Coping with scarcity: exchange and social storage. InEconomic archaeology: towards an integration of economic and social ap-proaches. A. Sheridan and Geoff Bailey, eds. Pp. 167–183. BAR InternationalSeries 96. Oxford: BAR.

Oswalt, Wendel H. 1976. An anthropological analysis of food-getting technology.New York: Wiley.

Parry, William J., and Robert L. Kelly. 1987. Expedient core technology andsedentism. In The organization of core technology. J. K. Johnson and TobiA. Morrow, eds. Pp. 285–309. Boulder, CO: Westview.

Pawlik, Alfred F., and J. P. Thissen. 2011. Hafted armatures and multi-com-ponent tool design at the Micoquian site of Inden-Altdorf, Germany. Journalof Archaeological Science 38:1699–1708.

Pettitt, Paul. 2010. The Palaeolithic origins of human burial. London: Routledge.Plisson, Hugue, and Silvie Beyries. 1998. Pointes ou outils triangulaires? donees

fonctionnelles dans le mousterian levantin. Paleorient 24:5–16, 23–24.Powell, Adam, Stephen Shennan, and Mark G. Thomas. 2009. Late Pleistocene

demography and the appearance of modern human behavior. Science 324:1298–1301.

Premo, Lukas S., and Steven L. Kuhn. 2010. Modeling effects of local ex-tinctions on culture change and diversity in the Paleolithic. PLoS ONE 5:e15582, doi:10.1671/journal.pone.0015582.

Pulliam, H. Ronald, and Christopher Dunford. 1980. Programmed tolearn: an essay on the evolution of culture. New York: Columbia Uni-versity Press.

Rabinovich, Rivka. 2003. The Levantine Upper Palaeolithic faunal record. InMore than meets the eye: studies on Upper Palaeolithic diversity in the NearEast. A. Nigel Goring-Morris and Anna Belfer-Cohen, eds. Pp. 33–48. Ox-ford: Oxbow.

Rabinovich, Rivka, and Erella Hovers. 2004. Faunal assemblages from AmudCave: preliminary results and interpretations. International Journal of Os-teoarchaeology 14:287–306.

Rabinovich, Rivka, and Eitan Tchernov. 1995. Chronological, paleoecologicaland taphonomical aspects of the Middle Paleolithic site of Qafzeh, Israel.In Archaeozoology of the Near East: proceedings of the second symposium ofthe archaeozoology of southwestern Asia and adjacent areas. H. Buitenhuisand Hans-Peter Uerpmann, eds. Pp. 5–44. Leiden: Backhuys.

Renfrew, Colin. 1978. The anatomy of innovation. In Approaches to socialarchaeology. Colin Renfrew, ed. Pp. 390–418. Cambridge, MA: HarvardUniversity Press.

Reynolds, Robert G. 1978. On modeling the evolution of hunter-gathererdecision-making systems. Geographical Analysis 10:31–46.

Richerson, Peter J., and Robert Boyd. 2005. Not by genes alone: how culturetransformed human evolution. Chicago: University of Chicago Press.

Roebroeks, Wil, Mark J. Sier, Trine K. Nielsen, Dimitri De Loecker, Josep M.Paresb, Charles E. S. Arps, and Herman J. Mucher. 2012. Use of red ochreby early Neandertals. Proceedings of the National Academy of Sciences of theUSA 109:1889–1894.

Roebroeks, Wil, and Paola Villa. 2011. On the earliest evidence for habitualuse of fire in Europe. Proceedings of the National Academy of Sciences of theUSA 108:5209–5214.

Rohling, Eelco J., Katharine M. Grant, Andrew P. Roberts, and Juan-CruzLarrasoana. 2013. Paleoclimate variability in the Mediterranean and RedSea regions during the last 500,000 years: implications for hominin migra-tions. Current Anthropology 54(suppl. 8):S183–S201.

Rolland, Nicholas. 2004. Was the emergence of home bases and domestic firea punctuated event? a review of the Middle Pleistocene record in Eurasia.Asian Perspectives 43:248–280.

Ronen, Avraham. 1979. Paleolithic industries. In The Quaternary of Israel.Aharon Horowitz, ed. Pp. 296–307. New York: Academic Press.

Rossano, Matt J. 2010. Making friends, making tools, and making symbols.Current Anthropology 51(suppl.):S89–S98.

Rots, Veerle, and Philip Van Peer. 2006. Early evidence of complexity in lithiceconomy: core-axe production, hafting and use at Late Middle Pleistocenesite 8-B-11, Sai Island (Sudan). Journal of Archaeological Science 33:360–371.

Rots, Veerle, Philip Van Peer, and Pierre M. Vermeersch. 2011. Aspects oftool production, use, and hafting in Palaeolithic assemblages from NortheastAfrica. Journal of Human Evolution 60:637–664.

Salomon, Helene, C. Vignaud, Y. Coquinot, L. Beck, Chris Stringer, D. Strivay,and Frencesco D’Errico. 2012. Selection and heating of colouring materialsin the Mousterian level of es-Skhul (c. 100,000 years BP, Mt. Carmel, Israel.Archaeometry 54:698–722.

Sandgathe, Dennis M., Harold L. Dibble, Paul Goldberg, Shannon P. Mc-Pherron, Alain Turq, Laura Niven, and Jamie Hodgkins. 2011. On the roleof fire in Neandertal adaptations in western Europe: evidence from Pechde l’Aze and Roc de Marsal, France. Paleoanthropology 2011:216–242.

Schick, Tamar, and Moshe Stekelis. 1977. Mousterian assemblages in KebaraCave, Israel. In Moshe Stekelis memorial volume: Eretz-Israe, vol. 13. BaruchArensburg and Ofer Bar-Yosef, eds. Pp. 97–149. Jerusalem: Israel Explo-ration Society.

Shahack-Gross, Ruth, Avner Ayalon, Paul Goldberg, Yuval Goren, Boaz Ofek,Rivka Rabinovich, and Erella Hovers. 2008. Formation processes of ce-mented features in karstic cave sites revealed using stable oxygen and carbonisotopic analyses: a case study at Middle Paleolithic Amud Cave, Israel.Geoarchaeology 23:43–62.

Sharon, Gonen, Leore Grosman, Hannah Fluck, Yoel Melamed, Yoel Rak,Rivka Rabinovich, and Maya Oron. 2010. The first two excavation seasonsat NMO: a Mousterian site at the bank of the Jordan river. Eurasian Pre-history 7:129–151.

Sharon, Gonen, and Maya Oron. 2013. The lithic tool arsenal of a Mousterianhunter. Quaternary International, forthcoming, doi:10.1016/j.quaint.2013.10.024.

Shea, John J. 1988. Spear points from the Middle Paleolithic of the Levant.Journal of Field Archaeology 15:441–450.

———. 1991. The behavioral significance of Levantine Mousterian variability.PhD dissertation, Harvard University.

———. 1998. Neandertal and early modern human behavioral variability: aregional scale approach to lithic evidence for hunting in the LevantineMousterian. Current Anthropology 39(suppl.):S45–S78.

———. 2003. Neandertals, competition, and the origin of modern humanbehavior in the Levant. Evolutionary Anthropology 12:173–187.

———. 2004. The Middle Paleolithic of the East Mediterranean Levant. Jour-nal of World Prehistory 17:313–394.

———. 2006. The origins of lithic projectile point technology: evidence from

S356 Current Anthropology Volume 54, Supplement 8, December 2013

Africa, the Levant, and Europe. Journal of Archaeological Science 33:823–846.

———. 2008. Transitions or turnovers? climatically-forced extinctions ofHomo sapiens and Neanderthals in the East Mediterranean Levant. Qua-ternary Science Reviews 27:2253–2270.

Shea, John J., Zachary Davis, and Kyle Brown. 2001. Experimental tests ofMiddle Paleolithic spear points using a calibrated crossbow. Journal of Ar-chaeological Science 28:807–816.

Shennan, Stephen. 2001. Demography and cultural innovation: a model andits implications for the emergence of modern human culture. CambridgeArchaeological Journal 11:5–16.

Shennan, Stephen, and James Steele. 2009. Introduction: demography andcultural macroevolution. Human Biology 81:105–119.

Shimelmitz, Ron, Ran Barkai, and Avi Gopher. 2011. Systematic blade pro-duction at late Lower Paleolithic (400–200 kyr) Qesem Cave, Israel. Journalof Human Evolution 61:458–479.

Shott, Michael. 1989. On tool-class use lives and the formation of archaeo-logical assemblages. American Antiquity 54:9–30.

Skinner, J. H. 1970. El Masloukh: a Yabrudian site in Lebanon. Bulletin duMusee de Beyrouth 23:143–172.

Soressi, Marie. 2005. Late Mousterian lithic technology: its implications forthe pace of the emergence of behavioural modernity and the relationshipbetween behavioural modernity and biological modernity. In From tools tosymbols. Lucinda Backwell and Francesco d’Errico, eds. Pp. 389–417. Jo-hannesburg: Witwatersrand University Press.

Speth, John D. 2004. Hunting pressure, subsistence intensification, and de-mographic change in the Levantine Late Middle Paleolithic. In Humanpaleoecology in the Levantine corridor. Naama Goren-Inbar and John D.Speth, eds. Pp. 149–166. Oxford: Oxbow.

———. 2006. Housekeeping, Neanderthal-style: hearth placement andmidden formation in Kebara Cave (Israel). In Transitions before thetransition: evolution and stability in the Middle Paleolithic and MiddleStone Age. Erella Hovers and Steven L. Kuhn, eds. Pp. 171–188. NewYork: Springer.

———. 2010. The paleoanthropology and archaeology of big-game hunting:protein, fat, or politics? New York: Springer.

———. 2012. Middle Palaeolithic subsistence in the Near East: zooarchae-ological perspectives: past, present, and future. Before Farming 2012(2), art.1.

Speth, John D., and Jamie Clark. 2006. Hunting and overhunting in theLevantine Late Middle Palaeolithic. Before Farming 2006(3), art. 1.

Speth, John D., Liliane Meignen, Ofer Bar-Yosef, and Paul Goldberg. 2012.Spatial organization of Middle Paleolithic occupation X in Kebara Cave(Israel): concentrations of animal bones. Quaternary International 247:85–102.

Speth, John D., and Eitan Tchernov. 2007. The Middle Paleolithic occupationsat Kebara Cave: a faunal perspective. In The Middle and Upper Paleolithicarchaeology of Kebara Cave, part I. Ofer Bar-Yosef and Liliane Meignen, eds.Pp. 165–260. Cambridge, MA: Peabody Museum of Archaeology and Eth-nology, Harvard University.

Steegmann, A. Theodore, Jr., Frank J. Cerny, and Trenton W. Holliday. 2002.Neandertal cold adaptation: physiological and energetic factors. AmericanJournal of Human Biology 14:566–583.

Sterelny, Kim. 2012. The evolved apprentice: how evolution made humansunique. Cambridge, MA: MIT Press.

Stiner, Mary C. 1994. Honor among thieves: a zooarchaeological study of Nean-dertal ecology. Princeton, NJ: Princeton University Press.

———. 2005. The faunas of Hayonim Cave, Israel: a 200,000-year record ofpaleolithic diet, demography, and society. American School of PrehistoricResearch Bulletin 48. Cambridge, MA: Peabody Museum of Archaeologyand Ethnology, Harvard University.

Stiner, Mary C. 2013. An unshakable Middle Paleolithic? trends versus con-servatism in the predatory niche and their social ramifications. CurrentAnthropology 54(suppl. 8):S288–S304.

Stiner, Mary C., Avi Gopher, and Ran Barkai. 2011. Hearth-side socioeco-nomics, hunting and paleoecology during the late Lower Paleolithic atQesem Cave, Israel. Journal of Human Evolution 60:213–233.

Stiner, Mary C., Natalie D. Munro, and Todd A. Surovell. 2000. The tortoiseand the hare: small-game use, the broad-spectrum revolution, and Paleo-lithic demography. Current Anthropology 41:39–73.

Stringer, Chris B., J. C. Finlayson, R. N. E. Barton, Yolanda Fernandez-Jalvo,I. Caceres, R. C. Sabin, E. J. Rhodes, et al. 2008. Neanderthal exploitation

of marine mammals in Gibraltar. Proceedings of the National Academy ofSciences of the USA 105:14319–14324.

Texier, Pierre-Jean, Guillaume Porraz, John Parkington, Jean-Phillipe Rigaud,Cedric Poggenpoel, Christopher Miller, Chantal Tribolo, et al. 2010. A How-iesons Poort tradition of engraving ostrich eggshell containers dated to60,000 years ago at Diepkloof Rock Shelter, South Africa. Proceedings of theNational Academy of Sciences of the USA 107:6180–6185.

Thiebaut, Celine, Emilie Claud, Vincent Mourre, Maria-Gema Chacon, Guil-laume Asselin, Michel Brenet, and Benoıt Paravel. 2010. The recycling andreuse of cores and bifaces during the Middle Paleolithic in western Europe.Palethnologie 2, 2010.

Thieme, Hartmut. 1997. Lower Palaeolithic hunting spears from Germany.Nature 385:807–810.

Tillier, Anne-Marie. 1990. Une controverse depassee: l’existence de pra-tiques funeraires au paleolithique moyen. Les Nouvelles de l’Archeologie40:22–24.

Tryon, Christian A., and J. Tyler Faith. 2013. Variability in the MiddleStone Age of eastern Africa. Current Anthropology 54(suppl. 8):S234–S254.

Tryon, Christian A., Sally McBrearty, and Pierre-Jean Texier. 2005. Le-vallois lithic technology from the Kapthurin Formation, Kenya: Acheu-lian origin and Middle Stone Age diversity. African Archaeological Re-view 22:199–229.

Tuffreau, Aalin. 1995. The variability of Levallois technology in northernFrance and neighbouring areas. In The definition and interpretation of Le-vallois technology. Harold L. Dibble and Ofer Bar-Yosef, eds. Pp. 413–427.Madison, WI: Prehistory Press.

Ugan, Andrew, Jason Bright, and Alan Rogers. 2003. When is technologyworth the trouble? Journal of Archaeological Science 30:1315–1329.

Vaesen, Krist. 2012. Cumulative cultural evolution and demography. PLoSONE 7:e40989.

Vaks, Anton, Miryam Bar-Matthews, Avner Ayalon, Alan Matthews, Lud-wig Halicz, and Amos Frumkin. 2007. Desert speleothems reveal cli-matic window for African exodus of early modern humans. Geology35:831–834.

Vanhaeren, Marian, Francesco d’Errico, Chris Stringer, Sarah L. James, Jon-athan A. Todd, and Henk K. Mienis. 2006. Middle Paleolithic shell beadsin Israel and Algeria. Science 312:1785–1788.

Van Peer, Philip. 1992. The Levallois reduction strategy. Monographs in WorldArchaeology 13. Madison, WI: Prehistory Press.

Villa, Paola, Paolo Boscato, Filomena Ranaldo, and Annamaria Ronchitelli.2009. Stone tools for the hunt: points with impact scars from a MiddlePaleolithic site in southern Italy. Journal of Archaeological Science 36:850–859.

Wadley, Lyn. 2005. Putting ochre to the test: replication studies of adhesivesthat may have been used for hafting tools in the Middle Stone Age. Journalof Human Evolution 49:587–601.

Wadley, Lyn, Tamaryin Hodgskiss, and Michael Grant. 2009. Implications forcomplex cognition from the hafting of tools with compound adhesives inthe Middle Stone Age, South Africa. Proceedings of the National Academyof Sciences of the USA 106:9590–9594.

Wallace, Ian J., and John J. Shea. 2006. Mobility patterns and core technologiesin the Middle Paleolithic of the Levant. Journal of Archaeological Science 33:1293–1309.

Walter, Phillipe. 2003. Characterisation des traces rouges et noires sur lescoquillage perfores de Qafzeh. In Echanges et diffusion dans la prehistoiremediterraneenne. Bernard Vandermeerch, ed. Pp. 122. Paris: Editions duComite des travaux historiques et scientifiques.

Weiner, Steve, Paul Goldberg, and Ofer Bar-Yosef. 2002. Three-dimensionaldistribution of minerals in the sediments of Hayonim Cave, Israel: diage-netic processes and archaeological implications. Journal of ArchaeologicalScience 29:1289–1308.

Weinstein-Evron, Mina, Guy Bar-Oz, Yossi Zaidner, Alexander Tsatskin, DotanDruck, Naomi Porat, and Israel Hershkovitz. 2003. Introducing MisliyaCave, Mount Carmel, Israel: a new continuous Lower/Middle Paleolithicsequence in the Levant. Eurasian Prehistory 1:31–55.

Whallon, Robert, William Lovis, and Robert Hitchcock, eds. 2011. Informationand its role in hunter-gatherer bands. Los Angeles: UCLA Cotsen Institutefor Archaeology.

White, Mark, Nick Ashton, and B. Scott. 2011. The emergence, diversity andsignificance of Mode 3 (prepared core) technologies. Developments in Qua-ternary Science 14:53–65.

Hovers and Belfer-Cohen Variability and Complexity S357

Wobst, H. Martin. 1974. Boundary conditions for Paleolithic social systems:a simulation approach. American Antiquity 39:147–178.

———. 1976. Locational relationships in Paleolithic society. Journal of HumanEvolution 5:49–58.

Wojtzak, Dorothea. 2011. Hummal (central Syria) and it eponymous industry.In The Lower and Middle Palaeolithic in the Middle East and neighbouringregions: Basel symposium, mai 8–10 2008. Jean-Marie Le Tensorer, RhetoJagher, and Marcel Otte, eds. Pp. 289–307. Etudes et recherches archeo-logiques de l’Universite de Liege 126. Liege: University of Liege Press.

Wurz, Sarah. 2013. Technological trends in the Middle Stone Age of SouthAfrica between MIS 7 and MIS 3. Current Anthropology 54(suppl. 8):S305–S319.

Wynn, Thomas, and Frederick L. Coolidge. 2012. How to think like a Nean-dertal. Oxford: Oxford University Press.

Yeshurun, Reuven, Guy Bar-Oz, and Mina Weinstein-Evron. 2007. Mod-ern hunting behavior in the Early Middle Paleolithic: faunal remainsfrom Misliya Cave, Mount Carmel, Israel. Journal of Human Evolution53:656–677.

Young, H. Peyton. 2012. The dynamics of social innovation. Proceedings ofthe National Academy of Sciences of the USA 108:21285–21291.

Zaidner, Yossi, Amow Frumkin, Nqomi Porat, Alexander Tsatskin, ReuvenYeshurun, and Lior Weissbrod. Forthcoming. A series of Mousterian oc-cupation in a new type of site: the Nesher Ramla karst depression, Israel.Journal of Human Evolution.

Zilhao, Joao, Diego E. Angelucci, Ernestina Badal-Garcıa, Francesco d’Errico,Floreal Daniel, Laure Dayet, Katerina Douka, et al. 2010. Symbolic use ofmarine shells and mineral pigments by Iberian Neandertals. Proceedings ofthe National Academy of Sciences of the USA 107:1338–1342.

S358 Current Anthropology Volume 54, Supplement 8, December 2013

� 2013 by The Wenner-Gren Foundation for Anthropological Research. All rights reserved. 0011-3204/2013/54S8-0014$10.00. DOI: 10.1086/673502

Paleolithic Cultures in ChinaUniqueness and Divergence

by Xing Gao

This paper presents an overview of the Chinese Paleolithic industries between 300 ka and 40 ka, a time span nowtermed the “later Early Paleolithic” (LEP) in the Chinese chronological scheme. It describes the unique features ofLEP remains in China compared with contemporaneous materials in Africa and western Eurasia as well as theinternal diversity and complexity of these Chinese Paleolithic assemblages. Basic features of LEP remains in Chinainclude the persistent and conservative pebble-tool and simple flake-tool traditions, the use of poor-quality localraw materials, tool fabrication on pebbles and direct use of unretouched flakes, opportunistic flaking, simple andcasual modification, and the lack of obvious temporal trends. The diversity and complexity of Chinese Paleolithiccultures as they are expressed in terms of the major difference between southern China’s pebble-tool tradition andnorthern China’s simple flake-tool tradition are also assessed. Based on such generalizations and analyses, a com-prehensive behavioral model is proposed to explain the unique features of LEP cultures in China and the alternativepathway of human evolution and adaptation in China during that period of time.

Introduction

Recent research and discussions concerning Pleistocene hu-man technological development and adaptive strategies havelargely concentrated on archaeological materials during theMiddle Late Stone Age and the Middle Upper Paleolithictransitions in Africa and western Eurasia, where scenarios ofearly modern human origins, dispersal, and their replacementof the Neanderthals are believed to have taken place. In thiswave of heated discussions and debates, China and East Asiakeep almost silent. While the Out of Africa theory enjoysoverwhelming support, the Continuity with Hybridizationtheory looks odd and outmoded, and the cultural remains ofHomo erectus and archaic Homo sapiens seem irrelevant andnegligible, for they might be outside the lineage leading di-rectly to living humans. Once again, as with the influentialMovius Line hypothesis that prevailed in the mid-twentiethcentury (Movius 1944, 1948), Paleolithic cultures in Chinaare found deep in a “backwater.”

Human evolution toward complexity and modernity mighthave taken different pathways in different regions. In thisregard, Asia has received far less attention than Africa andEurope in the search for human origins, but it is no longer

Xing Gao is a Research Member of the Laboratory of VertebrateEvolution and Human Origins at the Institute of VertebratePaleontology and Paleoanthropology of the Chinese Academy ofSciences (P.O. Box 643, Beijing 100044, China [gaoxing@ivpp.ac.cn]).This paper was submitted 3 VII 13, accepted 13 VIII 13, andelectronically published 14 XI 13.

considered to be of marginal importance. Indeed, a globalperspective on human origins cannot be properly understoodwithout a detailed consideration of the largest continent(Dennell 2009). While studies on African Middle Stone Ageand western Eurasian Middle Paleolithic industries may pro-duce insights about the emergence of modern human be-havior, technology, and the relationship/possible interactionbetween the intruding early modern human groups and theNeanderthals, research on contemporary East Asian Paleo-lithic remains may provide us a comparative data set of cul-tural and behavioral variability for ancient human groupsliving in different environmental and ecological zones, andthis may encourage us to look at the third lineage of humanevolution in late Middle Pleistocene. The unique feature and“conservative” progression of eastern Asian Paleolithic in-dustries in general and their obvious distinction from the Westseems to support the scenario of continuous evolution anddevelopment of local populations, which may present a majorchallenge to the Out of Africa or Total Replacement hypoth-esis (Gao et al. 2010).

In order to provide relevant information for this workshop,titled “Alternative Pathways to Complexity: Evolutionary Tra-jectories in the Middle Paleolithic and Middle Stone Age,”descriptions of the Chinese Paleolithic remains and discus-sions of related questions of this paper will focus mainly onthe time period between 300 ka and 40 ka, a cultural stagepreviously described as the later part of the Lower Paleolithicand the whole Middle Paleolithic, now reclassified as the laterEarly Paleolithic (LEP) by me and my colleagues (Gao 2000;Gao and Norton 2002).

Gao Paleolithic Cultures in China S359

Research Tradition and Conceptionconcerning the Chinese Early Paleolithic

The practice of Paleolithic research in China still to a certainextent differs from that of the West, including terminology,the classification-typology system, and the way data are pre-sented and interpreted. Therefore, a brief introduction to theresearch tradition and its conception may help Western schol-ars understand the content of this paper and research accom-plishments in this field in China.

A Shift from the Middle Paleolithic to the LEP in ChinesePaleolithic Research

Paleolithic archaeology in China was an adopted enterprisefrom the West in the 1920s and 1930s, and the three-stagePaleolithic cultural model was carried along with the Westernscientists who came to China to initiate this field and trainlocal scholars. Accordingly, the Chinese method copied theWestern model that was based on artifactual material indig-enous to western Eurasia, and the use of similar develop-mental stages implies that the cultural evolutionary trajectoryin China was similar to Africa and western Eurasia. However,this practice ran into problems when it became obvious thatin fact few similarities appear to exist between the Chinesematerials and those of the West (Gao and Olsen 1997; Ikawa-Smith 1978; Movius 1944) and that the Western “index”stone-tool types of different cultural stages are quite scarcein China and East Asia. Consequently, the derived Paleolithiccultural development periodization in China began to take adifferent approach. Two criteria have been utilized for defin-ing a distinct Middle Paleolithic in China: (1) age of site (i.e.,all archaeological materials dating from the late Middle–EarlyUpper Pleistocene—ca. 200–40 ka—were considered MiddlePaleolithic); and (2) association with archaic Homo sapiensremains.

Gradually, as the weakness of defining cultural stages basedon chronometric information and association with certainkinds of human fossils became obvious, researchers began toagree that such a practice must be based exclusively on thearchaeological record (Gao 1999). An analysis of four stone-tool criteria (raw material procurement, core reduction, re-touch, and typology) to determine whether a distinct MiddlePaleolithic existed in the Chinese record indicates that verylittle or very gradual change occurred in lithic technology andtypology between the Lower and Middle Paleolithic. Accord-ingly, there is little reason to retain the three-stage model ofcultural sequence. Instead, a two-stage progression is pro-posed consisting of the Early and the Late Paleolithic. Thetransition between these two cultural periods occurred withthe development of more refined lithic techniques (e.g., bladeand microblade technology) and the presence of ornaments,

art, and/or symbolism, indicators of modern human behavior(ca. 35–30 ka; Gao and Norton 2002).

Movius’s Partition of Two Paleolithic Traditions andIts Influence

Today, most archaeologists agree that the Early Paleolithicassemblages of China and East Asia differ in a number offundamental ways from Lower and Middle Paleolithic assem-blages of Africa and western Eurasia. For decades, Movius’spartition of two cultural traditions and his hypothetical in-terpretations of it have dominated discussions.

Movius proposed two technological traditions in the LowerPaleolithic: one is the Acheulean handaxe tradition of Africaand western Eurasia, characterized by handaxes and otherlarge bifaces. The other is the chopper-chopping tool traditionof East Asia, represented by simple core tools made of pebbles(Movius 1948). His explanation for this difference was that(1) East Asia is a marginal region of human biological andcultural evolution that somehow broke away from the main-stream of human development and maintained the technologyof the earliest phase of human culture in an isolated context,and (2) the quality of raw material in East Asia was so poorthat it would not permit the ancient population there to makebetter implements.

Even though Movius focused mainly on the Lower Pa-leolithic or the earlier period of the Chinese Early Paleolithicin the new scheme, his conceptualization has profound in-fluence on the study of other periods, and Paleolithic researchhas somehow been placed under its shadow since then. Forhalf a century, Chinese scholars have voiced some of thestrongest opposition to the Movius Line. They criticize Mov-ius’s conception of “chopper-chopping tool” in the East, ar-guing that such artifacts are not typical and dominant in theEast Asian Paleolithic complexes, that the Chinese and EastAsian Paleolithic assemblages are not simple and homoge-neous, and that a certain degree of cultural variability andinnovation are evident in the archaeological record. Chineseresearchers are also eager to demonstrate that some Westerncultural elements, such as Acheulean tool kits and the Lev-alloisian technological products, are also presented in the Eastand that therefore there are no fundamental differences be-tween the East and West and that East Asia was never a“cultural backwater” (Huang 1989a, 1989b, 1993; Huang,Hou, and Gao 2009). However, not everybody agrees withsuch a counterview. Some Western scholars still believe thatMovius’s “basic characterization of the major characteristicsof early stage technologies in eastern Asia still holds up”(Schick 1994:579), and some Chinese researchers also madesimilar statements that Movius’s basic observation and con-clusion were still applicable to the Chinese Paleolithic ma-terials before the Late Paleolithic (Lin 1994, 1996). In otherwords, attempts to support or invalidate the Movius Linehypothesis have become central to much Paleolithic researchin China.

S360 Current Anthropology Volume 54, Supplement 8, December 2013

The Controversial Hypothesis of Two Parallel PaleolithicTraditions in North China

An example of the Chinese archaeologists’ attempts to de-nounce the Movius theory and to demonstrate the diversityof Paleolithic cultures in the region is the proposition of twoparallel Paleolithic traditions in North China. This notionwas first put forth in 1972 by Jia and colleagues (Jia, Gai, andYou 1972) and was further developed later (Jia and Huang1985). Basically, it states that there are two parallel lithictraditions in North China. One is the Kehe-Dingcun Series,characterized by large chopper-chopping tools and triangularpoints; the other is the Zhoukoudian Locality 1-Shiyu Series,characterized by small flake tools such as scrapers and burins.The two “traditions” were postulated as extending in parallelfrom the Early Paleolithic all the way into the Late Paleolithicand even the Neolithic and developing into two differentagricultural patterns. This hypothesis had been influential andmade significant impact on Paleolithic research in China dur-ing the last three decades of the twentieth century, and someresearchers are still working in this domain up to the present.

Such a notion is quite troublesome. The sites of the two“traditions” are distributed basically in the same region, andit is difficult to imagine that two distinct cultural traditionscan exist side by side in the same area for about a millionyears. Studies have revealed that most of the sites of the large-tool tradition, including the key site Dingcun, are in factdominated by small flake tools (Zhang 1993). The most se-rious problem with the “large-tool tradition” is taphonomic:almost all the localities assigned to the “large-tool tradition”were fluvial sites exhibiting traces of disturbance and sec-ondary deposition, and many of these large stone tools wereselectively collected from the ground surface. Therefore, thesecollections could not represent a complete “assemblage” or“tool kit” and certainly not a “cultural tradition.” Anotherdrawback of this hypothesis is that it may well overlook orminimize variations within the specific lithic assemblages as-cribed to each of these so-called traditions and, at the sametime, potentially underestimate the similarities among in-dustries of different “traditions.” In short, the diversity andcomplexity of the lithic industries in China cannot be sum-marized simply by two unilinear traditions based solely ontypological and morphological analyses.

Communication Obstacles and the Distinct PaleolithicResearch Tradition in China

China has rich collections of Paleolithic remains. Many for-eign researchers are eager to access the collections and toestablish fruitful contact with colleagues there. However, theyoften feel frustrated; typical comments are “prehistorians out-side of China have found it difficult to obtain good infor-mation about the results of this surge in archaeological in-quiry, much less to synthesize a comprehensive understandingof Paleolithic trends in eastern Asia” (Schick and Dong 1993:

22) and “the number of well-documented Lower Paleolithicsites remains very small, and the credibility of some of theindustries must be questioned because of limited investigationor the lack of the firm chronometric data. Because of theseproblems, it is unreasonable to expect very sophisticated treat-ment of specific areal, temporal, and cultural topics. Studentswill find much of the literature very limited from theoreticaland methodological perspectives and rather vague when com-pared with that of better-studied areas” (Yi and Clark 1983:181).

Today, Chinese Paleolithic research and Chinese archae-ology as a whole still maintain an ambiguous relationshipwith Western practice and theory. Several factors could beresponsible for this. (1) Language: most of the site reportsand research papers by the Chinese researchers are publishedin Chinese journals, and the language barrier prevents Chineseand foreign scholars from sharing information and exchang-ing ideas freely. (2) Different research priorities: a large partof archaeological activities in China are undertaken as salvageprojects, and rescue of the artifacts or cultural relics and basicclassification and description rather than detailed analysis andtheoretical explanations are the principal tasks of the field-worker. (3) The profound impact of traditional epigraphy onmodern archaeological practice helps maintain a long-lastingand persistent classification and description tendency, and thepolitical situation, especially the adoption of Marxism andMaoism as notional ideology by the New China, strengthenedthe tradition. (4) Believing in the philosophy that scientificreasoning should be very cautious and that one has to ac-cumulate enough data before reaching meaningful conclu-sions, most Chinese scholars are reluctant or lack confidenceto touch theoretical issues, and if they have to do so, theywould like to take an inductive approach rather than beingdeductive. (5) A strict and somewhat exclusive governmentpolicy regarding foreigners’ involvement in archaeological re-search in China, especially fieldwork, makes it difficult forforeign scholars to get firsthand information and experienceand to establish long-term research programs there. Weshould also draw attention to the fact that compared withthe large size of the country and the rich archaeological ma-terials, the Paleolithic research community in China is verysmall, and the number of well-trained professional researchersfamiliar with global methodological and theoretical ap-proaches is inadequate. In addition, many Chinese researchersare accustomed to doing research in their own territory andseldom go beyond its border, even within China.

I would argue that the difficulties and differences in lan-guage and research practices are superficial. The more fun-damental reason that the Chinese record plays such a limitedrole in discussion of evolutionary trends within the Middleand Late Pleistocene is that the archaeological evidence is stilldifficult to reconcile with what is known from Europe andAfrica. Chinese and eastern Asian Paleolithic materials reallyare different from those of the West in many ways; theirunique types and morphological and developmental features

Gao Paleolithic Cultures in China S361

Figure 1. Key localities among major LEP sites in China. 1, Jinniushan; 2, Miaohoushan; 3, Xujiayao; 4, Zhoukoudian Locality 15;5, Dingcun; 6, Dali; 7, Longyadong; 8, Lingjing; 9, Jigongshan; 10, Jingshuiwan; 11, Guanyindong; 12, Dadong.

make Western typology and terminology difficult to apply,and many Western scholars do not fully understand them.They expect to find similar cultural remains and familiar re-search results, but they are often disappointed. We have failedto work out a typological-descriptive system and researchnorms that can be effectively applied to archaeological ma-terials from both the West and the East.

The Chinese LEP Archaeological Record:Uniqueness versus Divergence

Literally hundreds of archaeological sites estimated to belongto the 300–40 ka time span have been reported in China (fig.1). The exact figure is hard to calculate, for some sites wereassigned a wide chronological range, and it is difficult to saywhether they belong to this stage; and in some regions, manylocalities were identified and numbered, and it is difficult todetermine whether they belong to one site or discrete sites.I will describe a few key sites to provide some basic infor-mation.

Key Sites

Miaohoushan. Miaohoushan is a cave site situated nearBenxi City, Liaoning Province, in Northeast China (Zhang1989). It was discovered in 1978 and excavated in 1978, 1979,and 1980. A few human fossil fragments, probably of archaic

Homo sapiens, numerous mammalian fossils, some stone ar-tifacts, and ash and burned items were unearthed from severalcultural horizons. These cultural remains were generally datedto 140–250 ka, and even older ages were suggested (Wei 2009;Zhang et al. 2007). A total of 64 lithic artifacts have beenreported and analyzed, including simple cores, flakes, andretouched pieces. Direct hammer percussion was used as theprincipal method of core reduction; core preparation wasseldom applied. The tools were fabricated coarsely, and theartifacts vary greatly in size and morphology. Side scrapersare the dominant tools, and most of them are small; chopper-chopping tools take the second position, and some of themare very large.

Jinniushan. Jinniushan is a cave site complex located nearYingkou City, Liaoning Province, Northeast China. It wasdiscovered in 1974 and excavated in the 1970s and 1980s (Lu2004). The excavations resulted in the discovery of a partialskeleton of archaic Homo sapiens, numerous mammalian fos-sils, and stone artifacts. A few hearths covered by rocks be-lieved to be evidence of fire preservation, along with burntmaterials, were also unearthed. Several chronometric testshave been applied to the human fossil and artifact-bearinghorizons, and ages of 230–300 Ka, 228 Ka, and 187 Ka werereported (Chen, Yang, and Wu 1994). Only a part of theunearthed materials has been analyzed and published. Ac-

S362 Current Anthropology Volume 54, Supplement 8, December 2013

cording to the available information, the lithic assemblageincludes simple cores, flakes, scrapers, points, and burin. Corereduction was mainly carried out through direct hammer per-cussion without core preparation, and the bipolar techniquewas also used. The tool kit is dominated by side scrapers, andretouch on them is simple and casual. Most artifacts are smalland vary in size and morphology. Some chipped-bone toolswere also believed to be produced and used.

Zhoukoudian Locality 15. Zhoukoudian Locality 15 is partof the limestone cave complex of the Zhoukoudian site, lo-cated 50 km southwest of Beijing. The locality was discoveredin 1932 and excavated from 1935 to 1937. A large quantityof vertebrate fossils and lithic artifacts was unearthed. Faunalassemblage indicates a late Middle to early Upper Pleistoceneage, and limited uranium series dates estimated an age rangebetween 140 and 110 ka for the cultural horizon (Gao 2000).The rich lithic assemblage is composed of more than 10,000stone artifacts, including hammerstones, cores, flakes, re-touched pieces, and chunks, with the latter predominating.Hammer percussion was the principal flaking strategy, butbipolar flaking was also employed frequently, which made theassemblage unique in the Chinese LEP and a clear successorof the earlier “Peking Man” industry known from Zhoukou-dian Locality 1. In addition to the simple cores, regular discoidcores and heavily reduced polyhedral cores were also present,testifying to a sophisticated fashion of core reduction. Theretouched tools are mostly side scrapers; other tool types in-clude chopper-chopping tools, backed knives, points, awls,notches, and burins (fig. 2B). The tools are mostly fabricatedon flakes and are small in size, but the few pieces of backedknives or cleavers are large and distinctive. Most of the ar-tifacts were fabricated from locally available quartz, a rawmaterial source characterized by high abundance and lowworkability. An analysis of raw material utilization reveals thatsome simple but practical and efficient strategies were adoptedto make use of these raw materials: different materials wereprocured and consumed differently (table 1); the site wasprovisioned with abundant potential tool-making materials;numerous flakes were detached, but only a portion was se-lected for modification and utilization (Gao 2003).

Dali. Dali is an open-air site discovered in 1978 in ShanxiProvince, North China, and is best known for the presenceof an archaic Homo sapiens skull (Wu 2009), but some faunalremains and 564 pieces of stone artifacts were also collectedfrom sandy deposits. U-series and electron spin produced anage range between 380 and 140 ka (Chen, Yuan, and Gao1984; Yin et al. 2001). Core reduction at the site was foundto be conducted through hammer percussion and the bipolarmethod. Most of lithic artifacts are very small side scrapers,points, burins, and drills produced from quartzite, flint, andquartz materials. Retouch on these pieces is very simple andcasual, and some of them were heavily worn and difficult tostudy typologically and technologically.

Dingcun. The Dingcun site complex is located in ShanxiProvince, North China, and was originally discovered in 1953.The site is made up of a dozen separate open-air localities.Several excavations have been carried out at the site; a parietalskull and some isolated teeth of archaic Homo sapiens and alarge quantity of vertebrate fossils and lithic artifacts werecollected. U-series dates and lithostratigraphic and bio-stratigraphic reconstructions indicated an age range of 260–107 ka for the cultural remains (Norton, Gao, and Feng 2009).

For a long time, the rich lithic collection from the site hasbeen the center of discussion and debate in Paleolithic re-search in China. Jia nominated Dingcun as the representativesite of the “large chopper-chopping tool and triangular pointtradition” of North China, for most of the artifacts wereconsidered to be large ones, and some of them exhibiteddistinctive typological and technological features. Huang putforward the notion that many large pointed tools in the as-semblage were handaxes; therefore, Dingcun was the centerof a handaxe zone along the Fen and Wei rivers in NorthChina (Huang 1989a). However, close examination of thecollections reveals that the assemblages are dominated bysmall flake tools, mainly side scrapers and points, not largepebble tools, and there are no real handaxes from the site(Gao 2011). Still, the Dingcun industry shows some specialcharacteristics within the flake-tool techno-complex in Chinain that core reduction, even though still through simple hard-hammer percussion and possible anvil technique, seems tobe more sophisticated, and many large and regular flakes andtools were produced. Some large well-made triangular picksmade on flakes with unique morphology as well as chopper-chopping tools and spheroids were also present (fig. 2D). Amajor reason for the uniqueness of the Dingcun industrymight be the exploitation of high-quality dark hornfels avail-able in nearby river beds.

Xujiayao. Xujiayao is a fluvial-lacustrine open-air site lo-cated in the western margin of the Nihewan Basin in HebeiProvince, North China (Jia, Wei, and Li 1979). It was exca-vated three times in the late 1970s and more in recent years.Some fragmental archaic Homo sapiens fossils and an arrayof vertebrate fossils and Paleolithic materials were recovered.U-series, paleomagnetism, and optically stimulated lumines-cence (OSL) dating have been applied to the site, and datesof 125–104 ka, 117 ka, and 69 ka have been obtained, re-spectively (Chen et al. 1982; Liu, Su, and Jin 1992; Nagatomoet al. 2009). A recent study on 1,765 lithic artifacts unearthedin 1977 indicates that the assemblage includes cores, flakes,retouched pieces, chunks, and debris (Ma, Pei, and Gao 2011).Raw materials were mainly quartzite and quartz pebbles se-lected from nearby river beds, simple hammer percussion wasused for core reduction, and a certain number of discoid andpolyhedral cores were left behind. About half of the retouchedpieces are small side scrapers made on flake blanks, and sphe-roids constitute more than 27% of the assemblage, which isvery distinctive. Other tool types include point, notch, den-

Figure 2. Line drawings of stone artifacts from some key LEP sites in North China. A, Longyadong; B, Zhoukoudian Locality 15;C, Xujiayao; D, Dingcun.

S364 Current Anthropology Volume 54, Supplement 8, December 2013

Table 1. Raw material frequencies for artifacts by class from Zhoukoudian Locality 15

Class

Quartz Igneous Crystal Flint Sandstone Quartzite

N % N % N % N % N % N

Core 126 1.8 2 1 1Flake 393 5.7 113 1.6 15 .2 9 .1Bipolar 86 1.3 1Hammer 5 .1 2Chunk 4,730 68.9 32 .5 66 1.0 1Tool 1,198 17.4 54 .8 14 .2 12 .2 4 .1 1

ticulate, chopper-chopping tools, burin, and drill tools (fig.2C). Some bone tools were also recognized in previous studies.Because of the high density of equid and rhinoceros bonesand artifacts, particularly stone spheroids and bone tools, Xu-jiayao was interpreted as a horse kill site.

Lingjing. Lingjing is an open-air site located near XuchangCity, Henan Province, central China. The site was discoveredin the mid-1960s and excavated recently. A broken skullcapof possible archaic Homo sapiens was unearthed in associationwith a large quantity of mammalian fossil fragments andthousands of stone tools. OSL dates and biostratigraphic dataestimate the age of 110–80 Ka for the Paleolithic horizon.The lithic assemblage is dominated by small side scrapers,points, and drills made of vein quartz flakes, but some heavy-duty chopper-chopping tools and picks made of quartziteblank are also present (Li 2007). Core reduction and toolmanufacture are found to be simple and casual. Some mod-ified bone tools, mostly pointed ones, have been identifiedand analyzed (fig. 3), and use-wear analysis results suggestthat some bone tools were used for drilling, penetrating, andscraping animal substances and that some might have beenhafted during use (Li and Shen 2010).

A study of mortality profiles of the large herbivores fromthe site suggests that the accumulation of rich mammalianbone fragments is the result of human hunting and butch-ering. Aurochs (Bos primigenius) and horse (Equus caballus)are the major prey species, and the age structures of theseanimals can be best described as the “prime-dominated pat-tern.” This study confirms the well-established notions atmany Middle and Upper Paleolithic sites across Eurasia andAfrica that Middle Stone Age/Middle Paleolithic foragers werefully effective in hunting large prey species, particularly au-rochs and horse, which might indicate that the hunting be-haviors and subsistence strategies were not significantly dif-ferent between Middle Paleolithic and Upper Paleolithichumans in East Asia and hence suggest the early emergenceof modern human behaviors in this area (Zhang et al. 2009).

Sites in the Luonan Basin. Since 1995, more than 300 siteshave been found in the Luonan Basin of southern ShanxiProvince, central China. Among them, only Longyadong is acave site, and the others are open-air sites. Tens of thousandsof stone artifacts have been collected from different river ter-races along the Luonan River, and most of them are surface

finds (Wang 2005). Paleomagnetism and OSL dating suggestthat ancient humans stayed in the area off and on in the timespan of 800–140 ka (Lu et al. 2007). These sites have attracteda great deal of attention in China because of the discovery ofa certain number of Acheulean-like handaxes, cleavers, andpicks along with other large pebble tools and small flake tools(fig. 2A). Raw materials used for stone-tool manufacture areoverwhelmingly quartzite pebbles. Such a tool kit is quiteunique in central-southern China’s pebble-tool zone in whichsimple and large chopper-chopping tools dominated the in-dustries throughout the Paleolithic. Recent excavation andchronometric dating suggest that some of the Acheulean-styleartifacts might come from the Upper Pleistocene horizon,which may present challenges to the study of Paleolithic hu-man adaptation, migration, interaction, or convergent cul-tural development.

Jigongshan. The Jigongshan site is situated in the Jingzhoudistrict of Hubei Province along the Yangtze River. It wasdiscovered in 1984 and excavated in 1992. Numerous lithicmaterials were unearthed from two cultural horizons at thesite. The lower horizon was estimated to be of the early UpperPleistocene and the upper horizon of the late Upper Pleis-tocene. Lithic assemblages from the two horizons were dom-inated by simple cores, flakes, and chunks. Most of the toolsfrom the lower horizon were made of pebbles, and heavy-duty tools such as picks and chopper-chopping tools makeup the large majority of the assemblage (fig. 4D). Tools fromthe upper horizon were mostly made on flakes, and most ofthem are small irregular scrapers. According to the site report(Liu and Wang 2001), the “living floor” with a circular struc-ture composed of pebbles and artifacts was identified fromthe lower cultural horizon.

Sites in Sanxia (Three Gorges Region). More than 20 Pa-leolithic sites have been discovered and excavated in the ThreeGorges Region (Sanxia in Chinese) in Chongqing MunicipalCity, central China, in the past two decades, and some siteshave been dated to the time span of 140–70 ka, such asJingshuiwan (Gao and Pei 2010; Pei et al. 2006). Lithic ar-tifacts from these sites are typical of the pebble-tool industriesthat prevailed in southern and central China during the entirePleistocene. Assemblages are dominated overwhelmingly bylarge chopper-chopping tools and picks made of pebbles. Onlya small portion of the tools were fabricated on flake blanks.

Gao Paleolithic Cultures in China S365

Figure 3. Possible bone tools from Lingjing.

Raw materials exploited are mostly quartzite and quartzitesandstone. Core reduction was carried out using direct ham-mer percussion and a kind of special “throw and collisionmethod.” Only simple cores and flakes were produced. Re-touch of tools was simple and casual (fig. 4A).

Dadong (Grand Cave). The Dadong or Grand Cave site islocated in Guizhou Province, South China, and was discov-ered and excavated in 1990s. More than 2,000 artifacts havebeen unearthed (Huang, Hou, and Si 1997), and the age ofhuman occupation of the site has been estimated to be 260–142 ka (Wang et al. 2003). Raw materials exploited are mostlysmall flint nodules. Only direct hard-hammer percussion wasemployed for core reduction, and a few Levallois-like flakeswere reported. The assemblage is dominated by small flaketools, including side scrapers, drills, notches, denticulates, andend scrapers (fig. 4C). A few pieces of rhinoceros teeth wereidentified as modified into scraping tools.

Guanyindong. The Guanyindong cave site is also located inGuizhou Province, South China, and was discovered and ex-cavated in the 1960s (Li and Wen 1986). More than 3,000stone artifacts and numerous animal fossils were unearthedfrom two depositional units. Stone-tool raw materials ex-ploited are locally available flints. Direct hard-hammer per-cussion was believed to be used for core reduction and re-touch, and a few Levallois-like flakes were identified. A richvariety of tool types were recognized from the assemblage,including side scrapers, end scrapers, notches, denticulates,

points, drills, burins, and chopper-chopping tools; most ofthem are small flake tools (fig. 4B). The retouched piecesexhibit a simple and irregular mode of modification; someof them possess more than one cutting edge, and the edgesare usually thick and steep. The Guanyindong site was initiallyestimated to be of Middle Pleistocene age, and the culturalremains were believed to be comparable with those of thePeking Man site at Zhoukoudian. However, later chrono-metric dating placed the lower horizon at 50–140 ka and 180–240 ka and the upper horizon to be younger than 40 ka (Shenand Jin 1992).

Principal Features of the Chinese LEP Paleolithic Assemblages

It is obvious that Early Paleolithic industries in China havetheir unique features compared with contemporary culturalremains in Africa and western Eurasia. The most distinctivecultural characteristics of the Chinese LEP Paleolithic assem-blages can be summarized as follows.

1. Slow or conservative development process in that ModeI technology and assemblages prevailed for all the LEP. Mostof the assemblages consist of simple cores, irregular flakes,side scrapers, chopper-chopping tools, points, picks, and soforth. It is true that Acheulean-like tool kits, including hand-axes, cleavers, and picks were reported from some localitiesof the Luonan Basin and the Dingcun site. Some Acheulean-like assemblages from the Luonan Basin were estimated to beof the late Middle Pleistocene or even the upper Pleistocene.

Figure 4. Line drawings of stone artifacts from some key LEP sites in South China. A, Jingshuiwan; B, Guanyindong; C, Dadong;D, Jigongshan.

Gao Paleolithic Cultures in China S367

However, most of these artifacts are surface finds, and theage of such assemblages needs to be further analyzed. Fur-thermore, handaxes found in this region are mostly pointedand thick, and no trace of soft-hammer retouch and thinning(typical of contemporaneous industries in Europe, the NearEast, and Africa) can be observed from the samples. No realhandaxes or cleavers were collected from the Dingcun site,and no technological mark of soft-hammer flaking and thin-ning can be identified from the artifacts. Before 40 ka, noMousterian-style assemblages or blade technology can beidentified in the Chinese Paleolithic industries.

2. Only local raw materials were exploited for tool making,and they are mostly poor-quality quartz, quartzite, sandstone,and silicified limestone. At some sites chert and flint wereused, but they are usually of poor quality and in small nodules.No evidence of quarrying and long-distance transportationof high-quality raw material has been detected so far.

3. A variety of flaking methods used for core reduction isevident in these industries, including direct hard-hammerpercussion, bipolar, and block-on-block techniques. Coreswere rarely prepared, and no application of real Levalloisiantechnique was recognized. Simple cores with a few detach-flake scars and polyhedral ones are the most numerous, anddiscoidal cores appear in some assemblages. Flakes are usuallysmall, irregular, and vary in size and morphology. Evidencefor “predetermination” of flake shapes by core preparation islacking. No trace of soft-hammer flaking has been recognizedfor this period of time.

4. Tools are mostly simple, irregular, and casually modified,and some are difficult to classify into discrete types. In thenorth—that is, to the north of the Qinling Mountains andthe Huai River, a natural boundary that normally dividesChina into two ecological zones—industries were mostlydominated by small flake tools such as side scrapers, points,and drills supplemented by chopper-chopping tools, picks,and spheroids, while in the south, industries were dominatedby large pebble tools, especially chopper-chopping tools andpicks. Generally speaking, the degree of tool standardizationis pretty low, and flakes were frequently utilized without fur-ther modification. There are certainly exceptions. Some largeand well-made digging-cutting tools were collected from afew localities in the Dingcun site complex and the LuonanBasin, such as triangular picks, handaxes, and cleavers, andsome small flake tools, such as side scrapers and points, un-earthed from sites such as Zhoukoudian Location 15 andGuanyindong, are found to be fabricated delicately and skill-fully. It might mean that when raw material was suitable andnecessity arose, human groups living at these sites were ca-pable of producing regular, efficient and curated tools.

Any attempt to summarize the basic features of the ChineseLEP as a whole is inevitably oversimplified. While the factthat Paleolithic industries in China and East Asia are differentfrom those of the West has been recognized, regional diversityand internal complexity are also evident in these industries.China covers a vast and geographically/ecologically diverse

area, and it is more appropriate to think about regional find-ings according to paleoecological conditions rather than treat-ing the vast territory as a whole.

The complexity of Early Paleolithic industries in China andtheir regional variability have been realized in the past threedecades. Zhang Senshui’s observation of “two main ChinesePaleolithic industries with numerous local cultural variants”is a good example (Zhang 1999:198). He proposed that twoprincipal industries could be recognized from the ChinesePaleolithic remains, that is, the small flake tool industry inNorth China and the large pebble-tool industry in SouthChina. He called them the northern main industry (NMI)and the southern main industry (SMI), respectively. Withinthese two cultural zones, divided by the Huai River in theeast and the Qinling Mountains in the west, he further rec-ognized numerous local cultural variants. Major features ofNMI were summarized as including the domination of smallflake tools, mainly various kinds of side scrapers, points, awls,burins; large pebble tools, such as chopper-chopping tools,picks, and spheroids, were of secondary importance; directhammer percussion, the bipolar technique, and the block-on-block method were used for flaking; and the industries ex-hibited slow and conservative development during the EarlyPaleolithic but showed acceleration of technological devel-opment and innovation as well as the emergence of new tech-niques and tool types (such as blade and microblade toolsand technology) in the Late Paleolithic. Meanwhile, charac-teristics of SMI include the domination of large pebble tools,mainly chopper-chopping tools, picks, and handaxes; rarelypresent and poorly fabricated scrapers and points; hammer-percussion and block-on-block methods used for core re-duction; fewer clear tool types and coarser retouch comparedwith NMI; and a much stronger conservative developmentaltrend even all the way to the early Neolithic. On the basis ofthe partition of the north versus the south cultural zones andthe formation of many regional cultural variants, Zhangpointed out possible factors, such as environmental differ-ences and human migrations and interactions, but was shortof detailed analysis.

Simple stone-tool technology and assemblage and slow de-velopment of the Chinese LEP should not lead to the con-clusion that they are stagnant. In fact, changes and devel-opment in these industries are still evident through time.Zhang Senshui (1989) summarizes the major developmentaltrends of the Chinese Lower Paleolithic as follows. (1) Moreand more high-quality raw materials were exploited, especiallysilicified limestone and flint. (2) Direct hard-hammer per-cussion technique underwent a discernible process of mat-uration while block-on-block techniques diminished in im-portance. (3) Morphologically regular flakes and tools madeon them increased in number. (4) More flakes were used astool blanks, and even in the pebble-tool zone of southernChina, flake tools increased through time. (5) more tool typeswere added to the assemblages, and the discrepancy betweentool classes became clearer. (6) Chopper-chopping tools be-

S368 Current Anthropology Volume 54, Supplement 8, December 2013

came less common, and points and drills in accordance in-creased and became more regularized. (7) Through time,more delicately retouched tools increased in number. (8) Re-touch changed from multidirectional to mainly unidirectionalon the dorsal surface. Such technological and morphologicalchange or development is not as obvious and dramatic aswhat happened to the Early and Middle Stone Age or Lowerand Middle Paleolithic in the West. However, they provideus the opportunity to think about and reconstruct differentor alternative trajectories of human evolution during the re-mote past.

Discussion

For a long time, discussions of technological similarities anddifferences between Paleolithic tool traditions have been fre-quently based on assumptions regarding what biological af-finities may indicate about cultural ties and vice versa. Suchassumptions are not always borne out in the real world whenwe realize that biologically closely related populations some-times exhibit fundamental differences in their social structureand even language, and culture and technology can be trans-mitted quickly between geographically distant groups (Schick1994). Therefore, archaeologists must pursue alternative hy-potheses to explain cultural variability.

Other than the Movius Line theory, many hypotheses havebeen proposed to interpret the unique features of Paleolithicindustries in China and East Asia. A few researchers havesuggested that the main tools used by Paleolithic humans foradaptation in the region were those made from bamboo andwood materials, and the simple stone tools were actually usedto make such vegetal tools. Therefore, stone artifacts are notthe right indicator of human technological development andadaptive strategy in the Pleistocene (Pope 1989). Some pro-posed that the lack of high-quality raw material in East Asiawas the major obstacle for ancient populations living in theregion to develop more sophisticated lithic technology andmake stone tools as good as those of Africa and westernEurasia (Schick 1994). Some even suggested that when earlyhuman groups migrated into East Asia, they first encounteredtropical and subtropical ecological conditions in the southernpart of the territory. In such environments, foods were ob-tained mainly through gathering plant fruits and roots ratherthan hunting game; large hunting and butchering tools wereuseless, and simple pebble and flake tools took the dominantrole. Such a shift of survival conditions and adaptive strategiesbrought about fundamental changes to stone-tool technologyand the composition of tool kits, and the watershed in Pa-leolithic industries between the East and the West began toappear (Watanabe 1985).

The above hypotheses all offered some explanations on thedistinctive features of Paleolithic remains in China and EastAsia and cultural differences between the East and the Westin most of the Pleistocene. However, such discussions areoften confined to some isolated factors, such as geographic

isolation, restrictions of raw materials, and ecological/envi-ronmental conditions, and some theories are short of sup-porting evidence. The formation of a lithic industry or of acertain cultural tradition should be a complex process thatmight involve many influencing factors. Maybe it is time towork out a model that is more comprehensive and takes anintegrative approach to consider both environmental effectsand human behavior and adaptive strategies. I call it a “com-prehensive behavioral model.” The model may offer the fol-lowing observations and explanations on the stable devel-opment and unique features of Early Paleolithic industries inChina.

Stable Environments and Continuity of Human Evolution

During the LEP, China was under the control of monsoonclimates. Studies of Loess depositional sequence and faunalassemblages suggest that even if climatic fluctuations occurredperiodically, environmental conditions were relatively stablein the region, and most of the area was suitable for humanhabitation (Liu 2009). Rich and continuous archaeologicalrecords indicate that human evolution in the region was stableand uninterrupted and without large-scale population re-placement. Strong and stable cultural traditions were formedduring this process, and occasional outside intruders wereassimilated into the mainstream populations and lost theircultural identities.

Low-Intensity Resource Exploitation and High Mobility

Most of the LEP sites in China are seasonal, short-time oc-cupied ones, and artifacts collected from them are mostlysimple and share basic features of technology, typology, andmorphology, which may indicate that human groups livingin the region had a simple and “easy” hunting-gathering life-style. They kept the exploitation of natural resources at arather low intensity and seldom felt the pressure of innovatinglithic technology to procure difficult resources. In keepingsuch a lifeway, they moved frequently to other places to findnew food resources, and therefore they left identical artifactsand other materials at many sites in certain regions.

High Flexibility in Tool Technology and Adaptation

The lack of high-quality stone raw materials and suitable quar-rying sites forced Early Paleolithic humans living in the areato make best use of poor-quality and locally available rawmaterials. In dealing with such materials with great variabilityin lithology and morphology, these people learned to be highlyflexible and use simple but suitable and effective ways toproduce tool blanks and make stone tools. For instance, peo-ple living at Zhoukoudian Locality 15 relied heavily on thebipolar technique to make use of quartz nodules that wereabundant in the nearby river bed. People living at the Dingcunsite applied the hammer-percussion method skillfully to darkhornfels, a relatively high-quality local raw material, to detach

Gao Paleolithic Cultures in China S369

large and regular flakes and to produce large and sharp cuttingand digging tools. Still, people in the Sanxia region inventedthe unique “throw and collision method” to exploit highlypolished and rounded river pebbles. Such flexibility mighthave enforced ancient human survival capability and helpedthe Paleolithic traditions to be strong, stable, and full of vi-tality. If the Bamboo Tool hypothesis can be verified, it willbe a good example of the flexibility and intelligence of ancienthumans living in the region.

Conclusions

It is obvious that the LEP industries in China and East Asiaare different from those of contemporary Paleolithic remainsin Africa and western Eurasia in many ways. However, it isalso clear that the Lower Paleolithic world should not besimply divided into two different cultural/technologicaltraditions based solely on stone-tool technological and ty-pological comparisons. The variation in Paleolithic industriesbetween the West and the East is undeniable, but it shouldbe understood within a broad framework of universal culturaldiversity. It should be realized that while ancient hominidsin different parts of the world shared some basic lithic tech-nologies and produced and utilized similar stone tools (suchas core/flake tools), each group was unique in its methods ofsurvival and adaptation because of ecological context and rawmaterial availability and quality. What is more important isto look beyond lithic technological and typological variabilityand find the factors and dynamics behind such cultural dif-ferences and reconstruct different pathways of human evo-lution toward complexity and modernity.

Acknowledgments

I would like to thank Steve Kuhn and Erella Hovers for theirinvitation to the symposium “Alternative Pathways to Com-plexity: Evolutionary Trajectories in the Middle Paleolithicand Middle Stone Age” and their suggestions for and helpwith the revision of this paper. I also thank Leslie Aiello andLaurie Obbink for their organization of the symposium andtheir assistance. The research that led to the writing of thispaper was supported by the Chinese Academy of SciencesStrategic Priority Research Program (XDA05130202) and theMinistry of Science and Technology of China GroundworkProject (2007FY110200).

References CitedChen, Tiemei, Yang Quan, and Wu En. 1994. Antiquity of Homo sapiens in

China. Nature 368:55–56.Chen, Tiemei, Yuan SiXun, and Gao ShiJun. 1984. The study on uranium-

series dating of fossil bones as an absolute age sequence for the mainPaleolithic sites of North China. Acta Anthropologica Sinica 3(2):259–269.

Chen, Tiemei, Yuan SiXun, Gao ShiJun, Wang LiangXun, and Zhao GuiYing.1982. U-series dating on the mammalian fossils from the Xujiayao site.Acta Anthropologica Sinica 1(1):91–95.

Dennell, Robin. 2009. The Paleolithic settlement of Asia. New York: CambridgeUniversity Press.

Gao, Xing. 1999. A discussion of the “Chinese Middle Paleolithic.” Acta An-thropologica Sinica 18(1):1–16.

———. 2000. Interpretations of typological variability within Paleolithic re-mains from Zhoukoudian Locality 15, China. PhD dissertation, Universityof Arizona, Tucson.

———. 2003. A study of the lithic assemblage from Zhoukoudian Locality15. Acta Anthropologica Sinica 21(suppl.):31–52.

———. 2011. The nature of Paleolithic handaxes from China and its impli-cations for Lower Paleolithic cultural variation. In Handaxes in the ImjinBasin: diversity and variability in the East Asian Paleolithic. Seonbok Yi, ed.Pp. 193–217. Seoul: Seoul National University Press.

Gao, Xing, and Christopher J. Norton. 2002. A critique of the Chinese “MiddlePaleolithic.” Antiquity 76:397–412.

Gao, Xing, and John W. Olsen. 1997. Similarity and variation within the LowerPaleolithic: East Asia, western Europe and Africa compared. In Evidence forevolution. T. Yongsheng, ed. Pp. 63–76. Beijing: Ocean Press.

Gao, Xing, and Pei ShuWen. 2010. Footprints of ancient humans in the ThreeGorges region. Chengdu: Sichuan Publishing.

Gao, Xing, Zhang XiaoLing, Yang DongYa, Shen Chen, and Wu XinZhi. 2010.Revisiting the origin of modern humans in China and its implications forglobal human evolution. Science China (Earth Sciences) 40(9):1287–1300.

Huang, WeiWen. 1989a. Bifaces in China. Human Evolution 4(1):87–92.———. 1989b. The early Paleolithic of China. Quaternary Research 28:237–

242.———. 1993. Typology of Lower Paleolithic heavy-duty implements from

East and Southeast Asia: comments on the Movius’s typological system.Acta Anthropologica Sinica 12(4):297–304.

Huang, WeiWen, Hou YaMei, and Gao LiHong. 2009. “Western elements” inthe Chinese Paleolithic as viewed in a framework of early human culturalevolution. Acta Anthropologica Sinica 28(1):16–25.

Huang, WeWen, Hou YaMei, and Si XinQiang. 1997. Stone industry fromPanxian Dadong, a cave-site of southwestern China. Acta AnthropologicaSinica 16(3):171–192.

Ikawa-Smith, F., ed. 1978. Early Paleolithic in South and East Asia. The Hague:Mouton.

Jia, LanPo, Gai Pei, and You YuZhu. 1972. Excavation report on the Shiyusite in Shanxi. Acta Archaeologica Sinica (1):39–58.

Jia, LanPo, Wei, Qi, and Li, ChaoRong. 1979. Report on the excavation ofthe Hsuchiayao Man site in 1976. Vertebrata PalAsiatica 17:277–293. [InChinese.]

Jia, LanPo, and WeiWen Huang. 1985. On the recognition of China’s Pa-leolithic cultural traditions. In Paleoanthropology and Paleolithic archaeologyin the People’s Republic of China. Wu RuKang and John W. Olsen, eds. Pp.259–265. New York: Academic Press.

Li, YanXian, and Wen BenHeng. 1986. Guanyindong: a Lower Paleolithic sitein Qianxi county, Guizhou Province. Beijing: Cultural Relic Press.

Li, ZhanYang. 2007. A preliminary study on the stone artifacts of Lingjingsite excavated in 2005. Acta Anthropologica Sinica 26(2):138–154.

Li, ZhanYang, and Shen Chen. 2010. Use-wear analysis confirms the use ofPalaeolithic bone tools by the Lingjing Xuchang early human. Chinese Sci-ence Bulletin 55:2282–2289.

Lin, ShengLong. 1994. Restudy of nine hand-axe specimens and the appli-cability of Movius’ theory. Acta Anthropologica Sinica 13(3):189–208.

———. 1996. A comparative study on the Paleolithic technological modesbetween China and the West. Acta Anthropologica Sinica 15(1):1–20.

Liu, C., P. Su, and Z. Jin. 1992. Discovery of Blake Episode in the XujiayaoPaleolithic site, Shanxi, China. Scientia Geologica Sinica 1:87–95.

Liu, DeYin, and Wang YouPing. 2001. A preliminary report on the excavationof Jigongshan site. Acta Anthropologie Sinica 20(2):102–114.

Liu, DongSheng. 2009. Loess and arid environment. Hefei, China: Anhui Sci-entific and Technological Press.

Lu, HuaYu, Zhang HongYan, Wang SheJiang, Zhao CunFa, and Zhao Jun.2007. A preliminary survey on loess deposit in eastern Qinling Mountainin central China and its implication for estimating age of the Pleistocenelithic artifacts. Quaternary Sciences 27(4):550–567.

Lu Zune. 2004. Paleolithic archaeology in Liaoning Province. In Chinese ar-chaeological research in the 20th century (Paleolithic volume). Lu Zune, ed.Pp. 194–219. Beijing: Science Press. [In Chinese.]

Ma, Ning, Pei ShuWen, and Gao Xing. 2011. A preliminary study on thestone artifacts excavated from Locality 74093 of the Xijiayao Site in 1977.Acta Anthropologica Sinica 30(3):275–288.

S370 Current Anthropology Volume 54, Supplement 8, December 2013

Movius, H. L. 1944. Early man and Pleistocene stratigraphy in southern andeastern Asia. Papers of the Peabody Museum of American Archaeology andEthnology, vol. 19, no. 2. Cambridge, MA: Peabody Museum.

———. 1948. The Lower Paleolithic cultures of southern and eastern Asia.Transactions of the American Philosophical Society 38(4):329–420.

Nagatomo, Tsuneto, Yorinao Shitaoka, Hisae Namioka, Masatoshi Sagawa,and Wei Qi. 2009. OSL dating of the strata at Paleolithic sites in the NihewanBasin, China. Acta Anthropologica Sinica 28(3):276–284.

Norton, Christopher J., Gao Xing, and Feng XingWu. 2009. The East AsianMiddle Paleolithic reexamined. In Sourcebook of Paleolithic transition. MartaCamps and Parth Chauhan, eds. Pp. 245–254. Dordrecht: Springer.

Pei, ShuWen, Zhang JiaFu, Gao Xing, Zhou LiPing, Feng XingWu, and ChenFuYou. 2006. Optical dating of the Jingshuiwan Paleolithic site of ThreeGorges, China. Chinese Science Bulletin 51(11):1334–1342.

Pope, Geoffrey G. 1989. Bamboo and human evolution. Natural History89(10):48–57.

Schick, Kathy D. 1994. The Movius line reconsidered: perspectives on theearlier Paleolithic of eastern Asia. In Integrative paths to the past. Robert S.Corruccini and Russell L. Ciochon, eds. Pp. 569–596. Englewood Cliffs,NJ: Prentice Hall.

Schick, Kathy D., and Z. Dong. 1993. Early Paleolithic of China and easternAsia. Evolutionary Anthropology 2(1):22–35.

Shen, GuanJun, and Jin LinHong. 1992. U-series dating of speleothem samplesfrom Guanyindong Cave at Qianxi County, Guizhou Province. Acta An-thropologica Sinica 11(1):93–100.

Wang, Shejiang. 2005. Perspectives on hominid behaviour and settlement pat-terns: a study of the Lower Paleolithic sites in the Luonan Basin, China. Oxford:British Archaeological Reports.

Wang, Wei, Liu Jun, Hou YaMei, Lynne A. Shepartz, S. Miller-Antonio, W. J.

Rink, Xin-quiang Si, and Wei-wen Huang. 2003. Stratigraphic and paleoen-vironmental studies at the Dadong Cave, Panxian. Acta Anthropologica Sin-ica 22(2):131–138.

Watanabe H. 1985. The chopper-chopping tool complex of eastern Asia: anethnoarchaeological-ecological reexamination. Journal of AnthropologicalArchaeology 4:1–18.

Wei HaiBo. 2009. Continued research on the Miaohoushan site, LiaoningProvince. Acta Anthropologica Sinica 28(2):154–161.

Wu, XinZhi. 2009. A metrical study of the Dali cranium. Acta AnthropologicaSinica 28(3):217–236.

Yi, S., and G. A. Clark. 1983. Observations on the Lower Paleolithic of north-east Asia. Current Anthropology 24(2):181–202.

Yin, GongMing, Sun Yingjie, Ye YuGuang, and Liu Wu. 2001. The ESR ageof shells from the bed of Dali with human fossil. Acta Anthropologica Sinica20:34–38.

Zhang Li, Shen Guanjun, Fu Renyi, et al. 2007. Preliminary U-series datingof hominid locality Miaohoushan in Benxi, Liaoning Province. SoutheastCulture 2007(3):54–57.

Zhang, SenShui. 1989. The Lower Paleolithic cultures in North China. InAncient humans in China. Wu Rukang, Wu Xinzhi, and Zhang Senshui,eds. Pp. 97–158. Beijing: Science Press.

———. 1993. A study on the stone artifacts from Loc. 54:100 of the Dingcunsite. Acta Anthropologica Sinica 12(3):195–213.

———. 1999. On the important advancement of the Paleolithic archaeologyin China since 1949. Acta Anthropologica Sinica 18(3):193–214.

Zhang, ShuangQuan, Li ZhanYang, Zhang Ye, and Gao Xing. 2009. Mortalityprofiles of the large herbivores from the Lingjing, Xuchang man site, HenanProvince, and the early emergence of the modern human behaviors in EastAsia. Chinese Science Bulletin 54:3857–3863.

Current Anthropology Volume 54, Supplement 8, December 2013 S371

� 2013 by The Wenner-Gren Foundation for Anthropological Research. All rights reserved. 0011-3204/2013/54S8-0015$10.00. DOI: 10.1086/673388

Identifying Mechanisms behind MiddlePaleolithic and Middle Stone Age

Cultural Trajectories

by Francesco d’Errico and William E. Banks

A critical analysis of the debate that has surrounded the emergence of “modern behavior” during the last twodecades and new ways to study material culture and human-environment relationships allow us to design a novelapproach with which we can understand the mechanisms that have led human populations to develop the varietyof cultures that we recognize today. We propose a methodological framework that moves away from narrativeexplanations for the origin of “behavioral modernity” and instead focuses on the interplay between cultural adaptationand environmental change. We argue that by applying this approach to the many different instances of culturalchange as well as stasis that characterized the last 300 kyr of human societies we may identify the mechanisms thathave led us to become what we are and, if any, the underlying trends that guided this process.

Introduction

Twenty years ago, the path followed by humans to attain“behavioral modernity” appeared evident to most archaeol-ogists and paleoanthropologists. Best exemplified by the pub-lication of The Human Revolution monograph (Mellars andStringer 1989), this path was short, abrupt, exclusively as-sociated with anatomically modern humans (AMHs), and bestreflected in the European Upper Paleolithic archaeologicalrecord. However, apart from a possible neurological switch(Klein 2000, 2009), no solid cause was proposed that couldexplain why this should have happened where it did and insuch an instantaneous way. Subsequently, the discovery thatAMHs originated in Africa (Henn et al. 2011; Trinkaus 2005;Weaver and Roseman 2008) along with a growing body ofarchaeological evidence supporting the emergence of key cul-tural innovations in that continent before the purported Eu-ropean “revolution” ca. 40 ka gave rise to a different explan-atory model: “modern behavior” must have developedgradually in Africa as a consequence of the origin of ourspecies there and would have been expressed by a process ofgradual accretion of innovations observed in the African ar-

Francesco d’Errico is Director of Research and William E. Banksis Research Associate at the University of Bordeaux, Centre Nationalde la Recherche Scientifique-Unite Mixte de Recherche de laPrehistoire a l’Actuel: Culture, Environnement, Anthropologie,Equipe Prehistoire, Paleoenvironnement, Patrimoine (BatimentB18, Avenue des Facultes, 33405 Talence, France [f.derrico@pacea.u-bordeaux1.fr]). This paper was submitted 3 VII 13, accepted 1VIII 13, and electronically published 8 XI 13.

chaeological records over the past 300 kyr (Marean et al. 2007;McBrearty and Brooks 2000). The idea that the emergenceof cultures such as ours was abrupt nevertheless remained inplay, and some viewed innovations occurring in southernAfrica ca. 70 ka as the factor that allowed cognitively modernAMHs to qualitatively change their adaptive abilities and rap-idly expand out of that continent ca. 60 ka (Mellars 2006).In parallel, other researchers proposed that “modern” cog-nition was associated with various members of our lineage,not just AMHs, and that social and demographic factors,arguably triggered by climate change, could explain the asyn-chronous emergence, disappearance, and reemergence of keycultural innovations among both African Middle Stone Ageand Eurasian Middle Paleolithic populations (Conard 2008;d’Errico 2003; d’Errico and Stringer 2011; Hovers and Belfer-Cohen 2006; Langley, Clarkson, and Ulm 2008; Nowell 2010;Zilhao 2001, 2007). The partisans of this model relied oncultural innovations found in the Neanderthal archaeologicalrecord (burials, use of pigments, complex lithic and haftingtechnologies, and personal ornamentation at the end of theNeanderthal evolutionary trajectory) to counter the idea thatbehavioral modernity is unique to our species. The recentfinding that significant interbreeding occurred between Ne-anderthals, Denisovan, and modern populations originatingin Africa (Green et al. 2010; Meyer et al. 2012; Reich et al.2010) is used to support such a scenario because it blurspreviously accepted taxonomic boundaries between UpperPleistocene hominins. In parallel with these scenarios, whichare in one way or another anchored in the archaeological andgenetic records, some researchers have proposed that the evo-lution of inherent components of present-day modern cog-

S372 Current Anthropology Volume 54, Supplement 8, December 2013

nition may have played a key role in reaching the cognitivecapacity expressed in our species and the ensuing “complex-ification” of material culture. Altruism (Bowles 2009), en-hanced memory (Wynn and Coolidge 2010), complex lan-guage (Dunbar 2003), increased capacity for social learning(Mesoudi, Whiten, and Laland 2006; Richerson, Boyd, andBettinger 2009; Tomasello 1994), creation of adapted learningenvironments (Sterelny 2011), hierarchical mental construc-tions (Gibson 2007), and acquisition of syntactical language(Bickerton 1990) have each been proposed as the prime moverthat allowed AMHs to cross the threshold of behavioral mo-dernity. It has been argued, though, that the capacity for“modernity” resulting from a speciation event is a necessarybut not sufficient condition for the development of moderncultural traits. In this vein, a number of scholars have exploredthe roles played by population pressure (Compton 2011) aswell as demography and cultural transmission (Henrich 2004;Powell, Shennan, and Thomas 2009; Shennan 2001) in thespread and maintenance of cultural innovations. By applyingmodeling techniques, these latter authors argue that popu-lation size, density, rates of reproduction, and networks ofinformation exchange determine whether or not cultural in-novations can be diffused, maintained, and in some cases lost.The interest of their results is that they account for such eventswithout invoking speciation as a prime mover and that theirexpectation better fits the archaeological record than modelspredicting abrupt or incremental changes in behavior. How-ever, it has been argued that while demography may be keyto the diffusion and maintenance of innovative behaviors, itdoes not necessarily favor their emergence and acceptance.Each society has its own way of regulating the acceptance ofdeviations from a behavioral norm, which produces differentpostures toward innovation even when potentially advanta-geous (Bar-Yosef and Belfer-Cohen 2011). The spread of in-novations is also dependent on a society’s ability to createsettings of high-fidelity learning: even with sufficient popu-lation size, if such settings are absent, innovations may notbe maintained (Sterelny 2011). Another shortcoming, aspointed out by d’Errico and Stringer (2011), is that Powell,Shennan, and Thomas (2009) signal climate change as a ge-neric factor behind the increase or fragmentation of popu-lations in their demographic-based “mechanism,” but theyoffer no means of testing this, especially when consideringthat specific climate changes had very different effects in var-ious regions of the world. Implicit in their (Powell, Shennan,and Thomas 2009) contention that favors demography is theassumption that in order to produce the spread and main-tenance of innovations, this demographic-based mechanismis only relevant to AMH populations. It has been noticed thatcontrary to their assumption, such a mechanism is equallyapplicable to archaic hominins and that differences betweencultural trajectories of Neanderthals and AMHs may havebeen dependent on group size and rates of cultural exchangerather than on built-in differences in cognition (d’Errico andHenshilwood 2011). These criticisms serve to show that de-

mography is important but that it is only one factor in thespread and maintenance of innovations, not the prime movernor the unique explanatory mechanism behind their adop-tion.

In parallel with the proposition of these various explanatorymodels and triggers, a number of researchers have called intoquestion whether behavioral modernity is the appropriateconcept with which to identify when we became as we are.One such criticism has been directed at the reliance on ma-terial culture trait lists to recognize the crossing of the thresh-old to modernity. An early example of the use of a trait listis by Mellars (1989), who used disjointed elements of theUpper Paleolithic archaeological record to establish a sup-posed cognitive threshold between Neanderthal and AMH,suggesting that the latter had crossed the Rubicon to behav-ioral modernity by the beginning of the Upper Paleolithic.However, Deacon (1990) cautioned against the use of a Eu-rocentric yardstick for measuring cognitive modernity bypointing out that there were indications that these criteriawere not pertinent in other regions of the world, where thisprocess appeared to have followed a different tempo andmode. After the publication of the more comprehensive traitlist proposed by McBrearty and Brooks (2000), d’Errico(2003) noted that instead of being based on universal featuresfound in all present-day human cultures, this list was a mix-ture of traits that the authors had recognized in the MiddleStone Age and the Upper Paleolithic. This naturally led themto recognize behavioral modernity in the records on whichtheir list was based and create the misleading impression ofa gradual accretion of these traits in Africa. In his view thesetrait lists cannot necessarily be applied to human populationsthat lived in dramatically different environments and that haddifferent cultural trajectories. While proposing that behavioralmodernity played a useful role in highlighting the inconsis-tencies of the “human revolution” model and showing thatmore gradualistic patterns of cultural change existed outsideof Europe, d’Errico (2003) argued that behavioral modernitywas no longer a useful operational concept and that oneshould instead explore the emergence of cultural innovationin each region of the world irrespective of the taxonomicaffiliation of the population in question.

Zilhao (2011) points out that the real problem with be-havioral modernity is that it is based on the notion that dif-ferent species of Homo were each characterized by a uniqueset of behaviors and that modern behavior is unique to AMHs.He argues that attempts to define “modern human behavior”in opposition to “Neanderthal behavior” have consistentlyshown that some modern humans were “behaviorally Ne-anderthal” and some Neanderthals were “behaviorally mod-ern” (Zilhao 2006). In his view, there is no such thing asNeanderthal behavior because the archaeological record inboth Europe and the Near East demonstrates that Neanderthaladaptations span the entire range of ethnographically docu-mented settlement-subsistence strategies. Nowell (2010) alsohighlights the concept of modernity’s ambiguous status due

d’Errico and Banks Mechanisms behind Cultural Trajectories S373

to its routine association with the development of AMH so-cieties, which is at odds with the presence of many instancesof modernity in Neanderthal populations.

Henshilwood and Marean (2003) argue that behavioral mo-dernity is too loose a concept and that we should rely onclear evidence for symbolically mediated actions as the bestway to recognize societies that have acquired “fully symbolicsapiens behavior.” However, when applying this concept tothe archaeological record of southern Central Africa, Barham(2007) highlights how difficult it may be to identify the emer-gence of such a capacity in records that hint at the presenceof modernity but do not have the full suite of material culturethat supports its “formal” recognition. This problem may bedue to our lack of heuristic tools for recognizing fully symbolicbehavior when expressed in ways other than our preconceivednotions. His answer to this shortfall is to combine structuralistand ecological theory to recognize individual regional trajec-tories that could be influenced in part by historical contin-gencies. Belfer-Cohen and Hovers (2010) concur in that theyconsider behavioral modernity to be a loose and poorly op-erational concept that does not create a link between cognitivetheory and the typical work of archaeologists. They argue thatmodernity is multifaceted and cannot be boiled down to arigid checklist of presence/absence characteristics, and theypropose that we should focus instead on the circumstancesand contexts in which the phenotypic expressions of moder-nity, which were at first erratic, became fixed features in thearchaeological records of different regions of the world.

Additional criticism has been leveled at this concept byLangbroek (2011), who proposes that the concept of moder-nity as it is typically applied leads us to frame the evolutionof cognition in a unilinear and exponential manner when weshould rather envision this process as a branching one thatmirrors Darwinian evolution. Within such a framework, “cog-nitions” associated with different members of our lineagewould be submitted to varying selective pressures, therebycreating a variety of cognitive outcomes that cannot be clas-sified as more or less modern.

Shea (2011), like many others, criticizes the fact that be-havioral modernity is derived from a trait list founded on theUpper Paleolithic archaeological record, and he echoesd’Errico’s (2003:189) argument that archaeologists from dif-ferent cultural backgrounds could propose different featuresto define modernity. Furthermore, he adds that the concepthas been constructed on a paleoanthropological narrative tra-dition that implies the transformation from an inferior stateto a superior one. His answer to overcome the behavioralmodernity conundrum is to focus on “behavioral variability.”The idea is to take into account “modality, variance, skew,and other quantitative/statistical properties” (Shea 2011:2) tomeasure how successful specific cultural adaptations werefrom a cost-benefit standpoint in specific environmental set-tings. Shea’s view is anchored in the notion that whateverapproach one takes to address this issue, it is solely the busi-ness of Homo sapiens sapiens and does not concern archaic

hominins perceived by definition as a behaviorally differentbiological species and therefore irrelevant in this debate.

Comments on Shea’s paper (2011), which was publishedin Current Anthropology, welcome his proposed shift towardthe reconstruction of different cultural adaptations in specificenvironmental situations but point out that his search forbehavioral variability does not go very far. Comments byNicholas J. Conard and Rick Potts in particular highlight thefact that behavioral variability, as Shea describes it, is as vagueas behavioral modernity because he does not propose withthe concept any clear operational tools with which one canevaluate cultural adaptations within their respective environ-mental contexts. The tool he used, Clark’s system of tech-nological modes, appears to them as highly inappropriatebecause it is constructed on a logic of unilinear, incrementalevolution, which is what Shea (2011) is in principle trying tosteer away from. We would add another remark: by placingthe analysis of behavioral modernity in the realm of what heperceives as a single species, Shea has unwittingly found theprime mover or “donor” of his underlying definition of be-havioral modernity—the processes that led to the emergenceof our species in Africa.

From Pars destruens to Pars construens

The above review of the debate makes clear that behavioralmodernity and varying lists of cultural traits associated withit are not useful tools for establishing the way in which webecame what we are. Some consensus now exists that theevolution of human societies in the last 300 kyr has followeda multitude of paths, not necessarily progressive in nature, inwhich the material expression of modern cognition is rep-resented by different mosaics of cultural innovations. Focus-ing on regional trajectories appears to be the only way todocument cultural changes and ultimately the mechanismsbehind such changes. In doing so, we must seek ways tointegrate environmental, ecological, demographic, and socialfactors as well as historical contingencies in order to under-stand how human populations have developed and in somecases lost and reacquired cultural innovations that we rec-ognize to be the cornerstone of the human experience. Amongthose that accept this frame of thinking (d’Errico 2009; Hovers2009; Kuhn 2013; Stiner 2013; and comments in Shea 2011by Lawrence S. Barham, Nicholas J. Conard, James F.O’Connell, and Rick Potts), there is consensus that althoughthese are factors that played a role in the process of culturalinnovation, the way that they were organized and the interplaybetween them remains to be understood, and pertinent heu-ristic tools with which to interrogate the empirical evidenceare lacking.

Some might still question whether this endeavor should beconducted only in archaeological records associated withAMHs or should also include archaic hominins. The formerwould be a clear mistake in our view because it would con-strain, whether one would admit it or not, the analysis of

S374 Current Anthropology Volume 54, Supplement 8, December 2013

local trajectories within a conceptual framework in which keybehavioral innovations can only be the consequence of an-atomical modernity. By equating behavioral and anatomicalmodernity, no matter what “variability” or “cost-benefit” bal-ance is found, we will not get rid of the mind-set that bio-logical change is the prime mover. Such a stance would alsodeprive us of examining a significant number of cultural tra-jectories, thus compromising our ability to compare how dif-ferent populations reacted to comparable suites of externalstimuli. Encompassing all local trajectories is in our view thebest way to obtain a full picture of the many rapid culturalexperiments that are key features of the cultural evolution ofour lineage.

With this paper, we propose a methodological frameworkthat moves away from “narrative” explanations toward a focuson material culture and the evaluation of the potential in-terplay between cultural adaptation and environmentalchange. We think that by applying this approach to the manydifferent instances of cultural change as well as stasis thatcharacterize the last 300 kyr of human societies, we may iden-tify the mechanisms that have led us to become what we areand the underlying trends, if any, that guided this process.

Causes versus Mechanisms

We argue that the primary problem with most of the scenariosproposed to explain the emergence of behavioral modernityis that they are based on single-cause models. Such modelsare founded on the teleological notion that a unique causewill continuously act as the sole or the dominant factor inproducing the observed outcome. Some of the proposedcauses would have acted as long-term stimuli, such as altruismor enhanced working memory. Others, such as genetic mu-tation or a bottleneck event (e.g., Toba super eruption: Am-brose 1998), would have been short lived in nature and havehad a relatively immediate effect. Both types of causes, how-ever, are not sufficient to explain the complex paths and mul-titude of adaptations that a growing body of archaeologicaldata from Africa and Eurasia denotes.

Rather than causes, we need to focus on identifying themechanisms that have led different societies to develop spe-cific cultural adaptations as a means of coping with externalstimuli (both environmental and cultural). “Mechanism” isa term that has been defined in a number of ways, and Ku-orikoski (2009) points out that it is difficult to come up witha definition that satisfies all theoretical needs and potentialresearch practices. Drawing from the work of Bechtel andAbrahamsen (2005), Kuorikoski (2009), and MacHamer, Dar-den, and Craver (2000), we define a mechanism as a con-stellation of factors and components that through the processof their interaction with one another stimulates the trajectoryof a system. The investigation of mechanisms functions attwo different conceptual levels (Kuorikoski 2009). The firstconsists of examining a componential causal system by dis-assembling the role played by each component and factoring

in the multitude of interactions that occur between them. Thesecond is more abstract and seeks to encompass interactionsbetween factors and components with the goal of explainingsuch interactions with a simple model. The final goal is tomove from complexity to the proposition of a general ex-planatory law (e.g., the mechanism of natural selection inbiological evolution).

When put in the context of the debate surrounding MiddleStone Age and Middle Paleolithic cultural trajectories, the firstconcept of mechanism can be seen as a useful operationaltool, and the second points to the ultimate goal of identifyingthe long-term trends and rules that have shaped the culturalevolution of our lineage. One of the interests in approachingcultural evolution from the standpoint of mechanism is thatwe can describe cultural change dynamically through the anal-ysis of setup and termination conditions (Machamer, Darden,and Craver 2000). The former typically represent in the in-vestigation of mechanism the relevant components, their“structural” properties, spatial relations, and the causal factorsthought to influence the relationships between different com-ponents. Each component has some degree of variability andindependence such that identical setup conditions can resultin two systems following different trajectories. Between setupand termination, intervening factors or entities can influencethe interactions between components within the mechanism,thus influencing the direction the system follows. Terminationconditions are idealized states that represent a point fromwhich one can infer how the mechanism functioned. In thiscontext, “termination” is neither synonymous with equilib-rium nor a moment in which the process has necessarilyreached a terminal state. In reality, such states are idealizedconditions, and their choice by researchers is often deter-mined by the heuristic and analytical tools they have at theirdisposal to understand the interplay between components andcausal factors within a perceivable time frame. This movebetween setup and termination is analogous to the conceptof “adaptive cycle” that is used in resilience theory to char-acterize the dynamics of a socioecological system with respectto external stimuli and internal processes (Holling 1973;Schoon et al. 2011; Walker et al. 2004).

Setup conditions are of course the result of prior processes,so in the study of mechanisms in a componential causal sys-tem framework one is beginning the investigation at a specificpoint along a continuum. Human cultures follow continuoustrajectories in which the various factors that play a role inproducing change continuously interact. However, as archae-ologists, we identify in the archaeological record discrete andrecurrent associations of similar cultural items assumed toreflect cohesive adaptive systems (CASs). We define a CAS asa cultural entity characterized by shared and transmittedknowledge reflected by a recognizable suite of cultural traitsthat a population uses to operate within both cultural andenvironmental contexts. The cultural traits used to define aCAS can carry the same well-known ambiguities as those usedto define a techno-complex (Clarke 1968; Hodder 1991; Ren-

d’Errico and Banks Mechanisms behind Cultural Trajectories S375

frew 1977). The archaeological record represents a very palereflection of the features that constituted a past cohesive cul-tural group because human adaptation is shaped by culturalrules that govern aspects of human behavior such as kinshipsystems, marriage, gender politics, and symbolic or belief sys-tems. These rules define human societies as much as andprobably more than the environmental contexts they occupy.As archaeologists we have limited means with which to inferthose rules for societies of the remote past. It is, however,reasonable to think that those rules, or changes in those rules,shaped the material culture record that we have at our disposaland played a role in how a past group interacted with theirenvironment. The concept of CAS differs from that of atechno-complex (Childe 1929) in that environmental param-eters contribute to its definition. The difference lies in thefact that the approach that we detail below effectively explorespotential links between cultural traits and ecological param-eters and evaluates this relationship through time, therebyproviding us the potential to identify general, long-termtrends.

Traditionally, archaeologists working with the Middle Pa-leolithic and Middle Stone Age archaeological records haverelied on lithic technology and stone tool typology to definepast “cultures” and identify evolutionary trends. Regardlessof the multitude of reasons for this focus, examinations ofethnographically documented material cultures indicate thatstone tools represent just a small portion of the paraphernaliaused by a hunter-gatherer population and do not necessarilyreflect the complexity of cultural adaptation or its geography(Hayden 1979). Relying on just a single element of a technicalsystem to represent or infer the complex suite of behaviorsand social rules that characterize a past cultural adaption isclearly illusory. Data accumulated over the last decade in Af-rica and Eurasia on populations that lived there during thelast 300 kyr have broadened our understanding of these CASsby providing insight into a variety of domains beyond thosestrictly related to lithic technology. These include technolog-ical behaviors that certainly are expressions of salient featuresof those archaeological cultures. These behaviors include py-rotechnology (Brown et al. 2009; Mourre, Villa, and Hen-shilwood 2010), mastic production (Carciumaru 2012; Char-rie-Duhaut et al. 2013; d’Errico et al. 2012; Lombard 2012;Pawlik and Thissen 2011; Villa et al. 2012; Wadley, Hodgskiss,and Grant 2009), hafting techniques (Lombard 2005; Villa etal. 2009, 2012), projectile technology (Villa and Soriano 2010),techniques for small game capture (Stiner, Munro, and Su-rovell 2000; Wadley 2010c), use of poison in hunting (d’Erricoet al. 2012), bone tool production (Backwell, d’Errico, andWadley 2008; d’Errico, Backwell, and Wadley 2012; d’Errico,Borgia, and Ronchitellli 2012; d’Errico and Henshilwood2007), pigment processing and storage (d’Errico et al. 2010;Henshilwood et al. 2011), and use of plants (Mercader 2009;Wadley et al. 2011). Symbolically mediated behavior, whichappeared to be largely inexistent or largely out of our graspa decade ago for the time periods in question, is now well

documented and seems to be clustered in regional traditions.This is the case for personal ornamentation (Caron et al. 2011;d’Errico et al. 2009; Peresani et al. 2011; Vanhaeren et al.2006, 2013; Zilhao et al. 2010), symbolic use of pigments(Roebroeks et al. 2012; Watts 2010; Zilhao et al. 2010), graphicexpressions (d’Errico, Garcıa Moreno, and Rifkin 2012; Hen-shilwood, d’Errico, and Watts 2009; Mackay and Welz 2008;Texier et al. 2010), and mortuary practices (Grun et al. 2005;Pettitt 2011). Although open to debate, we may assume thatsuch CASs also included behavioral features that have notsurvived in the archaeological record and that made each ofthese societies unique in the cultural history of our lineage.

As discussed earlier, previous models proposed to explainthe emergence of these cultural features and innovations havetypically been monocausal in nature and have not been gearedtoward identifying potential mechanisms and long-termtrends. The evolution of a human society cannot be reducedto its demography; systems for transmitting and maintainingcultural innovations depend on a variety of factors. Recentstudies have pointed out that a number of ecological, his-torical, and psychological variables appear to condition therules that societies impose on individuals and the degree oftolerance a society has toward deviant behavior (Gelfand etal. 2011; Henrich, Heine, and Norenzayan 2010; Norenzayan2011). The cultural system that one inherits affects basic pro-cesses such as perception, reasoning, motivation, and coop-erative strategies. This seems to imply that each CAS is char-acterized by a different potential for cultural transmission,social learning, and the degree to which individuals are opento accept, maintain, and communicate innovations (Gelfandet al. 2011). It has also been argued that each individualsociety’s ability to maintain effective social learning environ-ments is key (Sterelny 2011). It is still unclear whether thereasons behind human behavioral variability are due to purelycultural processes leading to cultural divergence, to ecologicalconstraints, to gene-culture coevolution (Norenzayan 2011),or to some combination of these. Whatever the reason forthese differences, each society can be seen as a complex systemof attitudes and the potential they offer for change. In thiscontext, we use attitude to refer to the way in which thecollective worldviews of individuals in a social group influencebehavior. Bar-Yosef and Belfer-Cohen (2011) use such a ra-tionale to explore patterns of human expansion out of Africawith the debatable assumption that lithic technology is a fairreflection of these attitudes and therefore can be used as aproxy to trace successes and failures in hominin expansions.We argue that when identified and combined with the geo-graphic and environmental settings in which societies oper-ated, an array of behaviors related to technical and symbolicsystems and reflecting inherited knowledge can be viewed asthe setup conditions for the processes at work behind thesuite of cultural experiments that took place in regions ofAfrica and Eurasia between 300 ka and 10 ka.

The approach that we describe below entails means withwhich one can identify and follow the processes affecting CASs

S376 Current Anthropology Volume 54, Supplement 8, December 2013

between setup and termination conditions with no a prioriassumption on the trajectories followed by the system beingexamined. This approach is chosen to avoid the determinisminherent in a “single-cause” frame of thinking as well as therelativism implied in the behavioral variability approach. Weperceive the different regional trajectories as unique suites ofcultural experiments with their own distinct setup conditions.While these experiments may have components and processesin common, they did not necessarily play the same role withineach adaptive system.

Ideally in our field of study, understanding the processesat work between setup and termination conditions wouldrequire one to understand the role played by each componentand factor in a given region and time—environmentalchanges, adaptive system-specific material culture character-istics, inferred differences in social rules and attitudes, ge-ography—and to understand how they interacted over timeto produce the outcome that we describe as the terminationconditions at a particular point in time. This would provideinsight into the internal functioning of regional or individualcomponential causal systems. A second goal would be to un-derstand why for a different componential causal system ina different region and time similar factors interacted in dif-ferent ways to produce a different outcome. Through theidentification of commonalities and differences and by eval-uating the role played by specific components and factorswithin individual trajectories as well as the interplay betweenthem, one should be able to identify overarching trends inthe way in which systems operated. This process would allowus to move toward the goal of finding “the general law,” ifany, that operated behind the evolution of componentialcausal systems.

Regional Trajectories as Cultural Experiments

How do we put this approach into practice with the archae-ological, chronological, and paleoenvironmental data that wehave at hand? Gaining insight into the more remote aspectsof cultural systems, socially shared knowledge, and attitudestoward innovations is typically thought of as being a difficultendeavor for archaeologists. We argue, however, that usefulinferences can be made pertaining to these cognitive and so-cial domains through the detailed analysis of a wide range ofmaterial culture. When such inferences are placed within pa-leoenvironmental and landscape contexts, we possess an arrayof data that represent the setup conditions from which aninvestigation into mechanism can be launched. By applyingmethods that we detail below, one has the ability to identifythe key features and the degree of cohesiveness of these sys-tems and to track the way in which they evolved and possiblyresponded to environmental change through time. An interestin culture-environment interactions is by no means new, anda number of scholars have already argued that a culture’s“core” (Steward 1955; see also Odum 1971, and for a synthesisJohnson and Earle 2000) can be seen as a society’s means to

solve adaptive challenges. The problem one must address instudies of the distant past is how to operationalize this frameof thinking such that one gives social dimensions and materialculture the attention they deserve.

As is already evident to most archaeologists, an importantmethod for inferring behavior and cognition is to view ar-chaeological remains as representing an ordered chain ofevents, gestures, and processes belonging to a sequence ofactions that led to the transformation of a given material tothe finished form, that is, the chaıne operatoire (Lemonnier1986; Leroi-Gourhan 1964; Schlanger 1994). This concept isespecially pertinent because it permits one to infer from thefinished artifact, production waste, and potentially missingelements the socially shared and individual knowledge, co-operation, and amount of short- and long-term memory nec-essary for the functioning, maintenance, and transmission ofa given production sequence. It also allows one to understandto some extent the mental template of the actors.

Different classes of material culture possess different po-tentials to inform us on cultural cohesiveness, shared knowl-edge, and underlying cognitive processes. Until 10 years ago,identifying setup conditions of Middle Stone Age and MiddlePaleolithic “cultural experiments” would have meant focusingalmost exclusively on lithic technology, but now this has be-come a more complex and potentially informative endeavorconsidering the many categories of material remains that wehave recognized since then (bone tools, pigments, personalornaments, engravings, mastic compounds, poisons, contain-ers, use of plants and feathers, etc.). Because of their ubiquityand durability, lithic artifacts remain a valuable class of ma-terial for identifying the geographic extent of a CAS. In par-ticular, consistencies in lithic technologies and formal tooltypes have allowed researchers to identify discrete culturaladaptations and begin to piece together their geographic dis-tribution and chronological context in parts of Africa andEurasia (Barham 2001; Belfer-Cohen and Hovers 2010; Del-agnes and Meignen 2006; Discamps, Jaubert, and Bachellerie2011; Kuhn 2013; Soriano, Villa, and Wadley 2007; Villa etal. 2010, 2012; Wurz et al. 2003). The form that lithic rawmaterial acquisition takes (local or long-distance direct ac-quisition, trade, etc.) can be a reliable proxy for how an adap-tive system is linked to a given territory. The importanceassigned to specific types of raw material can provide insightinto how rigid or flexible a lithic technical system is and hownatural resources became key elements in the cultural system.Acquisition patterns can also reflect the presence of socialnetworks and their complexity. Analysis of debitage tech-niques can hint at the form and duration of apprenticeshiplikely necessary to produce certain classes of tools. Shapingtechniques to transform knapped blanks into finished tools,the degree of conformity to strict stylistic rules, and whetherformal tools were used for single or multiple functions allowus to understand the technical system’s degree of plasticityand to gauge the nature of information needed to maintainand transmit the required know-how. Finally, whether each

d’Errico and Banks Mechanisms behind Cultural Trajectories S377

phase of the chaıne operatoire is found in a unique locationor multiple locations can inform us not only about how apopulation organized its activities across the landscape butalso about the planning involved and the complexity of thecognitive maps that characterized an adaptive system.

During the last decade, improvements in excavation tech-niques and reappraisal of existing collections have identifieda variety of material culture remains other than lithics, andmethods have been developed to better infer their behavioralsignificance. During the Middle Stone Age and the end of theMiddle Paleolithic, we observe for the first time bone toolsproduced with techniques specifically conceived for this me-dium—such as grinding, scraping, and polishing—that al-lowed their final shape and size to be achieved with a highdegree of accuracy. This class of tool is considered particularlyappropriate for the characterization of technical systems,tracking technical changes through time, identifying regionalvariability, disentangling style from function, and inferringthe complexity inherent in a given adaptive system (Backwelland d’Errico 2005).

The production of complex compounds entails the orderedcombination and modification, often with the use of pyro-technology, of a variety of raw materials in order to producean end product that has physical properties not found in thenatural world. Beyond the important ability to develop suchtechnologies through experimentation, what is paramount isthe ability to maintain and transmit these innovations. Onceadopted, each recipe and the way it is employed can becomekey features of a CAS and have a feedback effect on otheraspects of the technical system as well as on how knowledgeis accumulated and shared. We now possess means to inferhow these compounds were made and used by specific humanpopulations in the Middle Paleolithic and the Middle StoneAge (Boeda et al. 1996; Carciumaru et al. 2012; Charrie-Duhaut et al. 2013; d’Errico et al. 2012; Hauck et al. 2013;Henshilwood et al. 2011; Mazza et al. 2006; Pawlik and This-sen 2011; Villa et al. 2012; Wadley 2010a; Wadley, Hodgskiss,and Grant 2009). Bitumen, birchbark tar, wood resin mixedwith hematite, hematite powder mixed with animal fat: weare starting to understand how different CASs have createdand incorporated comparable compound technologies as aresponse to specific needs in different environmental settings.

We are also moving beyond the simple recognition thatinstances of symbolic material culture in the form of personalornaments, engravings, pigment production and storage, anddecorated bone items are present in the archaeological recordof this period and reflect modern behavior to the explorationof the representativeness and significance of specific instances.First, one can wonder how much evidence has not been pre-served or, if preserved, been destroyed during excavation orgone unrecognized. It has been pointed out, for example, thatthe appearance of symbolic items in the archeological recordmay be largely conditioned by taphonomic processes (Barham2007). This problem can be partially overcome by criticallyexamining excavation techniques used at specific sites in the

past. Taphonomic analyses of the various categories of sym-bolic items may help to identify which classes of artifacts areespecially affected by taphonomic processes and infer whethersuch processes may have led to their disappearance at somesites. Cross-cultural analyses of the raw materials used toproduce symbolic items and the evaluation of their respectivedurability are also means to address the issue of the loss ofsome elements of symbolic material culture, particularly incontexts in which we observe other material remains pointingto the presence of symbolic mediated behaviors. These be-haviors may exist in a society that does not express themthrough purely symbolic items but rather embodies them infunctional items. These items express symbolic meaning bytheir adherence to strict stylistic norms that archaeologistsperceive as modern without having the means to disentanglefunctional from symbolic traits (Barham 2007). Taking intoaccount such items as components of a CAS is a means ofincorporating them, including their potential symbolic value,among the factors that played a role in the relationship be-tween culture and environment. In this way, potentially sym-bolic aspects of past material culture become full actors inthe exploration of mechanisms governing the evolution ofcultural systems.

Second, advanced analytical techniques contribute to thedisentanglement of accidental from purposeful behavior, thusallowing a precise documentation of the operational chainand underlying cognitive processes. Actualistic studies con-ducted with the aim of verifying the purposeful alteration ofpigment (Wadley 2010b) or shell bead color (d’Errico et al.2009; Kandel and Conard 2005) as well as assessing whetherplant remains are present naturally or because of specific hu-man actions such as use as bedding and as insect repellent(Wadley et al. 2011) are good examples of evaluating thebehavioral significance of past human agency. Archaeozoo-logical, taphonomic, chemical, technological, and functionalanalysis of symbolic items relying on actualistically establishedcriteria, when placed within the framework of a chaıne op-eratoire, have provided new means to identify and analyzeearly instances of symbolically mediated behavior. This pro-vides insights into the way in which material expressions ofsymbols were created, assembled, and displayed, how and forhow long they were used, and to what degree those earlysymbolic systems are comparable with those created by eth-nographically documented human societies. Adapted theo-retical frameworks are proposed to understand the amountand nature of information that one can convey through eachcategory of the identified type of symbolic material culture(Kuhn and Stiner 2007) and the function personal ornamentsmay have played in prehistoric societies (d’Errico and Van-haeren 2007). By establishing to what degree specific instancesof symbolic behavior are representative of a CAS and ex-ploring patterns of variability through time and space, we areable to verify whether these instances are inherent features ofthe system or simply the expression of subregional or suc-cessive cultural trajectories (Vanhaeren et al. 2013).

S378 Current Anthropology Volume 54, Supplement 8, December 2013

A key additional dimension that must be considered whenestablishing setup conditions of any cultural experiment isspace. This is particularly relevant in the case of hunter-gath-erer economies because we know, apart from a few cases inwhich particular adaptations allowed them to be sedentary orsemisedentary, that one key feature of these populations isthat they were mobile. Although hunter-gatherer economiesand their organization across the landscape can be broadlyclassified as logistical or residential (Binford 1980; Foley1992), we know that reality is more complex and must havebeen so in the past. We must delve deeper into each individualcase and attempt to understand the unique logic behind eachsettlement-subsistence system within its particular landscapeframework. We now have the means to reconstruct, at bothcontinental and regional scales, the effects that ice-sheet andsea-level changes (Lambeck, Esat, and Potter 2002) had onthe landscape. Information pertaining to environmental con-ditions has traditionally been reconstructed on the basis ofproxies from individual sites or groups of sites such as faunalor plant remains and soil characteristics, among others. Thesecertainly provide important information, but if one wishes toexplore the systemic and dynamic relationship between cul-ture and environment, one must move beyond the site scaleand find a way to examine this relationship at regional andcontinental scales.

An Integrative Approach

Research seeking to understand the relationship between hu-man and/or cultural evolution and climate change during theMiddle and Upper Pleistocene is becoming an increasinglywidespread field of study (Bocquet-Appel et al. 2005; Cartoet al. 2009; Compton 2011; deMenocal 2011; d’Errico andSanchez-Goni 2003; Discamps, Jaubert, and Bachellerie 2011;Gamble et al. 2004; Jacobs et al. 2008; Lowe et al. 2012; Maslinand Christensen 2007; Osborne et al. 2008; Richerson, Boyd,and Bettinger 2009; Sepulchre et al. 2007; Van Andel andDavies 2003). Most of this research has attempted to explorethe role played by climate on either long-term or suddenevolutionary/cultural changes/replacements, but they do notdetail means with which to verify the proposed causal con-nection or test alternative hypotheses. Only a handful of stud-ies have designed tools to test the relationship between thesefactors and explore their interaction at regional scales (e.g.,Bocquet-Appel and Tuffreau 2009). These studies attempt totest the correlation between a single and often qualitativeclimatic variable (e.g., isotopic stage, cold/warm) and thefunction of sites or the nature of the material record in agiven region. Such an approach, however, does not evaluatethe various components of a given CAS against multiple andquantitative climatic environmental variables. Other studiesuse a variety of modeling techniques to contrast Neanderthaland AMH population dynamics and interactions (Barton etal. 2011; Fabre et al. 2011), but they produce results that aredifficult to test against the archaeological record and do not

realistically incorporate the environmental conditions of eachphase that characterize the variable climatic conditions of thisperiod.

In a series of papers published during the last few years,we have outlined an approach with which we can explore theinteractions between CASs and paleoenvironment and un-derstand how environmental dynamics may have influencedthese adaptations and the distribution of prehistoric hunter-gatherer populations (Banks et al. 2008b, 2009, 2011; d’Erricoand Stringer 2011). This approach, termed eco-cultural nichemodeling (ECNM), integrates archaeological, chronological,geographic, and paleoclimatic data sets via biocomputationalarchitectures derived from biodiversity studies (Peterson etal. 2011) to estimate ecological niches and distributional areasoccupied by prehistoric hunter-gatherer populations and toidentify and quantify the environmental factors that shapedthese niches.

An eco-cultural niche is defined as the range of environ-mental conditions (i.e., the ecological niche) exploited by aCAS (see Banks, d’Errico, and Zilhao 2013 for a detaileddiscussion of eco-cultural niches). ECNM assumes that wecan characterize a past cultural niche by employing methodsused to reconstruct and study ecological niches. A key featureof such predictive architectures is that they can project theecological niche predicted for a climatic phase onto the en-vironmental conditions of a subsequent period. The resultingniche projection is compared with the locations of knownoccurrences for the latter period to see whether or not itsuccessfully predicts their presence within the niche. In thisway, one can evaluate whether an adaptive system, in the eventof its persistence, exploited the same ecological niche acrossdifferent climatic phases or significantly expanded or con-tracted it. This approach parallels the inquiry into the mech-anism driving the evolution of a componential causal systembecause it allows one to define the setup conditions of theprocess at work, in our case an individual cultural experiment,and evaluate its termination conditions (i.e., the attributesand distribution of a CAS at the end of the process) withineither the framework of environmental change or relative sta-sis.

For data inputs, ECNM requires the geographic coordinatesof archaeological sites bearing cultural features that are con-sidered distinctive of a particular CAS or consistent subsetswithin that system along with a set of raster geographic in-formation system data layers summarizing environmental di-mensions potentially relevant to shaping the eco-culturalniche exploited by the CAS as well as its spatial expressionduring a specific climatic phase. Geographic variables are as-sumed to have remained relatively constant over the past 300kyr, and thus one can use high-resolution present-day data(e.g., ETOPO1). Reconstructions of past sea-level fluctuationsat both general and regional scales are available and can beused to reconstruct coastlines and related paleogeography forthe region of study. Reconstructions of ice-sheet volume andlocation are available for most of the last climatic cycle and

d’Errico and Banks Mechanisms behind Cultural Trajectories S379

Figure 1. Schematic rendition of how one of the predictive architectures, in this case genetic algorithm for rule-set prediction,reconstructs an eco-cultural niche. 1, Occurrence data (i.e., location of archaeological sites belonging to a cohesive adaptive system)are resampled randomly by the algorithm to create training (b) and test data sets. An iterative process of rule generation andimprovement then follows in which an inferential tool is chosen from a suite of rule types and applied to the training data (S1–Si)and paleoenvironmental raster data layers (a) to develop specific rules (Stockwell and Peters 1999). These rules evolve to maximizepredictivity by several means (e.g., crossing over among rules), mimicking chromosomal evolution. Predictive accuracy is evaluatedbased on an independent subsample of presence data and a set of points sampled randomly from regions where the species hasnot been detected. 2, The resulting rule set defines the distribution of the subject in environmental dimensions (i.e., the ecologicalniche; Soberon and Peterson 2005), which is projected onto the landscape to estimate a potential geographic distribution (Peterson2003).

can be inferred, with some incertitude, for more ancient pe-riods. With respect to paleoclimatic variables (temperatureand precipitation), there exists a variety of modeling tech-niques for obtaining reconstructions that can be integratedinto a niche modeling approach. One can run (1) a coupledocean-atmosphere general circulation model (e.g., IPSL CM4and CM5A; Dufresne et al. 2013; Kageyama et al. 2013), (2)an atmosphere-only model with a slab ocean component (rep-resenting the top 50 m of the water column; Kang et al. 2008),or (3) an atmosphere-only model with imposed sea surfacetemperature (SST) values (Kageyama et al. 2005). With allthree, boundary conditions (orbital parameters, greenhousegas concentrations, ice-sheet volume) appropriate for the tar-geted climatic event are assigned. The atmosphere-only modelwith imposed SSTs also can be run with a refined resolution(∼50 km) over the region(s) of interest (see Banks et al. 2008a;Sepulchre et al. 2007).

The results from the different methods listed above can bestatistically downscaled (e.g., Vrac, Stein, and Hayhoe 2007)to increase the resolution of the simulated paleoclimatic data.A final option is to use a regional model forced by generalcirculation model outputs as boundary conditions, therebyproducing climatic simulations with a resolution as fine as 5km (Frei et al. 2006). This high level of resolution, or evenfiner when possible, is most appropriate for examining cul-tural and niche trajectories at a regional scale. The outputsof this simulation process can be used to force a dynamicglobal vegetation model (e.g., Orchidee, Spitfire) in order to

obtain reconstructions of vegetation cover compatible withthe targeted climate state. In this way, one obtains values forprecipitation, temperature (mean annual, coldest month,warmest month), and broad vegetation types. During thisprocess, outputs are compared with paleoenvironmental datato test whether the simulations capture past conditions sat-isfactorily, and if they do not, there exist means to improvesubsequent generations of simulations in an effort to bettercapture past paleoclimatic conditions.

Other methods exist, such as using statistical techniques toinfer past climatic conditions from a variety of vegetationdata (for a review, see Tingley et al. 2012), but so far theylack the spatial resolution required for our purposes. Ofcourse, models are just that, models, and one must keep inmind that they only approximate past climatic conditions.Also, our goals are different from those of paleoclimatologistswho seek to understand the functioning and evolution of theearth’s climate system. We wish to have at our disposal themost accurate simulation for paleoclimatic conditions at aspecific time in the past and seek to improve our means forevaluating the pertinence of paleoclimatic simulations. Be-cause of the pressure of the threat of global warming, climatemodeling is a rapidly evolving field of research, and it is clearthat means to evaluate the pertinence of high-resolution sim-ulations at regional scales will rapidly improve in comingyears. We will greatly benefit from such improvements.

A number of predictive modeling approaches are available(climatic envelope range, generalized linear model, general-

S380 Current Anthropology Volume 54, Supplement 8, December 2013

Figure 2. Schematic illustration of how, following a climaticchange, the conservation of an eco-cultural niche (a) may resultin either a contraction (b) or an expansion (c) of the niche’sgeographic range.

ized additive model, genetic algorithm for rule-set prediction,maximum entropy, ensemble approach; for a review see Ar-aujo and New 2007 as well as Pearson et al. 2006) for re-constructing the ecological niche of a CAS and its geographicdistribution using the above data. Araujo and New (2007)point out that ideally one should use multiple modeling meth-ods and compare their outputs. Therefore, in our previousapplications of ECNM we have employed genetic algorithm(genetic algorithm for rule-set prediction) and maximum en-tropy (Maxent) methods. At a very general level, these ar-chitectures first identify shared paleoenvironmental parame-ters among the geographic locations of archaeological sitesbelonging to the same culture and then find other geographicregions where these parameters are present, thus predictingthe total ecological range of the target population (fig. 1).When estimating ecological niches, it is important to considerthe geographic areas that would have been accessible via dis-persal to the population in question and that have been sam-pled such that occurrences could have been detected (Barveet al. 2011; see Banks, d’Errico, and Zilhao 2013 and Bankset al. 2013 for archaeological examples); this area is termed“M” in the BAM framework of Soberon and Peterson (2005).Once meaningful ecological niche estimations have been pro-duced, statistical methods are used to identify the environ-mental factors that shaped these niches and to measure theirbreadth. Similarly, a variety of methods exist (e.g., backgroundsimilarity test: Warren, Glor, and Turelli 2010; partial-ROCtest: Peterson, Papes, and Soberon 2008) to test whether twoCASs’ eco-cultural niches are significantly different or whether

they are interpredictive either within a single climatic eventor between two different events. Two niches are interpredictivewhen their observed degree of similarity is greater than wouldbe expected by chance.

The interest in following this approach lies in that we movebeyond the analysis of site distributions to that of niches andcan evaluate and quantify patterns of continuity or nicheshifts. Predictive algorithms allow us to evaluate whether aCAS has conserved, expanded, or contracted its ecologicalniche in the time span between two different points in time.It is noteworthy that such changes may not necessarily bereflected in observed changes of their geographic range. Inthe event of climatic change, if an adaptive system conservesits ecological niche and simply tracks its shifting footprint,this can result in either an expansion or contraction of itsgeographic range depending on the range that the relatedclimatic envelope occupies following the climatic shift (fig.2). With this approach, changes in eco-cultural niches areassessed without a priori assumptions on the role played byenvironment. ECNM can identify cases in which significantcultural change, potentially reflecting changes in social rulesand related organization along with niche shifts, appear to beunrelated to environmental variability. Also, it is often in-tuitively assumed that cultural innovations indicate an abilityto better exploit environments and increase a population’sgeographic range. However, in some instances innovationsmay reflect responses to environmental change such that theyallow a population to maintain its niche and avoid nichecontraction. Situations may exist in which cultural innova-tions are associated with a niche contraction. This may occureither because a CAS copes with climate change by targetinga smaller subset of its previous niche or because, in spite ofinnovations, the CAS is unable to maintain its previous niche.

We can envision four different scenarios in the evolutionof a CAS between setup and termination conditions (fig. 3).In the first scenario, the material culture characterizing a CASremains the same between these two points in time. Thesecond scenario features changes in some aspects of the ma-terial culture, but one can identify a clear continuity throughtime. From the perspective of resilience theory, these twosituations can be characterized as ecological resilience (Pee-ples, Barton, and Schmich 2006), which describes the situa-tion in which an adaptive system is changed and reorganizedwhile maintaining its key features and functions. For the thirdscenario, material culture disappears from the archaeologicalrecord, indicating that populations are no longer present inthe region or that they are archaeologically invisible. In thefourth scenario, the material culture associated with the ter-mination conditions is clearly different from that of setupconditions (e.g., a different techno-complex). Of course, therecognition of scenarios 1, 2, and 4 assumes that archaeol-ogists have the means to identify and follow the cultural pro-cesses at work in the various aspects of the technological andsymbolic domains. Scenarios 3 and 4 imply that at someundetermined point in time between setup and termination

Figure 3. Scenarios of a cohesive adaptive system’s (CAS’s) evolution between setup and termination conditions within the frameworkof climatic stability (A–D) or climate change (E–H). Columns depict setup conditions and one of four possible termination conditions:(1) niche conservation, (2) niche expansion, (3) niche contraction, or (4) disappearance. A and E illustrate situations in whichmaterial culture remains essentially unchanged (described as scenario 1 in the text). B and F depict situations in which the materialculture changes but clear continuity between setup and termination conditions can be recognized (scenario 2). In instances C andG, a CAS disappears or becomes archaeologically invisible at termination, representing either an extreme instance of niche contractionor migration (scenario 3). D and H are instances in which clear discontinuities in the material culture are observed between setupand termination (scenario 4). The potential for shifts in geographic range is indicated by displacement of the eco-cultural nichewithin the frame representing termination conditions.

S382 Current Anthropology Volume 54, Supplement 8, December 2013

Figure 4. Idealized example of a long-term regional cultural trajectory (a) composed of multiple stages in which terminationconditions become setup conditions for the subsequent stage, all of this encompassing multiple periods that may be characterizedby either relative climatic stability or climatic change. Trends in niche variability (b) synthesize long-term trends in the relationshipbetween cohesive adaptive systems and environmental variability at a regional scale.

conditions, a process or event occurred that led to the dis-appearance of the initial CAS or to its evolution into a sig-nificantly different cultural system, respectively.

Climate change may or may not have occurred betweensetup and termination conditions for the above scenarios. Theoutcome of these eight different trajectories (the four sce-narios with or without climate change) may involve nichestability, contraction, or expansion (i.e., 20 possible differentoutcomes; fig. 3). Once one has determined into which sce-nario the archaeological record under investigation falls,ECNM tools allow one to determine whether ecological fac-tors played a role in a given CAS’s trajectory or whether itwas influenced to a greater extent by cultural processes. Moreimportantly, in the first case, dedicated statistical tools allowone to quantify whether these changes are significantly dif-ferent from what would be expected by chance. For the sce-

narios taking place during a period of relative climatic stasis,comparing the initial and final eco-cultural niches and quan-tifying potential differences can be accomplished with back-ground similarity tests (Warren, Glor, and Turelli 2010; foran archaeological application see Banks et al. 2011). For sce-narios that occurred across a period characterized by a cli-matic change, the setup eco-cultural niche is projected ontothe climatic conditions of the termination period to evaluatewhether or not it changed or remained stable. These initialand projected eco-cultural niches can be compared with avariety of statistical methods (e.g., partial-ROC comparisons:Peterson, Papes, and Soberon 2008; cumulative binomial sta-tistic: Banks et al. 2008a; background similarity test: Warren,Glor, and Turelli 2010), again with the goal of evaluatingwhether or not there was an eco-cultural niche shift duringthe process.

d’Errico and Banks Mechanisms behind Cultural Trajectories S383

By determining whether or not an eco-cultural niche re-mained stable, along with an identification of the key envi-ronmental variables that played a role in the definition of theinitial and final niches, and in conjunction with a detailedreconstruction of the behaviors reflected in the material cul-ture record, we should be able to identify the different factorsand components involved in the mechanism at work behindthe evolution of a CAS. When one steps back to look at thebigger picture, each transition between setup and terminationrepresents a single stage within a long-term regional processin which the termination conditions of each stage becomethe setup conditions of the following stage (fig. 4). This sub-sequent stage in turn may not resemble the previous one andcan potentially fall into another of the 20 possible scenariosoutlined above. When viewed at a regional scale and acrossa long span of time encompassing a number of climaticphases, this process may reveal trends (e.g., random, punc-tuated, unilinear, exponential) in the way CASs respond toclimate variability. In this way trends may be revealed thatwould not be evident at the stage level of analysis. We may,for example, face at the stage level minor niche expansionthat is not statistically significant but that may become sowhen multiple, successive stages are considered. This couldreveal trends that underlie a long-term regional trajectory. Wedo not expect regional trends to mirror one another. It ispossible, however, that by comparing them, consistencies maybecome apparent. The identification of such consistenciesmay allow us to move from the analysis of multiple successivecomponential causal systems to the formulation of a generalor overarching theory that explains the mechanisms that gov-erned the evolution of human cultures and their relationshipto the environment before the development of production-based economies.

Conclusions

Documenting regional cultural trajectories is worthwhile, butit is not enough. While the direction that archaeological in-vestigations into behavioral modernity have taken in recentyears has been useful in that it has provided us with a wealthof detailed empirical data, this research has so far fallen shortof examining the dynamic relationships between the multi-tude of factors that were at play between human populationsand the environments within which they operated. Focusingon the various paths that human populations around theworld have taken to become what we are today certainlypresents the advantage of escaping previous gross approxi-mations or misleading and generalized scenarios accountingfor the origin of modernity. A growing body of evidence isdemonstrating that such scenarios, founded on single causes,do not account for the complex processes at work in eachregion of the world. This raises the question of how detailedcompilations of existing data and results from new researchon individual regions can lead us forward. Such approachescan quickly fall into the trap of cultural relativism: each cul-

ture is viewed as unique, and one implicitly assumes that ourjob as archaeologists is to simply document cultural variabilitywithout the ability to identify underlying trends in the be-havioral evolution of our lineage. Alternatively, intrinsic in anumber of regional examinations is the idea that one canbetter identify, at that scale, one or more points in time duringwhich significant behavioral transformations occurred thatlead human populations in that region to cross one or moreRubicons on the path to modernity. Proposing a best-fit factorat play in those passages from the available menu (demog-raphy, environment, cognitive changes, climate change, spe-ciation, etc.) often equates to transferring a single-cause sce-nario from a general to a local scale. The same holds true formore environmentally deterministic approaches seeking toreconstruct climatic changes at a regional scale with the aimof identifying a local prime mover behind a behavioral shiftidentified in that region: contemporaneity does not equate toa causal link.

We argue that meaningful advances in this field of studycannot be achieved without integrating detailed informationon past human behavior into a research strategy that allowsone to interrogate, rather than simply document, past materialculture with the aim of identifying short- and long-termmechanisms at work in the evolution of CASs within theirrespective, dynamic paleoenvironmental frameworks. To ef-fectively do so, we must apply the same methods to individualregional trajectories and conceive heuristic tools that enableus to quantitatively compare and evaluate different regionaltrajectories and their associated behavioral changes throughtime. Integration of paleoanthropological and paleogeneticdata can be important but should probably represent a laterstage of the inquiry into mechanisms, as implied by Lalueza-Fox (2013), rather than being used as a prime mover as isnow the case in the more popular single-cause models. Cog-nition does not exist in nature as a given but rather as theresult of a continuous interaction between conspecifics as wellas between them and the environment. Hypotheses on thebehavioral implications of genomic variability need to betested by finding ways to explore possible interactions betweenaptitudes and genes rather than attributing to our ancestorsan assumed cognitive potential based on our taxonomic read-ing of the fossil record.

Acknowledgments

We thank Erella Hovers and Steve Kuhn for their invita-tion to participate in the “Alternative Pathways to Complex-ity” symposium and Laurie Obbink for her valuable assis-tance. We also wish to thank Gauthier Devilder for his helpin creating figures 2–4. Finally, we thank the editors and ananonymous reviewer for their constructive comments. Theresearch that led us to the writing of this paper was supportedby the European Research Council (FP7/2007/2013/ERC

S384 Current Anthropology Volume 54, Supplement 8, December 2013

TRACSYMBOLS 249587) and the Project Origine II, Aqui-taine Region.

References CitedAmbrose, Stanley H. 1998. Late Pleistocene human population bottlenecks,

volcanic winter, and differentiation of modern humans. Journal of HumanEvolution 34:623–651.

Araujo, Miguel B., and Mark New. 2007. Ensemble forecasting of speciesdistributions. Trends in Ecology and Evolution 22:42–47.

Backwell, Lucinda, and Francesco d’Errico. 2005. The origin of bone tooltechnology and the identification of early hominid cultural traditions. InFrom tools to symbols: from early hominids to modern humans. Francescod’Errico and Lucinda Backwell, eds. Pp. 238–275. Johannesburg: Witwa-tersrand University Press.

Backwell, Lucinda, Francesco d’Errico, and Lyn Wadley. 2008. Middle StoneAge bone tools from the Howiesons Poort layers, Sibudu Cave, South Africa.Journal of Archaeological Science 35:1566–1580.

Banks, William E., Nicolas Antunes, Solange Rigaud, and Francesco d’Errico.2013. Ecological constraints on the first prehistoric farmers in Europe.Journal of Archaeological Science 40:2746–2753.

Banks, William E., Thierry Aubry, Francesco d’Errico, Joao Zilhao, AndresLira-Noriega, and A. Townsend Peterson. 2011. Eco-cultural niches of theBadegoulian: unraveling links between cultural adaptation and ecology dur-ing the Last Glacial Maximum in France. Journal of Anthropological Ar-chaeology 30:359–374.

Banks, William E., Francesco d’Errico, A. Townsend Peterson, Masa Kageyama,Adriana Sima, and Maria Sanchez-Goni. 2008a. Neanderthal extinction bycompetitive exclusion. PLoS ONE 3(12):e3972.

Banks, William E., Francesco d’Errico, A. Townsend Peterson, Marian Van-haeren, Masa Kageyama, Pierre Sepulchre, Gilles Ramstein, Anne Jost, andDaniel Lunt. 2008b. Human ecological niches and ranges during the LGMin Europe derived from an application of eco-cultural niche modeling.Journal of Archaeological Science 35:481–491.

Banks, William E., Francesco d’Errico, and Joao Zilhao. 2013. Human-climateinteraction during the Early Upper Paleolithic: testing the hypothesis of anadaptive shift between the Proto-Aurignacian and the Early Aurignacian.Journal of Human Evolution 64:39–55.

Banks, William E., Joao Zilhao, Francesco d’Errico, Masa Kageyama, AdrianaSima, and Annamaria Ronchitelli. 2009. Investigating links between ecologyand bifacial tool types in Western Europe during the Last Glacial Maximum.Journal of Archaeological Science 36:2853–2867.

Barham, Lawrence S. 2001. Central Africa and the emergence of regionalidentity in the Middle Pleistocene. In Human roots: Africa and Asia in theMiddle Pleistocene. Lawrence S. Barham and Kate Robson-Brown, eds. Pp.65–80. Bristol: Western Academic and Specialist Press.

———. 2007. Modern is as modern does? technological trends and thresholdsin the south-central African record. In Rethinking the human revolution.Paul Mellars, Katie Boyle, Ofer Bar-Yosef, and Chris Stringer, eds. Pp. 165–176. McDonald Institute Monographs. Cambridge: McDonald Institute forArchaeological Research.

Barton, C. Michael, Julien Riel-Salvatore, John M. Anderies, and GabrielPopescu. 2011. Modeling human ecodynamics and biocultural interactionsin the Late Pleistocene of western Eurasia. Human Ecology 39:705–725.

Barve, Narayani, Vijay Barve, Alberto Jimenez-Valverde, Andres Lira-Noriega,Sean P. Maher, A. Townsend Peterson, Jorge Soberon, and Fabricio Villa-lobos. 2011. The crucial role of the accessible area in ecological niche mod-eling and species distribution modeling. Ecological Modelling 222:1810–1819.

Bar-Yosef, Ofer, and Anna Belfer-Cohen. 2011. Following Pleistocene roadsigns of human dispersals across Eurasia. Quaternary International 285:30–43.

Bechtel, William, and Adele Abrahamsen. 2005. Explanation: a mechanisticalternative. Studies in History and Philosophy of Biological and BiomedicalSciences 36:421–441.

Belfer-Cohen, Anna, and Erella Hovers. 2010. Modernity, enhanced workingmemory, and the Middle to Upper Paleolithic record in the Levant. CurrentAnthropology 51(suppl. 1):S167–S175.

Bickerton, Derek. 1990. Language and species. Chicago: University of ChicagoPress.

Binford, Lewis R. 1980. Willow smoke and dogs’ tails: hunter-gatherer set-

tlement systems and archaeological site formation. American Antiquity 45:4–20.

Bocquet-Appel, Jean-Pierre, Pierre-Yves Demars, Lorette Noiret, and DmitryDobrowsky. 2005. Estimates of Upper Paleolithic meta-population size inEurope from archaeological data. Journal of Archaeological Science 32:1656–1668.

Bocquet-Appel, Jean-Pierre, and Alain Tuffreau. 2009. Technological responsesof Neanderthals to macroclimatic variations (240,000–40,000 BP). HumanBiology 81:287–307.

Boeda, Eric, Jacques Connan, Daniel Dessort, Sultan Muhesen, Norbert Mer-cier, Helene Valladas, and Nadine Tisnerat. 1996. Bitumen as a haftingmaterial on Middle Palaeolithic artefacts. Nature 380:336–338.

Bowles, Samuel. 2009. Did warfare among ancestral hunter-gatherers affectthe evolution of human social behaviors? Science 324:1293–1298.

Brown, Kyle S., Curtis W. Marean, Andy I. R. Herries, Zenobia Jacobs, ChantalTribolo, David Braun, David L. Roberts, Michael C. Meyer, and JocelynBernatchez. 2009. Fire as an engineering tool of early modern humans.Science 325:859–862.

Carciumaru, Marin, Rodica-Mariana Ion, Elena-Cristina Nitu, and Radu Ste-fanescu. 2012. New evidence of adhesive as hafting material on Middle andUpper Paleolithic artefacts from Gura Cheii-Rasnov Cave (Romania). Jour-nal of Archaeological Science 39:1942–1950.

Caron, Francois, Francesco d’Errico, Pierre Del Moral, Frederic Santos, andJoao Zilhao. 2011. The reality of Neanderthal symbolic behavior at theGrotte du Renne, Arcy-sur-Cure, France. PLoS ONE 6(6):e21545.

Carto, Shannon L., Andrew J. Weaver, Renee Hetherington, Yin Lam, andEdward C. Wiebe. 2009. Out of Africa and into an ice age: on the role ofglobal climate change in the late Pleistocene migration of early modernhumans out of Africa. Journal of Human Evolution 56:139–151.

Charrie-Duhaut, Armelle, Guillaume Porraz, Caroline R. Cartwright, MarinaIgreja, Jacques Connan, Cedric Poggenpoel, and Pirre-Jean Texier. 2013.First molecular identification of a hafting adhesive in the Late HowiesonsPoort at Diepkloof Rock Shelter (Western Cape, South Africa). Journal ofArchaeological Science 40:3506–3518.

Childe, V. Gordon. 1929. The Danube in prehistory. Oxford: Oxford UniversityPress.

Clarke, David L. 1968. Analytical archaeology. London: Methuen.Compton, John S. 2011. Pleistocene sea-level fluctuations and human evo-

lution on the southern coastal plain of South Africa. Quaternary ScienceReviews 30:506–527.

Conard, Nicholas J. 2008. A critical view of the evidence for a southern Africanorigin of behavioural modernity. South African Archaeological Society Good-win Series 10:175–179.

Deacon, Janette. 1990. Weaving the fabric of Stone Age research in southernAfrica. In A history of African archaeology. Peter Robertshaw, ed. Pp. 39–58. London: James Currey.

Delagnes, Anne, and Liliane Meignen. 2006. Diversity of lithic productionsystems during the Middle Paleolithic in France. In Transitions before thetransition: evolution and stability in the Middle Paleolithic and Middle StoneAge. Erella Hovers and Steven L. Kuhn, eds. Pp. 85–107. New York: Springer.

deMenocal, Peter B. 2011. Climate and human evolution. Science 331:540–542.

d’Errico, Francesco. 2003. The invisible frontier: a multiple species model forthe origin of behavioral modernity. Evolutionary Anthropology 12:188–202.

———. 2009. Le rouge et le noir: implications of early pigment use in Africa,the Near East, and Europe for the origin of cultural modernity. South AfricanArchaeological Bulletin 10:157–174.

d’Errico, Francesco, Lucinda Backwell, Paola Villa, Ilaria Degano, JeannetteJ. Lucejko, Marion K. Bamford, Thomas F. G. Higham, Maria Perla Col-ombini, and Peter B. Beaumont. 2012. Artifacts from Border Cave: earliestevidence of San material culture. Proceedings of the National Academy ofSciences of the USA 109:13214–13219.

d’Errico, Francesco, Lucinda Backwell, and Lyn Wadley. 2012. Indentifyingregional variability in Middle Stone Age bone technology: the case of SibuduCave. Journal of Archaeological Science 39:2479–2495.

d’Errico, Francesco, Valentina Borgia, and Annamaria Ronchitelli. 2012. Uluz-zian bone technology and its implications for the origin of behaviouralmodernity. Quaternary International 259:59–71.

d’Errico, Francesco, and Christopher S. Henshilwood. 2007. Additional evi-dence for bone technology in the southern African Middle Stone Age.Journal of Human Evolution 52:142–163.

———. 2011. The origin of symbolically mediated behavior. In Homo sym-bolicus: the dawn of language, imagination, and spirituality. Christopher S.

d’Errico and Banks Mechanisms behind Cultural Trajectories S385

Henshilwood and Francesco d’Errico, eds. Pp. 49–73, Amsterdam: JohnBenjamins.

d’Errico, Francesco, Renata Garcıa Moreno, and Riaan Rifkin. 2012. Tech-nological, elemental and colorimetric analysis of an engraved ochre frag-ment from the Middle Stone Age levels of Klasies River Cave 1, SouthAfrica. Journal of Archaeological Science 39:942–952.

d’Errico, Francesco, Helene Salomon, Colette Vignaud, and Chris Stringer.2010. Pigments from the Middle Palaeolithic levels of Es-Skhul (MountCarmel, Israel). Journal of Archaeological Science 37:3099–3110.

d’Errico, Francesco, and Maria Fernanda Sanchez Goni. 2003. Neanderthalextinction and the millennial scale climatic variability of the OIS 3. Qua-ternary Science Reviews 22:769–788.

d’Errico, Francesco, and Chris B. Stringer. 2011. Evolution, revolution orsaltation scenario for the emergence of modern cultures. PhilosophicalTransactions of the Royal Society B 366:1060–1069.

d’Errico, Francesco, and Marian Vanhaeren. 2007. Evolution or revolution?new evidence for the origin of symbolic behavior in and out of Africa. InRethinking the human revolution. Paul Mellars, Katie Boyle, Ofer Bar-Yosef,and Chris Stringer, eds. Pp. 275–286. McDonald Institute Monographs.Cambridge: McDonald Institute for Archaeological Research.

d’Errico, Francesco, Marian Vanhaeren, Nick Barton, Abdeljalil Bouzouggar,Henk Mienis, Daniel Richter, Jean-Jacques Hublin, Shannon P. McPherron,and Pierre Lozouet. 2009. Additional evidence on the use of personal or-naments in the Middle Paleolithic of North Africa. Proceedings of the Na-tional Academy of Sciences of the USA 106:16051–16056.

Discamps, Emmanuel, Jacques Jaubert, and Francois Bachellerie. 2011. Hu-man choices and environmental constraints: deciphering the variability oflarge game procurement from Mousterian to Aurignacian times (MIS 5–3) in southwestern France. Quaternary Science Reviews 30:2755–2775.

Dufresne, Jean-Louis, Marie-Alice Foujols, Sebastien Denvil, Arnaud Caubel,Olivier Marti, Yves Balkanski, Slimane Bekki, et al. 2013. Climate changeprojections using the IPSL-CM5 earth system model: from CMIP3 toCMIP5. Climate Dynamics 40:2123–2165.

Dunbar, Robin I. M. 2003. The social brain: mind, language, and society inevolutionary perspective. Annual Review of Anthropology 32:163–181.

Fabre, Virginie, Silvana Condemi, Anna Degioanni, and Estelle Herrscher.2011. Neanderthal versus modern humans: evidence for resource compe-tition from isotopic modelling. International Journal of Evolutionary Biology2011, doi:10.4061/2011/689315.

Foley, Robert L. 1992. Mobility/sedentism: concepts, archaeological measures,and effects. Annual Review of Anthropology 21:43–66.

Frei, Christoph, Regina Scholl, Sophie Fukutome, Jurg Schmidli, and Pier L.Vidale. 2006. Future change of precipitation extremes in Europe: intercom-parison of scenarios from regional climate models. Journal of GeophysicalResearch 111, doi:10.1029/2005JD005965.

Gamble, Clive, William Davies, Paul Pettitt, and Martin Richards. 2004. Cli-mate change and evolving human diversity in Europe during the last glacial.Philosophical Transactions of the Royal Society B 359:243–254.

Gelfand, Michele J., Jana L. Raver, Lisa Nishii, Lisa M. Leslie, Janetta Lun,Beng Chong Lim, Lili Duan, Assaf Almaliach, Soon Ang, and JakobinaArnadottir. 2011. Differences between tight and loose cultures: a 33-nationstudy. Science 332:1100–1104.

Gibson, Kathleen R. 2007. Putting it all together: a constructionist approachto the evolution of human mental capabilities. In Rethinking the humanrevolution. Paul Mellars, Katie Boyle, Ofer Bar-Yosef, and Chris Stringer,eds. Pp. 67–77. McDonald Institute Monographs. Cambridge: McDonaldInstitute for Archaeological Research.

Green, Richard E., Johannes Krause, Adrian W. Briggs, Tomislav Maricic, UdoStenzel, Martin Kircher, Nick Patterson, et al. 2010. A draft sequence ofthe Neandertal genome. Science 328:710–722.

Grun, Rainer, Chris Stringer, Frank McDermott, Roger Nathan, Naomi Porat,Steve Robertson, Lois Taylor, Graham Mortimer, Stephen Eggins, and Mal-colm McCulloch. 2005. U-series and ESR analyses of bones and teeth re-lating to the human burials from Skhul. Journal of Human Evolution 49:316–334.

Hauck, Thomas C., Jacques Connan, Armelle Charrie-Duhaut, Jean-Marie LeTensorer, and Heba al Sakhel. 2013. Molecular evidence of bitumen in theMousterian lithic assemblage of Hummal (Central Syria). Journal of Ar-chaeological Science 40:3252–3262.

Hayden, Brian. 1979. Palaeolithic reflections: lithic technology and ethnographicexcavation among Australian Aborigines. Englewood Cliffs, NJ: HumanitiesPress.

Henn, Brenna M., Christopher R. Gignoux, Matthew Jobin, Julie M. Granka,

J. M. Macpherson, Jeffrey M. Kidd, Laura Rodrıguez-Botigue, et al. 2011.Hunter-gatherer genomic diversity suggests a southern African origin formodern humans. Proceedings of the National Academy of Sciences of the USA108:5154–5162.

Henrich, Joseph. 2004. Demography and cultural evolution: how adaptivecultural processes can produce maladaptive losses: the Tasmanian case.American Antiquity 69:197–214.

Henrich, Joseph, Steven J. Heine, and Ara Norenzayan. 2010. The weirdestpeople in the world? Behavioral and Brain Sciences 33:61–135.

Henshilwood, Christopher S., Francesco d’Errico, Karen L. van Niekerk, YvanCoquinot, Zenobia Jacobs, Stein-Erik Lauritzen, Michel Menu, and RenataGarcıa-Moreno. 2011. A 100,000-year-old ochre-processing workshop atBlombos Cave, South Africa. Science 334:219–222.

Henshilwood, Christopher S., Francesco d’Errico, and Ian Watts. 2009. En-graved ochres from the Middle Stone Age levels at Blombos Cave, SouthAfrica. Journal of Human Evolution 57:27–47.

Henshilwood, Christopher S., and Curtis W. Marean. 2003. The origin ofmodern human behaviour: a review and critique of models and test im-plications. Current Anthropology 44:627–665.

Hodder, Ian. 1991. Reading the past: current approaches to interpretation inarchaeology. 2nd edition. Cambridge: Cambridge University Press.

Holling C. S. 1973. Resilience and stability of ecological systems. AnnualReview of Ecology and Systematics 4:1–23.

Hovers, Erella. 2009. The Middle-to-Upper Paleolithic transition: what news?In Sourcebook of Paleolithic transitions: methods, theories, and interpretations.Marta Camps and Parth Chauhan, eds. Pp. 455–462. New York: Springer.

Hovers, Erella, and Anna Belfer-Cohen. 2006. “Now you see it, now youdon’t”: modern human behavior in the Middle Paleolithic. In Transitionsbefore the transition: evolution and stability in the Middle Paleolithic andMiddle Stone Age. Erella Hovers and Steven L. Kuhn, eds. Pp. 295–304.New York: Springer.

Jacobs, Zenobia, Richard G. Roberts, Rex F. Galbraith, Hilary J. Deacon, RainerGrun, Alex Mackay, Peter Mitchell, Ralf Vogelsang, and Lyn Wadley. 2008.Ages for the Middle Stone Age of southern Africa: implications for humanbehavior and dispersal. Science 322:733–735.

Johnson, Allen W., and Timothy Earle. 2000. The evolution of human societies:from foraging group to agrarian state. 2nd edition. Stanford, CA: StanfordUniversity Press.

Kageyama, Masa, Pascale Braconnot, Laurent Bopp, Veronique Mariotti, TillaRoy, Marie-Noelle Woillez, Arnaud Caubel, et al. 2013. Mid-Holocene andlast glacial maximum climate simulations with the IPSL model. 2. Model-data comparisons. Climate Dynamics 40:2469–2495.

Kageyama, Masa, Nathalie Combourieu Nebout, Pierre Sepulchre, Odile Pey-ron, Gerhard Krinner, Gilles Ramstein, and Jean-Pierre Cazet. 2005. TheLast Glacial Maximum and Heinrich Event 1 in terms of climate and veg-etation around the Alboran Sea: a preliminary model-data comparison.Comptes Rendus Geoscience 337:983–992.

Kandel, Andrew W., and Nicholas J. Conard. 2005. Production sequences ofostrich eggshell beads and settlement dynamics in the Geelbek Dunes ofthe Western Cape, South Africa. Journal of Archaeological Science 32:1711–1721.

Kang, Sarah M., Isaac M. Held, Dargan M. W. Frierson, and Ming Zhao.2008. The response of the ITCZ to extratropical thermal forcing: idealizedslab-ocean experiments with a GCM. Journal of Climate 21:3521–3532.

Klein, Richard G. 2000. Archaeology and the evolution of human behavior.Evolutionary Anthropology 9:17–36.

———. 2009. The human career: human biological and cultural origins. Chi-cago: University of Chicago Press.

Kuhn, Steven L. 2013. Roots of the Middle Paleolithic in Eurasia. CurrentAnthropology 54(suppl. 8):S255–S268.

Kuhn, Steven L., and Mary C. Stiner. 2007. Paleolithic ornaments: implicationsfor cognition, demography and identity. Diogenes 54:40–48.

Kuorikoski, Jaakko. 2009. Two concepts of mechanism: componential causalsystem and abstract form of interaction. International Studies in the Phi-losophy of Science 23:143–160.

Lalueza-Fox, Carles. 2013. Agreements and misunderstandings among threescientific fields: paleogenomics, archaeology, and human paleontology. Cur-rent Anthropology 54(suppl. 8):S214–S220.

Lambeck, Kurt, Tezer M. Esat, and Emma-Kate Potter. 2002. Links betweenclimate and sea levels for the past three million years. Nature 419:199–206.

Langbroek, Marco. 2011. Trees and ladders: a critique of the theory of humancognitive and behavioural evolution in Palaeolithic archaeology. QuaternaryInternational 270:4–14.

S386 Current Anthropology Volume 54, Supplement 8, December 2013

Langley, Michelle C., Christopher Clarkson, and Sean Ulm. 2008. Behavioralcomplexity in Eurasian Neanderthal populations: a chronological exami-nation of the archaeological record. Cambridge Archaeological Journal 18:289–307.

Lemonnier, Pierre. 1986. The study of material culture today: toward ananthropology of technical systems. Journal of Anthropological Archaeology5:147–186.

Leroi-Gourhan, Andre. 1964. Le geste et la parole I: technique et langage. Paris:Albin Michel.

Lombard, Marlize. 2005. Evidence for hunting and hafting during the MiddleStone Age at Sibudu Cave, KwaZulu-Natal, South Africa. Journal of HumanEvolution 48:279–300.

———. 2012. Thinking through the Middle Stone Age of sub-Saharan Africa.Quaternary International 270:140–155.

Lowe, John, Nick Barton, Simon Blockley, Christopher Bronk Ramsey, VictoriaL. Cullen, William Davis, Clive Gamble, et al. 2012. Volcanic ash layersilluminate the resilience of Neanderthals and early modern humans tonatural hazards. Proceedings of the National Academy of Sciences of the USA109:13532–13537.

Machamer, Peter, Lindley Darden, and Carl F. Craver. 2000. Thinking aboutmechanisms. Philosophy of Science 67:1–25.

Mackay, Alex, and Aara Welz. 2008. Engraved ochre from a Middle Stone Agecontext at Klein Kliphuis in the Western Cape of South Africa. Journal ofArchaeological Science 35:1521–1532.

Marean, Curtis W., Miryam Bar-Matthews, Jocelyn Bernatchez, Erich Fisher,Paul Goldberg, Andy I. R. Herries, Zenobia Jacobs, et al. 2007. Early humanuse of marine resources and pigment in South Africa during the MiddlePleistocene. Nature 449:905–908.

Maslin, Mark A., and Beth Christensen. 2007. Tectonics, orbital forcing, globalclimate change, and human evolution in Africa: introduction to the Africanpaleoclimate special volume. Journal of Human Evolution 53:443–464.

Mazza, Paul Peter Anthony, Fabio Martini, Benedetto Sala, Maurizio Magi,Maria Perla Colombini, Gianna Giachi, Francesco Landucci, Cristina Le-morini, Francesca Modugno, and Erika Ribechini. 2006. A new Palaeolithicdiscovery: tar-hafted stone tools in a European Mid-Pleistocene bone-bear-ing bed. Journal of Archaeological Science 33:1310–1318.

McBrearty, Sally, and Alison S. Brooks. 2000. The revolution that wasn’t: anew interpretation of the origin of modern human behavior. Journal ofHuman Evolution 39:453–563.

Mellars, Paul. 1989. Major issues in the origin of modern humans. CurrentAnthropology 30:349–385.

———. 2006. Why did modern human populations disperse from Africa ca.60,000 years ago? a new model. Proceedings of the National Academy ofSciences of the USA 103:9381–9386.

Mellars, Paul, and Chris Stringer. 1989. The human revolution: behaviouraland biological perspectives on the origins of modern humans. Princeton, NJ:Princeton University Press.

Mercader, Julio. 2009. Mozambican grass seed consumption during the MiddleStone Age. Science 326:1680–1683.

Mesoudi, Alex, Andrew Whiten, and Kevin N. Laland. 2006. Towards a unifiedscience of cultural evolution. Behavioral Brain Sciences 29:329–383.

Meyer, M., M. Kircher, M.-T. Gansauge, H. Li, F. Racimo, S. Mallick, J. G.Schraiber, et al. 2012. A high-coverage genome sequence from an archaicDenisovan individual. Science: doi:10.1126/science.1224344.

Mourre, Vincent, Paola Villa, and Christopher S. Henshilwood. 2010. Earlyuse of pressure flaking on lithic artifacts at Blombos Cave, South Africa.Science 330:659–662.

Norenzayan, Ara. 2011. Explaining human behavioral diversity. Science 332:1041–1042.

Nowell, April. 2010. Defining behavioral modernity in the context of Nean-dertal and anatomically modern human populations. Annual Review ofAnthropology 39:437–452.

Odum, Eugene. 1971. Fundamentals of ecology. 3rd edition. Philadelphia: Saun-ders.

Osborne, Anne H., Derek Vance, Eelco J. Rohling, Nick Barton, Mike Rog-erson, and Nuri Fello. 2008. A humid corridor across the Sahara for themigration of early modern humans out of Africa 120,000 years ago. Pro-ceedings of the National Academy of Sciences of the USA 105:16444–16447.

Pawlik, Alfred F., Jurgen P. Thissen. 2011. Hafted armatures and multi-com-ponent tool design at the Micoquian site of Inden-Altdorf, Germany. Journalof Archaeological Science 38:1699–1708.

Pearson, Richard G., Wilfried Thuiller, Miguel B. Araujo, Enrique Martinez-Meyer, Lluıs Brotons, Colin McClean, Lera Miles, Pedro Segurado, Terence

P. Dawson, and David C. Lees. 2006. Model-based uncertainty in speciesrange prediction. Journal of Biogeography 33:1704–1711.

Peeples, Matthew A., C. Michael Barton, and Steven Schmich. 2006. Resiliencelost: intersecting land use and landscape dynamics in the prehistoric south-western United States. Ecology and Society 11(2), article 22. http://www.ecologyandsociety.org/vol11/iss2/art22.

Peresani, Marco, Ivana Fiore, Monica Gala, Matteo Romandini, and AntonioTagliacozzo. 2011. Late Neandertals and the intentional removal of feathersas evidenced from bird bone taphonomy at Fumane Cave 44 ky B.P., Italy.Proceedings of the National Academy of Sciences of the USA 108:3888–3893.

Peterson, A. Townsend. 2003. Predicting the geography of species’ invasionsvia ecological niche modeling. Quarterly Review of Biology 78:419–433.

Peterson, A. Townsend, Monica Papes, and Jorge Soberon. 2008. Rethinkingreceiver operating characteristic analysis applications in ecological nichemodeling. Ecological Modelling 213:63–72.

Peterson, A. Townsend, Jorge Soberon, Richard G. Pearson, Robert P. An-derson, Enrique Martınez-Meyer, Miguel Nakamua, and Miguel B. Araujo.2011. Ecological niches and geographic distributions. Monographs in Popu-lation Biology, no. 49. Princeton, NJ: Princeton University Press.

Pettitt, Paul B. 2011. The Palaeolithic origins of human burial. London: Rout-ledge.

Powell, Adam, Stephen Shennan, and Mark G. Thomas. 2009. Late Pleistocenedemography and the appearance of modern human behavior. Science 324:1298–1301.

Reich, David, Richard E. Green, Martin Kircher, Johannes Krause, Nick Pat-terson, Eric Y. Durand, Bence Viola, et al. 2010. Genetic history of anarchaic hominin group from Denisova Cave in Siberia. Nature 468:1053–1060.

Renfrew, A. Colin. 1977. Space, time and polity. In The Evolution of socialsystems. John Friedman and Michael J. Rowlands, eds. Pp. 89–114. London:Duckworth.

Richerson, Peter J., Robert Boyd, and Robert L. Bettinger. 2009. Culturalinnovations and demographic change. Human Biology 81:211–235.

Roebroeks, Wil, Mark J. Sier, Trine Kellberg Nielsen, Dimitri De Loecker,Josep Maria Pares, Charles E. S. Arps, and Herman J. Mucher. 2012. Useof red ochre by early Neandertals. Proceedings of the National Academy ofSciences of the USA 109:1889–1894.

Schlanger, Nathan. 1994. Mindful technology: unleashing the chaıne operatoirefor an archaeology of mind. In The ancient mind: elements of cognitivearchaeology. Colin Renfrew and Ezra B. W. Zubrow, eds. Pp. 143–151.Cambridge: Cambridge University Press.

Schoon, Michael, Christo Fabricius, John M. Anderies, and Margaret Nelson.2011. Synthesis: vulnerability, traps, and transformations: long-term per-spectives from archaeology. Ecology and Society 16(2), article 24. http://www.ecologyandsociety.org/vol16/iss2/art24.

Sepulchre, Pierre, Gilles Ramstein, Masa Kageyama, Marian Vanhaeren, Ger-hard Krinner, Marıa-Fernanda Sanchez-Goni, and Francesco d’Errico. 2007.H4 abrupt event and late Neanderthal presence in Iberia. Earth and Plan-etary Science Letters 258:283–292.

Shea, John J. 2011. Homo sapiens is as Homo sapiens was. Current Anthropology52:1–35.

Shennan, Stephen. 2001. Demography and cultural innovation: a model andits implications for the emergence of modern human culture. CambridgeArchaeological Journal 11:5–16.

Soberon, Jorge, and A. Townsend Peterson. 2005. Interpretation of models offundamental ecological niches and species’ distributional areas. BiodiversityInformatics 2:1–10.

Soriano, Sylvain, Paola Villa, and Lyn Wadley. 2007. Blade technology andtool forms in the Middle Stone Age of South Africa: the Howiesons Poortand post-Howiesons Poort at Rose Cottage Cave. Journal of ArchaeologicalScience 34:681–703.

Sterelny, Kim. 2011. From hominins to humans: how sapiens became behav-iourally modern. Philosophical Transactions of the Royal Society B 366:809–822.

Steward, Julian. 1955. Theory of culture change. Urbana: University of IllinoisPress.

Stiner, Mary C. 2013. An unshakable Middle Paleolithic? trends versus con-servatism in the predatory niche and their social ramifications. CurrentAnthropology 54(suppl. 8):S288–S304.

Stiner, Mary C., Natalie D. Munro, and Todd A. Surovell. 2000. The tortoiseand the hare: small-game use, the broad-spectrum revolution, and Paleo-lithic demography. Current Anthropology 41:39–73.

Stockwell, David, and David Peters. 1999. The GARP modelling system: prob-

d’Errico and Banks Mechanisms behind Cultural Trajectories S387

lems and solutions to automated spatial prediction. International Journalof Geographic Information Systems 13:143–158.

Texier, Pierre-Jean, Guillaume Porraz, John Parkington, Jean-Philippe Rigaud,Cedric Poggenpoel, Christopher Miller, Chantal Tribolo, et al. 2010. A How-iesons Poort tradition of engraving ostrich eggshell containers dated to60,000 years ago at Diepkloof Rock Shelter, South Africa. Proceedings of theNational Academy of Sciences of the USA 107:6180–6185.

Tingley, Martin P., Peter F. Craigmile, Murali Haran, Bo Li, Elizabeth Mann-shardt, and Bala Rajaratnam. 2012. Piecing together the past: statisticalinsights into paleoclimatic reconstructions. Quaternary Science Reviews 35:1–22.

Tomasello, Michael. 1994. The question of chimpanzee culture. In Chimpanzeecultures. Richard W. Wrangham, W. C. McGrew, Frans B. M. de Waal, andPaul Heltne, eds. Pp. 301–317. Cambridge: Harvard University Press.

Trinkaus, Erik. 2005. Early modern humans. Annual Review of Anthropology34:207–230.

Van Andel, Tjeerd H., and William Davies, eds. 2003. Neanderthals and modernhumans in the European landscape during the last glaciation: archaeologicalresults of the stage 3 project. Cambridge: McDonald Institute for Archaeo-logical Research.

Vanhaeren, Marian, Francesco d’Errico, Chris Stringer, Sarah L. James, Jon-athan Todd, and Henk K. Mienis. 2006. Middle Paleolithic shell beads inIsrael and Algeria. Science 312:1785–1788.

Vanhaeren, Marian, Francesco d’Errico, Karen L. van Niekerk, Christopher S.Henshilwood, and Rudolph M. Erasmus. 2013. Thinking strings: additionalevidence for personal ornament use in the Middle Stone Age at BlombosCave, South Africa. Journal of Human Evolution 64:500–517.

Villa, Paola, Marie Soressi, Christopher S. Henshilwood, and Vincent Mourre.2009. The Still Bay points of Blombos Cave (South Africa). Journal ofArchaeological Science 36:441–460.

Villa, Paola, and Sylvain Soriano. 2010. Hunting weapons of Neanderthalsand early modern humans in South Africa. Journal of Anthropological Re-search 66:5–38.

Villa, Paola, Sylvain Soriano, Nicolas Teyssandier, and Sara Wurz. 2010. TheHowiesons Poort and MSA III at Klasies River main site, Cave 1A. Journalof Archaeological Science 37:630–655.

Villa, Paola, Sylvain Soriano, Tsenka Tsanova, Ilaria Degano, Thomas F. G.Higham, Francesco d’Errico, Lucinda Backwell, Jeannette J. Luceiko, MariaPerla Colombini, and Peter B. Beaumont. 2012. Border Cave and the be-ginning of the Later Stone Age in South Africa. Proceedings of the NationalAcademy of Sciences of the USA 109:13208–13213.

Vrac, Mathieu, Michael Stein, and Katharine Hayhoe. 2007. Statistical down-scaling of precipitation through nonhomogeneous stochastic weather typ-ing. Climate Research 34:169–184.

Wadley, Lyn. 2010a. Compound-adhesive manufacture as a behavioral proxyfor complex cognition in the Middle Stone Age. Current Anthropology51(suppl. 1):S111–S119.

———. 2010b. A taphonomic study of ochre demonstrates post-depositionalcolor transformations. Journal of Taphonomy 8:243–254.

———. 2010c. Were snares used in the Middle Stone age and does it matter?a review and case study from Sibudu, South Africa. Journal of HumanEvolution 58:179–192.

Wadley, Lyn, Tamaryn Hodgskiss, and Michael Grant. 2009. Implications forcomplex cognition from the hafting of tools with compound adhesives inthe Middle Stone Age, South Africa. Proceedings of the National Academyof Sciences of the USA 106:9590–9594.

Wadley, Lyn, Christine Sievers, Marion Bamford, Paul Goldberg, FrancescoBerna, and Christopher Miller. 2011. Middle Stone Age bedding construc-tion and settlement patterns at Sibudu, South Africa. Science 334:1388–1391.

Walker, Brian, C. S. Holling, Stephen R. Carpenter, and Ann Kinzig. 2004.Resilience, adaptability and transformability in social-ecological systems.Ecology and Society 9(2), article 5. http://www.ecologyandsociety.org/vol9/iss2/art5.

Warren, Dan L., Richard E. Glor, and Michael Turelli. 2010. ENMTools: atoolbox for comparative studies of environmental niche models. Ecography33:607–611.

Watts, Ian. 2010. The pigments from Pinnacle Point Cave 13B, Western Cape,South Africa. Journal of Human Evolution 59:392–411.

Weaver, Timothy D., and Charles C. Roseman. 2008. New developments inthe genetic evidence for modern human origins. Evolutionary Anthropology17:69–80.

Wurz, Sara, N. J. le Roux, S. Gardner, and Hilary J. Deacon. 2003. Discrim-inating between the end products of the earlier Middle Stone Age sub-stages at Klasies River using biplot methodology. Journal of ArchaeologicalScience 30:1107–1126.

Wynn, Thomas, and Frederick L. Coolidge. 2010. Beyond symbolism andlanguage. Current Anthropology 51(suppl. 1):S5–S16.

Zilhao, Joao. 2001. Anatomically archaic, behaviorally modern: the last Nean-derthals and their destiny. Amsterdam: Stichting Nederlands Museum voorAnthropologie en Praehistoriae.

———. 2006. Neandertals and moderns mixed, and it matters. EvolutionaryAnthropology 15:183–195.

———. 2007. The emergence of ornaments and art: an archaeological per-spective on the origins of “behavioral modernity.” Journal of ArchaeologicalResearch 15:1–54.

———. 2011. The emergency of language, art, and symbolic thinking. InHomo symbolicus: the dawn of language, imagination and spirituality. Chris-topher S. Henshilwood and Francesco d’Errico, eds. Pp. 111–131. Amster-dam: John Benjamins.

Zilhao, Joao, Diego E. Angelucci, Ernestina Badal-Garcıa, Francesco d’Errico,Floreal Daniel, Laure Dayet, Katerina Douka, et al. 2010. Symbolic use ofmarine shells and mineral pigments by Iberian Neandertals. Proceedings ofthe National Academy of Sciences of the USA 107:1023–1028.

S388 Current Anthropology Volume 54, Supplement 8, December 2013

� 2013 by The Wenner-Gren Foundation for Anthropological Research. All rights reserved. 0011-3204/2013/54S8-0016$10.00. DOI: 10.1086/673881

Population Size as an Explanation forPatterns in the Paleolithic

Archaeological RecordMore Caution Is Needed

by Mark Collard, Briggs Buchanan, and Michael J. O’Brien

Recently it has become commonplace to use population size to explain patterns in the Paleolithic archaeologicalrecord. Several modeling studies support the idea that population size can affect cultural evolution, but the resultsof empirical studies are ambiguous. Here we report a study that used tool kit data from recent hunter-gatherers,in conjunction with correlation analysis and a global sample, a continental sample, and a regional sample. Theresults of the analyses do not support the hypothesis. Population size was correlated with some tool kit variablesin the global sample, but these relationships disappeared when two factors that have previously been found to affecthunter-gatherer tool kits—risk of resource failure and mobility—were controlled for. Population size was notcorrelated with the tool kit variables in the other samples. The regression analyses also did not support the populationsize hypothesis. Together, these results challenge the use of population size to explain patterns in the Paleolithicarchaeological record. Population size may explain some of the patterns in question, but this needs to be demonstratedthrough tests in which the population size hypothesis is explicitly pitted against competing hypotheses, such asadaptation to shifting ecological conditions.

Introduction

Recently a number of researchers have argued that populationsize might explain several long-debated patterns in the Pa-leolithic archaeological record. Shennan (2001), for example,has suggested that the so-called creative explosion of the lateMiddle Stone Age and Upper Paleolithic might have resultedfrom a large, climate-driven increase in population size. Sim-ilarly, Riede (2008) has argued that the emergence of theBromme and Perstunian technocomplexes in Northern Eu-rope during the Late Glacial period was driven by populationsize reduction associated with the Laacher See eruption. Pow-ell, Shennan, and Thomas (2009) have proposed that pop-ulation increase might also explain why many cultural in-

Mark Collard is Canada Research Chair and Professor, and BriggsBuchanan is Postdoctoral Fellow, at the Human Evolutionary StudiesProgram and Department of Archaeology of Simon Fraser University(8888 University Drive, Burnaby, British Columbia V5A 1S6, Canada[mcollard@sfu.ca]). Michael J. O’Brien is Dean of the College ofArts and Science, Professor of Anthropology, and Director of theMuseum of Anthropology in the Department of Anthropology ofthe University of Missouri (107 Swallow Hall, Columbia, Missouri65211-1440, U.S.A.). This paper was submitted 3 VII 13, accepted 4IX 13, and electronically published 20 XII 13.

novations seem to have appeared, disappeared, and thenreappeared during the late Pleistocene. Premo and Kuhn(2010) have argued that two key features of the Middle Pa-leolithic and Middle Stone Age archaeological records—anabsence of directional technological change and the reap-pearance of previously existing cultural behaviors—might bea function of a high rate of extirpation of small, isolatedgroups and subsequent repopulation.

Support for these hypotheses comes from a number offormal models that suggest population size can have a sig-nificant effect on the evolution of fitness-relevant culturaltraits. The earliest of these models was described by Shennan(2001). Shennan modified a population-genetics model toincorporate social learning among individuals and then car-ried out a series of simulation trials. He found that largerpopulations have a major advantage over smaller ones whenit comes to adaptive cultural innovation because of the de-creasing role of sampling effects as populations grow. Hisresults suggested that when effective population size is large,there is a far greater probability of fitness-enhancing culturalinnovations being maintained and deleterious ones being lostthan when effective populations are small. In the latter sit-uation, innovations that are maintained tend to be less ben-eficial in terms of reproduction and also less attractive forimitators.

Collard, Buchanan, and O’Brien Population Size and the Paleolithic Archaeological Record S389

Other models that demonstrate that population size canaffect the evolution of fitness-relevant traits have been re-ported by Henrich (2004); Powell, Shennan, and Thomas(2009); Mesoudi (2011); and Kobayashi and Aoki (2012).Henrich (2004) argued that population size can affect theprobability of more complex skills being invented and main-tained. In his model, learners preferentially copy the mostskilled practitioner in their population with some amount oferror. The probability distribution that determines theamount of error is such that a learner will only occasionallyarrive at a behavior that gives a better result than the previousbest. The likelihood of this occurring is dependent on pop-ulation size because in large populations even improbableevents occur now and again, and the larger the population,the more likely this is. Powell, Shennan, and Thomas (2009)implemented Henrich’s (2004) model with a spatially struc-tured metapopulation and found that contact and migrationaffect cultural evolution in a similar manner to increase inpopulation size. Mesoudi (2011) showed that Henrich’s(2004) results could be replicated when acquisition costs areallowed to increase as skill level increases. Kobayashi and Aoki(2012) modified Henrich’s (2004) model to examine the ef-fects of overlapping generations and found that the effects ofpopulation size on cultural evolution are amplified when gen-erational overlap is taken into account.

The idea that population size can affect cultural evolutionhas also been supported by formal and agent-based modelsinvolving selectively neutral traits. For example, Neiman(1995) investigated the amount of variation to be expectedin the decoration of a pottery assemblage if the motifs areneutral in terms of adaptation and showed that random loss,or “drift,” destroys variation more quickly in smaller popu-lations than in larger ones. More recently, Premo and Kuhn(2010) used an agent-based model to show that local groupextinction can reduce cultural richness and complexity evenwhen cultural traits do not affect fitness.

The situation with regard to empirical support for the pop-ulation size hypothesis is more complicated. To date, onlyPowell, Shennan, and Thomas (2009) have attempted to testthe hypothesis with Paleolithic archaeological data. They usedmolecular data to estimate when different regions of the worldwould have reached the same population density as Europeat the start of the Upper Paleolithic and then compared thoseestimates with the timing of the appearance of markers ofmodern behavior in the regions. Their results were mixed.They found a reasonable correspondence between the timingof the crossing of the density threshold and the timing of theappearance of markers of modern behavior in sub-SaharanAfrica, North Africa, and the Levant, but there was a consid-erable gap between the timing of the crossing of the densitythreshold and the appearance of markers of modern behaviorin southern, northern, and central Asia. As such, Powell,Shennan, and Thomas’s (2009) results only partially supportthe population size hypothesis.

A number of other empirical studies have a bearing on the

hypothesis (Collard, Kemery, and Banks 2005; Collard et al.2013a, 2013b; Kline and Boyd 2010; Neiman 1995; Nelson etal. 2011). Some studies support it. Neiman (1995) investigatedthe amount of variation to be expected in the decoration ofa pottery assemblage if the motifs are neutral in terms ofadaptation. He then analyzed rim decoration variation amongseven successive phases of the Woodland period in Illinois,United States, and found that it matched the expectations ofhis model. He concluded that the patterns of variation de-pended on changing levels of intergroup contact, whichstarted low, increased, and then declined again. Kline andBoyd (2010) examined the effect of population size on marineforaging tool kits of 10 recent nonindustrial farming popu-lations from Oceania and found that population size had asignificant effect on both the number of tools and the averagenumber of parts per tool. Collard et al. (2013a) applied simplelinear and stepwise multiple regression analysis to data from45 nonindustrial farming and pastoralist groups to test thehypothesis. Results of the analyses were consistent with thepredictions of the hypothesis: both the number of tools andthe number of tool parts were positively and significantlyinfluenced by population size in the simple linear regressionanalyses. The multiple regression analyses demonstrated thatthese correlations were independent of the effects of risk ofresource failure. Collard et al. (2013a) concluded from thisthat population size influences cultural evolution in recentnonindustrial food-producing populations.

Other empirical studies do not support the population sizehypothesis. Collard, Kemery, and Banks (2005) included pop-ulation size in a study designed to shed light on the driversof food-getting tool kit structure among recent hunter-gath-erers. The core of their data set comprised counts of thenumber of tools and tool parts in the tool kits of a worldwidesample of 20 recent hunter-gatherer populations. Their resultsdid not support the population size hypothesis. The onlyvariables that had a significant effect on the tool kit structuremeasures were measures of risk of resource failure, effectivetemperature, and net aboveground productivity. Read (2008)also tested the hypothesis as part of an investigation intofactors that drive variation in tool kit structure among recenthunter-gatherers. Like Collard, Kemery, and Banks (2005), hefound no support for the hypothesis. Nelson et al. (2011)used archaeological data to examine the relationship betweenregional population density and the number of pottery waresin the U.S. Southwest between 1000 CE and 1600 CE. Theyfound population density and pottery richness to be inverselycorrelated and argued that this indicates that social confor-mity becomes increasingly important as population densityincreases. Although Nelson and colleagues do not discuss thepopulation size hypothesis, their results are clearly not con-sistent with it. Finally, Collard et al. (2013b) tested the hy-pothesis as part of a study that focused on the drivers oftechnological richness among 85 recent hunter-gatherergroups from western North America. They found that thetotal number of material items and techniques was correlated

S390 Current Anthropology Volume 54, Supplement 8, December 2013

with both a proxy for environmental risk—mean rainfall forthe driest month—and population size. However, the direc-tion of the relationship was the opposite of the one predictedby the population size hypothesis (it was negative rather thanpositive).

Currently, then, the situation with respect to empirical sup-port for the population size hypothesis is mixed. Several stud-ies support it, several refute it, and one study has yieldedambiguous results. Given this, and given the potential im-portance of the hypothesis for our understanding of the Pa-leolithic and cultural evolution in general, we decided to re-visit the relationship between population size and the numberand intricacy of the food-getting tools used by recent hunter-gatherer populations. Collard et al. (2011) recently reportedan analysis that suggests the effect of risk of resource failureon hunter-gatherer food-getting technology is dependent onthe scale of risk differences among populations. They foundthat when there are large differences in risk of resource failureamong populations, risk has a significant effect on the numberand intricacy of the food-getting tools used by hunter-gath-erers. When differences in risk of resource failure are small,in contrast, risk does not have a significant effect on thestructure of hunter-gatherers’ food-getting tool kits. This find-ing suggests that the conflicting results of studies that havetested the population size hypothesis with ethnographic andarchaeological data may be more apparent than real. Specif-ically, it raises the possibility that the studies that have failedto support the hypothesis have done so because they haveemployed samples in which there are large risk differences,and therefore the impact of population size on tool kit struc-ture has been obscured by the impact of risk. With this inmind, we tested the hypothesis with samples spanning threelevels of among-population risk difference: a global sampleconsisting of populations from several continents, a conti-nental sample comprising populations from North America,and a regional sample made up of populations from the Pa-cific Northwest.

Material and Methods

Oswalt (1973, 1976) developed the method we used in thestudy. The method focuses on tools employed directly in theacquisition of food, which Oswalt termed “subsistants.” Os-walt divided subsistants into four categories: instruments,weapons, tended facilities, and untended facilities. Instru-ments are used to procure food that cannot run away orthreaten its pursuer, such as plants or sessile animals. A dig-ging stick is an example of an instrument. Weapons are de-signed to kill or maim potential prey that can escape or mayharm its pursuer. Weapons include boomerangs, crossbows,and harpoons. Facilities are structures that control the move-ment of animals or protect them to a human’s advantage,such as a fish weir or a livestock pen. Tended facilities requirecontinuous monitoring while in use (e.g., a fishhook), whereasuntended facilities are capable of functioning without a hu-

man present and require only occasional monitoring (e.g., adeadfall trap). Oswalt created a further distinction betweensimple and complex subsistants. Simple subsistants do notchange structurally during use, whereas complex subsistantshave multiple parts that change position relative to one an-other during use.

Oswalt (1973, 1976) devised three measures of tool kitstructure. One is the total number of subsistants, which Os-walt suggested is an indicator of the size of a tool kit. Otherresearchers have referred to the total number of subsistantsas tool kit “diversity” (Collard, Kemery, and Banks 2005; Col-lard et al. 2011; Shott 1986; Torrence 1983, 1989), but thisterm is potentially confusing. In ecology, “diversity” has twodimensions: “richness” and “evenness.” The former refers tothe number of taxa in a community, landscape, or region;the latter refers to how close the taxa in a community, land-scape, or region are in terms of numbers of individuals (Col-well 2009). The dimension of species diversity that the variable“total number of subsistants present in a tool kit” is akin tois clearly “species richness.” Thus, to reduce the potential forconfusion, here we refer to the total number of subsistantsas “tool kit richness” rather than “tool kit diversity.” Anotherof Oswalt’s measures of tool kit structure is the total numberof techno-units. Formally, a techno-unit is an “integrated,physically distinct, and unique structural configuration thatcontributes to the form of a finished artifact” (Oswalt 1976:38), but in simpler terms, techno-units are the different kindsof parts of a tool. The total number of techno-units includedin a tool kit is a measure of its “complexity” (Collard, Kemery,and Banks 2005; Collard et al. 2011; Oswalt 1973, 1976; Torr-ence 1983, 1989). Oswalt’s third measure of tool kit structureis the average number of techno-units per subsistant, whichis calculated by dividing the total number of techno-units ina tool kit by its richness. Again, this is a measure of tool kitcomplexity.

Using Oswalt’s (1973, 1976) method, we generated valuesfor total number of subsistants (STS), total number of techno-units (TTS), average number of techno-units per tool (AVE),and population size (POP) for a sample of 49 contact-erahunter-gatherer populations. Thirty of the populations arefrom North America, five are from South America, five arefrom Africa, five are from Asia, and four are from Oceania.The names and locations of the populations are given in table1. The majority of the tool kit data was taken from previousstudies that have used Oswalt’s (1973, 1976) method to quan-tify tool kit structure (Collard, Kemery, and Banks 2005; Col-lard et al. 2011). These data were supplemented with STS,TTS, and AVE values generated specifically for this study. Thesources from which the latter data were extracted vary in agefrom the late 1800s to the mid-20th century. The values forPOP were taken from Binford (2001). The POP data weretransformed to base e because the POP hypothesis predicts aconcave relationship between POP and tool kit richness andcomplexity. Log-transforming POP made the expected rela-tionship between POP and each tool kit measure a linear one.

Collard, Buchanan, and O’Brien Population Size and the Paleolithic Archaeological Record S391

Table 1. Names and locations of hunter-gatherergroups included in the samples

Name Location

Mbuti AfricaHadza Africa!Kung San AfricaNharo AfricaG/Wi AfricaPunan AsiaGreat Andamanese AsiaVeddas AsiaChenchu AsiaYukaghir AsiaCopper Inuit North America, arcticIglulik North America, arcticNetsilik North America, arcticAngmagsalik North America, arcticTareumiut North America, arcticTwana North America, Pacific NorthwestNootka North America, Pacific NorthwestQuinalt North America, Pacific NorthwestUpper Stalo North America, Pacific NorthwestCoast Salish North America, Pacific NorthwestMakah North America, Pacific NorthwestKwakiutl North America, Pacific NorthwestTlingit North America, Pacific NorthwestKlamath North America, plateauLillooet North America, plateauCoeur D’Alene North America, plateauOkanagan North America, plateauSanpoil-Nespelem North America, plateauShuswap North America, plateauOwens Valley Paiute North America, SouthwestSurprise Valley Paiute North America, SouthwestFort Nelson Slave North America, subarcticKaska North America, subarcticCarrier North America, subarcticLower Koyukon North America, subarcticChipewyan North America, subarcticIngalik North America, subarcticNabesna North America, subarcticTanaina North America, subarcticCaribou Inuit North America, subarcticTiwi OceaniaGroote-eylandt OceaniaNorthern Arenda OceaniaTasmanians OceaniaYaruro South AmericaSiriono South AmericaBotocudo South AmericaOna South AmericaYahgan South America

Some researchers contend that the technological variablesshould also be logged when testing the population size hy-pothesis. We tried this approach as well, and the results werenot qualitatively different.

In addition to generating technological and POP data, weobtained values for three measures of risk of resource failureand two measures of residential mobility. We did so because,as mentioned earlier, some previous tests of the POP hy-

pothesis found that POP did not affect tool kit structure whenmeasures of risk of resource failure and measures of mobilitywere included in the analysis (Collard, Kemery, and Banks2005; Read 2008). The proxies for risk of resource failure wereeffective temperature (ET), net aboveground productivity(NAGP), and mean rainfall for the wettest month of the year(RHIGH). Also known as “warmth,” ET was developed tobetter understand the effect of temperature on the distributionof living and fossil plants (Bailey 1960). It is defined as thetemperature characteristic of the start and finish of the periodin which plant growth occurs (Bailey 1960). NAGP is theamount of new cell life that is added to a given location byphotosynthesis and growth in a year (measured in grams persquare meters per year; Binford 2001). The measures of res-idential mobility we included were number of residentialmoves per year (NOMOV) and total distance moved per yearduring residential moves (DISMOV). The values for ET,NAGP, RHIGH, NOMOV, and DISMOV were obtained fromBinford (2001).

After compiling the data set, we used the Kolmogorov-Smirnov test to assess how closely the variables approximatea normal distribution. In the global sample, ET, NAGP,RHIGH, and DISMOV were found to have distributions thatdeparted significantly from normal and thus were trans-formed. We transformed NAGP, RHIGH, and DISMOV usingthe natural log. To transform ET, we applied the Box-Coxpower transformation in Minitab. Because a negative valuewas selected as the λ parameter (�1.910), all values were thensubtracted from 1. In the North American sample, the onlyvariable whose distribution departed significantly from a nor-mal distribution was DISMOV. We transformed it using thenatural log. None of the distributions were significantly dif-ferent from normal in the Pacific Northwest sample.

After completing the transformations, we carried out threesets of analyses. In the first, we used simple correlation analysisto assess the direction and strength of the correlation betweenpopulation size and each of the three tool kit variables (STS,TTS, AVE). Here, as in the other tests, we began with theglobal sample, then analyzed the North American sample, andthen the Pacific Northwest sample. The test prediction wasthat the relationships between tool kit variables and popu-lation size should be both positive and statistically significant.Because multiple tests were conducted, Benjamini and Yek-utieli’s (2001) method of significance level correction was usedto reduce Type I error rates. We employed this method ratherthan the better-known Bonferroni correction because it hasbeen shown to balance the reduction of Type I and Type IIerror rates better than the Bonferroni correction (Narum2006).

In the second set of analyses, partial correlation analysiswas used to assess the direction and strength of the correlationbetween population size and each of the three tool kit variableswhile controlling for the risk variables (ET, NAGP, RHIGH)and the mobility variables (NOMOV, DISMOV). The testprediction was the same as the one in the previous set of

S392 Current Anthropology Volume 54, Supplement 8, December 2013

Table 2. Summary of results of simple correla-tion analyses carried out to assess the strength ofthe relationship between population size (POP)and tool kit richness and complexity

Sample and variablescorrelated r P

Global ( ):n p 49STS, POP .353 .013a

TTS, POP .361 .011a

AVE, POP .179 .219North America ( ):n p 30

STS, POP .318 .086TTS, POP .383 .037AVE, POP .230 .221

Pacific Northwest ( ):n p 14STS, POP .455 .102TTS, POP .525 .054AVE, POP .204 .484

Note. STS p total number of subsistants; TTS p totalnumber of techno-units; AVE p average number of techno-units per tool.a Significant correlation using Benjamini and Yekutieli’s(2001) alpha correction (the critical value for three tests is

).α p .027

Table 3. Summary of results of partial correlation analysescarried out to assess the strength of the relationship be-tween population size (POP) and tool kit richness andcomplexity while controlling for variables that have previ-ously been found to influence the richness and complexityof hunter-gatherer tool kits

Sample, variables correlated, andvariables controlled for r P

Global ( ):n p 49STS, POP:

ET,a NAGP,b DISMOV,b RHIGH,b NOMOV .246 .107TTS, POP:

ET,a NAGP,b DISMOV,b RHIGH,b NOMOV .248 .104AVE, POP:

ET,a NAGP,b DISMOV,b RHIGH,b NOMOV �.017 .911North America ( ):n p 30

STS, POP:ET, NAGP, DISMOV,b RHIGH, NOMOV .289 .160

TTS, POP:ET, NAGP, DISMOV,b RHIGH, NOMOV .349 .087

AVE, POP:ET, NAGP, DISMOV,b RHIGH, NOMOV .178 .395

Pacific Northwest ( ):n p 14STS, POP:

ET, NAGP, DISMOV, RHIGH, NOMOV .428 .250TTS, POP:

ET, NAGP, DISMOV, RHIGH, NOMOV .179 .491AVE, POP:

ET, NAGP, DISMOV, RHIGH, NOMOV .339 .372

Note. STS p total number of subsistants; TTS p total number oftechno-units; AVE p average number of techno-units per tool; ET peffective temperature; NAGP p net aboveground productivity; RHIGHp rainfall for the wettest month of the year; NOMOV p number ofresidential moves per year; DISMOV p total distance moved per yearduring residential moves.a Transformed with Box-Cox method and reciprocal taken before anal-ysis; see “Material and Methods” section for details.b Converted to natural logarithm before analysis; see “Material and Meth-ods” section for details.

analyses: relationships between the tool kit variables and pop-ulation size should be both positive and statistically signifi-cant.

In the third set of analyses, we used standard multipleregression analysis to assess the importance of population sizeas an influence on tool kit richness and complexity comparedwith the risk and mobility variables. The tool kit variableswere the dependent variables, and population size, ET, NAGP,RHIGH, DISMOV, and NOMOV were the independent var-iables. When the variance inflation factor (VIF) for two ormore variables exceeded 10, the variable with the highest VIFwas removed and the analysis rerun. PASW (SPSS) 19 wasused to carry out all the analyses.

Results

Results of the simple correlation analyses are summarized intable 2. In the analyses of the global sample, STS and TTSwere significantly correlated with POP, but AVE was not. Noneof the tool kit variables were significantly correlated with POPin the analyses of the North American sample or for the PacificNorthwest sample.

Table 3 summarizes results of the partial correlation anal-yses. None of the tool kit variables were significantly corre-lated with POP in the analyses of the global sample. Theresults of the analyses using the North American sample weresimilar: None of the tool kit variables were significantly cor-related with POP. The results of the analyses that focused onthe Pacific Northwest sample were consistent with the resultsof the analyses of the other two samples. Once again, noneof the tool kit variables were significantly correlated with POP.

The results of the regression analyses are summarized in

tables 4–6. The only significant influences on tool kit richnessand diversity were the risk variables ET and RHIGH and themobility variables NOMOV and DISMOV. POP was not asignificant influence on any of the tool kit variables in anyof the samples; it consistently had one of the lowest stan-dardized beta coefficients; and it was always either the fourthor fifth lowest of the six independent variables.

Discussion and Conclusions

POP was correlated with two of the three tool kit variablesin the global sample, but these relationships disappeared whenpartial correlation analysis was used to control for risk ofresource failure and mobility, both of which have previouslybeen found to influence tool kit richness and complexityamong hunter-gatherers. POP was not correlated with any ofthe tool kit variables in the North American and PacificNorthwest samples regardless of which form of correlationanalysis was used. The regression analyses were consistent with

Collard, Buchanan, and O’Brien Population Size and the Paleolithic Archaeological Record S393

Table 4. Summary of results of standard multiple regression analyses using the global sample ( ) carried out ton p 49assess the relative importance of various variables as drivers of tool kit richness and complexity

Dependentvariable Full model POP ET NAGP RHIGH NOMOV DISMOV

STS F p 6.199,df p6, 42,P p .000,a

r2 p .470

β p .195,P p .107,VIF p 1.116

β p �.658,P p .002,b

VIF p 3.310

β p .204,P p .379,VIF p 4.183

β p �.264,P p .039,b

VIF p 1.215

β p �.265,P p .143,VIF p 2.495

β p .011,P p .954,VIF p 2.940

TTS F p 7.745,df p 6, 42,P p .000,a

r2 p .525

β p .187,P p .104,VIF p 1.116

β p �.740,P p .000,b

VIF p 3.310

β p .271,P p .220,VIF p 4.183

β p �.298,P p .015,b

VIF p 1.215

β p �.226,P p .186,VIF p 2.495

β p �.146,P p .428,VIF p 2.940

AVE F p 5.469,df p 6, 42,P p .000,a

r2 p .439

β p �.014,P p .911,VIF p 1.116

β p �.544,P p .013,b

VIF p 3.310

β p �.012,P p .959,VIF p 4.183

β p �.222,P p .088,VIF p 1.215

β p �.097,P p.598,VIF p 2.495

β p �.409,P p .045,b

VIF p 2.940

Note. STS p total number of subsistants; POP p population size; TTS p total number of techno-units; AVE p average number of techno-unitsper tool; ET p effective temperature; NAGP p net aboveground productivity; RHIGH p rainfall for the wettest month of the year; NOMOV pnumber of residential moves per year; DISMOV p total distance moved per year during residential moves; VIF p variance inflation factor.a Significant correlation using Benjamini and Yekutieli’s (2001) alpha correction (the critical value for three tests is ).α p .027b Significant at .P ≤ .05

the results of the partial correlation analyses: POP was not asignificant influence on any of the tool kit variables in anyof the three samples and consistently had one of the loweststandardized beta coefficients. Thus, the analyses did not sup-port the POP hypothesis. Even when the influence of thefactor that previously has been found to most affect tool kitrichness and complexity among hunter-gatherers—risk of re-source failure—was minimized, there was no evidence thatPOP influenced tool kit richness and complexity.

This means that there are now four empirical studies thatsupport the POP hypothesis (Collard et al. 2013a; Kline andBoyd 2010; Neiman 1995; Powell, Shennan, and Thomas2009) and four that do not (Collard, Kemery, and Banks 2005;Collard et al. 2013b; Nelson et al. 2011; this study). There aretwo basic potential explanations for this disagreement. Oneis that the studies that have failed to support the hypothesissuffer from shortcomings that are sufficiently serious to haveresulted in Type II errors; that is, the studies’ failure to supportthe hypothesis is a false negative. The other is that the resultsof the studies that have failed to support the hypothesis arereliable and the hypothesis needs modification.

Regarding the first possibility, there are three potentialshortcomings that need to be evaluated. One is the accuracyof the population estimates. Henrich (2006), Kline and Boyd(2010), and Boyd, Richerson, and Henrich (2013) have arguedthat the studies by Collard, Kemery, and Banks (2005) andRead (2008) failed to support the hypothesis because they didnot take into account cultural transmission among popula-tions and therefore did not accurately measure the effectivePOP for cultural traits. This is unlikely. Population values forboth studies were generated in the same way as the valuesused by Collard et al. (2013a), which supported the hypoth-esis. That Collard et al.’s (2013a) study supported the hy-pothesis implies that the method of collecting population data

is adequate. Additionally, one of the other studies that havefailed to support the POP hypothesis (Collard et al. 2013b)cannot be criticized for not taking into account cultural trans-mission among populations. It controlled for cultural trans-mission and still failed to find support for the hypothesis.Thus, use of inadequate estimates of POP seems unlikely toexplain the failure of Collard, Kemery, and Banks’s (2005),Read’s (2008), Collard et al.’s (2013b), and this study to sup-port the POP hypothesis.

A second potential shortcoming that needs to be consideredis sample size. In principle, it is possible that the studies thathave failed to support the hypothesis have done so becausethe samples they used were too small to pick up the influenceof POP, but this seems unlikely. Samples employed in thestudies that have tested the hypothesis with ethnographic dataand found support for it comprised 10 and 45 populations,respectively (Collard et al. 2013a; Kline and Boyd 2010). Themajority of the samples that have failed to support the hy-pothesis are larger than Kline and Boyd’s (2010) sample, andsome of them are larger than Collard et al.’s (2013a) sample.The samples used by Collard, Kemery, and Banks (2005) andby Read (2008) comprised 20 hunter-gatherer populations.Collard et al.’s (2013b) sample consisted of 85 populations.Samples employed in the study reported here range in sizefrom 14 populations to 49 populations. As such, it is unlikelythat small sample size explains the failure of the studies ofCollard, Kemery, and Banks (2005), Read (2008), Collard etal. (2013b), and the one reported here to support the POPhypothesis.

A third potential shortcoming concerns sample bias. Inaddition to suggesting that Collard, Kemery, and Banks’s(2005) population estimates are inaccurate, Kline and Boyd(2010) and Boyd, Richerson, and Henrich (2013) claim thatCollard and colleagues’ results are unreliable because North

S394 Current Anthropology Volume 54, Supplement 8, December 2013

Table 5. Summary of results of standard multiple regression analyses using the North American sample ( ) carriedn p 30out to assess the relative importance of various variables as drivers of tool kit richness and complexity

Dependentvariable Full model POP ET NAGP RHIGH NOMOV

STS F p 1.526,df p 5, 24,P p.219,r2 p .241

β p .239,P p .218,VIF p 1.125

β p �.251,P p .319,VIF p 1.923

β p .810,P p .094,VIF p 6.813

β p �.479,P p .241,VIF p 5.026

β p �.116,P p .583,VIF p 1.371

TTS F p 3.351,df p 5, 24,P p .020,a

r2 p .411

β p .283,P p .102,VIF p 1.125

β p �.441,P p .054,VIF p 1.923

β p .675,P p .112,VIF p 6.813

β p �.291,P p .416,VIF p 5.026

β p �.337,P p .079,VIF p 1.371

AVE F p 3.892,df 5, 24,P p .010,a

r2 p .448

β p .146,P p .372,VIF p 1.125

β p �.493,P p .028,b

VIF p 1.923

β p .117,P p .771,VIF p 6.813

β p .151,P p .662,VIF p 5.026

β p �.500,P p .010,b

VIF p 1.371

Note. DISMOV was excluded from the analysis because of multicollinearity with NOMOV (see “Material and Methods” for details). STS p totalnumber of subsistants; POP p population size; TTS p total number of techno-units; AVE p average number of techno-units per tool; ET peffective temperature; NAGP p net aboveground productivity; RHIGH p rainfall for the wettest month of the year; NOMOV p number ofresidential moves per year; DISMOV p total distance moved per year during residential moves; VIF p variance inflation factor.a Significant using Benjamini and Yekutieli’s (2001) alpha correction (the critical value for three tests is ).α p .027b Significant at .P ≤ .05

American populations dominate their sample. Collard et al.’s(2013b) sample consists solely of North American popula-tions, and North American populations also dominate thesample used in the study reported here. Thus, Kline andBoyd’s (2010) and Boyd, Richerson, and Henrich’s (2013)concerns can be extended to the other studies that have failedto support the hypothesis, but there are reasons to think theirconcerns are unwarranted. To begin with, there is an impor-tant corollary to the idea that the studies that have failed tosupport the hypothesis have done so because North Americansamples dominate the samples. The corollary is that the hy-pothesis does not apply to North American populations butrather to populations from other regions of the world. Thus,even if the sample-bias argument were correct, it would re-quire us to revise the POP hypothesis to explain its failure toapply to North America populations. In other words, thesample-bias argument simply changes the scope of the ref-utation of the hypothesis rather than explaining away thefailure to support the hypothesis.

Another reason to reject the claims of Kline and Boyd(2010) and Boyd, Richerson, and Henrich (2013) is that weobtained similar results with a balanced global sample to ouroriginal global sample. We used a random-number generatorto select one population from each of five culture regionsrepresented among the North American populations in thesample. We then deleted the other 25 North American pop-ulations. This left us with 24 populations: five from NorthAmerica, five from South America, five from Africa, five fromAsia, and four from Oceania. We then repeated the partialcorrelation analyses in which we correlated STS, TTS, andAVE with POP while controlling for ET, NAGP, RHIGH,NOMOV, and DISMOV. STS was not positively and signifi-cantly correlated with POP ( , ), nor wasr p 0.193 P p .428

TTS ( , ) or AVE ( , ).r p 0.185 P p .448 r p �0.100 P p .683Thus, the balanced global sample did not support the POPhypothesis. This indicates that the failure of this study tosupport the hypothesis cannot be explained away as a con-sequence of sample bias and suggests the same holds for theother studies that have not supported it.

It appears, then, that the disagreement between the em-pirical studies that support the hypothesis and those that donot is not a consequence of the latter studies suffering fromshortcomings that are sufficiently serious to have resulted inType II errors. Rather, it appears that the disagreement issubstantive.

What might account for the disagreement? So far, we havebeen able to identify four potential answers to this question.One concerns the mode of production. Samples that havesupported the hypothesis comprise populations that wereheavily dependent on domesticated species (Collard et al.2013a; Kline and Boyd 2010; Neiman 1995), whereas the ma-jority of samples that have refuted the hypothesis consist ofpopulations that relied primarily on wild resources (Collard,Kemery, and Banks 2005; Collard et al. 2013b; Read 2008).Consequently, it could be that mode of production mediatesthe effect of POP on cultural evolution such that the tech-nology of food producers is more affected by POP than byrisk, whereas the technology of hunter-gatherers is more af-fected by risk than by POP. The problem with this proposalis that Nelson et al.’s (2011) data relate to small-scale farminggroups. This makes the idea that mode of production mediatesthe effect of POP on cultural evolution less plausible giventhat it means that groups with a food-producing mode ofproduction both support and refute the hypothesis.

A second possibility is that there is a threshold effect inthe influence of POP on technology. Populations that support

Collard, Buchanan, and O’Brien Population Size and the Paleolithic Archaeological Record S395

Table 6. Summary of results of standard multiple regression analyses using the Pacific Northwest sample ( ) carriedn p 14out to assess the relative importance of various variables as drivers of tool kit richness and complexity

Dependentvariable

Fullmodel POP ET NAGP RHIGH NOMOV DISMOV

STS F p 1.753,df p 6, 7,P p .240,r2 p .600

β p .344,P p .250,VIF p 1.323

β p �.181,P p .535,VIF p 1.354

β p .348,P p .553,VIF p 5.479

β p .085,P p .883,VIF p 5.444

β p �1.028,P p .160,VIF p 7.488

β p 1.620,P p .059,VIF p 9.034

TTS F p 1.289,df p 6, 7,P .370,r2 p .525

β p .447,P p .179,VIF p 1.323

β p .196,P p .539,VIF p 1.354

β p .839,P p .211,VIF p 5.479

β p �.345,P p .588,VIF p 5.444

β p �.803,P p .297,VIF p 7.488

β p 1.063,P p .217,VIF p 9.034

AVE F p 1.862,df p 6, 7,P p .217,r2 p .615

β p .258,P p .372,VIF p 1.323

β p .608,P p .061,VIF p 1.354

β p .780,P p .198,VIF p 5.479

β p �.610,P p .302,VIF p 5.444

β p .210,P p .753,VIF p 7.488

β p �.681,P p .366,VIF p 9.034

Note. STS p total number of subsistants; POP p population size; TTS p total number of techno-units; AVE p average number of techno-unitsper tool; ET p effective temperature; NAGP p net aboveground productivity; RHIGH p rainfall for the wettest month of the year; NOMOV pnumber of residential moves per year; DISMOV p total distance moved per year during residential moves; VIF p variance inflation factor.

the hypothesis and those that do not overlap in terms of size,but several of the former are much larger than the largest ofthe latter. Thus, it could be that POP does not have a sig-nificant effect on cultural evolution until it is greater than avalue close to or above the upper end of the populations thatsupport the POP hypothesis, which is ∼12,000 people. How-ever, the modeling work of Shennan (2001) and Henrich(2004) suggests that the effect of POP on cultural evolutionshould be greatest when POP is less than a few thousand, soa threshold effect where the POP holds for larger populationsbut not for smaller ones also seems unlikely to be the expla-nation for the disagreement.

The other two potential explanations involve social factors.Henrich (2010) has argued that norms and institutions thatfoster sharing can positively affect the spread of inventionswithin a population. Thus, it could be that sharing normsand institutions can mediate the effects of POP on culturalevolution such that a small population with numerous and/or strong sharing norms and institutions is equivalent or evenbetter in terms of its ability to retain beneficial inventionsthan a large population with few and/or weak sharing normsand institutions. If this is the case, then it is possible that thedisagreement among the studies is the result of populationsthat support the hypothesis having fewer and/or weaker shar-ing norms and institutions than populations that do not sup-port it. Another possibility is that the disagreement is a con-sequence of differences in degree of task specialization.Recently, Bentley and O’Brien (2011) argued that the effectof POP documented by Henrich (2004) and Powell, Shennan,and Thomas (2009) depends on two strong assumptions: (1)the skill level of the most skilled member of the group isseveral times greater than the skill level of the average groupmember, and (2) all learners can identify and copy the mostskilled member of the group. Bentley and O’Brien demon-strate that the POP effect is reduced if skill level is normallydistributed within a group and/or if people are less selective

about whom they copy. Indeed, they show that in certaincircumstances (e.g., if individuals copy the most popular be-havior or copy from each other at random), cultural com-plexity can increase or decrease regardless of POP. One cor-ollary of their findings is that the POP effect is likely to bemediated by degree of task specialization. Given that skill levelis primarily a result of practice time (Ericsson and Charness1994), task specialization can be expected to increase the dif-ference in skill level between the most skilled individual withina group and the majority of group members. This means thatthe POP effect should be more pronounced in populationswith more task specialization than in populations with lesstask specialization. Thus, it is possible that the disagreementamong the studies that have tested the POP hypothesis is aconsequence of populations that support the hypothesis hav-ing more task specialization than populations that do notsupport it. At the moment, we are not in a position to de-termine which, if either, of these hypotheses is correct. Doingso will require further modeling work and cross-cultural stud-ies.

Together, the study reported here and the other studies thathave failed to support the POP hypothesis have implicationsfor interpreting the Paleolithic archaeological record. As wenoted earlier, it has become commonplace to use POP toexplain patterns in the record, but given that the record wasproduced exclusively by hunter-gatherers, the failure of thestudy reported here and the studies of Collard, Kemery, andBanks (2005), Read (2008), and Collard et al. (2013b) tosupport the hypothesis challenges these interpretations. If therichness and complexity of the technology of recent hunter-gatherers are not affected by POP, there would seem to belittle reason to expect changes in POP to be a broadly usefulhypothesis for explaining patterns in the Paleolithic archae-ological record given that it was produced exclusively byhunter-gatherers. The same holds for stability in POP. POPchange/stability may explain some of the patterns the Pale-

S396 Current Anthropology Volume 54, Supplement 8, December 2013

olithic archaeological record, but this needs to be demon-strated on a case-by-case basis through tests in which the POPhypothesis is pitted against competing hypotheses. As men-tioned earlier, several studies have suggested that environ-mental risk is the primary driver of technological richnessand complexity among recent hunter-gatherers (Collard,Kemery, and Banks 2005; Collard et al. 2013b; Read 2008;Torrence 1983, 1989). Thus, adaptation to environmentalconditions is one hypothesis that should be included in suchtests. Another factor that should probably be taken into ac-count is social conformity, given that it has the capacity toaffect collective action (Nelson et al. 2011). Regardless ofwhich competing hypotheses are considered, simply attrib-uting patterns in the Paleolithic archaeological record to POPis not a defensible course of action.

Acknowledgments

We thank Steven Kuhn and Erella Hovers for the invitationto participate in the “Alternative Pathways to Complexity:Evolutionary Trajectories in the Middle Paleolithic and Mid-dle Stone Age” symposium and Leslie Aiello and Laurie Ob-bink of the Wenner-Gren Foundation and the other partic-ipants in the symposium for making it such an enjoyable andintellectually stimulating experience. We also thank MichaelKemery and Jesse Morin for assistance with data collection.Mark Collard is supported by the Canada Research ChairsProgram, the Social Sciences and Humanities Research Coun-cil, the Canada Foundation for Innovation, the British Co-lumbia Knowledge Development Fund, and Simon FraserUniversity. Briggs Buchanan is supported by the Universityof Missouri and Simon Fraser University. The opinions ex-pressed in this paper do not necessarily reflect the views ofthe aforementioned funding bodies.

References CitedBailey, Harry P. 1960. A method of determining the warmth and temperateness

of climate. Geografiska Annaler 42:1–16.Benjamini, Yoav, and Daniel Yekutieli. 2001. The control of false discovery

rate under dependency. Annals of Statistics 29:1165–1188.Bentley, R. Alexander, and Michael J. O’Brien. 2011. The selectivity of social

learning and the tempo of cultural evolution. Journal of Evolutionary Psy-chology 9:125–141.

Binford, Lewis R. 2001. Constructing frames of reference: an analytical methodfor theory building using ethnographic and environmental data sets. Berkeley:University of California Press.

Boyd, Robert, Peter J. Richerson, and Joseph Henrich. 2013. The culturalevolution of technology: facts and theories. In Cultural evolution: society,technology, language, and religion. Peter J. Richerson and Morten H. Chris-tiansen, eds. Pp. 119–142. Cambridge, MA: MIT Press.

Collard, Mark, Briggs Buchanan, Jesse Morin, and Andre Costopolous. 2011.What drives the evolution of hunter-gatherer subsistence technology? areanalysis of the risk hypothesis with data from early contact era populations

in the Pacific Northwest. Philosophical Transactions of the Royal Society B366:1129–1138.

Collard, Mark, Briggs Buchanan, Michael J. O’Brien, and Jonathan Scholnick.2013a. Risk, mobility, or population size? drivers of technological diversityamong recent western North American hunter-gatherers. PhilosophicalTransactions of the Royal Society B 368:20120412.

Collard, Mark, Michael Kemery, and Samantha Banks. 2005. Causes of toolkitvariation among hunter-gatherers: a test of four competing hypotheses.Canadian Journal of Archaeology 29:1–19.

Collard, Mark, April Ruttle, Briggs Buchanan, and Michael J. O’Brien. 2013b.Population size and cultural evolution in non-industrial food-producingsocieties. PLoS ONE 8:e72628.

Colwell, Robert K. 2009. Biodiversity: concepts, patterns and measurement.In The Princeton guide to ecology. Simon A. Levin, ed. Pp. 257–263. Prince-ton, NJ: Princeton University Press.

Ericsson, K. Anders, and Neil Charness. 1994. Expert performance: its struc-ture and acquisition. American Psychologist 49:725–747.

Henrich, Joseph. 2004. Demography and cultural evolution: why adaptivecultural processes produced maladaptive losses in Tasmania. American An-tiquity 69:197–218.

———. 2006. Understanding cultural evolutionary models: a reply to Read’scritique. American Antiquity 71:771–782.

———. 2010. The evolution of innovation-enhancing institutions. In Inno-vation in cultural systems: contributions from evolutionary anthropology. Mi-chael J. O’Brien and Stephen J. Shennan, eds. Pp. 203–223. Cambridge,MA: MIT Press.

Kline, Michelle A., and Robert Boyd. 2010. Population size predicts tech-nological complexity in Oceania. Proceedings of the Royal Society B 277:2559–2564.

Kobayashi, Yutaka, and Kenichi Aoki. 2012. Innovativeness, population sizeand cumulative cultural evolution. Theoretical Population Biology 82:38–47.

Mesoudi, Alex. 2011. Variable cultural acquisition costs constrain cumulativecultural evolution. PLoS ONE 6:e18239.

Narum, Shawn R. 2006. Beyond Bonferroni: less conservative analyses forconservation genetics. Conservation Genetics 7:783–787.

Neiman, Fraser D. 1995. Stylistic variation in evolutionary perspective: in-ferences from decorative diversity and interassemblage distance in IllinoisWoodland ceramic assemblages. American Antiquity 60:7–36.

Nelson, Margaret C., Michelle Hegmon, Stephanie R. Kulow, Matthew A.Peeples, Keith W. Kintigh, and Ann P. Kinzig. 2011. Resisting diversity: along-term archaeological study. Ecology and Society 16(1):25.

Oswalt, Wendell H. 1973. Habitat and technology: the evolution of hunting.New York: Holt, Rinehart, & Winston.

———. 1976. An anthropological analysis of food-getting technology. New York:Wiley.

Powell, Adam, Stephen J. Shennan, and Mark G. Thomas. 2009. Late Pleis-tocene demography and the appearance of modern human behavior. Science324:1298–1301.

Premo, Luke S., and Steven L. Kuhn. 2010. Modeling effects of local extinctionson culture change and diversity in the Paleolithic. PLoS ONE 5:e15582.

Read, Dwight. 2008. An interaction model for resource implement complexitybased on risk and number of annual moves. American Antiquity 73:599–625.

Riede, Felix. 2008. The Laacher See-eruption (12,920 BP) and material culturechange at the end of the Allerød in Northern Europe. Journal of Archae-ological Science 35:591–599.

Shennan, Stephen J. 2001. Demography and cultural innovation: a model andsome implications for the emergence of modern human culture. CambridgeArchaeological Journal 11:5–16.

Shott, Michael. 1986. Technological organization and settlement mobility: anethnographic examination. Journal of Anthropological Research 42:15–51.

Torrence, Robin. 1983. Time budgeting and hunter-gatherer technology. InHunter-gatherer economy in prehistory: a European perspective. Geoff Bailey,ed. Pp. 11–22. Cambridge: Cambridge University Press.

———. 1989. Re-tooling: towards a behavioral theory of stone tools. In Time,energy and stone tools. Robin Torrence, ed. Pp. 57–66. Cambridge: Cam-bridge University Press.

Current Anthropology Volume 54, Supplement 8, December 2013 S397

� 2013 by The Wenner-Gren Foundation for Anthropological Research. All rights reserved. 0011-3204/2013/54S8-0017$10.00. DOI: 10.1086/673264

Measuring the Complexity ofLithic Technology

by Charles Perreault, P. Jeffrey Brantingham,Steven L. Kuhn, Sarah Wurz, and Xing Gao

CA� Online-Only Material: Supplementary Table

Assessments of the complexity of lithic technologies coming from different time periods, regions, or hominid speciesare recurrent features of the literature on Paleolithic archaeology. Yet the notion of lithic complexity is often definedintuitively and qualitatively, which can easily lead to circular arguments and makes difficult the comparison ofassemblages across different regions and time periods. Here we propose, in the spirit of Oswalt’s techno-units, thatthe complexity of lithic technology can be quantified by counting the procedural units involved in tool manufacture.We define procedural units as mutually exclusive manufacturing steps that make a distinct contribution to thefinished form of a technology. As a proof of concept, we use the procedural-unit approach to measure the complexityof 13 Paleolithic assemblages. While preliminary, these results provide a quantitative benchmark confirming thatlithic technological complexity increased throughout the Paleolithic period. The method to measure lithic complexityoutlined here will allow us to revisit several claims made about change in technological complexity during humanevolution.

Introduction

Arguments about the complexity of lithic technologies comingfrom different time periods, regions, or hominid species arerecurrent features of the literature on Paleolithic archaeology.Technological complexity—or technological “simplicity,”“crudeness,” “refinement,” “sophistication,” or “advance-ment”—are thought to have had important effects on foragerpopulations, including broadening of ecological niches (Shea

Charles Perreault is Assistant Professor in the Department ofAnthropology at the University of Missouri, Columbia (107 SwallowHall, Columbia, Missouri 65211, U.S.A. [perreaulc@missouri.edu]).P. Jeffrey Brantingham is Professor in the Department ofAnthropology at the University of California, Los Angeles (341Haines Hall, Box 951553, Los Angeles, California 90095, U.S.A.[branting@ucla.edu]). Steven L. Kuhn is Professor in the School ofAnthropology at the University of Arizona (Emil W. Haury Building,P.O. Box 210030, Tucson, Arizona 85721, U.S.A. [skuhn@email.arizona.edu]). Sarah Wurz is Senior Researcher at the Institutefor Human Evolution of the University of the Witwatersrand (PrivateBag 3, PO WITS, 2050, Johannesburg, South Africa [sarah.wurz@wits.ac.za]). Xing Gao is Vice Director of the Institute of VertebratePaleontology and Paleoanthropology at the Chinese Academy ofSciences (P.O. Box 643, Beijing 100044, China [gaoxing@ivpp.ac.cn]).This paper was submitted 3 VII 13, accepted 18 VII 13, andelectronically published 24 X 13.

and Sisk 2010) or increasing the productivity and therebyaiding the dispersal of our species out of Africa (Mellars2006b). Several processes can also drive changes in techno-logical complexity. Studying changes in technological com-plexity thus allows us to make inferences about how theseprocesses unfolded during human evolution. For instance,putative increases in technological complexity have been in-terpreted as signaling changes in cognitive abilities (e.g., Am-brose 2001, 2010; Coolidge and Wynn 2009; de Beaune 2004;Foley 1987; Foley and Lahr 2003; Haidle 2010; Mellars 1989,2006b; Wadley 2010) as well as the extent to which our an-cestors relied on social learning (Foley and Lahr 2003; Richer-son and Boyd 2005). In line with theories of artifact designand technological organization—which predict that techno-logical complexity will vary with factors such as prey choice,time budgeting, risk, labor costs, and mobility pattern (e.g.,Bleed 1986; Bousman 1993; Kelly 1995; Osborne 1999; Oswalt1976; Torrence 1983, 1989)—changes in lithic technologycomplexity have also been seen as a response to change inclimate and environment (Mellars 1989, 2006a, 2006b; Sheaand Sisk 2010), in variation in energetic constraints and timebudgeting (Shea and Sisk 2010), in hunting strategies (Mellars1989), or in lithic raw material quality (Brantingham et al.2000; Mellars 2006a, 2006b; Pope 1989; Schick 1994). Culturaltraditions or ethnic groups have also been interpreted as thecause of spatial variation in lithic complexity (Movius 1944,1948; Schick 1994). Finally, there is even greater need for

S398 Current Anthropology Volume 54, Supplement 8, December 2013

rigorous measures of technological complexity in order to testincreasingly influential models linking demography and cul-tural evolution (Beatty 1995; Foley and Lahr 2003; Mellars2006a; Powell, Shennan, and Thomas 2009; Premo and Kuhn2010; Reiter 2000; Shennan 2001).

Consideration of technological complexity in the Paleo-lithic record is weakened, however, by a lack of explicit def-inition of complexity. Often one technology is simply de-scribed as more complex than another without explanation.Such intuitive notions of technological complexity are dan-gerous because they lead easily to circular arguments. Forinstance, methods such as prismatic blade technology havebeen assumed to be more complex that the Levallois methodor bifacial shaping by the simple virtues of appearing later intime or being associated with modern humans (see Bar-Yosefand Kuhn 1999).

The complexity of Paleolithic technologies is also rarelyquantified. Complexity is sometimes defined by the presenceor absence of qualitative features, such as composite tools(Ambrose 2001, 2010; Coolidge and Wynn 2009), the use ofcompound adhesives (Wadley 2010), the use of a compositetool to make another composite tool (Lombard and Haidle2012), or technologies that store energy exosomatically, suchas the bow and arrow (Shea and Sisk 2010). But the usefulnessof these qualitative definitions of technological complexity islimited because they are based on time-specific, region-spe-cific, or culture-specific traits. They do not allow for the com-parison of two assemblages if both either lack or possess thesetraits. In other words, if the transition from simple to complextechnologies is marked by the advent of composite tools, thenhow does the complexity of two assemblages that lack com-posite tools compare? Finally, these qualitative definitions alsofail to measure how much more complex these traits are. Forinstance, how much of a leap in technological complexitydoes the use of a composite tool really imply?

Given its importance in Paleolithic research, we have muchto gain from developing a more objective way of measuringtechnological complexity that is quantitative and not boundto a specific time period, region, culture, or indeed any specifictechnology. We need a way to measure technological com-plexity that will allow us to detect an increase in complexitymarked by, for example, the advent of composite tools butwithout treating composite tools as a qualitative break. Finally,our system of measurement should allow us to compare notjust the complexity of different lithic assemblages but also oflithic and nonlithic technologies. We should be able to com-pare the complexity of Oldowan tools not only with that ofEuropean Upper Paleolithic blades but also with that of aBoeing 747.

Here we present a quantitative and widely applicable ap-proach to measuring technological complexity in the archae-ological record. We propose that technological complexity canbe measured by counting the procedural units, or manufac-turing “building blocks,” represented in an assemblage. Belowwe describe this method, followed by its application to a series

of Paleolithic assemblages as a proof of concept. Our goalhere is not to test any specific hypothesis about the mecha-nisms driving changes in technological complexity but ratherto develop a tool that can be used to test such hypotheses ina more rigorous manner than has been done so far.

Measuring Lithic Technology Complexity

We define technological complexity as the minimum amountof information that is needed to manufacture a product. Thisdefinition is in line with other formalized definitions of com-plexity (Shannon and Weaver 1949). Computer scientists, forexample, have defined the complexity of an algorithm as theshortest string length, or the smallest number of bits of in-formation, that is necessary to describe it (Chaitin 1970). Thisinformation criterion is analogous to the various measures ofrichness used to describe biological systems that are definedas the number of unique types of some constituent presentwithin an aggregate group. For instance, at the level of theorganism, biological complexity has been measured as thecount of cell types (Bonner 1988). At the level of an ecosystem,biological complexity has been measured as the count ofunique species it contains (Bonner 1988). Finally, the com-plexity of animal behavior has also been estimated by countingthe number of elemental “building blocks” that is associatedwith a specific behavior or, at a larger scale, as the numberof acts in a species’ behavioral repertoire (Sambrook andWhiten 1997; Whiten et al. 1999).

In the same spirit, we argue that the complexity of a tech-nology can be measured by counting the number of elementalbuilding blocks associated with it. We call these buildingblocks “procedural units.” We define procedural units as mu-tually exclusive manufacturing steps that make a distinct con-tribution to the finished form of the product of a technology.Focusing on lithic technology, the count of procedural unitspresent in a tool reduction sequence is a measure of com-plexity because it reflects the minimum amount of infor-mation that is needed to carry it out to a successful end.

This procedural-unit approach to stone-tool complexityparallels Oswalt’s “techno-units” (Oswalt 1976). Oswalt as-sessed the complexity of food-getting technologies by count-ing (1) the number of tool types present in a tool kit, whichhe called “subsistants,” and (2) the number of integrated andphysically distinct structures that contribute to the finishedform of a tool, which he called “techno-units.” Oswalt’smethod is powerful because it allows for the measurement oftechnological complexity cross-culturally. It has been appliedto ethnographic data to test a wide range of hypotheses, in-cluding hypotheses about the ecological determinants of tech-nological complexity (Collard, Kemery, and Banks 2005; Col-lard et al. 2011; Shott 1986; Torrence 1983, 1989, 2000) andthe effect of demography on the evolution of technologies(Collard, Kemery, and Banks 2005; Collard et al. 2011; Klineand Boyd 2010; Oswalt 1976).

Oswalt (1976:229–230) recognized that there are problems

Perreault et al. Measuring the Complexity of Lithic Technology S399

with applying his concept of complexity to the archaeologicalrecord. First, the number of tool types present in a prehistorictool kit is difficult to assess because the actual function oftools is difficult to infer. Moreover, when tool function canbe determined, there is often overlap in the functional tasksaccomplished with what appear to be morphologically distincttool types. Second, the techno-unit approach will lead to un-derestimating the complexity of prehistoric technologies be-cause of preservation biases. This is especially true for thePaleolithic record, were the range of preserved material isnarrow. For instance, in most situations a complex technologysuch as the bow and arrow would leave in the archaeologicalrecord only one techno-unit, the stone projectile point. Yetwith full preservation, it is clear that bow-and-arrow tech-nologies are more complex technology than simple handheldscrapers, a 1-techno-unit technology. Attempting to inferwhat techno-units are missing in an archaeological assemblagefor taphonomic reasons is not a solution because any giventechnology can contain a varying number of techno-units.For instance, Oswalt’s data set (1976) contains examples ofbows that range in complexity from 2 to 10 techno-units andof arrows that range from 2 to 13 techno-units. Finally, be-cause Oswalt’s approach focuses on the finished product, itfails to capture variation in the complexity involved in themaking of different tools. For example, producing a prismaticblade can be more complex than making a digging stick, andyet both could be seen as simple, 1-techno-unit technologieswhen analyzed using Oswalt’s method.

We can avoid to a certain extent these problems by focusingon the chaınes operatoires of technologies rather than on thefinished products. Counting the procedural units present ina reduction sequence or in an assemblage allows us to avoidhaving to identify functionally distinct tool types as well ashaving to make inferences about tool parts that are missingfrom the assemblage. The idea of comparing technologiesbased on the number of manufacture steps they contain isnot new (see, e.g., Ambrose 2001; Gowlett 1996) and is similarto the “cognigrams” method developed by Haidle (Haidle2009, 2010; Lombard and Haidle 2012). Cognigrams can pro-vide rich insights into the nature of technologies and humanbehaviors, but their value in comparative research is limitedby the fact that they are not easily quantifiable.

Lithic Procedural Units

We have assembled a list of potential procedural units showinghow the complexity of lithic technology can be measured. Inaccordance with our definition of procedural units, we haveorganized our list of lithic procedural units by reduction steps:preliminary treatment of raw material, core preparation tech-niques, blank production techniques, product shaping, andcore rejuvenation. Such division is necessary because the sameprocedural unit, such as the use of a hard hammer, can servedistinct functions depending on the reduction step for whichit is called into action. Each distinct usage should be counted

as a different procedural unit. Conversely, the repeated useof a hard hammer within the same reduction step, such asin the shaping of a core, should not be counted as multipleindependent procedural units; each hammer blow serves thesame function, and the number of blows struck during thepreparation of the core may vary from one core to anotherwithout it affecting the nature of the technology. This is anal-ogous to Oswalt’s (1976:52) recommendation that physicallydistinct elements serving the same purpose, such as the ballsof a bola or the teeth of a rake, be counted as only one techno-unit. Researchers thus need to decide on a case-by-case basiswhether the same procedural unit at different reduction stepsconstitutes functionally distinct manufacturing steps or not.Taking all these things into account and focusing on uncon-troversial features that are commonly discussed in the lithicanalysis literature, we have identified 35 procedural units thatmay be associated with lithic technology.

Preliminary Steps

(1) Raw material treatment: evidence of heat treatment

Core Preparation Techniques

(2) Decortification: cortex is removed(3) Shaping of platform: platform intentionally prepared

by flaking(4) Shaping of flaking surface: face of flake removal is in-

tentionally prepared by flaking(5) Shaping of nonflaking surface: nonactive part of the

core is shaped(6) Blades: crested or debordante blades used to align face

of flake removalCore shaping techniques: (7) hard hammer percussion

used, (8) soft hammer percussion used, (9) bipolar hammerpercussion used, (10) indirect hammer percussion used, (11)pressure hammer percussion used, (12) pecking hammer per-cussion used

Blank Production Techniques

(13) Hard hammer percussion used(14) Soft hammer percussion used(15) Bipolar hammer percussion used(16) Indirect hammer percussion used(17) Pressure hammer percussion used(18) Ochre applied (has traces of ochre)Platform treatment: (19) abrasion (platform is rubbed/

abraded), (20) overhang removal (small flakes are removedfrom core face adjacent to platform), (21) faceting (smallflakes are removed from the platform)

Product Shaping

Edge shaping: (22) retouched edge (edge of final productis retouched)

S400 Current Anthropology Volume 54, Supplement 8, December 2013

Prehensile modification: (23) backing (a sharp era of finalproduct is backed for manual prehension or hafting), (24)notching (final product is notched), (25) tanging (final prod-uct is tanged)

Surface shaping: (26) unifacial retouch (overall morphol-ogy of the blank altered by unifacial flaking), (27) bifacialretouch (overall morphology of the blank altered by bifacialflaking), (28) grinding (ventral/dorsal surface of the finalproduct is ground)

(29) Bulbar thinning: flat flakes removed from bulbar faceof the blank

Core Rejuvenation

(30) Tablette: platforms reshaped by removal of single largeflakes

(31) Outrepassee: overpassed/plunging flakes struck inten-tionally to shape distal end of core

(32) Debordante: flakes or blades struck along edge of coreto reshape the flaking surface

(33) Secondary crest: secondary crest used to reshape coreface

(34) Rotation of core: subsequent flakes removed fromnonopposed platforms

(35) Face shaping: core reshaped by lateral flaking betweenremovals or series of removals

Using such a list, the complexity of chaınes operatoires canbe translated into a number, the sum of procedural unitspresent in the reduction sequence, that summarizes their com-plexity and that can be compared across cases.

The list above is by no means exhaustive. Lithic technologycan potentially include more procedural units. For instance,in some archaeological contexts it might be useful to alsocount the use of binding material for composite tools as wellas the type of binding material used (gum, ochre, fat, wax,resin). The list can also be extended to encompass within thesame analysis other material and technologies, such as bonetools, ochre pigments, and shell beads.

In this regard, it is important to recognize that our unitof analysis is the procedural unit independent of its content.In that sense, it does not matter whether a core is shaped bya combination of hard hammer and indirect percussion orby a combination of hard and soft hammer: what matters isthat in both cases, the shaping of the core involves two pro-cedural units. This is analogous to what ecologists do whenthey contrast ecosystems by comparing their species richness,that is, the total count of species present in each ecosystem.This allows for ecosystems that have few or no species incommon to be compared. Similarly, and similar to Oswalt’stechno-units, the count of procedural units provides us witha common measurement unit that allows for the comparisonof different technologies. This is also why the list above cannotbe used as a universal checklist: whether or not an item onthe list constitutes a procedural unit really depends onwhether it constitutes a mutually exclusive manufacturing step

that makes a distinct contribution to the finished form of theproduct of a particular technology. It is this focus on thedefinition of a procedural unit rather than on its content thatallows for comparison among a wide range of lithic and non-lithic technologies across cultures, time, space, and species.

Many aspects of the way procedural units are counted de-pend on the research question asked, such as whether idio-syncratic units, or units that contribute to the decorative as-pects of a technology, should be included or excluded fromthe analysis. There is no single set of units that applies uni-versally, so it is important that these decisions be reported inpublications in order to increase the replicability of analysesand to facilitate the comparison of published results. Finally,the analysis of technological complexity can be conducted atdifferent scales because the procedural-unit approach can beused to measure the complexity of individual chaınes opera-toires as well as that of assemblages as a whole. This flexibilityis useful because there are many archaeological contexts inwhich chaınes operatoires cannot be easily reconstructed.

Proof of Concept

We have conducted a series of experiments to evaluate thevalidity of the procedural-unit approach to lithic technologyby measuring the complexity of 13 lithic assemblages. Morespecifically, we are interested in detecting a temporal trendin lithic complexity through the Paleolithic period. We havepurposely tried to sample assemblages coming from a widerange of time periods and spatial locations. Our sample thusincludes sites dating from the Lower Paleolithic (Early StoneAge) to the Upper Paleolithic (Late Stone Age) and rangingfrom South Africa to Turkey. Given the temporal and spatialrange it covers, our sample is too small to draw any definitiveconclusion about patterns of complexity during the Paleo-lithic period. However, it is a useful proof of concept for theprocedural-unit approach because it shows that (1) themethod can capture variation in lithic complexity and (2) themethod is sufficiently robust in the face of variation in howprocedural units are defined.

Different analysts may count procedural units in a lithicassemblage differently. To examine this problem, we simulatedthe noise that could be generated by different analysts withdifferent views on what constitutes a procedural unit by lump-ing and splitting the list presented above. In other words, weask to what extent does the temporal pattern of lithic com-plexity that we can observe in the Paleolithic record dependon whether the assemblages have been analyzed by a “splitter”or by a “lumper.” More saliently, what if a mixture of “lump-ers” and “splitters” analyzed the assemblages compared, aswould be the case in a data set assembled from variouspublications? To answer these questions, we counted the pro-cedural units of the 13 Paleolithic assemblages using two dif-ferent lists. First, the assemblages were analyzed using theextended list presented above. In our view, this is the list thatbest captures what we mean by procedural units in the context

Perreault et al. Measuring the Complexity of Lithic Technology S401

of lithic technology. Second, we produced a second estimateof the complexity of the assemblages using a shorter versionof the above list. This second estimate represents how a con-servative lithic analyst who prefers to err on the side of cautionmight count procedural units. It excludes nine variables thatmay be deemed ambiguous and too difficult to identify, suchas heat treatment of raw material or the rotation of the coreduring core rejuvenation. It also lumps 14 procedural unitsinto four units. For instance, it lumps together hard hammer,soft hammer, indirect, and pressure flaking into a single cat-egory: “unipolar percussion.” This “conservative” list containsa total of 16 procedural units as opposed to 38 proceduralunits for the “nonconservative” list.

Table 1 summarizes our sample of assemblages as well astheir complexity relative to the two lists (see supplementarymaterial, available online, for details). The counts of proce-dural units are reported for assemblages as a whole ratherthan for individual chaınes operatoires. These counts thus rep-resent the complexity of the tool kits used by the populationsthat produced these assemblages rather than the complexityof specific tools they used. Table 1 shows that the procedural-unit approach does capture variation in the complexity oflithic technologies. Using the conservative list, the counts ofprocedural units range from four, with the Lower Paleolithicassemblage of Tabun Cave in Israel, to 11, with the MiddleStone Age assemblage of Klasies River Mouth in South Africa.With the nonconservative list, the counts of procedural unitsrange from six, with the two Oldowan assemblages, to 23,with the Mousterian assemblage of Amud Cave.

Figure 1 shows the count of procedural units of the assem-blages plotted against the midpoint of their age on a log-logscale (the two Oldowan assemblages, A.L. 894 and A.L. 666,are plotted against their terminus ante quem date, 12.35 mya).A linear regression analysis in which the dependent variableis the logarithm of the count of procedural units and theindependent variable is the logarithm of the midpoint of theage of the assemblage suggests that the complexity of lithictechnologies increases steadily through time. The slope β ofthe best-fit linear model for the nonconservative count ofprocedural units is �0.314 ( ), which means that theP ! .001size of the procedural-unit inventory shrinks by about 3% asthe age of the material increases by 10%. The model explains68% of the variance (adjusted ). The relationship2R p 0.68between age and lithic complexity is also detectable when theconservative view of procedural units is adopted, although itis not as strong: (a decrease in complexity ofβ p �0.184about 2% per 10% increase in age; , adjusted 2P ! .005 R p

). These results suggest that the temporal pattern of in-0.49crease in complexity of lithic technology through time isstrong enough to be detected by both conservative lumpersand nonconservative splitters. But what if both conservativeand nonconservative estimates were combined in the sameanalysis? To test this possibility, we created 1,000 data sets byselecting randomly, for each assemblage, a count of proceduralunits from the conservative or the nonconservative view. Each

data set thus contains a different mixture of conservative andnonconservative analytical decisions simulating the effect ofoperator variation on estimates of lithic complexity. Runninglinear regression analysis on these data sets, we find that theresults are affected by this sampling procedure but not soprofoundly as to disguise the basic patterning. The effect sizeof age on the count of procedural units across these 1,000regression analyses ranges from �0.11 to �0.38 with an av-erage of �0.25, and the adjusted R2’s of these different linearmodels range from 0.13 to 0.73 with an average of 0.39 (fig.2).

Although there is a general increase in lithic technologicalcomplexity over the Pleistocene, the rate of increase in com-plexity is greater for the subset of Middle Paleolithic andMiddle Stone Age assemblages ( ; see fig. 3). The slopen p 8of the best-fit linear model is steeper ( and �0.7,β p �0.46

and 0.003 for the conservative and nonconservativeP p .004counts, respectively), and the variance explained by the modelis greater (adjusted and 0.88, respectively). This2R p 0.74suggests that the increase in lithic complexity within the Mid-dle Stone Age/Middle Paleolithic might have been more reg-ular than it is across the whole Paleolithic period (thus betterdescribed by a linear model on a log-log scale). This resultcontrasts with other studies that have found no general tech-nological trends within the Middle Paleolithic period (see dela Torre 2013; Kuhn 2013). Assuming that this long-termtrend for increasing complexity through time holds as moreassemblages are added to our sample, this analysis is in linewith the view that behavioral complexity increased graduallyand cumulatively through the Middle Paleolithic/MiddleStone Age and well before 50 ka (e.g., Brown et al. 2009;d’Errico and Henshilwood 2007; d’Errico et al. 2005; Mareanet al. 2007; McBrearty and Brooks 2000).

Overall, our analysis suggests that trends such as the in-crease of lithic complexity through the Paleolithic period maybe detected in a robust manner even in the face of divergencebetween analysts in the definition of procedural units. Eventhough our results are preliminary, we find it intriguing thatthe complexity of these lithic assemblages align along the sametrend line even though they were produced over thousandsof years by different hominid species coming from variousparts of Africa and Eurasia and given the presumed sensitivityof lithic technology to environmental, demographic, behav-ioral, economical, and cognitive factors.

Discussion and Conclusion

Our main argument in this paper is that the complexity oflithic technology can be measured by counting the proceduralunits, that is, the mutually exclusive manufacturing steps thatcontribute to the finished form of the technology. Because itis quantitative and independent of any specific technology,time period, region, or culture, this approach can also beapplied to nonlithic technologies and could help solve manyof the issues associated with previous attempts at measuring

Tabl

e1.

Cou

nt

ofpr

oced

ura

lu

nit

sof

13P

aleo

lithi

cas

sem

blag

es

Ass

embl

age

Per

iod

Age

Cou

nt

ofpr

oced

ura

lu

nit

s

Ref

eren

ces

Con

serv

ativ

elis

tN

onco

nse

rvat

ive

list

A.L

.89

4,E

thio

pia

Low

erP

aleo

lith

ic(O

ldow

an)

12.

35m

ya5

6H

over

s20

09a;

E.

Hov

ers

and

A.

Dav

id-

szon

,u

npu

blis

hed

man

usc

ript

A.L

.66

6,E

thio

pia

Low

erP

aleo

lith

ic(O

ldow

an)

12.

35m

ya5

6E

.H

over

san

dA

.D

avid

szon

,u

npu

blis

hed

man

usc

ript

;K

imbe

let

al.

1996

Gn

Jh-0

3,K

enya

Ear

lySt

one

Age

(Ach

eule

an)

ca.

500–

284

ka8

14C

orn

elis

sen

1992

;Le

akey

etal

.19

69;

McB

rear

ty19

99;

Tryo

n,

McB

rear

ty,

and

Texi

er20

05Ta

bun

Cav

e,U

nit

XI-

XIV

,Is

rael

Ear

lyP

aleo

lith

icca

.45

0–22

5ka

47

Yan

mbu

rgaz

Cav

e,Tu

rkey

Ear

lyM

iddl

eP

aleo

lith

ic27

5ka

67

Ku

hn

,A

rseb

uk,

and

How

ell

1996

Tabu

nC

ave,

Un

itIX

,Is

rael

Ear

lyM

iddl

eP

aleo

lithi

c(M

oust

eria

n)

ca.

225–

100

ka6

13R

.Sh

imel

mit

zan

dS.

L.K

uhn

,u

npu

blis

hed

man

usc

ript

Gh

Jh-7

4,K

enya

Mid

dle

Ston

eA

geca

.20

0ka

511

Tryo

n20

03,

2006

Kla

sies

Riv

erM

outh

,K

lasi

esst

age,

Sou

thA

fric

aM

iddl

eSt

one

Age

115–

100

ka9

19W

urz

2002

a,20

02b

Qaf

zeh

Cav

e(l

ayer

sX

XIV

-III

),Is

rael

Mid

dle

Pal

eolit

hic

(Mou

ster

ian

)92

ka10

19H

over

s20

09b;

Hov

ers

and

Rav

eh20

00K

lasi

esR

iver

Mou

th,

Mos

sel

Bay

stag

e,So

uth

Afr

ica

Mid

dle

Ston

eA

ge10

0–80

ka9

15W

urz

2002

a,20

02b

Am

ud

Cav

eB

4-B

1,Is

rael

Mid

dle

Pal

eolit

hic

(Mou

ster

ian

)68

–55

ka11

23A

lper

son

-Afi

lan

dH

over

s20

05;

Eks

htai

n20

06;

God

er19

97;

Hov

ers

1998

,20

07;

Hov

ers

etal

.20

11K

lasi

esR

iver

Mou

th,

How

ieso

nP

oort

stag

e,So

uth

Afr

ica

Mid

dle

Ston

eA

ge62

–58

ka9

22W

urz

2002

a,20

02b

Uca

gızl

ıC

ave,

Ahm

aria

nla

yers

B,

B1-

3,C

,Tu

rkey

Upp

erP

aleo

lithi

cca

.35

–32

ka9

15K

uh

net

al.

2009

Perreault et al. Measuring the Complexity of Lithic Technology S403

Figure 1. “Conservative” and “nonconservative” count of procedural units for 13 Paleolithic assemblages (see table 1) plotted againsttheir age. Note the logarithmic scale (base 10) on both axes. Assemblages with a range of ages are plotted against the midpoint ofthe age range. The line represents the best-fit linear model using least squares regression (conservative: β p �.184, standard error

, , , adjusted ; nonconservative: B p β p �.314, , , , adjusted2[SE] p 0.05 P p .005 n p 13 R p 0.49 SE p 0.06 P p .0003 n p 13). The dashed lines represent the 95% confidence bands, and the dotted lines represent the 95% prediction bands.2R p 0.68

Figure 2. Relative frequency histogram of the effect size and variance explained of the best-fit linear model using least squaresregression calculated over 1,000 data sets generated by randomly selecting for each of the 13 assemblages either the conservativeor the nonconservative count of procedural units (see table 1).

complexity in the archaeological record. As such, this methodwill allow us to revisit several claims about the complexity ofPaleolithic technologies. For instance, do Clark’s technologicalmodes really represent an increase in complexity (Foley andLahr 2003)? Are Upper Paleolithic tools more complex thanMiddle Paleolithic ones (Mellars 1989), or is the Levalloismethod more complex than the production of blades fromprismatic cores (Bar-Yosef and Kuhn 1999)? Our preliminaryanalysis suggests the existence of a long-term trend towardgreater complexity in the evolution of lithic technologiesthroughout the Paleolithic period but that there is a partic-ularly sharp increase in complexity within the Middle Pa-leolithic and the Middle Stone Age. A larger sample of as-

semblages will allow us to verify this finding and comparethe long-term rates of change in technological complexitybetween Europe, Africa, and Asia as well as between differentperiods of the Paleolithic.

Nonetheless, there are several caveats to the procedural-unit approach. For instance, it fails to capture some aspectsof what is commonly meant by “technological complexity,”such as the level of skills involved in the manufacturing oftools or the complexity that emerges from the hierarchicalorganization of some manufacturing procedures (Byrne2007). A second caveat is that the procedural-unit approachis not completely impervious to taphonomic issues. For in-stance, the complexity of lithic technologies could plateau at

S404 Current Anthropology Volume 54, Supplement 8, December 2013

Figure 3. “Conservative” and “nonconservative” count of procedural units for eight Middle Paleolithic/Middle Stone Age assemblagesplotted against their age. Note the logarithmic scale (base 10) on both axes. Assemblages with a range of ages are plotted againstthe midpoint of the age range. The line represents the best-fit linear model using least squares regression (conservative: β p �.46,standard error , , , adjusted ; nonconservative: B p β p �.70, , , ,2[SE] p 0.1 P p .004 n p 8 R p 0.74 SE p 0.1 P p .0003 n p 13adjusted ). The dashed lines represent the 95% confidence bands, and the dotted lines represent the 95% prediction bands.2R p 0.88

certain points in time while the complexity of other tech-nological parts made of perishable material may continue toincrease. This means that the procedural-unit approach canbe biased by false negatives and as such provides us only witha lower bound for technological complexity. Moreover, dif-ferent classes of material, such as lithic and ceramic, can alsobe subject to different taphonomic processes and thereforebe less amenable to comparison. And like other measures ofrichness in the archaeological record (e.g., Cannon 2001;Cochrane 2003; Meltzer, Leonard, and Stratton 1992; Rhode1988), the count of procedural units is likely dependent onsample size.

Another source of problems is the effect of subjective in-ferences on the count of procedural units. Some proceduralunits are harder than others to identify in an archaeologicalassemblage, and not every analyst will be comfortable withthe inferential leap required to mark them as “present”—theuse of a soft hammer versus indirect percussion comes tomind here. Other operator errors can affect the replicabilityof the method, such as disagreement between operators onwhether something like rotating the core during core reju-venation constitutes a procedural unit or not. As such, theprocedural-unit approach to lithic complexity suffers fromthe same kinds of subjectivity that prevail in lithic analysis ingeneral.

Our proof-of-concept study, however, does provide ten-tative evidence that the method is sufficiently robust to sub-jectivity in the definition of lithic procedural units. We wereable to detect a temporal trend in the complexity of Paleolithiclithic technology even when the list of procedural units ex-amined was reduced by 54%. Even though the proceduralunits discussed in this paper are uncontroversial and fre-

quently discussed in the archaeological literature, it would beuseful to verify the reliability of the approach by measuringthe level of agreement between different operators and ex-perimental context. But the problem of subjectivity can alsobe addressed statistically. By analyzing large samples we canaverage out the noise created by factors that are independentof the variable of interest, such as operator errors, between-operator disagreements, or differential preservation. This isnot an attempt to avoid the issue of subjectivity in definingprocedural units. On the contrary, the problem of noise inempirical data is not specific to the procedural-unit approach:it is a problem that every scientific discipline faces, and themost powerful way we have to deal with it is to collect largersamples. The standard error of the mean, , where�S p s / nx

s is the standard deviation of the sample and n the size ofthe sample, is a good measure of the effect of errors on themean of a sample.1 Because the standard error of the meandecreases in proportion to the square root of sample size, forany given amount of error in a sample there will always bea sample size that is large enough to estimate accurately the

1. In an ideal and purely mechanistic world, , where y is they p xvariable of interest (e.g., lithic complexity) and x is the predictor (e.g.,the age of assemblages). But we live in a world where , wherey p x � ε

ε is the noise in the data. In most cases this error will be random withrespect to x; some of us will tend to overestimate the number of pro-cedural units present in assemblages while others will tend to underes-timate it. The standard error of the mean, a measure of the effects oferrors on the mean of a sample, is used to calculate the confidence intervalwithin which the true mean of a population lies. For example, if theaverage complexity in a sample of Middle Stone Age assemblages of sizen is x with a standard deviation of s, then there is a 95% chance thatthe true mean complexity in the Middle Stone Age lies within the intervaldefined by .�x � 1.96(s / n)

Perreault et al. Measuring the Complexity of Lithic Technology S405

true mean complexity of a sample. The same line of reasoningcan be extended to regression analysis. In sum, the influenceof operator errors can be dealt with statistically. This allowsfor the procedural-unit approach to remain a useful metricby which the complexity of prehistoric technologies can bequantified and compared and to test a wide range of hy-potheses about what drives changes in technological com-plexity.

Acknowledgments

We thank Christian Tryon and Erella Hovers for providingus with their data. We also thank the Wenner-Gren Foun-dation, Leslie Aiello, the participants of the Wenner-Grensymposium no. 145, and the reviewers for their commentsand criticisms of our paper.

References CitedAlperson-Afil, N., and Erella Hovers. 2005. Differential use of space at the

Neandertal site of Amud Cave, Israel. Eurasian Prehistory 3:3–22.Ambrose, Stanley H. 2001. Paleolithic technology and human evolution. Sci-

ence 291(5509):1748–1753.———. 2010. Coevolution of composite-tool technology, constructive mem-

ory, and language. Current Anthropology 51(suppl. 1):S135–S147.Bar-Yosef, O., and S. L. Kuhn. 1999. The big deal about blades: laminar

technologies and human evolution. American Anthropologist 101(2):322–338.

Beatty, J. 1995. The evolutionary contingency thesis. In Concepts, theories, andrationality in the biological sciences. G. Wolters and J. G. Lennox, eds. Pp.45–81. Pittsburgh, PA: University of Pittsburgh Press.

Bleed, Peter. 1986. The optimal design of hunting weapons: maintainabilityor reliability. American Antiquity 51(4):737–747.

Bonner, John Tyler. 1988. The evolution of complexity by means of naturalselection. Princeton, NJ: Princeton University Press.

Bousman, C. Britt. 1993. Hunter-gatherer adaptations, economic risk and tooldesign. Lithic Technology 18(1/2):59–86.

Brantingham, P. Jeffrey, John W. Olsen, Jason A. Rech, and Andrei I. Kri-voshapkin. 2000. Raw material quality and prepared core technologies innortheast Asia. Journal of Archaeological Science 27(3):255–271.

Brown, Kyle S., Curtis W. Marean, Andy I. R. Herries, Zenobia Jacobs, ChantalTribolo, David Braun, David L. Roberts, Michael C. Meyers, and JocelynBernatchez. 2009. Fire as an engineering tool of early modern humans.Science 325(5942):859–862.

Byrne, R. W. 2007. Culture in great apes: using intricate complexity in feedingskills to trace the evolutionary origin of human technical prowess. Philo-sophical Transactions of the Royal Society B 362(1480):577–585.

Cannon, Michael D. 2001. Archaeofaunal relative abundance, sample size, andstatistical methods. Journal of Archaeological Science 28(2):185–195.

Chaitin, G. J. 1970. On the difficulty of computations. IEEE Transactions onInformation Theory IT-16:5–9.

Cochrane, Grant W. G. 2003. Artefact attribute richness and sample sizeadequacy. Journal of Archaeological Science 30(7):837–848.

Collard, Mark, Briggs Buchanan, Jesse Morin, and Andre Costopoulos. 2011.What drives the evolution of hunter-gatherer subsistence technology? areanalysis of the risk hypothesis with data from the Pacific Northwest.Philosophical Transactions of the Royal Society B 366(1567):1129–1138.

Collard, Mark, Michael Kemery, and Samantha Banks. 2005. Causes of toolkitvariation among hunter-gatherers: a test of four competing hypotheses.Canadian Journal of Archaeology 29:1–19.

Coolidge, Frederick L., and Thomas Wynn. 2009. The rise of Homo sapiens:the evolution of modern thinking. Malden, MA: Wiley-Blackwell.

Cornelissen, E. 1992. Site GnJh-17 and its implications for the archaeology ofthe Middle Kapthurin Formation, Baringo, Kenya, vol. 133 of Annales, scienceshumaines. Tervuren: Musee Royale de l’Afrique Centrale.

de Beaune, Sophie A. 2004. The invention of technology: prehistory andcognition. Current Anthropology 45(2):139–162.

de la Torre, Ignacio, Jorge Martınez-Moreno, and Rafael Mora. 2013. Changeand stasis in the Iberian Middle Paleolithic: considerations on the signifi-cance of Mousterian technological variability. Current Anthropology54(suppl. 8):S320–S336.

d’Errico, Francesco, and Christopher Henshilwood. 2007. Additional evidencefor bone technology in the southern African Middle Stone Age. Journal ofHuman Evolution 52(2):142–163.

d’Errico, Francesco, Christopher Henshilwood, Marian Vanhaeren, and Karenvan Niekerk. 2005. Nassarius kraussianus shell beads from Blombos Cave:evidence for symbolic behaviour in the Middle Stone Age. Journal of HumanEvolution 48(1):3–24.

Ekshtain, R. 2006. A comparative study of knapping accidents in the LevantineMousterian. Jerusalem: Hebrew University of Jerusalem.

Foley, Robert. 1987. Another unique species. New York: Longman Scientific &Technical.

Foley, Robert, and Marta Mirazon Lahr. 2003. On stony ground: lithic tech-nology, human evolution, and the emergence of culture. Evolutionary An-thropology: Issues, News, and Reviews 12(3):109–122.

Goder, M. 1997. Technological and typological aspects of Layer B1 in the LateMousterian site at Amud Cave. MA thesis, Hebrew University of Jerusalem.

Gowlett, John. 1996. The frameworks of early hominid social systems: howmany useful parameters of archaeological evidence can we isolate? In Thearchaeology of human ancestry: power, sex and tradition. J. Steele and S.Shennan, eds. Pp. 135–183. London: Routledge.

Haidle, Miriam Noel. 2009. How to think a simple spear. In Cognitive ar-chaeology and human evolution. S. de Beaune, F. L. Coolidge, and T. Wynn,eds. Pp. 57–73. Cambridge: Cambridge University Press.

———. 2010. Working-memory capacity and the evolution of modern cog-nitive potential. Current Anthropology 51(suppl. 1):S149–S166.

Hovers, Erella. 1998. The lithic assemblages of Amud Cave: implications forthe end of the Mousterian in the Levant. In Neanderthals and modernhumans in Western Asia. T. Akazawa, K. Aoki, and O. Bar-Yosef, eds. Pp.143–163. New York: Plenum.

———. 2007. The many faces of cores-on-flakes: a perspective from theLevantine Mousterian. In Cores or tools? alternative approaches to stone toolanalysis. S. P. McPherron, ed. Pp. 42–74. Cambridge: Cambridge Scholars.

———. 2009a. Learning from mistakes: flaking accidents and knapping skillsin the assemblage of A. L. 894 (Hadar, Ethiopia). In The cutting edge: newapproaches to the archaeology of human origins. K. Schick and N. Toth, eds.Pp. 137–150. Gosport, IN: Stone Age Institute.

———. 2009b. The lithic assemblages of Qafzeh Cave. New York: OxfordUniversity Press.

Hovers, Erella, A. Malinsky-Buler, M. Goder-Goldberger, and R. Ekshtain.2011. Capturing a moment: identifying short-lived activity locations inAmud Cave, Israel. In The Lower and Middle Palaeolithic in the Middle Eastand neighbouring regions. J.-M. Le Tensorer, R. Jagher, and M. Otte, eds.Pp. 110–114. Liege: University of Liege Press.

Hovers, Erella, and A. Raveh. 2000. The use of a multivariate display techniquein the analysis of inter-assemblage lithic variability: a case study from QafzehCave, Israel. Journal of Archaeological Science 27(11):1023–1038.

Kelly, Robert L. 1995. The foraging spectrum: diversity in hunter-gatherer life-ways. Washington, DC: Smithsonian Institution.

Kimbel, W. H., R. C. Walter, D. C. Johanson, K. E. Reed, J. E. Aronson, Z.Assefa, C. W. Marean, et al. 1996. Late Pliocene Homo and Oldowan stonetools from the Hadar Formation (Kada Hadar Member), Ethiopia. Journalof Human Evolution 31:549–561.

Kline, Michelle A., and Robert Boyd. 2010. Population size predicts tech-nological complexity in Oceania. Proceedings of the Royal Society B 277(1693):2559–2564.

Kuhn, Steven L. 2013. Roots of the Middle Paleolithic in Eurasia. CurrentAnthropology 54(suppl. 8):S255–S268.

Kuhn, Steven L., Guven Arsebuk, and F. Clark Howell. 1996. The MiddlePleistocene lithic assemblage from Yarimburgaz Cave, Turkey. Paleorient22(1):31–49.

Kuhn, Steven L., Mary C. Stiner, Gulec Erksin, Ismail Ozer, Hakan Yılmaz,Ismail Baykara, Aysen Acıkkol, et al. 2009. The Early Upper Paleolithicoccupations at Ucagızlı Cave (Hatay, Turkey). Journal of Human Evolution56:87–113.

Leakey, M., P. V. Tobias, J. E. Martyn, and R. E. F. Leakey. 1969. An Acheulianindustry with prepared core technique and the discovery of a contemporary

S406 Current Anthropology Volume 54, Supplement 8, December 2013

hominid mandible at Lake Baringo, Kenya. Proceedings of the PrehistoricSociety 3:48–76.

Lombard, Marlize, and Miriam Noel Haidle. 2012. Thinking a bow-and-arrowset: cognitive implications of Middle Stone Age bow and stone-tipped arrowtechnology. Cambridge Archaeological Journal 22(2):237–264.

Marean, C. W., Miryam Bar-Matthews, Jocelyn Bernatchez, Eric Fisher, PaulGoldberg, Andy I. R. Herries, Zenobia Jacobs, et al. 2007. Early human useof marine resources and pigment in South Africa during the Middle Pleis-tocene. Nature 449(7164):905–908.

McBrearty, Sally. 1999. The archaeology of the Kapthurin Formation. In LateCenozoic environments and hominid evolution: a tribute to Bill Bishop. P.Andrews and P. Banham, eds. Pp. 143–156. London: Geological Society.

McBrearty, Sally, and Alison S. Brooks. 2000. The revolution that wasn’t: anew interpretation of the origin of modern human behavior. Journal ofHuman Evolution 39(5):453–563.

Mellars, Paul. 1989. Technological changes across the Middle-Upper Palaeo-lithic transition: economic, social and cognitive perspectives. In The humanrevolution: behavioural and biological perspectives on the origins of modernhumans. P. Mellars and C. Stringer, eds. Pp. 339–365. Princeton, NJ: Prince-ton University Press.

———. 2006a. Going east: new genetic and archaeological perspectives onthe modern human colonization of Eurasia. Science 313(5788):796–800.

———. 2006b. Why did modern human populations disperse from Africaca. 60,000 years ago? a new model. Proceedings of the National Academy ofSciences, U.S.A. 103(45):9381–9386.

Meltzer, David J., Robert D. Leonard, and Susan K. Stratton. 1992. The re-lationship between sample size and diversity in archaeological assemblages.Journal of Archaeological Science 19(4):375–387.

Movius, Hallam, L. 1944. Early man and Pleistocene stratigraphy in southernand eastern Asia. Papers of the Peabody Museum of Archaeology and Eth-nology, vol. 19. Cambridge, MA: Peabody Museum of American Archae-ology and Ethnology.

———. 1948. The Lower Paleolithic cultures of southern and eastern Asia.Transactions of the American Philosophical Society 38:329–420.

Osborne, A. J. 1999. From global models to regional patterns: possible de-terminants of Folsom hunting weapon design diversity and complexity. InFolsom lithic technology: explorations in structure and variation. D. S. Amick,ed. Pp. 188–213. Ann Arbor, MI: International Monographs in Prehistory.

Oswalt, Wendell H. 1976. An anthropological analysis of food-getting technology.New York: Wiley-Interscience.

Pope, G. G. 1989. Bamboo and human evolution. Natural History 10:48–57.Powell, Adam, Stephen Shennan, and Mark G. Thomas. 2009. Late Pleistocene

demography and the appearance of modern human behavior. Science324(5932):1298–1301.

Premo, L. S., and Steven L. Kuhn. 2010. Modeling effects of local extinctionson culture change and diversity in the Paleolithic. PLoS ONE 5(12):e15582.

Reiter, Paul. 2000. From Shakespeare to Defoe: malaria in England in theLittle Ice Age. Emerging Infection Diseases 6(1):1–11.

Rhode, David. 1988. Measurement of archaeological diversity and the sample-size effect. American Antiquity 53(4):708–716.

Richerson, Peter J., and Robert Boyd. 2005. Not by genes alone: how culturetransformed human evolution. Chicago: University of Chicago Press.

Sambrook, T., and A. Whiten. 1997. On the nature of complexity in cognitiveand behavioural science. Theory and Psychology 7(2):191–213.

Schick, Kathy D. 1994. The Movius Line reconsidered: perspectives on theearlier Paleolithic of eastern Asia. In Integrative paths to the past: paleo-anthropological advances in honor of F. Clark Howell. R. S. Corruccini andR. L. Ciochon, eds. Pp. 569–596. Englewood Cliffs, NJ: Prentice Hall.

Shannon, C. E., and W. Weaver. 1949. The mathematical theory of commu-nication. Urbana: University of Illinois Press.

Shea, John J., and Matthew L. Sisk. 2010. Complex projectile technology andHomo sapiens dispersal into western Eurasia. Paleoanthropology (2010):100–122.

Shennan, S. 2001. Demography and cultural innovation: a model and itsimplications for the emergence of modern human culture. Cambridge Ar-chaeological Journal 11(1):5–16.

Shott, Michael J. 1986. Technological organization and settlement mobility:an ethnographic examination. Journal of Anthropological Research 42:15–51.

Torrence, R. 1983. Time budgeting and hunter-gatherer technology. In Hunter-gatherer economy in prehistory: a European perspective. G. Bailey, ed. Pp. 11–22. Cambridge: Cambridge University Press.

Torrence, Robin. 1989. Re-tooling: towards a behavioral theory of stone tools.In Time, energy and stone tools. R. Torrence, ed. Pp. 57–66. Cambridge:Cambridge University Press.

———. 2000. Hunter-gatherer technology: macro- and microscale ap-proaches. In Hunter-gatherers: an interdisciplinary perspective. C. Panter-Brick, R. H. Layton, and P. Rowley-Conwy, eds. Pp. 99–143. Cambridge:Cambridge University Press.

Tryon, Christian A. 2003. The Acheulian to Middle Stone Age transition:tephrostratigraphic context for archaeological change in the Kapthurin For-mation. PhD dissertation, University of Connecticut, Storrs.

———. 2006. “Early” Middle Stone Age lithic technology of the KapthurinFormation (Kenya). Current Anthropology 47:367–375.

Tryon, Christian A., Sally McBrearty, and P. J. Texier. 2005. Levallois lithictechnology from the Kapthurin Formation, Kenya: Acheulian origin andMiddle Stone Age diversity. African Archaeological Review 22(4):199–229.

Wadley, Lyn. 2010. Compound-adhesive manufacture as a behavioral proxyfor complex cognition in the Middle Stone Age. Current Anthropology51(suppl. 1):S111–S119.

Whiten, A., J. Goodall, W. C. McGrew, T. Nishida, V. Reynolds, Y. Sugiyama,C. E. G. Tutin, R. W. Wrangham, and C. Boesch. 1999. Cultures in chim-panzees. Nature 399(6737):682–685.

Wurz, Sarah. 2002a. The Middle Stone Age sequence at Klasies River, SouthAfrica. PhD dissertation, University of Stellenbosch.

———. 2002b. Variability in the Middle Stone Age lithic sequence, 115,000–60,000 years ago at Klasies River, South Africa. Journal of ArchaeologicalScience 29(9):1001–1015.