CS5 CAv53nS6 Cover CAsupplement PMS8641 · 2015-11-07 · Ecological Energetics in Early Homo...

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

HUMAN BIOLOGY AND THE ORIGINS OF HOMO

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 2012Volum

e53

Supp

lemen

t 6Pag

esS267–S496

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

Early Homo: Who, When, and Where

Environmental and Behavioral Evidence

Dental Evidence for the Reconstruction of Diet in African Early Homo

Body Size, Body Shape, and the Circumscription of the Genus Homo

Ecological Energetics in Early Homo

Effects of Mortality, Subsistence, and Ecology on Human Adult Height

Plasticity in Human Life History Strategy

Conditions for Evolution of Small Adult Body Size in Southern Africa

Growth, Development, and Life History throughout the Evolution of Homo

Body Size, Size Variation, and Sexual Size Dimorphism in Early Homo

Male Life History, Reproductive Effort, and Evolution of the Genus Homo

Evolution of Cooperation among Mammalian Carnivores

How Our Ancestors Broke through the Gray Ceiling

The Capital Economy in Hominin Evolution

Origins and Evolution of Genus Homo

GUEST EDITORS: SUSAN ANTÓN AND LESLIE C. AIELLO

Humanness and Potentiality: Revisiting the Anthropological Object in the Context of New Medical Technologies. Klaus Hoeyer and Karen-Sue Taussig, eds.

Alternative Pathways to Complexity: Evolutionary Trajectories in the Middle Paleolithic and Middle Stone Age. Steven L. Kuhn and Erella Hovers, 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.

Previously Published Supplementary Issues

Wenner-Gren Symposium Series Editor: Leslie AielloWenner-Gren Symposium Series Managing Editor: Victoria MalkinCurrent Anthropology Editor: Mark AldenderferCurrent Anthropology Managing Editor: Lisa McKamyBook Reviews Editor: Holley MoyesCorresponding Editors: Claudia Briones (IIDyPCa-Universidad Nacional de Rıo Negro, Argentina; [email protected]), Annede Sales (Centre National de la Recherche Scientifique, France; [email protected]), Michalis Kontopodis (Humboldt Univ-ersitat zu Berlin, Germany; [email protected]), Jose Luis Lanata (Universidad Nacional de Rıo Negro San Carlosde Bariloche, Argentina; [email protected]), David Palmer (Hong Kong University, China; [email protected]), Zhang Yinong(Shanghai University, China; [email protected])

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: [email protected])

Individual subscription rates for 2013: $71 print � elec-tronic, $42 print-only, $41 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: $300 to $600 (print � electronic), $255 to$510 (e-only). Institutional print-only is $300. 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-U.S. subscrip-tions may apply. Free or deeply discounted access is availableto readers 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: [email protected]

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.

� 2012 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.

mailto:[email protected]







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


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

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


Current AnthropologyVolume 53 Supplement 6 December 2012


Human Biology and the Origins of Homo

Leslie C. AielloHuman Biology and the Origins of Homo: Wenner-GrenSymposium Supplement 6 S267

Introduction

Leslie C. Aiello and Susan C. AntonHuman Biology and the Origins of Homo: AnIntroduction to Supplement 6 S269

Setting the Stage

Susan C. AntonEarly Homo: Who, When, and Where S278

Richard PottsEnvironmental and Behavioral Evidence Pertaining to theEvolution of Early Homo S299

Food, Morphology, and Locomotion

Peter S. UngarDental Evidence for the Reconstruction of Diet in AfricanEarly Homo S318

Trenton W. HollidayBody Size, Body Shape, and the Circumscription of theGenus Homo S330

Herman PontzerEcological Energetics in Early Homo S346

Body Size and Growth

Andrea Bamberg Migliano and Myrtille GuillonThe Effects of Mortality, Subsistence, and Ecology onHuman Adult Height and Implications for HomoEvolution S359

Christopher W. Kuzawa and Jared M. BraggPlasticity in Human Life History Strategy: Implications forContemporary Human Variation and the Evolution ofGenus Homo S369

Susan PfeifferConditions for Evolution of Small Adult Body Size inSouthern Africa S383

Gary T. SchwartzGrowth, Development, and Life History throughout theEvolution of Homo S395

J. Michael PlavcanBody Size, Size Variation, and Sexual Size Dimorphism inEarly Homo S409

Richard G. Bribiescas, Peter T. Ellison, and Peter B. GrayMale Life History, Reproductive Effort, and the Evolutionof the Genus Homo: New Directions and Perspectives S424

Models for Cooperation, Sociality, Life History,Body Size, and Brain Size

Jennifer E. Smith, Eli M. Swanson, Daphna Reed, andKay E. HolekampEvolution of Cooperation among Mammalian Carnivoresand Its Relevance to Hominin Evolution S436

Karin Isler and Carel P. van SchaikHow Our Ancestors Broke through the Gray Ceiling:Comparative Evidence for Cooperative Breeding in EarlyHomo S453

Jonathan C. K. WellsThe Capital Economy in Hominin Evolution: HowAdipose Tissue and Social Relationships ConferPhenotypic Flexibility and Resilience in StochasticEnvironments S466

New Perspectives on the Evolution of Homo

Susan C. Anton and J. Josh SnodgrassOrigins and Evolution of Genus Homo: New Perspectives S479

Current Anthropology Volume 53, Supplement 6, December 2012 S267

� 2012 by The Wenner-Gren Foundation for Anthropological Research. All rights reserved. 0011-3204/2012/53S6-0001$10.00. DOI: 10.1086/667709

Human Biology and the Origins of HomoWenner-Gren Symposium Supplement 6

by Leslie C. Aiello

Human Biology and the Origins of Homo is the 143rd Wenner-Gren Symposium and the sixth to be published as an open-access supplement of Current Anthropology. The symposiumwas organized by Susan C. Anton (New York University) andLeslie C. Aiello (Wenner-Gren Foundation) and was heldMarch 4–11, 2011, at the Tivoli Palacio de Seteais in Sintra,Portugal (fig. 1).

Wenner-Gren symposia are intensive week-long workshopmeetings that traditionally focus on big questions in the fieldof anthropology, and the origin of Homo is currently one ofthe biggest questions in the field of hominin paleontology.Although Homo erectus has been known since the 1890s (Pith-ecanthropus erectus; Dubois 1894) and Homo habilis was an-nounced almost 50 years ago (Leakey, Tobias, and Napier1964), new fossil discoveries in the last decade have compli-cated our understanding of early Homo and challenged ourlong-held assumptions about its similarities and differencesto the australopiths as well as to later members of our genus.This necessarily influences our interpretations for the originand evolution of Homo and also highlights the need for a newframework for interpretation of the hard evidence.

The purpose of this symposium was to meet this challenge.The aims were to assess what is currently known about thefossil evidence and the environmental context of early Homoand to set the stage for integrated, multidisciplinary studiesto provide a framework for interpretation of the hard evi-dence. The basic premise of the meeting was that it is essentialto have a solid understanding of how and why modern hu-mans and other animals vary in order to understand theadaptive shifts involved with the evolution of Homo. Partic-ipants in the symposium included paleoanthropologists, hu-man biologists, behavorialists, and modelers, and, to ourknowledge, this is the first time that such a varied multidis-ciplinary group has gathered to focus attention on a majorquestion in hominin evolution.

The collection of papers is introduced by Aiello and Anton(2012), who summarize the current state of our knowledgeabout the origin of Homo, integrate the varied contributions,and give a taste of the potential that this approach has for

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.).

human evolution. The concluding paper by Anton and Snod-grass (2012) draws from the wealth of ideas in this collectionand provides a fresh perspective on three important shifts inhuman evolutionary history: (1) the emergence of Homo; (2)the transition between non-erectus early Homo and H. erectus;and (3) the appearance of regional variation in H. erectus. Itconcludes with a new positive feedback model for the originand evolution of Homo that involves critical elements suchas cooperative breeding, changes in diet, body composition,and extrinsic mortality risk that drive life history.

This symposium builds on the long history of the Wenner-Gren Foundation with hominin evolution. Foundation in-terest began in the 1940s with “The Early Man in Africa”program (1947–1955) that was initiated by Fr. Teilhard deChardin to call attention to the extraordinary significance ofthe human origins in southern Africa, to date the southernAfrica cave deposits, and to facilitate multidisciplinary teamresearch. The “Origins of Man” program followed (1965–1972), under the guidance of Walter William (Bill) Bishop,C. K. (Bob) Brain, J. Desmond Clark, Francis Clark Howell,Louis Leakey, and Sherwood Washburn, and the Foundationcontinues to be an enthusiastic supporter of human originsresearch (see Wood 2011 for a history of this support).

Since the late 1950s, the Foundation has held a number ofpaleoanthropological symposia that have led to landmarkpublications. These include Social Life of Early Man (Wash-burn 1961), African Ecology and Human Evolution (Howelland Bourliere 1963), Classification and Human Evolution(Washburn 1964), Background to Evolution in Africa (Bishopand Clark 1967; Clark 1967), Man the Hunter (Lee and Devore1968), Calibration of Hominoid Evolution: Recent Advances inIsotopic and Other Dating Methods as Applicable to the Originof Man (Bishop and Miller 1972), After the Australopithecines:Stratigraphy, Ecology, and Culture Change in the Middle Pleis-tocene (Butzer and Isaac 1975), Earliest Man and Environmentsin the Lake Rudolf Basin: Stratigraphy, Paleoecology, and Evo-lution (Coppens et al. 1976), and Early Hominids of Africa(Jolly 1978).

To continue this tradition, the Wenner-Gren Foundationis always looking for big questions and innovative new di-rections in all areas of anthropology for future Foundation-sponsored and Foundation-organized symposium meetingsand eventual CA publication. We encourage anthropologiststo contact us with their ideas for future meetings. Information

S268 Current Anthropology Volume 53, Supplement 6, December 2012

Figure 1. Participants in the symposium “Human Biology and the Origins of Homo.” Front: Laurie Obbink, Susan Anton, LeslieAiello, Rick Potts, Andrea Migliano, Jonathan Wells, Peter Ungar. Middle: Katie MacKinnon, Jennifer Smith, Karin Isler, KarenSteudel, Susan Pfeiffer. Back: Josh Snodgrass, Trent Holliday, Gary Schwartz, Tom Schoenemann, Herman Pontzer, Carel van Schaik,Mike Plavcan, Chris Kuzawa, Chris Rainwater, Rick Bribiescas. A color version of this photo appears in the online edition of CurrentAnthropology.

about the Wenner-Gren Foundation and the Symposium pro-gram can be found on the Foundation’s Web site (http://wennergren.org/programs/international-symposia).

References CitedAiello, Leslie C., and Susan C. Anton. 2012. Human biology and the origins

of Homo: an introduction to supplement 6. Current Anthropology 53(S6):S269–S277.

Anton, Susan C., and J. Josh Snodgrass. 2012. Origins and evolution of genusHomo: new perspectives. Current Anthropology 53(S6):S479–S496.

Bishop, Walter William, and J. Desmond Clark, eds. 1967. Background toevolution in Africa. Chicago: University of Chicago Press.

Bishop, Walter William, and J. A. Miller, eds. 1972. Calibration of hominoidevolution: recent advances in isotopic and other dating methods as applicableto the origin of man. Edinburgh: Scottish Academic Press.

Butzer, Karl W., and Glynn L. Isaac, eds. 1975. After the australopithecines:stratigraphy, ecology, and culture change in the middle Pleistocene. New York:de Gruyter.

Clark, J. Desmond, ed. 1967. Atlas of African prehistory. Chicago: Universityof Chicago Press.

Coppens, Yves, Francis Clark Howell, Glynn L. Isaac, and Richard E. F. Leakey,eds. 1976. Earliest man and environments in the Lake Rudolf Basin: stratig-raphy, paleoecology, and evolution. Prehistory Archeology and Ecology Series.Chicago: University of Chicago Press.

Dubois, Eugene. 1894. Pithecanthropus erectus, eine menschenaehnliche Uber-egangsform aus Java. Batavia: Landesdruckerei.

Howell, Francis Clark, and Francois Bourliere, eds. 1963. African ecology andhuman evolution. Viking Fund Publications in Anthropology, no. 36 (Wen-ner-Gren Foundation for Anthropological Research). Chicago: Aldine.

Jolly, Clifford J. ed. 1978. Early hominids of Africa. New York: St. Martin’s.Leakey, Louis S. B., Phillip V. Tobias, and John R. Napier. 1964. A new species

of the genus Homo from Olduvai Gorge. Nature 202:7–9.Lee, Richard B., and Irven DeVore, eds. 1968. Man the hunter. Chicago: Aldine.Washburn, Sherwood L., ed. 1961. Social life of early man. Viking Fund

Publications in Anthropology, no. 31 (Wenner-Gren Foundation for An-thropological Research) Chicago: Aldine.

———. 1964. Classification and human evolution. Viking Fund Publicationsin Anthropology, no. 37 (Wenner-Gren Foundation for AnthropologicalResearch) Chicago: Aldine.

Wood, Bernard. 2011. Wiley-Blackwell encyclopedia of human evolution. Chich-ester, UK: Wiley-Blackwell.

http://wennergren.org/programs/international-symposia

http://wennergren.org/programs/international-symposia



Human Biology and the Origins of HomoAn Introduction to Supplement 6

by Leslie C. Aiello and Susan C. Anton

New fossil discoveries relevant to the origin of Homo have overturned conventional wisdom about the nature ofthe australopiths and early Homo, and particularly Homo erectus (including Homo ergaster). They have eroded priorassumptions about the differences between these genera and complicated interpretations for the origin and evolutionof Homo. This special issue surveys what is now known about the fossil evidence and the environmental contextof early Homo. It also moves beyond the hard evidence and sets the stage for integrated, multidisciplinary studiesto provide a framework for interpretation of the hard evidence. The underlying premise is that to understand theadaptive shifts at the origin of Homo, it is essential to have a solid understanding of how and why modern humansand other animals vary. Contributors to this issue include paleoanthropologists, human biologists, behavorialists,and modelers. We tasked each with bringing her or his special expertise to bear on the question of the origins andearly evolution of Homo. The papers in this collection are a product of a week-long Wenner-Gren symposium heldin March 2011, and this introduction integrates this work and its significance for Homo.

What We Once Knew . . .

The origin of Homo holds particular sway for us and has oftenbeen seen as the point in our evolution when the balance tipsfrom a more ape-like to a more human-like ancestor. By theturn of this century, a conventional wisdom had grown uparound the origin of Homo and particularly Homo erectus thatcast this species as the first hominin to take important bio-logical and behavioral steps in the direction of modern hu-mans (Anton 2003; Shipman and Walker 1989). Homo erectuswas envisioned as a large-brained, small-toothed, long-legged,narrow-hipped, and large-bodied hominin with relatively lowsexual dimorphism. By virtue of a higher-quality, perhapsanimal-based diet, H. erectus is said to have ranged farther,cooperated more, and quickly dispersed from Africa (Aielloand Key 2002; Anton, Leonard, and Robertson 2002; Mc-Henry and Coffing 2000; Walker and Leakey 1993). The pau-city of early Homo fossils of Homo habilis sensu lato (includingHomo rudolfensis) meant that comparisons of Australopithecus((Paranthropus) were made to H. erectus (including Homoergaster) rather than to other early Homo. And the distinctionsbetween Australopithecus and Homo were perhaps overem-

Leslie C. Aiello is President of the Wenner-Gren Foundation forAnthropological Research (470 Park Avenue South, New York, NewYork 10016, U.S.A. [[email protected]]). Susan C. Anton is aProfessor in the Center for the Study of Human Origins, Departmentof Anthropology, New York University (Rufus D. Smith Hall, 25Waverly Place, New York, New York 10003, U.S.A. [[email protected]]). This paper was submitted 12 XII 11, accepted 8VII 12, and electronically published 27 IX 12.

phasized by the diminutive size of the most complete Aus-tralopithecus skeleton (A.L. 288-1; Lucy), on the one hand,and the surprisingly large size of the most complete H. erectusskeleton (KNM-WT 15000; Nariokotome boy), on the other(e.g., Ruff 1993). The comparisons between H. erectus andHomo sapiens were so strongly drawn that the inclusion inthe genus of some of the earliest species, such as H. habilisand H. rudolfensis, was seriously questioned on the basis oftheir more australopith-like postcranial skeleton, among otherthings (Wood and Baker 2011; Wood and Collard 1999, 2007).

The fossil record never ceases to upset conventional wis-dom, and over the past 2 decades, new discoveries from Eastand South Africa, Georgia, and even Indonesia have chal-lenged these stark distinctions between Australopithecus andH. erectus and within non-erectus early Homo. In particular,new small-bodied and small-brained finds from the Republicof Georgia and Kenya call to question claims for universallylarge size in H. erectus (e.g., Gabunia et al. 2000; Potts et al.2004; Simpson et al. 2008; Spoor et al. 2007) and focus ourattention instead on the range of variation within that taxon.This variation in H. erectus has most often been referred toas sexual dimorphism and/or regional/climatic adaptations(Anton 2008; Spoor et al. 2007), although short-term accom-modations and phenotypic plasticity are likely to have playedan important role (see Anton 2013). And larger-sized, longer-legged Australopithecus have been found (Haile-Selassie et al.2010), as have members of that genus who may share somepostcranial characteristics with Homo (Asfaw et al. 1999; Ber-ger et al. 2010; Kibii et al. 2011; Kivell et al. 2011; Zipfel etal. 2011). Additionally, new fossil remains of non-erectusHomo and new work on previously known remains emphasize



the diversity of the early members of the genus and the waysin which they differ from Australopithecus (Blumenschine etal. 2003; Spoor et al. 2007).

Yet despite this increased appreciation of variation in earlyfossil Homo, little time has been spent evaluating the rela-tionships between morphology and behaviors in extant taxa,especially modern humans, in different ecological circum-stances. We maintain that these are the data that are essentialto create a more nuanced understanding of the implicationsand expectations of anatomical changes at the origin of ourgenus—an understanding that goes beyond simple assump-tions of sexual or climatic variation.

Topic and Rationale

The increasing number of early Homo fossil finds and theirdiversity in size and shape suggest that this discussion sur-rounding the origin and evolution of early Homo is likely toform a major focus for the next decade of paleoanthropo-logical work. As such, our goal was to bring together humanbiologists, behaviorists, modelers, and fossil experts to inte-grate the rich extant data sets with the new details of the fossilrecord. Understanding the adaptive shifts at the origin ofHomo is dependent on a solid understanding of how and whymodern humans vary and particularly on the relationshipbetween human behavior, human morphology, and humanlifestyle and life history variation.

Among the highly variable features in living humans arefeatures such as body size and aspects of life history thatseparate us from other primates. In many cases, human var-iation in, for example, growth rates, fertility, and perhaps evenlifespan, can be traced to such environmental or behavioralfactors as nutritional sufficiency and unavoidable (extrinsic)mortality. Such phenotypic plasticity provides a more rapidresponse to environmental challenges than does geneticchange, but the fact that genetic change can follow has beenlong suggested in human biology (e.g., Kuzawa and Bragg2012). Thus, understanding the causes of human phenotypicplasticity can provide important clues to understanding bothwithin and between species variation in the morphology ofour hominin ancestors.

Unavoidably, the biology of early Homo will be unlike thatof ourselves, however—and therefore, primate and mam-malian trends are also important to understand. And becauseat some point cooperation in hunting or breeding becameimportant to survival, considering how both carnivory andcooperation influence life history, body size, and body shapeand whether they leave a detectable signal are important con-siderations as well.

We set out, then, to probe the meaning of the newly iden-tified ranges of variation in size and shape in early Homobased on empirical evidence of how extant humans, non-human primates, and social carnivores respond energetically,physiologically, and socially to changes in resource availabilityand to stress from climatic, environmental, and other factors.

We argue that understanding the response of extant organ-isms, especially humans, in shifting environments providesan ideal basis for understanding the integration of bioculturalresponses to environmental constraints. The application ofthese data in light of the known fossil record can help us tounderstand these past populations, their constraints and adap-tive strategies. By mining the rich data sets of our subspe-cialties, we sought to forge a stronger and more nuancedunderstanding of the adaptive shifts that can—or cannot—be inferred at the base of our genus and to set out a seriesof hypotheses and predictions to be tested against future fossiland archaeological data. The results of an intense 5-day Wen-ner-Gren Symposium in Sintra, Portugal, in March 2011 andour follow-up analyses are presented in this special issue.

Setting the Stage

We begin the volume, as we did the symposium, by reassessingthe fossil foundation of what we now know regarding genusHomo. Anton (2012) provides an overview of the genus andits species and the differences between Australopithecus andHomo. The first recognizable members of the genus Homoappear at approximately 2.3 Ma, suggesting that the genusevolved earlier, but substantial fossil evidence does not appearuntil about 2.0 Ma. Her paper focuses our attention on theimportance of individual fossil data points for understandingdiversity within and between groups, and she concludes thata strong case can be made for at least three different morphsbetween 2.0 and 1.5 Ma: an 1813-group, a 1470-group, andHomo erectus (including Homo ergaster). She avoids the useof taxonomic names for the 1813-group and 1470-group be-cause of uncertainty over group affiliation of type specimensfor early Homo species (e.g., Homo habilis and Homo rudol-fensis; Leakey et al. 2012). Her paper also provides an intro-duction to what is now known about the distinct featuresthat separate the three different morphs from each other andfrom Australopithecus (see Anton 2012, tables 1–8, and Antonand Snodgrass 2012, tables 1–6, for membership of thesemorphs and distinguishing morphological and inferred be-havioral features). Additionally, she suggests that, on average,early Homo is larger of body and brain than Australopithecus,and H. erectus is larger than other early Homo. That said, thesurprising facts, particularly to those who have been involvedin paleoanthropology for a considerable time, are the degreeof diversity within the morphs and that, in some ways, themorphs are more similar to each other than was previouslyimagined. For example, all early Homo, including H. erectus,may exhibit substantial amounts of sexual dimorphism, andH. erectus is less fully modern in body proportions than hasbeen previously claimed. These themes and their implicationsare further plumbed in the contributions by Holliday (2012),Pontzer (2012), and Plavcan (2012).

Fossils cannot be understood and interpreted without theircontext, and Potts (2012) provides an overview of the envi-ronmental and archaeological background for the evolution

Aiello and Anton Origins of Homo S271

of Homo in eastern Africa between 3.0 and 1.5 Ma. He con-cludes that there were major episodes of moist-arid variabilityduring this period, superimposed on an overall drying trend.The first appearance of Homo at approximately 2.3 Ma (Kim-bel et al. 1996; and of the Oldowan at approximately 2.58Ma; Semaw et al. 2003; but see McPherron et al. 2010) aswell as the proliferation of the genus after 2.0 Ma coincidewith particularly high levels of climate variability, suggestingthat adaptive plasticity in its broadest developmental, physi-ological, and behavioral manifestations was integral to theevolution of Homo. For example, stone tools, which have astrong stratigraphic persistence in the archaeological recordafter 2.0 Ma, provide an efficient behavioral mechanism toenhance foraging ability, enabling predictable returns in achanging environment. But they also pose an energetic chal-lenge of material transport over distances as great as 1–13 kmby about 2.0 Ma (Braun et al. 2008).

Food, Morphology, and Locomotion

One means of offsetting the energetic cost of tool and rawmaterial transport as well as increased body and brain size isdietary expansion to higher-quality food resources, whichmight involve access to animal resources (as well as a widerrange of plant food; Aiello and Wells 2002; Aiello and Wheeler1995; Leonard and Robertson 1997). Such resources wouldalso serve to buffer environmental instability and resultingchanges of food resources across space and over time (Potts2012). Some direct evidence for a dietary shift in early Homo(including even more substantial changes in Homo erectus)relative to the diet of Australopithecus is provided by Ungar(2012), who reviews dental macro and micro anatomy andwear. In particular, all early Homo teeth are most similar toextant animals that do not use fracture-resistant foods. Inother words, the genus seems not to have used particularlyhard-brittle foods or especially tough foods. However, withinearly members of the genus, there are some differences thatsuggest a broader subsistence base for H. erectus that includedmore tough foods than other early Homo. This dental evidenceis consistent with increased meat eating (or eating other non-brittle foods) and tool use in food preparation (perhaps evencooking) over the condition in Australopithecus, with abroader range of foods eaten by H. erectus than other earlyHomo. These results could be a hard tissue signal of dietaryand behavioral plasticity to temper environmental vacillation.

Like the dietary results, other papers suggest that someadaptations once thought to appear with H. erectus arise atthe origin of the genus or even earlier. Holliday (2012) pro-vides new analyses and an overview of our current knowledgeof body size and body proportions. The unexpected outcomeis that our prior understanding that H. erectus was uniqueamong the early hominins in having long legs and a narrow,heat-adapted body is wrong. Leg length scales with body mass,and large-bodied australopiths have long, human-like legs andhuman-like thoraces, while new analyses of the H. erectus

pelvis demonstrate that its bi-illiac breadth was broad andaustralopith-like (Simpson et al. 2008; but see Ruff 2010).Because locomotor efficiency is primarily a function of relativelimb length (Pontzer 2012), and because limb length is al-lometrically related to body size in all hominins, larger-bodiedindividuals of any taxon would be more efficient in walkingand running, with faster optimal speeds and increased ab-solute speed. Long leg length and arm length also have ther-moregulatory advantages in hot climates, which would be adistinct advantage either during locomotion or at rest.

Given the similar scaling relationships between limb lengthsand body size across hominins, Holliday concludes that thereis little evidence of a major locomotor shift between Aus-tralopithecus and early Homo (including H. erectus), a pointthat is shown by the analyses presented by Pontzer (2012) aswell. However, they note that a significant difference remainsbetween these genera in terms of mean body size; early Homois approximately 33% larger than Australopithecus, and H.erectus is approximately 15% larger than other early Homo,even when the recently discovered small H. erectus fossils (e.g.,from Georgia, Kenya, Tanzania, and perhaps Ethiopia) andlarge Australopithecus are included in estimates.

Despite having similar proportions as earlier hominins, anumber of symposium contributions emphasize that largersize itself has important energetic, locomotor, and survivalconsequences for Homo. Holliday (2012) and Pontzer (2012)point out that across mammals, larger body size equates witha larger home range size, which would be exaggerated furtherif Homo was also more carnivorous (Anton, Leonard, andRobertson 2002). Pontzer (2012) develops the implicationsof this by demonstrating that across mammals there is noselection for greater locomotor efficiency (as proxied bychanges in limb proportions) in those species with largerhome range sizes. Instead, species that travel farther adopt ahigh-throughput strategy (increased daily energy expenditurein relation to body size and a correspondingly greater repro-ductive investment), resulting in greater lifetime reproductiveoutput. This suggests that in Homo, as in other far-rangingmammals, there must have been an increased energy budgetto provide for increased brain and body growth and repro-duction. Outside of a more calorie-rich diet at the carnivorousend of the omnivorous spectrum, Pontzer (2012) argues thatan increased daily energy expenditure would suggest greaterfood availability, perhaps implying the origins of food sharing.These results are consistent with the symposium papers onliving humans and extant carnivores by Migliano and Guillon(2012), Kuzawa and Bragg (2012), and Smith and colleagues(2012).

Body Size and Growth

While height in humans is influenced by a number of en-vironmental and idiosyncratic factors (see references in Ku-zawa and Bragg 2012; Migliano and Guillon 2012), it isachieved through a combination of speed and duration of


growth, which is in turn dependent on resource availabilityand mortality probability (Kuzawa and Bragg 2012; Miglianoand Guillon 2012). This may provide a clue for possible in-terpretations of size variation in Homo. In human popula-tions, the greater the probability of mortality, the earlier isthe age of maturity (and the shorter the period of growth),to ensure maximum reproductive output. Based on the anal-ysis of small-scale human societies, Migliano and Guillon(2012) demonstrate that the main determinant in height var-iability in humans is the probability of mortality—the lowerthe mortality probability, the taller the populations. They alsoshow that environment has an important effect, although dietis not significant in their analyses. This probably results fromdata limitations in their large comparative analysis.

Kuzawa and Bragg (2012) emphasize this nutritional sideof the equation and argue that nutritional abundance is as-sociated with a faster growth rate, earlier maturity, and largeradult sizes. Because males are affected more than females,sexual size dimorphism is increased with greater nutritionalabundance. Nutritional stress has the opposite effects—slowergrowth, later maturity, and reduced size dimorphism. If mor-tality rates were high and precluded later maturity, smallerbody size would be expected.

Pfeiffer’s (2012) work on the bioarchaeological record ofthe small-bodied KhoeSan, however, reminds us of the mul-tifactorial effect on size and of the issue that small size maybe the default in the absence of selective factors for largersize. That is, bigger may not always be better. In her particularcase study she finds no evidence for the traditional drivers ofsmall size: nutritional insufficiency, early maturation, highextrinsic mortality, or climate change. In light of this, shesuggests instead that the long-term relaxation of selection forlarge size was allowed due to the relative isolation of theKhoeSan and therefore an absence of competition with large-sized human populations. This might favor increasingly largemale size and shape changes to the pelvis that accommodatedrelatively large infants in small mothers, which might oth-erwise favor large females. It is important, then, to considerthe multiple and sometimes conflicting causes for size changein light of their influence on developmental plasticity.

At present we lack the detailed data from the fossil recordto assess intraspecific differences in growth and developmentwith the aim of inferring the possible roles of nutrition, mor-tality probability, or other factors in shaping the observed sizedifferences in the hominins. However, because the size varia-tion, particularly in Homo erectus, is similar to that found inmodern humans (Migliano and Guillon 2012), there is everyreason to assume that similar factors were in play and that sizedifferences in the hominins also reflect an adaptive plasticitysimilar to that observed in modern humans. Migliano andGuillon (2012), Kuzawa and Bragg (2012), and Bribiescas andcolleagues (2012) also emphasize the important role of behav-ior, and particularly cooperation, in buffering both nutritionalsufficiency and mortality probability, thereby setting the stagefor body size increase. A series of alternative scenarios for con-

sidering especially the regional variation in H. erectus derivedfrom these principles are developed in detail in the concludingpaper of this issue (Anton and Snodgrass 2012).

Although more data are sorely needed, what we know aboutthe tempo and mode of growth in the early hominins, basedon rates of dental maturation, is summarized by Schwartz(2012). These data suggest that both Australopithecus and earlyHomo have more rapid maturation than Homo sapiens, whichcould reflect environments of higher extrinsic mortality. How-ever, there are probably differences between the genera as well.Dental eruption in Australopithecus and Paranthropus is com-parable to, or faster than, Gorilla gorilla beringei, the fastestof the living great apes. Dental eruption in H. erectus is equiv-alent to Pongo pygmaeus pygmaeus, the slowest of the livinggreat apes, and just below the large range of eruption ages inmodern humans. The somewhat extended developmentalschedule of H. erectus relative to Australopithecus is consistentwith mortality reduction and increased body size. There isstill much to learn, but one definite conclusion of Schwartz’ssynthesis is that the full suite of modern human life historywith extended periods of growth and development was notpresent in early Homo, including H. erectus, and possibly didnot appear in its modern form until much later in time.

While these papers address size differences between humanpopulations and what we know about the tempo of homininmaturation, Plavcan (2012) raises the important issue of sex-ual size dimorphism. He provides a detailed overview of whatis currently known (and knowable) about sexual size dimor-phism in hominins and living primates and provides a num-ber of caveats in relation to interpretation of the evidence. Ithas long been assumed that sexual dimorphism is a featureobservable in hominins that can be directly and causally re-lated to social behavior across primates, and particularly tomale competition over mates (Leigh 1992; Plavcan 2001; Plav-can and van Schaik 1997a, 1997b). However, Plavcan cautionsthat although all highly dimorphic primate species are po-lygynous, the inverse relationship is not straightforward, andany hominin inferences can only be made with extreme cau-tion.

A particularly important aspect of his research is the focuson female size. He argues that female size represents the op-timum for the particular environment, and male size repre-sents a trade-off between the costs of deviating from thisoptimum and the benefits of larger size in mate competition.Female size change alone does not result in marked changesin dimorphism across extant species. Although there is somevariation within species, without a change in mating com-petition, which drives male size increase or decrease, an in-crease in female size should simply be tracked by an equivalentincrease in male size. Beyond this, he notes that size variationwithin H. erectus is “unremarkable” relative to human levelsand that our understanding of sexual size dimorphisms re-quires better understanding of temporal changes in male andfemale size relative to each other in both humans and non-human primates. He makes a specific call for further system-


atic research on intraspecific geographic and sexual variationin primates.

The evolution of male life history trade-offs has not beena major focus in hominin evolution, outside of the relativelysimple association between reduced dimorphism and a moveaway from polygynous social systems and strong size-drivenmating competition. However, Bribiescas and colleagues(2012) argue that it is central to the evolution of Homo.Consistent with Plavcan’s interspecific analyses demonstratingno correspondence between reduced dimorphism and anyparticular social system, Bribiescas and colleagues (2012)point to the fact that humans, with their relatively reduceddimorphism, are unique among the apes not only in thedegree of their paternal parenting behavior but also in itsvariation. There is variation in the amount of time and energyinvested and in the type of offspring care, provisioning, andother involvement. Paternal parenting behavior is dependenton context, including the availability of other caregivers suchas grandmothers or siblings.

The important point is the move away from energy in-vestment in large male size that permits not only energy al-location to parenting behavior but also to other aspects oflife history including fertility and longevity. One particularlynovel aspect of their work is the argument that increased malefertility at older ages may have contributed to the emergenceof female longevity and the evolution of the female postre-productive lifespan and increased female reproductive effortthrough grandmothering and child care. This provides analternative, or perhaps complementary, explanation to theGrandmothering Hypothesis (Hawkes et al. 1998, 2011; Ka-chel, Premo, and Hublin 2011a, 2011b; O’Connell, Hawkes,and Blurton Jones 1999).

It is now clear that we cannot be sure whether there wasany significant difference in dimorphism between Australo-pithecus and Homo (either early Homo or H. erectus ; Anton2012; Holliday 2012; Plavcan 2012; Pontzer 2012). However,if we could be sure that dimorphism was reduced in H. erectus,we would have direct evidence of a change in mating behaviorleading to a reduction of sexual selection acting on male size,and likely involving a major change in social organizationinvolving increased levels of cooperation and allocare, andperhaps increased longevity. Given the importance of under-standing the relationship between male and female size, itwould be useful to expend the effort to understand morecritically the relationship between skeletal and body mass di-morphism and how it varies among modern human popu-lations.

Models for Cooperation, Sociality, LifeHistory, Body Size, and Brain Size

Although we cannot be confident that there was a change insexual size dimorphism associated with the evolution ofHomo, the idea that cooperation and food sharing may havebeen major distinguishing factors between the Australopithe-

cus and early Homo was a common theme that developedfrom a variety of perspectives throughout the symposium.Many of the previously mentioned papers suggest that theabilities to maximize food and to limit predation are criticalto increasing body and brain size. Food sharing and conse-quent group cooperation are means of achieving this. In lightof the several lines of evidence pointing toward the impor-tance of sociality and cooperation, several symposium par-ticipants explored the correlates of such behavior in extantorganisms as well as the concept of cooperation as a meansof expanding an organism’s “capital.”

Nonhuman primates provide the logical starting point be-cause of their close phylogenetic relationship to humans. Theydemonstrate the roots of human evolutionary plasticity par-ticularly in dietary/niche expansion, extended life history, andincreasing social complexity with extensive cooperation andcommunication (Anton and Snodgrass 2012). These abilitiesprovide the basis for the elaborate niche construction ob-served in humans that involves accelerating biocultural com-plexity and an increasing reliance on cooperation in all aspectsof hominin life (Fuentes, Wyczalkowski, and MacKinnon2010; Odling-Smee, Laland, and Feldman 2003).

However, primates are not the only models for homininbehavior, and useful insights can be drawn from, for example,other large-brained animals such as dolphins, cooperativebreeders among all orders, and large-bodied mammal speciesthat inhabit woodland and savanna environments. In the late1960s, Schaller and Lowther (1969) wrote a now classic paperon the relevance of carnivore behavior to the study of earlyhominins. Their basic premise was that to understand socialityin hominins, it would be productive to draw inferences fromanimals that are ecologically similar, such as social carnivores,as well as animals that were closely related, such as the pri-mates. At the time, this work represented a major innovationin the interpretation of early hominin behavior. Smith andcolleagues (2012) take up where Schaller and Lowther left offmore than 40 years ago to demonstrate commonalities inbehavior and morphology between humans and the Carni-vora. Sociality among carnivores is the exception rather thanthe rule, but significant features related to sociality and co-operation in the Carnivora include cursorial hunting of largegame in open habitats, a relatively tall body build (shoulderheight in relation to body mass), reduced sexual dimorphism,larger brains, a high reproductive output (in this case largerlitters), allocare of infants, increased weaning age, and largerpopulation density. Many of these features (with the possibleexception of reduced sexual dimorphism) are reminiscent ofthe morphology of Homo and may help to infer behavior andlife history of early Homo (including Homo erectus). Indeed,many of these features in social carnivores help to reinforcethe inferred relationships between dietary change, morphol-ogy, and cooperation reached by other symposium partici-pants using other data sets.

Brain size expansion in Homo may provide another, in-dependent avenue for inferring the presence of cooperative


breeding. Isler and van Schaik (2012) draw on their previouswork on the Expensive Brain Hypothesis (Isler and van Schaik2009) and on a large comparative mammalian database tounderscore the relationship between brain size increase andcooperation in the form of allocare. They argue that acrossmammals, large brain size is generally correlated with a re-duction in population growth rate. This correlation is drivenby the extended ontogenetic periods necessary for the growthof larger-bodied and larger-brained offspring and the corre-spondingly longer interbirth intervals required. Allocare pro-vides extra resources to the mother, resulting in early weaningof the infant and a shorter interbirth interval. An importantaspect of their work is the prediction based on extant primatesthat a mean endocranial capacity of 600–700 cc would be the“gray ceiling” beyond which cooperation in the form of allo-care would be essential if the population were to grow fastenough to replace itself and avoid extinction. The great apesseem to be at the very farthest extension of this relationship,just barely allowing population replacement without allocare.The fact that the mean brain size of the largest-bodied Aus-tralopithecus species (478 cc in Australopithecus afarensis; Hol-loway, Broadfield, and Yuan 2004) converges on that of mod-ern apes suggests that cooperative breeding probably had notyet appeared in these early hominins. However, by the timeof H. erectus, with brain sizes uniformly over 700 cc, repro-ductive cooperation would seem to have been a necessity.

Allocare is undoubtedly a key element that enabled hom-inins to break through the gray ceiling, but it is only oneelement of capital in hominin evolution (Kaplan, Lancaster,and Robson 2003; Kaplan et al. 2000). Wells (2012) sees capitalas a key and integrating concept that brings together manyof the themes of intraspecific and interspecific variation andplasticity that emerged from the symposium. He defines cap-ital as a generalized energy currency that can be expended ina variety of ways to increase adaptive flexibility. This resultsin the fact that humans are “uniquely under-committed” toany specific niche.

Wells (2012) talks about social capital as well as physicalcapital. Social capital is facilitated by larger brain sizes, storedin social relationships and variably expended to achieve pred-ator protection or to enable food security (particularlythrough sexual division of labor and allocare). Physical capitalis stored within the body as adipose tissue or extracorporeallyin food hoards and similarly used to avoid predators andprovide sufficient nutrition. Because it is difficult to assessfrom the fossil record, adipose tissue has received relativelylittle attention in human evolution. Yet it is a major featuredistinguishing us (and our sexual dimorphism; see Anton andSnodgrass 2012; Plavcan 2012) from nonhuman primates andone that buffers the costs of reproduction against food short-ages in fluctuating environments (Knott 1998; Kuzawa 1998).It also is the source of signaling molecules responsible forenergy trade-offs between competing biological functionssuch as growth, immune function, and reproduction (Wells2009). It may be one of the major factors responsible for

differences in life history strategies among human popula-tions, and it has been correlated with large brain sizes acrossmammals (Navarrete, van Schaik, and Isler 2011). It is thusimportant to develop creative ways to infer adiposity fromfossil record.

Storing energy in generalized currencies (social relation-ships and adipose tissue) means that various aspects of lifehistory (growth, reproduction, and immunity) can be“funded” according to the state of the environment and over-all energy availability (Wells 2012). If conditions demand it,one aspect may be prioritized at the expense of others, re-sulting in the life history variation and its outcomes in featuressuch as growth rates, adult body size, fertility, and possiblylifespan that are observed in modern humans (Bribiescas,Ellison, and Gray 2012; Kuzawa and Bragg 2012; Miglianoand Guillon 2012). The symposium provided a vehicle forbringing together the disparate data sets of our subdisciplinesinto a framework that suggests ways in which variation isproduced, organized, and interrelated in the extant world.

What We Know Now and What We Hope toKnow in the Future

The final paper of the volume takes up the challenge of usingthis framework to generate means of assessing the variationobserved in the fossil record and the biocultural relationshipbetween hominin morphology, hominin behavior, and thefluctuating environment of the African Pliocene and Pleis-tocene (Anton and Snodgrass 2012). What we know is thatearly Homo existed in a highly variable environment (a non-equilibrium ecosystem) that may have placed adaptive plas-ticity at a premium. The type of body size variation observedin early Homo is consistent with the range observed in modernhumans that is mediated by life history differences in growthand development that are dependent on energy availabilityand mortality probability (Kuzawa and Bragg 2012; Miglianoand Guillon 2012; Migliano, Vinicius, and Lahr 2007). Dietarydifferences, involving increased dietary breadth and a morecarnivorous diet, are also evident between these hominins andare consistent with greater adaptive plasticity than inferredfor the australopiths. Comparative studies suggest that larger-bodied hominins would have had to adopt a high energythroughput strategy, and this, together with the increase inbrain size, would presuppose increased cooperation in theform of allocare and sexual division of labor. The increasedcooperation and sociality would also be significant in groupprotection in a relatively dangerous terrestrial environmentand create a relatively safe “niche” that would be consistentwith later maturation (and perhaps increased longevity) thatbegins to be evident in the dental maturation evidence forHomo erectus in relation to the other early hominins.

These results provide the basis for a model for the evolutionof Homo involving an integrated feedback loop that drove lifehistory evolution and contributed to cultural change (Antonand Snodgrass 2012). The central elements of this model are


cooperative behavior, diet, cognitive abilities, and extrinsicmortality risk. This model also generates a number of testablehypotheses (e.g., to explain size variation in the hominins),but each requires additional data from both the fossil and themodern records to test. Increased body size across genera maybe hypothesized to signal either decreased extrinsic mortality,increased nutritional sufficiency, or both. On the one hand,specific predictions are made about rates of growth, timingof weaning, and timing of growth cessation in each scenario—so additional growth data, especially from dental developmentand especially for the earliest Homo, are needed. But beyondsimply “finding more fossils,” additional means of assessing“hard tissue” growth in extant populations in ways that arecomparable to fossil samples are needed—and the need toconsider intrapopulational variation among extant primateand nonprimate mammals to the level that it is done in humanpopulations is also required. There are issues of scale andcomparability in our data sets that can only be remedied bylong-term analyses of extant populations.

The increase in brain and body size between other earlyHomo and H. erectus again suggests decreased extrinsic mor-tality and/or increased nutrition. These features may reflectadaptations to the higher mortality rates in terrestrial envi-ronments and perhaps cooperative hunting—as shown in thesocial carnivores. The extended developmental schedule of H.erectus is consistent with mortality reduction, possibly as theresult of behavioral changes involving some form of coop-eration. Here, along with additional fossil data and more com-parable data sets, a key endeavor will be mining the rich socialcarnivore data sets and creating new ones that consider aspectsof morphology, behavior, and variation. Regarding extrinsicmortality, new ways can and should be developed to inter-rogate the archaeological and extant records for predator loadand the degree to which this can be assessed for differenthominin species. Novel means should also be developed forassessing the ways in which population variation and physicalcharacter development (including size and dimorphism) areinfluenced by the high degree of climate variability that char-acterized this period of evolutionary history.

For example, it has been hypothesized for the small-sizedDmanisi sample that nutritional insufficiency and perhapsisolation resulted in short-term accommodations or adapta-tions (e.g., Anton 2003, 2013; see also Migliano and Guillon2012). If this is the case, one would expect lower growth ratesand a longer period of maturation in relation to larger-bodiedH. erectus. Alternatively, if the short stature was due to a highmortality environment, one would expect the smaller speci-mens to have more rapid growth rates and a shorter periodof maturation. Clearly this requires greater detail than wecurrently possess regarding H. erectus growth, but it providesa place to start.

While we end, as many such symposia do, with a decidedplea for more fossil remains in different localities, we hopeto move beyond that to an agenda of integrated and multi-disciplinary studies to provide a framework against which to

test these predictions. For example, to understand the factorssurrounding the evolution of Homo, it will be essential totarget intra- and interpopulational research on energetic andlife history variation in various climatic, nutritional, and mor-tality environments. We need to tease out the complex in-terrelationships among these and other variables to under-stand the fundamental correlates of body size, brain size, andsexual dimorphism (Kuzawa and Bragg 2012; Smith et al.2012). Among many other things, we want to know, for ex-ample, how skeletal dimorphism tracks body mass dimor-phism across populations, especially in humans (Plavcan2012), and how individual skeletal features are influenced inmales and females in differing circumstances (Bribiescas, El-lison, and Gray 2012; Kuzawa and Bragg 2012).

Our thinking about the origins of Homo has continued tochange since Homo habilis was announced (Leakey, Tobias,and Napier 1964), the almost complete Nariokotome H.erectus skeleton was discovered (Brown et al. 1985), andDmanisi and other material changed our ideas about variationin H. erectus (Gabunia et al. 2000; Potts et al. 2004; Simpsonet al. 2008; Spoor et al. 2007). New fossils will undoubtedlycontinue to be uncovered. However, this material cannot beinterpreted in a vacuum, and the more we know about intra-and interspecific variation in modern humans and other an-imals, the stronger the foundation we have for a rich under-standing of our evolutionary past.

Acknowledgments

We would like to thank the Wenner-Gren Foundation for theopportunity to hold this symposium and to publish the resultsas an open-access supplementary issue of Current Anthro-pology. We would also like to thank all of the participants forlively and stimulating discussion and debate over a 6-dayperiod in March 2011 at the Tivoli Palacio de Seteais Hotelin Sintra, Portugal. This experience will be fondly remem-bered for a long time to come. We would like to give specialthanks to Chris Rainwater (New York University), who servedas the rapporteur for the meeting; to Emily Middletown (NewYork University), who provided invaluable assistance in help-ing to edit and prepare all of the manuscripts for publication;and to Lisa McKamy and the editorial and production staffat the University of Chicago Press for their help in bringingthis issue to fruition. The meeting would not have been assuccessful as it was without the deft organizational skills ofLaurie Obbink, the Wenner-Gren Foundation Conference As-sociate, and for this we are most grateful.

References CitedAiello, Leslie C., and Catherine Key. 2002. Energetic consequences of being

a Homo erectus female. American Journal of Human Biology 14:551–565.Aiello, Leslie C., and Jonathan C. K. Wells. 2002. Energetics and the evolution

of the genus Homo. Annual Review of Anthropology 31:323–338.Aiello, Leslie C., and Peter Wheeler. 1995. The expensive-tissue hypothesis:


the brain and the digestive system in human and primate evolution. CurrentAnthropology 36:199–221.

Anton, Susan C. 2003. Natural history of Homo erectus. Yearbook of PhysicalAnthropology 46:126–170.

———. 2008. Framing the question: diet and evolution in early Homo. InPrimate craniofacial function and biology: papers in honor of Bill Hylander.Christopher J. Vinyard, Christine E. Wall, and Matthew J. Ravosa, eds. Pp.443–482. New York: Springer.

———. 2012. Early Homo: who, when, and where. Current Anthropology53(suppl. 6):S278–S298.

———. 2013. Homo erectus and related taxa. In Companion to paleoanthro-pology. David Begun, ed. Hoboken, NJ: Wiley-Blackwell.

Anton, Susan C., William R. Leonard, and Marcia Robertson. 2002. An eco-morphological model of the initial hominid dispersal from Africa. Journalof Human Evolution 43:773–785.

Anton, Susan C., and J. Josh Snodgrass. 2012. Origins and evolution of genusHomo: new perspectives. Current Anthropology 53(suppl. 6):S479–S496.

Asfaw, Berhane, Tim White, Owen Lovejoy, Bruce Latimer, Scott Simpson,and Gen Suwa. 1999. Australopithecus garhi: a new species of early hominidfrom Ethiopia. Science 284:629–635.

Berger, Lee R., Darryl de Ruiter, Steven E. Churchill, Peter Schmid, KristianJ. Carlson, Paul H. G. M. Dirks, and Job M. Kibii. 2010. Australopithecussediba: a new species of Homo-like australopith from South Africa. Science328:195–204.

Blumenschine, Robert J., Charles C. Peters, Fidelis T. Masao, Ron J. Clarke,Alan L. Deino, Richard L. Hay, Carl C. Swisher, et al. 2003. Late PlioceneHomo and hominid land use from western Olduvai Gorge, Tanzania. Science299:1217–1221.

Braun, David R., Thomas Plummer, Peter Ditchfield, Joseph V. Ferraro, DavidMaina, Laura C. Bishop, and Richard Potts. 2008. Oldowan behavior andraw material transport: perspectives from the Kanjera Formation. Journalof Archaeological Science 35:2329–2345.

Bribiescas, Richard G., Peter T. Ellison, and Peter B. Gray. 2012. Male lifehistory, reproductive effort, and the evolution of the genus Homo. CurrentAnthropology 53(suppl. 6):S424–S435.

Brown, Frank, John Harris, Richard Leakey, and Alan Walker. 1985. EarlyHomo erectus skeleton from west Lake Turkana, Kenya. Nature 316:788–792.

Fuentes, Agustın, Matthew A. Wyczalkowski, and Katherine C. MacKinnon.2010. Niche construction through cooperation: a nonlinear dynamics con-tribution to modeling facets of the evolutionary history in the genus Homo.Current Anthropology 51:435–444.

Gabunia, Leo, Abesalom Vekua, David Lordkipanidze, Carl C. Swisher, ReidFerring, Antje Justus, Medea Nioradze, et al. 2000. Earliest Pleistocenecranial remains from Dmanisi, Republic of Georgia: taxonomy, geologicalsetting, and age. Science 288:1019–1025.

Haile-Selassie, Yohannes, Bruce M. Latimer, Mulugeta Alene, Alan L. Deino,Luis Gibert, Stephanie M. Melillo, Beverly Z. Saylor, Gary R. Scott, and C.Owen Lovejoy. 2010. An early Australopithecus afarensis postcranium fromWoranso-Mille, Ethiopia. Proceedings of the National Academy of Sciences ofthe USA 107:12121–12126.

Hawkes, Kristen, Peter S. Kim, Brett Kennedy, Ryan Bohlender, and JohnHawks. 2011. A reappraisal of grandmothering and natural selection. Pro-ceedings of the Royal Society B 278:1936–1938.

Hawkes, Kristen, James F. O’Connell, Nicholas G. Blurton Jones, Helen Al-varez, and Eric L. Charnov. 1998. Grandmothering, menopause, and theevolution of human life histories. Proceedings of the National Academy ofSciences of the USA 95:1336–1339.

Holliday, Trenton W. 2012. Body size, body shape, and the circumscriptionof the genus Homo. Current Anthropology 53(suppl. 6):S330–S345.

Holloway, Ralph L., Douglas C. Broadfield, and Michael S. Yuan. 2004. Thehuman fossil record: brain endocasts: the paleoneurological evidence. New York:Wiley.

Isler, Karin, and Carel P. van Schaik. 2009. The expensive brain: a frameworkfor explaining evolutionary changes in brain size. Journal of Human Evo-lution 57:392–400.

———. 2012. How our ancestors broke through the gray ceiling: comparativeevidence for cooperative breeding in early Homo. Current Anthropology53(suppl. 6):S453–S465.

Kachel, A. Friederike, Luke S. Premo, and Jean-Jacques Hublin. 2011a. Grand-mothering and natural selection. Proceedings of the Royal Society B 278:384–391.

———. 2011b. Invited reply: grandmothering and natural selection revisited.Proceedings of the Royal Society B 278:1939–1941.

Kaplan, Hillard, Kim Hill, Jane Lancaster, and Magdalena A. Hurtado. 2000.A theory of human life history evolution: diet, intelligence, and longevity.Evolutionary Anthropology 2000:156–185.

Kaplan, Hillard, Jane Lancaster, and Arthur Robson. 2003. Embodied capitaland the evolutionary economics of the human life span. In Life span: evo-lutionary, ecological, and demographic perspectives. James R. Carey and Shri-pad Tuljapurkar, eds. Pp. 152–182. New York: Population Council.

Kibii, Job M., Steven E. Churchill, Peter Schmid, Kristian J. Carlson, NichelleD. Reed, Darryl J. de Ruiter, and Lee R. Berger. 2011. A partial pelvis ofAustralopithecus sediba. Science 333:1407–1411.

Kimbel, William H., Robert C. Walter, Donald C. Johanson, Kaye E. Reed,James L. Aronson, Zelalem Assefa, Curtis W. Marean, et al. 1996. LatePliocene Homo and Oldowan tools from the Hadar Formation (Kada HadarMember), Ethiopia. Journal of Human Evolution 31:549–561.

Kivell, Tracy L., Job M. Kibii, Steven E. Churchill, Peter Schmid, and Lee R.Berger. 2011. Australopithecus sediba hand demonstrates mosaic evolutionof locomotor and manipulative abilities. Science 333:1411–1417.

Knott, Cheryl D. 1998. Changes in orangutan caloric intake, energy balance,and ketones in response to fluctuating fruit availability. International Journalof Primatology 19:1061–1079.

Kuzawa, Christopher W. 1998. Adipose tissue in human infancy and child-hood: an evolutionary perspective. Yearbook of Physical Anthropology 41:177–209.

Kuzawa, Christopher W., and Jared M. Bragg. 2012. Plasticity in human lifehistory strategy: implications for contemporary human variation and theevolution of genus Homo. Current Anthropology 53(suppl. 6):S369–S382.

Leakey, Louis S., Philip V. Tobias, and John R. Napier. 1964. A new speciesof the genus Homo from Olduvai Gorge. Nature 202:7–9.

Leakey, Meave G., Fred Spoor, M. Christopher Dean, Craig S. Feibel, SusanC. Anton, Christopher Kiarie, and Louise N. Leakey. 2012. New fossils fromKoobi Fora in northern Kenya confirm taxonomic diversity in early Homo.Nature 488:201–204, doi:10.1038/nature11322.

Leigh, Steven R. 1992. Patterns of variation in the ontogeny of primate bodysize dimorphism. Journal of Human Evolution 23:27–50.

Leonard, William R., and Marcia L. Robertson. 1997. Comparative primateenergetics and hominid evolution. American Journal of Physical Anthropology102:265–281.

McHenry, Henry M., and Katherine Coffing. 2000. Australopithecus to Homo:transformations in body and mind. Annual Review of Anthropology 29:125–146.

McPherron, Shannon P., Zeresenay Alemseged, Curtis W. Marean, JonathanG. Wynn, Denne Reed, Denis Geraads, Rene Bobe, and Hamdallah A.Bearat. 2010. Evidence for stone-tool-assisted consumption of animal tissuesbefore 3.39 million years ago at Dikika, Ethiopia. Nature 466:857–860.

Migliano, Andrea Bamberg, and Myrtille Guillon. 2012. The effects of mor-tality, subsistence, and ecology on human adult height and implications forHomo evolution. Current Anthropology 53(suppl. 6):S359–S368.

Migliano, Andrea B., Lucio Vinicius, and Marta Mirazon Lahr. 2007. Lifehistory trade-offs explain the evolution of human pygmies. Proceedings ofthe National Academy of Sciences of the USA 104:20216–20219.

Navarrete, Ana, Carel P. van Schaik, and Karin Isler. 2011. Energetics and theevolution of human brain size. Nature 480:91–93.

O’Connell, James F., Kristin Hawkes, and Nicolas B. Blurton Jones. 1999.Grandmothering and the evolution of Homo erectus. Journal of HumanEvolution 36:461–485.

Odling-Smee, F. John, Kevin N. Laland, and Marcus W. Feldman. 2003. Nicheconstruction: the neglected process in evolution. Monographs in PopulationBiology 37. Princeton, NJ: Princeton University Press.

Pfeiffer, Susan. 2012. Conditions for evolution of small adult body size insouthern Africa. Current Anthropology 53(suppl. 6):S383–S394.

Plavcan, J. Michael. 2001. Sexual dimorphism in primate evolution. AmericanJournal of Physical Anthropology 116:25–53.

———. 2012. Body size, size variation, and sexual size dimorphism. CurrentAnthropology 53(suppl. 6):S409–S423.

Plavcan, J. Michael, and Carel P. van Schaik. 1997a. Interpreting hominidbehavior on the basis of sexual dimorphism. Journal of Human Evolution32:345–374.

———. 1997b. Intrasexual competition and body weight dimorphism in an-thropoid primates. American Journal of Physical Anthropology 103:37–68.

Pontzer, Herman. 2012. Ecological energetics in early Homo. Current An-thropology 53(suppl. 6):S346–S358.


Potts, Richard. 2012. Environmental and behavioral evidence pertaining tothe evolution of early Homo. Current Anthropology 53(suppl. 6):S299–S317.

Potts, Richard, Anna K. Behrensmeyer, Alan Deino, Peter Ditchfield, andJennifer Clark. 2004. Small mid-Pleistocene hominin associated with EastAfrican Acheulean technology. Science 305:75–78.

Ruff, Christopher B. 1993. Climatic adaptation and hominid evolution: thethermoregulatory imperative. Evolutionary Anthropology 2:53–60.

———. 2010. Body size and body shape in early Homo: implications of theGona pelvis. American Journal of Physical Anthropology 58:166–178.

Schaller, George B., and Gordon R. Lowther. 1969. Relevance of carnivorebehavior to study of early hominids. Southwestern Journal of Anthropology25:307–341.

Schwartz, Gary T. 2012. Growth, development, and life history throughoutthe evolution of Homo. Current Anthropology 53(suppl. 6):S395–S408.

Semaw, Sileshi, Michael J. Rogers, Jay Quade, Paul R. Renne, Robert F. Butler,Manuel Dominguez-Rodrigo, Dietrich Stout, William S. Hart, Travis Pick-ering, and Scott W. Simpson. 2003. 2.6-million-year-old stone tools andassociated bones from OGS-6 and OGS-7, Gona, Afar, Ethiopia. Journal ofHuman Evolution 45:169–177.

Shipman, Pat, and Alan Walker. 1989. The costs of becoming a predator.Journal of Human Evolution 18:373–392.

Simpson, Scott W., Jay Quade, Naomi E. Levin, Robert Butler, GuillaumeDupont-Nivet, Melanie Everett, and Sileshi Semaw. 2008. A female Homoerectus pelvis from Gona, Ethiopia. Science 322:1089–1092.

Smith, Jennifer E., Eli M. Swanson, Daphna Reed, and Kay E. Holekamp.

2012. Evolution of cooperation among mammalian carnivores and its rel-evance to hominid evolution. Current Anthropology 53(suppl. 6):S436–S452.

Spoor, Fred, Meave G. Leakey, Patrick N. Gathogo, Frank H. Brown, SusanC. Anton, Ian McDougall, Christopher Kiarie, F. Kyalo Manthi, and LouiseN. Leakey. 2007. Implications of new early Homo fossils from Ileret, eastof Lake Turkana, Kenya. Nature 448:688–691.

Ungar, Peter S. 2012. Dental evidence for the reconstruction of diet in Africanearly Homo. Current Anthropology 53(suppl. 6):S318–S329.

Walker, Alan, and Richard Leakey, eds. 1993. The Nariokotome Homo erectusskeleton. Cambridge, MA: Harvard University Press.

Wells, Jonathan C. K. 2009. The evolutionary biology of human body fat: thriftand control. Cambridge: Cambridge University Press.

———. 2012. The capital economy in hominin evolution: how adipose tissueand social relationships confer phenotypic flexibility and resilience in sto-chastic environments. Current Anthropology 53(suppl. 6):S466–S478.

Wood, Bernard, and Jennifer Baker. 2011. Evolution in the genus Homo.Annual Review of Ecology, Evolution, and Systematics 42:47–69.

Wood, Bernard A., and Mark C. Collard. 1999. The human genus. Science284:65–71.

———. 2007. Defining the genus Homo. In Handbook of paleoanthropology.Winfried Henke, Helen Rothe, and Ian Tattersall, eds. Pp. 1575–1610. Ber-lin: Springer.

Zipfel, Bernhard, Jeremy M. DeSilva, Robert S. Kidd, Kristian J. Carlson,Steven E. Churchill, and Lee R. Berger. 2011. The foot and ankle of Aus-tralopithecus sediba. Science 333:1417–1420.



Early HomoWho, When, and Where

by Susan C. Anton

The origin of Homo is argued to entail niche differentiation resulting from increasing terrestriality and dietarybreadth relative to the better known species of Australopithecus (A. afarensis, A. anamensis, A. africanus). I reviewthe fossil evidence from ∼2.5 to 1.5 Ma in light of new finds and analyses that challenge previous inferences.Minimally, three cranial morphs of early Homo (including Homo erectus) exist in eastern Africa (1.9–1.4 Ma), withat least two in southern Africa. Because of taphonomic damage to the type specimen of Homo habilis, in East Africatwo species with different masticatory adaptations are better identified by their main specimen (i.e., the 1813 groupand the 1470 group) rather than a species name. Until recently, the 1470 group comprised a single specimen. SouthAfrican early Homo are likely distinct from these groups. Together, contemporary early H. erectus and early Homoare bigger than Australopithecus (∼30%). Early H. erectus (including recently discovered small specimens) is largerthan non-erectus Homo (∼15%–25%), but their size ranges overlap. All early Homo are likely to exhibit substantialsexual dimorphism. Early H. erectus is less “modern” and its regional variation in size more substantial than previouslyallowed. These findings form the baseline for understanding the origin of the genus.

The origins of the genus Homo and the factors that may haveled to its appearance remain murky. In the past two decades,the idea of increased behavioral flexibility in our early fore-bears (Potts 1988) and increased diet quality and ranging(Anton, Leonard, and Robertson 2002; Shipman and Walker1989) have become cornerstones of how we understand theorigin and evolution of Homo before we left Africa. Theseideas in turn have emphasized the importance of enlargingbrain and body size, decreases in sexual dimorphism, some-what expanded ontogenetic periods, increases in energetic re-quirements, and increased cooperation during the first millionyears or so of our history (Aiello and Key 2002; Dean andSmith 2009; Dean et al. 2001). Yet more recent fossil findscall into question some of these trends. New fossils in EastAfrica, Georgia, and Indonesia suggest large ranges of size andperhaps shape variation in Homo erectus sensu lato and hintat local adaptation and short-term accommodation as an im-portant yet underappreciated contributor to the morpholog-ical picture seen in the fossil record (Anton et al. 2007; Pottset al. 2004; Rightmire, Lordkipanidze, and Vekua 2006; Simp-son et al. 2008; Spoor et al. 2007). Although the record ispatchier, there is also considerable variation among earlierHomo, and arguably several species are represented (Curnoe2010; Leakey, Tobias, and Napier 1964; Leakey et al. 2012;

Susan C. Anton is Professor of Anthropology at the Center for theStudy of Human Origins, Department of Anthropology, New YorkUniversity (25 Waverly Place, New York, New York 10003, U.S.A.[[email protected]]). This paper was submitted 12 XII 11,accepted 9 VII 12, and electronically published 3 XII 12.

Schrenk, Kullmer, and Bromage 2007; Stringer 1986; Wood1991) even if there is little agreement as to their composition.

Any understanding of the origin and evolution of Homo mustbuild from the primary data of the fossil record. To frame whatcomes later in this special issue, here I discuss how the genusmay be defined relative to other hominin genera. And withinour genus I consider the morphology, location, and age of theindividual representatives of early Homo up to and includingH. erectus. The taxa of interest here are those commonly referredto as Homo habilis sensu lato and H. erectus sensu lato (table1; fig. 1). The former group is often split into multiple taxa,usually H. habilis and Homo rudolfensis, but others, such asHomo microcranous (for KNM-ER 1813) and Homo gauten-gensis (for Stw 53) have also been suggested. The H. erectusgroup is also sometimes split into Homo ergaster for the earlyAfrican and Georgian material and H. erectus for the Asian,although a consensus seems to be building for recognizing justone species, H. erectus (Anton 2003; Baab 2008; Rightmire1990). Given this flux in species composition, I advocate thatparticular attention must be paid to individual fossil data andranges of variation in size and shape in order to build an explicitpicture of what we can and cannot know about fossil taxa andhow what we think we know changes depending on the in-cluded data points of a taxon.

Recognizing Early Homo

A biological genus comprises closely related species, and al-though the protocol for doing this is not strictly codified in


Anton Early Homo: Who, When, and Where S279

Table 1. Fossils attributed as type specimens to named species of early Homo

Umbrella species andtype specimen Species name Publication

Countryof type

Homo habilis sensu lato:OH 7 Homo habilis Leakey, Tobias, and Napier 1964 TanzaniaKNM-ER 1470 Homo rudolfensis Alexeev 1986 KenyaKNM-ER 1813 Homo microcranous Ferguson 1995 KenyaStw 53 Homo gautengensis Curnoe 2010 South Africa

Homo erectus sensu lato:Trinil 2 (Pithecanthropus) erectus Dubois 1894 IndonesiaZhoukoudian 1 Sinanthropus pekinensis Black 1927 ChinaNgandong 1 Homo soloensis Openoorth 1932 IndonesiaPerning 1 Homo modjokertensis Von Koenigswald 1936 IndonesiaSwartkrans 15 Telanthropus capensis Broom and Robinson 1949 South AfricaTernifine 1 Atlanthropus mauritanicus Arambourg 1954 AlgeriaOH 9 Homo leakeyi Heberer 1963 TanzaniaKNM-ER 992 Homo ergaster Groves and Mazek 1975 KenyaDmanisi 2600 Homo georgicus Gabunia et al. 2002 Georgia

Note. Species listed by umbrella taxon in chronological order of publication. The most commonly used species appear in bold.

zoological nomenclature, it has been argued that a genus“should be defined as a species, or monophylum, whose mem-bers occupy a single adaptive zone” (Wood and Collard 1999:66). Such a definition combines the cladistic requirement ofmonophyly for genera with a means of deciding (the adaptivezone) where to recognize the base of the genus. Such a def-inition is thus both prudent and pragmatic if somewhat prob-lematic to apply. The main problem is how to assess the“adaptive zone,” especially in light of the probability that thefull suite of characters associated with such a zone is likelyto have evolved in a mosaic fashion rather than appearingfull blown at the base of the genus.

Several morphological differences distinguish fossil mem-bers of the genus Homo from those of Australopithecus andParanthropus,1 including reduction in tooth and jaw size, re-organization of craniofacial morphology, and perhaps changesin body shape and size (Kimbel 2009; Rightmire and Lord-kipanidze 2009; Wood 2009). And these physical differenceshave been taken to suggest underlying adaptive shifts at theorigin of the genus Homo, most or all of which have energeticand life history implications (e.g., McHenry and Coffing2000). Thus, the adaptive zone of Homo has been variouslydefined, implicitly or explicitly, to relate either to cranial ex-pansion and masticatory diminution (e.g., Kimbel 2009; Kim-bel, Johanson, and Rak 1997; Kimbel et al. 1996; Leakey,Tobias, and Napier 1964) and/or to increased locomotor ef-ficiency and ranging (Wood 2009; Wood and Collard 1999)relative to Australopithecus.

Using the above definition of a genus, Wood and Collard(1999; see also Collard and Wood 2007; Wood 2009; Wood

1. While it is recognized that Australopithecus may be paraphyletic,for the purposes of the comparisons in this paper, the genus is consideredto exclude Paranthropus species but to include the best represented speciescommonly assigned to Australopithecus, that is, A. anamensis, A. afarensis,A. garhi, A. africanus, and A. sediba. When the data for specific com-parisons come from a single species, that species is indicated by name.

and Baker 2011) proposed to remove two species, Homo ha-bilis and Homo rudolfensis, from the genus and place theminto Australopithecus. Their criteria for distinguishing speciesof Homo from those of Australopithecus were based on findingsix classes of characteristics that were more similar to thecondition in Homo sapiens than to that of Australopithecusafricanus (the type species of each genus). The first criterionis monophyly. The last is an extended period of growth anddevelopment. The remaining four criteria are more explicitlyrelated to the adaptive zone; three to reconstructions of bodymass, shape, and proportions and one to jaw and tooth pro-portions as scaled to body-size-adjusted brain size. Three ofthe six are not assessable in H. rudolfensis. And those that areassessable in H. rudolfensis or H. habilis are contested (seeHolliday 2012). That aside, judging inclusion in a genus basedon the association with its most derived member would seemto preclude the possibility of mosaic evolution in its earliermembers. Thus, the specific ways in which H. rudolfensis andH. habilis are like other members of the genus, such as relativebrain size, are trumped by their dissimilarity to H. sapiens.

While monophyly must be maintained and identifyingadaptive zones will always be somewhat subjective, I favorrecognizing and fully weighting the incipient characters of theadaptive zone; Wood and Baker (2011) refer to this as a“bottom-up” approach. Genus Homo is recognized, then, onthe basis of the following mostly derived craniodental char-acters relative to Australopithecus as have been outlined andmore fully described by others.

1. Cranial expansion. Size-adjusted capacity relative to orbitsize is above 2.7 (Collard and Wood 2007; Wood and Collard1999). In addition or independently, there may be evidenceof endocranial expansion or asymmetry relative to Austral-opithecus (LeGros Clark 1964; Rightmire and Lordkipanidze2009).

2. Shape of the face and palate. The palate is deep and broad.The anterior maxillary profile, as seen from above, is round

Figure 1. Temporal and geographic distribution of early Homo and early Homo erectus localities and some important specimensdiscussed in the text. On the far left is the geomagnetic polarity timescale, with normal periods in black and reversed in white.Radiometric time is indicated in millions of years on the far right. Within regional columns, solid lines on either side of site namesindicate time spans suggested by multiple individuals from a site; dashed-and-dotted lines indicate possible time range around asingle or a few specimens. In the Africa columns, sites are grouped from left to right as South Africa, Malawi, Kenya, Tanzania,and Ethiopia; A.L. p Afar Locality; OG p Olorgesailie; OH p Olduvai Hominid; numbers on the Koobi Fora lines p KNM-ERnumbers; WT p West Turkana specimen numbers; Stw p Sterkfontein. The recently described specimens of the 1470 group ofearly Homo are KNM-ER 60000, 62000, and 62003 (Leakey et al. 2012).


to square (not triangular). The subnasal prognathism is mild,the nasoalveolar clivus is sharply angled to the nasal floor,and the nasal margin is everted (Kimbel 2009; Kimbel, Jo-hanson, and Rak 1997; Kimbel et al. 1996).

3. Size and shape of the dentition. The canine crown issymmetrical. Premolars lack substantial molarization incrown and root form (buccolingually narrow) but are notsectorial. The molars, especially the first, are somewhat me-siodistally elongated but may retain a large crown-base area.M2 is “rhomboidal” in shape (dominated by mesial cusps;Bromage, Schrenk, and Zonneveld 1995; Johanson et al. 1987;Kimbel 2009; Kimbel, Johanson, and Rak 1997; Kimbel et al.1996; Wood 1991).

Postcranial differences are not used here to distinguishHomo and Australopithecus because few postcranial remainsare certainly associated with species-diagnostic cranial re-mains of early Homo. Additionally, those that are do notsupport a major locomotor difference between H. habilissensu lato on the one hand and H. erectus (p ergaster) onthe other (see Holliday 2012 contra Wood and Collard 1999).Using this standard, H. habilis and H. rudolfensis are recog-nized as Homo because both differ in distinct craniodentalways from Australopithecus (see below). In these differences,H. habilis and H. rudolfensis trend toward the condition inlater members of genus Homo.

Homo before 2 Ma

A small number of fossil remains of older than 2 Ma in EastAfrica satisfy the above criteria; however, none of these canbe confidently attributed to species (fig. 1). The oldest fossilHomo is likely to be the A.L. 666 maxilla from Hadar, Ethiopia,which is minimally 2.33 Ma (Kimbel et al. 1996). This spec-imen differs from Australopithecus in the anatomical suites ofcharacters mentioned in items 2 and 3 above. A similarly agedisolated molar from West Turkana, Kenya, is also likely earlyHomo (Prat et al. 2005). Both have some affinities with laterearly Homo from Kenya, such as KNM-ER 1813. Isolateddental remains from the Omo, Ethiopia (2–2.4 Ma), also likelyrepresent early members of Homo and bear some similaritiesto later teeth of KNM-ER 1802 from Kenya (Suwa, White,and Howell 1996). The Uraha-501 (UR-501) mandible fromMalawi has been argued to be of similar or slightly older age;its age, based on faunal correlations, may be as young as 1.9Ma or as old as 2.5 Ma (Bromage, Schrenk, and Zonneveld1995; Kimbel 2009). The inclusion of UR-501 in Homo isbased on both molar and premolar morphology (item 3) andmandibular anatomy. This specimen has strong morpholog-ical affinities of the symphysis and corpus with the KNM-ER1802 mandible (Schrenk, Kullmer, and Bromage 2007). Thesesimilarities suggest the two are likely to belong to the samespecies of Homo, whatever that may be.

In South Africa, fewer fossils from the period 2.6 to 2.0Ma or older have been suggested to represent either earlyHomo or a species of Australopithecus derived in the direction

of Homo (Curnoe 2010; Dean and Wood 1982; Kimbel andRak 1993; Moggi-Cecchi, Tobias, and Beynon 1998). Theseinclude the cranial base, Sts 19, and juvenile cranial fragments,Stw 151, from Sterkfontein. However, both are commonlyattributed to Australopithecus africanus (Dean and Wood1982; Spoor 1993).

It is unclear which species is directly ancestral to Homo(Kimbel 2009). However, the origin of the lineage is likely tobe at 2.5 Ma or earlier given that by 2.3 Ma there is incipientevidence of two dental morphs. Based on the current record,the earliest accepted Homo appear to be in the northern partof eastern Africa; however, this does not preclude an ancestorfrom another part of the continent.

Non-erectus Early Homo (2.0–1.44 Ma)

Members of “non-erectus” Homo are better represented after2 Ma and range in age from about 2.0 to 1.44 Ma (Feibel,Brown, and McDougall 1989; Spoor et al. 2007). They arebest known from Kenya and Tanzania, although at least oneSouth African morph also appears to be present (Curnoe2010; Grine 2005; Grine et al. 2009; Hughes and Tobias 1977;Moggi-Cecchi, Grine, and Tobias 2006).

Early Homo: Taxa, Individuals, and Anatomy

Homo habilis (Leakey, Tobias, and Napier 1964) was the firstnon-erectus species of early Homo to be recognized, and formany scholars it represents a single species that is largerbrained and smaller toothed than Australopithecus yet smallerbrained and slightly larger toothed than early African Homoerectus. Apart from size, the teeth differ in shape from Aus-tralopithecus and Paranthropus in the ways discussed in thedefinition above, among others. For a time, all other early,relatively small-brained African Homo specimens werelumped into this species (e.g., Boaz and Howell 1977; Hughesand Tobias 1977; Johanson et al. 1987; Leakey, Clarke, andLeakey 1971; Wood 1991).

The discussion in the mid-1980s regarding whether the H.habilis hypodigm varied too much to constitute a single spe-cies is well known (Lieberman, Pilbeam, and Wood 1988;Stringer 1986; Wood 1985). The debate has focused heavilyon issues of brain size, with the most complete of the KoobiFora specimens, KNM-ER 1470 and 1813, representing theextremes of cranial capacity (750 vs. 510 cm3). For those whoinclude these fossils in a single species, the size and shapedifferences between the two are explained as sexual dimor-phism (Howell 1978; Tobias 1991); however, the very differentfacial structures are more problematic than is absolute sizefor the inclusion of both specimens in a single species (Leakeyet al. 2012; Wood 1991, 1992).

Subsequently, these fossils have been used as type specimensfor other species. In 1986, KNM-ER 1470 became the typeof Australopithecus rudolfensis (Alexeev 1986); Wood (1991,1992) provided substantial anatomical reasoning to support

Table 2. Elements commonly associated with KNM-ER 1470 as Homo rudolfensis

Specimens Element Main attributes and reasons for previous association with KNM-ER 1470a

Type specimen:KNM-ER 1470 Cranium � partial tooth roots Large cranial vault (750 cm3); flat face with anterior placement of zygomatic takeoff (mesial end M1);

P3 with three roots “incompletely divided double root”; P4 probably double rooted (Wood 1991:74).

Craniodental specimens:KNM-ER 1590 Upper dentition and cranial fragments Large vault size and shape and large tooth crown size (Wood 1991:251). However, root form cannot be

compared as these are yet to form in this subadult, and crown form is not preserved in 1470. Thisspecimen formed the basis for arguments of large crown size in H. rudolfensis (see Leakey et al. 2012).

KNM-ER 3732 Partial calotte and left zygoma Large vault size, size and robusticity of the frontal bone’s contribution to the face, and the anteriorinclination of the malar region (Wood 1991:251). However, no teeth are present and supraorbital andorbital region differs from 1470. Zygomatic and palatal portions are incompletely preserved. Does notadd to our knowledge of crown morphology.

KNM-ER 3891 Cranial fragments including maxilla fragments Anterior takeoff of the zygomatic at distal P4; three-rooted P3 and P4; large entoglenoid (althoughconstruction differs from that of 1470; Wood 1991:135, 251). This fragmentary specimen may affili-ate with 1470 but does not add to our knowledge of crown morphology or cranial size.

Mandibulodental specimens:b

KNM-ER 819 Mandible fragment Similar in size and shape to 1802, sharing everted base and P roots. Superficial resemblance to Paran-thropus boisei, but lacks extreme specializations (Wood 1991:251). Ties to KNM-ER 1802 but notnecessarily 1470 (see Leakey et al. 2012).

KNM-ER 1482 Mandible Similar in size and shape to 1802. Similar superficial association with P. boisei as above (Wood 1991:251). Ties to 1470 (see Leakey et al. 2012).

KNM-ER 1483 Mandible fragment Large corpus size, and symphysis align with mandible 1802, but premolar roots are simpler (Wood1991:251). Ties to KNM-ER 1802 less strong for corpus shape and premolar form.

KNM-ER 1801 Mandible fragment Similarities to KNM-ER 1802 in corpus, symphysis, P roots and M3 (Wood 1991). Homo � Australo-pithecus boisei-like features affine it with 1802 (Wood 1991:189). Ties to 1470 (Leakey et al. 2012).

KNM-ER 1802 Mandible Mandibular robusticity and dental size and root complexity consistent with the inference that 1470was adapted to a heavy masticatory pattern (Wood 1991:251). No direct tie to KNM-ER 1470; but,as noted by Wood, Stringer argued, “The anterior placement of the root of the ascending ramus andextramolar sulcus may be consistent with KNM-ER 1470 zygoma position” (Stringer 1986). How-ever, see Leakey et al. (2012).

UR-501 Mandible Mandibular morphology similar to KNM-ER 1802 with especial reference to the broad Ps and plate-like P roots (Bromage, Schrenk, and Zonneveld 1995). Strong ties to KNM-ER 1802 but not neces-sarily 1470.

Postcranial specimens:KNM-ER 813 Talus More similar to human than australopithecine tali (Wood 1992). Tie to KNM-ER 1470 is based on

presumed size. Could be other early Homo.KNM-ER 1472 Femur Larger size and more humanlike anatomy than femora from Olduvai (Wood 1992). Tie to KNM-ER

1470 is based on presumed size. Could be other early Homo.KNM-ER 1481A, B Femur/tibia Larger size and more humanlike anatomy than femora from Olduvai (Wood 1992). Tie to KNM-ER

1470 is based on presumed size. Could be other early Homo.

Note. The list of referred specimens is taken from the most commonly followed delineation of H. rudolfensis, that presented by Wood (1991, 1992), with the additional inclusion ofUR-501, the mandible from Malawi that is considered nearly identical to but with smaller premolars than KNM-ER 1802 (Bromage, Schrenk, and Zonneveld 1995).a Main attributes or reason for association with KNM-1470 as provided by original author. Italicized comments provided by Anton and/or Leakey et al. (2012).b Note added in proof: I consider the newly described cranial (KNM-ER 62000) and mandibular (KNM-ER 60000 and 62003) specimens to associate with KNM-ER 1470 to theexclusion of KNM-ER 1802 for the reasons articulated by Leakey et al. (2012).


the specific designation and its association with genus Homo.He also suggested the possibility that a number of other cra-nial, mandibular, dental, and postcranial remains from KoobiFora might be included as well (see table 2). In 1995, KNM-ER 1813 became the type specimen of Homo microcranous(Ferguson 1995). Because the issue of affinity of OH 7, andhence the H. habilis name, has never been adequately ad-dressed (see below), it is most prudent to set it aside for themoment and work from the specimens that are more com-plete. So as not to confound the argument, I will refer hereto these specimens and any associated fossils as the 1813 and1470 groups, respectively. The main specimens assigned tothese groups and their earlier attributions are listed in table3.

As was argued in the initial proposal for two taxa, thereare important differences in facial structure and perhaps den-tal size between the two specimens (Wood 1991, 2009). Re-gardless of how it is hafted to the vault (Bromage et al. 2008),the KNM-ER 1470 face is quite flat, with a forward-facingmalar region and relatively deep and tall zygomatics that areanteriorly inclined. The anterior tooth row and the maxillarybone that holds it are somewhat retracted and narrow acrossthe canines (Wood 1985, 1991). Although the anatomy showssuperficial similarities to the Paranthropus face, zygomaticposition causes the facial flatness in Paranthropus, whereasthe midface itself is flat in KNM-ER 1470. Wood suggestedadditional parallels between the two, including postcaninemegadontia (albeit less marked in KNM-ER 1470). He thusdescribed the 1470 morph as a large-brained but large-toothedearly Homo. Alternatively, the KNM-ER 1813 face is moreconservative in its structure, with a moderately prognathicmidface, a rounder anterior maxilla, and somewhat more pos-teriorly positioned but more vertical zygomatic arches. TheKNM-ER 1813 face also houses relatively smaller teeth.

If we accept that the structural differences in the face in-dicate two different species, the constituents of the groupscan only be built based on direct anatomical associationsbetween other fossils and either 1470 or 1813. KNM-ER 1470,while preserving a face and vault, lacks maxillary dental crownmorphology and does not preserve a mandible. Table 2 de-lineates the specimens typically associated with KNM-ER 1470(aka Homo rudolfensis) and the original reason for so doing(mostly after Wood 1991, 1992). The large size of the vault,tooth roots, and flat anterior face suggested that large indi-viduals with a “heavy” masticatory pattern might be the bestfit. However, there are no discrete, derived anatomical ar-guments beyond vault size and inferred dental size for linkingKNM-ER 1470 with KNM-ER 1590, until recently the onlycomplete maxillary dentition assigned to H. rudolfensis. Al-ternatively, the table highlights the strong arguments, includ-ing premolar root form, for linking some of the mandibularremains (e.g., KNM-ER 819, UR 501) with one another andwith the more complete mandible KNM-ER 1802. However,there are again no linkages between these mandibles andKNM-ER 1590 or 1470 except for size. Despite this, KNM-

ER 1590 and KNM-ER 1802 have become the de facto ex-amples for H. rudolfensis maxillary and mandibular dentalmorphology, respectively (e.g., Schrenk, Kullmer, and Brom-age 2007; Spoor et al. 2007; Wood and Richmond 2000).While parsimony may suggest that this KNM-ER 1802/UR-501 group of mandibles may go with the 1470 face, there isno particular anatomical argument that they must.2

Alternatively, because of the better preservation of KNM-ER 1813, the 1813 morph can be extended to include othermaxillae from Kenya and Tanzania (e.g., OH 13, OH 62, OH65, KNM-ER 1805). OH 65 was initially aligned with KNM-ER 1470 largely on the basis of size and malar position (Blu-menschine et al. 2003; Clarke 2012). However, zygomatic rootposition, midfacial prognathism, anterior maxillary contours,arcade shape, and tooth position differ so markedly betweenthe two as to preclude their inclusion in the same group tothe exclusion of the 1813-group specimens (see Rightmireand Lordkipanidze 2009; Spoor et al. 2007).3 Thus, if twomorphs are accepted, OH 65 must be placed with the 1813group. Because KNM-ER 1813 also retains some maxillaryteeth, the maxillary dental anatomy of the group is knownand can be compared for consistency with those dentitionsthat are housed in the maxilla mentioned above. And becausesome of these other maxillae have associated mandibles (e.g.,OH 13; KNM-ER 1805), the mandibular dental anatomy andbony morphology can be directly seen and linked to isolatedmandibles. The 1813 group can thus be formed on the basisof direct anatomical ties.

The affinity of the type of H. habilis, OH 7, to either ofthese groups remains unclear. OH 7 is a difficult type fromwhich to judge the anatomy of the species because it com-prises a subadult, a partial mandible that is also deformedtaphonomically, a set of parietals that indicate approximatevault size but provide little definitive anatomy, and some iso-lated hand bones presumed to be the same individual. Thescenario that has found the most favor links OH 7 with the1813 group, thus named H. habilis (e.g., Schrenk, Kullmer,and Bromage 2007; Wood 1991). The 1470 group (whichaccording to those authors includes the 1802/UR-501 man-dibles) is then H. rudolfensis. OH 7 has often been suggestedto align with the 1813 group on dental crown anatomy, al-

2. Note added in proof: Recently described Kenyan fossils (KNM-ER60000, 62000, 62003) confirm facial differences between KNM-ER 1470and 1813 but not absolute molar size differences. These fossils also seemto exclude the 1802 group of mandibles from the 1470 group (Leakeyet al. 2012).

3. The main issue here is the contention that OH 65’s lower nasalregion; roots of the zygoma; and broad, flat, nasoalveolar clivus weremost similar to KNM-ER 1470’s. However, this is not the case. OH 65is much more prognathic subnasally and has more posteriorly positionedzygomatic roots than does KNM-ER 1470. Additionally, OH 65’s na-soalveolar clivus is arched at the alveolar margin where 1470 is flat, andits canine alveoli are not part of the anterior tooth row, whereas thoseof 1470 are. In short, OH 65 is a large version of KNM-ER 1813 andshows none of the structural features that are critical to erecting the 1470morph.

S284

Table 3. Main fossil specimens attributed to earliest or early Homo by original attribution and species group in this paper

Specimen Element Common attribution Species/group

12.0 Ma:East Africa:

A.L. 666-1 Palate Homo sp. aff. H. habilis (Kimbel et al. 1996) Early HomoKNM-WT 42718 Molar Homo sp. aff. H. habilis (Prat et al. 2005) Early HomoOmo E-G teeth Miscellaneous teeth Homo sp. aff. H. rudolfensis (Suwa, White, and Howell 1996) Early HomoUR-501a Mandible H. rudolfensis (Bromage, Schrenk, and Zonneveld 1995) 1802 groupKNM-BC 1 Temporal Homo sp. indet. (Hill et al. 1992), Hominidae gen. sp. indet. (Martyn and Tobias 1967) ?

South Africa:Sts 19 Cranial base Homo sp. (Kimbel and Rak 1993) ?Stw 151 Cranial fragments—juvenile Derived toward an early Homo condition (Moggi-Cecchi, Tobias, and Beynon 1998) ?

2.0–1.5 Ma non-erectus Homo:East Africa:

KNM-ER 1470 Cranium H. (A.) rudolfensis (Alexeev 1986) 1470 groupKNM-ER 1590 Cranial fragments/teeth H. rudolfensis (Wood 1991) Early HomoKNM-ER 3732 Partial cranium H. rudolfensis (Wood 1991) Early HomoKNM-ER 3891 Cranial fragments/maxilla H. rudolfensis (Wood 1991) Early HomoKNM-ER 819 Mandible fragment H. rudolfensis (Wood 1991) 1802 groupKNM-ER 1482b Mandible H. rudolfensis (Wood 1991) Early Homob

KNM-ER 1483 Mandible fragment H. rudolfensis (Wood 1991) 1802 groupKNM-ER 1801b Mandible partial H. rudolfensis (Wood 1991) Early Homob

KNM-ER 1802 Mandible H. rudolfensis (Wood 1991) 1802 groupKNM-ER 1501 Mandible partial H. habilis (Wood 1991:270) 1813 groupKNM-ER 1813 Cranium H. habilis (Howell 1978; Wood 1991) 1813 groupKNM-ER 1805 Calvaria � maxilla � mandible H. habilis (Wood 1991) 1813 groupKNM-ER 3735 Cranial fragments � partial skeleton H. habilis (Leakey et. al. 1989) Early Homo; ?1813 groupKNM-ER 42703c Maxilla fragment H. habilis (Spoor et al. 2007) 1813 groupOH 7 Mandible � cranial fragments H. habilis (Leakey, Tobias, and Napier 1964) H. habilisOH 13 Maxilla, mandible, teeth, cranial

fragmentsH. habilis (Leakey, Tobias, and Napier 1964) 1813 group

OH 16 Maxillary and mandibular teeth �cranial fragments

H. habilis (Leakey, Tobias, and Napier 1964) Early Homo

OH 24 Cranium trampled H. habilis (Leakey, Clarke, and Leakey 1971) 1813 groupOH 62 Maxilla � fragmentary skeleton H. habilis (Johanson et al. 1987) 1813 groupOH 65 Maxilla � dentition Similar to 1470, which they include in H. habilis (Blumenschine et al. 2003; Clarke

2012)1813 group

S285

South Africa:Stw-53d Reconstructed partial cranium Homo aff. habilis (Hughes and Tobias 1977); some view as Australopithecus africanus

(Kuman and Clarke 2000) or type of H. gautengensis (Curnoe 2010)?Early Homo

Stw-80 Mandible partial, crushed � teeth Homo possibly aff. ergaster (Kuman and Clarke 2000) Early HomoSK 27 Crushed cranium (juvenile) � teeth Homo (Howell 1978); H. habilis (Curnoe and Tobias 2006) Early Homo

2.0–1.5 Ma Homo erectus-like:East Africa

KNM-ER 730 Occipital, parietal, frontal,mandible, partial

Homo aff. erectus (Wood 1991) H. erectus

KNM-ER 820 Mandible—subadult Homo aff. erectus (Wood 1991) H. erectusKNM-ER 992 Mandible Homo aff. erectus (Wood 1991) H. erectusKNM-ER 1808 Skeleton � cranial fragments H. erectus (Walker, Zimmerman, and Leakey 1982) H. erectusKNM-ER 3733 Cranium H. erectus (Leakey and Walker 1985) H. erectusKNM-ER 3883 Cranium H. erectus (Leakey and Walker 1985) H. erectusKNM-ER 42700 Calvaria H. erectus (Spoor et al. 2007) H. erectusKNM-WT 15000 Skull � skeleton H. erectus (Walker and Leakey 1993) H. erectusOH 9c Calvaria H. leakeyi (Heberer 1963); H. erectus (Rightmire 1979) H. erectusOH 12c Cranial fragments H. erectus (Rightmire 1979) H. erectusDakac Calvaria H. erectus (Asfaw et al. 2002) H. erectusKGA 10-1c Mandible partial H. erectus (Asfaw et al. 1992) H. erectus

South Africa:SK-847 Partial face Homo sp. indet. aff. erectus (Clarke, Howell, and Brain 1970); H. erectus (Kimbel,

Johanson, and Rak 1997; Tobias 1991; Walker 1981); Homo aff. habilis (Grine 2001)Homo aff. erectus

SK-15c Partial mandible H. erectus (Howell 1978; Robinson 1961); Homo sp. indet. (Grine 2001) Homo aff. erectusSK-45 Partial mandible H. erectus (Robinson 1961); H. habilis (Howell 1978); Homo sp. indet. (Grine 2001) Homo

Georgia:Dmanisi (multiple) Crania, mandibles, postcrania H. erectus (p H. ergaster; Gabunia et al. 2000; Rightmire and Lordkipanidze 2009) H. erectus

Note. Most isolated teeth, mandible fragments, and specimens younger than 1.5 Ma are omitted. Newly described specimens from the 1470 group (KNM-ER 60000, 62000, 62003) arenot included, but see Leakey et al. (2012).a May be younger than 2.0 Ma.b Following Leakey et al. (2012), KNM-ER 1482 and 1801 likely belong to the 1470 group along with newly described KNM-ER 60000, 62000, and 62003.c Is or may be younger than 1.5 Ma.d May be older than 2.0 Ma.


though it should again be noted that no dental remains arefirmly associated with the 1470 group (the calvaria retainsonly partial roots), and it is unclear, therefore, how or whetherthe groups differ in dental occlusal anatomy or size. Alter-natively, OH 7’s roughly estimated cranial capacity (690 cm3;Tobias 1991) perhaps aligns it with KNM-ER 1470. If OH 7is so placed, that group (p KNM-ER 1470 only � OH 7)then becomes H. habilis, leaving the 1813 group as H. mi-crocranous. Present data are, in my opinion, insufficient tochoose between these scenarios or to show to which the 1802/UR-501 group links. While future evidence may prove con-vincing, it is also possible that 1802/UR-501 represents a thirdmorph and that the affinities of OH 7 are with any of thethree (see also Leakey et al. 2012).

It is generally agreed that there is some fossil evidence forat least one member of non-erectus early Homo in SouthAfrica; however, there is no consensus concerning which spe-cies are present or whether these also occur in East Africa(see also Grine 2005; Grine et al. 2009). The remains in ques-tion date (roughly) between 2.0 and 1.5 Ma and come fromSterkfontein, Swartkrans, and Drimolen. The Swartkrans re-mains are most frequently linked with H. erectus and arediscussed below. The Sterkfontein remains include isolatedteeth, two partial mandibles, and the cranium Stw 53, whosetaxonomic identification and reconstruction is heavily con-tested (e.g., Curnoe 2010; Curnoe and Tobias 2006; Grine2001; Grine et al. 2009; Kuman and Clarke 2000). The Sterk-fontein cranial remains have been variously affiliated withHomo (aff. habilis or sp. indet.; Kuman and Clarke 2000 forthe mandibles; Curnoe and Tobias 2006; Grine et al. 2009)or Australopithecus (Kuman and Clarke 2000 for StW 53;Clarke 2012). The isolated teeth from Sterkfontein and Dri-molen lend strong support to the identification of an as yetunnamed early non-erectus Homo in South Africa (Grine etal. 2009). Because they do not affiliate strongly with the EastAfrican teeth, following Grine (2001; Grine et al. 2009) theyare considered here simply as early non-erectus Homo.

Postcranial remains of early East African Homo are few,and fewer still are certainly associated with one of the cranialmorphs (table 4). OH 7 has a partial hand. The OH 62 partialskeleton can be associated with the 1813 group on the basisof maxillary morphology. The KNM-ER 3735 fragmentaryskeleton is Homo, based on anatomy of the posttoral sulcus,mandibular fossa, and zygomatic, but of uncertain affinity.Isolated elements of the lower limb—such as OH 8 (foot),OH 35 (distal tibia), and various hind limb fragments fromKoobi Fora—have been tentatively assigned to Homo. How-ever, for these isolated remains there is always some questionas to which hominin they belong (see Gebo and Schwartz2006; Wood and Constantino 2007). If OH 8 is early Homo,the ankle and “close-packed” arches may indicate a patternof bipedalism similar to Homo sapiens (see Harcourt-Smith2007; Harcourt-Smith and Aiello 2004), although primitiveelements are retained as well. The OH 7 hand retains theprimitive condition of the carpals and curvature of the pha-

langes, which presumably indicate some arboreal behavior(Tocheri et al. 2007). However, other aspects of especially thethumb suggest precision grip abilities more derived towardhuman than ape capabilities (Susman and Creel 1979). Thesizes and proportions of the limbs of the associated skeletonsare discussed below.

Non-erectus Early Homo: Size and Proportions

Given that at least two different facial morphs seem to coexistin time and space in East Africa, what can be said about sizeand shape of non-erectus Homo? As currently constructed, the1470 group has a larger cranial capacity (750 cm3) than doesthe 1813 group (510–675 cm3) with the caveat that OH 7 hasnot been assigned to either group (tables 5, 6). Additionally,more fragmentary remains, such as the OH 65 maxilla, con-form well to the shape of the KNM-ER 1813 palate and teethbut are quite a bit bigger and could (but are not required to)imply a larger cranial capacity for that group. Similarly, KNM-ER 1590, a fragmentary juvenile specimen whose parietalssuggest a large cranial capacity, has been included by somein the 1470 group based on size and the presumption thatthis group is large and the 1813 group small, but it couldinstead affiliate with the 1813 group if size differences are notsubstantiated. Yet even without such size extension of the 1813group, brain size among the well-preserved early Homo in-dividuals (KNM-ER 1805; OH 13, 16, 24) is fairly continu-ously distributed between the two end members, KNM-ER1813 and 1470 (table 6). Such distribution suggests that av-erage ranges of cranial capacity will vary depending on whois included in each group. On the basis of facial morphology,KNM-ER 1805 and OH 13 and 24 should be included in the1813 group (OH 16 does not preserve facial anatomy), andthus the upper end of this group is as much as 670 cm3,substantially larger than KNM-ER 1813. Regardless ofwhether there is one or more species of early non-erectusHomo, brain size across the entire group ranges from 510 to750 cm3.

Dental size cannot be compared between the groups be-cause only the 1813 group has complete teeth that are certainlyassociated with it; however, dental size can be assessed acrossthe entire early Homo sample and for the 1813 group alone(table 7). KNM-ER 1470 has been inferred to be large toothedon the basis of preserved tooth roots. However, this inferenceis overstated given that especially the postcanine roots areobserved low on the root and in oblique section. The max-illary dental metrics generally quoted for this species are fromKNM-ER 1590, which has extremely large teeth but as dis-cussed above has no firm anatomical tie to KNM-ER 1470.Similarly, the mandibular dental metrics are from KNM-ER1802, which has large teeth but also has no firm anatomicaltie to KNM-ER 1470. Thus, the argument that the 1470 groupis large toothed is circular because it is made on the basis ofspecimens that have been placed in the group because of largedental size resulting in the conclusion that the group has large

Table 4. Main postcranial specimens attributed in this paper to earliest or early Homo in Africa and Georgia by original attribution and species groupused in this paper

Specimen Element Common attribution Species/groupBody mass

estimate (kg)a

East Africa:KNM-ER 1472 Femur Homo rudolfensis (Wood 1992) ?Early Homo 49.6KNM-ER 1481A, B Femur, tibia H. rudolfensis (Wood 1992) ?Early Homo 57KNM-ER 3228 Os coxae Homo ?erectus (Rightmire 1990) or H. rudolfensis (McHenry and Coffing 2000) Early Homo 63.5OH 35 Tibia, distal H. habilis Early Homo 31.8OH 62 Maxilla � fragmentary skeleton H. habilis (Johanson et al. 1987) 1813 group ?OH 28 Os coxae H. erectus (Rightmire 1990) H. erectus 54OH 34 Femur H. erectus (Rightmire 1990) H. erectus 51KNM-ER 736 Femur H. erectus (Rightmire 1990) H. erectus 68.4KNM-ER 737 Femur H. erectus (Rightmire 1990) H. erectus ?KNM-ER 1808 Multiple cranial and postcranial elements H. erectus (Walker, Zimmerman, and Leakey 1982) H. erectus 63.4KNM-WT 15000 Skeleton � skull H. erectus (Walker and Leakey 1993) H. erectus 51Gona Pelvis H. erectus (Simpson et al. 2008) ?Homo 39.7

South Africa:SK-1896 Femur, distal Homo aff. erectus (Susman, de Ruiter, and Brain 2001) Homo aff. erectus 57SK-2045 Radius, proximal Homo aff. erectus (Susman, de Ruiter, and Brain 2001) 53–58SKX-10924 Humerus, distal (small) Homo aff. erectus (Susman, de Ruiter, and Brain 2001) Homo aff. erectus (30)SKW(SKX) 34805 Humerus, distal (large) Homo aff. erectus (Susman, de Ruiter, and Brain 2001) H. aff. erectus ?

Georgia:Dmanisi (small) Multiple elements H. erectus (p H. ergaster; Gabunia et al. 2000; Rightmire and Lordkipanidze

2009)H. erectus 40.7

Dmanisi (large) Multiple elements H. erectus (p H. ergaster; Gabunia et al. 2000; Rightmire and Lordkipanidze2009)

H. erectus 48.8

Note. Only those bones useful in establishing stature or body weight are listed. Hand and foot elements excluded.a In addition to sources in a previous column of this table, body mass estimates follow table 8, Holliday (2012), Pontzer (2012), and Ruff, Trinkaus, and Holliday (1997).


Table 5. Comparative brain and body size of Australopithecus and Homo

South Africa East Africa South Africa East Africa/Georgia

A. sediba A. africanus A. afarensisEarly

non-erectus Homo H. aff. erectus Early H. erectus

Brain sizea 420 (MH 1) 571 (Stw 505) 550 (A.L. 444-2) 510 (1813) . . . 638 (D3444)485 (Sts 5) 485 (A.L. 333-45) 580 (1805) 655 (D2282)443 (MLD 37/38) 400 (A.L. 162-28) 595 (OH 24) 690 (42700)385 (Sts 60) 630 (OH 16) 727 (OH 12)410 (Sts 71) 660 (OH 13) 775 (D2880)

680 (OH 7) 804 (3883)750 (1470) 848 (3733)

909 (15000)995 (Daka)

1,067 (OH 9)Mean 420 454–461 478 629 . . . 810 (w/Dmanisi)

863 (East Africa only)CV . . . 15.9 15.7 12.2 . . . 17.8 (w/Dmanisi)

15.9 (East Africa)Body mass/femur

lengthb 35.7 (MH 2) 45.4/433.5 (Stw 99) 50.1/382 (A.L. 333-3) 63.5 (3228) 57 (SK 1896) 68.4 (736)31.5 (MH 1) 41.3 (Stw 443) 48.2 (A.L. 333x-26) 49.6/401 (1472) 53–58 (SK 2045) 63.4/485 (1808)

40.7 (Stw 311) 45.6/375 (A.L. 827-1) 57/396 (1481) [30] (SKX 10924) 54/456 (OH 28)38.4 (Sts 340) 45.4 (KSD-VP-1/1) 31.8 (OH 35) 51/429 (15000)37.9 (Stw 389) 42.6 (A.L. 333-7) 31 (OH 8) 51/432 (OH 34)34.2 (Stw 25) 41.4 (A.L. 333-4) —/315 (OH 62) 48.8 (Dmanisi large)32.7 (Stw 392) 40.2 (A.L. 333-w-56) 40.7 (Dmanisi small)32.5 (TM 1513) 33.5 (A.L. 333-8) 39.7 (Gona)30.5 (Stw 102) 28/281 (A.L. 288-1)30.3/276 (Sts 13, 34) 27.1 (A.L. 129a)27.5 (Stw 347)23.3 (Stw 358)

Mean body mass 33 34 40 44 44 52 (w/Dmanisi)55 (East Africa only)

CV body mass/femur length 9/— 18.7/— 20.2/16.3 33/13 32/— 19.3/8.7 (w/Dmanisi)

18.8/5.8 (East Africa)

Note. Specimen numbers are in parentheses. Brain size in cm3; mass in kilograms; femur length in millimeters.a Endocranial capacity for A. sediba from Berger et al. (2010), for A. africanus from Neubauer et al. (2012), for A. afarensis from Holloway andYuan (2004), and individually for Homo as indicated in table 6 of this paper.b Body mass estimates follow Pontzer (2012) and tables 4 and 8 of this paper. Femur length CVs are raw values not corrected for dimensionality.

teeth. Across all of early Homo, then, molar size is somewhatdiminished over the condition in Australopithecus and some-what, but not significantly, larger than the condition in earlyH. erectus (table 5; Anton 2008). There is some suggestionthat there may be a large-toothed morph, but it is unclearwhether this morph belongs to the 1470 group or not (seenote in table 7).

Dental proportions and occlusal morphology can also bedescribed for the 1813 group, which differs from Australo-pithecus and H. erectus. Whether this morphology is uniqueto this group of early Homo is unknown. The 1813 groupshows buccolingual narrowing of its cheek teeth, especiallythe molars, relative to Australopithecus, and its M2s are mostlyrhomboidal in form. The third molar is large relative to M2,however, in contrast to the condition in H. erectus (Spoor et

al. 2007). There is also some evidence of difference betweenthe 1802 group and the 1813 group in both dental and man-dibular morphology. The former have relatively broader pre-molars with greater talonid development and differentlyshaped roots, and the base of the mandible is everted and thesymphysis more vertical than the apparent condition in the1813 group (Anton 2008; Schrenk, Kullmer, and Bromage2007, table 9.1); however, it is unclear to what extent thesedifferences may reflect intraspecific idiosyncratic variation inmandibular size and robusticity.

Body size cannot be compared between the groups becauseonly the 1813 group has associated postcranial remains (i.e.,OH 62 on maxillary form; table 3). OH 62 has been inter-preted as small bodied and with relatively long and strongarms, but the specimen is very fragmentary (Haeusler and

Table 6. Size comparisons of individual fossils of Homo habilis sensu lato and early Homo erectus sensu lato

Specimena

Group (1813or 1470)

Geologicalage (Ma)

Presumedsexb

Brainsize (cm3)

Mean body mass (kg)from orbit area (k/a)c

Range body mass (kg)from orbit area (k/a)c

Mass (kg)from

postcrania M1 area M1 area M2 area M2 area

Homo habilis s.l.:KNM-ER 1813 Homo microcranous 1813 1.9 Female 510 34.9/31.0 24.3–50/36.9–35.5 . . . 1,560 . . . 1,640 . . .KNM-ER 1805 1813 1.9 Male 580 . . . . . . . . . 1,770 . . . 1,730 . . .OH 24 1813 1.8 Female 595 30.3/36.3 21.1–43.4/31.5–41.8 . . . 1,790 . . . 1,890 . . .OH 16 ? 1.8 ? 625–638 . . . . . . . . . 2,010 1,870 2,020 2,330OH 13 1813 1.6 Female 650–675 . . . . . . . . . 1,640 1,510 1,810 1,700OH 7 Homo habilis ? 1.8 Male 647–690 . . . . . . . . . . . . 1,760 . . . 2,110KNM-ER 60000 1470 1.8 ? . . . . . . . . . . . . . . . 1,460 . . . 1,790KNM-ER 1590 ? 1.85 ? . . . . . . . . . . . . 2,090 . . . 2,570 . . .KNM-ER 62000 1470 1.9 Female . . . . . . . . . . . . 1,850 . . . 2,016 . . .KNM-ER 1470 Homo rudolfensis 1470 2.03 Male 750 45.5/77.4 33.8–69.9/64.1–93.6 . . . . . . . . . . . . . . .

Homo erectus s.l.:Dmanisi D3444 1.7 Male 638 . . . . . . . . . . . . . . . . . . . . .D2282/D211 1.7 Female 655 . . . . . . . . . 1,560 1,550 1,560 1,420KNM-ER 42700 1.55 Female 690 . . . . . . . . . . . . . . . . . . . . .OH 12 1.2 Male 727 . . . . . . . . . . . . . . . . . . . . .D2280 1.7 Male 775 . . . . . . . . . . . . . . . . . . . . .KNM-ER 3883 1.5 Male 804 57.4/83 39.9–57.5/68.2–101 kg . . . . . . . . . . . . . . .KNM-ER 3733 1.8 Female 848 59.2/88.8 41–85.3/72.5–108.7 kg . . . . . . . . . 1,860 . . .KNM-WT 15000 1.5 Male 909 59.9/ . . . 41.5–86.4/ . . . 51 1,490 1,410 1,500 1,520Daka 1.0 ? 995 . . . . . . . . . . . . . . . . . . . . .OH 9 1.5 Male 1,067 . . . . . . . . . . . . . . . . . . . . .

Note added in proof. Data from recently described fossils from the 1470 group (Leakey et al. 2012) are included here but could not be included in text discussion.a Species names for which these serve as type specimens are noted by the specimen number.b The sex of all specimens is unknown. These estimates represent the most frequent inferences and should be considered uncertain at best.c Body mass estimates from orbital area are from k p Kappelman (1996) and a p Aiello and Wood (1994).


Table 7. Dentognathic summary statistics for Homo habilis sensu lato and early Homo erectus from Africa and Georgia

Africa Africa and Georgia

KNM-ER 1590 KNM-ER 1802 1813 group Homo habilis sensu lato Early Homo erectus

M1 buccolingual . . . 130 119.3/6.7 (3) 123.8/9.0 (8) 119.2/7.6 (5)M1 mesiodistal . . . 148 137.6/0.6 (3) 138.5/6.1 (8) 132.2/6.5 (5)M1 area . . . 1,920 1,610/118 (3) 1,705/181 (8) 1,592/131 (6)M1 buccolingual 148 . . . 132.3/2.9 (6) 134.9/6.0 (8) 129.2/6.8 (4)M1 mesiodistal 142 . . . 128.3/5.1 (6) 132.3/8.7 (8) 126.2/3.8 (5)M1 area 2,090 . . . 1,695/88 (6) 1,784/181 (8) 1,630/127 (5)Corpus height M1 . . . 38 29.3/2.7 (3) 32.6/4.3 (11) 29.4/2.9 (6)Corpus breadth M1 . . . 23 18.3/2.3 (3) 20.8/3.68 (10) 20.1/0/9 (6)Symphyseal height . . . 36 25 33.2/6.0 (5) 33.4/2.6 (4)Symphyseal depth (labiolingual) . . . 24.5 18 21.7/3.2 (5) 18.5/2.4 (4)

Note. Individual fossils KNM-ER 1590 and 1802 and the 1813 group are also presented. Measurements in millimeters; mean/SD; (n).Note added in proof. Teeth of the 1470 group are now known from recently described fossils, and, while dental proportions differ, teeth are notalways large. See Leakey et al. (2012) for discussion. New fossils are not included in these statistics, but see table 6 for some raw measures.

McHenry 2004; Richmond, Aiello, and Wood 2002; see alsoHolliday 2012). KNM-ER 3735 is larger than OH 62 but morefragmentary and is not definitively assigned to either subgroup(Haeusler and McHenry 2007; Leakey et al. 1989). The rangeof body weights in these two specimens is 30–46 kg (66–101lb.), respectively (tables 5, 8; Johanson et al. 1987; Leakey etal. 1989). Inferred stature is very approximately 118–145 cm(3′11′′–4′8′′). Three other relatively large femora may be as-signed to Homo sp. and would support the larger end of thissize range (149 cm/57 kg; Holliday 2012) if their attributionis correct. However, two of the femora (KNM-ER 1472 and1481) are often attributed either to the 1470 group (becausethey are large; Wood 1992) or H. erectus (because they aresomewhat flattened in the subtrochanteric area and are large;Kennedy 1983a, 1983b; but see Trinkaus 1984), and they arealso attributed to H. habilis sensu stricto by others (e.g.,Schrenk, Kullmer, and Bromage 2007). One large partial oscoxae (KNM-ER 3228) has a geological age of 1.95 Ma. Thisspecimen is usually considered Homo and possibly H. erectus(Anton 2003; Rightmire 1990), although McHenry and Coff-ing (2000) suggest it may represent a large-bodied non-erectusHomo (their H. rudolfensis). Based on estimates of femoralhead size, a body mass estimate of 60–65 kg has been sug-gested (Ruff, Trinkaus, and Holliday 1997). Collectively, theseremains suggest that the largest end of the non-erectus earlyHomo body size range is just under 5 ft. tall (150 cm), andthe average weight is 44 kg (range 31–65; tables 4–6, 8; seeHolliday 2012; Pontzer 2012).

There is much discussion as to whether the limb propor-tions of early Homo are as or even more primitive than Aus-tralopithecus afarensis and therefore whether they differentiateearly Homo from H. erectus. Based on cross-sectional strengthmeasures, the OH 62 humerus is relatively stronger comparedwith its femur than is true of recent humans and is like thoseof Pan (Ruff 2010; see also Richmond, Aiello, and Wood2002). Thus, OH 62, but not later Homo, likely participatedin substantial arboreal locomotion as well as terrestrial (bi-pedal) locomotion. This would be supported by the OH 7

wrist bones, if they are of the same taxon (Tocheri et al. 2007).Alternatively, hind-limb elongation remains debated becauseof uncertainties in reconstructions of long bone lengths fromthe highly fragmentary OH 62 and KNM-ER 3735. At leastone set of researchers argues that hind-limb elongation mayhave been present (Haeusler and McHenry 2004; Reno et al.2005; but see Korey 1990; Richmond, Aiello, and Wood 2002).And the relatively long distal tibia of OH 35 may supportthis idea (Harcourt-Smith 2007). However, others have re-constructed the lengths differently and found the primitivecondition (i.e., long humerus, short femur; Richmond, Aiello,and Wood 2002). More recent work, however, suggests thathind-limb length proportions do not actually differ betweenAustralopithecus and Homo (see Holliday 2012; Pontzer 2012).So, while strength proportions appear to link OH 62 withAustralopithecus rather than later Homo, hind-limb elongationrelative to body size would appear to be the same in all genera.

It will be clear from the small number of fossils in eachgroup and the disagreement about taxonomic assignmentsthat assessing sexual dimorphism will be nearly impossiblefor early Homo. Ideally, we should identify males and femalesby focusing on discrete characters that are independent ofoverall body size, such as canine size and robusticity, and thenassess male and female mean values from these independentlyassigned subgroups. The 1813 group is the only one in whichit is possible to try to assess sexual dimorphism in this way,but such characters are few. Historically, the development ofcranial crests in KNM-ER 1805 over the condition in KNM-ER 1813 has been used to suggest the former is male and thelatter female (Wood 1991:84). None of the other known spec-imens in the group exhibit crests, leaving KNM-ER 1805 asthe only male candidate at the moment. Cranial capacity dif-fers little between the two specimens (510 vs. 580 cm3). How-ever, molar occlusal areas are larger in KNM-ER 1805. Yetthey are no larger than, say, OH 24, which lacks crests andis thus a presumed female. Here we face the issue of thepossibility of large females and small males and the inadequatesampling of the fossil record. For the postcranial skeleton, if


Table 8. Body size in East African adult early non-erectus Homo and early East African and Georgian Homo erectusindividuals and isolated elements

Africa Georgia

Taxon

Early Homoadult

OH 62

Early Homoadult

KNM-ER 3735

Adultisolated

Homo sp.a

H. erectusadult

KNM-ER 1808b

H. erectusadult

individualsc

InferredH. erectusisolated

elementsd

H. erectusadult

individualse

Brain size (cc) . . . . . . 510–750 . . . 909� 690–1,067 638–775x p 629 x p 863 x p 686

Femur length (mm) [280–374] . . . 350 (1503) 480 . . . . . . 386400 (1472)395 (1481)

. . . . . . x p 383.75 x p 498 x p 454.5 . . .473

Range (n) [280–374] (1) . . . 350–400 (4) 480 (1) 480–517 (2) 430–500 (4) 386 (1)466–480 (2)

Stature (cm):Mean 118 . . . 145 173 179 174 155

170Range (n) . . . . . . 173–185 (2) 161–186 (4) 145–166

168–173 (2)Body mass (kg):

Mean orbit . . . . . . 36.9 k . . . . . . 58.8 k . . .48.2 a 85 a

Femur head mass estimates . . . . . . 49.6 (1472) 63.4 51 (15000) 49.6 (D4507)57 (1481)

Mean postcrania 33 46 48 59 64.5 (2) 54.75 48.8 large63 57 (2) 40.2 small

Range (n) . . . . . . 21–77 . . . 63–68 (2) 51–68 (4) 40–49.6 (3)55–63 (2)

a KNM-ER 1503 is assigned to either Homo habilis or Australopithecus boisei, as are most specimens of this age from Koobi Fora (McHenry 1991);“k” and “a” refer to orbitally estimated mean body masses from Kappelman (1996) and Aiello and Wood (1994); femur head estimates from Holliday(2012); parenthetical number next to femur value is the specimen number providing that value. Square brackets are estimates.b KNM-1808 mean body mass values from Ruff and Walker (1993), top, or Ruff, Trinkaus, and Holliday (1997), bottom.c Adult individuals are KNM-1808, and adult estimates for KNM WT 15000 femur, stature, and body mass following either Ruff and Walker (1993),top, or Graves et al. (2010); Ohman et al. (2002) provide even smaller stature estimates for KNM-WT 15000.d Femora: KNM-ER 736 and 737, OH 28 and 34.e Mass data are for large and small adult individuals (see Pontzer 2012), and range estimates include other elements (Holliday 2012).

the small OH 62 skeleton is female and the larger KNM-ER3735 is male, then if they are typical of their sexes, sexualdimorphism in body mass could be as much as 1.5 (male/female weight dimorphism as compared with chimpanzees atabout 1.3 and H. sapiens about 1.1). Alternatively, one coulduse CVs from the certain members of the 1813 group, al-though this may give an artificially low value because it is notclear how to handle some of the larger specimens, such asKNM-ER 1590. And using the overall CV also implicitly as-sumes that size differences and sex differences are the samething. In the absence of discrete characters to indicate sex,this may be a necessary assumption, but it is worth notingthe confounding issue (see also Plavcan 2012). CVs for upperand lower first molar area are 5.2 and 7.3, respectively, andCVs for capacity are 9.9. These values are lower than thosefor early African and Georgian H. erectus, particularly forcranial capacity. If we take instead all the early non-erectusHomo, regardless of group affiliation, the cranial CVs are stilllower than those for H. erectus, but their body mass CVs are

significantly greater (table 5). Plavcan (2012) provides a fullerdiscussion of the options for considering size dimorphism inthese taxa.

Homo erectus sensu lato (1.9–? Ma)

In the early part of its range with which we are concerned,Homo erectus overlapped for nearly half a million years withother groups of early Homo in Africa, principally the 1813group, whose last appearance datum is 1.44 Ma (fig. 1; seeSpoor et al. 2007). The earliest H. erectus (at about 1.8–1.9Ma) are found in Koobi Fora, Kenya, and the species persistsin Africa until about the Brunhes-Matuyama boundary (0.78Ma; fig. 1; Asfaw et al. 2002; Feibel, Brown, and McDougall1989; Potts et al. 2004). Homo erectus is best known fromKenya and Tanzania, although Ethiopian and South Africanspecimens also exist (e.g., Asfaw et al. 2002; Robinson 1953).The earliest African H. erectus quickly dispersed into Asia by1.7–1.8 Ma (Gabunia et al. 2000; Swisher et al. 1994). The


species persists in Asia and Island Southeast Asia through atleast the middle if not the late Pleistocene (Indriati et al. 2011;Shen et al. 2009; Swisher et al. 1996).4 Given interest here inthe origin and evolution of Homo, I consider only the earlyAfrican and Georgian record.

Homo erectus: Taxa, Individuals, and Anatomy

Like Homo habilis, H. erectus sensu lato is an umbrella taxonthat may include nested sets of other taxa (table 1). Thosewho split the larger taxon in two usually refer to the African/Georgian members as Homo ergaster based on the type spec-imen KNM-ER 992 (Groves and Mazek 1975).5 Most recently,Homo georgicus (Gabunia et al. 2002) was named to accom-modate morphology in the Georgian remains that is arguedto combine more primitive traits (especially of the face andbrain size) than in the earliest African H. erectus. However,this variation is easily encompassed within early African H.erectus (Anton 2003; see also Baab 2008; Rightmire and Lord-kipanidze 2009), which also shows clear signs of size-relatedshape changes with cranial capacity (Anton et al. 2007; Spooret al. 2007). And the derived characters shared by the Georgianand recently discovered small-brained African specimensunite them taxonomically (Spoor et al. 2007). For the purposeof discussion here, it makes little difference whether the FarEast Asian and African/Georgian hominins are seen as re-gional demes of one species, as two distinct species, or evenmultiple species, although I favor the former interpretation(table 3; Anton 2003).

As will be clear from the discussion of the genus, H. erectusis now considered to take the first major anatomical andbehavioral steps in the direction of a “modern human” bodyplan (Anton 2003; Anton et al. 2007; Walker and Leakey1993). Although the species was not identical to Homo sapiensin size or shape, H. erectus bodies and brains were larger andtheir teeth and especially jaws were somewhat diminished insize, on average, compared with those of earlier members ofHomo (Anton 2008). However, their teeth were larger andtheir brains smaller than in later Homo. Their lower-limbskeleton was relatively elongated compared with body massover the condition in Pan, and their upper limbs were some-what foreshortened over the condition in Australopithecus andperhaps other early Homo (Holliday 2012; Pontzer 2012).That said, newly discovered small-sized individuals from

4. Throughout the paper I use the traditional chronological delineationof the Plio/Pleistocene boundary as occurring at 1.8 Ma, the onset ofsevere northern hemispheric glaciation.

5. While this is the most common nomen, it should be noted thatmany of the often included specimens represent types for earlier namedspecies (i.e., Homo (T.) capensis [Swartkrans 15; Robinson 1953]; Homo(At.) mauritanicus [Ternifine 1; Arambourg 1954], and Homo leakeyi [OH9; Heberer 1963]; table 1); thus, the earliest of those included shouldprovide the group name. This has not been the case, however, becausethese earlier types were not included when the species was named andare not consistently included in Homo ergaster by all scholars (e.g., Wood1991:276).

Georgia, Kenya, Tanzania, and perhaps Ethiopia suggest sub-stantial overlap in absolute size with earlier Homo species(tables 5–7; Anton 2004; Gabunia et al. 2000; Potts et al. 2004;Rightmire 1979; Simpson et al. 2008; Spoor et al. 2007; Vekuaet al. 2002).

Although absolutes of size do not differ, some proportionsdo, and so individuals of H. erectus are relatively easy todifferentiate from all other early Homo on the basis of cra-niodental remains. Early H. erectus tends to have somewhatsmaller occlusal areas and fewer roots than other early Homobut relatively larger crowns and more complex roots (espe-cially premolar) than do modern humans (table 7; Gabuniaet al. 2000, 2001; Indriati and Anton 2008). Crowns, especiallymolars, tend to be buccolingually narrower compared withlength and less bulbous than in early Homo, with cusp apicescloser to the outer margins of the tooth than in other earlyHomo (Anton 2008; Indriati and Anton 2008). Homo erectusalso shows size reduction along the molar row with the thirdmolar reduced or similar in size to M2 (Spoor et al. 2007).And early H. erectus jaws are relatively more lightly built withnarrower extramolar sulci than in early Homo (Grine 2001).The H. erectus symphysis is thinner (anteroposterior), thegenioglossus pit is relatively lower, and the postincisive plane,although obliquely oriented, is not quite so pronounced asin other early Homo (Anton 2008).

Although general vault thickness scales with cranial capacityin H. erectus and other early Homo, H. erectus shows species-typical examples of thickening (Anton et al. 2007; Spoor etal. 2007). These include (1) essentially continuous supraor-bital tori of variable thickness associated with a posttoral shelf/sulcus that may be continuous, (2) occipital tori that arecontinuous but somewhat variably expressed, often contin-uous with the angular tori and mastoid crests and often as-sociated with a supratoral sulcus, (3) angular tori, and (4)midline (sagittal, bregmatic, and frontal) keels.

Homo erectus also differs from other early Homo and mod-ern humans in other aspects of the cranium. The occipitalsquama is relatively short, and the petrous temporal is moresagittally oriented and angled relative to the tympanic portion(i.e., petrotympanic angle reduced; Rightmire and Lordki-panidze 2009; Weidenreich 1943; although the base of earlierHomo is not well known). The glenoid fossa is relativelybroader anteroposteriorly (compared with mediolaterally)than in other early Homo (Spoor et al. 2007). The face isdescribed as more similar in proportions to modern humansthan with other early Homo (Bilsborough and Wood 1988;Wood and Richmond 2000); however, the positioning andform of the zygomatics and supraorbital torus is more similarto the 1813 group than to KNM-ER 1470, as is relative facialbreadth, which is greatest at the midface in 1470 but at thesuperior face in the other groups.

While there again is some difference of opinion as to tax-onomic affinities of the South African fossils, craniodentalremains from Swartkrans are likely to represent either H.erectus (as discussed above) or something very erectus-like.


Two mandibles, SK 15 and SK 45, were the first at Swartkransto be recognized as Homo (originally as Telanthropus capensis;Broom and Robinson 1949; Grine 2001; Grine et al. 2009;Robinson 1961). The most compelling evidence, however, isthe SK 847 partial face that shows strong affinities with earlyEast African H. erectus (Anton 2003; Clarke, Howell, andBrain 1970; Kimbel, Johanson, and Rak 1997; Walker 1981).Given this strong resemblance, the postcranial remains atSwartkrans that differ from those of Paranthropus are mostusually assigned to Homo aff. erectus (Susman, de Ruiter, andBrain 2001).

The postcranial record for early East African H. erectus isfar better than that of other early Homo. In Africa, a singlewell-preserved skeleton, KNM-WT 15000, is the main datapoint, but this information is augmented by a second partialskeleton, KNM-ER 1808, and a number of large isolated el-ements (tables 4, 8). We should be cautious that the isolatedelements could, however, represent other early Homo. Ad-ditionally, in Georgia both adult and subadult skeletal ele-ments, some associated, are known. Although little that isspecies specific can be attributed to the postcrania of H.erectus, there is much that differs from Australopithecus. Forexample, H. erectus has enlarged articular surface areas of longbones, thick cortical bone particularly in the lower limb, andan anteroposteriorly flattened (platymeric) femur (Weiden-reich 1941), deep trochlea of the distal femur (Tardieu 1998,1999), double meniscal attachments of the proximal tibia (butsee Dugan and Holliday 2009), reoriented pelves that are per-haps less broad (but see Holliday 2012) but certainly morecapacious (Ruff 2010), a marked iliac pillar (i.e., acetabulo-cristal buttress), and medial torsion of the ischial tuberosity(Day 1971; Rose 1984). Thus, postcranially, H. erectus differsfrom modern humans mostly in primitive characters, someof which are derived relative to nonhuman primates and oth-ers of which may originate at the origin of Homo (e.g., Mc-Henry, Corruccini, and Howell 1976; Trinkaus 1984). Alter-natively, a number of aspects of the postcranial skeleton—including most of the hand and foot—are better known inother early Homo than in H. erectus.

Homo erectus: Size and Proportions

Early H. erectus from Georgia and East Africa are moderatelybigger brained than other early Homo in East Africa. Adultcranial capacity ranges from 638 cm3 to a maximum of 1,067cm3 (tables 5, 6; Holloway 1983; Spoor et al. 2007; Vekua etal. 2002). Several characters scale with cranial capacity, in-cluding cranial vault shape; smaller crania are more globular(see Anton et al. 2007; Spoor et al. 2007). Although theirranges overlap, the Georgian sample is smaller (638–775 cm3)than the African (690–1,067), which may speak to issues ofresource scarcity, extrinsic mortality, or climatic adaptation(seasonality).

Dental and mandibular size are smaller in early H. erectusthan in the entire early Homo group (table 7). However, the

differences are smaller between early H. erectus and the 1813group (excluding KNM-ER 1802 and 1590) and are not sta-tistically significant for individual teeth in any event (Anton2008). Relative molar cusp proportions and molar size rela-tionships do, however, sort early H. erectus from early Homo(Grine et al. 2009; Spoor et al. 2007).

Body size range is quite substantial as well (tables 4, 5, 8).South African and East African H. erectus are similar in size,although there are only a few South African remains for whichbody size can be estimated. The Georgian remains are 17%–24% smaller (40–50 kg, 146–166 cm) on average than earlyEast African H. erectus (51–68 kg, 160–185 cm) dependingon whether the Gona pelvis is included. This difference maybe the result of a categorization bias in Africa that has tendedto place smaller isolated postcrania into early Homo and largerinto H. erectus. The Gona pelvis (Simpson et al. 2008), ifsubstantiated as H. erectus, would lower the body size rangefor Africa to perhaps as little as 120 cm/39.7 kg. In light ofthe small-headed remains from Ileret (Spoor et al. 2007),Olorgesailie (Potts et al. 2004), and Olduvai Gorge (OH 12;Anton 2004), a small-bodied H. erectus in Ethiopia seemsplausible. However, the Gona pelvis is not associated withcranial remains, and Ruff (2010) argues that the pelvis is morelikely Australopithecus or Paranthropus. Without Gona, earlyH. erectus stature estimates range from 145 to 185 cm (5′3′′–6′1′′) and body mass estimates from 40 to 68 kg (88 to 150lb.). It should be noted, however, that although the remainsare fragmentary, Susman, de Ruiter, and Brain (2001) havesuggested that female Homo aff. erectus at Swartkrans may beas small as 30 kg (table 4).

Several lines of evidence suggest that early H. erectus wasan accomplished striding biped with little arboreal locomotionin its repertoire. Hind-limb elongation is present in the as-sociated skeletons from Africa (McHenry and Coffing 2000)and Georgia (Lordkipanidze et al. 2007). Whether this elon-gation began with H. erectus, at the base of the genus, or evenat the base of the hominins, is a matter of debate (see Holliday2012; Pontzer 2012), although it now seems likely that thisis a hominin adaptation. However, the forelimb is at leastsomewhat reduced in length compared with overall body sizein H. erectus and with the condition in Australopithecus andperhaps early Homo. Cross-sectional properties of the hindlimb and forelimb indicate different patterns of strength be-tween Australopithecus afarensis and early H. erectus (Ruff2008, 2009), suggesting that like humans, H. erectus was apredominantly terrestrial biped. The foot of H. erectus, whichis recently known from Dmanisi, supports this notion byshowing evidence of both transverse (metatarsal torsion) andlongitudinal (first metatarsal base width) arches (Lordkipan-idze et al. 2007; Pontzer et al. 2010). The hand skeleton islargely unknown.

The shoulder and trunk appear to exhibit both primitiveand derived conditions. The shoulder girdle retains an inter-mediate condition: The glenoid fossa of the scapula is orientedmore superiorly (Dmanisi and KNM-WT 15000), the clavicle


is relatively short compared with body size (proxied by hu-meral length), and humeral torsion is not as great as in mod-ern humans. Thus, the scapula was likely placed relatively lessdorsally than in recent humans (Larson et al. 2007), whichhas implications for throwing and possibly suggests less sta-bility while running (Larson 2009). Like Australopithecus, thelumbar vertebral bodies of both Dmanisi and Nariokotomeare small relative to body weight, unlike the condition inmodern humans (Latimer and Ward 1993; Lordkipanidze etal. 2007). Additionally, the African remains may suggest theprimitive condition of diaphragmatic placement (Williams2011) even though they appear to retain five lumbar vertebrae,as is the modal condition in humans (Haeusler, Martelli, andBoeni 2002; Haeusler, Schiess, and Boeni 2011; Williams2011).

Other thoracic and pelvic features appear derived. The tho-rax is broad superiorly and narrow inferiorly (based on theangulation of KNM-WT 15000’s ribs to thoracic vertebrae;Jellema, Latimer, and Walker 1993). This shape suggests thatH. erectus had a relatively small gut (Aiello and Wheeler 1995),which has implications for diet quality and possible foragingshifts. Additionally, a number of pelvic features of H. erectushave been argued to be more similar to H. sapiens than to A.afarensis and possibly H. habilis (Ruff and Walker 1993).However, some of these features are from os coxae of indef-inite species attribution, such KNM-ER 3228, and may alsobe seen in non-Homo pelves such as the A. sediba adult, MH2. Furthermore, Holliday (2012) contends that pelvic nar-rowing is not seen until H. sapiens.

Until recently, sexual dimorphism was thought to besmaller in H. erectus than in Australopithecus because of dif-ferentially large H. erectus females (Aiello and Key 2002; Leon-ard and Robertson 1997; McHenry 1992; McHenry and Coff-ing 2000; Ruff 2002; but see Susman, de Ruiter, and Brain2001). Current evidence suggests, however, that cranial andbody mass dimorphism may have been as large as in earlierhominins (Pontzer 2012; Spoor et al. 2007). However, thisconclusion is dependent on the frame of comparison. Plavcan(2012) shows that there is a significant temporal component-to-size variation in Homo. And he shows that when comparedwith intraspecific variation in extant apes, H. erectus size var-iation (and perhaps dimorphism) is not particularly remark-able. To complicate matters, the assemblages make clear thatnot all small-sized individuals are females. In the small-sizedDmanisi population, some cranially robust probable male re-mains (D3444; 638 cm3) are absolutely small (Lordkipanidzeet al. 2006). Similarly, at Olduvai, OH 9 (1,067 cm3) and OH12 (727 cm3) differ greatly in size but not robusticity (i.e.,cranial thickness, superstructure development), which I in-terpret as within-sex variation. Nonetheless, cranial and bodymass CVs are similar to one another (around 15%–19%) andnot substantially different than species of Australopithecus orearly Homo (12%–20%; table 5). Given small and uncertainsamples, the values should not be weighed too heavily; none-

theless, they do not suggest a reduction of dimorphism in H.erectus over other groups.

Given that by 1.8 or 1.9 Ma H. erectus coexists with twogroups of early Homo, the origin of the taxon must predatethis by some time. Craniofacial affinities are strongest withthe 1813 group, suggesting that the A.L. 666 maxilla is onepossible source population.

Summary of Shifts in Homo

Early Homo appears in the record by 2.3 Ma. By 2.0 Ma atleast two facial morphs of early Homo (1813 group and 1470group) representing two different adaptations are present (ta-ble 3). The 1813 group survives until at least 1.44 Ma. EarlyHomo erectus represents a third more derived morph yet andone that is of slightly larger brain and body size but somewhatsmaller tooth size. South African remains of early Homo arepresent; however, they likely represent a separate species fromthose in East Africa (see Grine et al. 2009).

Small cranial remains from Georgia and Africa provideevidence of substantial individual and perhaps populationalsize variation within early H. erectus and indicate overlappingranges of brain size with other early Homo. However, evenwith these new discoveries, H. erectus had a larger range (638–1,067) and average ( cm3) of cranial capacity than didx p 810other early Homo (510–750, cm3; Anton et al. 2007).x p 629Currently, the overlap of cranial ranges is greater than is thebody size overlap; however, this may reflect sampling biasbetween cranial and postcranial remains.

Although the fossil evidence is limited, average body andbrain size increase appears to be an important shift betweenearly Homo and Australopithecus and again between H. erectus(sensu lato) and other early Homo (H. habilis sensu lato).Holliday (2012) and Pontzer (2012) document ∼33% increasein average mass estimates between the genera Australopithecusand early Homo (all taxa from 2.0 to 1.5 Ma inclusive of H.erectus) and about a 10% increase between early non-erectusHomo and Australopithecus. Body size estimates from post-cranial specimens that can be certainly assigned to H. erectusfrom Africa and Georgia yield adult stature estimates betweenabout 145 and 185 cm and adult body mass estimates ofbetween 40 and 65 kg (tables 5, 8; Graves et al. 2010; Lord-kipanidze et al. 2007; McHenry 1992, 1994; Ruff and Walker1993). The lower end of the range may decrease to as littleas 120 cm and 30 kg if the Gona pelvis and Swartkrans post-crania are certainly assigned to H. erectus. The sparser evi-dence for early non-erectus Homo overlaps the lower end ofthis range (118–150 cm and 30–60 kg) but is about 15%smaller than the combined early H. erectus mean (Georgia �Africa) and 37% smaller than the early African H. erectusmean. Dimorphism as proxied by CVs seems no less thanearlier Australopithecus, but the variables used and the scaleof comparison seem to influence the results (table 5; and seePlavcan 2012).

Average differences in body size have implications for life


history and ranging that may be of particular importance toniche differentiation in Africa. The overall larger size of earlyH. erectus and their different patterns of postcranial strengthif not length may indicate larger home range sizes and perhapsmore open habitat for H. erectus, all of which may entailgreater daily energy requirements (Aiello and Key 2002; An-ton, Leonard, and Robertson 2002; Leonard and Robertson1997; Ruff 2009; Steudel-Numbers 2006). Based on life historycorrelates in modern humans (Kuzawa and Bragg 2012), dif-ferences in average body size among and between taxa mayalso signal decreased extrinsic mortality rates and/or increasednutritional sufficiency in the larger-bodied morph, includingsome combination of decreased predator and parasite load orsusceptibility and increased diet quality.

Acknowledgments

I am grateful to many colleagues worldwide for their discus-sions and for access to their specimens and field localities. Iam especially grateful to F. Spoor, M. C. Dean, L. N. Leakey,R. E. F. Leakey, and M. G. Leakey for ongoing discussions onearly Homo. I thank my co-organizer, Leslie Aiello, with whomI have had the great good luck to work closely for nearly 2years on this workshop. I thank Leslie and the workshopparticipants for engaging the topic with such enthusiasm andfor their inspiration and support. I am forever grateful toLaurie Obbink for her life perspective and good will. Reviewercomments improved this work. All errors in fact and judg-ment remain wholly my own. Emily Middleton and the Cur-rent Anthropology staff provided expert editorial help.

References CitedAiello, Leslie C., and Catherine Key. 2002. Energetic consequences of being

a Homo erectus female. American Journal of Human Biology 14:551–565.Aiello, Leslie C., and Peter Wheeler. 1995. The expensive tissue hypothesis:

the brain and digestive system in human and primate evolution. CurrentAnthropology 36:199–221.

Aiello, Leslie C., and Bernard A. Wood. 1994. Cranial variables as predictorsof hominine body mass. American Journal of Physical Anthropology 95:409–426.

Alexeev, Valeri. 1986. The origin of the human race. Moscow: Progress.Anton, Susan C. 2003. Natural history of Homo erectus. Yearbook of Physical

Anthropology 46:126–170.———. 2004. The face of Olduvai Hominid 12. Journal of Human Evolution

46:337–347.———. 2008. Framing the question: diet and evolution in early Homo. In

Primate craniofacial function and biology: papers in honor of Bill Hylander.Christopher J. Vinyard, Christine E. Wall, and Matthew J. Ravosa, eds. Pp.443–482. New York: Springer Science.


Anton, Susan C., Fred Spoor, Connie D. Fellmann, and Carl C. Swisher III.2007. Defining Homo erectus: size considered. In Handbook of paleoanthro-pology, vol. 3. Winfried Henke and Ian Tattersall, eds. Pp. 1655–1693. Berlin:Springer.

Arambourg, Camille. 1954. L’hominien fossile de Ternifine (Algerie). ComptesRendus de l’Academie des Sciences 239:893–895.

Asfaw, Berhane, Yonas Beyene, Gen Suwa, Robert C. Walter, Tim D. White,Giday WoldeGabriel, and Yohannes Yemene. 1992. The earliest Acheuleanfrom Konso-Gardula. Nature 360:732–735.

Asfaw, Berhane, William H. Gilbert, Yonas Beyene, William K. Hart, Paul R.Renne, Giday WoldeGabriel, Elizabeth S. Vrba, and Tim D. White. 2002.Remains of Homo erectus from Bouri, Middle Awash, Ethiopia. Nature 416:317–320.

Baab, Karen L. 2008. The taxonomic implications of cranial shape variationin Homo erectus. Journal of Human Evolution 54:827–847.

Berger, Lee R., Darryl de Ruiter, Steven E. Churchill, Peter Schmid, KristianJ. Carlson, Paul H. G. M. Dirks, and Job M. Kibii. 2010. Australopithecussediba: a new species of Homo-like australopith from South Africa. Science328:195–204.

Bilsborough, Alan, and Bernard A. Wood. 1988. Cranial morphometry of earlyhominids: facial region. American Journal of Physical Anthropology 76:61–86.

Black, Davidson. 1927. Further hominid remains of Lower Quaternary agefrom Chou Kou Tien deposit. Nature 120:954.


Boaz, Noal T., and F. Clark Howell. 1977. A gracile hominid cranium fromupper Member G of the Shungura Formation, Ethiopia. American Journalof Physical Anthropology 46:93–108.

Bromage, Timothy G., James J. McMahon, J. Frances Thackeray, Ottmar Kull-mer, Russel Hogg, Alfred L. Rosenberger, Friedemann Schrenk, and DonaldH. Enlow. 2008. Craniofacial architectural constraints and their importancefor reconstructing the early Homo skull KNM-ER 1470. Journal of ClinicalPaediatric Dentistry 33:43–54.

Bromage, Timothy G., Friedemann Schrenk, and Franz W. Zonneveld. 1995.Paleoanthropology of the Malawi Rift: an early hominid mandible fromthe Chiwondo Beds, northern Malawi. Journal of Human Evolution 28:71–108.

Broom, Robert, and John T. Robinson. 1949. A new type of fossil man. Nature164:322–323.

Clarke, Ron J. 2012. A Homo habilis maxilla and other newly-discoveredhominid fossils from Olduvai Gorge, Tanzania. Journal of Human Evolution63(2):418–428.

Clarke, Ron J., F. Clark Howell, and C. K. Brain. 1970. New finds at theSwartkrans australopithecine site: more evidence of an advanced hominidat Swartkrans. Nature 225:1219–1222.

Collard, Mark, and Bernard A. Wood. 2007. Defining the genus Homo. InHandbook of paleoanthropology, vol. 3. Winfried Henke and Ian Tattersall,eds. Pp. 1575–1610. Berlin: Springer.

Curnoe, Darren. 2010. A review of early Homo in southern Africa focusingon cranial, mandibular and dental remains, with the description of a newspecies (Homo gautengensis sp. nov.). Homo 61:151–177.

Curnoe, Darren, and Phillip V. Tobias. 2006. Description, new reconstruction,comparative anatomy, and classification of the Sterkfontein Stw 53 cranium,with discussions about the taxonomy of other southern African early Homoremains. Journal of Human Evolution 50:36–77.

Day, Michael H. 1971. Postcranial remains of Homo erectus from Bed IV,Olduvai Gorge, Tanzania. Nature 232:383–387.

Dean, M. Christopher, Meave G. Leakey, Don Reid, Friedemann Schrenk,Gary T. Schwartz, Christopher Stringer, and Alan Walker. 2001. Growthprocesses in teeth distinguish modern humans from Homo erectus andearlier hominins. 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. Vertebrate Paleontology and Pa-leoanthropology Series. Berlin: Springer.

Dean, M. Christopher, and Bernard A. Wood. 1982. Basicranial anatomy ofPlio-Pleistocene hominids from East and South Africa. American Journal ofPhysical Anthropology 59:157–174.

Dubois, Eugene. 1894. Pithecanthropus erectus: eine menschenaehnlichUebergangsform aus Java. Batavia: Landsdrukerei.

Dugan, James, and Trenton W. Holliday. 2009. Utility of the lateral meniscalnotch in distinguishing hominin taxa. Journal of Human Evolution 57:773–776.

Feibel, Craig S., Frank H. Brown, and Ian McDougall. 1989. Stratigraphiccontext of fossil hominids from the Omo group deposits: northern TurkanaBasin, Kenya and Ethiopia. American Journal of Physical Anthropology 78:595–622.

Ferguson, Walter W. 1995. A new species of the genus Homo (primates:


Hominidae) from the Plio/Pleistocene of Koobi Fora, in Kenya. Primates36:69–89.

Gabunia, Leo, Susan C. Anton, David Lordkipanidze, Abesalom Vekua, CarlC. Swisher, and Antje Justus. 2001. Dmanisi and dispersal. EvolutionaryAnthropology 10:158–170.

Gabunia, Leo, Marie-Antoinette de Lumley, Abesalom Vekua, David Lord-kipanidze, and Henry de Lumley. 2002. Decouverte d’un nouvel hominidea Dmanisi (Transcaucasie, Georgie). Comptes Rendu Palevol 1:242–253.


Gebo, Daniel L., and Gary T. Schwartz. 2006. Foot bones from Omo: impli-cations for hominid evolution. American Journal of Physical Anthropology129:499–511.

Graves, Rhonda R., Amy C. Lupo, Robert C. McCarthy, Daniel J. Wescott,and Deborah L. Cunningham. 2010. Just how strapping was KNM-WT15000? Journal of Human Evolution 59:542–554.

Grine, Frederick E. 2001. Implications of morphological diversity in earlyHomo crania from eastern and southern Africa. In Humanity from Africannaissance to coming millennia: colloquia in human biology and palaeoan-thropology. Phillip V. Tobias, Michael A. Raath, Jacopo Moggi-Cecchi, andGerald A. Doyle, eds. Pp. 107–115. Florence: Florence University Press.

———. 2005. Early Homo at Swartkrans, South Africa: a review of the evi-dence and an evaluation of recently proposed morphs. South African Journalof Science 101:43–52.

Grine, Frederick E., Heather F. Smith, Christopher P. Heesy, and Emma J.Smith. 2009. Phenetic affinities of Plio-Pleistocene Homo fossils from SouthAfrica: molar cusp proportions. In The first humans: origin and early evo-lution of the genus Homo. Frederick E. Grine, John G. Fleagle, and RichardE. Leakey, eds. Pp. 49–62. Vertebrate Paleontology and PaleoanthropologySeries. Berlin: Springer.

Groves, Colin P., and Vratislav Mazek. 1975. An approach to the taxonomyof the Hominidae: gracile Villafranchian hominids of Africa. Caspos proMineralogii a Geologii 20:225–247.

Haeusler, Martin, Sandra A. Martelli, and Thomas Boeni. 2002. Vertebraenumbers of the early hominid lumbar spine. Journal of Human Evolution43:621–643.

Haeusler, Martin, and Henry M. McHenry. 2004. Body proportions of Homohabilis reviewed. Journal of Human Evolution 46:433–465.

———. 2007. Evolutionary reversals of limb proportions in early hominids?evidence from KNM-ER 3735 (Homo habilis). Journal of Human Evolution53:383–405.

Haeusler, Martin, Regula Schiess, and Thomas Boeni. 2011. New vertebraland rib material point to modern bauplan of the Nariokotome Homo erectusskeleton. Journal of Human Evolution 61:575–582.

Harcourt-Smith, William E. H. 2007. The origins of bipedal locomotion. InHandbook of paleoanthropology, vol. 3. Winfried Henke and Ian Tattersall,eds. Pp. 1483–1518. Berlin: Springer.

Harcourt-Smith, William E. H., and Leslie C. Aiello. 2004. Fossils, feet andthe evolution of bipedal locomotion. Journal of Anatomy 5:403–416.

Heberer, Gerhard. 1963. Ueber einen neuen archanthropin Typus aus derOldoway-Schlucht. Zeitschrift fur Morphologie und Anthropologie 53:171–177.

Hill, Andrew, Steven Ward, Alan Deino, Garniss Curtis, and Robert Drake.1992. Earliest Homo. Nature 355:719–722.


Holloway, Ralph. 1983. Human paleontological evidence relevant to languagebehavior. Human Neurobiology 2:105–114.

Holloway, Ralph L., and Michael S. Yuan. 2004. Endocranial morphology ofA.L. 444-2. In The skull of Australopithecus afarensis. William H. Kimbel,Yoel Rak, and Donald C. Johanson, eds. Pp. 123–125. New York: OxfordUniversity Press.

Howell, F. Clark. 1978. Hominidae. In Evolution of African mammals. VincentJ. Maglio and H. B. S. Cooke, eds. Pp. 154–248. Cambridge, MA: HarvardUniversity Press.

Hughes, Alun, and Phillip V. Tobias. 1977. A fossil skull probably of the genusHomo from Sterkfontein, Transvaal. Nature 265:310–312.

Indriati, Etty, and Susan C. Anton. 2008. Earliest Indonesian facial and dentalremains from Sangiran, Java: a description of Sangiran 27. AnthropologicalScience 116:219–229.

Indriati, Etty, Carl C. Swisher III, Christopher Lepre, Rhonda L. Quinn, Rusyad

Suriyanto, Agus T. Hascaryon, Rainer Grun, et al. 2011. The age of the 20Meter Solo River Terrace, Java, Indonesia and the survival of Homo erectusin Asia. PLoS ONE 6(6):e21562, doi:10.1371/journal.pone.0021562.

Jellema, Lyman M., Bruce Latimer, and Alan Walker. 1993. The rib cage. InThe Nariokotome Homo erectus skeleton. Alan Walker and Richard Leakey,eds. Pp. 294–325. Cambridge, MA: Harvard University Press.

Johanson, Donald C., Fidelis T. Masao, Gerry G. Eck, Tim D. White, RobertC. Walter, William H. Kimbel, Berhane Asfaw, Paul Manega, Prosper Ndes-sokia, and Gen Suwa. 1987. New partial skeleton of Homo habilis fromOlduvai Gorge, Tanzania. Nature 327:205–209.

Kappelman, John. 1996. The evolution of body mass and relative brain sizein fossil hominids. Journal of Human Evolution 30:243–276.

Kennedy, Gail E. 1983a. A morphometric and taxonomic assessment of ahominine femur from the lower member Koobi Fora, Lake Turkana. Amer-ican Journal of Physical Anthropology 61:429–431.

———. 1983b. Some aspects of the evolution of fossil and recent homininefemora. Journal of Human Evolution 12:587–616.

Kimbel, William H. 2009. The origin of Homo. In The first humans: originand early evolution of the genus Homo. Frederick E. Grine, John G. Fleagle,and Richard E. Leakey, eds. Pp. 31–38. Vertebrate Paleontology and Pa-leoanthropology Series. Dordrecht: Springer.

Kimbel, William H., Donald C. Johanson, and Yoel Rak. 1997. Systematicassessment of a maxilla of Homo from Hadar, Ethiopia. American Journalof Physical Anthropology 103:235–262.

Kimbel, William H., and Yoel Rak. 1993. The importance of species taxa inpaleoanthropology and an argument for the phylogenetic species concept.In Species, species concepts, and primate evolution. William Kimbel andLawrence Martin, eds. Pp. 461–484. New York: Plenum.

Kimbel, William H., Robert C. Walter, Donald C. Johanson, Kaye E. Reed,James L. Aronson, Z. Assefa, Curtis W. Marean, et al. 1996. Late PlioceneHomo and Oldowan tools from the Hadar Formation (Kada Hadar Mem-ber), Ethiopia. Journal of Human Evolution 31:549–561.

Korey, Kenneth A. 1990. Deconstructing reconstruction: the OH 62 humer-ofemoral index. American Journal of Physical Anthropology 83:25–33.

Kuman, Kathleen, and Ron J. Clarke. 2000. Stratigraphy, artefact industriesand hominid associations for Sterkfontein, Member 5. Journal of HumanEvolution 38:827–847.


Larson, Susan G. 2009. Evolution of the hominin shoulder: early Homo. InThe first humans: origin and early evolution of the genus Homo. FrederickE. Grine, John G. Fleagle, Richard E. Leakey, eds. Pp. 65–75. Dordrecht:Springer.

Larson, Susan G., William L. Jungers, Michael Morwood, Thomas Sutikna,Jatmiko, E. Wahyu Saptomo, Rokus Awe Due, and Tony Djubiantono. 2007.Homo floresiensis and the evolution of the hominin shoulder. Journal ofHuman Evolution 53:718–731.

Latimer, Bruce, and Carol V. Ward. 1993. The thoracic and lumbar vertebrae.In The Nariokotome Homo erectus skeleton. Alan Walker and Richard Lea-key, eds. Pp. 266–293. Cambridge, MA: Harvard University Press.

Leakey, Louis S. B., Phillip V. Tobias, and John Napier. 1964. A new speciesof the genus Homo from Olduvai Gorge. Nature 202:7–9.

Leakey, Meave G., Fred Spoor, M. Christopher Dean, Craig S. Feibel, SusanC. Anton, Christopher Kiarie, and Louise N. Leakey. 2012. New early Homofossils from Koobi Fora, northern Kenya, highlight facial diversity. Nature488:201–204.

Leakey, Mary D., Ron J. Clarke, and Louis S. B. Leakey. 1971. New hominidskull from Bed I, Olduvai Gorge, Tanzania. Nature 233:317–323.

Leakey, Richard E., and Alan Walker. 1985. Further hominids from the Plio-Pleistocene of Koobi Fora, Kenya. American Journal of Physical Anthropology67:135–163.

Leakey, Richard E., Alan Walker, Carol V. Ward, and Hannah M. Grausz. 1989.A partial skeleton of a gracile hominid from the Upper Burgi Member ofthe Koobi Fora Formation, east Lake Turkana, Kenya. In Hominidae: pro-ceedings of the 2nd International Congress of Human Paleontology. GiacomoGiacobini, ed. Pp. 167–173. Milan: Jaca.

LeGros Clark, William E. 1964. The fossil evidence for human evolution. Chi-cago: University of Chicago Press.


Lieberman, Daniel E., David R. Pilbeam, and Bernard A. Wood. 1988. A


probabilistic approach to the problem of sexual dimorphism in Homo ha-bilis: a comparison of KNM-ER 1470 and KNM-ER 1813. Journal of HumanEvolution 17:503–512.

Lordkipanidze, David, Tea Jashashvili, Abesalom Vekua, Marcia S. Ponce deLeon, Christophe P. E. Zollikofer, G. Philip Rightmire, Herman Pontzer, etal. 2007. Postcranial evidence from early Homo from Dmanisi, Georgia.Nature 449:305–310.

Lordkipanidze, David, Abesalom Vekua, Reid Ferring, G. Philip Rightmire,Christophe P. E. Zollikofer, Marcia S. Ponce de Leon, and Jordi Agusti.2006. A fourth hominin skull from Dmanisi, Georgia. Anatomical Record288A:1146–1157.

Martyn, John, and Philip V. Tobias. 1967. Pleistocene deposits and new fossillocalities in Kenya. Nature 215:476–479.

McHenry, Henry M. 1991. Petite bodies of the “robust” australopithecines.American Journal of Physical Anthropology 86:445–454.

———. 1992. Body size and proportions in early hominids. American Journalof Physical Anthropology 87:407–431.

———. 1994. Behavioral ecological implications of early hominid body size.Journal of Human Evolution 27:77–87.


McHenry, Henry M., Robert S. Corruccini, and F. Clark Howell. 1976. Analysisof an early hominid ulna from the Omo Basin, Ethiopia. American Journalof Physical Anthropology 44:295–304.

Moggi-Cecchi, Jacopo, Frederik E. Grine, and Philip V. Tobias. 2006. Earlyhominid dental remains from Members 4 and 5 of the Sterkfontein For-mation (1966–1996 excavations): catalogue, individual associations, mor-phological descriptions and initial metrical analysis. Journal of Human Evo-lution 50:239–328.

Moggi-Cecchi, Jacopo, Philip Tobias, and A. Beynon. 1998. The mixed den-tition and associated skull fragments of a juvenile fossil hominid fromSterkfontein, South Africa. American Journal of Physical Anthropology 106:425–465.

Neubauer, Simon, Philip Gunz, Gerhard W. Weber, and Jean-Jacques Hublin.2012. Endocranial volume of A. africanus: new CT based estimates and theeffects of missing data and small sample size. Journal of Human Evolution62:498–510.

Ohman, James C., C. Wood, Bernard Wood, Robin Crompton, Michael Gun-ter, Yu Li, Russell Savage, and Weijie Wang. 2002. Stature-at-death of KNM-WT 15000. Human Evolution 17:129–142.

Openoorth, W. F. F. 1932. Homo (Javanthropus) soloensis, een pleistoceenemensch van Java. Weteschappelijke Mededeelingen: Dienst van den Mijnbouwin Nederlandisch-Indie 20:49–74.

Plavcan, J. Michael. 2012. Body size, size variation, and sexual size dimorphismin early Homo. Current Anthropology 53(suppl. 6):S409–S423.


Pontzer, Herman, David A. Raichlen, Robert W. Shumaker, Cara Ocobock,and Serge A. Wich. 2010. Metabolic adaptation for low energy throughputin orangutans. Proceedings of the National Academy of Sciences of the USA107(32):14048–14052.

Potts, Richard. 1988. Environmental hypotheses of human evolution. Yearbookof Physical Anthropology 41:93–136.


Prat, Sandrine, J. Jean-Philip Brugal, Jean-Jacques Tiercelin, Jean-Alix Barrat,Marcel Bohn, Anne Delagnes, Sonia Harmand, et al. 2005. First occurrenceof early Homo in the Nachukui Formation (West Turkana, Kenya) at 2.3–2.4 Myr. Journal of Human Evolution 49:230–240.

Reno, Philip L., David DeGusta, Maria A. Serrat, Richard S. Meindl, Tim D.White, Robert B. Eckhardt, Adam J. Kuperavage, Karol Galik, and C. OwenLovejoy. 2005. Plio-Pleistocene hominid limb proportions: evolutionaryreversals of estimation errors? Current Anthropology 46:575–588.

Richmond, Brian G., Leslie C. Aiello, and Bernard A. Wood. 2002. Earlyhominin limb proportions. Journal of Human Evolution 43:529–548.

Rightmire, G. Philip. 1979. Cranial remains of Homo erectus from Beds II andIV, Olduvai Gorge, Tanzania. American Journal Physical Anthropology 51:99–116.

———. 1990. The evolution of Homo erectus. Cambridge: Cambridge Uni-versity Press.

Rightmire, G. Philip, and David Lordkipanidze. 2009. Comparisons of early

Pleistocene skulls from East Africa and the Georgian Caucasus: evidencebearing on the origin and systematic of genus Homo. In The first humans:origin and early evolution of the genus Homo. Frederick E. Grine, John G.Fleagle, and Richard E. Leakey, eds. Pp. 39–48. Vertebrate Paleontology andPaleoanthropology Series. Dordrecht: Springer.

Rightmire, G. Philip, David Lordkipanidze, and Abesalom Vekua. 2006. An-atomical descriptions, comparative studies and evolutionary significance ofthe hominin skulls from Dmanisi, Republic of Georgia. Journal of HumanEvolution 50:115–141.

Robinson, John T. 1953. The nature of Telanthropus capensis. Nature 171:33.———. 1961. The australopithecines and their bearing on the origin of man

and of stone tool-making. South African Journal of Science 57:3–13.Rose, Michael D. 1984. A hominine hip bone, KNM-ER 3228, from east Lake

Turkana, Kenya. American Journal of Physical Anthropology 63:371–378.Ruff, Christopher B. 2002. Variation in human body size and shape. Annual

Review of Anthropology 31:211–232.———. 2008. Femoral/humeral strength in early African Homo erectus. Jour-

nal of Human Evolution 54:383–390.———. 2009. Relative limb strength and locomotion in Homo habilis. Amer-

ican Journal of Physical Anthropology 138:90–100.———. 2010. Body size and body shape in early hominins: implications of

the Gona pelvis. Journal of Human Evolution 58:166–178.Ruff, Christopher B., Erik Trinkaus, and Trenton W. Holliday. 1997. Body

mass and encephalization in Pleistocene Homo. Nature 387:173–176.Ruff, Christopher B., and Alan Walker. 1993. Body size and body shape. In

The Nariokotome Homo erectus skeleton. Alan Walker and Richard Leakey,eds. Cambridge, MA: Harvard University Press.

Schrenk, Friedemann, Ottmar Kullmer, and Timothy Bromage. 2007. Theearliest putative Homo fossils. In Handbook of paleoanthropology, vol. 3.Winfried Henke and Ian Tattersall, eds. Pp. 1611–1631. Berlin: Springer.

Shen, Guanjen, Xing Gao, Bin Gao, and Darryl E. Granger. 2009. Age ofZhoukoudian Homo erectus determined with 26Al/10Be burial dating. Nature458:198–200.


Simpson, Scott W., Jaye Quade, Naomi E. Levin, Robert Butler, GuillaumeDupont-Nivet, Melanie Everett, and Sileshi Semaw. 2008. A female Homoerectus pelvis from Gona, Ethiopia. Science 322:1089–1092.

Spoor, Fred. 1993. The comparative morphology and phylogeny of the humanbony labyrinth. PhD dissertation, Utrecht University.

Spoor, Fred, Meave G. Leakey, Patrick N. Gathogo, Frank H. Brown, SusanC. Anton, Ian McDougall, Christopher Kiarie, F. Kyalo Manthi, and LouiseN. Leakey. 2007. Implications of new early Homo fossils from Ileret, eastof Lake Turkana, Kenya. Nature 448:688–691.

Steudel-Numbers, Karen L. 2006. Energetics in Homo erectus and other earlyhominins: the consequences of increased lower-limb length. Journal of Hu-man Evolution 51:445–453.

Stringer, Christopher B. 1986. The credibility of Homo habilis. In Topics inprimate and human evolution. Bernard A. Wood, Lawrence B. Martin, andPeter Andrews, eds. Pp. 266–294. New York: Liss.

Susman, Randal L., and Norman Creel. 1979. Functional and morphologicalaffinities of the subadult hand (O.H. 7) from Olduvai Gorge. AmericanJournal of Physical Anthropology 51:311–332.

Susman, Randal L., Darryl de Ruiter, and C. K. Brain. 2001. Recently identifiedpostcranial remains of Paranthropus and early Homo from Swartkrans cave,South Africa. Journal of Human Evolution 41:607–629.

Suwa, Gen, Tim D. White, and F. Clark Howell. 1996. Mandibular postcaninedentition from the Shungura Formation, Ethiopia: crown morphology, tax-onomic allocations, and Plio-Pleistocene hominid evolution. American Jour-nal of Physical Anthropology 101:247–282.

Swisher, Carl C., III, Garniss H. Curtis, Teuku Jacob, Ann G. Getty, AgusSuprijo, and Widiasmoro. 1994. Age of the earliest known hominids inJava, Indonesia. Science 263:1118–1121.

Swisher, Carl C., III, W. Jack Rink, Susan C. Anton, Henry P. Schwarcz, GarnissH. Curtis, Agus Suprijo, and Widiasmoro. 1996. Latest Homo erectus, inJava: potential contemporaneity with Homo sapiens in southeast Asia. Sci-ence 274:1870–1874.

Tardieu, Christine. 1998. Short adolescence in early hominids: infantile andadolescent growth of the human femur. American Journal of Physical An-thropology 107:163–178.

———. 1999. Ontogeny and phylogeny of femoro-tibial characters in humansand hominid fossils: functional influence and genetic determinism. Amer-ican Journal of Physical Anthropology 110:365–377.


Tobias, Phillip V. 1991. The skulls, endocasts and teeth of Homo habilis, vol.4 of Olduvai Gorge. Cambridge: Cambridge University Press.

Tocheri, Matthew W., Caley Orr, Susan G. Larson, Thomas Sutikna, Jatmiko,E. Wahyu Saptomo, Rokus Awe Due, Tony Djubiantono, Michael J. Mor-wood, and William L. Jungers. 2007. The primitive wrist of Homo floresiensisand its implications for hominin evolution. Science 317:1743–1745.

Trinkaus, Erik. 1984. Does KNM-ER 1481 establish Homo erectus at 2.0 myB.P.? American Journal of Physical Anthropology 64:137–139.

Vekua, Abesalom, David Lordkipanidze, G. Philip Rightmire, Jordi Agusti,Reid Ferring, Givi Maisuradze, Alexander Mouskhelishvili, et al. 2002. Anew skull of early Homo from Dmanisi Georgia. Science 297:85–89.

Von Koengiswald, G. H. Ralph. 1936. Ein fossiler hominide aus dem Alt-pleistocan Ostjavas. De Ingenieur in Nederlandsch-Indie 8:149–158.

Walker, Alan. 1981. The Koobi Fora hominids and their bearing on the originsof the genus Homo. In Homo erectus: papers in honor of Davidson Black.Becky A. Sigmon and Jerry S. Cybulski, eds. Pp. 63–94. Cambridge, MA:Harvard University Press.

Walker, Alan, and Richard Leakey, eds. 1993. The Nariokotome Homo erectusskeleton. Cambridge, MA: Harvard University Press.

Walker, Alan, Michael R. Zimmerman, and Richard E. F. Leakey. 1982. Apossible case of hypervitaminosis A in Homo erectus. Nature 296:248–250.

Weidenreich, Franz. 1941. The extremity bones of Sinanthropus pekinensis.Palaeontologia Sinica, n.s. D, no. 5. Peking: Geological Survey of China.

———. 1943. The skull of Sinanthropus pekinensis. Palaeontologia Sinica, n.s.D, no. 10. Pehpei, Chungking: Geological Survey of China.

Williams, Scott A. 2011. Modern or distinct axial bauplan in early hominins?comments on Haeuesler et al. 2011. Journal of Human Evolution, doi:10.1016/j.jhevol.2012.01.007.

Wood, Bernard A. 1985. Early Homo in Kenya and its systematic relationships.In Ancestors: the hard evidence. Eric Delson, ed. Pp. 206–214. New York:Liss.

———. 1991. Hominid cranial remains, vol. 4 of Koobi Fora research project.Oxford: Clarendon.

———. 1992. Origin and evolution of the genus Homo. Nature 355:783–790.———. 2009. Where does the genus Homo begin, and how would we know?

In The first humans: origin and early evolution of the genus Homo. FrederickE. Grine, John G. Fleagle, and Richard E. Leakey, eds. Pp. 17–28. VertebratePaleontology and Paleoanthropology Series. Dordrecht: Springer.

Wood, Bernard, and Jennifer Baker. 2011. Evolution in the genus Homo.Annual Review of Ecology Evolution and Systematics 42:47–69.

Wood, Bernard, and Mark C. Collard. 1999. The human genus. Science 284:65–71.

Wood, Bernard, and Paul Constantino. 2007. Paranthropus boisei: fifty yearsof fossil evidence and analysis. Yearbook of Physical Anthropology 50:106–132.

Wood, Bernard, and Brian Richmond. 2000. Human evolution: taxonomy andpaleobiology. Journal of Anatomy 197:19–60.



Environmental and Behavioral EvidencePertaining to the Evolution of Early Homo

by Richard Potts

East African paleoenvironmental data increasingly inform an understanding of environmental dynamics. This un-derstanding focuses less on habitat reconstructions at specific sites than on the regional trends, tempo, and amplitudesof climate and habitat change. Sole reliance on any one indicator, such as windblown dust or lake sediments, givesa bias toward strong aridity or high moisture as the driving force behind early human evolution. A synthesis ofgeological data instead offers a new paleoenvironmental framework in which alternating intervals of high and lowclimate variability provided the dynamic context in which East African Homo evolved. The Oldowan behavioralrecord presents further clues about how early Homo and Homo erectus responded to East African environmentalchange. Shifting conditions of natural selection, which were triggered by climatic variability, helped shape theadaptability of Oldowan hominins. Together, the behavioral and environmental evidence indicates the initial adaptivefoundation for the dispersal of H. erectus and the persistence of Homo. In particular, overall dietary expansion madepossible by the making and transport of stone tools compensated for increased locomotor and foraging costs andprovided effective behavioral-ecological responses to resource instability during the early evolution of Homo.

The interval from ∼3.0 to 1.5 million years ago (Ma) broadlydefines when Homo originated and Homo erectus first evolvedand dispersed beyond Africa (Anton 2003; Anton and Swisher2004; Kimbel 2009; Kimbel et al. 1996; Pickering et al. 2011).Despite small samples of hominin fossils, particularly in theinterval between 3 and 2 Ma, there is a growing body ofAfrican climatic and paleohabitat data relevant to the earlyevolution of Homo (e.g., Cerling et al. 2011; deMenocal 2011;Potts 2007; Trauth et al. 2007). This time interval in Africaalso preserves the oldest definite stone tool flaking and thespread of innovations such as carcass processing, overall di-etary expansion, and the transport of resources across ancientlandscapes (de Heinzelin et al. 1999; Domınguez-Rodrigo etal. 2005; Plummer 2004; Potts 1991; Semaw et al. 2003). Thegoals of this paper are (1) to characterize the paleoclimateand overall environmental dynamics in which early Homoevolved, (2) to consider how well the prevailing environ-mental hypotheses of human evolution explain the adapta-tions of early Homo, and (3) to examine archaeologically vis-ible behaviors as part of the emerging adaptability of Homo.

Paleoenvironmental research on human evolution has longemphasized the reconstruction of habitat, where the aim isto portray the main type of vegetation or climatic conditionin which early hominins lived. However, the leading edge of

Richard Potts is Director of the Human Origins Program, NationalMuseum of Natural History, Smithsonian Institution (P.O. Box37012, Washington, DC 20013-7012, U.S.A. [[email protected]]). Thispaper was submitted 12 XII 11, accepted 10 VII 12, and electronicallypublished 27 XI 12.

paleoenvironmental research has changed in recent years to-ward a focus on environmental dynamics, that is, the natureand tempo of environmental change that resulted from cli-mate variability, specifically the nonlinear interaction of in-solation cycles, and the episodic effects of faulting and vol-canism (e.g., deMenocal 1995, 2004; Feibel 1997; Potts 1998a,2007; Trauth et al. 2010). The interaction of these factors isespecially apparent in the paleoenvironmental records of theEast African Rift System.

Throughout the period of hominin evolution, environ-mental dynamics inevitably altered the local and regionalabundance of water, vegetation, and food sources. This re-shaping of the overall landscape and of the time and spacedistribution of resources had a pervasive influence on eco-logical opportunities, competition, mortality, and reproduc-tive success. It also stimulated repeated population divergenceand coalescence, the subsequent degree of allopatry of pop-ulations, and thus eventual speciation in hominins and otherorganisms (e.g., Potts 1996b, 1998b; Vrba 1985, 1995b).

A rich array of paleoenvironmental data sets offers timesequences of regional resource opportunities and stresses thatinfluenced the prospects for lineage divergence and the ben-efits and costs of adaptive strategies (ecological, social, de-velopmental, and reproductive) associated with the origin andearly evolution of Homo. Although the fossil records of south-ern and other regions of Africa offer a rich body of evidenceof hominin evolution, Plio-Pleistocene climate and vegetationdata for these parts of the continent typically offer only short-term snapshots or combine lengthy or unknown periods oftime (i.e., they are highly time averaged); in addition, they



Figure 1. Oxygen isotope curve (d18O) for the past 10 Myr (data from Zachos et al. 2001). Arrow 1, beginning around 6 Ma, warm-cold climate fluctuation became more dramatic; compare pre– with post–6 Ma intervals. Arrow 2, beginning 3.0–2.8 Ma, glacialfluctuations strengthened and included the onset of Northern Hemisphere glaciations. Note that the interval 3.0–2.4 Ma wascharacterized by both overall cooling and an increase in the amplitude of oscillation. Arrow 3, from the mid-Pliocene through thePleistocene, the range of climate variability increased dramatically; this observation suggests that organismal features that heightenedthe ability to adjust to ecological dynamics and uncertainty (i.e., adaptability) were at a premium. The genus Homo, and eventuallythe adaptations characteristic of Homo sapiens, evolved during the strongest fluctuations. A color version of this figure is availablein the online edition of Current Anthropology.

are patchily distributed in time and space and often less pre-cisely calibrated than those in East Africa. For this combi-nation of reasons, the Plio-Pleistocene environmental syn-thesis developed here is mainly drawn from and is relevantto the fossil-rich, well-calibrated stratigraphic records of EastAfrica.

Environmental Dynamics in WhichEarly Homo Evolved

In this paper, climate and vegetation data for the intervalfrom 3.0 to 1.5 Ma come from several main sources: stableisotopes, eolian dust and plant biomarker records, northeastAfrican sapropels, and East African lake sediments. By com-bining the variety of environmental data sets, I briefly sum-marize here four principal developments in the East Africanand global environmental system during the evolution of earlyHomo and Homo erectus.

The Gradual Onset of Continental Ice Sheets (∼3.1–2.5 Ma)Denotes the Development of a Periodically Cooler,Drier, and Glaciated Planet

Heating of Earth’s surface, the amount and distribution ofmoisture, and the strength of ocean currents that redistributeheat are all influenced by solar insolation—the amount ofsolar radiation reaching marine and land surfaces (Ruddiman2001). Insolation is regulated by three large-scale orbital cy-cles: “eccentricity” (the shape of Earth’s orbit around the sun;

∼100,000-yr [100-kyr] and ∼413-kyr periods), “obliquity”(the angle of Earth’s axis of rotation relative to the sun; ∼41-kyr period), and “precession” (an effect of Earth’s axial wobblethat creates a progression of the seasons relative to how closethe planet is to the sun; ∼19–23-kyr periods). Interactionsamong these cycles and with millennial-scale and shorter-termsources of variability create a dynamic system of insolationamplification and damping prone to both predictable cyclesand nonlinear threshold-type change, especially over thecourse of the past 3 million years (Myr), since the onset ofNorthern Hemisphere glaciation.

Measurement of stable oxygen isotopes (d18O) in oceanbenthic foraminifera provides a global record of the trend,amplitude, and periodicity in temperature and glacial ice var-iability (fig. 1). Because glaciation also lowers sea level world-wide, glacial conditions (high d18O) can also reduce the mois-ture (largely originating from the ocean) that reachescontinental interiors.

Between 2.8 and 2.4 Ma, an important shift took place inthe dominant period of climate oscillation, from the 19–23-kyr mode to the 41-kyr mode, evident in d18O variability. Thisshift in the dominant periodicity coincided with an increasein climate variability. Another notable shift in the dominantmode, from 41- to 100-kyr periodicity, occurred around 0.8Ma, accompanied again by an increase in amplitude. Theevolution of mid- and late Pleistocene Homo, including Homosapiens, was associated with the highest-amplitude fluctua-tions in marine d18O (fig. 1).

Potts Environments of Early Homo S301

A Net Increase in Aridity and C4 Grasses Occurred over thePast 4 Myr, with a Large Increase Possibly ∼2.8 Ma andAnother ∼1.8 Ma

Continental sedimentary records indicate considerable dryingof East Africa and expansion of grasslands around 2.8–2.5Ma and between 2.0 and 1.7 Ma. These aridity pulses andthe timing of substantial increases in grass abundance havebest been documented by three types of data: eolian dustobtained from deep-sea cores off the northeast coast of Africa(deMenocal 1995, 2004), molecular plant biomarkers ob-tained from these same drill cores (Feakins, deMenocal, andEglinton 2005), and the isotopic composition of East Africasoil carbonates associated with early hominin sites (e.g., Cer-ling 1992; Cerling et al. 2011).

Drill cores obtained from the Atlantic and northwest IndianOceans have provided nearly continuous records of the pro-duction and atmospheric transport of mineral dust from theAfrican continent. Large rainfall seasonality generates thedust. Summer eolian dust plumes in northeast Africa, whichare tied to Indian Ocean monsoonal surface winds, are ex-ported to the Arabian Sea and the Gulf of Aden (Clemens1998; deMenocal 2004). Measurement of wind-borne dust indrill cores from these areas shows that large-amplitude ariditycycles were associated with the onset of Northern Hemisphereglacial cycles around 2.8 Ma and that major shifts in wind-blown dust variability occurred at 2.8–2.6 Ma, 1.8–1.6 Ma,and again at 1.0–0.8 Ma (deMenocal 1995, 2004, 2011). Onthis basis, deMenocal has emphasized the importance of thearidity trend in African climate, which was superimposed onwet-dry cycles. Fossil pollen recovered from terrestrial settingsand marine drill cores are consistent with the substantial arid-ity intervals, especially between 1.8 and 1.6 Ma. Drier vege-tation is recorded in Rift Valley lowlands and across northwestAfrica by 2.4 Ma, and arid vegetation intensified ∼1.8 Ma(Bonnefille 1995; Leroy and Dupont 1994).

Deep-sea records of African dust are complemented by theanalysis of terrestrial plant biomarkers, which are derivedfrom waxy lipids abraded from plant leaf surfaces and trans-ported by wind to marine sediments. Molecular biomarkersin deep-sea cores offer a rich record of terrestrial vegetation.A study by Feakins, deMenocal, and Eglinton (2005) showsthat long-chain n-alkanoic acids from the leafy waxes providea reliable indicator of terrestrial vegetation. Carbon isotopic(d13C) analysis of these molecules sampled through time fromSite 231 in the Gulf of Aden demonstrates the expansion ofgrasslands across northeast Africa from the Miocene throughearly Pleistocene. As with the eolian dust record, the analysisof plant biomarkers shows considerable variability in anygiven time interval; however, an emphasis on the average (fig.3 in Feakins, deMenocal, and Eglinton 2005) indicates a sub-stantial shift toward C4 grasses relative to C3 woody vegetationbetween 3.4 and 2.4 Ma, with an ongoing trend toward grass-dominated habitat registered at 1.7 Ma.

Study of paleosols (buried ancient soils) has produced the

third type of data set that focuses on East African aridity.Paleosols can preserve organic residues and carbonate depositsthat form under certain environmental conditions. The d13Cof these ancient soil components has been used to infer therelative proportions of C3 (wooded) and C4 (grassy) signalsof vegetation that grew in a limited area (∼10 m2) averagedover many years (e.g., Ambrose and Sikes 1991; Cerling 1992;Kingston 2007).

Figure 2 illustrates d13C data over the past 7 Myr, based onthe data set compiled by Kingston (2007). An overall increasein C4 biomass occurs within a mixed vegetation settingthroughout the time period of human evolution. Compilationof d13C data from Turkana and Olduvai (Cerling 1992; Cerlingand Hay 1986; Wynn 2000), in particular, indicates a long-term trend toward aridity and open habitat in East Africa.However, inspection of broader compilations (e.g., Cerling etal. 2011; Kingston 2007; Levin et al. 2004; see fig. 2) suggeststhe wide range of vegetation settings that accompanied thearidity trend, even from 3.0 to 1.5 Ma, the period with thestrongest turn from wooded to grassland settings. A d13C anal-ysis of soil carbonates from Gona, Ethiopia, also indicates anoverall expansion of C4 grass biomass from 4.5 to 1.5 Ma,although considerable heterogeneity in the vegetation is evi-dent throughout the sequence (Levin et al. 2004). While theonset of a “savanna” ecosystem consisting of more grass thantrees is evident in the Middle Awash of Ethiopia by around2.6 Ma (Quade et al. 2004), the oldest isotopic evidence ofopen grassland in East Africa is from ∼2.0 Ma at the site ofKanjera South, Kenya (Plummer et al. 2009).

Deep Lakes Formed in East Africa between ∼2.7 and 2.5 Maand between ∼1.9 and 1.7 Ma as End Members of StrongMoist-Arid Oscillations

Sedimentary sequences preserving evidence of hominins andother fauna include deposits, most notably lake sediments,that are sensitive indicators of past climate. In recent years,Trauth and colleagues have produced an analytical synthesisof lake deposits in East African basins. They conclude thatlong phases of high moisture characterized East Africa at threeimportant times—2.7–2.5 Ma, 1.9–1.7 Ma, and 1.1–0.9 Ma—and that these times were critical intervals in human evolution(Trauth et al. 2005). In publications since 2005, these re-searchers have placed greater emphasis on the fact that theprolonged high-moisture phases actually occurred during“periods of extreme climate variability” (Trauth et al. 2007:475) related to Earth’s eccentricity cycle (i.e., its modulationof precession, especially at the ∼413-kyr period; Trauth et al.2007). In other words, cycles of deep lakes and strong aridityoccurred during those ∼200-kyr-long intervals.

The most precisely dated evidence of periodic lake diato-mites is from the Tugen Hills, where 40Ar/39Ar dates and orbitaltuning of the sedimentary sequence point to five deep-lakecycles over a period of ∼100–115 kyr between ∼2.68 and 2.58Ma (Deino et al. 2006; Kingston et al. 2007). As to the period


Figure 2. Stable carbon isotope data (d13C) for the past 7 Myr, based on the data set compiled by Kingston (2007). An overallincrease in C4 biomass occurred. Although some d13C data sets from Turkana and Olduvai indicate a strong aridity trend andgrassland expansion associated with the evolution of early Homo and Homo erectus, a wider data set sampling many East Africansites demonstrates the considerable heterogeneity of vegetation throughout the key time interval. A color version of this figure isavailable in the online edition of Current Anthropology.

between 1.9 and 1.7 Ma, the major lake phase postulated byTrauth et al. (2005, 2007) corresponds to the Lorenyang Lakeof the Koobi Fora Formation, Turkana basin, and Beds I andlower II, Olduvai Gorge. Analyses of the Lorenyang Lake (aprecursor to the present Lake Turkana) indicate a period ofrelative stability surrounding a large lake from ∼2.0 to 1.85Ma followed by stronger fluctuations between low and highlake levels and substantial change in landscape features be-tween 1.85 and 1.7 Ma (Feibel, Harris, and Brown 1991; Joor-dens et al. 2011; Lepre et al. 2007; Quinn et al. 2007). AtOlduvai, Ashley (2007) postulates five episodes of lake ex-pansion and contraction between ∼1.85 and 1.74 Ma as pre-cessionally controlled climate affected the amount of rainfalland the profile of resources available to organisms. Deino(2012) has refined the Bed I Olduvai chronology and showsa more complex picture of wet-dry cycles in relation to orbitalprecession. These interpretations of strong wet-dry variabilityare consistent with a synthesis of fossil pollen, stable isotopes,microfauna, and other evidence from this oldest part of theOlduvai sequence (Potts and Teague 2010).

The manifestation of large, deep lakes in East Africa appears

at odds with the general aridity trend emphasized bydeMenocal (1995, 2004) in his study of the windblown dustrecord. In fact, an alternative view concerning the timing ofEast African aridity is offered by Trauth, Larrasoana, andMudelsee (2009). Their statistical reanalysis of eolian dust dataindicates that significant aridification of East Africa did notbegin ∼2.8 Ma. Rather, heightened aridity is evident only intimes of highly variable climate (strong wet-dry oscillations)in East Africa, especially ∼1.8 Ma, with a further drying trendthat began ∼1.5 Ma and reached a peak starting ∼1.0 Ma,again associated with magnified wet-dry fluctuation (see alsoOwen, Potts, and Behrensmeyer 2009).

Several Very Long Periods (Lasting ∼130–330 Kyr Each) ofMagnified Moist-Arid Variability Occurred in the Periodbetween 3.0 and 1.5 Ma

A strong stepwise increase in monsoonal variability (moist-arid fluctuation) occurred in northeast Africa ∼3 Ma. Thisincrease in moisture variability is recorded in the Mediter-ranean record of sapropels, which are dark, organic-rich sed-


Figure 3. Alternating high– and low–climate variability intervalsfrom ∼3.2 to 1.4 Ma. Low variability is defined by mean orbitaleccentricity or 1 standard deviation (SD) below meane ≤ 0.0145e for the past 5 Myr (R. Potts and P. deMenocal, unpublisheddata). The time interval, duration, and mean e for each intervalare shown. SDs of dust flux measured in Arabian Sea core 721(deMenocal 1995) test the validity and robustness of this high-/low-variability framework. Asterisks indicate that the directionof change in the SD matches (13 out of 17 pairs of intervals[76%]) the predicted widening and narrowing of climate vari-ability from one interval to the next. Boldface indicates intervalsin which low variability dominates for more than 30,000 yearsor high variability dominates for more than 100,000 years. Notethat the critical interval ∼2.85–2.08 Ma (highlighted) exhibits thelengthiest eras of both high and low climate variability, that is,prolonged intervals of wide fluctuation interspersed with rela-tively stable climate. A color version of this figure is available inthe online edition of Current Anthropology.

imentary layers deposited periodically on the sea floor as theresult of heightened runoff of the surface water from the Nilecatchment. The succession of Pliocene and Pleistocene sap-ropels suggests that climate variability in the northeast quad-rant of the African continent has been sensitive to orbitalprecession, with moist-arid cycles recorded nearly every 19–23 kyr (deMenocal 2004; Rossignol-Strick 1983). Sapropelstudies have thus documented that African monsoonal cli-mate is driven largely by precession. Geochemical studies in-dicate that the humid period in each cycle may have rangedfrom ∼4 to 12 kyr in duration (Wehausen and Brumsack1999).

Furthermore, according to Mediterranean sapropel records,a relatively low level of monsoonal variability was expressedfrom ∼4.4 to 3.0 Ma, followed immediately by a sharp ex-pansion in the range of fluctuation. This expansion coincideswith the last appearance datum (LAD) of Australopithecusafarensis and the current first appearance data (FAD) for Par-anthropus, ∼2.7 Ma, and Homo (sensu stricto), arguably by∼2.4 Ma.

While the rhythm of low-latitude climate variability is or-bital precession with periods of ∼19 and 23 kyr, the amplitudeof tropical moist-dry variability is strongly affected by orbitaleccentricity with its dual periods of ∼100 and ∼413 kyr. Theintersection of the precession and eccentricity curves (essen-tially the interaction of four sine waves) results in a predictedsequence of high and low climate variability for African lowlatitudes. This predictive framework of alternating high- andlow-amplitude climate variability “packets,” each lasting 104–105 years, is evident in the eolian dust records of deMenocal(1995) and is recognized by other authors (e.g., Campisanoand Feibel 2007; Deino et al. 2006; Kingston 2007; Trauth etal. 2007).

The recognition of high- and low-variability intervals offersa novel framework in which to examine East African climatechange. Figure 3 shows the time intervals in the period from3.0 to 1.5 Ma that are defined by high or low eccentricity.During high eccentricity, moist-arid variability is most pro-nounced (i.e., a high–climate variability interval), whereasrelatively stable climate occurs at low eccentricity. High var-iability dominates this key interval, and three of the mostprolonged periods of high climate oscillation in tropical Africaover the past 5 Myr are predicted for this interval, with du-rations of 326 kyr (∼2.79–2.47 Ma), 288 kyr (∼2.37–2.08 Ma),and 192 kyr (∼1.89–1.69 Ma).

Figure 4 shows a broader time perspective by showing theeight longest eras of high climate variability, based on orbital-eccentricity values, during the past 5 Myr, that is, the pre-diction of when moist-arid variability prevailed for the longestperiods in East Africa. Several important FADs and LADs inAfrican human evolutionary history are situated in these in-tervals. It should be noted, however, that the dates of thesefirst and last appearances are likely to shift as new fossil andarchaeological discoveries are made.

Figure 5 shows a plot of sapropel spectral reflectance from

2.7 to 1.5 Ma; the intensity of this reflectance records thestrengthening and reduction in moist-arid variability. Analysisof sapropel color (spectral reflectance) confirms that mon-soonal intensity was magnified in alternating “packets” of highand low climate variability as predicted by eccentricity-mod-ulated precession over long periods of time. The bottom offigure 5 situates the oldest known Oldowan archaeologicaloccurrence (Gona) in a high-variability interval. Oldowansites throughout the period occurred in both moist and aridenvironments and in prolonged eras of high or low climatevariability.


Figure 4. Longest-duration high–climate variability intervals (the longest time intervals of predicted high climate variability in EastAfrica over the past 5 Myr). The intervals range from 326 to 192 kyr in duration. These are the most prolonged eras of environmentalinstability in East African hominin evolutionary history. Several of the most prominent events in hominin evolution appear to haveoccurred in these intervals. FAD p first appearance datum; LAD p last appearance datum.

This synthesis of environmental data sheds light on thelarger physical and biotic context of hominin populations inand beyond the localities that have produced fossils so far.Over the past three decades, a small number of key hypotheseshave proposed critical linkages between climate, resources,ecological and social interactions, natural selection, and spe-ciation. These hypotheses offer robust and testable ideas re-garding the environmental factors involved in the evolutionof early Homo.

Environmental Drivers of Evolution in EarlyHomo: Cooling, Aridity, Moisture,Warmth, or Variability?

The environmental hallmarks and data sets reviewed in thepreceding section have led researchers to propose a variety ofenvironmental explanations concerning the origin of Homoand the early evolution of Homo erectus. The following sum-marizes the most prominent hypotheses.

Cooling

This idea has been around the longest, probably as a resultof the fact that glacial climate has, since the 1800s, beenunderstood as an important context of human evolution. Thecausal influence of climate cooling on African hominin evo-lution was formalized by Vrba in her influential turnover-pulse hypothesis (Vrba 1985, 1988, 1995a, 1995b) and habitattheory (Vrba 1992). Vrba proposed that global cooling was a

driver of evolutionary change across a wide range of taxa,that is, a concentrated pulse of extinction and speciationevents in lineages that, in Africa, favored cool- and arid-adapted taxa at the expense of warm- and moist-adaptedlineages, particularly evident in bovids and rodents. Coolingthus set the climatic and ecological contexts in which thegenus Homo (as well as Paranthropus) originated, around 2.5Ma, or more broadly between 2.8 and 2.4 Ma (Vrba 1988).In a more speculative development, Vrba (1994) posited thatcooling produced conditions that favored the retention ofjuvenile neurocranial traits through neotenic processes, whichpromoted brain enlargement in Homo.

Aridity

The long-term aridity trend in Africa was established by tec-tonic uplift, with the formation of the East African Rift Systemhaving a dominant effect in that portion of the continent(Sepulchre et al. 2006). Aridity, the expansion of open habitat,and, ultimately, the formation of grasslands were paralleltrends. Vrba’s most influential papers seemed to consolidatethe savanna hypothesis of early human evolution: global cool-ing led to African drying and the spread of grass-dominatedhabitats. The savanna hypothesis is the long-established ideathat dry, open, grassy settings provided the stage on whichthe human evolutionary drama unfolded. In a series of work-shops and conferences organized by Vrba (e.g., Vrba et al.1995), a number of researchers offered paleontological and


Figure 5. Sapropel variability, showing the presence of alternating intervals of high and low climate variability, from 2.7 to 1.5 Ma(right to left). Spectral reflectance is a measure of dark versus light color, with higher values for lighter stratigraphic layers (aridtimes, prevalent during less variable periods) and lower values for darker strata (moist times, especially prevalent during high moist-arid variability). Variability, noted in the upper right, provides examples of well-defined time periods of high and low climatefluctuation. Malapa, SK, ST, Wonderwerk, and A.Hanech are key southern and northern African Oldowan sites and age estimates.The remaining points and age estimates represent key East African Oldowan archaeological sites, starting with the oldest documentedso far (OGS-6/7 at Gona, ∼2.58 Ma). Spectral reflectance data are from ODP Site 967A (Hilgen et al. 1999; P. deMenocal, personalcommunication). A color version of this figure is available in the online edition of Current Anthropology.

climate data sets that fell in line with Vrba’s idea that some-thing important happened in African climate and mammalianevolution between 2.8 and 2.4 Ma. Among those data setswas deMenocal’s eminent analysis of eolian dust input fromnortheast Africa to the Arabian Sea and the Gulf of Aden.His efforts (e.g., deMenocal 1995) drew further attention tothe East African drying trend with a proposal that there wasa stepwise increase in dust production on the African con-tinent and dust input to nearby deep-sea records ∼2.8–2.5Ma. The conclusion reached by deMenocal and others is thatHomo arose as an integral part of the arid- and grassland-adapted African biota and that the major adaptations of Homo(e.g., stone tools, meat eating, brain size increase, geographicdispersal) were responses to the increase in aridity. The useby deMenocal of Cerling’s East African compilation of pa-leosol d13C, which portrays an increase in grass proportionsover time, adds evidence to the aridity hypothesis. A com-parison of paleosol d13C for the entire Turkana basin and thelower Awash basin, Ethiopia, by Levin et al. (2011) confirmsan overall increase in C4 (grass) vegetation in floodplain set-tings between 4 million and 700,000 years ago.

Moisture

In considerable contrast, Trauth, Maslin, and colleagues (Mas-lin and Trauth 2009; Trauth et al. 2005, 2007, 2010) haveargued that lengthy periods of high moisture and climate-driven production of deep lakes were the primary drivers ofPliocene and Pleistocene human evolution in East Africa.They envisioned three main intervals of climatic moisture thatwere associated with key events in human evolution. Specifi-cally, they considered earliest Paranthropus and Homo to beassociated with the first moisture phase, from 2.7 to 2.5 Ma;they placed the earliest H. erectus, its dispersal beyond Africa,and the origin of the Acheulean in the second phase, from1.9 to 1.7 Ma; and they asserted that the extinction of Par-anthropus and further expansion of H. erectus matched upwith the third moisture phase, from 1.1 to 0.9 Ma. As notedabove, this hypothesis actually emphasizes that lake devel-opment was part of a highly variable climate system; thus,these authors actually consider the moisture hypothesis as asubset of the variability selection hypothesis (Trauth et al.2005, 2007; see “Variability,” below).

Trauth et al. (2010) expanded the moisture hypothesis by


combining tectonic evidence for north-south rift formationin East Africa with lake history to postulate that large lakesimposed a barrier to populations on opposite sides (east vs.west) of those lakes, whereas lake contraction and drying ofthe rift floor created refugia on opposite rift shoulders. Thissequence thus promoted vicariance and allopatric speciation.The authors thus consider large lakes as “amplifiers” that driveevolutionary change.

Warmth

This hypothesis derives largely from Passey et al. (2010), whodocument what they consider to be extraordinarily persistenthigh-temperature soils in the Turkana basin, Kenya, over thepast 4 Myr. Soil carbonates were analyzed using an innovativemethod known as the “clumped-isotope thermometer” (Ghoshet al. 2006:1441) to investigate the temperature history of theTurkana basin. This method uses the distribution of 13C-18Obonds in paleosol carbonates as a proxy for soil temperatureduring carbonate formation. For the Turkana basin, the resultsindicate persistent hot temperatures above 30�C and oftenabove 35�C for the past 4 Myr, similar to those in the presentTurkana area, which is one of the hottest places on Earth.Although Passey et al. (2010) note that these surprising tem-perature estimates apply only to periods of soil carbonate for-mation, they claim that such periods make up most of the pastseveral million years in the Turkana basin. The authors linkthis finding to the evolution of human thermophysiology, spe-cifically, the change in body proportions associated with earlyH. erectus and “a long-standing human association with mar-ginal environments” (Passey et al. 2010:11245).

Variability

While recognizing that environmental variability is more thanmere noise in an overall trend, each of the previous fourhypotheses emphasizes mainly one dimension of the globalor African environmental system and treats it as the dominantadaptive challenge and evolutionary force relevant to the evo-lution of early Homo and H. erectus. An integrated treatmentof the environmental data sets, however, begs the question asto what the principal environmental signal in this dynamicera of Earth’s climate history truly was. Was it cooling orwarmth, aridity or monsoons? The variability hypothesis, alsoknown as variability selection, cuts across this possible im-passe by highlighting not any one trend, habitat, or extremityof climate oscillation but rather the evolutionary effect ofenvironmental dynamics itself (Potts 1996a, 1996b, 1998a,1998b, 2007). It is the variability across a large set of envi-ronmental axes that led to highly varying adaptive conditionsover time and space: warm-cool, wet-dry, high or low resourceabundances, concentrated or patchy resource distributions,high or low parasite loads, high or low predation risks, denseor dispersed species populations, and strong or weak inter-species competition.

The variability selection idea points to this spectrum ofenvironmental dynamics in creating a signal that can promptadaptive change. It seeks to answer how a population of or-ganisms can change over time via a process of adaptation tothe variability—that is, to the temporal range—of environ-mental dynamics in adaptive settings. Furthermore, it positsthat adaptation to environmental dynamics fosters plasticity,adaptive versatility, or—perhaps the most encompassing termat many levels of biological organization—adaptability (Potts1998b, 2002, 2007; see fig. 6).

Based on evidence that well-defined eras of pronouncedclimate variability occurred between 3.0 and 1.5 Ma, coupledwith episodic revisions of landscapes due to tectonic events,the idea is that African environmental dynamics created highlydiverse conditions of natural selection that led to positiveselection for genetic combinations favoring adaptability andthus the ability of certain organisms to adjust to environ-mental change, move to new habitats, and respond in novelways to their surroundings. Adaptive change in response toenvironmental dynamics thus engendered responsiveness atmany biological levels—from molecular, cellular, and physi-ological to developmental, social, and ecological dimensionsof life (Potts 2002). The possibility presented by the variabilityselection hypothesis is that the evolution of early Homo andH. erectus was embedded in environmental instability and canbe explained by selection that improved the ability of certainhominin populations (ultimately species) to adjust to varia-tion in their adaptive setting.

Variability selection as a viable process of evolutionarychange has recently been tested by Grove (2011), who useda single-locus genetic model originally suggested in Potts(1996a, 1996b, 1998b). In Grove’s simulations, “versatilist”alleles that build genetic combinations favoring plasticity wereunable to increase when the fluctuating environment wasmodeled as a sine wave. However, in an empirical environ-ment based on d18O for the past 5 Myr, variability selectionwas inevitable as versatile strategies of adaptation and behav-ioral plasticity were favored (Grove 2011). This test of vari-ability selection as an evolutionary process aligns with pointsmade by deMenocal (2011): in a situation where seasonal- toorbital-scale fluctuation is regular or even in tempo and am-plitude, variability alone is unlikely to have served as a se-lection agent; however, progressively larger degrees of envi-ronmental variability, evident as an overall trend in the d18Ocurve (see fig. 1) and in particular intervals during the Pli-ocene and Pleistocene (figs. 3, 4), may result in new adap-tations that promote adaptable behavior, including successfulresponses and dispersal to novel environments.

Geographic Variation and the PossibleRole of Refugia

According to several recent studies, the expression of climaticconditions varied considerably across different areas of EastAfrica. The Omo-Turkana basin has especially contributed to


Figure 6. Three possible outcomes of population evolution in a time series of environmental dynamics typical of the Plio-Pleistocene.The ability to move and track habitat change geographically (narrow lines) or to expand the degree of adaptive versatility is importantfor any lineage to persist. Extinction occurs if species populations have specific dietary/habitat adaptations (i.e., a narrow band of“adaptive versatility”; highlighted bands) and cannot relocate to a favored habitat. In the hypothetical situation (rightmost band)where adaptive versatility expands, migration and dispersal may occur independently of the timing and direction of environmentalchange. The evolution of adaptive versatility is the impetus behind the variability selection idea. A color version of this figure isavailable in the online edition of Current Anthropology.

this emerging picture of geographic variation. In a study ofsoil d13C and d18O by Levin et al. (2011) encompassing ∼5,000km2, the northern floodplains of the ancestral Omo Riverbetween ∼2.9 and 2.0 Ma were dominated by woody C3 veg-etation indicative of a seasonally moist, riparian woodland,at the same time as the broader southern floodplains haddrier, less productive soils populated by grassy C4 vegetation.The Awash basin, by comparison, was more arid than theOmo-Turkana basin, and it supported a greater proportionof C4 vegetation during this same interval (Levin et al. 2011).These comparisons suggest that different regions, both be-tween and within basins, manifested varied degrees of sen-sitivity to the large-scale climate shifts that affected East Africaoverall during the late Pliocene.

Furthermore, a study by Joordens et al. (2011) introducesthe intriguing idea of a habitat refugium as a magnet forhominin populations during the Plio-Pleistocene. The authorsdevelop a new type of climate record, a strontium isotopic(87Sr/86Sr) indicator applied to lake fish fossils. Their studydemonstrates, first, that precessional moist-arid variation didindeed affect the lake that existed in the Turkana basin be-tween 2.0 and 1.7 Ma but, second, that the basin retainedpermanent water and moist wooded habitats throughout theperiod from ∼2.0 to 1.85 Ma. The authors show that homininfossils occur during both wet and dry phases of the long-termmonsoonal cycles and further suggest that hominins weredrawn to continuously well-watered habitats over this long

period. In other words, for about 150,000 years the easternTurkana basin served as a refugium for hominins and otherpermanent water-dependent fauna. This refugium was buf-fered from severe lake level fluctuations and drought thataffected other regions of East Africa.

The idea that an area on the order of 102–103 km2 wasprotected over the long term from extreme fluctuations is animportant consideration. Could the reliable resources of arefugium have provided a more significant context for theevolution of early Homo or Homo erectus than climate-drivenextremes of food and water fluctuation over the broader EastAfrican region?

This question cannot yet be answered; however, the moist,wooded refugium suggested by Joordens et al. (2011) appearsnot to have extended past 1.85 Ma. Isotopic comparisonacross the Omo-Turkana basin indicates a major restructuringof hydrology and vegetation toward arid C4 habitats beginningno later than 1.9 Ma (Levin et al. 2011). In fact, Quinn et al.(2007) characterize the entire period from 2.0 to 1.75 Ma asone of important vegetational change in the Turkana basin,trending from relatively closed savanna woodland towardopen, low-tree shrub savanna. These authors also note thatthe shift in average floral composition between 2.0 and 1.75Ma coincides with high species turnover, the principal findingby Behrensmeyer et al. (1997). Furthermore, this finding isconsistent with the study by Lepre et al. (2007) at East Tur-kana, which detects evidence of heightened environmental


variability from 1.87 to 1.48 Ma; high wet-dry monsoonalvariability is predicted for much of this period (fig. 3).

Do Environmental Drivers Serve as AdequateEvolutionary Explanations?

In brief, the answer to this question is “no” or, at least, “notif any given environmental hypothesis invokes a fairly sim-plistic notion of correlation.” Tests of correlation seek to de-termine how key evolutionary events map onto large-scaleenvironmental patterns. The value of correlation hypothesesis that they usually integrate all the evidence from paleoan-thropological sites pertinent to a particular evolutionary eventand then examine how well the patterning of evolutionarychange matches global or continental environmental trends.Information from deep-sea cores has proved especially usefulin testing correlation hypotheses. Such tests do not, however,formally establish which particular environments homininsactually encountered in their local settings or the processesby which hominin populations responded to novel survivalchallenges in those settings.

Two additional factors, discussed below, have become crit-ical in developing environmental explanations for the evo-lution of early Homo: evidence of evolutionary change incontemporaneous large mammals and the development ofcompelling, testable models that specify how particular typesof environmental change can incite evolutionary change.

Faunal Change Can Test the Causes of EvolutionaryChange in Homo

The study of faunal communities can provide an importanttest of how environmental dynamics influenced the evolutionof Homo. Each lineage represents a natural experiment in howexternal environment may have prompted certain evolution-ary changes. Responses across an array of large mammals—for example, the success of arid-adapted taxa over moist-adapted ones, the emergence of dental hypertrophy in mul-tiple animal lineages, an increase in body size across carniv-orous taxa, or a widening of geographic ranges across a varietyof taxa—can offer clues as to the nature of the adaptive chal-lenges and the conditions that shaped human evolution, in-cluding the origin and early evolution of the genus Homo.For these tests to be effective, the relevant faunal lineagesmust occur in the same times and places in which homininslived and evolved. This point has been recognized for sometime, particularly for the interval between about 3.6 and 1.8Ma.

Analysis of a wide range of large mammal fossil assemblagesfrom eastern and southern Africa by Reed (1997), for ex-ample, showed that grazing adaptations fluctuated within nar-row limits (10%–25%) between 3.6 and ∼2.0 Ma. Only byabout 1.8 Ma did an increase in the percentage of grazingspecies occur, exceeding 30%. Reed’s finding corresponds wellto information from d13C (Cerling 1992; Levin et al. 2011),

African dust variability (deMenocal 1995, 2004), and evidenceof species turnover in the Turkana basin (Behrensmeyer etal. 1997). Reed (1997) concludes that the climatic and eco-logical transition at ∼1.8 Ma corresponds to the early evo-lution of Homo erectus, and thus only with this later speciesof Homo (after 2.0 Ma) do we have a hominin adapted toarid and open landscapes.

A different finding by Bobe and Eck (2001) came frommeasuring the relative abundances of bovids recovered fromthe Omo Shungura Formation between 3.4 and 1.9 Ma. Theresults indicate a climatic shift toward increased aridity be-ginning ∼2.8 Ma and intensifying by 2.3 Ma, an age for ar-idification considerably earlier than that reached by Reed(1997). Later analysis of bovids, suids, and cercopithecids inthe Turkana basin (Bobe and Behrensmeyer 2004) gave evi-dence of (1) an initial increase in grassland-adapted mammalsat ∼2.5 Ma, (2) fluctuation in the abundance of arid andmoist taxa until ∼2.0 Ma, and then (3) a strong excursiontoward open-country grazers at ∼1.8 Ma. On the basis ofthese findings, Bobe and Behrensmeyer (2004) favor a com-bination of explanations that invoke both aridity and vari-ability selection as evolutionary drivers in the Turkana basinbetween 2.5 and 1.8 Ma.

In a more refined analysis, Bobe et al. (2007) confirmedthat the abundance of grazing bovids underwent an overallincrease in the Turkana basin between 3.0 and 1.0 Ma. How-ever, different patterns of fluctuation in the percentage ofgrazers occurred in different parts of this large basin; in fact,grazing bovids underwent an overall decline in West Turkanabetween 3.0 and 1.5 Ma. According to the authors, the overalltrend nonetheless indicates that East African landscapes in-habited by bovids and hominins became more open, arid,and seasonal; the moist and more vegetated end of the habitatspectrum became particularly limited after 2.0 Ma. They cau-tioned, however, that different patterns of faunal change inseparate parts of the same region point to the difficulty ofestablishing definitive correlations between climatic and fau-nal change.

Several other notable studies occur in a compendium fo-cused on the African Pliocene faunal evidence (Bobe, Alem-seged, and Behrensmeyer 2007). Frost (2007), for example,examined evolutionary change in the Cercopithecidae andfound no support for a turnover pulse from 2.8 to 2.5 Ma,the time range predicted by Vrba (1995b) for speciation andextinction caused by global cooling and the spread of grass-land habitats in Africa. Frost did find, however, a clusteringof first and last appearances of monkey lineages ∼2.0 Ma,which could be associated with a number of environmentalhappenings at that time, including East African aridification.

In a study of evolutionary change in Plio-Pleistocene car-nivores (members of the Carnivora), Lewis and Werdelin(2007) noted that the earliest known appearance of stone tools∼2.6 Ma had no obvious effect on the carnivore guild. How-ever, a drop in carnivore speciation rate and a rise in theirextinction rate after 1.8 Ma and a pronounced decrease in


carnivore lineages after 1.5 Ma could be ascribed to the emer-gence of H. erectus, climate change, and a drop in overall preyspecies richness.

From these examples and others (e.g., Behrensmeyer et al.1997; Reed 2008; Vrba 1995a), it is evident that most of whatis termed “hominin paleoecology” to date has involved eitherthe analysis of lineage turnover or habitat reconstruction.Questions about which species commonly occurred togetherand which taxa had particularly strong associations with hom-inins have yet to gain much attention, even though suchstudies would truly reflect “paleoecology” in terms of docu-menting species co-occurrences and potential interactions.Such studies would offer the strongest clues regarding theadaptive milieu that influenced early hominin populations.Some notable exceptions exist, however, drawn mainly fromthe late Pliocene Turkana basin and the mid-Pleistocene Olor-gesailie basin (Bobe and Behrensmeyer 2004; Bobe, Behrens-meyer, and Chapman 2002; Potts 2007). These studies suggestthe fluidity of species composition in faunal communitiesdated between 2.8 and 1.8 Ma (Turkana basin) and between1.0 and 0.6 Ma (Olorgesailie basin). These findings are con-sistent with environmental variability’s role in causing theassembly and disassembly of ecological communities and thecontinual shifting of the adaptive conditions associated withthe evolution of Homo.

An Understanding of Evolutionary Processes Is Also Vital

Any evolutionary explanation must posit explicitly what evo-lutionary processes were at work in order to evaluate thefeasibility of each explanation and to lay out potential testsof the explanation. Vrba (1985, 1992, 1995b) has been expliciton this point by asking what evolutionary mechanisms areinvolved in translating from external environment to evolu-tionary responses. She posits, for example, that cooling re-sulted in directional selection and adaptive evolutionary re-sponses to cooler, drier habitat and in speciation that favoredorganisms possessing such adaptations. Other researchers im-plicitly invoke habitat-specific directional selection as the pri-mary adaptive process, although very little attempt is madeto state how directional selection related to aridity, moisture,or high temperature translated into evolutionary change atthe critical junctures in the environmental system.

The idea of variability selection has followed Vrba’s leadby focusing on an evolutionary process as the foundation forexplaining hominin and faunal evolution. As noted above,variability selection makes a specific connection between (1)evidence of increased environmental variability, (2) the effectof this increase on resources and adaptive settings at varioustemporal scales, and (3) the evolutionary change that mayresult from the inconsistency of natural selection over time,which is posited to favor adaptive versatility over habitat-specific solutions to survival problems.

A focus on evolutionary processes leads to many questionsas to how climate and overall environmental change actually

influenced speciation and adaptive shifts in early Homo(deMenocal 2011). The work has hardly begun to relate eventhe most highly resolved climate records to shifts in season-ality, landscapes, and resource abundances, that is, the thingsthat matter to organisms. So, for example, are evolutionarytransitions driven largely by stresses associated with resourcescarcity, opportunities related to resource abundance, or un-certainties linked to seasonal unpredictability and longer-termlandscape remodeling? Were adaptations in early Homo largelyhabitat specific (e.g., solutions to a well-defined and consistentset of environmental problems), or did they reflect increasesin behavioral plasticity and adaptability (e.g., solutions tohighly dynamic settings and environmental novelty)?

Furthermore, it is hard to say whether significant evolu-tionary shifts were initiated in highly localized settings (e.g.,refugia), where intra- and interspecific interactions played aprominent role, or over much of an evolving lineage’s geo-graphic range, where broader climatic and tectonic effects mayhave been important. Finally, were adaptations in early Homoa response largely to external (climate- and tectonism-mediated) factors or to social and competitive factors thatwere consistent across a diverse spectrum of habitats? It iscounterproductive to frame this last matter as an either/orquestion; rather, it is better to see environmental, resource,social, and competitive factors as interrelated and potentiallyreinforcing rather than at odds with one another in explaininghow evolutionary change occurred.

Archaeological Behaviors and the EmergingAdaptability of Homo

The behavioral and ecological adaptations of hominins,viewed through the archaeological record, offer their owndirect clues as to the nature of responses by early Homo andHomo erectus to environmental challenges and can thus pointto whether and how cooling, warming, aridity, moisture, orvariability influenced evolutionary change. This section iden-tifies certain key adaptations evident in the behavioral recordof artifacts, sites, and archeofaunas and examines them in thelight of the changes in the East African environmental systembetween 3 and 1.5 Ma.

Transporting Rocks and the Oldest Known StoneToolmaking, ∼2.6–2.0 Ma

During this interval, groups of one or more hominin speciesbegan to seek out rocks across distances of up to severalkilometers to flake and to assist in processing food. From anadaptive-strategy standpoint, this behavior is more peculiarthan is commonly perceived, largely because the costs of learn-ing about and keeping track of good-quality stone, walkingconsiderable distances to get it, and carrying rocks of up toseveral kilograms across the landscape were likely quite high,especially since rocks by themselves have no caloric or nu-tritional value. It is not the manipulation and flaking of stone


that is unexpected in hominins so much as the dedication totransporting sufficiently large amounts of rock, detectable to-day as concentrations, over considerable distances from placeswhere that rock naturally occurred.

Between ∼2.6 and 2.3 Ma, the use of flaked stone tools inacquiring food entailed relatively short-distance transport ofresources, typically tens to hundreds of meters (e.g., Delagnesand Roche 2005; Goldman-Neuman and Hovers 2009; Semawet al. 2003). Longer-distance transport is well documentedlater. By 2.0–1.95 Ma, at the site of Kanjera South (Braun etal. 2008), hominin toolmakers were moving certain types ofstone over total distances of at least 12–13 km from theirclosest rock sources. The energetic and potential social bene-fits of processing particular foods using stone tools obviouslycompensated for the costs of rock transport involved in suchcumulative distances.

The question, then, is “Under what environmental andselective conditions did this complex of activities—rock trans-port, precise flaking, and mental mapping of stone and fooddistributions—catch on and become conspicuous in the rep-ertoire of certain hominin populations?” The oldest knownsite that preserves stone flaking—Gona, Ethiopia—is asso-ciated with d13C evidence of a prominent grassy componentin a mixed vegetation setting (Quade et al. 2004). On thisbasis, it could be claimed that stone toolmaking was an ad-aptation associated with aridity and increasingly open vege-tation.

However, as noted above, this oldest archaeological site issituated in a period of pronounced moist-arid variability. Fig-ure 5, furthermore, opens the question as to why Oldowantool behavior eventually spread across East Africa and else-where and became an important behavior in the genus Homo.The emergence of the Oldowan and its dispersal in an era ofwidely shifting environments and diverse food regimes pointsto an interpretation that contrasts with the long-standing arid-ity-grassland explanation. It is feasible that stone tools suc-ceeded largely because carrying stone made it possible to bringtools and foods that required tool processing together in thesame places—and to do this consistently across diverse hab-itats, even as the distribution and abundance of food resourcesvaried over time and place. Stone transport not only incurredlarge energetic costs but also enabled consistent and predict-able returns in response to a varying environment. As a result,the effort involved in making sharp edges and using rocksfor crushing provided a resilient means of processing achangeable array of foods across a wide range of habitats.

By ∼2.0–1.95 Ma, therefore, we see definitive evidence ofpersistent Oldowan toolmaking (and stone transport) in anarid grassland setting recorded at Kanjera South (Plummeret al. 2009) and in a nearly contemporaneous moist, woodedhabitat recorded at FwJj20, East Turkana (Braun et al. 2010).In this light, the Oldowan repertoire of behavior emerged asa characteristic of the genus Homo (even if adopted initiallyby non-Homo populations) because it enabled dietary and

foraging adaptability in the face of Plio-Pleistocene environ-mental dynamics.

An important further implication regarding stone transportconcerns its effect on the energetic costs of locomotion. Thistopic has not received much consideration since a “nullmodel” of Oldowan stone and food transport costs was de-veloped in Potts (1988). In light of current ideas about lo-comotor costs (e.g., Bramble and Leiberman 2004), the pointto consider is that the cost of stone transport for particularactivities must be factored into interpretations about endur-ance running versus walking in early Homo.

Based on weights of Oldowan tools from Bed I Olduvai(e.g., Potts 1988, 1991), an estimate of the minimum weightof the basic Oldowan equipment—a single hammerstone plusa number of basalt/trachyte/phonolite cores sufficient to yieldenough sharp flakes (minimally 50–100) to cut the hide anddisarticulate a fleshed wildebeest-sized ungulate—wouldcome to ∼4–9 kg. (An average of ∼6 flake scars per Oldowancore would require at least 8–16 cores [for ∼50 to ∼100 flakes],assuming that all flake scars define usable flakes, plus onehammerstone, with calculated weights of 375–500 g per coreand an average of 0.5–1 kg per hammerstone [Potts 1991].)Repeated trials in an East African field situation (Olorgesailie,Kenya), involving well-conditioned individuals carrying 5–10kg of rocks in backpacks over uneven terrain from outcrops5 km away, suggest that running is feasible for 100 m or sobefore the rate of fatigue strongly urges the desire to walk,which can be accomplished comfortably for the entire trialdistance. While this does not cast doubt on whether earlyHomo or H. erectus ran long distances, it is reasonable to askwhat the individual might accomplish after such a run if aminimal butchery tool kit involving 8–16 sizeable rocks werenot also transported. Social running is feasible, although itwould be speculative at best to wonder whether each indi-vidual runner would carry a similarly sized tool kit. Or per-haps running in order to capture and butcher animals wasnot the point. Adequate quantitative experiments yielding re-alistic expectations about the addition of stone transport tolocomotor costs have yet to be carried out (see Pontzer 2012).

Dietary Expansion Involving Access toMeat/Marrow Resources

Because of the relative rarity and unpredictability of preyspecies and animal tissues across the landscape relative toplant foods, carnivory typically leads to larger foraging dis-tances and substantially increased energy budgets (Carbone,Teacher, and Rowcliffe 2007; Kelt and Van Vuren 1999; Nagy,Girard, and Brown 1999). That Oldowan hominins coupledthese latter costs with the added costs of transporting therocks needed to process carcasses is reason to suspect thatthe survival and energetic returns on accessing animal tissueswere substantial. Access to nutritious resources in mammalianbones and organs as well as to a wider range of plant foods


is likely to have played a fundamental role in offsetting thecosts of transporting stone.

The expansion in diet implied by access to animal fat andprotein is also thought to have played a role in a number ofadaptive changes in the evolution of Homo (Aiello andWheeler 1995; Anton, Leonard, and Robertson 2002; Plum-mer 2004; Shipman and Walker 1989). The possible statureincrease in early Homo and by 1.9–1.7 Ma in some popula-tions of H. erectus could have had the dual advantage ofincreasing the foraging range and facilitating the search forcarcasses, while animal protein and fatty tissues could havefueled body and brain growth. These foraging strategies ap-pear to have been advantageous in an arid, open habitat. Atthe same time, the expansion of diet to include animal foodswas an advantageous buffer against changing templates offood resources across an array of habitats. That is, the adop-tion of animal foods likely offered a useful means of ecologicaland dietary adjustment even if meat/fat acquisition was notthe goal of all or most occasions of stone tool use. Thisapproach to buffering environmental instability would haveproved equally useful when moving through unfamiliar hab-itats, thus facilitating range expansion and dispersal (Anton,Leonard, and Robertson 2002).

As hominins ventured farther into the ecological arena oflarge carnivores, seeking out and carrying portions of animalcarcasses had its risks (Blumenschine 1991; Donadio andBuskirk 2006). These risks were conceivably lowered whenprey density was high and competition for meat relatively low.Evidence from a variety of sites—Kanjera South, FwJj20 inEast Turkana, and FLK Zinj at Olduvai—implies, however,that by 2.0–1.8 Ma, the success of Oldowan toolmakers incompeting for animal tissues applied across a broad spectrumof environmental conditions and probably a variety of eco-logically competitive situations. This could mean that earlyHomo and especially H. erectus succeeded in the carnivoreend of an omnivorous dietary spectrum not solely in arid,grassland habitats but across the gamut of arid-moist con-ditions and the changing template of seasonality associatedwith high climate variability.

Delayed Consumption of Food: An ExtraordinaryDevelopment in Hominin Behavior with BroadImplications regarding Sociality

Relatively dense concentrations of butchered faunal remainsassociated with plentiful tools made from rocks 2–13 km fromtheir sources are known by ∼2.0–1.8 Ma (Braun et al. 2008,2010; Plummer 2004; Plummer et al. 2009). These sites appearto reflect a critical step in foraging and sociality, namely, adelay in eating the marrow- and meat-rich carcass portionsthat were transported. This behavior is odd, given the em-phasis on “eat-as-you-go” foraging in almost all animals ex-cept when provisioning young. Although Isaac’s (1984) em-phasis on home bases involving male-female division of labor,pair bonding, and other elements of human behavior has been

well dissected (Binford 1981; Potts 1988), the repetitive actof carrying and aggregating rich packets of fatty tissue andprotein almost certainly had important social implications(see below).

The spread or elaboration of this behavior (the transportingof food) is not well calibrated, but its appearance by ∼2 Mafalls in an era when stable and highly variable climate regimesalternated with one another, evident in the predictive frame-work of East African monsoonal variability (fig. 3) and in thesapropel record (fig. 5). Ultimately, this strategy of dual trans-port of stone and food resources was expanded to diversehabitats populated by Oldowan toolmakers, including earlyHomo and H. erectus, and was elaborated in the relativelyvariable settings of East Africa between 1.9 and 1.5 Ma as wellas across the diverse habitats of northern and southern Africaand from the Caucasus to eastern Eurasia (Potts and Teague2010).

Oldowan Behavior and Its Implications concerningMortality in Early Homo

It is commonly assumed that the venture into the ecologicaldomain of large African carnivores, made possible by profi-cient stone flaking, entailed substantive risks of injury anddeath and probably led to a marked rise in extrinsic mortalityin Oldowan toolmakers. There is evidence, for example inBed I Olduvai (Domınguez-Rodrigo, Barba, and Egeland2007; Potts 1988), of substantive carnivore involvement inbone assemblages where aggregations of stone tools also occur,which is indicative of a spatially focused overlap where tool-making hominins and meat-eating carnivores conducted theirbusiness. The overlap of carnivore tooth marks and toolbutchery marks, first reported in Bed I Olduvai (Potts andShipman 1981), might at first suggest a powerful way to in-vestigate whether carnivores or hominins held the upper handin their competitive interactions over carcasses. Such tooth/cut mark overlaps appear to be quite rare, however, and areunlikely to produce definitive, statistically robust results acrossa variety of sites. Furthermore, no systematic study has yetbeen carried out comparing the frequency of carnivore toothmarks on bones of Australopithecus, early Homo, and earlyH. erectus; the findings of such a study would not necessarilymeasure carnivore-caused mortality but may imply how com-monly carnivores had access to hominin bones and the degreeof overlap and potential risk that hominins incurred beforeand after the emergence of stone tool flaking.

The assumption that stone tool–assisted carnivory incurredhigher mortality costs thus has yet to be demonstrated. Thispoint leaves open the reasonable possibility that episodic andeventually persistent stone toolmaking and carnivory oc-curred only when the mortality costs due to predation haddecreased, compared with that for earlier pre-Oldowan hom-inins (Lewis and Werdelin 2007). We may call this the“decreased-mortality hypothesis” of Oldowan behavior, whichsees carnivory as an integral yet variably expressed aspect of


the adaptive repertoire in early Homo and H. erectus. Such adecrease could have been sustained by shifts in social group-ing, cooperation, vigilance, and signaling, that is, a variety offeasible behavioral changes that could have negated the sup-posed rise in mortality risk due to predation (Gursky-Doyenand Nekaris 2006). Conditions favoring such social-basedstrategies likely had to be met as Oldowan hominins carriedcarcass parts and invested in strategies of food and stonetransport that repeatedly concentrated social encounters incertain spots on the landscape, whether through central-placeor multiple-place foraging (Isaac 1984; Potts 1988, 1991).

A decrease in extrinsic mortality due to predation could haveprovided the context in which intrinsic mortality (due to aging)became a more prominent factor in life history (Kuzawa andBragg 2012). In other words, decreased mortality due to pre-dation served as either a release or the impetus for the pro-longation of maturation and an increase in longevity. The lattereffects, furthermore, would have improved the opportunitiesfor alloparenting by older siblings and postreproductive adults.This train of cause and effect is currently speculative yet hasthe following potentially testable implications: (1) taphonomicindicators of decreased hominin-carnivore interaction (at oddswith the usual assumption of increased predation risk due tooverlapping carnivory); (2) prolongation of dental develop-ment/maturation (see Schwartz 2012), correlated with de-creased hominin-carnivore interaction; and (3) body size in-crease, which is predicted to accompany decreased extrinsicmortality (see Kuzawa and Bragg 2012; Migliano and Guillon2012; Pontzer 2012).

Increase in Body Size

An increase in body size in early H. erectus is documented byfossils such as KNM-WT 15000 (∼1.53 Ma) and KNM-ER1808 (∼1.7 Ma). The robust innominate KNM-ER 3228 (Ruffet al. 1993) dated ∼1.92 Ma (Joordens et al. 2011) and femora(e.g., KNM-ER 1481) dated ∼1.89 Ma from the Turkana basin(Ruff and Walker 1993) further suggest that body enlargementhad begun by at least 1.9 Ma, although the taxon in whichthis occurred is as yet unknown (see Holliday 2012; Pontzer2012). It seems unlikely, however, that the body size increasein Plio-Pleistocene Homo was registered across all populationsor environmental contexts. Since several intervals of height-ened moist-arid fluctuation (with intervening low-variabilityperiods) characterized the time between 2.0 and 1.7 Ma, it ispossible that the increase in body size reflects the importanceof plasticity in body growth trajectories. This proposed risein plasticity would have made feasible a broader spectrum ofbody sizes, from small to large, in response to resource avail-ability and other environmental factors. Evidence of smallindividuals of early H. erectus (sensu lato) at Dmanisi (Lord-kipanidze et al. 2007; Pontzer et al. 2010) may indicate thatthe apparent evolution of large body size in this species ac-tually reflects the evolution of plasticity in body growth inaccord with a variability selection scenario rather than a re-

sponse solely to arid, hot, open landscapes (see also Anton2012; Anton and Snodgrass 2012; Kuzawa and Bragg 2012;Migliano and Guillon 2012).

Persistence and Spread of Oldowan BehaviorsBeginning by 2.0 Ma

An important shift in the archaeological record of Oldowanbehavior appears to be focused in the interval from about 2.0to 1.8 Ma. Before 2.0 Ma, archaeological sites (defined byclusters of stone artifacts) are distributed over time in anepisodic pattern. That is, within a given stratigraphic se-quence, sites typically occur within a very confined strati-graphic interval and a narrow and repetitive range of geo-logical settings. By contrast, after 2.0 Ma, Oldowan behaviorbecomes considerably more persistent, with clusters foundmore frequently across consecutive layers within a given strat-igraphic sequence. In addition, Oldowan tools are foundthroughout a varied spectrum of habitats and over widergeographic areas both within Africa and, for the first time,in Eurasia.

In the series of 2.58-Ma sites from Gona, Ethiopia, forexample, the three most studied lithic assemblages (EG-10,EG-12, and OGS-7) come from the same stratigraphic inter-val, described by Stout et al. (2010) as within the secondfining-upward sequence above the base of the Busidima For-mation. Each of about a dozen Gona sites, distributed from∼2.58 to 2.53 Ma and from ∼2.17 to 2.0 Ma, occurs in fining-upward sediments overlying large cobble conglomerates(Quade et al. 2008). Site location thus seems to have beenconditioned by proximity to conglomerates, which were thesources for on-site or very localized (e.g., ∼20-m distance inthe case of OGS-7) stone flaking (Stout et al. 2005, 2010).

Oldowan tools associated with the Hata Member at Bouri,dating to ∼2.5 Ma, are described as rare and scattered surfaceartifacts with no concentrations observed. Several mammalianbones bearing cut marks and hammerstone impact scars havealso been described from surface and excavated finds withina single stratigraphic horizon across more than 2 km of out-crop (de Heinzelin et al. 1999).

The next-oldest archaeological sites documented so far—from Hadar and Omo, Ethiopia, and Lokalalei, West Turkana,Kenya—are dated between ∼2.36 and 2.32 Ma. The two Hadarsites known so far (within the Makaamatilu basin), A.L. 894and A.L. 666, are separated vertically by ∼2 m within thelowermost portion of the Busidima Formation (Goldman-Neuman and Hovers 2011). Five archaeological sites reportedin Omo Shungura Member F are distributed though ∼35-mthickness of Member F; here, the repetitive characteristic isthat all of the Member F sites occur within small channeldeposits or associated floodplain silts of a braided stream (dela Torre 2004; Howell, Haesaerts, and de Heinzelin 1987). Inthe West Turkana sequence, two sites, the slightly older Loka-lalei 1 and the younger Lokalalei 2C, are almost contempo-raneous (Delagnes and Roche 2005), and the next Oldowan


occurrences reported so far in the basin are more than 300,000years younger, in the Upper Burgi Member at East Turkana.

There is little doubt that Oldowan assemblages dated be-tween 2.6 and 2.0 Ma provide solid evidence of intentionalflaking, selective use of raw materials, and a well-developedsense of fracture mechanics and planning during the processof stone flaking (e.g., Delagnes and Roche 2005; Goldman-Neuman and Hovers 2011; Stout et al. 2005). The boundarydefined here beginning ∼2.0 Ma, nonetheless, marks a laterinterval of Oldowan sites that occur in a wider variety ofgeological, geographical, and environmental contexts. For ex-ample, excavations at Kanjera South, Kenya, have uncovereda continuous stratigraphic distribution of Oldowan artifactson the order of 102–103 years long, dated between 2.0 and1.95 Ma, associated with evidence of persistent carnivory(across multiple stratigraphic layers) in the form of cut andpercussion marks on small and medium-sized mammalianskeletal remains (Ferraro 2007; Plummer 2004; Plummer etal. 2009). M. D. Leakey’s excavations in Bed I Olduvai, fur-thermore, documented a nearly continuous distribution ofOldowan stone tools from ∼1.85 to 1.70 Ma (Deino 2012;Hay 1976; Leakey 1971). The temporal distribution of sitesin figure 5, moreover, illustrates the expansion of Oldowansites within and beyond East Africa starting ∼2 Ma, after anoticeable sampling hiatus between roughly 2.3 and 2.0 Ma.

In South Africa, new age estimates and a synthesis of data(Herries, Curnoe, and Adams 2009; Herries and Shaw 2011)indicate that the oldest currently documented Oldowan toolsin this part of Africa are between 2.0 and 1.78 Ma, includingSterkfontein M5A (1.8–1.5 Ma), Wonderwerk Cave (1.95–1.78 Ma), Malapa (∼1.98 Ma), and Swartkrans M1 (∼2.0 Ma,although this date is considered relatively poorly constrained).This revised chronology for the oldest archaeological finds inSouth Africa, along with the presence of a sequence of stonetool layers dated ∼1.8 Ma at Aın Hanech, Algeria (Sahnouniet al. 2002), further suggests that the making of stone toolshad become a regular part of the behavioral repertoire ofdispersing populations of Homo by ∼2.0 Ma. The sequenceof Oldowan sites now documented at Dmanisi between 1.85and 1.78 Ma and the archaeological sequence beginning by1.66 Ma in the Nihewan basin, China, further confirms thepersistence of hominin toolmakers across a wide variety ofclimatic and ecological settings (Ferring et al. 2011; Zhu etal. 2004).

It is unclear whether the stratigraphic persistence of theOldowan within East Africa beginning roughly 2 Ma and itsspread to other regions can be attributed to the emergenceof H. erectus or whether these phenomena are the product oftwo contemporaneous Oldowan toolmaking species, Homohabilis and H. erectus. It does, nonetheless, denote a shift inthe regularity of Oldowan toolmaking from ∼2.0 to 1.8 Ma,the persistence of Oldowan hominins across numerous en-vironmental transitions within East Africa, and the spread ofthese populations into novel environments and diverse cli-matic regimes. All of this occurred after the early development

of the Oldowan within highly dynamic East African environ-ments between 2.6 and 2.0 Ma and appears to have been inplace by the time of the aridity pulse in East Africa.

Adaptability as a Framework for the Analysis ofAdaptations in Homo

This analysis of the environmental and behavioral adaptationsof early Homo and H. erectus point to the importance ofadaptability in diet, foraging, and mobility, that is, resiliencein the face of moist or arid habitats, abundant or scarce re-sources, and large lakes or dry landscapes. The picture thatemerges is one of shifting variability in the ecological milieusuperimposed on an overall drying trend across the time in-terval from 3 to 1.5 Ma. If this current understanding ofenvironmental dynamics is correct, it suggests that adapta-bility may be expected in many aspects of the biology of earlyH. erectus and possibly its immediate precursor, especially anovel degree of developmental and physiological plasticity anda spectrum of life history trajectories that promoted the finetuning of biological adaptations in Homo to the dynamics ofits surroundings (Potts 2002). The idea that particular com-binations of genes favoring plasticity can be filtered and se-lected because of the instability of conditions of natural se-lection, especially in times of increased seasonal- toorbital-scale climate variability, is consistent with the hypo-thetical process of variability selection (e.g., deMenocal 2011;Grove 2011; Potts 1996a, 1996b, 1998b; Trauth et al. 2007,2010).

Conclusion

A few final points arise from this analysis of the East Africanenvironmental and behavioral contexts of early Homo andHomo erectus. They are as follows.

Building a Synthesis of Environmental Data

A growing array of environmental indicators—exemplified byd18O, d13C, eolian dust, plant biomarkers, lake sediments, andsapropels—offers insights into the evolutionary context ofPlio-Pleistocene Africa. On first inspection, these records ap-pear to contradict one another. Eolian dust and d13C (in-cluding its application to plant biomarkers) appear to be mostsensitive to the arid aspects of climate variability. This makessense, given that the most arid times in Africa would correlatewith strongest dust productivity and would lead to the pre-cipitation of carbonate nodules in soils, the primary focus ofstable carbon isotope analyses. During times of large or deeplakes, when aquatic deposits are dominant, soils and terrestrialfossil animals (whose teeth also provide d13C data) are oftennot even recorded in sedimentary exposures. In a parallel vein,analyses that focus on lake sediments are sensitive to thewettest times. Deep-sea d18O offers a picture of global oceantemperature and tends to highlight global cooling and a trendtoward rising amplitude in cold-warm or glacial-interglacial


oscillation over the past 6 Myr. Passey et al.’s (2010) paleo-temperature proxy based on d13C of Turkana basin paleosolsis at odds with the global picture of overall cooling and pro-nounced temperature variability. This apparent contradictionmay result from the special conditions under which carbonateprecipitates in paleosols, and thus Passey’s method may sam-ple primarily the arid and hot end member in the range ofenvironmental variability. Finally, sapropels capture an in-triguing pattern in wet-dry climate variability that does notappear in the global ocean record—that is, an alternationbetween high and low climate variability—rather than a con-tinual rise in variability.

The sapropel record in particular appears to offer a prom-ising way of examining the entire moist-arid spectrum ofnortheast African climate. The records of sapropels, eoliandust, plant biomarkers, and lake sediments, furthermore, canall be reconciled with one another through the high/low cli-mate variability framework described in this paper. As notedby Feakins, deMenocal, and Eglinton (2005) with regard tobiomarker isotopes, the overall aridity trend in Africa seemsto be made up of high and low variability intervals; this meansthat an overall aridity trend can be seen as a matter of sam-pling: “The data . . . suggest that changes in northeast Africanvegetation were primarily related to the amplitude of sub-tropical orbital insolation variations, since large-amplitudevegetation variability ca. 3.7 and ca. 1.4 Ma coincided withhigh orbital precession variability, and low variability ca. 2.4Ma occurred during a precessional minimum” (Feakins,deMenocal, and Eglinton 2005:979–980). Likewise, dust rec-ords show that alternating high and low climate variability isa crucial, newly recognized dimension in tropical African en-vironment pertinent to hominin evolution.

The current limitation of the sapropel, dust, and biomarkerrecords as applied to questions of human evolution is thatthey are focused exclusively in eastern and northeastern Af-rica. Development of long-term stratigraphic records of cli-matic, environmental, and ecological dynamics in otherregions of Africa remains a critical need.

Evolutionary Interpretations That Emphasize Habitat orResource Stability Lack Support

At present, the most impressive feature of East African climaterecords during the evolution of early Homo and H. erectus isthe series of lengthy eras of pronounced moist-arid variability.In an era characterized by alternations between high and lowvariability, the dominance of high climate variability between3.0 and 1.5 Ma makes it almost inevitable that the most no-table first and last appearances in the hominin and faunalrecords are associated with instability in the conditions ofnatural selection.

Interpretations that invoke habitat and resource stability asan important factor in the evolution of early Homo or in theorigination and dispersal of H. erectus are at odds with thissynthetic picture of East African climate variability. The cur-

rent FADs for these events occur during intervals of strongclimate variability, which, along with local tectonic and vol-canic events, caused substantial modifications to resourcelandscapes. The growing body of data from East African hom-inin-occupied basins confirms that landscapes and resourcesunderwent recurrent, high-amplitude alterations during thefocal time interval. Within this dynamic context, of growinginterest is the presence of refugia that offered basic needs ofwater and food options during eras of strong climate vari-ability throughout East Africa.

Evolutionary Interpretations That Emphasize AdaptiveVersatility Are Well Supported

The environmental and behavioral evidence summarized herehighlights the importance of ecological adaptability and phys-iological plasticity as an element, perhaps even the centralissue, in the evolution of Homo. An evolved responsivenessto environmental variability would seem to have played acentral role in the adaptive and phylogenetic history of Plio-Pleistocene Homo. It is important, nonetheless, to recognizethat Pleistocene Homo underwent critical behavioral, ecolog-ical, and life history transitions over the past 1 Myr. It is thussensible to avoid the temptation of attributing all that is im-portant in the adaptive history of Homo sapiens to the originof the genus or of its longest-enduring lineage, H. erectus.

Acknowledgments

I am grateful to the following colleagues for their collabo-ration and shared data over the past decade: Kay Behrens-meyer, Chris Campisano, Andy Cohen, Alan Deino, PeterdeMenocal, Tim Eglinton, Sarah Feakins, Craig Feibel, BernieOwen, and Tom Plummer. I also thank Jennifer Clark forassistance with the figures and John Kingston for enablingthe use of his data compilation in figure 2. Research reportedhere was supported by the Peter Buck Fund for Human Or-igins Research, the Ruth and Vernon Taylor Foundation, Na-tional Science Foundation Hominid Program grant BCS-0128511, and the Smithsonian Institution’s Human OriginsProgram. Logistical support and collaboration with the Na-tional Museums of Kenya have provided a strong base forseveral of the studies reported here. I am grateful to LeslieAiello and Susan Anton, with assistance from Laurie Obbink,for organizing the Wenner-Gren symposium in Sintra, Por-tugal, and to the symposium participants for stimulating dis-cussions.

References CitedAiello, Leslie C., and Peter Wheeler. 1995. The expensive-tissue hypothesis:


Ambrose, Stanley H., and Nancy E. Sikes. 1991. Soil organic carbon isotopicevidence for vegetation change in the Kenya Rift Valley. Science 253:1402–1405.


Anton, Susan C. 2003. Natural history of Homo erectus. Yearbook of PhysicalAnthropology 46:126–170.


Anton, Susan C., William R. Leonard, and Marcia L. Robertson. 2002. Anecomorphological model of the initial hominid dispersal from Africa. Jour-nal of Human Evolution 43:773–785.


Anton, Susan C., and Carl C. Swisher III. 2004. Early dispersals of Homo fromAfrica. Annual Review of Anthropology 33:271–296.

Ashley, Gail M. 2007. Orbital rhythms, monsoons, and playa lake response,Olduvai basin, equatorial East Africa (ca. 1.85–1.74 Ma). Geology 35:1091–1094.

Behrensmeyer, Anna K., Nancy E. Todd, Richard Potts, and Geraldine E.McBrinn. 1997. Late Pliocene faunal turnover in the Turkana basin, Kenyaand Ethiopia. Science 278:1589–1594.

Binford, Lewis R. 1981. Bones: ancient men and modern myths. New York:Academic Press.

Blumenschine, Robert J. 1991. Hominid carnivory and foraging strategies, andthe socio-economic function of early archaeological sites. PhilosophicalTransactions of the Royal Society B: Biological Sciences 334:211–221.

Bobe, Rene, Zeresenay Alemseged, and Anna K. Behrensmeyer, eds. 2007.Hominin environments in the East African Pliocene: an assessment of the faunalevidence. New York: Springer.

Bobe, Rene, and Anna K. Behrensmeyer. 2004. The expansion of grasslandecosystems in Africa in relation to mammalian evolution and the origin ofthe genus Homo. Palaeogeography, Palaeoclimatology, Palaeoecology 207:399–420.

Bobe, Rene, Anna K. Behrensmeyer, and Ralph E. Chapman. 2002. Faunalchange, environmental variability and late Pliocene hominin evolution.Journal of Human Evolution 42:475–497.

Bobe, Rene, Anna K. Behrensmeyer, Gerald G. Eck, and John M. Harris. 2007.Patterns of abundance and diversity in late Cenozoic bovids from the Tur-kana and Hadar basins, Kenya and Ethiopia. In Hominin environments inthe East African Pliocene: an assessment of the faunal evidence. Rene Bobe,Zeresenay Alemseged, and Anna K. Behrensmeyer, eds. Pp. 129–157. NewYork: Springer.

Bobe, Rene, and Gerald G. Eck. 2001. Response of African bovids to Plioceneclimatic change. Paleobiology 27(suppl. 2):1–47.

Bonnefille, Raymonde. 1995. A reassessment of the Plio-Pleistocene pollenrecord of East Africa. In Paleoclimate and evolution, with emphasis on humanorigins. Elizabeth S. Vrba, George H. Denton, Timothy C. Partridge, andLloyd H. Burckle, eds. Pp. 299–310. New Haven, CT: Yale University Press.

Bramble, Dennis M., and Daniel E. Lieberman. 2004. Endurance running andthe evolution of Homo. Nature 432:345–352.

Braun, David R., John W. K. Harris, Naomi E. Levin, Jack T. McCoy, AndyI. R. Herries, Marion K. Bamford, Laura C. Bishop, Brian G. Richmond,and Mzalendo Kibunjia. 2010. Early hominin diet included diverse terres-trial and aquatic animals 1.95 Ma in East Turkana, Kenya. Proceedings ofthe National Academy of Sciences of the USA 107:10002–10007.


Campisano, Christopher J., and Craig S. Feibel. 2007. Connecting local en-vironmental sequences to global climate patterns: evidence from the hom-inin-bearing Hadar Formation, Ethiopia. Journal of Human Evolution 53:515–527.

Carbone, Chris, Amber Teacher, and J. Marcus Rowcliffe. 2007. The costs ofcarnivory. PLoS Biology 5(2):e22, doi:10.1371/journal.pbio.0050022.

Cerling, Thure E. 1992. Development of grasslands and savannas in East Africaduring the Neogene. Palaeogeography, Palaeoclimatology, Palaeoecology 97:241–247.

Cerling, Thure E., and Richard L. Hay. 1986. An isotopic study of paleosolcarbonates from Olduvai Gorge. Quaternary Research 25:63–78.

Cerling, Thure E., Jonathan G. Wynn, Samuel A. Andanje, Michael I. Bird,David Kimutai Korir, Naomi E. Levin, William Mace, Anthony N. Macharia,Jay Quade, and Christopher H. Remien. 2011. Woody cover and homininenvironments in the past 6 million years. Nature 476:51–56.

Clemens, Steven C. 1998. Dust response to seasonal atmospheric forcing:proxy evaluation and forcing. Paleoceanography 13:471–490.

de Heinzelin, Jean, J. Desmond Clark, Tim White, William Hart, Paul Renne,

Giday WoldeGabriel, Yonas Beyene, and Elisabeth Vrba. 1999. Environmentand behavior of 2.5-million-year-old Bouri hominids. Science 284:625–629.

Deino, Alan L. 2012. 40Ar/39Ar dating of Bed I, Olduvai Gorge, Tanzania, andthe chronology of early Pleistocene climate change. Journal of Human Evo-lution 63:251–273.

Deino, Alan L., John D. Kingston, Jonathan M. Glen, Robert K. Edgar, andAndrew Hill. 2006. Precessional forcing of lacustrine sedimentation in thelate Cenozoic Chemeron basin, Central Kenya Rift. Earth and PlanetaryScience Letters 247:41–60.

Delagnes, Anne, and Helene Roche. 2005. Late Pliocene hominid knappingskills: the case of Lokalalei 2C, West Turkana, Kenya. Journal of HumanEvolution 48:435–472.

de la Torre, Ignacio. 2004. Omo revisited. Current Anthropology 45:439–465.deMenocal, Peter B. 1995. Plio-Pleistocene African climate. Science 270:53–

59.———. 2004. African climate change and faunal evolution during the Plio-

cene-Pleistocene. Earth and Planetary Science Letters 220:3–24.———. 2011. Climate and human evolution. Science 331:540–541.Domınguez-Rodrigo, Manuel, Rebeca Barba, and Charles P. Egeland. 2007.

Deconstructing Olduvai: a taphonomic study of the Bed I sites. Dordrecht:Springer.

Domınguez-Rodrigo, Manuel, Travis Rayne Pickering, Sileshi Semaw, andMichael J. Rogers. 2005. Cutmarked bones from Pliocene archaeologicalsites at Gona, Ethiopia: implications for the function of the world’s oldeststone tools. Journal of Human Evolution 48:109–121.

Donadio, Emiliano, and Steven W. Buskirk. 2006. Diet, morphology, andinterspecific killing in carnivora. American Naturalist 167:524–536.

Feakins, Sarah J., Peter B. deMenocal, and Timothy I. Eglinton. 2005. Bio-marker records of late Neogene changes in northeast African vegetation.Geology 33:977–980.

Feibel, Craig S. 1997. Debating the environmental factors in hominid evo-lution. Geological Society of America Today 7(3):1–7.

Feibel, Craig S., John M. Harris, and Francis H. Brown. 1991. Palaeoenvi-ronmental context for the late Neogene of the Turkana basin. In The fossilungulates: geology, fossil artiodactyls, and palaeoenvironments, vol. 3 of KoobiFora research project. John M. Harris, ed. Pp. 321–370. Oxford: Clarendon.

Ferraro, Joseph V. 2007. Broken bones and smashed stones: on the foragingecology of Oldowan hominins. PhD dissertation, University of California,Los Angeles.

Ferring, Reid, Oriol Oms, Jordi Agustı, Francesco Berna, Medea Nioradze,Teona Shelia, Martha Tappen, Abesalom Vekua, David Zhvania, and DavidLordkipanidze. 2011. Earliest human occupations at Dmanisi (GeorgianCaucasus) dated to 1.85–1.78 Ma. Proceedings of the National Academy ofSciences of the USA 108:10432–10436.

Frost, Stephen R. 2007. African Pliocene and Pleistocene cercopithecid evo-lution and global climatic change. In Hominin environments in the EastAfrican Pliocene: an assessment of the faunal evidence. Rene Bobe, ZeresenayAlemseged, and Anna K. Behrensmeyer, eds. Pp. 51–76. New York: Springer.

Ghosh, Prosenjit, Jess Adkins, Hagit Affek, Brian Balta, Weifu Guo, Edwin A.Schauble, Dan Schrag, and John M. Eiler. 2006. 13C-18O bonds in carbonateminerals: a new kind of paleothermometer. Geochimica et CosmochimicaActa 70:1439–1456.

Goldman-Neuman, Talia, and Erella Hovers. 2009. Methodological issues inthe study of Oldowan raw material selectivity: insights from A. L. 894(Hadar, Ethiopia). In Interdisciplinary approaches to the Oldowan. ErellaHovers and David R. Braun, eds. Pp. 71–84. Dordrecht: Springer.

———. 2011. Raw material selectivity in Late Pliocene Oldowan sites in theMakaamitalu basin, Hadar, Ethiopia. Journal of Human Evolution 62:353–366.

Grove, Matt. 2011. Speciation, diversity, and Mode 1 technologies: the impactof variability selection. Journal of Human Evolution 61:306–319.

Gursky-Doyen, Sharon, and K. A. I. Nekaris, eds. 2006. Primate anti-predatorstrategies. New York: Springer.

Hay, Richard L. 1976. Geology of the Olduvai Gorge. Berkeley: University ofCalifornia Press.

Herries, Andy I. R., Darren Curnoe, and Justin W. Adams. 2009. A multi-disciplinary seriation of early Homo and Paranthropus bearing palaeocavesin southern Africa. Quaternary International 202:14–28.

Herries, Andy I. R., and John Shaw. 2011. Palaeomagnetic analysis of theSterkfontein palaeocave deposits: age implications for the hominin fossilsand stone tool industries. Journal of Human Evolution 60:523–539.

Hilgen, Frederik J., Hayfaa Abdul Aziz, Wout Krijgsman, Cor G. Langereis,Lucas J. Lourens, Johan E. Meulenkamp, Isabella Raffi, et al. 1999. Present


status of the astronomical (polarity) time-scale for the Mediterranean LateNeogene. Philosophical Transactions of the Royal Society A: Mathematical,Physical and Engineering Sciences 357:1931–1947.


Howell, F. Clark, Paul Haesaerts, and Jean de Heinzelin. 1987. Depositionalenvironments, archeological occurrences and hominids from members Eand F of the Shungura Formation (Omo basin, Ethiopia). Journal of HumanEvolution 16:665–700.

Isaac, Glynn L. 1984. The archaeology of human origins: studies of the LowerPleistocene in East Africa. Advances in World Archaeology 3:1–87.

Joordens, Josephine C. A., Hubert B. Vonhof, Craig S. Feibel, Lucas J. Lourens,Guillaume Dupont-Nivet, Jeroen H. J. L. van der Lubbe, Mark J. Sier, GarethR. Davies, and Dick Kroon. 2011. An astronomically-tuned climate frame-work for hominins in the Turkana basin. Earth and Planetary Science Letters307:1–8.

Kelt, Douglas A., and Dirk Van Vuren. 1999. Energetic constraints and therelationship between body size and home range area in mammals. Ecology80:337–340.

Kimbel, William H. 2009. The origin of Homo. In The first humans. FrederickE. Grine, John G. Fleagle, and Richard E. Leakey, eds. Pp. 31–37. New York:Springer.

Kimbel, William H., Robert C. Walter, Donald C. Johanson, Kaye E. Reed,James L. Aronson, Zelalem Assefa, Curtis W. Marean, et al. 1996. LatePliocene Homo and Oldowan tools from the Hadar Formation (Kada Hadarmember), Ethiopia. Journal of Human Evolution 31:549–561.

Kingston, John D. 2007. Shifting adaptive landscapes: progress and challengesin reconstructing early hominid environments. Yearbook of Physical An-thropology 50:20–58.

Kingston, John D., Alan L. Deino, Robert K. Edgar, and Andrew Hill. 2007.Astronomically forced climate change in the Kenyan Rift Valley 2.7–2.55Ma: implications for the evolution of early hominin ecosystems. Journal ofHuman Evolution 53:487–503.

Kuzawa, Christopher W., and Jared M Bragg. 2012. Plasticity in human lifehistory strategy: implications for contemporary human variation and theevolution of genus Homo. Current Anthropology 53(suppl. 6):S369–S382.

Leakey, Mary D. 1971. Olduvai Gorge, vol. 3. Cambridge: Cambridge Uni-versity Press.

Lepre, Christopher J., Rhonda L. Quinn, Josephine C. A. Joordens, Carl C.Swisher III, and Craig S. Feibel. 2007. Plio-Pleistocene facies environmentsfrom the KBS member, Koobi Fora Formation: implications for climatecontrols on the development of lake-margin hominin habitats in the north-east Turkana basin (northwest Kenya). Journal of Human Evolution 53:504–514.

Leroy, Suzanne, and Lydie Dupont. 1994. Development of vegetation andcontinental aridity in northwestern Africa during the Late Pliocene: thepollen record of ODP Site 658. Palaeogeography, Palaeoclimatology, Palaeo-ecology 109:295–316.

Levin, Naomi E., Francis H. Brown, Anna K. Behrensmeyer, Rene Bobe, andThure E. Cerling. 2011. Paleosol carbonates from the Omo group: isotopicrecords of local and regional environmental change in East Africa. Palaeo-geography, Palaeoclimatology, Palaeoecology 307:75–89.

Levin, Naomi E., Jay Quade, Scott W. Simpson, Sileshi Semaw, and MichaelRogers. 2004. Isotopic evidence for Plio-Pleistocene environmental changeat Gona, Ethiopia. Earth and Planetary Science Letters 219:93–110.

Lewis, Margaret E., and Lars Werdelin. 2007. Patterns of change in the Plio-Pleistocene carnivorans of eastern Africa. In Hominin environments in theEast African Pliocene: an assessment of the faunal evidence. Rene Bobe, Zere-senay Alemseged, and Anna K. Behrensmeyer, eds. Pp. 77–105. New York:Springer.

Lordkipanidze, David, Tea Jashashvili, Abesalom Vekua, Marcia S. Ponce deLeon, Christoph P. E. Zollikofer, G. Philip Rightmire, Herman Pontzer, etal. 2007. Postcranial evidence from early Homo from Dmanisi, Georgia.Nature 449:305–310.

Maslin, Mark A., and Martin H. Trauth. 2009. Plio-Pleistocene East Africanpulsed climate variability and its influence on early human evolution. InThe first humans: origin and early evolution of the genus Homo. FrederickE. Grine, John G. Fleagle, and Richard E. Leakey, eds. Pp. 151–158. NewYork: Springer.


Nagy, Ken A., Isabelle A. Girard, and Tracey K. Brown. 1999. Energetics of

free-ranging mammals, reptiles, and birds. Annual Review of Nutrition 19:247–277.

Owen, R. Bernhart, Richard Potts, and Anna K. Behrensmeyer. 2009. Replyto the comment on “Diatomaceous sediments and environmental changein the Pleistocene Olorgesailie Formation, southern Kenya Rift Valley” byR. Bernhart Owen, Richard Potts, Anna K. Behrensmeyer, and Peter Ditch-field. Palaeogeography, Palaeoclimatology, Palaeoecology 282:147–148.

Passey, Benjamin H., Naomi E. Levin, Thure E. Cerling, Francis H. Brown,and John M. Eiler. 2010. High-temperature environments of human evo-lution in East Africa based on bond ordering in paleosol carbonates. Pro-ceedings of the National Academy of Sciences of the USA 107:11245–11249.

Pickering, Robyn, Paul H. G. M. Dirks, Zubair Jinnah, Darryl J. de Ruiter,Steven E. Churchill, Andy I. R. Herries, Jon D. Woodhead, John C. Hell-strom, and Lee R. Berger. 2011. Australopithecus sediba at 1.977 Ma andimplications for the origins of the genus Homo. Science 333:1421–1423.

Plummer, Thomas W. 2004. Flaked stones and old bones: biological andcultural evolution at the dawn of technology. Yearbook of Physical Anthro-pology 47:118–164.

Plummer, Thomas W., Peter W. Ditchfield, Laura C. Bishop, John D. Kingston,Joseph V. Ferraro, David R. Braun, Fritz Hertel, and Richard Potts. 2009.Oldest evidence of toolmaking hominins in a grassland-dominated ecosys-tem. PLoS ONE 4(9):e7199, doi:10.1371/journal.pone.0007199.


Pontzer, Herman, Campbell Rolian, G. Philip Rightmire, Tea Jashashvili, Mar-cia S. Ponce de Leon, David Lordkipanidze, and Christoph P. E. Zollikofer.2010. Locomotor anatomy and biomechanics of the Dmanisi hominins.Journal of Human Evolution 58:492–504.

Potts, Richard. 1988. Early hominid activities at Olduvai. New York: Aldine.———. 1991. Why the Oldowan? Plio-Pleistocene toolmaking and the trans-

port of resources. Journal of Anthropological Research 47:153–176.———. 1996a. Evolution and climate variability. Science 273:922–923.———. 1996b. Humanity’s descent: the consequences of ecological instability.

New York: Morrow.———. 1998a. Environmental hypotheses of hominin evolution. Yearbook of

Physical Anthropology 41:93–136.———. 1998b. Variability selection in hominid evolution. Evolutionary An-

thropology 7:81–96.———. 2002. Complexity and adaptability in human evolution. In Probing

human origins. Morris Goodman and Anne Moffat, eds. Pp. 33–57. Cam-bridge, MA: American Academy of Arts and Sciences.

———. 2007. Environmental hypotheses of Pliocene human evolution. InHominin environments in the East African Pliocene: an assessment of the faunalevidence. Rene Bobe, Zeresenay Alemseged, and Anna K. Behrensmeyer,eds. Pp. 25–50. New York: Springer.

Potts, Richard, and Pat Shipman. 1981. Cutmarks made by stone tools onbones from Olduvai Gorge, Tanzania. Nature 291:577–580.

Potts, Richard, and Robin Teague. 2010. Behavioral and environmental back-ground to “Out of Africa I” and the arrival of Homo erectus in East Asia.In Out of Africa I: the first hominin colonization of Eurasia. John G. Fleagle,John J. Shea, Frederick E. Grine, Andrea L. Baden, and Richard E. Leakey,eds. Pp. 67–85. New York: Springer.

Quade, Jay, Naomi Levin, Sileshi Semaw, Dietrich Stout, Paul Renne, MichaelRogers, and Scott Simpson. 2004. Paleoenvironments of the earliest stonetoolmakers, Gona, Ethiopia. Geological Society of America Bulletin 116:1529–1544.

Quade, Jay, Naomi E. Levin, Scott W. Simpson, Robert Butler, William C.McIntosh, Sileshi Semaw, Lynnette Kleinsasser, Guillaume Dupont-Nivet,Paul Renne, and Nelia Dunbar. 2008. The geology of Gona, Afar, Ethiopia.In The geology of early humans in the Horn of Africa. Jay Quade and JonathanC. Wynn, eds. Special Paper 446. Pp. 1–31. Boulder, CO: Geological Societyof America.

Quinn, Rhonda L., Christopher J. Lepre, James D. Wright, and Craig S. Feibel.2007. Paleogeographic variations of pedogenic carbonate d13C values fromKoobi Fora, Kenya: implications for floral compositions of Plio-Pleistocenehominin environments. Journal of Human Evolution 53:560–573.

Reed, Kaye E. 1997. Early hominid evolution and ecological change throughthe African Plio-Pleistocene. Journal of Human Evolution 32:289–322.

———. 2008. Paleoecological patterns at the Hadar hominin site, Afar Re-gional State, Ethiopia. Journal of Human Evolution 54:743–768.

Rossignol-Strick, Martine. 1983. African monsoons, an immediate climateresponse to orbital insolation. Nature 304:46–49.


Ruddiman, William F. 2001. Earth’s climate: past and future. New York: W.H. Freeman.

Ruff, Christopher B., Erik Trinkaus, Alan Walker, and Clark Spencer Larsen.1993. Postcranial robusticity in Homo. 1. Temporal trends and mechanicalinterpretation. American Journal of Physical Anthropology 91:21–53.

Ruff, Christopher B., and Alan Walker. 1993. Body size and body shape. InThe Nariokotome Homo erectus skeleton. Alan C. Walker and Richard E.Leakey, eds. Pp. 234–265. Cambridge, MA: Harvard University Press.

Sahnouni, Mohamed, Djillali Hadjouis, Jan van der Made, Abd-El-Kader Der-radji, Antoni Canals, Mohamed Medig, Hocine Belahrech, Zoheir Hari-chane, and Merouane Rabhi. 2002. Further research at the Oldowan siteof Aın Hanech, north-eastern Algeria. Journal of Human Evolution 43:925–937.


Semaw, Sileshi, Michael J. Rogers, Jay Quade, Paul R. Renne, Robert F. Butler,Manuel Domınguez-Rodrigo, Dietrich Stout, William S. Hart, Travis Pick-ering, and Scott W. Simpson. 2003. 2.6-million-year-old stone tools andassociated bones from OGS-6 and OGS-7, Gona, Afar, Ethiopia. Journal ofHuman Evolution 45:169–177.

Sepulchre, Pierre, Gilles Ramstein, Frederic Fluteau, Mathieu Schuster, Jean-Jacques Tiercelin, and Michel Brunet. 2006. Tectonic uplift and easternAfrica aridification. Science 313:1419–1423.


Stout, Dietrich, Jay Quade, Sileshi Semaw, Michael Rogers, and Naomi E.Levin. 2005. Raw material selectivity of the earliest stone tool makers atGona, Afar, Ethiopia. Journal of Human Evolution 48(4):365–380.

Stout, Dietrich, Sileshi Semaw, Michael J. Rogers, and Dominique Cauche.2010. Technological variation in the earliest Oldowan from Gona, Afar,Ethiopia. Journal of Human Evolution 58:474–491.

Trauth, Martin H., Juan C. Larrasoana, and Manfred Mudelsee. 2009. Trends,rhythms and events in Plio-Pleistocene African climate. Quaternary ScienceReviews 28:399–411.

Trauth, Martin H., Mark A. Maslin, Alan L. Deino, Annett Junginger, MosesLesoloyia, Eric O. Odada, Daniel O. Olago, Lydia A. Olaka, Manfred R.Strecker, and Ralph Tiedemann. 2010. Human evolution in a variable en-vironment: the amplifier lakes of eastern Africa. Quaternary Science Reviews29:2981–2988.

Trauth, Martin H., Mark A. Maslin, Alan L. Deino, and Manfred R. Strecker.2005. Late Cenozoic moisture history of East Africa. Science 309:2051–2053.

Trauth, Martin H., Mark A. Maslin, Alan L. Deino, Manfred R. Strecker,Andreas G. N. Bergner, and Miriam Duhnforth. 2007. High- and low-latitude forcing of Plio-Pleistocene East African climate and human evo-lution. Journal of Human Evolution 53:475–486.

Vrba, Elisabeth S. 1985. Environment and evolution: alternative causes of thetemporal distribution of evolutionary events. South African Journal of Science81:229–236.

———. 1988. Late Pliocene climatic events and hominid evolution. In Evo-lutionary history of the “robust” australopithecines. Frederick E. Grine, ed.Pp. 405–426. New York: Aldine.

———. 1992. Mammals as a key to evolutionary theory. Journal of Mam-malogy 73:1–28.

———. 1994. An hypothesis of heterochrony in response to climatic coolingand its relevance to early hominid evolution. In Integrative paths to the past.Robert S. Corruccini and Russell L. Ciochon, eds. Pp. 345–376. EnglewoodCliffs, NJ: Prentice Hall.

———. 1995a. The fossil record of African antelopes (Mammalia, Bovidae)in relation to human evolution and paleoclimate. In Paleoclimate and evo-lution, with emphasis on human origins. Elizabeth S. Vrba, George H. Den-ton, Timothy C. Partridge, and Lloyd H. Burckle, eds. Pp. 385–424. NewHaven, CT: Yale University Press.

———. 1995b. On the connections between paleoclimate and evolution. InPaleoclimate and evolution, with emphasis on human origins. Elizabeth S.Vrba, George H. Denton, Timothy C. Partridge, and Lloyd H. Burckle, eds.Pp. 24–45. New Haven, CT: Yale University Press.

Vrba, Elizabeth S., George H. Denton, Timothy C. Partridge, and Lloyd H.Burckle, eds. 1995. Paleoclimate and evolution, with emphasis on humanorigins. New Haven, CT: Yale University Press.

Wehausen, Rolf, and Hans-Jurgen Brumsack. 1999. Cyclic variations in thechemical composition of eastern Mediterranean Pliocene sediments: a keyfor understanding sapropel formation. Marine Geology 153:161–176.

Wynn, Jonathan G. 2000. Paleosols, stable carbon isotopes, and paleoenvi-ronmental interpretation of Kanapoi, Northern Kenya. Journal of HumanEvolution 39:411–432.

Zachos, James, Mark Pagani, Lisa Sloan, Ellen Thomas, and Katharina Billups.2001. Trends, rhythms, and aberrations in global climate 65 Ma to present.Science 292:686–693.

Zhu, Rixiang, Richard Potts, Fei Xie, Kenneth A. Hoffman, Chenglong L.Deng, C. D. Shi, Yongxin X. Pan, et al. 2004. New evidence regarding theearliest human presence at high northern latitudes in northeast Asia. Nature431:559–562.



Dental Evidence for the Reconstruction ofDiet in African Early Homo

by Peter S. Ungar

The reconstruction of diet is important for understanding the paleoecology and evolution of early hominins. Thispaper reviews and colligates the fossil evidence for diets of early Homo (Homo habilis, Homo rudolfensis, Homoerectus), particularly that related to tooth size, shape, structure, and wear. Technological innovations and new findshave led to improved understandings of feeding adaptations and food preferences in the earliest members of ourgenus. Differences in dental topography between these species and the australopiths, for example, have been doc-umented, as have differences in microwear textures between H. habilis and H. erectus. These and other lines ofevidence suggest a probable shift in diet in early Homo, and especially H. erectus, compared with their australopithforebears, with a broadened subsistence base to include foods with a wider range of fracture properties. Studies todate also make clear that while much remains to be done, early hominin teeth hold the potential to provide moredetail about diet and confidence in our reconstructions as samples increase, our understanding of functional mor-phology improves, and other methods of analysis are applied to the fossils we have.

Introduction

Diet is the most direct connection an organism has with itsenvironment. Changing environments, with associated newchallenges and opportunities, have surely driven changes inearly hominin diets and with them the evolution of our genus.There are many ideas concerning the role of diet changes inthe transition to early Homo. These include elegant and well-reasoned models based on nutritional studies combined withdirect analogy to living peoples or nonhuman primates or oncontextual evidence such as archeological remains and pa-leoenvironmental indicators (e.g., Aiello and Wheeler 1995;Isaac 1978; O’Connell, Hawkes, and Jones 1999; Wranghamet al. 1999; Zihlman and Tanner 1978). Some have suggestedthat meat eating was key to human evolution, and others haveopined that plant foods such as underground storage organswere more important. Yet others have proposed that foodpreparation with tools and cooking were critical elements forthe evolution of human diet.

Another issue that has been raised revolves around whereto draw the line between australopith-like and humanlikediets. Leakey, Napier, and Tobias (1964) drew that line withHomo habilis. By including this presumed-toolmaking hom-inin in Homo, these authors recognized H. habilis as belonging

Peter Ungar is Distinguished Professor and Chair, Department ofAnthropology, University of Arkansas (Old Main 330, Fayetteville,Arkansas 72701, U.S.A. [[email protected]]). This paper wassubmitted 12 XII 11, accepted 14 V 12, and electronically published23 VIII 12.

to the same adaptive zone as later members of the genus (seeMayr 1950). Wood and Collard (1999) later challenged this,suggesting, in part on the basis of presumed dietary adap-tations, that the shift really began with Homo erectus. Theyeven proposed that both H. habilis and Homo rudolfensis betransferred to the genus Australopithecus on this basis (seeAnton 2012; Holliday 2012).

Models for key changes in the evolution of human diet canbe viewed as hypotheses, some of which may be testable usingthe fossils themselves (Ungar, Grine, and Teaford 2006). Thereare many potential lines of evidence to examine, includingdental microwear and the sizes, shapes, internal architecture,and chemical compositions of the teeth and jaws (see Ungarand Sponheimer 2011 for review). This paper will review fourof the more commonly examined lines of evidence—toothsize, shape, structure, and microwear—and what these mighttell us about diets of the earliest members of our genus. Thefocus will be on early Homo (i.e., fossils often attributed toH. habilis, H. rudolfensis, and H. erectus) from Africa andwhether these provide evidence for dietary differences amongthese species and between them and the australopiths (Aus-tralopithecus spp., Paranthropus spp.). While data are ex-tremely limited and interpretations must be considered ten-tative, the fossil evidence is consistent with a broadening ofthe subsistence base by early Homo and especially H. erectusto include more tough foods.

Tooth Size

Tooth size has been considered an important proxy for earlyhominin diets since Robinson (1954) noted variation among


Ungar Diet in Early Homo S319

species more than half a century ago. He argued that thelarge, flat premolars and molars of Paranthropus robustus werewell suited to crushing and grinding tough vegetation, whereasthe larger front teeth and smaller cheek teeth of Australo-pithecus africanus are consistent with “a more nearly omniv-orous diet, which may have included a fair proportion offlesh” (Robinson 1954:328). He also noted that “Telanthropus”(Homo erectus) from Swartkrans evinced more humanlikedental proportions with smaller molars than either of theaustralopiths from South Africa. And Leakey, Napier, andTobias (1964) included small molars relative to Australo-pithecus as part of their revised diagnosis of the genus Homowhen they first described Homo habilis.

Groves and Napier (1968) followed with a comparativestudy of incisor-to-molar row-length ratios (incisor-molar in-dex). They found that fruit-eating chimpanzees have relativelylarger incisors and smaller molars than do more folivorousgorillas, with orangutans intermediate in both incisor-molarindex and diet. Relatively large incisors were thought to beadapted to husking and ingesting fruits, whereas larger molarswere associated with grinding coarse vegetation. They thenreported that Australopithecus had a similar incisor-molar in-dex to the gorilla, and Paranthropus had an even smaller value.Homo habilis had an incisor-molar index in the range ofchimpanzees and orangutans.

Jolly (1970) soon after proposed his seed-eater hypothesiswherein, and by analogy with the modern gelada, the largemolar teeth and smaller incisors of Paranthropus were inter-preted as adaptations to consume small tough seeds. Aus-tralopithecus and especially H. habilis were said to have movedfrom grass seeds to the consumption of more meat and gath-ered vegetable foodstuffs given that “the trend to back-toothdominance has been partially reversed” (Jolly 1970:23). Whilea specialization on grass seeds by earlier hominins has sincebeen considered unlikely (Dunbar 1976), this idea has beeninfluential in stimulating further research, especially that usingextant analogues as models for early hominin diets. And theidea that some hominins (i.e., Paranthropus boisei) regularlyconsumed C4 resources, such as tropical grasses, has recentlyreceived support from other studies (e.g., Ungar and Spon-heimer 2011).

Incisor Allometry

Most of the work that followed considered front teeth andback teeth separately, as it would otherwise be unclear onwhich part of the dentition selection has acted. Hylander’s(1975) study of anthropoid incisors provides a case in point.He found that residuals from a regression line of incisor rowwidth (taken as a proxy for incisor size) plotted against bodymass predict diet such that frugivorous cercopithecines haverelatively larger front teeth than do folivorous colobines (seealso Goldstein, Post, and Melnick 1978). Frugivorous squirrelmonkeys also have relatively larger front teeth than do morefolivorous howlers, and the same pattern holds when com-

paring gibbons to siamangs and chimpanzees to gorillas. Pri-mates feeding on large, husked fruits benefit from larger in-cisors to process them, whereas those that feed on smallerobjects (e.g., berries, leaves) do not require big front teeth.Larger incisors also increase functional life given wear asso-ciated with increased use. These results have been more orless confirmed in subsequent studies of anthropoids, althoughthe importance of comparing closely related species has beennoted; for example, independent of diet, platyrrhines as agroup have smaller incisors than do catarrhines (Eaglen 1984;see also Ungar 1996). Further, the relationship between dietand incisor size in strepsirhines is not nearly as clear (Eaglen1986).

So where do early hominins, and especially early Homospecies, fit in incisor size studies? This is not an easy questionto answer. Efforts to compare incisor sizes (usually measuredas mesiodistal diameters of I1s) of early hominins are com-plicated by small samples (Ungar, Grine, and Teaford 2006).We can count the number of I1s reported for some hominintaxa with the fingers of one hand, which presents a formidablechallenge given typical size variation of about �20% for ex-tant hominoids (Plavcan 1990). An even greater challenge isthe paucity of associated craniodental and postcranial remainsavailable on which to base body-mass estimates. Indeed, asSmith (1996) has noted, confidence intervals for recon-structed weight averages of most hominin species are so greatthat our allometric studies must be approached with cautionif not skepticism. McHenry’s (1994) average body weight es-timate for H. habilis is kg for males and51.6 � 22.6 31.5 �

kg for females. And for Homo rudolfensis, we have no22.5idea about body size because there are no published postcra-nial bones definitively associated with this hominin, thoughit is common to assign KNM-ER 1472 and KNM-ER 1481(from Koobi Fora, Kenya) to it (Kimbel 2009; Wood 1992).In this light, the value of placing estimates for individualhominin species on a regression plot of tooth size againstbody mass for extant primates is not entirely clear.

Despite these constraints, however, it may still be “heuris-tically interesting” (McHenry and Coffing 2000) to compareincisor size with estimated body weights, understanding thatresults must be considered tentative at best. Figure 1 repre-sents a regression of I1 breadth for a variety of extant catar-rhines with 95% confidence limits as indicated. This confirmsthat relative incisor size does indeed reflect diet differencesamong living Old World higher primates. Average values forhominins (both incisor size and body weight estimates) areplotted for comparison. Data for Australopithecus anamensis,Australopithecus afarensis, and Australopithecus africanus fallon or near the regression line connecting gibbons and gorillas,suggesting moderate incisor size. Australopithecus spp. haveneither the large incisors of chimpanzees and orangutans northe small ones of extant folivorous colobines. Homo habilis,and H. rudolfensis both have enlarged incisors relative to Aus-tralopithecus spp., and relative incisor size in H. erectus isintermediate between those of H. habilis and H. rudolfensis


Figure 1. Incisor allometry. The dashed lines indicate 95% confidence limits of the least squares regression. This figure is modifiedfrom Teaford, Ungar, and Grine (2002) with original data and taxonomic attributions from Coffing et al. (1994), Jungers (1988),Leakey et al. (1995), Ungar and Grine (1991), and Wood (1991).

on the one hand and Paranthropus and modern humans onthe other. If these values are accurate, they suggest an increasein incisor size with the earliest members of the genus Homofollowed by a decrease in H. erectus and ultimately Homosapiens. It should be reiterated, however, that tiny samples( each for H. habilis and H. erectus, for H. ru-n p 2 n p 1dolfensis) and uncertain weight estimates remain a seriouslimitation (see Teaford, Ungar, and Grine 2002).

Molar Allometry

Studies of molar allometry face the same limitations as thosefor incisors (small samples, questionable weight estimates)but are also challenged by uncertain form-function relation-ships. That said, molar size has been considered an importantindicator of adaptive zone (e.g., Leakey et al. 2001; Wood andCollard 1999), and many have suggested an increase in molarsize over time in the australopiths followed by a decrease overtime in the genus Homo (e.g., see Brace, Smith, and Hunt1991; McHenry and Coffing 2000; Teaford and Ungar 2000).As illustrated in figure 2, this holds whether we considerabsolute cheek-tooth surface area or a megadontia quotientthat incorporates reconstructed body size. There has been littleagreement, however, as to with whom the decrease in me-gadontia starts. According to Wood and Collard (1999), H.habilis and H. rudolfensis retain australopith-size teeth, andmolar reduction begins with H. erectus. According to Mc-Henry and Coffing (2000), however, H. habilis and especiallyH. rudolfensis show some reduction of occlusal area relativeto body size when compared with the australopiths. And asAnton (2008) has noted, there is substantial overlap between

Homo habilis s.l. and H. erectus, particularly when specimensfrom outside of Africa are incorporated into analyses.

The traditional explanation for megadontia in the australo-piths, and especially Paranthropus, has been that enlargedcheek teeth provide more surface area to process larger quan-tities of lower-quality and mechanically challenging foods.While Pilbeam and Gould (1974) suggested that differencesin tooth size between hominins could be explained by met-abolic scaling (i.e., larger species might need relatively largerteeth given metabolic requirements), Kay (1975, 1985) hasshown that for primates with given food preferences, toothsize scales isometrically with body size. Thus, differences inrelative tooth size between hominins probably do reflect dif-ferences in diet, especially given the likely overlap in bodymass between hominin species (Jungers 1988; McHenry andCoffing 2000). Indeed, Lucas (2004) has argued from a bio-mechanical perspective that cheek-tooth size should relate tothe external properties of foods, including ingested particlesize, shape, and abrasiveness. A diet dominated by smallerparticles or thinner ones with less volume (sheets or rods)should select for larger teeth to increase the probability offracture. Likewise, abrasive foods should select for larger teethto increase surface area for wear.

But what would explain decreasing molar sizes in the genusHomo? Several hypotheses have been proposed. The tradi-tional explanation is that tooth size decreases as selective pres-sures to maintain larger teeth are relaxed; foods processedwith tools and by cooking require less chewing (e.g., see Brace,Smith, and Hunt 1991). Reduction in cheek-tooth size hasbeen attributed to mutation, drift, or even selection given


Figure 2. Cheek-tooth occlusal areas and megadontia quotients of early hominins. Occlusal areas (the sum of the products ofmesiodistal and buccolingual diameters of P4, M1, and M2) are presented above (A), and megadontia quotients (occlusal areasdivided by 12.15 # body mass0.86) are illustrated below (B). The values in these graphs are from McHenry and Coffing (2000),and taxonomic attributions are as presented by those authors.

savings of energy and raw materials (Brace 1963; Jolly 1970;Smith 1982). Another popular idea is that decreased chewingresulting from tool use led to a reduction of mechanicalstresses on the mandible that are necessary for alveolar bonegrowth (Oppenheimer 1964). Teeth would have evolved tobecome smaller to avoid dental crowding, which can lead toimpaction, malocclusion, and a predisposition for periodontaldisease (see Calcagno and Gibson 1991; Jungers 1978; Sofaer,Bailit, and MacLean 1971). Lucas et al. (2009) suggested yetanother possibility related to the notion that tooth size affectsthe rate of food processing. Smaller cheek teeth could slowdown oral processing to avoid a “potential avalanche of food”given the consumption of items requiring less chewing, suchas meat or plant parts prepared with tools or by cooking.

We may be able to gain some insights by considering re-lationships between occlusal surface area and diet in livingprimates. Given that leaves are tough and that thin sheetsrequire thorough chewing, they should select for larger teeth,and indeed folivores do have longer molars than closely re-lated frugivores for most primate groups (Kay 1977; Vinyard

and Hanna 2005). This does not hold for cercopithecoids,however, as colobines have smaller molars than cercopithe-cines (Kay 1977). It is also unclear why for many primatespecies, males have relatively smaller cheek teeth than females(Harvey, Clutton-Brock, and Kavanagh 1978). Because rela-tive molar size does not track broad diet category the sameway in all groups of extant primates, it is difficult to use thistrait to retrodict diets of fossil hominins, at least until we canexplain differences between higher-level taxa in the patternsobserved (Kay and Cartmill 1977).

So what might explain the unexpected results for cerco-pithecoids? Perhaps the tendency toward smaller teeth in co-lobines relates to the need to avoid dental crowding givenshorter faces. Of course, this raises the question of whethertooth size drives jaw length or whether it is the other wayaround (Brace, Smith, and Hunt 1991). Perhaps there is amodular developmental link between jaw length and molarsize (see Vinyard and Hanna 2005). Indeed, Workman et al.(2002) found that for mice, many quantitative trait loci af-fecting tooth size and jaw shape are the same. This may have


implications for fossil hominins, for which associations be-tween jaw length and tooth size have been noted for sometime (Sofaer 1973). And as McCollum and Sharpe (2001)have argued, there is likely a developmental link between toothform and skull form. Until we have a better understandingof relationships between tooth size and function in livingprimates, however, we are probably best off using other linesof evidence to infer the diets of fossil hominins.

Occlusal Morphology

As Aristotle noted more than 2 millennia ago in De partibusanimalium, “teeth have one invariable office, namely the re-duction of food” (in Ogle’s translation, 1912). Teeth are toolsto burst cell walls and to fracture and fragment foods toincrease the surface area exposed to digestive enzymes. Andbecause natural selection theory dictates that animals shouldevolve the best tools possible for the job, teeth should beformed in a manner suited to overcoming the mechanicaldefenses of the particular foods that an animal eats (Ungar2010).

Dental functional morphologists consider two basic typesof mechanical defense: stress limited and displacement limited(see Lucas 2004; Ungar and Lucas 2010). Foods protected bystress-limited defenses tend to be strong and stiff; it requiressubstantial force per unit area to initiate a crack in them.These foods are also often brittle, so they do not require agreat deal of work to spread a crack once it starts. Examplesinclude hard-brittle nuts and bone. Foods protected by dis-placement-limited defenses, on the other hand, are tough orductile; initiating a crack in such foods is often less of achallenge than propagating it. Examples include many leavesand animal flesh. Some foods, such as softer fruits, are in-termediate in their fracture properties, and many are com-posites with individual elements varying in their mechanicaldefenses. Different dental tool kits are most appropriate forefficient processing of these different types of food. Hard-object feeders, for example, should have blunt but dome-shaped cusps to concentrate force on a small area while atthe same time protecting the tooth itself from fracture. Theseshould oppose concave surfaces formed by basins or staggeredcusps to prevent energy loss due to the spread or movementof food. Folivores and other tough-food feeders, in contrast,would be better served by shear-like offset opposing bladesor crests acting as wedges to create tension at the tips ofadvancing cracks.

Studies of dental morphology have borne out these pre-dictions. Folivorous and insectivorous primates tend to havelonger shearing crests relative to tooth length than do closelyrelated frugivorous species (Kay and Covert 1984; Strait 1993).Further, among frugivores, hard-object feeders have the short-est shearing crests (Meldrum and Kay 1997). As with studiesof incisor size, shearing-crest lengths need to be comparedamong closely related species as, for example, cercopithecoidshave longer crests than platyrrhines independent of diet (Kay

and Ungar 1997). The fact that teeth change shape as theywear must also be considered; shearing crests can be difficultto measure when the landmarks used to define them areobliterated with use. As such, whole-surface characterizationsof dental topography can be especially valuable for distin-guishing species on the basis of diet. Indeed, dental topo-graphic analysis has shown that folivorous monkeys and apeshave, at a given stage of gross wear, steeper sloping surfaceswith greater occlusal relief than do closely related frugivores(Bunn and Ungar 2009; M’Kirera and Ungar 2003; Ungar andBunn 2008; Ungar and M’Kirera 2003).

So what about early hominins? Are there differences inshearing-crest lengths or slope and relief between early Homoand the australopiths that preceded them? While homininspecies have been distinguished on the basis of occlusal form(see Bailey and Wood 2007 and references therein), little re-search to date has focused on the functional morphology oftheir crown shapes. Studies of shearing-crest length do notwork well with early hominins because their molars typicallyhave bulbous cusps that lack distinct crests for measurement.Also, because crest lengths are difficult or impossible to mea-sure on worn or damaged teeth, there is an insufficient sampleof early Homo specimens for such analyses, especially giventhat studies typically focus on a single tooth type (e.g., M2s)for comparability of results. Still, it has been noted that whileearly Homo species did not have the long sharp crests seenin extant folivores, their cheek teeth do appear to be lessbunodont than those of australopiths (Teaford, Ungar, andGrine 2002; Wood and Strait 2004).

This has been confirmed with a dental topographic analysis(Ungar 2004, 2007) of early Homo from Africa, though thepaucity of available specimens, even worn ones, has made itnecessary to combine early members of the genus into a singlesample (see table 1 for taxonomic attributions of individualspecimens). Results of comparisons of dental topography ofM2s of early Homo with Australopithecus afarensis and extantchimpanzees and gorillas are illustrated in figure 3. The com-bined early Homo sample falls between the two African apesin average surface slope and topographic relief for all but themost worn specimens, and relief is significantly less than thatof gorillas. On the other hand, the average occlusal slope forearly Homo is significantly greater than that for A. afarensis.Degree of difference in surface slope and occlusal relief valuesat given stages of wear for the two hominin samples are onthe same order as or slightly less than those between gorillasand chimpanzees. This suggests a degree of difference in di-etary adaptation between early Homo and A. afarensis com-parable with or slightly less than that seen between the extantAfrican apes (with appropriate caveats for sample size andcombining samples).

So what do differences in occlusal topography between earlyHomo and Australopithecus mean in terms of dietary adap-tations? Chimpanzees and gorillas consume many foods incommon where the two are sympatric; they differ mostly attimes of fruit scarcity, when gorillas fall back on tougher, more


Table 1. Specimens of early Homo used in analyses

Analysis

Species Occlusal morphology Microwear texture

Homo erectus KNM-ER 806, KNM-ER 992, KNM-WT 15000, OH 22 KNM-BK 8518, KNM-ER 807, KNM-ER 820, KNM-ER 992,KNM-ER 1808, KNM-WT 15000, OH 60, SK 15

Homo habilis OH 16 OH 4, OH 7x, OH 15, OH 16, OH 41, OH 65, OH 67, OH69, OH 70, Stw 15

Homo rudolfensis KNM-ER 1506, KNM-ER 1802Homo sp. KNM-ER 3734

Note. Analyses are described in the text. Attributions follow authors referenced in each section (see Anton 2012 for additional discussion).

Figure 3. Dental topographic analysis data. Mean occlusal relief(A) and slope (B) values by wear stage (see Ungar 2004) for taxaas indicated in the legend. The data illustrated are from Ungar(2004), with early Homo specimen attributions as indicated intable 1.

fibrous plant parts such as leaves and stems, whereas chim-panzees continue to exploit available ripe, succulent fruits(e.g., Remis 1997; Tutin et al. 1991). It may be that earlyHomo and A. afarensis diets likewise differed subtly, with sub-stantial overlap in food types, or at least in their fractureproperties. The blunt-cusped molars of Australopithecuswould have been somewhat better suited to crushing hard-brittle foods because crown shape would have made theseteeth resistant to fracture under heavy stress (Berthaume etal. 2010). Early Homo cheek teeth have less bulbous cuspsand more sloping occlusal surfaces that would have allowedthem to shear or slice tough items more efficiently. This ofcourse does not tell us whether differences in dental-dietaryadaptations between early Homo and the australopiths reflect

differences in preferred foods or fallback items, though dentalmicrowear (see below) may provide additional evidence tohelp us resolve this.

Dental Structure

Studies of early hominin dental structure are beginning toexperience a renaissance of sorts given both new methods ofcharacterization and new theories for interpreting it. Researchhas focused on the thickness of the enamel cap and on thelayout of its microstructure (see Swain and Xue 2009; Teaford2007b for review).

Enamel Thickness

For more than half a century, the australopiths have beenrecognized to have had a relatively thick cap of enamel cov-ering their teeth (e.g., Robinson 1956). Simons and Pilbeam(1972) suggested that this was an adaptation to lengthen theuse life of the dentition given rapid wear with the consump-tion of tough, grit-laden foods on the ground (see also Machoand Spears 1999). Kay (1981) countered that there is no ten-dency among living apes or Old World monkeys for terrestrialspecies to have thicker enamel than arboreal ones. And thefact that orangutans, the most arboreal of the great apes, notonly have thicker tooth enamel (at least by traditional mea-sures) than African apes but tend to have less worn teeth thanchimpanzees or gorillas suggests that this trait need not be acompensatory mechanism for increased wear (Dean, Jones,and Pilley 1992; but see Kono 2004).

The alternative explanation is that thickened enamel inearly hominins provided structural reinforcement tostrengthen teeth against breakage given forces generated by adiet of hard foods (Kay 1981). Tooth and food are in a “deathmatch” as nature selects for strength in both—teeth mustbreak foods without themselves being broken (Ungar 2008).And indeed, primates that consume hard objects tend to havethicker molar enamel than do closely related forms that eatsofter foods (Dumont 1995). This makes sense, especiallygiven that tooth crowns are bilayered with hard-brittle enameloverlaying more compliant dentin; hard-object feeders shouldhave thicker enamel because heavy loads would make thin


coats more prone to flex and cause tensile stresses leading tocracks in teeth (Lucas et al. 2008).

The relationship between enamel thickness and diet is nota simple one, however. It has become increasingly clear thatthe distribution of enamel covering a tooth, not just its av-erage thickness, is important for understanding function (e.g.,Macho and Thackeray 1992; Schwartz 2000). Dental sculptingprovides a case in point. Many mammals have teeth thatactually require wear to function properly (see Fortelius 1985;Ungar 2010). Because enamel is harder than dentin, wear cancause a sharp edge to form where the two tissue types meeton the occlusal surface (Kono 2004; Shimizu 2002; Ungar andM’Kirera 2003). Thinner enamel can result in quicker dentinexposure to facilitate fracture of tough foods. Thus, the mor-phology of the enamel-dentin junction (EDJ) can literallyguide wear to sculpt occlusal surfaces. Exciting new studiesusing x-ray microcomputed tomography (micro-CT) to mapthe distribution of enamel across tooth crowns show greatpromise to help us better understand both form and function(e.g., Gantt et al. 2006; Kono 2004; Kono and Suwa 2008;Olejniczak et al. 2008b; Smith and Tafforeau 2008).

While researchers have begun to map enamel distributionand EDJ morphology in fossil hominins (Olejniczak et al.2008a; Suwa et al. 2009), such studies have not yet beenextended to early Homo. Nevertheless, there has long beenthe general impression among researchers that early membersof our genus had thinner enamel than did the australopiths(see Wallace 1975). Beynon and Wood’s (1986) measurementsof naturally fractured specimens from East Africa support thisassertion in that Homo erectus has the absolutely thinnestenamel of Plio-Pleistocene hominins they analyzed. On theother hand, a recent study of enamel thickness across thegenus using micro-CT on intact specimens indicates sub-stantial variation in early Homo molar teeth. Indeed, somespecimens from South Africa have average and relative enamelthickness values comparable to those of Paranthropus (Smithet al. 2012). Further study and comparisons among taxa willhopefully provide new insights into the functional implica-tions of enamel thickness and distribution in early Homo.

Enamel Microstructure

It is also clear that the physical properties of enamel can varyacross a tooth crown (Cuy et al. 2002). This is due in partto chemical composition but also in large measure to mi-crostructure. Enamel prisms and crystallites within them canbe arranged in many different ways to limit the developmentand spread of cracks through a crown. For example, layersof prisms can decussate or wiggle about in waves between theEDJ and crown surface. Adjacent layers can interweave andbe stacked horizontally, vertically, or in a zigzag fashion toprevent the spread of cracks and strengthen a tooth(Koenigswald and Clemens 1992; Maas and Dumont 1999;Rensberger 2000). This has important implications for struc-tural integrity of enamel under different loading regimes. It

even has implications for crown sculpting with wear, as prismsaligned parallel to an abrasion vector are more resistant thanthose perpendicular to it—witness the enamel ridges on theocclusal surfaces of rhinoceros teeth resulting from differingprism orientations (Rensberger and Koenigswald 1980). Finiteelement modeling has recently been applied to better under-stand effects of prism orientation on surface stresses and wearresistance (Shimizu, Macho, and Spears 2005).

As with studies of enamel thickness, though, only limitedwork has been published on the functional implications ofenamel microstructure variation in early hominins (Machoand Shimizu 2009; Macho et al. 2005). While this researchdocuments variation between species, it has been limited tonaturally fractured surfaces of australopith teeth; no suchfunctional studies have yet been published for early Homo.There remains a great deal of background work to be done,but the potential of dental microstructure for functional stud-ies is clear. Phase contrast x-ray synchrotron microtomog-raphy, which allows imaging of dental microstructure withsubmicron resolution, offers a particularly promising ap-proach (Tafforeau and Smith 2008).

Dental Microwear

Tooth size, shape, and structure can all potentially tell ussomething about hominin dietary adaptation. That said, mor-phological specializations should reflect mechanically chal-lenging foods rather than those that require little work tofracture if both are needed for survival, regardless of whichis eaten more frequently (e.g., Constantino et al. 2009; Ungar2004; Wood and Strait 2004). It should not matter what yourteeth look like if you eat gelatin most of the time, but if youhave to crush hard nuts 5 days out of the year to survive,your teeth had better be able to crush hard nuts. Dentaladaptations can tell us something about the capabilities ofteeth but not how frequently they were used on given typesof food. This is an important distinction, as a tooth may beoverbuilt for most of the foods an animal eats.

In contrast to adaptive lines of evidence, dental microwearreflects the fracture properties of foods that animals eat on aday-to-day basis. Dental microwear analysis is the study ofmicroscopic patterns of wear that form on teeth as the resultof food acquisition and processing. The basic idea is that hard-brittle items crushed between lower and upper teeth shouldcreate pits, whereas tough foods sheared as opposing surfacesslide past one another should result in scratches. And nu-merous studies have shown that hard-object feeding primatestend to have molar microwear surfaces dominated by largepits, folivores have more scratches, and soft-fruit eaters ormixed feeders are intermediate (e.g., see Teaford 1991, 2007a;Ungar et al. 2007). Studies also show that microwear featuresare replaced as they wear away and that patterns can changeover the course of days, weeks, or months (Teaford and Oyen1988). As such, if we have sufficient numbers of individualssampled over time, we should be able to infer something


about dietary preferences and perhaps even foraging strategiesof a fossil species. The combination of adaptive evidence andmicrowear may then provide important insights not just intodiet but into how nature selects for tooth size, shape, andstructure (Ungar 2009).

Microwear research on early hominins at first focused onthe australopiths (e.g., Grine 1981; Ryan 1981; Walker 1981),but studies have recently been expanded to include earlyHomo. Ungar et al. (2006) found that a sample of Homohabilis, African Homo erectus, and early Homo specimens ofuncertain taxonomic affinity from Sterkfontein Member 5and Swartkrans Member 1 had average microwear pit-scratchratios intermediate between those of living hard-object feedersand folivores. Early Homo pit widths also fell near the middleof the extant primate range, suggesting that fossil individualssampled did not specialize on extremely hard, stiff, or toughfoods,1 at least not in the days or weeks before death. Theseresults have been confirmed by microwear-texture analysis, a3-D automated approach to whole-surface characterizationthat combines white-light confocal profilometry and scale-sensitive fractal analysis (Scott et al. 2006; Ungar et al. 2003).The early Homo sample as a whole shows moderate microw-ear-texture complexity (complex surfaces tend to be morepitted) and fill volume (an indication of feature size) both interms of central tendencies and dispersion. Early Homo av-erage values are intermediate between extant hard-object feed-ers and folivores (Ungar and Scott 2009).

A more recent study on an expanded sample of Africanearly Homo specimens by Ungar et al. (2012) suggests dif-ferences between H. habilis and H. erectus (see table 1 fortaxonomic attributions of specimens included in this analysisand fig. 4 for illustration of results). Homo habilis has a higheraverage scale of maximum complexity and textural fill volume(both indicative of fewer small features; see Scott et al. 2006for details) and less variance in texture complexity than H.erectus. In fact, H. erectus has remarkably high variance intexture complexity, matched only by Paranthropus robustusamong the early hominins. The average complexity of H.erectus is lower than that of the P. robustus, however, sug-gesting that while the two may have had similar ranges offood hardness, the latter ate more hard-brittle foods on av-erage. These results indicate that early Homo as a group prob-ably did not regularly consume foods that were either espe-cially hard or tough but that H. erectus may have had asomewhat broader diet, at least in terms of food fractureproperties, than did H. habilis. Microwear-texture data fortwo H. erectus specimens from Dmanisi are consistent withthese results, as both Georgian specimens have complexityvalues within the interquartile range of African H. erectus butabove that for H. habilis (Pontzer et al. 2011).

1. Small pits can form with the consumption of tough foods as prismsare “plucked” from their surrounding matrix because of friction (Teafordand Runestad 1992).

Discussion

The most obvious conclusion we can draw from a review ofthe fossil evidence for diet in African early Homo is that thereis not much of it, and what we do have is not very compelling.We lack the large samples of teeth for individual species thatare available for Australopithecus afarensis and the South Af-rican australopiths. The fossil record for postcranial remains,especially those associated with teeth, is even worse. These donot engender confidence in our studies of dental allometry,especially for incisors, for which size estimates are based onsamples as small as one or two individuals per species. Andfor cheek teeth, because the relationship between occlusal areaand diet in extant catarrhines is not consistent among higher-level taxa, our interpretations would be limited even if ourdata were reliable. In addition, studies of some functionallyrelevant aspects of tooth form (i.e., enamel distribution acrossthe crown and microstructure) have not yet been publishedfor early Homo, and those that have (i.e., occlusal topographyand enamel thickness) must await larger samples for com-parisons among Homo habilis, Homo rudolfensis, and Homoerectus. These things do not bode well for testing hypothesesconcerning dietary shifts from the australopiths to earlyHomo, the tempo of presumed shifts, or in which species theyfirst occurred.

While we can say little with confidence, speculation basedon some data is likely better than speculation based on none.If incisor allometry results hold, H. habilis and H. rudolfensis(as identified by Wood 1991, 1992) had larger front teeth thaneither the australopiths or African H. erectus. One possibleinterpretation is a shift toward foods requiring more incisalpreparation in H. habilis and H. rudolfensis followed by adecrease in front tooth size due to either a change in diet inH. erectus or a change in selective pressures related to in-creased extraoral food processing.

As far as the back teeth are concerned, there is a tendencytoward increasing occlusal area through time in the austra-lopiths then a decrease through time in Homo. That said, H.habilis and H. rudolfensis appear to have retained large cheekteeth, in the size range of Australopithecus, with fossils com-monly referred to H. rudolfensis having an occlusal area av-erage comparable to that of Paranthropus robustus. Homoerectus, on the other hand, shows a substantive decrease incheek-tooth size, with values approaching those of modernhumans. Some have suggested that smaller cheek teeth mightreflect selection to slow down the process of digestion or toreduce dental crowding given a shorter jaw. Interpretationsof variation among hominins in occlusal surface area willremain uncertain, however, until we better understand rela-tionships between diet and cheek-tooth area in living pri-mates.

We may be on firmer ground regarding interpretations oftopographic relief, though small samples prevent comparisonsof early Homo species to one another. Early Homo as a grouphas more occlusal topographic relief than the australopiths,

Figure 4. Dental microwear-texture data. Box-and-whiskers plots of microwear data for area-scale fractal complexity (A), scale of maximum complexity (B), and textural fill volume(C). The hinges mark the first and third quantiles, the horizontal lines between them are medians, each whisker represents a value 1.5 times the interquartile range, asterisks areoutliers, and circles are far outliers. The data illustrated are from Ungar et al. (2012), with early Homo specimen attributions as indicated in table 1.


though these teeth are much more bunodont than those ofgorillas or tough-food specialist monkeys.

And early Homo molars tend to have unremarkable mi-crowear texture complexity and anisotropy, suggesting thatmost individuals examined to date did not have diets dom-inated by either especially hard-brittle or very tough foods inthe days or weeks before death. Homo erectus specimens dovary substantially in complexity across surfaces, however, im-plying variation in food hardness; they also have texture at-tributes suggesting smaller average features compared with H.habilis (Ungar et al. 2012), such as adhesion pits that havebeen associated with tough-food consumption.

In sum, there is some evidence for a change in dietaryadaptations with the earliest members of the genus Homo, atleast in incisor size and perhaps molar occlusal slope andrelief. This might suggest a shift toward foods requiring moreincisal preparation and molar shearing, perhaps including dis-placement-limited items such as tough-plant products or an-imal tissues. More substantial change seems to have comewith H. erectus, which has both smaller incisors and smallermolar teeth compared with H. habilis and H. rudolfensis. Abroader range of microwear texture complexity values in H.erectus compared with H. habilis accords with the consump-tion of a wider variety of foods, and smaller average featuresize is consistent with the incorporation of more tough itemsin the diet.

Are these lines of evidence consistent with increased meateating or tool use in food preparation? The short answer isyes; each of these might have played a role. Larger samplesand more work on the fossils it is hoped will allow us tochoose among existing models or lead to new ones. The avail-able evidence suggests a shift in diet in early Homo and es-pecially H. erectus with broadening of the subsistence base toinclude at least some more tough foods. It also makes clearthat while much remains to be done, early hominin teethhold the promise to provide more detail about diet and con-fidence in our reconstructions as samples increase, our un-derstanding of functional morphology improves, and othermethods of analysis are applied to the fossils we have.

Acknowledgments

I thank Leslie Aiello and Susan Anton for inviting me toparticipate in the workshop that led to this paper. I am alsograteful for their comments and those of the reviewers of anearlier version of this paper. I thank the collaborators withwhom my part of the work reviewed here was done, especiallyFred Grine, Mark Teaford, Alan Walker, Robert Scott, FrancisM’Kirera, Jessica Scott, Kristin Krueger, Matt Sponheimer,Christopher Brown, and Toby Bergstrom. Finally, I am grate-ful to the many curators who have granted me access to theircollections in the United States, Europe, and Africa and boththe U.S. National Science Foundation and the L. S. B. LeakeyFoundation for their generous support.

References CitedAiello, Leslie C., and Peter Wheeler. 1995. The expensive-tissue hypothesis:


Anton, Susan C. 2008. Framing the question: diet and evolution in early Homo.In Primate craniofacial function and biology: papers in honor of Bill Hylander.Christopher J. Vinyard, Christine E. Wall, and Matthew J. Ravosa, eds. Pp.443–482. New York: Springer.


Bailey, Shara E., and Bernard A. Wood. 2007. Trends in postcanine occlusalmorphology within the hominin clade: the case of Paranthropus. In Dentalperspectives on human evolution: state of the art research in dental paleoan-thropology. Shara E. Bailey and Jean-Jacques Hublin, eds. Pp. 33–52. Dor-drecht: Springer.

Berthaume, Michael, Ian R. Grosse, Nirdesh D. Patel, David S. Strait, SarahWood, and Brian G. Richmond. 2010. The effect of early hominin occlusalmorphology on the fracturing of hard food items. Anatomical Record 293:594–606.

Beynon, A. D., and B. A. Wood. 1986. Variations in enamel thickness andstructure in East African hominids. American Journal of Physical Anthro-pology 70:177–193.

Brace, C. Loring. 1963. Structural reduction in evolution. American Naturalist97:39–49.

Brace, C. Loring, B. Holly Smith, and Kevin D. Hunt. 1991. What big teethyou had grandma! human tooth size, past and present. In Advances indental anthropology. Marc A. Kelley and Clark S. Larsen, eds. Pp. 33–57.New York: Wiley-Liss.

Bunn, Jonathan M., and Peter S. Ungar. 2009. Dental topography and dietsof four Old World monkey species. American Journal of Primatology 71:466–477.

Calcagno, James M., and Kathleen R. Gibson. 1991. Selective compromise:evolutionary trends and mechanisms in hominid tooth size. In Advancesin dental anthropology. Marc A. Kelley and Clark S. Larsen, eds. Pp. 59–76. New York: Wiley-Liss.

Coffing, Katherine, Craig Feibel, Meave Leakey, and Alan Walker. 1994. Four-million-year-old hominids from East Lake Turkana, Kenay. American Jour-nal of Physical Anthropology 93:55–65.

Constantino, Paul J., Peter W. Lucas, James J. W. Lee, and Brian R. Lawn.2009. The influence of fallback foods on great ape tooth enamel. AmericanJournal of Physical Anthropology 140:653–660.

Cuy, Janet L., Adrian B. Mann, Kenneth J. Livi, Mark F. Teaford, and TimothyP. Weihs. 2002. Nanoindentation mapping of the mechanical properties ofhuman molar tooth enamel. Archives of Oral Biology 47:281–291.

Dean, M. Christopher, Martin E. Jones, and J. Richard Pilley. 1992. The naturalhistory of tooth wear, continuous eruption and periodontal disease in wildshot great apes. Journal of Human Evolution 22:23–39.

Dumont, Elizabeth R. 1995. Enamel thickness and dietary adaptation amongextant primates and chiropterans. Journal of Mammalogy 76:1127–1136.

Dunbar, Robin I. M. 1976. Australopithecine diet based on a baboon analogy.Journal of Human Evolution 5:161–167.

Eaglen, Robert H. 1984. Incisor size and diet revisited: the view from a plat-yrrhine perspective. American Journal of Physical Anthropology 64:263–275.

———. 1986. Morphometrics of the anterior dentition in strepsirhine pri-mates. American Journal of Physical Anthropology 71:185–201.

Fortelius, Mikael. 1985. Ungulate cheek teeth: developmental, functional andevolutionary interrelations. Acta Zoologica Fennica 180:1–76.

Gantt, David G., John Kappleman, Richard A. Ketcham, Marden E. Alder,and Thomas H. Deahl. 2006. Three-dimensional reconstruction of enamelthickness and volume in humans and hominoids. European Journal of OralSciences 114:360–364.

Goldstein, Serge, David Post, and Don Melnick. 1978. Analysis of cerco-pithecoid odontometrics. 1. Scaling of maxillary dentition. American Journalof Physical Anthropology 49:517–532.

Grine, Frederick E. 1981. Trophic differences between “gracile” and “robust”australopithecines: a scanning electron microscope analysis of occlusalevents. South African Journal of Science 77:203–230.

Groves, Colin P., and John R. Napier. 1968. Dental dimensions and diet inaustralopithecines. Proceedings of the VIIIth International Congress of An-thropology and Ethnological Sciences 3:273–276.

Harvey, Paul H., Timothy H. Clutton-Brock, and Michael Kavanagh. 1978.Sexual dimorphism in primate teeth. Journal of Zoology 186:475–485.



Hylander, William L. 1975. Incisor size and diet in anthropoids with specialreference to Cercopithecidae. Science 189:1095–1198.

Isaac, Glynn L. 1978. Food sharing and human evolution: archeological evi-dence from the Plio-Pleistocene of East Africa. Journal of AnthropologicalResearch 34:311–325.

Jolly, Clifford J. 1970. The seed-eaters: a new model of hominid differentiationbased on a baboon analogy. Man 5:5–26.

Jungers, William L. 1978. Canine reduction in early hominids. Current An-thropology 19:155–156.

———. 1988. New estimations of body size in australopithecines. In Evolu-tionary history of the “robust” australopithecines. Frederick E. Grine, ed. Pp.115–126. New York: de Gruyter.

Kay, Richard F. 1975. Allometry and early hominids (comment). Science 189:63.

———. 1977. Evolution of molar occlusion in Cercopithecidae and earlycatarrhines. American Journal of Physical Anthropology 46:327–352.

———. 1981. The nut-crackers: a new theory of the adaptations of the Ra-mapithecinae. American Journal of Physical Anthropology 55:141–151.

———. 1985. Dental evidence for the diet of Australopithecus. Annual Reviewof Anthropology 14:315–341.

Kay, Richard F., and Matt Cartmill. 1977. Cranial morphology and adaptationsof Palaechthon nacimienti and other Paromomyidae (Plesiadapoidea, Pri-mates), with a description of a new genus and species. Journal of HumanEvolution 6:19–53.

Kay, Richard F., and Herbert H. Covert. 1984. Anatomy and behavior ofextinct primates. In Food acquisition and processing in primates. David J.Chivers, Bernard A. Wood, and Alan Bilsborough, eds. Pp. 467–508. NewYork: Plenum.

Kay, Richard F., and Peter S. Ungar. 1997. Dental evidence for diet in someMiocene catarrhines with comments on the effects of phylogeny on theinterpretation of adaptation. In Function, phylogeny and fossils: Miocenehominoids and great ape and human origins. D. R. Begun, C. Ward, and M.Rose, eds. Pp. 131–151. New York: Plenum.

Kimbel, William H. 2009. The origin of Homo. In The first humans: originand early evolution of the genus Homo. F. E. Grine, J. G. Fleagle, and R. E.Leakey, eds. Pp. 31–38. New York: Springer.

Koenigswald, Wighert V., and W. A. Clemens. 1992. Levels of complexity inthe microstructure of mammalian enamel and their application in studiesof systematics. Scanning Microscopy 6:195–218.

Kono, Reiko T. 2004. Molar enamel thickness and distribution patterns inextant great apes and humans: new insights based on a 3-dimensional wholecrown perspective. Anthropological Science 112:121–146.

Kono, Reiko T., and Gen Suwa. 2008. Enamel distribution patterns in extanthuman and hominoid molars: occlusal versus lateral enamel thickness. Bul-letin of the National Museum of Natural Sciences D 34:1–9.

Leakey, Louis S. B., John R. Napier, and Phillip V. Tobias. 1964. New speciesof genus Homo from Olduvai Gorge. Nature 202:7–9.

Leakey, Meave G., Craig S. Feibel, Ian McDougall, and Alan Walker. 1995.New four-million-year-old hominid species from Kanapoi and Allia Bay,Kenya. Nature 376:565–571.

Leakey, Meave G., Fred Spoor, Frank H. Brown, Patrick N. Gathogo, Chris-topher Kiarie, Louise N. Leakey, and Ian McDougall. 2001. New hominingenus from eastern Africa shows diverse middle Pliocene lineages. Nature410:433–440.

Lucas, Peter, Paul Constantino, Bernard Wood, and Brian Lawn. 2008. Dentalenamel as a dietary indicator in mammals. BioEssays 30:374–385.

Lucas, Peter W. 2004. Dental functional morphology: how teeth work. New York:Cambridge University Press.

Lucas, Peter W., Zhongquan Sui, Kay Y. Ang, Hugh Tiang, Wah Tan, SheauH. King, Brooke Sadler, and Neeraja Peri. 2009. Meals versus snacks andthe human dentition and diet during the Paleolithic. In The evolution ofhominin diets: integrating approaches to the study of Palaeolithic subsistence.Jean-Jacques Hublin and Michael P. Richards, eds. Pp. 31–41. Dordrecht:Springer.

Maas, Mary C., and Elizabeth R. Dumont. 1999. Built to last: the structure,function, and evolution of primate dental enamel. Evolutionary Anthro-pology 8:133–152.

Macho, Gabrielle A., and Daisuke Shimizu. 2009. Dietary adaptations of SouthAfrican australopiths: inference from enamel prism attitude. Journal of Hu-man Evolution 57:241–247.

Macho, Gabrielle A., Daisuke Shimizu, Yong Jiang, and Iain R. Spears. 2005.

Australopithecus anamensis: a finite element approach to studying the func-tional adaptations of extinct hominins. Anatomical Record 283A:310–318.

Macho, Gabrielle A., and Iain R. Spears. 1999. Effects of loading on thebiochemical behavior of molars of Homo, Pan, and Pongo. American Journalof Physical Anthropology 109:211–227.

Macho, Gabrielle A., and J. Francis Thackeray. 1992. Computed tomographyand enamel thickness of maxillary molars of Plio-Pleistocene hominids fromSterkfontein, Swartkrans, and Kromdraai (South Africa): an exploratorystudy. American Journal of Physical Anthropology 89:133–143.

Mayr, Earnst. 1950. Taxonomic categories in fossil hominids. Cold SpringHarbor Symposia on Quantitative Biology 15:109–118.

McCollum, Melanie A., and Paul T. Sharpe. 2001. Developmental geneticsand early hominid craniodental evolution. BioEssays 23:481–493.

McHenry, Henry M. 1994. Behavioral ecological implications of early hominidbody size. Journal of Human Evolution 27:77–87.


Meldrum, D. Jeffery, and Richard F. Kay. 1997. Nuciruptor rubricae, a newpitheciin seed predator from the Miocene of Colombia. American Journalof Physical Anthropology 102:407–427.

M’Kirera, Francis, and Peter S. Ungar. 2003. Occlusal relief changes with molarwear in Pan troglodytes troglodytes and Gorilla gorilla gorilla. American Jour-nal of Primatology 60:31–41.

O’Connell, James F., Kristin Hawkes, and Nicholas G. B. Jones. 1999. Grand-mothering and the evolution of Homo erectus. Journal of Human Evolution36:461–485.

Olejniczak, Anthony J., Tanya M. Smith, Matthew M. Skinner, Frederick E.Grine, Robin N. M. Feeney, J. Francis Thackeray, and Jean-Jacques Hublin.2008a. Three-dimensional molar enamel distribution and thickness in Aus-tralopithecus and Paranthropus. Biology Letters 4:406–410.

Olejniczak, Anthony J., Paul Tafforeau, Robin N. M. Feeney, and LawrenceB. Martin. 2008b. Three-dimensional primate molar enamel thickness. Jour-nal of Human Evolution 54:187–195.

Oppenheimer, Armand. 1964. Tool use and crowded teeth in Australopithe-cinae. Current Anthropology 5:419–421.

Pilbeam, David, and Stephen J. Gould. 1974. Size and scaling in humanevolution. Science 186:892–901.

Plavcan, J. Michael. 1990. Sexual dimorphism in the dentition of extant an-thropoid primates. PhD dissertation, Duke University.

Pontzer, Herman, Jessica R. Scott, David Lordkipanidze, and Peter S. Ungar.2011. Dental microwear texture analysis and diet in the Dmanisi hominins.Journal of Human Evolution 61:683–687.

Remis, Michelle J. 1997. Western lowland gorillas (Gorilla gorilla gorilla) asseasonal frugivores: use of variable resources. American Journal of Pri-matology 43:87–109.

Rensberger, John M. 2000. Dental constraints in the early evolution of mam-malian herbivory. In Evolution of herbivory in terrestrial vertebrates. Hans-Dieter Sues, ed. Pp. 144–167. Cambridge: Cambridge University Press.

Rensberger, John M., and Wighert V. Koenigswald. 1980. Functional andphylogenetic interpretation of enamel microstructure in rhinoceroses.Paleobiology 6:477–495.

Robinson, John T. 1954. Prehominid dentition and hominid evolution. Evo-lution 8:324–334.

———. 1956. The dentition of the australopithecines. Transvaal MuseumMemoirs 9:1–179.

Ryan, Alan S. 1981. Anterior dental microwear and its relationship to dietand feeding behavior in three African primates (Pan troglodytes troglodytes,Gorilla gorilla gorilla, and Papio hamadryas). Primates 22:533–550.

Schwartz, Gary T. 2000. Taxonomic and functional aspects of the patterningof enamel thickness distribution in extant large-bodied hominoids. Amer-ican Journal of Physical Anthropology 111:221–244.

Scott, Robert S., Peter S. Ungar, Torbjorn S. Bergstrom, Christopher A. Brown,Benjamin E. Childs, Mark F. Teaford, and Alan Walker. 2006. Dental mi-crowear texture analysis: technical considerations. Journal of Human Evo-lution 51:339–349.

Shimizu, Daisuke. 2002. Functional implications of enamel thickness in thelower molars of red colobus (Procolobus badius) and Japanese macaque(Macaca fuscata). Journal of Human Evolution 43:605–620.

Shimizu, Daisuke, Gabrielle A. Macho, and Iain R. Spears. 2005. Effect ofprism orientation and loading direction on contact stresses in prismaticenamel of primates: implications for interpreting wear patterns. AmericanJournal of Physical Anthropology 126:427–434.


Simons, Elwyn L., and David Pilbeam. 1972. Hominoid paleoprimatology. InThe functional and evolutionary biology of primates. Russell Tuttle, ed. Pp.36–62. New York: Aldine-Atherion.

Smith, Patricia. 1982. Dental reduction, selection or drift? In Teeth: form,function, and evolution. Bjorn Kurten, ed. Pp. 366–379. New York: ColumbiaUniversity Press.

Smith, Richard J. 1996. Biology and body size in human evolution: statisticalinference misapplied. Current Anthropology 37:451–481.

Smith, Tanya M., Anthony J. Olejniczak, John P. Zermeno, Paul Tafforeau,Matthew M. Skinner, Almut Hoffmann, Jakov Radovcic, et al. 2012. Var-iation in enamel thickness within the genus Homo. Journal of Human Evo-lution 62:395–411.

Smith, Tanya M., and Paul Tafforeau. 2008. New visions of dental tissueresearch: tooth development, chemistry, and structure. Evolutionary An-thropology 17:213–226.

Sofaer, Jeffrey A. 1973. A model relating developmental interaction and dif-ferential evolutionary reduction of tooth size. Evolution 27:427–434.

Sofaer, Jeffrey A., Howard L. Bailit, and C. J. MacLean. 1971. A developmentalbasis for differential tooth reduction during hominid evolution. Evolution25:509–517.

Strait, Suzanne G. 1993. Differences in occlusal morphology and molar sizein frugivores and faunivores. Journal of Human Evolution 25:471–484.

Suwa, Gen, Reiko T. Kono, Scott W. Simpson, Berhane Asfaw, C. OwenLovejoy, and Tim D. White. 2009. Paleobiological implications of the Ar-dipithecus ramidus dentition. Science 326:94–99.

Swain, Michael V., and Jing Xue. 2009. State of the art of micro-CT appli-cations in dental research. International Journal of Oral Science 1:177–188.

Tafforeau, Paul, and Tanya A. Smith. 2008. Nondestructive imaging of hom-inoid dental microstructure using phase contrast x-ray synchrotron micro-tomography. Journal of Human Evolution 54:272–278.

Teaford, Mark F. 1991. Dental microwear: what can it tell us about diet anddental function? In Advances in dental anthropology. Marc A. Kelley andClark S. Larsen, eds. Pp. 341–356. New York: Wiley-Liss.

———. 2007a. What do we know and not know about dental microwear anddiet? In Evolution of the human diet: the known, the unknown, and theunknowable. Peter S. Ungar, ed. Pp. 106–132. New York: Oxford UniversityPress.

———. 2007b. What do we know and not know about diet and enamelstructure? In Evolution of the human diet: the known, the unknown, and theunknowable. Peter S. Ungar, ed. Pp. 56–76. New York: Oxford UniversityPress.

Teaford, Mark F., and Ordean J. Oyen. 1988. In vivo and in vitro turnoverin dental microwear. American Journal of Physical Anthropology 75:279.

Teaford, Mark F., and Jacqueline A. Runestad. 1992. Dental microwear anddiet in Venezuelan primates. American Journal of Physical Anthropology 88:347–364.

Teaford, Mark F., and Peter S. Ungar. 2000. Diet and the evolution of theearliest human ancestors. Proceedings of the National Academy of Sciences,U.S.A. 97:13506–13511.

Teaford, Mark F., Peter S. Ungar, and Frederick E. Grine. 2002. Paleontologicalevidence for the diets of African Plio-Pleistocene hominins with specialreference to early Homo. In Human diet: its origin and evolution. Peter S.Ungar and Mark F. Teaford, eds. Pp. 143–166. Westport, CT: Bergin &Garvey.

Tutin, Caroline E. G., Michel Fernandez, M. Elizabeth Rogers, Elizabeth A.Williamson, and William C. McGrew. 1991. Foraging profiles of sympatriclowland gorillas and chimpanzees in the Lope Reserve, Gabon. PhilosophicalTransactions of the Royal Society B 334:179–186.

Ungar, Peter S. 1996. Relationship of incisor size to diet and anterior toothuse in sympatric Sumatran anthropoids. American Journal of Primatology38:145–156.

———. 2004. Dental topography and diets of Australopithecus afarensis andearly Homo. Journal of Human Evolution 46:605–622.

———. 2007. Dental topography and human evolution: with comments onthe diets of Australopithecus africanus and Paranthropus robustus. In Dentalperspectives on human evolution: state of the art research in dental anthro-pology. Shara Bailey and Jean-Jacques Hublin, eds. Pp. 321–344. New York:Springer.

———. 2008. Materials science: strong teeth, strong seeds. Nature 452:703–705.

———. 2009. Tooth form and function: insights into adaptation through theanalysis of dental microwear. In Comparative dental morphology. ThomasKoppe, Georg Meyer, and Kurt W. Alt, eds. Pp. 38–43. Basel: Karger.

———. 2010. Mammal teeth: origin, evolution, and diversity. Baltimore: JohnsHopkins University Press.

Ungar, Peter. S., Christopher A. Brown, Torbjorn S. Bergstrom, and AlanWalker. 2003. Quantification of dental microwear by tandem scanning con-focal microscopy and scale-sensitive fractal analyses. Scanning 25:185–193.

Ungar, Peter S., and Jonathan M. Bunn. 2008. Primate dental topographicanalysis and functional morphology. In Technique and application in dentalanthropology. Joel D. Irish and Gregory C. Nelson, eds. Pp. 253–265. NewYork: Cambridge University Press.

Ungar, Peter S., and Frederick E. Grine. 1991. Incisor size and wear in Aus-tralopithecus africanus and Paranthropus robustus. Journal of Human Evo-lution 20:313–340.

Ungar, Peter S., Frederick E. Grine, and Mark F. Teaford. 2006. Diet in earlyHomo: a review of the evidence and a new model of adaptive versatility.Annual Review of Anthropology 35:209–228.

Ungar, Peter S., Frederick E. Grine, Mark F. Teaford, and Sireen El Zaatari.2006. Dental microwear and diets of African early Homo. Journal of HumanEvolution 50:78–95.

Ungar, Peter S., Kristin L. Krueger, Robert J. Blumenschine, Jackson Njau,and Robert S. Scott. 2012. Dental microwear texture analysis of homininsrecovered by the Olduvai Landscape Paleoanthropology Project, 1995–2007.Journal of Human Evolution. Forthcoming, doi:10.1016/j.jhevol.2011.04.006.

Ungar, Peter S., and Peter W. Lucas. 2010. Tooth form and function in bi-ological anthropology. In A companion to biological anthropology. Clark S.Larsen, ed. Pp. 516–529. Malden, MA: Wiley-Blackwell.

Ungar, Peter S., and Francis M’Kirera. 2003. A solution to the worn toothconundrum in primate functional anatomy. Proceedings of the NationalAcademy of Sciences, U.S.A. 100:3874–3877.

Ungar, Peter S., and Robert S. Scott. 2009. Dental evidence for diets of earlyHomo. In The first humans: origins of the genus Homo. Frederick E. Grine,Richard E. Leakey, and John G. Fleagle, eds. Pp. 121–134. New York:Springer.

Ungar, Peter S., Robert S. Scott, Jessica R. Scott, and Mark F. Teaford. 2007.Dental microwear analysis: historical perspectives and new approaches. InTechnique and application in dental anthropology. Joel D. Irish and GregoryC. Nelson, eds. Pp. 389–425. Cambridge: Cambridge University Press.

Ungar, Peter S., and Matthew J. Sponheimer. 2011. Hominin diets. In Acompanion to paleoanthropology. D. R. Begun, ed. New York: Wiley-Black-well.

Vinyard, Christopher J., and Jandy Hanna. 2005. Molar scaling in strepsirrhineprimates. Journal of Human Evolution 49:241–269.

Walker, Alan. 1981. Diet and teeth: dietary hypotheses and human evolution.Philosophical Transactions of the Royal Society B 292:57–64.

Wallace, John A. 1975. Dietary adaptations of Australopithecus and early Homo.In Paleoanthropology, morphology and paleoecology. Russell H. Tuttle, ed.Pp. 202–223. The Hague: Mouton.

Wood, Bernard. 1991. Hominid cranial remains, vol. 4 of Koobi Fora researchproject. Oxford: Oxford University Press.

———. 1992. Origin and evolution of the genus Homo. Nature 355:783–790.Wood, Bernard A., and Mark Collard. 1999. The human genus. Science 284:

65–71.Wood, Bernard, and David Strait. 2004. Patterns of resource use in early

Homo and Paranthropus. Journal of Human Evolution 46:119–162.Workman, Michael S., Larry J. Leamy, Eric J. Routman, and James M. Che-

verud. 2002. Analysis of quantitative trait locus effects on the size and shapeof mandibular molars in mice. Genetics 160:1573–1586.

Wrangham, Richard W., James H. Jones, Greg Laden, David Pilbeam, andNancy Lou Conklin-Brittain. 1999. The raw and the stolen: cooking andthe ecology of human origins. Current Anthropology 40:567–594.

Zihlman, Adrienne, and Nancy M. Tanner. 1978. Gathering and hominidadaptation. In Female hierarchies. Lionel Tiger and Heather T. Fowler, eds.Pp. 163–194. Chicago: Beresford Book Service.



Body Size, Body Shape, and theCircumscription of the

Genus Homo

by Trenton W. Holliday

Since the 1984 discovery of the Nariokotome Homo erectus/Homo ergaster skeleton, it has been almost axiomaticthat the emergence of Homo (sensu stricto) was characterized by an increase in body size to the modern humancondition and an autapomorphic shift in body proportions to those found today. This was linked to a behavioralshift toward more intensive carnivory and wider ranging in the genus Homo. Recent fossil discoveries and reanalysisof the Nariokotome skeleton suggest a more complex evolutionary pattern. While early Homo tend to be largerthan Australopithecus/Paranthropus, they were shorter on average than people today. Reanalysis of the Nariokotomepelvis along with the discovery of additional early and middle Pleistocene pelves indicate that a narrow bi-iliac(pelvic) breadth is an autapomorphy specific to Homo sapiens. Likewise, it appears that at least some early Homo(even those referred to H. ergaster/H. erectus) were characterized by higher humero-femoral indices than the H.sapiens average. All these data suggest a pattern of mosaic postcranial evolution in Homo with implications for theincreased ranging/carnivory model of the origin of Homo as well as for which species are included within the Homohypodigm.

Introduction

The oldest and most persistent questions in paleoanthropol-ogy are those that pertain directly to us. Specifically, the twolongest-lived questions in our discipline are (1) the origin ofour own species, Homo sapiens, which occurred sometimearound the end of the middle Pleistocene; and (2) the originof our own genus, Homo, which most likely occurred in thelate Pliocene. The questions surrounding the origins of thesetaxa are not merely of phylogenetic bookkeeping interest. Be-cause they involve our direct ancestors, we tend to imbuethese questions with a suite of functional/adaptive explana-tions (e.g., intelligence and/or spoken language in the case ofmodern human origins and increased encephalization, largerbody and home-range size, more efficient bipedality, and in-creased carnivory in the case of the origins of the genusHomo). In light of humanity’s global distribution, invokingadaptation when our own species/genus’s evolution is con-cerned is certainly warranted. Our widespread geographic dis-tribution is even more impressive when one considers thatamong the African hominids (sensu lato), humans are evo-

Trenton W. Holliday is Professor of Anthropology at TulaneUniversity (101 Dinwiddie Hall, 6823 St. Charles Avenue, NewOrleans, Louisiana 70118, U.S.A. [[email protected]]). This paperwas submitted 12 XII 11, accepted 11 VI 12, and electronicallypublished 14 IX 12.

lutionarily the only successful taxon, because the other Africanapes appear poised for extinction in the wild, and importantly,their decline is not likely merely the result of recent habitatdestruction due to an ever-expanding human population butrather may have more to do with their overreliance on a highlyK-selected reproductive “strategy” coupled with a decreasedpopulation growth rate associated with larger brains (Islerand van Schaik 2012; Lovejoy 1981, 2009).

This paper studies early (ca. 2.0–1.5 million years ago[mya]) Homo specifically with regard to its postcranial skel-eton, body size, and limb and body proportions. The studyof limb and body proportions of early Homo in particularhas led to extensive debate as to whether the earliest membersof the genus had limb and body proportions similar to thoseof Australopithecus (who are presumed to have included moreof an arboreal component in their locomotor behavior) or,in contrast, whether their limb and body proportions weremore similar to those of modern humans, which could betaken as indicative of fully committed terrestrial bipedality(Bramble and Lieberman 2004; Haeusler and McHenry 2004,2007; Hartwig-Scherer and Martin 1991; Holliday and Fran-ciscus 2009; McHenry and Berger 1998a, 1998b; Pontzer 2012;Pontzer et al. 2010; Reno et al. 2005; Richmond, Aiello, andWood 2002). Observed (or presumed) differences in bodysize and proportions between Homo sapiens and some earlyPleistocene hominins referred to Homo have even played arole in some researchers’ call to remove key specimens (and


Holliday Body Size and Shape in Early Homo S331

even presumed species-level taxa) from the genus Homo al-together (Collard and Wood 2007; Wood and Collard 1999a,1999b).

Circumscribing the Genus Homo

In order to discuss body size and shape in the genus Homo,it is critical to determine the composition of that genus. Inzoology, there are generally four nonmutually exclusive setsof criteria by which species taxa are considered congeneric:(1) recency of common ancestry, (2) ecological/adaptive sim-ilarity, (3) genetic divergence, or (4) morphological similar-ities. However, there is little agreement among zoologists asto how exactly this is done, and, as recently pointed out byCollard and Wood (2007), relatively little attention has beenpaid by taxonomists as to how the genus category is opera-tionally defined. Ernst Mayr (1942:283), one of the foundersof the school of evolutionary taxonomy, defined a genus as“one species or a group of species of presumably commonphylogenetic origin, separated by a decided gap from othersimilar groups.” Mayr’s definition therefore allows for “grade”characters (or at a minimum, perceived “gaps” between them)to be involved in circumscribing taxa, making it unacceptableto strict practitioners of phylogenetic systematics (cladistics)as developed by Willi Hennig.

Hennig (1966) argued that all biological taxa be strictlymonophyletic (what Mayr [2000] suggests should instead bereferred to as “holophyletic”); that is, taxa at all levels shouldbe made up of all the descendents of an ancestral taxon re-gardless of their plesiomorphic versus apomorphic (i.e.,“grade”) status. Of course, the criterion of monophyly alonedoes not help one distinguish a monophyletic genus from amonophyletic tribe, subfamily, or family. Hennig (1966)therefore suggested that time since divergence be the secondmajor criterion for circumscribing genera, but he also arguedthat rather than using a universal “one size fits all” time frame(i.e., all genera everywhere should be 8–10 million years old),temporal criteria should be specific to each particular bio-logical group studied. While admittedly subjective, this wouldseem appropriate given vast differences in divergence timesamong long-recognized “good” zoological higher taxa. Forexample, molecular data suggest that the two genera of Asiatichorseshoe crabs (Tachypleus and Carcinoscorpius) divergedfrom their North American cousins (genus Limulus) ca. 45–60 mya (Avise, Nelson, and Sugita 1994), which may be earlierthan the divergence of the Old World anthropoids (infraorderCatarrhini) from the New World monkeys (infraorder Pla-tyrrhini) and certainly millions of years earlier than diver-gence of the family Hominidae (apes and humans) from theCercopithecidae (Old World monkeys).

Thus, it is evident that neither the phylogenetic systematicdescription of genera nor the definition of genera employedby evolutionary taxonomists is without its flaws, but it shouldbe possible to combine the strengths of both into a bettermeans by which to define generic taxa. Applying this logic,

Wood and Collard (1999a, 1999b) used an operational defi-nition of the genus category that combined what they viewedas the stronger aspects of both Mayr’s (1942) and Hennig’s(1966) approaches. They recognized that a strict “Hennigian”approach to defining genera gave no real guidance as to howto circumscribe genera versus other higher Linnaean cate-gories and that Mayr’s (1942) approach was faulty in that itcould lead to the recognition of paraphyletic genera—taxathat are biologically “unreal” in that they do not reflect thetrue evolutionary relationships of the taxa in question. In-stead, Wood and Collard (1999b:66) define a genus as “aspecies, or monophylum, whose members occupy a singleadaptive zone.” Importantly, their definition of a genus doesnot restrict any particular “adaptive zone” to a single genus,instead allowing “for the possibility that species assigned todifferent genera will occupy the same adaptive zone,” but itdoes prevent “species in the same genus from occupying dif-ferent adaptive zones” (Collard and Wood 2007:1584). An“adaptive zone” is characterized by Wood and Collard (1999a:202) as being related to an organism’s phenotype and itsability to “maintain homeostasis, acquire food, and produceoffspring.” While recognizing that not all of these aspects ofthe phenotype are easily recognizable in the fossil record,Wood and Collard (1999a, 1999b) maintain that certain fea-tures, such as the size of the masticatory apparatus, relativebrain size, ontogenetic pattern, body size and shape, and lo-comotor behavior all leave phenotypic traces observable inhominin fossils.

In circumscribing the genus Homo, then, Wood and Collard(1999a, 1999b) argued that two main criteria must be met:(1) that the genus Homo be monophyletic (holophyletic) and(2) that all its members share a common adaptive strategy orzone. To test the first criterion, Wood and Collard (1999a,1999b) reported the results of multiple cladistic analyses, someof which failed to consistently link Homo habilis and Homorudolfensis as sister taxa to Homo sapiens to the exclusion ofthe australopiths.

For the second (i.e., the adaptive-zone) criterion, Woodand Collard evaluated whether a presumed member speciesof the genus Homo was more similar to the type species ofthe genus Homo (H. sapiens) or Australopithecus (Australo-pithecus africanus) for the following adaptive complexes: (1)body size, (2) body shape, (3) locomotion, (4) mastication,(5) growth and development, and (6) relative brain size.1 Theyfound that Homo ergaster, Homo erectus, Homo heidelbergensis,and Homo neanderthalensis were more similar to H. sapiensthan A. africanus for all, or at least the vast majority, of theseadaptive complexes (the clade including these five species will

1. Anton (2012) points out that comparing 2.0–1.5-million-year-oldfossil hominins to modern humans (Homo sapiens) in order to refer themto Homo (or not) is problematic in that it will tend to disregard similaritiesshared between Homo habilis/Homo rudolfensis and Homo ergaster/Homoerectus, instead emphasizing differences between earliest Homo and themost derived taxon in the clade (H. sapiens), a taxon that also happensto be far removed from early Homo in time.


here be referred to as Homo [sensu stricto]). In contrast,according to their analyses, H. rudolfensis and H. habilis weredecidedly more Australopithecus-like. Wood and Collardtherefore came to the conclusion that these two species failedboth the phylogenetic and the adaptive-zone criteria for in-clusion in the genus Homo and suggested that they insteadbe referred to the genus Australopithecus.

More recently, Collard and Wood (2007) redid their 1999analyses taking into account new fossil data as well as morerecently published cladistic analyses. As before, the cladisticanalyses remained somewhat inconsistent in clustering H. ha-bilis and H. rudolfensis with later Homo to the exclusion ofother taxa. Based on new data, they did find some changesin adaptive-zone patterning from their 1999 results (partic-ularly with regard to the pattern of growth and developmentof H. ergaster; and see Schwartz 2012); however, their ultimateconclusion remained that neither H. habilis nor H. rudolfensisshould be included in the Homo hypodigm (Collard andWood 2007).

This paper evaluates five reflections of body size and shapein early Homo (sensu lato) that figure into the adaptive-zonehalf of the circumscription of the genus Homo. The featuresevaluated here allow the examination of three of the six adap-tive “complexes” as defined by Wood and Collard (1999a,1999b): (1) body size, (2) body shape, and (3) locomotion.The current analyses also include recently published data notavailable to Wood and Collard (1999a, 1999b) or Collard andWood (2007). This paper will attempt to assess whether (1)the emergence of the genus Homo is associated with an in-crease in locomotor efficiency specifically as related to ter-restrial bipedality and its relationship to ranging behavior,and (2) H. habilis and H. rudolfensis should be removed fromthe genus Homo based on Wood and Collard’s adaptive-zonecriteria for body size and shape and locomotion. The fivespecific postcranial features to be examined are (1) relativelower limb length, (2) humero-femoral proportions, (3) rel-ative forearm length, (4) relative pelvic breadth, and (5) bodysize as reflected in predicted body mass and a proxy for stat-ure.

Relative Lower Limb Length

Longer limbs are known to increase the efficiency of animallocomotion by reducing the number of strides necessary tocover any given distance (Jungers 1982; Pontzer 2007, 2012;Steudel-Numbers 2006; Steudel-Numbers and Tilkens 2004).In this light, it has long been argued that Australopithecus hadrelatively short lower limbs while Homo (sensu stricto) wascharacterized by longer lower limbs (Bramble and Lieberman2004; Jungers 1982, 1991a; Jungers and Stern 1983; Pontzeret al. 2010; Richmond, Aiello, and Wood 2002). However,there are many recent data that challenge this assumption.First, this view, while prevalent, has been largely based on thediminutive A.L. 288-1 Australopithecus afarensis (“Lucy”)specimen (Jungers 1982, 1991a; Jungers and Stern 1983). The

problem with this line of evidence is that lower limb lengthshows positive allometry in humans (Auerbach and Sylvester2011; Holliday and Franciscus 2009; and see Pontzer 2012)such that the expectation is for small-bodied hominins suchas “Lucy,” including members of the genus Homo, to haveshorter lower limbs (Franciscus and Holliday 1992; Hollidayand Franciscus 2009; but see Haeusler and McHenry 2004).The discovery of diminutive late Pleistocene (but clearly genusHomo) fossils at the site of Liang Bua on Flores (referred tothe species Homo floresiensis; Brown et al. 2004), which arecharacterized by relatively short lower limbs (Jungers et al.2009), provides further support for this idea no matter whatthose specimens’ ultimate taxonomic status within our genus(Holliday and Franciscus 2009).

Likewise, there have been numerous recent discoveries and/or reanalyses of larger-bodied members of the genus Aus-tralopithecus who (unlike “Lucy”) are characterized by rela-tively elongated lower limbs. These include the A. afarensisspecimens A.L. 333-3, A.L. 827-1 (Harmon 2005; Kimbel andDelezene 2009), and KSD-VP-1/1 (Haile-Selassie et al. 2010)and the (presumed) Australopithecus africanus specimen StW99 (T. W. Holliday, unpublished data). The overall pattern oflower limb length relative to size in hominins is illustratedby the scatterplot in figure 1, which shows femoral lengthregressed on femoral head diameter. The former measurementreflects lower limb length, while the latter reflects overall bodymass. Note that Homo sapiens is largely separated from Panand Gorilla in bivariate space in that for any given femoralhead size, our species is expected to have a much longer femur.With two exceptions, all of the fossil hominins fall within the95% confidence limits about the H. sapiens individuals forthis relationship and fall outside the 95% confidence limitsabout Pan and Gorilla. The exceptions, notably, are the di-minutive A.L. 288-1 (“Lucy”) and Liang Bua 1 (holotype ofH. floresiensis) specimens, which fall among the chimpanzees.The slightly larger (but still diminutive by modern humanstandards) A.L. 152-2 A. afarensis individual falls at the limitsof the H. sapiens sample and just beyond the 95% confidencelimits about the chimpanzee individuals. Note, however, thatthis individual falls almost directly on a modern human“Pygmy” individual’s values for both measurements. In con-trast, the larger-bodied australopiths (A.L. 333-3, A.L. 827-1,KSD-VP-1/1, and StW 99), while characterized by smallerfemoral heads than most members of the genus Homo (longknown to be an australopith characteristic; Galik et al. 2004;Harmon 2009; Jungers 1988, 1991a; Kennedy 1983; Lovejoy,Heiple, and Burstein 1973; Napier 1964; Richmond and Jung-ers 2008; Robinson 1972; Ruff 1988; Walker 1973), nonethe-less fall among the H. sapiens sample for this relationship andbeyond the confidence limits about the African ape samples,as do the early Homo specimens KNM-ER 1472 and 1481,KNM-WT 15000 (the “Nariokotome Boy”), and Dmanisi4507.

A scaling difference in femoral head size between Homoand the australopiths has also been documented (Jungers


Figure 1. Scatterplot of femoral length regressed on femoral head diameter for Pleistocene/Holocene Homo sapiens, Pan, Gorilla,and Pliocene/Pleistocene early hominins (and Liang Bua 1). Australopiths are indicated by open squares, fossil Homo by filledsquares. The ordinary least squares regression lines for the comparative samples are represented by solid lines with the 95% confidencelimits for the individuals indicated about them. The reduced major axis (RMA) regression lines for the comparative samples arethe dashed lines. Homo sapiens RMA formula: ; Pan RMA formula: ; Gorilla RMA formula:y p 8.074x � 77.175 y p 7.159x � 56.2

.y p 6.439x � 53.28

1988; Ruff 1990). It is probable that this difference is in someway related to a shift in locomotor repertoire between Homoand the australopiths, but the exact nature of such a shift isuncertain, especially when one considers that there are ap-parent differences in upper versus lower limb articular andlength proportions even within the genus Australopithecus (seebelow). Despite this caveat, the pattern of relative lower limblength revealed in figure 1 is mirrored in plots for which bodymass and not femoral head diameter is plotted along the X-axis (Franciscus and Holliday 1992; Holliday and Franciscus2009; Pontzer 2012; Pontzer et al. 2010).

Humero-Femoral Proportions

The discovery of the A.L. 288-1 specimen, with its nearlyintact humerus and femur, led to the confirmation that Aus-tralopithecus afarensis was characterized by very differenthumero-femoral proportions than the genus Homo (or, forthat matter, the genus Pan), with longer humeri relative totheir femora than Homo and much shorter humeri relativeto their femora than Pan (Johanson et al. 1982). What fol-

lowed were years of intense debate as to whether the humero-femoral proportions of “Lucy” were due to a long humerus,a short femur, or a combination of both (Franciscus andHolliday 1992; Jungers 1982, 1991a; Jungers and Stern 1983;Wolpoff 1983). It now seems probable that the last of thesealternative explanations best accounts for the data (Hollidayand Franciscus 2012).

It has also been argued that the humero-femoral propor-tions of Australopithecus africanus and/or Homo habilis wereas different, or more different, from modern humans thanthose of A. afarensis (Hartwig-Scherer and Martin 1991; Mc-Henry and Berger 1998a). Unfortunately, in the case of H.habilis, at least, this supposition is based on faulty data—specifically the femoral length of the OH 62 specimen, whichcannot be reliably reconstructed with any degree of confidence(see below). Another presumed H. habilis partial skeleton,KNM-ER 3735, is even less complete (Haeusler and McHenry2007). Similarly, there is a relative dearth of associated A.africanus skeletons, and those that are available (e.g., Sts 14and StW 431) lack complete limb bones.


Figure 2. Scatterplot of humeral length regressed on femoral length for Pleistocene/Holocene Homo sapiens, Pan, Gorilla, andPliocene/Pleistocene early hominins (and Liang Bua 1). Australopiths are indicated by open squares, fossil Homo by filled squares.The ordinary least squares regression lines for the comparative samples are represented by solid lines with the 95% confidence limitsfor the individuals indicated about them. The reduced major axis (RMA) regression lines for the comparative samples are the dashedlines. Homo sapiens RMA formula: ; Pan RMA formula: ; Gorilla RMA formula:y p 0.748x � 14.432 y p 1.01x � 4.15 y p

.1.092x � 26.715

Patterning in humero-femoral proportions among Africanhominids (sensu lato) is shown in figure 2, which is a scat-terplot of humeral length regressed on femoral length. As haslong been appreciated, gorillas have the longest humeri rel-ative to their femora, followed by chimpanzees, while humanshave the shortest humeri relative to the length of their femora(or the longest femora relative to their humeri). The firstpattern evident in the graph is that the humero-femoral pro-portions of OH 62 are impossible to assess (as was arguedby Haeusler and McHenry 2004, 2007; Korey 1990; Reno etal. 2005; Richmond, Aiello, and Wood 2002). The primaryproblem in the assessment of OH 62’s humero-femoral pro-portions is unlikely to be its humerus; Haeusler and McHenry(2004:446) point out that the OH 62 humerus is “relativelywell preserved, lacking only the proximal and distal extrem-ities. Its length can be estimated with only minor error.” Theyestimate humeral length of the specimen at 270 mm. Herean estimate of 264 mm from Johanson et al. (1987) is used(these two estimates differ from each other by only 2%).

Despite the fact that Korey (1990) argues there is a highdegree of variability in humeral length estimation for OH 62,the more intractable problem with assessing its humero-fem-

oral proportions lies with its femur. As noted by Reno et al.(2005), it is likely that less than half of the femur’s length ispreserved, and among hominids (sensu lato) there is no pre-dictable relationship between proximal femoral length andmaximum (or bicondylar) femoral length. Because Johansonet al. (1987) argued that the OH 62 shaft was smaller andless robust than that of A.L. 288-1, some researchers (e.g.,Haeusler and McHenry 2004; Richmond, Aiello, and Wood2002) have taken to using Lucy’s femoral length of 280 mmas the minimum length for the OH 62 femur, a conventionfollowed here. Haeusler and McHenry (2004) argue, however,that the longer (albeit damaged; see Green, Gordon, and Rich-mond 2007) OH 34 femur is a better analog for the OH 62femur than is A.L. 288-1. They therefore used their estimatedOH 34 femoral length of 374 mm as the maximum (potential)femoral length for OH 62; the same is done here. This equateswith nearly 100 mm of difference between the minimum andmaximum estimates of the OH 62 femoral length. Thus, evenin the absence of error in humeral length, we can say nothingabout OH 62’s humero-femoral proportions because, as isevident in figure 2, they could range from somewhat morechimpanzee-like than “Lucy” (as was maintained by Hartwig-


Figure 3. Histogram of humero-femoral indices, Pleistocene/Holocene Homo sapiens sample with the values (or in the case of BOU-VP-12/1, the range of potential values) of fossil hominins indicated by arrows. Homo sapiens mean p 71.5; SD p 2.3.

Scherer and Martin 1991) to falling almost directly on theHomo sapiens reduced major axis (RMA) and ordinary leastsquares (OLS) regression lines.

Note, too, that as was the case for the previous relationship,the only fossil hominins to fall squarely outside of the H.sapiens sample in figure 2 are the diminutive A.L. 288-1 andLB1 specimens. The limb-segment lengths of the 2.5-million-year-old taxonomically unassigned remains from Bouri(BOU-VP-12/1) were estimated anatomically by Asfaw et al.(1999) at 226–236 mm for the humerus and 335–348 mmfor the femur, which places them squarely among H. sapiensin bivariate space (although there is likely a problem withthese estimates; see below). The early Pleistocene Homo re-mains from Dmanisi and KNM-WT 15000 also fall amongthe more recent humans, although Dmanisi 4167/4507 fallson the 95% confidence limits about the recent human in-dividuals. This graphic result is, however, somewhat differentfrom that depicted in figure 3b of Lordkipanidze et al. (2007),where Dmanisi 4167/4507 appears to fall in the middle of therecent human scatter for the humerus length to femoral lengthbivariate relationship. On closer inspection of Lordkipanidzeet al. (2007), however, it becomes evident that the 295-mmmaximum humeral length of Dmanisi 4507 is erroneouslyrepresented as ca. 275 mm in their figure 3b.

A histogram of humero-femoral indices of the H. sapienssample is presented in figure 3 with the positions of the fossilsthat preserve both a humerus and a femur indicated by ar-rows. As has been noted multiple times (e.g., Brown et al.2004; Johanson et al. 1982; Jungers 1982; McHenry 1978),A.L. 288-1 and LB1 lie far (5.6 and 6.6 standard deviations,

respectively) above the Homo sapiens mean. While fallingwithin the anatomically modern human range, Dmanisi 4507and KNM-WT 15000 still have high index values because theylie at the modern human 97th and 84th percentiles, respec-tively. In fact, a special case of Student’s t-test for comparinga single individual to a sample (Sokal and Rohlf 1981) findsthat Dmanisi 4507 is statistically significantly different (at

) from modern humans ( , ),P ! .05 t p 2.134 P p .033s

whereas KNM-WT 15000 is not ( , ). Givent p 1.014 P p .311s

that Dmanisi dates to ca. 1.77 mya (Gabunia, Vekua, andLordkipanidze 2000) and KNM-WT 15000 dates to ca. 1.53mya (Brown and McDougall 1993), if each is reflective of itspopulation, and assuming that these skeletons represent alineage in the broadest sense, then it may be that humero-femoral indices decreased through time in early Homo.

In light of the above patterning, the extremely low humero-femoral index values (ranging from below the first to the 32ndpercentiles of H. sapiens) of the 2.5-million-year-old BOU-VP-12/1 specimen appear particularly incongruous. Thiscould be due to the humerus and femur coming from differentindividuals or (more likely) one or both of the estimated long-bone lengths being erroneously reconstructed. If the latter isthe case, the most probable scenario, given the preservationof the remains shown in figure 4 of Asfaw et al. (1999), isthat the humeral length of the specimen has been underpre-dicted, because the femur preserves the diaphysis from theinferior margin of the neck to an area just proximal to thelateral epicondyles, and therefore its length should be accu-rately reconstructed. In contrast, while on the humerus aproximal portion of the medial epicondyle is preserved (giving


Figure 4. Scatterplot of radius length regressed on femoral length for Pleistocene/Holocene Homo sapiens, Pan, Gorilla, and Pliocene/Pleistocene early hominins (and Liang Bua 1). Australopiths are indicated by open squares, fossil Homo by filled squares. Theordinary least squares regression lines for the comparative samples are represented by solid lines with the 95% confidence limitsfor the individuals indicated about them. The reduced major axis (RMA) regression lines for the comparative samples are the dashedlines. Homo sapiens RMA formula: ; Pan RMA formula: ; Gorilla RMA formula:y p 0.623x–33.815 y p 1.175x–70.195 y p

.0.93x � 3.713

one a solid idea of where the distal end of the bone shouldbe), it is impossible to ascertain from the figure alone howmuch of the proximal diaphysis is present.

Relative Forearm Length

One of the key features long thought to distinguish the aus-tralopiths from the genus Homo is that the former are saidto have had longer (and therefore more “apelike”) forearmsthan the latter (Asfaw et al. 1999; Haeusler and McHenry2004; Howell and Wood 1974; Kimbel, Johanson, and Rak1994; McHenry 1978; but see Drapeau and Ward 2007; Haeus-ler and McHenry 2007). A scatterplot of radius length re-gressed on femoral length is shown in figure 4. As expected,for any given femoral length, humans have much shorter radiithan the African apes. Note, too, that for this relationship,gorilla radii appear to lie along a continuation of the chim-panzee trajectory, something that was not the case for thehumerus.

With regard to the fossils, the left ulna of A.L. 288-1 ispreserved in two pieces that make only ambiguous contact

with each other. Hausler (2001) believes the bony contactbetween these two pieces to be good, and when they arejoined, the ulnar length of 191 mm suggests to him a radiuslength of ca. 181 mm. However, Richmond, Aiello, and Wood(2002) published a “humanlike” radius length of 174 mm for“Lucy” based on Schmid’s (1983) reconstruction (this valueclosely corresponds to the radius length [175 mm] estimatedfrom a 191-mm ulnar length for a sample of recent humans;T. W. Holliday, unpublished data). Kimbel, Johanson, andRak (1994) generate a radius length of 206 mm for the spec-imen, and Asfaw et al. (1999) generated a “chimpanzee-like”A.L. 288-1 radius length of 215 mm. Here I use this lastestimate as the specimen’s maximum length and 174 mm asits minimum. As seen in figure 4, the result is a radius thatperhaps falls just outside the modern human sample’s 95%confidence limits but is almost certainly longer than wouldbe expected for a modern human of its diminutive size. Yeteven the longest length estimate for the specimen does notfall near the Pan 95% confidence limits, a result consistentwith the findings of Haeusler and McHenry (2007) andDrapeau and Ward (2007), who found no evidence of chim-


panzee-like antebrachial elongation in Australopithecus afa-rensis.

As was done for A.L. 288-1, the radius length of LB1 wasestimated from its nearly complete right ulna at 190 mm(Morwood et al. 2005), placing it near the middle of thepredicted range of A.L. 288-1 and well beyond the 95% con-fidence limits for both Homo and Pan, albeit closer to theformer than to the latter. The taxonomically unassigned BOU-VP-12/1 specimen falls beyond the range of both Pan andHomo for this bivariate relationship, which might be expected,given its early date. However, recall that at least some of theestimated limb-segment lengths for this specimen are suspect.The radius length (255 mm, predicted from ulnar length byWalker and Leakey [1993]) of KNM-WT 15000 falls just be-yond the 95% confidence limits about the Homo sapiens in-dividuals while lying within the scatter of the H. sapiens sam-ple. The radio-femoral (RL/FL # 100) index for KNM-WT15000 (59.0) falls at the 98th percentile of the H. sapienssample ( ) and is statistically significantly differentn p 1,060from that sample at ( ; ). WhileP ! .05 t p 2.000 P p .0457s

this result is based on a juvenile (who is potentially patho-logical; Ohman et al. 2002), it does suggest that relative fore-arm length may have decreased over time in Homo.

Relative Pelvic Breadth

Narrow pelves have long been known to be a human autap-omorphy. They have been hypothesized to be related to ob-stetrical selection (Lovejoy 1988; Simpson et al. 2008), re-duced gut size (Aiello and Wheeler 1995), or thermal adap-tation to hot climates (Weaver and Hublin 2009). This lastexplanation seems unlikely in light of the discovery of thetropical yet wide Gona BSN49/P27 pelvis, which has beenattributed to Homo (Simpson et al. 2008; but see Ruff 2010).It is, however, possible that a narrower pelvis in humans isrelated to increasing locomotor efficiency, given that a broadpelvis is less biomechanically stable and less efficient in bipedallocomotion (Bramble and Lieberman 2004; Langdon 2005;Rosenberg and Trevathan 2007), and in light of the fact thatmore cursorial members of the Canidae, including dog breedssuch as greyhounds, tend to have narrower pelves than lesscursorial canids (Carrier, Chase, and Lark 2005; Schutz et al.2009). This hypothesis enjoys little experimental support,however, because in laboratory settings pelvic breadth is notassociated with locomotor efficiency in humans (Pontzer2012) and given the fact that effective limb length alone ex-plains 98% of the variance in locomotor efficiency across awide range of terrestrial animals (Pontzer 2007).

A scatterplot of bi-iliac (pelvic) breadth regressed on ace-tabular height is presented in figure 5. The former is a measureof body breadth while the latter is a pelvic reflection of bodymass. Note that Homo sapiens and Pan troglodytes have nearlyparallel regression lines, with the chimpanzees differing frommodern humans in that for any given acetabular size they are

expected to have a wider bi-iliac breadth. Irrespective of theirgeneric classification, the fossil hominins tend to fall into twoclusters: a small-bodied cluster and a large-bodied cluster. Thethree Australopithecus specimens for which bi-iliac breadthcan be reconstructed (A.L. 288-1 [Australopithecus afarensis],Sts 14 [Australopithecus africanus], and MH2 [Australopithecussediba]) fall among the wide-trunked chimpanzees in bivariatespace for this relationship, although Sts 14 and MH2 also falljust within the 95% confidence limits about the H. sapiensindividuals. The recently described 1.0-million-year-oldBSN49/P27 pelvis from Ethiopia, referred to Homo by Simp-son et al. (2008), has a wider pelvis than expected for evena Pan individual of its body size. The Gona pelvis lies closestin bivariate space to the female Levantine Neanderthal spec-imen Tabun C1, the bi-iliac breadth of which was acquiredfrom the virtual reconstruction of that pelvis by Weaver andHublin (2009). This specimen also falls well beyond the 95%confidence limits of the H. sapiens individuals.

The larger-sized fossil hominins on the right side of theplot include KNM-WT 15000, two middle Pleistocene hom-inins (Jinnuishan and Atapuerca Pelvis 1), and two Nean-derthals (La Chapelle-aux-Saints 1 and Kebara 2). The Na-riokotome specimen falls at the margins of the H. sapienssample, but its position in bivariate space is less reliable thanthe other fossils because of two factors. The first is that itsacetabular height was estimated from femoral head diameter.The second is that its predicted adult bi-iliac breadth is es-timated by Ruff (2010), and there is considerable debate asto how much further growth would be expected for the spec-imen (Graves et al. 2010) and even as to how wide the Na-riokotome pelvis was in its juvenile state (Ruff 1995, 2010;Simpson et al. 2010). The two middle Pleistocene Homo spec-imens, Atapuerca Pelvis 1 and the Jinnuishan pelvis, haveextremely wide bi-iliac breadths such that they fall well beyondthe 95% confidence limits about the H. sapiens individuals.

The La Chapelle-aux-Saints 1 bi-iliac breadth used here(292 mm) is taken from Trinkaus’s (2011) recent reconstruc-tion. While Neanderthals are known for their wide bi-iliacbreadths (Churchill 1998; Holliday 1997; Ruff 2002; Steeg-mann, Cerny, and Holliday 2002), among the large Nean-derthals included in the plot, La Chapelle-aux-Saints 1 fallswithin the H. sapiens’s scatter and in fact falls below the H.sapiens RMA line. This indicates that La Chapelle has a re-markably narrow pelvis for nonmodern Homo. This is in partdue to the specimen’s remarkably narrow sacrum (falling atthe 15th percentile of the H. sapiens [ ] sample). Then p 726specimen also lacks the marked iliac flare seen in Kebara 2.In contrast, Kebara 2, with a bi-iliac breadth of 313 mm, fallsjust beyond the 95% confidence limits of the H. sapiens sam-ple.

Body Size (Mass and Stature)

It has long been posited that body size increased with theemergence of the genus Homo. Body size is an important


Figure 5. Scatterplot of bi-iliac breadth regressed on acetabular height for Pleistocene/Holocene Homo sapiens, Pan, and Pliocene/Pleistocene early hominins. Australopiths are indicated by open squares, fossil Homo by filled squares. The ordinary least squaresregression lines for the comparative samples are represented by solid lines with the 95% confidence limits for the individuals indicatedabout them. The reduced major axis (RMA) regression lines for the comparative samples are the dashed lines. Homo sapiens RMAformula: ; Pan RMA formula: .y p 4.833x � 3.279 y p 5.429x � 33.197

variable because of its behavioral, biological, and ecologicalcorrelates (Calder 1984; Damuth and MacFadden 1990; Jung-ers 1985; Millien and Bovy 2010; Schmidt-Nielsen 1984).Across taxa, body mass is generally considered to be the mostappropriate measure of an animal’s overall size (Darveau etal. 2002; Garcia and da Silva 2006; West, Brown, and Enquist1997), but predicting body mass from skeletons involves theintroduction of a level of error that some find problematic(Smith 1996). It has been shown that among primates, limbarticular dimensions, especially articular breadths, are amongthe best predictors of body mass (Jungers 1990; McHenry1992; Ruff 2003; Ruff, Walker, and Teaford 1989). In esti-mating body mass for A.L. 288-1 using an all-hominoid re-gression, Jungers (1990) found that femoral head diametershowed the lowest standard error (SE) of the estimate andlowest percentage predicted error of six lower limb articularmeasurements, and importantly, McHenry (1992) found thathuman regression formulas likely outperformed cross hom-inoid predictive formulas in estimating fossil hominin bodymass. For this reason, a recent human least squares regressionformula (arguably the most appropriate formula for predic-

tion; see Hens, Konigsberg, and Jungers 2000; Jungers 1982;Martin, Genoud, and Hemelrijk 2005; Smith 2009) from Mc-Henry (1992) is used here to predict body mass from femoralhead diameter.

From this, the mean body mass of Australopithecus afarensisis estimated at ca. 41 kg, while those of Australopithecus af-ricanus, Australopithecus (P.) robustus, and Australopithecus(P.) boisei are slightly smaller, at ca. 37.3, 37.0, and 38.5 kg,respectively. In contrast, early (ca. 1.8–1.5 mya) Homo isfound to be considerably (ca. �33%) heavier, at ca. 54.5 kg.This is close to the mean of the recent low-latitude humansample (55.1 kg) although less heavy than that of recent high-latitude humans (59.9 kg) and considerably lighter than Homoneanderthalensis (ca. 73.6 kg) or late Pleistocene Homo sapiens(64.1 kg). These and other differences are best visualized inthe box plots in figure 6. There is a clear dichotomy in es-timated body mass between the members of the genera Aus-tralopithecus and Paranthropus (on the left side of the graph)and Homo on the right, a result similar to that obtained byPontzer (2012). Note, however, that while the earliest undis-puted members of the genus Homo (at least those that pre-


Figure 6. Box plots of estimated body mass for fossil hominin and recent human samples.

serve a femoral head) are larger than almost all of the indi-vidual australopiths, overall size in early Homo remainssmaller than that of Holocene or late Pleistocene H. sapienswith the exception of modern-day “Pygmoid” groups, whothemselves show a reduction in body size likely due to selec-tion related to climate, diet, high mortality rates, and/or islandbiogeography (Bailey 1991; Cavalli-Sforza 1986; Hiernaux andFroment 1976; Kuzawa and Bragg 2012; Lomolino 2005; Mig-liano and Guillon 2012; Migliano, Vinicius, and Mirazon Lahr2007).

Because of the importance of body mass in biology, thereis a rich literature on the estimation of body mass in fossilhominins (Hartwig-Scherer 1993; Jungers 1990; Kappelman1996; McHenry 1992; Ruff 1990; Ruff, Trinkaus, and Holliday1997). As mentioned before, some (e.g., Smith 1996) feel thatbecause of the introduction of prediction error, body massin fossils should not be estimated but rather comparable mea-surements that are known to be highly correlated with bodymass should be compared (although this alternative methodis not without its own problems; see Smith 1980). It shouldtherefore be noted that while body mass estimation formulasbased on femoral head size show low SEs of the estimate,they are subject to error and interpretation.

One potential problem with body mass estimation fromfemoral head size lies in the apparent difference in relativefemoral head size between Australopithecus (smaller femoralheads) and Homo (larger femoral heads). Given this differ-ence, should body mass in Australopithecus be estimated fromthe femoral head using a nonhuman-hominoid formula orrather a human and/or an all-hominoid one? In the smaller-body size range, this question is largely academic, but it be-comes problematic for larger-sized australopiths. For exam-ple, two estimates of body mass from femoral head diameter

for the small A.L. 288-1 specimen, based on the all-hominoidversus human regressions, are both 27.9 kg according to Mc-Henry (1992). Likewise, Jungers (1990) arrives at a similarestimate of 29.6 kg (only 6% higher) using a nonhuman-hominoid formula. In contrast, McHenry’s (1992) leastsquares femoral head estimate of body mass for the largerA.L. 333-3 specimen is 50 kg (based on a human formula),and 37% higher, at 68.6 kg, for an all-hominoid formula.Worse still, the nonhuman-hominoid least squares formulareported in Jungers (1991b) produces an estimated body massof 81.9 kg for the specimen, 64% higher than McHenry’s(1992) human estimate. The body mass estimates reportedhere must therefore be approached with caution. Despite thiscaveat, other researchers, using different means of body massestimation, have found a similar dichotomy in body massbetween australopiths and Homo erectus (Kappelman 1996;McHenry 1988; and see Pontzer 2012).

For recent humans, in addition to body mass, stature (orstanding height) is also frequently used as a proxy for bodysize (Auerbach and Sylvester 2011; Malina et al. 2004; Ruff2007; Zakrewski 2003). The problem with stature as a mea-surement is that with skeletal data it is always estimated witherror. Given potential differences in limb : trunk proportionsbetween living humans and at least some australopiths (Fel-desman, Kleckner, and Lundy 1990; Franciscus and Holliday1992; Jungers and Stern 1983; but see Wolpoff 1983), andgiven the fact that cranial height for most fossil hominins isshorter than that of modern humans (Graves et al. 2010; Ruffand Walker 1993), it is probably better to avoid estimatingstature in fossil hominins and instead analyze one of its majorconstituents. In this study, femoral length is used as a proxyfor stature. While an imperfect reflection of stature, it is thelongest limb bone making up a portion of stature, and among


Figure 7. Box plots of femoral length for fossil hominin and recent human samples.

the long bones, it tends to show the lowest SE of the estimatein stature prediction (Trotter 1970).

The distribution of femoral lengths of fossil and recenthominins is shown in the box plot presented in figure 7. Aswas the case with body mass, there is a dichotomy betweenAustralopithecus on the one hand (here represented solely bythe species A. afarensis) and Homo on the other. All of theundisputed early Homo specimens fall well within the sizerange of recent humans, and most have femora that are longerthan those of modern-day “Pygmoid” peoples (the earlyHomo mean and median lie just below a single large “Pyg-moid” outlier). In contrast, two of the five A. afarensis spec-imens fall well below the modern human range of variation,and for the three who fall within the modern human rangeof variation, none falls outside the “Pygmoid” range of var-iation.

Summary

With regard to the emergence of our genus, the postcranialmorphology of early Homo appears to have been mosaic innature. Of the features examined here, the most significantdifference between Australopithecus and early Homo is an in-crease in body size as reflected in both body mass and stature(importantly, this finding excludes OH 62, the femoral lengthof which remains unknown). In contrast, differences in limbproportions and body shape between the two genera are morenuanced. For example, the differences in lower limb lengthrelative to overall size between Australopithecus and earlyHomo are not as great as was once thought but rather therelatively short femora of some specimens of both genera (e.g.,

A.L. 288-1 or LB1) appear to be primarily a function of theirdiminutive body size (Franciscus and Holliday 1992; Hollidayand Franciscus 2009; Pontzer 2012). Likewise, while thehumero-femoral indices of early Homo are not as elevated asthose observed in Australopithecus afarensis (or likely thoseof Australopithecus africanus; see McHenry and Berger 1998a,1998b), it is evident that at least some specimens referred toearly Homo (e.g., Dmanisi 4167/4507) have longer humerirelative to their femora than do later members of the genus.This suggests that contra Asfaw et al. (1999), modern hu-manlike humero-femoral indices were likely not characteristicof hominins 2.5 mya but rather the low index values calculatedfor the taxonomically unassigned BOU-VP-12/1 specimen arethe result of erroneous reconstruction of at least one of thespecimen’s limb bone lengths (most likely its humerus).

It is more difficult to evaluate relative forearm length inHomo given a dearth of individuals who preserve antebrachialelements. In general the genus Australopithecus, including thelate species Australopithecus sediba (Berger et al. 2010), ap-pears to have had relatively longer forearms than modernhumans. At the same time, however, these forearms wereconsiderably shorter than those of Pan. While one must re-main mindful that KNM-WT 15000 is a juvenile, its predictedradius length relative to the length of its femur falls near theupper margins of the combined late Pleistocene/HoloceneHomo sapiens sample, which may suggest that early membersof the genus Homo, including Homo ergaster/Homo erectus,were characterized by longer forearms than later members ofthe genus.

In terms of relative pelvic breadth, early Homo is similarto Australopithecus in that it is characterized by wide bi-iliac


breadths. This plesiomorphic character is also observed inmiddle Pleistocene Homo and Homo neanderthalensis, sug-gesting that the relatively narrow bi-iliac breadth seen in peo-ple today is an autapomorphy of H. sapiens.

What emerges from these analyses, then, is that the originof the genus Homo is not a marked postcranial shift from anaustralopith-like morphological pattern to an H. sapiens–likeone, but as has been noted by others (e.g., Berger et al. 2010;Kivell et al. 2011; Lordkipanidze et al. 2007; Pontzer 2012;Pontzer et al. 2010), it is more complex. For example, lowerlimb length appears to be as long in Australopithecus as it isin early Homo once one corrects for differences in body size.This brings up the fact that one long-recognized differencebetween the australopiths and Homo that remains borne outby the current analyses is that Homo tends to be much largerin body size than Australopithecus or Paranthropus. This pat-tern holds despite recent work (Graves et al. 2010; Ohmanet al. 2002) finding that the original projection of the adultstature of KNM-WT 15000 (Ruff and Walker 1993) was fartoo tall, or the revelation that many of the 1.77-million-year-old Homo specimens from Dmanisi were of relatively dimin-utive size (Lordkipanidze et al. 2007), or even the discoveryof relatively long-legged australopiths such as KSD-VP-1/1(Haile-Selassie et al. 2010). It is a robust pattern: while onaverage smaller than many people today, early Homo wassignificantly larger than the australopiths in both body massand stature. And because among mammals, home-range sizeis to a great extent a function of body size (Eisenberg 1990;Harestad and Bunnell 1979; Kelt and Van Vuren 2001; McNab1963; Vieira and Cunha 2008), it is almost certain that Homoranged more widely than Australopithecus, a notion under-scored by the presence of Homo (and apparent lack of Aus-tralopithecus or Paranthropus) outside of Africa.

Increased ranging in early Homo is most likely tied to anincrease in carnivory (Aiello and Key 2002; Martin 1981;McHenry 1994; O’Connell, Hawkes, and Blurton Jones 1999).First, mammalian carnivores need larger home ranges thancomparably sized herbivores or omnivores (Gittleman andHarvey 1982; Harestad and Bunnell 1979; Lindstedt, Miller,and Buskirk 1986) and tend to exploit much greater dayranges than noncarnivores as well (Carbone et al. 2005). Sec-ond, larger bodies (and bigger brains; Aiello and Wheeler1995) would likely only be made possible via an increase innutritional quality of the diet, and animal muscle and fat arevery high-caloric food sources known to have been exploitedby early Homo. Evidence in favor of increased carnivory inearly Homo include the 1.7-million-year-old partial H. erg-aster/H. erectus skeleton KNM-ER 1808, argued by Walker,Zimmerman, and Leakey (1982) to show pathological bonesuggestive of hypervitaminosis A, which can be caused viaingestion of carnivore livers. Likewise, from an archaeologicalperspective, the Acheulean Industry, which is now known todate as early as 1.76 mya (Lepre et al. 2011), is temporallycorrelated with zooarchaeological evidence from multiple sites

indicative of a shift toward consistent early access to carcassesby hominins (Cachel and Harris 1998).

Despite the important adaptive implications surroundingincreased ranging activity in early Homo, aside from a fewanatomical features of the foot (DeSilva et al. 2012; Pontzeret al. 2010; Zipfel et al. 2011), pelvis (Berger et al. 2010; Kibiiet al. 2011), and some features related to head carriage (Bram-ble and Lieberman 2004), there is little evidence of a majorlocomotor shift between Australopithecus and Homo. Insteadit appears that Homo augmented an existing australopith pat-tern almost exclusively via an increase in body size (and seePontzer 2012).

Following these observations, at least from a postcranialadaptive-zone perspective, there are insufficient data to war-rant the removal of Homo habilis and Homo rudolfensis fromthe genus Homo as recommended by Wood and Collard (Col-lard and Wood 2007; Wood and Collard 1999a, 1999b). Ofthe cranial remains assigned to these two species, only two(OH 62 and KNM-ER 3735) have associated postcrania, andin both cases, the postcranial remains are extremely frag-mentary and therefore relatively uninformative (Haeusler andMcHenry 2007). In this light, if any of the long (but isolated)femora dating to ca. 1.8 mya in Africa is referred to either ofthese species, then we have probably underestimated bothbody size and lower limb length in these two taxa. Addition-ally, because H. ergaster/H. erectus may have retained a some-what elongated upper limb and an equally broad pelvis, itappears that many of the perceived proportional differencesbetween H. habilis and/or H. rudolfensis on the one hand andspecimens assigned to H. ergaster or H. erectus on the otherare diminished.2

As for the monophyletic nature of the genus Homo, whilesome recent cladistic analyses (e.g., Cameron and Groves2004) fail to cluster H. habilis and H. rudolfensis with H.sapiens to the exclusion of any australopith taxon, it is im-portant to note that Cameron and Groves’s (2004) result isentirely due to the inclusion of Kenyanthropus platyops in theanalyses, a paleospecies whose taxonomic validity has beenquestioned (White 2003). Simply removing this species fromCameron and Groves’s (2004) analysis makes Homo (sensulato) a monophylum. More importantly, the most rigorouscladistic analyses of fossil hominins done to date, includingsome analyses that include K. platyops (Strait and Grine 2004)show Homo (sensu lato) to be a monophylum. In any case,the close phylogenetic relationship of Homo and the austral-opiths (it seems unlikely that they split much before 2.5 mya)probably lent itself to some (albeit low) level of interbreedingbetween these taxa, which in turn could produce confusingpatterns of synapomorphy (Holliday 2003, 2006; Jolly 2001).

In conclusion, then, while the emergence of Homo (sensu

2. There are also numerous craniodental characters that support theinclusion of Homo habilis and Homo rudolfensis in the genus Homo(Anton 2012; Bromage, Schrenk, and Zonneveld 1995; Kimbel, Johanson,and Rak 1997; Rightmire and Lordkipanidze 2009).


lato) is an important chapter in human evolution—one likelyinvolving increased ranging, carnivory, and encephalization—it is also the case that we now have the resolution in the fossilrecord to actually capture the gradual mosaic nature of thatemergence. The long-standing idea of the origin of Homo(sensu stricto) as a punctuated event involving the geologicallyrapid emergence of hominins who (at least from the neckdown) were H. sapiens–like seems less applicable now thanit did a decade ago.

Acknowledgments

I would like to thank Leslie Aiello and Susan Anton for in-viting me to participate in the Wenner-Gren symposium, andto them and all the other participants, my thanks for makingthe symposium such a phenomenal and unforgettable expe-rience. Thanks also to Tim Weaver for supplying a bi-iliacbreadth estimate for Tabun C1 and to Erik Trinkaus for insightinto the morphology of the La Chapelle-aux-Saints pelvis.Thanks, too, to Campbell Rolian for generously providingsome of the Pan and Gorilla data included in these analyses.This work was supported in part by the National ScienceFoundation, the Leakey Foundation, Tulane University, andthe University of the Witwatersrand.

References CitedAiello, Leslie C., and Cathy Key. 2002. The energetic consequences of being



Anton, Susan C. 2012. Early Homo: who, when, and where. Current Anthro-pology 53(suppl. 6):S278–S298.

Asfaw, Berhane, Tim White, Owen Lovejoy, Bruce Latimer, Scott Simpson,and Gen Suwa. 1999. Australopithecus garhi: a new species of early hominidfrom Ethiopia. Science 284:629–635.

Auerbach, Benjamin M., and Adam D. Sylvester. 2011. Allometry and apparentparadoxes in human limb proportions: implications for scaling factors.American Journal of Physical Anthropology 144:382–391.

Avise, John C., William S. Nelson, and Hiroaki Sugita. 1994. A speciationalhistory of “living fossils”: molecular evolutionary patterns in horseshoecrabs. Evolution 48:1986–2001.

Bailey, Robert C. 1991. The comparative growth of Efe Pygmies and Africanfarmers from birth to age 5 years. Annals of Human Biology 18:113–120.

Berger, Lee R., Darryl J. de Ruiter, Steven E. Churchill, Peter Schmid, KristianJ. Carlson, Paul H. G. M. Dirks, and Job M. Kibii. 2010. Australopithecussediba: a new species of Homo-like australopith from South Africa. Science328:195–204.


Bromage, Timothy G., Friedemann Schrenk, and Frans W. Zonneveld. 1995.Paleoanthropology of the Malawi Rift: an early hominid mandible fromthe Chiwondo Beds, northern Malawi. Journal of Human Evolution 22:79–108.

Brown, Francis H., and Ian McDougall. 1993. Geological setting and age. InThe Nariokotome Homo erectus skeleton. Alan Walker and Richard Leakey,eds. Pp. 9–20. Cambridge, MA: Harvard University Press.

Brown, Peter, Thomas Sutikna, Michael J. Morwood, Raden P. Soejono, Jat-miko, E. Wayhu Saptomo, and Rokus Awe Due. 2004. A new small-bodiedhominin from the late Pleistocene of Flores, Indonesia. Nature 431:1055–1061.

Cachel, Susan, and John W. K. Harris. 1998. The lifeways of Homo erectusinferred from archaeology and evolutionary ecology: a perspective fromEast Africa. In Early human behaviour in global context: the rise and diversityof the lower Palaeolithic record. Michael D. Petraglia and Ravi Korisettar,eds. Pp. 108–132. New York: Routledge.

Calder, William A. 1984. Size, function and life history. Cambridge, MA: Har-vard University Press.

Cameron, David W., and Colin P. Groves. 2004. Bones, stones and molecules:“Out of Africa” and human origins. Amsterdam: Elsevier.

Carbone, Chris, Guy Cowlishaw, Nick J. B. Isaac, and J. Marcus Rowcliffe.2005. How far do animals go? determinants of day range in mammals.American Naturalist 165:290–297.

Carrier, David R., Kevin Chase, and Karl G. Lark. 2005. Genetics of canidskeletal variation: size and shape of the pelvis. Genome Research 15:1825–1830.

Cavalli-Sforza, Luigi L. 1986. African Pygmies. New York: Academic Press.Churchill, Steven E. 1998. Cold adaptation, heterochrony, and the Neandertals.

Evolutionary Anthropology 7:46–61.Collard, Mark, and Bernard Wood. 2007. Defining the genus Homo. In Hand-

book of paleoanthropology, vol. 3 of Phylogeny of hominids. Winfried Henkeand Ian Tattersall, eds. Pp. 1575–1611. Berlin: Springer.

Damuth, John, and Bruce J. MacFadden, eds. 1990. Body size in mammalianpaleobiology: estimation and biological implications. Cambridge: CambridgeUniversity Press.

Darveau, Charles-A., Raul K. Suarez, Russel D. Andrews, and Peter W. Ho-chachka. 2002. Allometric cascade as a unifying principle of body masseffects on metabolism. Nature 417:166–170.

DeSilva, Jeremy M., Bernhard Zipfel, Robert S. Kidd, Kristian J. Carlson,Steven E. Churchill, and Lee R. Berger. 2012. The primitive aspects of thefoot and ankle of Australopithecus sediba: presented at the eighty-first annualmeeting of the American Association of Physical Anthropologists. AmericanJournal of Physical Anthropology 147(suppl. 54):129.

Drapeau, Michelle S. M., and Carol V. Ward. 2007. Forelimb segment lengthproportions in extant hominoids and Australopithecus afarensis. AmericanJournal of Physical Anthropology 132:327–343.

Eisenberg, John F. 1990. The behavioral/ecological significance of body sizein the Mammalia. In Body size in mammalian paleobiology: estimation andbiological implications. John Damuth and Bruce J. MacFadden, eds. Pp. 25–37. Cambridge: Cambridge University Press.

Feldesman, Marc R., J. Geoffrey Kleckner, and John K. Lundy. 1990. Femur/stature ratio and estimates of stature in mid and late-Pleistocene fossilhominids. American Journal of Physical Anthropology 83:359–372.

Franciscus, Robert G., and Trenton W. Holliday. 1992. Hindlimb skeletalallometry in Plio-Pleistocene hominids with special reference to AL-288-1(‘‘Lucy”). Bulletins et memoires de la Societe d’anthropologie de Paris 22:1–16.

Gabunia, Leo, Abesalom Vekua, and David Lordkipanidze. 2000. The envi-ronmental contexts of early human occupation of Georgia (Transcaucasia).Journal of Human Evolution 38:785–802.

Galik, Karol, Brigitte Senut, Martin Pickford, Dominique Gommery, JacquesTreil, Adam J. Kuperavage, and Robert B. Eckhardt. 2004. External andinternal morphology of the BAR 1002000 Orrorin tugenensis femur. Science305:1450–1453.

Garcia, Guilherme J. M., and Jafferson Kamphorst Leal da Silva. 2006. In-terspecific allometry of bone dimensions: a review of the theoretical models.Physics of Life Reviews 3:188–209.

Gittleman, John L., and Paul H. Harvey. 1982. Carnivore home-range size,metabolic needs and ecology. Behavioral Ecology and Sociobiology 10:57–63.

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, David J., Adam D. Gordon, and Brian G. Richmond. 2007. Limb-sizeproportions in Australopithecus afarensis and Australopithecus africanus.Journal of Human Evolution 52:187–200.


———. 2007. Evolutionary reversals of limb proportions in early hominids?evidence from KNM-ER 3735 (Homo habilis). Journal of Human Evolution53:383–405.

Haile-Selassie, Yohannes, Bruce M. Latimer, Mulugeta Alene, Alan L. Deino,Luis Gibert, Stephanie M. Melillo, Beverly Z. Saylor, Gary R. Scott, and C.Owen Lovejoy. 2010. An early Australopithecus afarensis postcranium from


Woranso-Mille, Ethiopia. Proceedings of the National Academy of Sciences ofthe USA 107:12121–12126.

Harestad, Alton S., and Fred L. Bunnell. 1979. Home range and body weight:a reevaluation. Ecology 60:389–402.

Harmon, Elizabeth H. 2005. A comparative analysis of femoral morphologyin Australopithecus afarensis: implications for the evolution of bipedal lo-comotion. PhD dissertation, Arizona State University, Tempe.

———. 2009. The shape of the early hominin femur. American Journal ofPhysical Anthropology 139:154–171.

Hartwig-Scherer, Sigrid. 1993. Body weight prediction in early fossil hominids:towards a taxon-“independent” approach. American Journal of Physical An-thropology 92:17–36.

Hartwig-Scherer, Sigrid, and Robert D. Martin. 1991. Was “Lucy” more humanthan her “child”? observations on early hominid postcranial skeletons. Jour-nal of Human Evolution 21:439–449.

Hausler, Martin. 2001. New insights into the locomotion of Australopithecusafricanus: implications of the partial skeleton Stw 431 (Sterkfontein, SouthAfrica). PhD dissertation, Universitat Zurich.

Hennig, Willi. 1966. Phylogenetic systematics. D. Dwight Davis and RainerZangerl, trans. Urbana: University of Illinois Press.

Hens, Samantha M., Lyle W. Konigsberg, and William L. Jungers. 2000. Es-timating stature in fossil hominids: which regression model and referencesample to use? Journal of Human Evolution 38:767–784.

Hiernaux, Jean, and Alain Froment. 1976. The correlations between anthro-pobiological and climatic variables in sub-Saharan Africa: revised estimates.Human Biology 48:757–767.

Holliday, Trenton W. 1997. Postcranial evidence of cold adaptation in Eu-ropean Neandertals. American Journal of Physical Anthropology 104:245–258.

———. 2003. Species concepts, reticulation, and human evolution. CurrentAnthropology 44:653–673.

———. 2006. Neanderthals and modern humans: an example of a mam-malian syngameon? In Neanderthals revisited: new approaches and perspec-tives. Katerina Harvati and Terry Harrison, eds. Pp. 281–298. New York:Springer.

Holliday, Trenton W., and Robert G. Franciscus. 2009. Body size and itsconsequences: allometry and the lower limb length of Liang Bua 1 (Homofloresiensis). Journal of Human Evolution 57:253–257.

———. 2012. Humeral length allometry in African hominids (sensu lato) withspecial reference to A.L. 288-1 and Liang Bua 1. Paleoanthropology 2012:1–12.

Howell, F. Clark, and Bernard A. Wood. 1974. Early hominid ulna from theOmo Basin, Ethiopia. Nature 249:174–176.

Isler, Karin, and Carel P. van Schaik. 2012. How our ancestors broke throughthe gray ceiling: comparative evidence for cooperative breeding in earlyHomo. Current Anthropology 53(suppl. 6):S453–S465.

Johanson, Donald C., C. Owen Lovejoy, William H. Kimbel, Tim D. White,Steven C. Ward, Michael E. Bush, Bruce M. Latimer, and Yves Coppens.1982. Morphology of the Pliocene partial skeleton (A.L. 288-1) from theHadar Formation, Ethiopia. American Journal of Physical Anthropology 57:403–451.

Johanson, Donald C., Fidelis T. Masao, Gerald G. Eck, Tim D. White, RobertC. Walter, William H. Kimbel, Berhane Asfaw, Paul Manega, Prosper Ndes-sokia, and Gen Suwa. 1987. New partial skeleton of Homo habilis fromOlduvai Gorge, Tanzania. Nature 327:205–209.

Jolly, Clifford J. 2001. A proper study for mankind: analogies from papioninmonkeys and their implications for human evolution. Yearbook of PhysicalAnthropology 44:177–204.

Jungers, William L. 1982. Lucy’s limbs: skeletal allometry and locomotion inAustralopithecus afarensis. Nature 297:676–678.

———. 1985. Size and scaling in primate biology. New York: Plenum.———. 1988. Relative joint size and hominoid locomotor adaptations with

implications for the evolution of hominid bipedalism. Journal of HumanEvolution 17:247–265.

———. 1990. Problems and methods in reconstructing body size in fossilprimates. In Body size in mammalian paleobiology: estimation and biologicalimplications. John Damuth and Bruce J. MacFadden, eds. Pp. 103–118.Cambridge: Cambridge University Press.

———. 1991a. A Pygmy perspective on body size and shape in Australo-pithecus afarensis (AL 288-1, “Lucy”). In Origine(s) de la bipedie chez leshominides. Brigitte Senut and Yves Coppens, eds. Pp. 215–224. Paris: Cahiersde paleoanthropologie.

———. 1991b. Scaling of postcranial joint size in hominoid primates. HumanEvolution 6:391–399.

Jungers, William L., Susan G. Larson, William Harcourt-Smith, Michael J.Morwood, Thomas Sutikna, Rokhus Awe Due, and Tony Djubiantono.2009. Descriptions of the lower limb skeleton of Homo floresiensis. Journalof Human Evolution 57:538–554.

Jungers, William L., and Jack T. Stern Jr. 1983. Body proportions, skeletalallometry and locomotion in the Hadar hominids: a reply to Wolpoff.Journal of Human Evolution 12:673–684.

Kappelman, John. 1996. The evolution of body mass and relative brain sizein fossil hominids. Journal of Human Evolution 30:243–276.

Kelt, Douglas A., and Dirk H. Van Vuren. 2001. The ecology and macroecologyof mammalian home range area. American Naturalist 157:637–645.

Kennedy, G. E. 1983. Some aspects of femoral morphology in Homo erectus.Journal of Human Evolution 12:587–616.

Kibii, Job M., Steven E. Churchill, Peter Schmid, Kristian J. Carlson, NichelleD. Reed, Darryl J. de Ruiter, and Lee R. Berger. 2011. A partial pelvis ofAustralopithecus sediba. Science 333:1407–1411.

Kimbel, William H., and Lucas K. Delezene. 2009. “Lucy” redux: a review ofresearch on Australopithecus afarensis. Yearbook of Physical Anthropology 52:2–48.

Kimbel, William H., Donald C. Johanson, and Yoel Rak. 1994. The first skulland other new discoveries of Australopithecus afarensis at Hadar, Ethiopia.Nature 368:449–451.

———. 1997. Systematic assessment of a maxilla of Homo from Hadar, Ethi-opia. American Journal of Physical Anthropology 103:235–262.

Kivell, Tracy L., Job M. Kibii, Steven E. Churchill, Peter Schmid, and Lee R.Berger. 2011. Australopithecus sediba hand demonstrates mosaic evolutionof locomotor and manipulative abilities. Science 333:1411–1416.

Korey, Kenneth A. 1990. Deconstructing reconstruction: the OH 62 humero-femoral index. American Journal of Physical Anthropology 83:25–34.

Kuzawa, Christopher W., and Jared M. Bragg. 2012. Plasticity in humanlife history strategy: implications for contemporary human variation andthe evolution of genus Homo. Current Anthropology 53(suppl. 6):S369–S382.

Langdon, John H. 2005. The human strategy: an evolutionary perspective onhuman anatomy. Oxford: Oxford University Press.

Lepre, Christopher J., Helene Roche, Dennis V. Kent, Sonia Harmand, RhondaL. Quinn, Jean-Philippe Brugal, Pierre-Jean Texier, Arnaud Lenoble, andCraig S. Feibel. 2011. An earlier origin for the Acheulian. Nature 477:82–85.

Lindstedt, Stan L., Brian J. Miller, and Steven W. Buskirk. 1986. Home range,time, and body size in mammals. Ecology 67:413–418.

Lomolino, Mark V. 2005. Body size evolution in insular vertebrates: generalityof the island rule. Journal of Biogeography 32:1683–1699.


Lovejoy, C. Owen. 1981. The origin of man. Science 211:341–350.———. 1988. Evolution of human walking. Scientific American 259:118–

125.———. 2009. Reexamining human origins in light of Ardipithecus ramidus.

Science 326:74e1–74e8.Lovejoy, C. Owen, Kingsbury G. Heiple, and Albert H. Burstein. 1973. The

gait of Australopithecus. American Journal of Physical Anthropology 38:757–780.

Malina, Robert M., Maria Eugenia Pena Reyes, Swee Kheng Tan, Peter H.Buschang, Bertis B. Little, and Sławomir Koziel. 2004. Secular change inheight, sitting height and leg length in rural Oaxaca, southern Mexico: 1968–2000. Annals of Human Biology 31:615–633.

Martin, Robert A. 1981. On extinct hominid population densities. Journal ofHuman Evolution 10:427–428.

Martin, Robert D., Michel Genoud, and Charlotte K. Hemelrijk. 2005. Prob-lems of allometric scaling analysis: examples from mammalian reproductivebiology. Journal of Experimental Biology 208:1731–1747.

Mayr, Ernst. 1942. Systematics and the origin of species. New York: ColumbiaUniversity Press.

———. 2000. A critique from the biological species concept perspective: whatis a species, and what is not? In Species concepts and phylogenetic theory: adebate. Quentin D. Wheeler and Rudolf Meier, eds. Pp. 93–100. New York:Columbia University Press.


McHenry, Henry M. 1978. Fore- and hindlimb proportions in Plio-Pleistocenehominids. American Journal of Physical Anthropology 49:15–22.

———. 1988. New estimates of body weight in early hominids and theirsignificance to encephalization and megadontia in “robust” australopithe-cines. In Evolutionary history of the “robust” australopithecines. Fred E. Grine,ed. Pp. 133–148. New York: Aldine de Gruyter.



McHenry, Henry M., and Lee R. Berger. 1998a. Body proportions in Aus-tralopithecus afarensis and A. africanus and the origin of the genus Homo.Journal of Human Evolution 35:1–22.

———. 1998b. Limb lengths in Australopithecus and the origin of the genusHomo. South African Journal of Science 94:447–450.

McNab, Brian K. 1963. Bioenergetics and the determination of home rangesize. American Naturalist 97:133–140.


Migliano, Andrea Bamberg, Lucio Vinicius, and Marta Mirazon Lahr. 2007.Life-history trade-offs explain the evolution of human Pygmies. Proceedingsof the National Academy of Sciences of the USA 104:20216–20219.

Millien, Virginie, and Helene Bovy. 2010. When teeth and bones disagree:body mass estimation of a giant extinct rodent. Journal of Mammalogy 91:11–18.

Morwood, Michael J., Peter Brown, Jatmiko, Thomas Sutikna, E. WahyuSaptomo, Kira E. Westaway, Rokus Awe Due, et al. 2005. Further evidencefor small-bodied hominins from the late Pleistocene of Flores, Indonesia.Nature 437:1012–1017.

Napier, John R. 1964. The evolution of bipedal walking in the hominids.Archives de Biologie (Liege) 75:673–708.

O’Connell, James F., Kristen Hawkes, and Nicholas G. Blurton Jones. 1999.Grandmothering and the evolution of Homo erectus. Journal of HumanEvolution 36:461–485.

Ohman, James C., C. Wood, Bernard Wood, Robin H. Crompton, MichaelM. Gunther, Yu Li, Russell Savage, and Weijie Wang. 2002. Stature-at-deathof KNM-WT 15000. Human Evolution 17:129–142.

Pontzer, Herman. 2007. Effective limb length and the scaling of locomotorcost in terrestrial animals. Journal of Experimental Biology 210:1752–1761.

———. 2012. Ecological energetics in early Homo. Current Anthropology53(suppl. 6):S346–S358.

Pontzer, Herman, Campbell Rolian, G. Philip Rightmire, Tea Jashashvili, Mar-cia S. Ponce de Leon, David Lordkipanidze, and Christoph P. E. Zollikofer.2010. Locomotor anatomy and biomechanics of the Dmanisi hominins.Journal of Human Evolution 58:492–504.

Reno, Philip L., David DeGusta, Maria A. Serrat, Richard S. Meindl, Tim D.White, Robert B. Eckhardt, Adam J. Kuperavage, Karol Galik, and C. OwenLovejoy. 2005. Plio-Pleistocene hominid limb proportions: evolutionaryreversals or estimation errors? Current Anthropology 46:575–588.

Richmond, Brian G., Leslie C. Aiello, and Bernard A. Wood. 2002. Earlyhominin limb proportions. Journal of Human Evolution 43:529–548.

Richmond, Brian G., and William L. Jungers. 2008. Orrorin tugenensis femoralmorphology and the evolution of hominin bipedalism. Science 319:1662–1665.

Rightmire, G. Philip, and David Lordkipanidze. 2009. Comparisons of earlyPleistocene skulls from East Africa and the Georgian Caucasus: evidencebearing on the origin and systematics of genus Homo. In The first humans:origin and early evolution of the genus Homo. Fred E. Grine, John G. Fleagle,and Richard E. Leakey, eds. Pp. 39–48. Vertebrate Paleontology and Paleo-anthropology Series. New York: Springer.

Robinson, John T. 1972. Early hominid posture and locomotion. Chicago: Uni-versity of Chicago Press.

Rosenberg, Karen R., and Wenda R. Trevathan. 2007. An anthropologicalperspective on the evolutionary context of preeclampsia in humans. Journalof Reproductive Immunology 76:91–97.

Ruff, Christopher B. 1988. Hindlimb articular surface allometry in Homi-noidea and Macaca, with comparisons to diaphyseal scaling. Journal ofHuman Evolution 17:687–714.

———. 1990. Body mass and hindlimb bone cross-sectional and articulardimensions in anthropoid primates. In Body size in mammalian paleobiology:estimation and biological implications. John Damuth and Bruce J. Mac-Fadden, eds. Pp. 119–149. Cambridge: Cambridge University Press.

———. 1995. Biomechanics of the hip and birth in early Homo. AmericanJournal of Physical Anthropology 98:527–574.

———. 2002. Variation in human body size and shape. Annual Review ofAnthropology 31:211–232.

———. 2003. Long bone articular and diaphyseal structure in Old Worldmonkeys and apes. 2. Estimation of body mass. American Journal of PhysicalAnthropology 120:16–37.

———. 2007. Body size prediction from juvenile skeletal remains. AmericanJournal of Physical Anthropology 133:698–716.

———. 2010. Body size and body shape in early hominins: implications ofthe Gona pelvis. Journal of Human Evolution 58:166–178.

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

Ruff, Christopher B., and Alan Walker. 1993. Body size and body shape. InThe Nariokotome Homo erectus skeleton. Alan Walker and Richard Leakey,eds. Pp. 234–265. Cambridge, MA: Harvard University Press.

Ruff, Christopher B., Alan C. Walker, and Mark F. Teaford. 1989. Body mass,sexual dimorphism and femoral proportions of Proconsul from Rusinga andMfangano Islands, Kenya. Journal of Human Evolution 18:515–536.

Schmid, Peter. 1983. Eine Rekonstruktion des Skelettes von A.L. 288-1 (Hadar)und deren Konsequenzen. Folia Primatologica 40:282–306.

Schmidt-Nielsen, Knut. 1984. Scaling: why is animal size so important? Cam-bridge: Cambridge University Press.

Schutz, Heidi P., David Polly, Jonathan D. Krieger, and Robert P. Guralnick.2009. Differential sexual dimorphism: size and shape in the cranium andpelvis of gray foxes (Urocyon). Biological Journal of the Linnean Society 96:339–353.



Simpson, Scott W., Linda B. Spurlock, C. Owen Lovejoy, and Bruce Latimer.2010. A new reconstruction of the KNM-WT 15000 juvenile male pelvis.Presented at the seventy-ninth annual meeting of the American Associationof Physical Anthropologists. American Journal of Physical Anthropology141(suppl. 50):189.

Smith, Richard J. 1980. Rethinking allometry. Journal of Theoretical Biology87:97–111.

———. 1996. Biology and body size in human evolution: statistical inferencemisapplied. Current Anthropology 37:451–481.

———. 2009. Use and misuse of the reduced major axis for line-fitting.American Journal of Physical Anthropology 140:476–486.

Sokal, Robert R., and F. James Rohlf. 1981. Biometry. 2nd edition. San Fran-cisco: W. H. Freeman.

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.

Steudel-Numbers, Karen L. 2006. Energetics in Homo erectus and other earlyhominins: the consequences of increased lower limb length. Journal of Hu-man Evolution 51:445–453.

Steudel-Numbers, Karen L., and Michael J. Tilkens. 2004. The effect of lowerlimb length on the energetic cost of locomotion: implications for fossilhominins. Journal of Human Evolution 47:95–109.

Strait, David S., and Fred E. Grine. 2004. Inferring hominoid and early hom-inid phylogeny using craniodental characters: the role of fossil taxa. Journalof Human Evolution 47:399–452.

Trinkaus, Erik. 2011. The postcranial dimensions of the La Chapelle-aux-Saints1 Neandertal. American Journal of Physical Anthropology 145:461–468.

Trotter, Mildred. 1970. Estimation of stature from intact long bones. In Per-sonal identification in mass disasters. T. Dale Stewart, ed. Pp. 71–83. Wash-ington, DC: National Museum of Natural History.

Vieira, Marcus Vinıcius, and Andre de Almeida Cunha. 2008. Scaling bodymass and use of space in three species of marsupials in the Atlantic Forestof Brazil. Austral Ecology 33:872–879.

Walker, Alan C. 1973. New Australopithecus femora from East Rudolf, Kenya.Journal of Human Evolution 2:545–555.

Walker, Alan C., and Richard E. F. Leakey. 1993. The postcranial bones. InThe Nariokotome Homo erectus skeleton. Alan Walker and Richard Leakey,eds. Pp. 95–160. Cambridge, MA: Harvard University Press.

Walker, Alan C., Michael R. Zimmerman, and Richard E. F. Leakey. 1982. Apossible case of hypervitaminosis A in Homo erectus. Nature 296:248–250.

Weaver, Timothy D., and Jean-Jacques Hublin. 2009. Neandertal birth canal


shape and the evolution of human childbirth. Proceedings of the NationalAcademy of Sciences of the USA 106:8151–8156.

West, Geoffrey B., James H. Brown, and Brian J. Enquist. 1997. A generalmodel for the origin of allometric scaling laws in biology. Science 276:122–126.

White, Tim. 2003. Early hominids: diversity or distortion? Science 299:1994–1997.

Wolpoff, Milford H. 1983. Lucy’s little legs. Journal of Human Evolution 12:443–453.

Wood, Bernard, and Mark Collard. 1999a. The changing face of Homo. Evo-lutionary Anthropology 8:195–207.

———. 1999b. The human genus. Science 284:65–71.Zakrewski, Sonia R. 2003. Variation in ancient Egyptian stature and body

proportions. American Journal of Physical Anthropology 121:219–229.Zipfel, Bernhard, Jeremy M. DeSilva, Robert S. Kidd, Kristian J. Carlson,

Steven E. Churchill, and Lee R. Berger. 2011. The foot and ankle of Aus-tralopithecus sediba. Science 333:1417–1420.



Ecological Energetics in Early Homo

by Herman Pontzer

CA� Online-Only Material: Supplement A

Models for the origin of the genus Homo propose that increased quality of diet led to changes in ranging ecologyand selection for greater locomotor economy, speed, and endurance. Here, I examine the fossil evidence for postcranialchange in early Homo and draw on comparative data from living mammals to assess whether increased diet qualityhas led to selection for improved locomotor performance in other lineages. Body mass estimates indicate earlyHomo, both males and females, were approximately 33% larger than australopiths, consistent with archeologicalevidence indicating an ecological change with the origins of our genus. However, many of the postcranial featuresthought to be derived in Homo, including longer hind limbs, are present in Australopithecus, challenging thehypothesis that early Homo is marked by significant change in walking and running performance. Analysis of energybudgets across mammals suggests that the larger body mass and increased diet quality in early Homo may reflectan increase in the hominin energy budget. Expanding the energy budget would enable greater investment inreproduction without decreasing energy available for larger brains or increased activity. Food sharing and increasedadiposity, which decrease variance in food energy availability, may have been integral to this metabolic strategy.

Introduction

In describing the first specimens of Homo habilis 5 decadesago, Leakey, Tobias, and Napier (1964) argued for two distinctgrades within the hominin lineage. Australopithecus species ofsouth and east Africa—with their large molars and ape-sizebrains—were viewed as more primitive in their cognitive abil-ities and behavior, while H. habilis—with its larger brain,smaller molars, and dexterous hands capable of making theearliest stone tools—marked the beginnings of modern hu-man ability and behavior (Leakey, Tobias, and Napier 1964).While Leakey, Tobias, and Napier (1964:9) viewed Australo-pithecus and Homo as “two branches of Hominidae evolvingside by side,” subsequent work has suggested that the originof our genus marks an ecological shift in foraging behaviorand diet from an earlier australopith strategy that persistedthroughout much of the Pliocene (Aiello and Wheeler 1995;Bramble and Lieberman 2004; Conroy and Pontzer 2012;Leonard and Robertson 1997; O’Connell, Hawkes, andBlurton Jones 1999; Shipman and Walker 1989). In this re-view, I briefly outline current ecological models for the evo-lution of the genus Homo and examine whether the proposedchanges in foraging behavior are consistent with comparative

Herman Pontzer is Assistant Professor in the Department ofAnthropology, Hunter College (728 North Building, 695 ParkAvenue, New York, New York 10065, U.S.A. [[email protected]]). This paper was submitted 12 XII 11, accepted 15VI 12, and electronically published 3 X 12.

data from living mammals and the fossil evidence of post-cranial anatomy and locomotor performance in Plio-Pleis-tocene hominins.

Researchers have long focused on foraging behavior in re-constructing hominin ecology and evolution (Bramble andLieberman 2004; Dart 1949, 1957; Darwin 1871; Hawkes etal. 1998; Lee and Devore 1968; Lovejoy 1981). Unlike theother great apes, which travel modest distances each day insearch of plant foods and must fend for themselves, humanforagers range widely, hunt regularly, and share religiously(Marlowe 2005). Early models of hominin evolution proposedthat these defining aspects of modern human foraging be-havior arose early in the hominin lineage. Darwin (1871:40)offered that our species’ “preeminent success in the battle oflife” was in large part due to our progenitors’ bipedal posture,which freed their hands and allowed them “to defend them-selves with stones or clubs, to attack prey, or otherwise obtainfood.” Dart’s (1949, 1957) osteodontokeratic interpretationsof Australopithecus behavior followed, with the Man the Hun-ter paradigm emerging in the 1960s (Lee and Devore 1968).

As late as the 1980s, models of early hominin evolutionsuggested that Australopithecus, then the oldest hominin genusknown, engaged in hunting and food sharing (Carrier 1984;Lovejoy 1981). However, over the past 3 decades, analyses ofaustralopith dental and postcranial anatomy have typicallyportrayed them as semiarboreal and vegetarian (Conroy andPontzer 2012). Analyses of australopith locomotor anatomy,based largely on the A.L. 288 Australopithecus afarensis skel-eton (e.g., Stern and Susman 1983; see Ward 2002), have



Pontzer Ecological Energetics in Early Homo S347

suggested that they were too arboreal and too inefficient onthe ground—in essence, too apelike—to be plausible hunter-gatherers (but see Latimer 1991). Instead, recent models ofhominin evolution have proposed that the roots of modernhuman foraging ecology lie in the origin of the genus Homo,recalling the distinction made by Leakey and colleagues nearly50 years ago (Aiello and Wheeler 1995; Bramble and Lieber-man 2004; Leakey, Tobias, and Napier 1964; O’Connell,Hawkes, and Blurton Jones 1999; Shipman and Walker 1989).

Ecological Models for the Origin of Homo

Because of the inclusion of meat, underground storage organs(USOs), and cooking, the diet quality (i.e., kcal/g of food; seeLeonard and Robertson 1997) of modern human foragers isgreater than that of living apes, and most ecological modelsfor the origin of Homo suggest that a shift toward modernhuman diet quality was instrumental in shaping the evolutionof our genus (Aiello and Wheeler 1995; Bramble and Lie-berman 2004; Leonard and Robertson 1997; O’Connell,Hawkes, and Blurton Jones 1999; Shipman and Walker 1989).However, current models differ in which aspects of modernhuman foraging behavior they view as seminal.

Many ecological models for the evolution of the genusHomo emphasize the adoption of hunting and scavenging(Aiello and Wheeler 1995; Bramble and Lieberman 2004;Shipman and Walker 1989). In these hunting and scavengingmodels, meat and marrow provided valuable energetic andnutritional rewards for early Homo while simultaneously cre-ating selection pressure for locomotor and cognitive adap-tations to pursue prey or monopolize fresh carcasses. Thesemodels often envision the pursuit of prey and carcasses oc-curring under the midday sun of equatorial Africa, producingadditional selection pressure for effective thermoregulation(Bramble and Lieberman 2004; Ruff and Walker 1993; Walker1993). Proponents of these hunting and scavenging modelsview the tall, long-limbed proportions of Homo erectus (epit-omized by the KNM-WT 15000 skeleton) as evidence forimproved running and thermoregulatory ability and the in-crease in brain size and reduction in tooth size as evidencefor the inclusion of energy-rich meat/marrow into the diet(Aiello and Wheeler 1995; Bramble and Lieberman 2004;Shipman and Walker 1989).

Others have suggested that hunting and scavenging wereof negligible importance in the evolution of our genus andhave instead focused on the gathering of USOs, food sharing,and cooking (O’Connell, Hawkes, and Blurton Jones 1999;O’Connell et al. 2002; Wrangham et al. 1999). In these USOmodels for the origin of Homo, climatic changes leading toincreased aridity in the early Pliocene led to an increasedreliance on USOs, which are rich in calories and nutrientsand available even in dry seasons (O’Connell, Hawkes, andBlurton Jones 1999). The difficulty in locating and harvestingthese underground foods led to selection favoring the pro-visioning of children, which in turn resulted in longer female

life spans and increased body size (O’Connell, Hawkes, andBlurton Jones 1999). In a variant of this model, the adoptionof cooking plays a key role by unlocking otherwise indigestiblecarbohydrates in USOs and alters the social dynamics of earlyHomo by creating a home base where males and females wouldcook and share food (Wrangham 2009; Wrangham et al.1999).

Several studies have modeled the energetic consequencesof improved diet quality and larger body size of PleistoceneHomo (Aiello and Key 2002; Aiello and Wheeler 1995; Leon-ard and Robertson 1997; Steudel-Numbers 2006). Perhapsthe most influential has been the expensive tissue hypothesisof Aiello and Wheeler (1995), which proposes that the ad-dition of readily digestible and energy-rich meat into the dietof early Homo decreased their required gut size, freeing met-abolic energy to fuel a larger brain. Other studies have focusedprimarily on differences in daily energy expenditure (DEE;kcal/day) between Australopithecus and Homo, suggesting thatthe larger body size in H. erectus and an increase in rangingactivity associated with a higher-quality diet would have sub-stantially increased daily energy requirements for PleistoceneHomo (Aiello and Key 2002; Leonard and Robertson 1997;Steudel-Numbers 2006). This increase in DEE is generallyviewed as increasing the challenge of meeting food require-ments in Pleistocene Homo, although one analysis by Steudel-Numbers (2006) has suggested that this increase in DEEwould be partially offset by an improvement in locomotoreconomy due to increased hind-limb length in Homo.

Testing Ecological Models for theOrigin of Homo

The models briefly outlined above each provide a compellingreconstruction of the origin and evolution of our genus. De-spite their differences, these models share a common logicalframework and set of assumptions that can be tested usingcomparative data from living mammals as well as evidencefrom the hominin fossil record. Recent work in the homininfossil record, locomotor energetics, ranging ecology, and theevolution of mammalian metabolic strategies provide the op-portunity to reexamine these ecological models with new data.

In discussing the transition from Australopithecus to Homo,it is necessary to address the taxonomic placement of theoldest taxon in the genus Homo, Homo habilis. Wood andCollard (1999) have suggested that the primitive morphologyof H. habilis more accurately places it within Australopithecus.Similarly, many of the ecological models discussed above viewHomo erectus as the earliest species to exhibit the morpho-logical and behavioral traits they view as critical and definitiveof the genus. The ecological reconstruction and taxonomicplacement of H. habilis affect the timing but not the natureof the proposed transition from Australopithecus to Homo;instead, the assessment of H. habilis affects whether it isviewed as a latest exemplar of the Australopithecus grade or


the earliest exemplar of the Homo grade. This timing is dis-cussed below.

For the purposes of this analysis, I use a simplified taxo-nomic scheme that combines specimens assigned to H. habilisand Homo rudolfensis into a single taxon, Homo habilis sensulato, and similarly places specimens previously assigned toHomo ergaster into H. erectus. Given the difficulty in assigningpostcranial elements to species and the grade-level focus ofthese analyses, further distinction seems unwarranted. Fur-ther, I focus only on gracile Australopithecus taxa (Australo-pithecus afarensis, Australopithecus africanus, Australopithecusgarhi, and Australopithecus sediba) because the postcranialevidence for robust forms is poor and because current hy-potheses (both phylogenetic and ecological) view robustforms as an evolutionary side branch rather than ancestral toHomo (Conroy and Pontzer 2012).

Ecological models of past evolutionary events are inherentlycomplex. In order to organize and facilitate discussion, I haveoutlined five broad points on which current ecological modelsfor the origin of Homo rest. Some points are more vital tosome models than to others, but all draw on these five pointsto some degree. Points 1–3 concern the evolution of homininpostcranial anatomy and the biomechanical effects of thesechanges on locomotor performance. Points 4–5 concern pro-posed links between ecological change and evolutionary pres-sures on locomotor performance. I use this broad distinctionto organize the analyses and discussion below.

1. The postcranial anatomy of early Homo (at least by earlyHomo erectus) shows significant departures from the postcranialanatomy of earlier hominins. All of the ecological models out-lined above draw on fossil analyses indicating that the post-cranial anatomy of early Homo differs from that of Austral-opithecus. Derived aspects of modern human postcranialanatomy thought to be evident in early Homo include largerbody size (especially females), longer hind limbs, shorter fore-limbs, a narrow pelvis, shorter phalanges, and a stiff springlikeplantar arch. A long list of derived postcranial changes inearly Homo was given by Bramble and Lieberman (2004),who viewed these traits as critical for long-distance running.As mentioned above, the earliest evidence for larger femalebody size and some derived postcranial morphology is oftennoted in early Homo erectus rather than Homo habilis.

2. Postcranial changes evident in early Homo are productsof natural selection, not neutral processes such as drift. Theecological models discussed above view the ecological tran-sition from Australopithecus to Homo as a selection event. Thebehavioral and morphological traits adopted by early Homoare viewed as a response to changing climatic conditions andthe transition to a new foraging regime.

3. The derived postcranial characters evident in early Homoimprove locomotor performance (speed, economy, endurance).Hunting and scavenging models typically emphasize the im-portance of derived postcranial traits in early Homo for run-ning performance (Bramble and Lieberman 2004; Shipmanand Walker 1989). Others have suggested these traits are ad-

aptations for walking long distances to gather plant foods

(Isbell et al. 1998; O’Connell, Hawkes, and Blurton Jones

1999).

4. Evolutionary increases in diet quality result in increased

ranging activity. Ecological reconstructions of Plio-Pleistocene

hominins suggest an increase in ranging activity as a conse-

quence of changing foraging behavior in Homo (Anton, Leon-

ard, and Robertson 2002; Leonard and Robertson 1997; Steu-

del-Numbers 2006). In hunting and scavenging models, the

pursuit of prey or fresh carcasses requires increased running

speed (Shipman and Walker 1989) or endurance (Bramble

and Lieberman 2004). The USO models do not explicitly link

the change in diet quality to an increase in ranging but suggest

that increased ranging would be a consequence of the drier

climate and decreased food availability (O’Connell, Hawkes,

and Blurton Jones 1999).

5. Evolutionary increases in ranging activity lead to corre-

sponding increases in locomotor performance (speed, economy,

endurance) in mammals. Ecological models for the evolution

of our genus view changes in foraging behavior (point 4) as

resulting in a new set of selection pressures on locomotor

performance (point 2). Hunting and scavenging models are

typically very explicit in outlining the aspects of performance

most affected. In discussing the evolutionary consequences of

a transition to hunting and scavenging, Shipman and Walker

(1989) emphasized the importance of running speed. More

recently, Bramble and Lieberman (2004) argued that the crit-

ical aspect of performance for early Homo was running en-

durance, the ability to run at a moderate speed for long pe-

riods in order to exhaust prey or outpace other scavengers to

distant carcasses. The USO models for the evolution of Homo

suggest increased ranging would lead to improved economy

and endurance as well, although these models typically em-

phasize walking performance (Isbell et al. 1998; O’Connell,

Hawkes, and Blurton Jones 1999).

Postcranial Change in Homo (Points 1–3)

Ecological models for the evolution of our genus draw on

postcranial differences between Homo and Australopithecus.

Dozens of morphological features have been cited by previous

studies as distinguishing the locomotor anatomy of austral-

opiths from modern humans (see Bramble and Lieberman

2004; Stern 2000; Ward 2002). Here, I focus primarily on

morphological features that have been shown through ex-

perimental testing to significantly affect locomotor perfor-

mance. The discussion below is organized by the demon-

strated contribution of each trait to locomotor performance

or foraging ecology beginning with traits known to have the

greatest effect. The evolutionary forces shaping these changes

(point 2) are briefly discussed afterward.


Figure 1. Estimated body mass for fossil hominins: Australopi-thecus (triangles), Homo (filled circles), and modern humans andNeanderthals (open circles; from Ruff, Trinkaus, and Holliday1997; see table 1). Recent finds are labeled KSD-VP-1/1 Aus-tralopithecus afarensis specimen (K), Australopithecus sediba (S),three Dmanisi specimens (D), and the Gona pelvis (N). Estimatedages for individual australopith specimens have been adjustedslightly to improve symbol clarity.

Body Size

The USO models for the evolution of Homo draw on theevidence for increasing body size—particularly female bodysize—reported for Homo erectus (McHenry 1994). Increasedfemale body size is taken as evidence for delayed female lifehistory schedules (i.e., later age at maturity and longer lifespan) associated with the advent of provisioning by postre-productive females (O’Connell, Hawkes, and Blurton Jones1999). The cooking hypothesis also views increased femalebody size as evidence for increased diet quality and a changingsocial landscape that favored a decrease in sexual dimorphism(Wrangham et al. 1999).

Figure 1 shows estimated body masses for Plio-Pleistocenehominins. Body mass estimates were taken from the literatureor calculated from reported femoral head diameters (or, forthe MH-1 Australopithecus sediba tibia, from the tibiotalarsurface dimensions) using the intra-Homo least squares re-gression in McHenry (1992). Data and sources are shown intable 1 and in the extended version of this table (table A1 inCA� online supplement A).

As reported previously (McHenry 1992, 1994), there isgood evidence for an increase in body mass from Australo-pithecus to Homo. Mean body mass for the pooled sample ofHomo specimens (48.8 kg, , ) was 32%SD � 11.3 n p 15greater than in Australopithecus ( , ). Body36.8 � 7.4 n p 24masses for specimens identified as male in the Homo sample( , ) were 34% larger than those in Australo-56.4 � 7.8 n p 9pithecus ( , ), and masses for purported fe-42.2 � 4.9 n p 13males in the Homo sample ( , ) were 32%40.7 � 9.3 n p 7larger than those of Australopithecus ( , ).30.5 � 3.7 n p 11Given the inherent error in estimating mass and the smallsamples available, further parsing into individual species isstatistically unpalatable. Indeed, sample sizes become so min-iscule that the confidence intervals for mean male and meanfemale body mass overlap (see McHenry 1994). Recent finds,while consistent with the general conclusion of increasingbody size in Homo, advise caution in estimating the degreeof dimorphism in Homo habilis or early H. erectus. Smallspecimens from Gona and Dmanisi indicate that the rangeof body mass in early Pleistocene Homo may not have beensubstantially less than that of Australopithecus (fig. 1; see alsoAnton 2012; Anton and Snodgrass 2012).

As discussed in USO models for the origin of Homo, adultbody size is correlated with life history schedules (Charnovand Berrigan 1993; O’Connell, Hawkes, and Blurton Jones1999), and sexual dimorphism is correlated with mating andreproductive strategies in primates (Plavcan 2012; Wranghamet al. 1999). The DEE (kcal/day) also increases with body size(Leonard and Robertson 1997; Nagy, Girard, and Brown 1999;see also Anton and Snodgrass 2012). Body mass also has adirect effect on locomotor cost, discussed below, but notablyhas no apparent effect on running speed. Maximum runningspeed in mammals is independent of body mass for speciesgreater than 10 kg (Garland 1983b).

Limb Length

Increased relative hind-limb length in Homo has been citedas evidence for increased cursoriality, improving both walkingand (especially) running performance in our genus (Brambleand Lieberman 2004; Isbell et al. 1998; Jungers 1982; Ruffand Walker 1993). Indeed, limb length is one of four primarydeterminants of locomotor cost (kcal/m) in terrestrial ani-mals, the others being body mass, the effective mechanicaladvantage (EMA) of the limb joints, and the fascicle lengthsof limb muscles (Pontzer, Raichlen, and Sockol 2009). Themetabolic cost of walking and running derives from the vol-ume of muscle activated in each step to support body weight;consequently, larger animals spend more energy to walk andrun (Kram and Taylor 1990; see Pontzer, Raichlen, and Sockol2009). Limb length, EMA, and muscle length largely deter-mine the volume of muscle activated to support each gramof body mass. Animals with longer limbs, greater EMA (i.e.,more straight-legged postures), and shorter muscle fasciclesuse less energy per gram of body mass to walk and run(Pontzer 2007; Pontzer, Raichlen, and Sockol 2009; Roberts,Chen, and Taylor 1998). The effects of limb length and EMAare evident within our own species. Humans with longer hindlimbs use less energy to walk and run (DeJaeger, Willems,and Heglund 2001; Steudel-Numbers and Tilkens 2004; Steu-del-Numbers, Weaver, and Wall-Scheffler 2007), and our useof more flexed hind limbs while running decreases the EMAof our knee and hip joints such that the energy cost (kcal/m) of running is greater than the cost of walking (Bieweneret al. 2004; Pontzer, Raichlen, and Sockol 2009). Notably, thenumber of limbs used in locomotion (e.g., whether a species


Table 1. Estimated body masses and femur and tibia lengths for Plio-Pleistocene hominins

Species and specimen Mass (kg) Sex Femur (mm) Tibia (mm) Source

Ardipithecus ramidus:ARA-VP-6/500 51.0a F . . . 262.0 Lovejoy et al. 2009

Australopithecus anamensis:KNM KP 29285 51.0a M . . . . . . Leakey et al. 1995

Australopithecus afarensis:KSD-VP-1/1 45.4 M . . . 355.0 Haile-Selassie et al. 2010AL 827-1 45.6 M 368–382 . . . T. Holliday, personal communicationAL 333-3 50.1 M 373–391 . . . McHenry 1992; T. Holliday, personal communicationAL 333-4 41.3 M . . . . . . McHenry 1992AL 333-7 42.6 M . . . . . . McHenry 1992AL 333-w-56 40.2 M . . . . . . McHenry 1992AL 333x-26 48.2 M . . . . . . McHenry 1992AL 129a 27.1 F . . . . . . McHenry 1992AL 333-6 33.5 F . . . . . . McHenry 1992AL 288-1 28.0 F 281.0 241.0 McHenry 1992; Pontzer et al. 2010b

Australopithecus africanus:Sts 34 38.4 M . . . . . . McHenry 1992Stw 99 45.4 M 433.5 . . . McHenry 1992; T. Holliday, personal communicationStw 311 40.7 M . . . . . . McHenry 1992Stw 389 37.9 M . . . . . . McHenry 1992Stw 443 41.3 M . . . . . . McHenry 1992Sts 14, 34 30.3 F 276.0 . . . Steudel-Numbers and Tilkens 2004Stw 392 32.7 F . . . . . . McHenry 1992TM 1513 32.5 F . . . . . . McHenry 1992Stw 25 34.2 F . . . . . . McHenry 1992Stw 102 30.5 F . . . . . . McHenry 1992Stw 347 27.5 F . . . . . . McHenry 1992Stw 358 23.3 F . . . . . . McHenry 1992

Australopithecus sediba:MH 1 31.5 M . . . 262–269 T. Holliday, personal communicationMH 2 35.7 F . . . . . . Berger et al. 2010

Australopithecus garhi:BOU-VP-12/1 30–43a . . . 335.0 . . . Steudel-Numbers and Tilkens 2004

Homo habilis:KNM ER 3228 63.5 M . . . . . . Ruff, Trinkaus, and Holliday 1997KNM ER 1472 49.6 M 401.0 . . . Steudel-Numbers and Tilkens 2004KNM ER 1481 57.0 M 396.0 . . . Steudel-Numbers and Tilkens 2004OH-35 31.8 F . . . . . . McHenry 1992OH-8 31.0 F . . . . . . McHenry 1992OH-62 33.0a . . . 315.0 . . . McHenry 1991, 1992

Homo erectus:KNM WT 15000 51.0 M 429.0 380.0 Pontzer et al. 2010bDmanisi large adult 48.8 M 382.0 306.0 Pontzer et al. 2010bKNM ER 736 68.3 M . . . . . . Ruff, Trinkaus, and Holliday 1997KNM ER 1808 63.4 M 485.0 . . . Ruff, Trinkaus, and Holliday 1997Dmanisi subadult 49.4b M . . . . . . Pontzer et al. 2010bBSN49/P27 39.7 F . . . . . . Simpson et al. 2008Dmanisi small adult 40.2 F . . . . . . Pontzer et al. 2010bOH 34 51.0 F 432.0 . . . Steudel-Numbers and Tilkens 2004OH 28 54.0 F 456.0 . . . Steudel-Numbers and Tilkens 2004

a Not included in body mass comparisons of Australopithecus and Homo.b Estimated adult mass.

is bipedal or quadrupedal) has no effect on locomotor cost(Pontzer 2007; Pontzer, Raichlen, and Sockol 2009; Sockol,Raichlen, and Pontzer 2007; Taylor, Heglund, and Maloiy1982).

Longer limbs—both forelimb and hind limb—are alsoknown to improve an individual’s ability to dissipate heat andthermoregulate in hot environments by increasing the ratio

of surface area to body mass (see Tilkens et al. 2007 andreferences therein). Small increases in core body temperaturecan have catastrophic effects on the brain and other organs,making adaptations for heat dissipation critical in hot envi-ronments (Schmidt-Nielsen 1999). Overheating also curtailsranging activity, as animals must stop walking and runningif the heat produced from muscle activity threatens to raise


Figure 2. Estimated hind-limb length versus estimated body mass for fossil hominins. Shaded areas indicate �2 SDs of residualsfrom the human and African ape ordinary least squares regression trend lines. Error ranges (gray) represent �25 mm estimatedhind-limb length and �5% estimated body mass; see text for details. A color version of this figure is available in the online editionof Current Anthropology.

the body temperature. Thus, in addition to improving lo-comotor economy, longer hind limbs in early Homo wouldlikely improve endurance by preventing overheating (Brambleand Lieberman 2004; Ruff and Walker 1993).

In assessing the evidence for increased hind-limb length inHomo, one must account for its correlation with body mass.Figure 2 shows estimated hind-limb lengths (femur � tibia)plotted against estimated body mass for Plio-Pleistocene hom-inins. A sample of modern humans ( , including mea-n p 110surements of Pygmy skeletons generously providedn p 24by W. Jungers), chimpanzees ( ), and gorillas (n p 60 n p

) is shown for comparison (data from Pontzer et al. 2010a).22The non-Pygmy human sample is from the Hamman-Toddcollection at the Cleveland Museum of Natural History andconsists of adults who died in the Cleveland area in the 1900s.While detailed genealogical data are not available for theHamman-Todd sample, 60% ( ) of the individuals inn p 52this data set are identified as “black,” while 40% ( ) aren p 34identified as “white.” Individuals identified as “black” had agreater crural index (tibia/femur length: mean )0.85 � 0.03than “white” individuals ( ), suggesting the “black”0.82 � 0.03and “white” populations represent more equatorial (presum-ably African) and more northerly (presumably European)populations, respectively. However, while the difference inmean crural index was significant ( , t-test) it wasP ! 0.001relatively small, and the ranges for these populations (“black”:0.79–0.91, “white”: 0.76–0.90) largely overlap. Mean crural

index for the Pygmy sample ( , range 0.81–0.89)0.85 � 0.02was similar to the “black” sample ( , t-test).P p 0.83

Fossil dimensions were taken from the literature (exceptA.L. 333-3, A.L. 827-1, StW 99, and MH-1, which were gen-erously provided by T. Holliday). Data and sources are givenin table 1. Where only tibia or femur lengths are known (thecase for most early hominins including early Homo), hind-limb length was estimated using a crural index (tibia/femurlength) 0.85. This value (0.85) is equal to the mean cruralindex for fossil Homo sapiens in the sample ( ,0.85 � 0.03

) and near the means for the modern H. sapiens samplen p 12( , ) and combined Pan and Gorilla sample0.84 � 0.03 n p 110( , ). To account for potential variation in0.84 � 0.02 n p 82crural index among early hominins, an error range of �25mm was plotted with the estimated hind-limb length, equiv-alent to a range of crural indexes from 0.80 to 0.90 for thesefossils; this range is equivalent to that of the “black” andPygmy samples, which is appropriate because the specimensfor which femur or tibia length was estimated are all African(table 1). For incomplete specimens, when a maximum andminimum estimated length were available, the error rangeplotted in figure 2 was further expanded to reflect this rangeof uncertainty.

Figure 2 includes the Ardipithecus ramidus skeleton (ARA-VP-6/500; Lovejoy et al. 2009), the recent Australopithecusafarensis skeleton from Ethiopia (KSD-VP-1/1; Haile-Selassieet al. 2010), Australopithecus garhi (BOU-VP-12/1; Asfaw et


al. 1999), A. sediba (MH-1; Berger et al. 2010), and the larger

adult H. erectus specimen from Dmanisi (Lordkipanidze et

al. 2007). The range of body mass estimates for the A. garhi

specimen (30–43 kg) was based on the male and female means

for Australopithecus (see above). As with hind-limb length

estimates, a range of �5% of estimated body mass was plotted

for early hominin specimens in figure 2 to reflect some degree

of uncertainty in mass.

Recent finds strongly challenge previous reconstructions of

Australopithecus as having short hind limbs. Previous assess-

ments of australopith hind-limb length focused on A.L. 288

and Sts 14, the only Australopithecus specimens for which

reliable estimates of limb length and body mass were obtain-

able (Jungers 1982; McHenry 1991, 1992; Pontzer et al. 2010a;

Steudel-Numbers 2006). These specimens have shorter hind

limbs than expected for a human of similar mass (fig. 2; see

Jungers 1982). However, hind-limb lengths for these speci-

mens are ambiguous because the range of limb lengths seen

in modern humans overlaps considerably with that of African

apes at body sizes below 30 kg (fig. 2). Inclusion of new

specimens suggests hind-limb length in Australopithecus is in

fact similar to modern humans. Hind-limb lengths for the

large-bodied A. afarensis specimen (KSD-VP-1/1), A. sediba,

and A. garhi all fall within modern human range, with KSD-

VP-1/1 clearly distinguished from the African apes (fig. 2).

These results are robust to error in hind-limb length and body

mass estimation; australopith hind-limb lengths relative to

body mass remain in the modern human range even when

hind-limb length and body mass estimates are varied sub-

stantially (fig. 2).

Hind-limb length within the genus Homo warrants dis-

cussion. Contrary to reconstructions of H. habilis as having

short hind limbs, specimens of H. habilis—including OH

62—fall clearly within the range of modern humans. The

hind-limb proportions (relative to body mass) of H. habilis

combined with evidence for a springlike plantar arch in this

species (see Harcourt-Smith and Aiello 2004; and below) sug-

gest the bipedal capabilities of H. habilis would have been

similar to those of H. erectus. Late Pleistocene hominins ex-

hibit a substantial degree of variation in relative hind-limb

length (fig. 2). Several Neanderthal specimens, including all

of the European specimens and two Middle Eastern speci-

mens, are more than 2 SDs below the range for modern

human hind-limb length. While their relatively short tibiae

contribute to a shorter hind limb, it is their total hind-limb

length, not their crural index, that places Neanderthals outside

the modern human range: the range of crural indexes among

the Neanderthal sample (0.76–0.81) falls within the observed

range in the modern human sample, but their relative hind-

limb lengths do not. As discussed by Weaver and Steudel-

Numbers (2005), the short hind limbs of Neanderthals may

have increased their daily foraging costs.

Plantar Arch and Achilles Tendon

During running, the limbs of terrestrial animals act likesprings. With each step the ligaments, tendons, and musclesstore energy as the limb flexes under body weight and thenrelease this strain energy to help propel the body into thenext step (Biewener et al. 2004). In modern humans, muchof this springlike work is performed by the plantar arch andthe Achilles tendon (Alexander 1991). Together, the plantararch and Achilles tendon convert over 50% of the energystored as strain into kinetic energy, reducing the amount ofmuscular work and metabolic energy needed to power ourstride (Alexander 1991). A rigid midfoot also appears to im-prove walking economy by increasing the efficiency withwhich the foot pushes off the ground at toe off: when therigidity of the foot is effectively compromised by walking overa soft surface, the energy cost of walking increases (Lejeune,Willems, and Heglund 1998). Bramble and Lieberman (2004)have argued that a springlike Achilles tendon and plantar archevolved early in the genus Homo as adaptations for endurancerunning.

The evolutionary history of these traits is difficult to assess.The presence of an elongated, humanlike Achilles tendon can-not be discerned from fossil remains using current techniques.However, it is notable that gibbons have an elongate Achillestendon (Payne et al. 2006), indicating that its presence neednot coincide with habitual endurance running. The plantararch is more amenable to measurement in the fossil record,although associated foot bones are rare. Previous analyses ofA. afarensis and A. africanus foot morphology have suggestedthat these species lacked the derived tarsal morphology, par-ticularly of the navicular and cuboid, associated with thespringlike modern human plantar arch, and some have evenargued that these species retained an opposable hallux (Har-

court-Smith and Aiello 2004). However, more recent workhas provided evidence that australopiths may have had a plan-tar arch. The hominin footprints at Laetoli, usually attributedto A. afarensis, suggest a foot that is functionally similar inmany ways to that of modern humans, with a stiff midfoot,adducted hallux, and at least some arching (Tuttle, Webb, andBaksh 1991). Ward, Kimbel, and Johanson (2011) have re-cently described a fourth metatarsal from Hadar, attributedto A. afarensis, that has a humanlike degree of torsion thatalong with its proximal articular morphology suggests thepresence of an arch. Evidence of an arch appears less ambig-uous in early Homo. Specimens of H. habilis (OH-8) and H.erectus (Dmanisi) show strong evidence for the presence of aplantar arch in these taxa (Harcourt-Smith and Aiello 2004;Pontzer, Raichlen, and Sockol 2009). Thus, while it remainsplausible that the springlike plantar arch typical of modernhumans arose with the genus Homo, recent morphologicalevidence as well as mechanical analyses of the Laetoli trackwaysuggest this morphology may extend back to Australopithecus.


Stabilization and Stress Reduction

Numerous changes in the size and morphology of the limbjoints and inferred changes in the size of the muscles thatstabilize the trunk have been argued to reflect an increase interrestrial locomotion, particularly running, in Homo (seeBramble and Lieberman 2004; Stern 2000; Ward 2002). Forexample, Bramble and Lieberman (2004) have suggested thatthe larger hind-limb joints and inferred increases in the sizeof the erector spinae and gluteus maximus muscles in earlyHomo serve to improve trunk stability and reduce mechanicalstress on the joints during long-distance running. While thesetraits may in fact distinguish Australopithecus from Homo,their effects on locomotor performance are difficult to assess;none have been tested experimentally. Further, in light of theevidence for larger body size in early Homo (fig. 1), it ispossible that some of these changes are allometric effects ofincreased mass.

Bramble and Lieberman (2004) suggest that these traitsimprove endurance, implicitly defined as the ability to run ata moderate to high speed for a long period of time (up to 1hour or more), by improving stability of the trunk and re-ducing mechanical stress and fatigue in the hind-limb joints.However, the only morphological traits known to affect run-ning endurance across mammals are the mass and mito-chondrial density of the limb muscles (Weibel et al. 2004);the capacity of the liver and limb muscles to store glycogenmay also constrain endurance ability (see Rapoport 2010).Hind-limb muscle mass is greater in modern humans thanin living apes (Payne et al. 2006), as expected given our bipedalposture, but it is unclear whether australopiths, which weresimilarly dependent on their hind limbs for weight support,had similar hind-limb muscle mass. Current techniques arenot capable of assessing mitochondrial density in extinct hom-inins. More work is needed to assess hind-limb muscle prop-erties in Plio-Pleistocene hominins and to determine the effectof other implicated anatomical variables on locomotor per-formance.

Interpreting Postcranial Anatomy in Early Homo

The evidence outlined above challenges the hypothesis thatwalking and running performance improved substantiallywith the origin of the genus Homo. The similarity in hind-limb length between Australopithecus and Homo suggests thatthe derived longer hind limb typical of modern humans wasalready present in australopiths nearly 4 million years ago.Similarly, while australopith foot anatomy remains a subjectof debate, there is evidence of a rigid, possibly springlike,plantar arch in both the skeletal anatomy of the A. afarensisfoot (Ward, Kimbel, and Johanson 2011) and the Laetolitrackway (Tuttle, Webb, and Baksh 1991). Analyses of theAustralopithecus pelvis and the footprints at Laetoli as well ascomputer simulations of australopith gait have previously in-dicated that Australopithecus most likely used a relatively

straight-legged walking gait similar to modern humans(Pontzer, Raichlen, and Sockol 2009; Raichlen et al. 2010;Sockol, Raichlen, and Pontzer 2007; Wang et al. 2004). Thus,with the possible exception of the plantar arch, the locomotoranatomy of Australopithecus, at least as it pertains to walkingand running performance, appears to have been functionallyequivalent to that of early Homo.

Perhaps the clearest signal for postcranial change in earlyHomo is an increase in body mass of roughly 33% comparedwith australopiths (fig. 1). However, while an increase in bodymass may signal an ecological change (see below), experi-mental and comparative evidence suggest increased size wouldnot have improved walking and running performance. Largeranimals use more energy to walk and run (Taylor, Heglund,and Maloiy 1982), and the proportion of DEE spent on travelalso tends to increase with body size (Garland 1983a). Asnoted above, maximum running speed among mammals isindependent of body mass above 10 kg (Garland 1983b), al-though it should be noted that among modern human ath-letes, sprinters are generally taller and heavier than distancerunners (Weyand and Davis 2005). Indeed, larger body sizein early Homo, absent an increase in relative hind-limb length,may be particularly difficult to reconcile with endurance run-ning models. Among modern human athletes, endurance run-ning appears to favor shorter, lighter individuals, while sprint-ers are heavier and taller (Weyand and Davis 2005). Further,because larger body size tends to reduce the ratio of surfacearea to body mass (Ruff 1994), increased body size in earlyHomo would likely diminish its ability to shed heat, contrabehavioral reconstructions suggesting intense activity in theheat of the day.

Evidence for increased body mass does provide some sup-port for hypotheses suggesting a change in ecology in earlyHomo, but this support is tempered by three observationsfrom the fossil and comparative record. First, with the inclu-sion of newer specimens of early Homo (fig. 1), there is noevidence that female size increases more than male size, whichindicates that an emphasis on increasing female size in Homomay be unwarranted. Second, while larger species tend tohave slower life histories (Charnov and Berrigan 1993), thereis a considerable degree of variation in this relationship, andanalyses of hominin tooth formation suggest that growth ratesand thus life history schedules in early Pleistocene Homo weresimilar to those of Australopithecus (Dean and Smith 2009;Dean et al. 2001), though perhaps somewhat slower in H.erectus (see Schwartz 2012). Third, while larger species tendto have larger energy budgets, there is also a considerabledegree of variation in this relationship. Measurements of DEEacross a broad range of taxa indicate a sixfold range of var-iation in DEE even after accounting for the effects of bodymass and phylogeny (Nagy, Girard, and Brown 1999). Thislast point is discussed below.

These analyses have omitted discussion of forelimb lengthin Australopithecus and Homo because arm length has noknown effect on walking or running performance (other than


its effect on body mass). However, it may be that the primarychange in limb length with the genus Homo is shorter armsrather than longer legs (Holliday 2012). Analyses of the in-termembral index of Australopithecus suggest substantiallylonger arms than in Homo, although whether this is a prim-itive retention or an adaptation to climbing remains a matterof debate (Ward 2002). Homo habilis has also been argued tohave long arms, based primarily on the OH 62 specimen,although this reconstruction is debated (Haeusler and Mc-Henry 2004). Thus, rather than an increase in walking andrunning performance, the origin of Homo, or perhaps H.erectus, may mark a decrease in arboreal ability. This scenariois consistent with shorter, straighter phalanges evident inHomo (Aiello and Dean 1990); with the pattern of limb ro-busticity in African H. erectus (but not OH-62; Ruff 2009);and perhaps with the larger body size evident in early Homo,but it warrants further scrutiny of the hominin forelimb. Hereagain the small body size of A.L. 288, the primary specimenused to calculate australopith intermembral index, may con-flate the difference in proportions with a difference in size.Analysis of the large-bodied KSD-VP-1/1 specimen suggeststhat forelimb and hind-limb proportions of A. afarensis maybe more similar to modern humans than previously thought(Haile-Selassie et al. 2010).

Given the evidence for postcranial change, or lack thereof,in the origin of Homo, it is important to address whetherreported changes in locomotor anatomy may reflect neutralevolutionary processes rather than selection (point 2). Recentcomparisons of the human and Neanderthal cranium providean important reminder that differences in hominin skeletalanatomy often thought to reflect natural selection may in factreflect neutral processes (Weaver 2009). With this in mind,it is notable that the majority of postcranial traits distinguish-ing Australopithecus from Homo have not been tested withregard to their effect on locomotor performance. Traits withknown effects on walking and running performance, namelyhind-limb length, appear to have remained stable over thepast 4 million years of hominin evolution, perhaps changingin the transition from Ardipithecus to Australopithecus (fig.2). Similarly, the reorganization of the hominin pelvis forbipedalism, a likely product of natural selection (Grabowski,Polk, and Roseman 2011), appears to predate the genus Aus-tralopithecus and the transition to Homo. Thus, it is difficultto reject the hypothesis that the hominin postcranial skeletonhas largely been under stabilizing selection since the middlePliocene, with neutral evolutionary forces leading to smallchanges in joint morphology over time. The strongest can-didate for postcranial change under natural selection in theevolution of Homo may be body mass (fig. 1).

Foraging Ecology and the Evolution ofLocomotor Performance (Points 4 and 5)

As discussed above, current ecological models for the evo-lution of the genus Homo envision an increase in diet quality

(Bramble and Lieberman 2004; Leonard and Robertson 1997;

O’Connell, Hawkes, and Blurton Jones 1999; Shipman and

Walker 1989; Wrangham et al. 1999). In these models, par-

ticularly those emphasizing hunting and scavenging, higher-

quality diet in early Homo provided the benefit of increased

energy availability but required a substantial increase in rang-

ing activity. The USO models generally frame this increase in

ranging activity as an increase in the daily distance traveled

(O’Connell, Hawkes, and Blurton Jones 1999; see also Isbell

et al. 1998), while hunting and scavenging models emphasize

not only increased travel distance but also the need to travel

quickly to run down prey or monopolize carcasses (Bramble

and Lieberman 2004; Shipman and Walker 1989). Increased

ranging activity is in turn thought to increase DEE and to

impose a new set of selection pressures on hominin locomotor

performance (Bramble and Lieberman 2004; Leonard and

Robertson 1997; Steudel-Numbers 2006).

Comparative studies of living mammals suggest that in-

creased diet quality leads to increased daily travel distance.

Carbone et al. (2005), in a phylogenetically controlled mul-

tivariate analysis of daily travel distance in 200 mammal spe-

cies, showed that daily travel distance increases significantly

with diet quality. While variation in ranging distance is con-

siderable, faunivores travel farther, on average, than similarly

sized frugivores, which in turn travel farther than herbivores

(Carbone et al. 2005); on average, carnivores travel four times

farther each day than similarly sized herbivores (Garland

1983b). An analysis by Anton and colleagues showed that

home range size among primates increases with both body

mass and diet quality (Anton, Leonard, and Robertson 2002).

Applying results from extant primates to fossil hominins, they

estimated that home ranges for Homo erectus would have been

10 times larger than those of Australopithecus (Anton, Leon-

ard, and Robertson 2002).

Yet increased travel distance does not appear to result in

improved locomotor economy, speed, or endurance. Com-

parative studies of locomotor cost indicate that the economy

of carnivores is no different than that of artiodactyls or other

herbivores (Taylor, Heglund, and Maloiy 1982). Further, limb

length in “cursorial” species is no different than that of other

mammals and is unrelated to daily travel distance (Harris and

Steudel 1997; Steudel and Beattie 1993). Similarly, maximum

running speed does not differ between carnivores and ar-

tiodactyls, even among predator-prey pairs (Garland 1983a;

Shipman and Walker 1989). Maximum aerobic power, a re-

liable measure of endurance, has not been investigated to

determine whether it correlates with diet quality or ranging

ecology in mammals, but several herbivores, including An-

tilocapra and Equus, have relatively high maximum aerobic

power for their body size (Weibel et al. 2004), suggesting that

diet quality and aerobic performance are not strongly posi-

tively correlated.


Limitations of Current Ecological Models

The lack of correspondence between ranging activity and lo-comotor performance, particularly locomotor economy, runscounter to the expectations of most ecological models for theevolution of Homo. However, these ecological models rest ontwo critical assumptions that are not well supported by theavailable comparative evidence. The first is that selection forimproved locomotor economy is a relatively strong forceshaping locomotor anatomy, particularly in cursorial species.In fact, analysis of foraging efficiencies among modern mam-mals suggests that selection pressure for improved locomotoreconomy is probably quite low relative to other selection pres-sures in most species because they already obtain a high rateof energy return while foraging (Pontzer 2012). In an analysisof foraging return rates for 228 mammal species, the medianestimated foraging efficiency was , or 56 calories of food56 : 1energy obtained for every 1 calorie spent on locomotion(Pontzer 2012). With such high foraging efficiencies, evenlarge reductions in locomotor cost have relatively small effectson net energy intake (i.e., gross energy intake minus travelcost). For an organism obtaining a foraging efficiency of

, a 20% reduction in locomotor cost, requiring sub-40 : 1stantial anatomical change, would improve net energy intakeby only 0.5% (Pontzer 2012). Even when foraging efficiencyis halved, to , a 20% reduction in locomotor cost would20 : 1yield only a 1% gain in net intake (Pontzer 2012). Thus, giventhe myriad selection pressures acting on the hominin post-cranial skeleton, it is likely that selection for improved econ-omy remained relatively weak throughout the Plio-Pleistoceneregardless of any changes in foraging ecology.

A second problematic assumption of many ecological mod-els is that they generally view DEE in hominins and otherlineages in a “zero-sum game” framework in which any in-crease in the energy spent on one activity must be matchedby a corresponding decrease in the energy spent on another(e.g., Aiello and Wheeler 1995; Charnov and Berrigan 1993;Isler and van Schaik 2009). Zero-sum frameworks have hadsuccess in explaining large-scale trends in life history (e.g.,Charnov and Berrigan 1993) and brain size among primatesand other animals (Fish and Lockwood 2003; Isler and vanSchaik 2006, 2009), although some studies have found littleevidence for energetic trade-offs (Barrickman and Lin 2010;Bordes, Morand, and Krasnov 2011; Jones and MacLarnon2004). In the context of foraging ecology, zero-sum gamemodels predict that longer daily travel distances increase theportion of the daily energy budget spent on travel, whichdetracts from the energy spent on other activities and in turnleads to selection to improve locomotor economy and restoreenergy expenditure to those activities.

While this logic is compelling, it is not supported by dataon locomotor cost or anatomy; as noted above, daily traveldistance is not correlated with locomotor efficiency amongmammals. In fact, humans themselves do not appear to fitthe predictions of a zero-sum game framework, casting some

doubt on the applicability of this approach to hominin evo-lution. Despite having the largest, most metabolically expen-sive brains and longest daily travel distances of any primatespecies, human foragers outpace all other hominoids in termsof reproductive output and maximum life span (Hawkes etal. 1998; Isler and van Schaik 2012).

Recent work on mammalian metabolic strategies offers analternative to zero-sum game approaches. Nagy, Girard, andBrown (1999), in reviewing measurements of DEE in wildpopulations of 79 mammal species, noted that there is a six-fold range of variation in DEE among species even after con-trolling for body mass and phylogeny. This variation in DEEappears to reflect evolved strategies for energy throughput(McNab 1986; Pontzer and Kamilar 2009; Sibly and Brown2007). In habitats where food is abundant, species may adopthigh-throughput (i.e., high DEE) strategies that increase foodrequirements but also provide more energy for reproduction.Alternatively, in habitats where food availability is highly var-iable or where foraging increases the risk of predation, speciesmay evolve a low-throughput strategy that reduces the energyrequirements even at the cost of lower reproductive output.In a test of this hypothesis, Pontzer and Kamilar (2009) con-ducted a phylogenetically controlled multivariate study ofdaily travel distance and reproductive output in a sample of110 mammal species. While there was considerable variationamong species, daily travel distance was found to be positivelyassociated with lifetime reproductive output among mam-mals, suggesting that species that travel farther generally doso as part of a high-throughput strategy of increased DEEand reproductive investment (Pontzer and Kamilar 2009).Rather than a zero-sum game framework in which DEE isrelatively constant, these results suggest that species’ metabolicstrategies are labile over evolutionary time, with DEE shrink-ing or expanding in response to environmental pressures.

Ecological Implications of Increasing Diet Quality and DEE

A dynamic view of mammalian metabolic strategies focusingon throughput rather than efficiency and trade-offs changesthe way one interprets the evidence for increased diet qualityand ranging activity in Homo. Rather than presenting an eco-logical or energetic cost, increased travel distance and bodymass in early Homo may reflect an improved ability to procurefood energy and a subsequent expansion of the energy budget.Indeed, comparisons with the limited data available for apeDEE suggests humans may have evolved larger energy budgetsat some point in our lineage (Pontzer et al. 2010b), and asdiscussed above, modeling studies suggest this may have oc-curred with Homo (Aiello and Key 2002; Leonard and Rob-ertson 1997; Steudel-Numbers 2006). Expansion of the dailyenergy budget would make more energy available for braingrowth, reproduction, and other investments without nec-essarily resulting in increased selection for locomotor per-formance.

Greater DEE in early Homo would suggest an increase in


food availability either through an increase in abundance ora decrease in variability (McNab 1986; Pontzer and Kamilar2009; Sibly and Brown 2007). Given the evidence for im-proved diet quality, particularly the inclusion of meat, anincrease in abundance is unlikely because higher-quality foodsare generally less abundant (Carbone et al. 2005), requiringlarger home ranges (Anton, Leonard, and Robertson 2002).Instead, the combination of increased DEE and higher dietquality suggests that early Homo evolved strategies for de-creasing variance in food intake. Decreased variability mightbe achieved in a number of ways, but one intriguing possibilityis the advent of food sharing, which would reduce day-to-day variance in food availability. The USO and cooking mod-els (O’Connell, Hawkes, and Blurton Jones 1999; Wranghamet al. 1999) and hunting and scavenging models proposingexploitation of large game (e.g., Bramble and Lieberman 2004;Bunn and Pickering 2010; Shipman and Walker 1989) allimplicate food sharing as a key derived ecological feature ofearly Homo. Increasing adiposity, which is evident in modernhumans, would also serve to buffer variability in food avail-ability by providing energy reserves during periods of foodshortage (see Wells and Stock 2007). If the increase in bodysize and diet quality in early Homo is read as evidence forincreased energy throughput, it may indicate that provisioningand food sharing, and perhaps increased body fat, were criticalearly adaptations in the evolution of our genus.

Summary and Conclusion

Recent fossil discoveries provide a new perspective on eco-logical models for the evolution of our genus. Comparisonsof postcranial morphology suggest that adaptations for im-proved walking and running performance predate the originof Homo. Indeed, locomotor performance in Australopithecusmay have been equivalent to that of early Homo, includingHomo erectus. The strongest case for postcranial change inearly Homo is an increase in body mass, but recent findschallenge previous reconstructions of decreasing dimorphismin H. erectus, suggesting instead that variation in body sizeremained substantial throughout the early Pleistocene. Theevidence for increased body size is consistent with models ofecological change in early Homo, but the relative stasis inlocomotor morphology runs counter to models suggesting amarked change in ranging behavior and locomotor perfor-mance between Australopithecus and Homo.

The increase in body size evident in early Homo suggestsan increase in DEE as discussed by previous studies of energyexpenditure in fossil hominins (Aiello and Key 2002; Leonardand Robertson 1997; Steudel-Numbers 2006). However, farfrom increasing the challenge of finding sufficient food, com-parative studies of living mammals suggest an expansion ofthe daily energy budget would likely reflect an improved abil-ity to obtain food energy reliably and an increase in repro-ductive investment. This view is consistent with models em-phasizing the importance of provisioning in early Homo,

which decrease variance in food availability. A cooperativeforaging strategy would have pervasive effects on the socialand nutritional ecology of early Homo (O’Connell, Hawkes,and Blurton Jones 1999; Wrangham et al. 1999). Food sharingcould also mitigate the ecological risk of seeking high-value,high-risk foods such as meat, and indeed the earliest con-firmed evidence of butchery is associated with Homo habilis.Future efforts to reconstruct the evolution of our genusshould seek to examine evidence for food sharing in the earlyPleistocene.

Acknowledgments

I thank Leslie Aiello and Susan Anton for inviting me toparticipate in this symposium, and I thank them and the othersymposium attendees as well as anonymous reviewers forcomments, conversations, and insights that improved thisstudy considerably. William Jungers generously shared mea-surements of Pygmy femoral dimensions, and Trent Hollidaygenerously shared measurements of hominin fossils.

References CitedAiello, Leslie, and Christopher Dean. 1990. An introduction to human evolu-

tionary anatomy. London: Academic Press.Aiello, Leslie C., and Catherine Key. 2002. Energetic consequences of being



Alexander, R. McNeill. 1991. Energy-saving mechanisms in walking and run-ning. Journal of Experimental Biology 160:55–69.


Anton, Susan C., William R. Leonard, and Marcia L. Robertson. 2002. Anecomorphological model of the initial hominid dispersal from Africa. Jour-nal of Human Evolution 43:773–785.


Asfaw, Berhane, Tim D. White, C. Owen Lovejoy, Bruce M. Latimer, ScottW. Simpson, and Gen Suwa. 1999. Australopithecus garhi: a new species ofearly hominid from Ethiopia. Science 284:629–635.

Barrickman, Nancy L., and Maggie J. Lin. 2010. Encephalization, expensivetissues, and energetics: an examination of the relative costs of brain size instrepsirrhines. American Journal of Physical Anthropology 143:579–590.


Biewener, Andrew A., Claire T. Farley, Thomas J. Roberts, and Marco Temaner.2004. Muscle mechanical advantage of human walking and running: im-plications of energy cost. Journal of Applied Physiology 97:2266–2274.

Bordes, Frederic, Serge Morand, and Boris R. Krasnov. 2011. Does investmentinto “expensive” tissue compromise anti-parasitic defence? testes size, brainsize and parasite diversity in rodent hosts. Oecologia 165:7–16.


Bunn, Henry T., and Travis R. Pickering. 2010. Bovid mortality profiles inpaleoecological context falsify hypotheses of endurance running-huntingand passive scavenging by early Pleistocene hominins. Quaternary Research74:395–404.

Carbone, Chris, Guy Cowlishaw, Nick J. B. Isaac, and Marcus Rowcliffe. 2005.How far do animals go? determinants of day range in mammals. AmericanNaturalist 165:290–297.


Carrier, David R. 1984. The energetic paradox of human running and hominidevolution. Current Anthropology 25:483–495.

Charnov, Eric L., and David Berrigan. 1993. Why do female primates havesuch long lifespans and so few babies? or life in the slow lane. EvolutionaryAnthropology 1:191–194.

Conroy, Glenn C., and Herman Pontzer. 2012. Reconstructing human origins.3rd edition. New York: Norton.

Dart, Raymond A. 1949. The predatory implemental technique of the aus-tralopithecines. American Journal of Physical Anthropology 7:1–16.

———. 1957. The osteodontokeratic culture of Australopithecus prometheus.Memoir (Transvaal Museum) 10. Pretoria: Transvaal Museum.

Darwin, Charles D. 1871. The descent of man, and selection in relation to sex.London: J. Murray.

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: originsand evolution of the genus Homo. Fred E. Grine, ed. Pp. 101–120. NewYork: Springer.

DeJaeger, Dominique, Patrick A. Willems, and Norman C. Heglund. 2001.The energy cost of walking in children. European Journal of Physiology 441:538–543.

Fish, Jennifer L., and Charles A. Lockwood. 2003. Dietary constraints onencephalization in primates. American Journal of Physical Anthropology 120:171–181.

Garland, Theodore, Jr. 1983a. The relation between maximal running speedand body mass in terrestrial mammals. Journal of Zoology (London) 199:157–170.

———. 1983b. Scaling the ecological cost of transport to body mass in ter-restrial mammals. American Naturalist 121:571–587.

Grabowski, Mark W., John D. Polk, and Charles C. Roseman. 2011. Divergentpatterns of integration and reduced constraint in the human hip bone andthe origins of bipedalism. Evolution 65:1336–1356.



Harcourt-Smith, William E. H., and Leslie C. Aiello. 2004. Fossils, feet andthe evolution of human bipedal locomotion. Journal of Anatomy 204:403–416.

Harris, Michelle, and Karen Steudel. 1997. The selection of hind limb lengthin the Carnivora: the influence of daily movement distance, home rangearea, prey size, latitude, and prey capture method. Journal of Zoology (Lon-don) 241:381–408.

Hawkes, Kristen, James F. O’Connell, Nicholas G. Blurton Jones, Eric L. Char-nov, and Helen Alvarez. 1998. Grandmothering, menopause, and the evo-lution of human life histories. Proceedings of the National Academy of Sciencesof the USA 95:1336–1339.


Isbell, Lynne A., Jill D. Pruetz, Michelle Lewis, and Truman P. Young. 1998.Locomotor activity differences between sympatric patas monkeys (Erythro-cebus patas) and vervet monkeys (Cercopithecus aethiops): implications forthe evolution of long hindlimb length in Homo. American Journal of PhysicalAnthropology 105:199–207.

Isler, Karin, and Carel P. van Schaik. 2006. Costs of encephalization: the energytrade-off hypothesis tested on birds. Journal of Human Evolution 51:228–243.

———. 2009. The expensive brain: a framework for explaining evolutionarychanges in brain size. Journal of Human Evolution 57:392–400.


Jones, Kate E., and Ann M. MacLarnon. 2004. Affording larger brains: testinghypotheses of mammalian brain evolution on bats. American Naturalist164:E20–E31, doi:10.1086/421334.

Jungers, William L. 1982. Lucy’s limbs: skeletal allometry and locomotion inAustralopithecus afarensis (A.L. 288-1). Nature 297:676–678.

Kram, Rodger, and C. Richard Taylor. 1990. Energetics of running: a newperspective. Nature 346:265–267.

Latimer, Bruce. 1991. Locomotor adaptations in Australopithecus afarensis: theissue of arboreality. In Origine(s) de la bipedie chez les Hominides. YvesCoppens and Brigitte Senut, eds. Pp. 169–176. Paris: CRNS.

Leakey, Louis S. B., Philip V. Tobias, and John R. Napier. 1964. A new speciesof genus Homo from Olduvai Gorge. Nature 202:7–9.

Leakey, Meave G., Craig S. Feibel, Ian MacDougall, and Alan Walker. 1995.New four-million-year-old hominid species from Kanapoi and Allia Bay,Kenya. Nature 376:565–571.

Lee, Richard B., and Irven Devore. 1968. Man the hunter. Piscataway, NJ:Aldine Transaction.

Lejeune, Thierry M., Patrick A. Willems, and Norman C. Heglund. 1998.Mechanics and energetics of human locomotion on sand. Journal of Ex-perimental Biology 201:2071–2080.



Lovejoy, C. Owen. 1981. The origin of man. Science 211:341–350.Lovejoy, C. Owen, Gen Suwa, Scott W. Simpson, Jay H. Matternes, and Tim

D. White. 2009. The great divides: Ardipithecus reveals the postcrania ofour last common ancestor with African apes. Science 326:73.

Marlowe, Frank W. 2005. Hunter-gatherers and human evolution. Evolution-ary Anthropology 14:54–67.

McHenry, Henry M. 1991. Femoral lengths and stature in Plio-Pleistocenehominids. American Journal of Physical Anthropology 85:149–158.



McNab, Brian K. 1986. The influence of food habits on the energetics ofeutherian mammals. Ecology Monographs 56:1–19.

Nagy, Kenneth A., I. A. Girard, and Tracey K. Brown. 1999. Energetics offree-ranging mammals, reptiles, and birds. Annual Review of Nutrition 19:247–277.

O’Connell, James F., Kristen Hawkes, and Nicholas G. Blurton Jones. 1999.Grandmothering and the evolution of Homo erectus. Journal of HumanEvolution 36:461–485.

O’Connell, James F., Kristen Hawkes, Karen D. Lupo, and Nicholas G.Blurton-Jones. 2002. Male strategies and Plio-Pleistocene archaeology. Jour-nal of Human Evolution 43:831–872.

Payne, Rachel C., Robin H. Crompton, Karin Isler, Russell Savage, Evie E.Vereecke, Michael M. Gunther, Susannah K. Thorpe, and Kristiaan D’Aout.2006. Morphological analysis of the hindlimb in apes and humans. 1. Musclearchitecture. Journal of Anatomy 208:709–724.

Plavcan, J. Michael. 2012. Body size, size variation, and sexual size dimor-phism. Current Anthropology 53(suppl. 6):S409–S423.

Pontzer, Herman. 2007. Limb length and the scaling of locomotor cost interrestrial animals. Journal of Experimental Biology 210:1752–1761.

———. 2012. Relating ranging ecology, limb length, and locomotor economyin terrestrial animals. Journal of Theoretical Biology 296:6–12.

Pontzer, Herman, and Jason M. Kamilar. 2009. Great ranging associated withgreater reproductive investment in mammals. Proceedings of the NationalAcademy of Sciences of the USA 106:192–196.

Pontzer, Herman, David A. Raichlen, Robert W. Shumaker, Cara Ocobock,and Serge A. Wich. 2010a. Metabolic adaptation for low energy throughputin orangutans. Proceedings of the National Academy of Sciences of the USA107:14048–14052.

Pontzer, Herman, David A. Raichlen, and Michael D. Sockol. 2009. The met-abolic cost of walking in humans, chimpanzees, and early hominins. Journalof Human Evolution 56:43–54.

Pontzer, Herman, Campbell Rolian, G. Philip Rightmire, Tea Jashashvili, Mar-cia S. Ponce de Leon, David Lordkipanidze, and Christoph P. E. Zollikofer.2010b. Locomotor anatomy and biomechanics of the Dmanisi hominins.Journal of Human Evolution 58:492–504.

Raichlen, David A., Adam D. Gordon, William E. H. Harcourt-Smith, AdamD. Foster, and W. Randall Haas Jr. 2010. Laetoli footprints preserve earliestdirect evidence of human-like bipedalism. PLoS ONE 5:e9769.


Rapoport, Benjamin I. 2010. Metabolic factors limiting performance in mar-athon runners. Public Library of Science, Computational Biology 6:e1000960.

Roberts, Thomas J., Michael S. Chen, and C. Richard Taylor. 1998. Energeticsof bipedal running. 2. Limb design and running mechanics. Journal ofExperimental Biology 201:2753–2762.

Ruff, Christopher. 2009. Relative limb strength and locomotion in Homohabilis. American Journal of Physical Anthropology 138:90–100.

Ruff, Christopher B. 1994. Morphological adaptation to climate in modernand fossil hominids. 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.

Ruff, Christopher B., and Alan Walker. 1993. Body size and shape. In TheNariokotome Homo erectus skeleton. Alan Walker and Richard Leakey, eds.Pp. 234–265. Cambridge, MA: Harvard University Press.

Schmidt-Nielsen, Knut. 1999. Animal physiology: adaptation and environment.5th edition. Cambridge: Cambridge University Press.



Sibly, Richard M., and James H. Brown. 2007. Effects of body size and lifestyleon evolution of mammal life histories. Proceedings of the National Academyof Sciences of the USA 104:17707–17712.


Sockol, Michael D., David A. Raichlen, and Herman Pontzer. 2007. Chim-panzee locomotor energetics and the origin of human bipedalism. Pro-ceedings of the National Academy of Sciences of the USA 30:12265–12269.

Stern, Jack T., Jr. 2000. Climbing to the top: a personal memoir of Austral-opithecus afarensis. Evolutionary Anthropology 9:113–133.

Stern, Jack T., Jr., and Robert L. Susman. 1983. Locomotor anatomy of Aus-tralopithecus afarensis. American Journal of Physical Anthropology 60:279–317.

Steudel, Karen, and Jeanne Beattie. 1993. Scaling of cursoriality in mammals.Journal of Morphology 217:55–63.

Steudel-Numbers, Karen L. 2006. Energetics in Homo erectus and other earlyhominins: the consequences of increased lower-limb length. Journal of Hu-man Evolution 51:445–453.

Steudel-Numbers, Karen L., and Michael J. Tilkens. 2004. The effect of lowerlimb length on the energetic cost of locomotion: implications for fossilhominins. Journal of Human Evolution 47:95–109.

Steudel-Numbers, Karen L., Timothy D. Weaver, and Cara M. Wall-Scheffler.2007. The evolution of human running: effects of changes in lower-limblength on locomotor economy. Journal of Human Evolution 53:191–196.

Taylor, C. Richard, Norman C. Heglund, and Geoffrey M. O. Maloiy. 1982.Energetics and mechanics of terrestrial locomotion. 1. Metabolic energyconsumption as a function of speed and body size in birds and mammals.Journal of Experimental Biology 97:1–21.

Tilkens, Michael J., Cara Wall-Scheffler, Timothy D. Weaver, and Karen Steu-del-Numbers. 2007. The effects of body proportions on thermoregulation:an experimental assessment of Allen’s rule. Journal of Human Evolution 53:286–291.

Tuttle, Russel H., David M. Webb, and Michael Baksh. 1991. Laetoli toes andAustralopithecus afarensis. Human Evolution 6:193–200.

Walker, Alan. 1993. Perspectives on the Nariokotome discovery. In The Na-riokotome Homo erectus skeleton. Alan Walker and Richard Leakey, eds.Pp. 411–432. Cambridge, MA: Harvard University Press.

Wang, Weijie, Robin H. Crompton, Tanya S. Carey, Michael M. Gunther, YuLi, Russell Savage, and William I. Sellers. 2004. Comparison of inverse-dynamics musculo-skeletal models of A.L. 288-1 Australopithecus afarensisand KNM-WT 15000 Homo ergaster to modern humans, with implicationsfor the evolution of bipedalism. Journal of Human Evolution 47:453–478.

Ward, Carol V. 2002. Interpreting the posture and locomotion of Australo-pithecus afarensis: where do we stand? Yearbook of Physical Anthropology 35:185–215.

Ward, Carol V., William H. Kimbel, and Don C. Johanson. 2011. Completefourth metatarsal and arches of the foot of Australopithecus afarensis. Science331:750–753.

Weaver, Timothy D. 2009. The meaning of Neandertal skeletal morphology.Proceedings of the National Academy of Sciences of the USA 106:16028–16033.

Weaver, Timothy D., and Karen L. Steudel-Numbers. 2005. Does climate ormobility explain the differences in body proportions between Neandertalsand their Upper Paleolithic successors? Evolutionary Anthropology 14:218–223.

Weibel, Ewald R., Leonardo D. Bacigalupe, Beat Schmitt, and Hans Hoppeler.2004. Allometric scaling of maximal metabolic rate in mammals: muscleaerobic capacity as determinant factor. Respiratory Physiology and Neuro-biology 140:115–132.

Wells, Jonathan C. K., and Jay T. Stock. 2007. The biology of the colonizingape. Yearbook of Physical Anthropology 50:191–222.

Weyand, Peter G., and J. Adam Davis. 2005. Running performance has astructural basis. Journal of Experimental Biology 208:2625–2631.

Wood, Bernard, and Mark Collard. 1999. The human genus. Science 284:65–71.

Wrangham, Richard. 2009. Catching fire: how cooking made us human. NewYork: Basic.

Wrangham, Richard W., James Holland Jones, Greg Laden, David Pilbeam,and NancyLou Conklin-Brittain. 1999. The raw and the stolen: cookingand the ecology of human origins. Current Anthropology 40:567–594.



The Effects of Mortality, Subsistence, andEcology on Human Adult Height and

Implications for Homo Evolution

by Andrea Bamberg Migliano and Myrtille Guillon

CA� Online-Only Material: Supplement A

The increase in body size observed with the appearance and evolution of Homo is most often attributed to ther-moregulatory and locomotor adaptations to environment; increased reliance on animal protein and fat; or increasedbehavioral flexibility, provisioning, and cooperation leading to decreased mortality rates and slow life histories. Itis not easy to test these hypotheses in the fossil record. Therefore, understanding selective pressures shaping heightvariability in living humans might help to construct models for the interpretation of body size variation in thehominins. Among human populations, average male height varies extensively (145 cm–183 cm); a similar range ofvariation is found in Homo erectus (including African and Georgian samples). Previous research shows that heightin human populations covaries with life history traits and variations in mortality rates and that different environmentsaffect adult height through adaptations related to thermoregulation and nutrition. We investigate the interactionsbetween life history traits, mortality rates, environmental setting, and subsistence for 89 small-scale societies. Weshow that mortality rates are the primary factor shaping adult height variation and that people in savanna areconsistently taller than people in forests. We focus on relevant results for interpreting the evolution of Homo bodysize variability.

Body size is one of the major features that distinguishes aus-tralopiths from early Homo and early Homo from Homoerectus (Anton 2012; Holliday 2012; Pontzer 2012). Some ofthe most important questions about the evolution of Homoconcern the reasons behind the observed size increase betweenthese taxa and the variation within them.

Paleoanthropologists have offered a number of hypothesesfor body size increase in hominins. For example, Wheeler(1992) proposed that body size and body proportions werespecific thermoregulatory adaptations to the environmentsencountered by the hominins, as have Vrba (1996), Passey etal. (2010), and Trauth et al. (2010). A number of authors alsohave related diet to increased body size (e.g., Aiello andWheeler 1995; Carmody and Wrangham 2009). AndO’Connell, Hawkes, and Blurton Jones (1999) argue that theevolution of larger body size in H. erectus was originally as-

Andrea Bamberg Migliano is Lecturer in Evolutionary Anthropologyin the Department of Anthropology at University College London(14 Taviton Street, London WC1H 0BW, U.K. [[email protected]]). Myrtille Guillon is a graduate student in the Department ofAnthropology at University College London (Gower Street, LondonWC1E 6BT, U.K). This paper was submitted 12 XII 11, accepted 9VII 12, and electronically published 27 XI 12.

sociated with reduced mortality rates (associated with in-creased alloparenting) in comparison with earlier membersof Homo and australopithecines.

These hypotheses are largely mutually exclusive, and theexplanation for size variation in the hominins undoubtedlyinvolves a complex interaction between such factors as cli-mate, mortality rates, and nutrition. By building a detailedunderstanding of the causes of body size differences in mod-ern human populations, we believe that we will be in a muchstronger position to generate testable hypotheses for body sizechanges during hominin evolution.

In modern humans there is good evidence of the relation-ship between body size and climate. For example, Roberts(1973) demonstrated body size increase with distance fromthe equator. Size variation has also been linked directly tothermoregulatory adaptations. Specifically, smaller body size(both in height and weight) may help to reduce heat pro-duction in hot and humid environments (Cavalli-Sforza 1986)while larger bodies may help to conserve heat in cold envi-ronments (Bergmann 1847), a pattern that is generally foundamong other mammals (Ashton, Tracy, and de Queiroz 2000).

With the publication of Charnov’s (1992) general life his-tory model, growing attention has been directed to mortalityrisk as a factor potentially shaping body size evolution. As




death restricts the amount of time available to an organism,

time and energy invested into one process (growth, reproduc-

tion, and maintenance) cannot be invested in another. In other

words, not all processes can be simultaneously maximized

(Charnov 1992; Stearns 1992). Because larger individuals tend

to have a higher energy capture rate during growth and thus

a higher production rate at adulthood and a higher energy

budget to be invested into reproduction (Stearns 1992), more

time allotted to growth (i.e., late maturation) will tend to be

favored. However, there are potential costs to delaying matu-

ration as it increases the likelihood of death before reproduc-

tion. For this reason, low mortality rates favor delayed matu-

ration and large body size, and high mortality favors earlier

reproduction and growth termination and hence small body

size. For this reason, mortality rates are likely to determine the

pace of life histories, the balance of investment in growth versus

reproduction, and variation in adult body size (Harvey and

Clutton-Brock 1985; Harvey and Purvis 1999; Harvey and

Zammuto 1985; Promislow and Harvey 1990).

This general relationship between mortality and growth

should influence body size variability not only across taxa but

also among human populations (Adair 2007; Kuzawa and

Bragg 2012; Migliano, Vinicius, and Lahr 2007; Walker et al.

2006). One example of this relationship in humans is the

short stature of Pygmies, which we found to be best explained

as a consequence of an accelerated life history (early growth

cessation) caused by high mortality rates in a nutritionally

stressful environment (Migliano 2005; Migliano, Vinicius, and

Lahr 2007, 2010; Stock and Migliano 2009).

Diet and nutrition are also important factors affecting

growth and consequently adult height in modern humans

(Bailey 1991; Golden 1991). For example, malnourished chil-

dren suffering from protein or calorie deficiency grow slowly,

delay maturation, and achieve shorter stature (Akachi and

Canning 2007; Cameron 1991; Danubio and Sanna 2008;

Lampl, Johnston, and Malcolm 1978; Silventoinen 2003). In-

terestingly, these factors affect adult body size through dif-

ferent mechanisms; while malnourishment will lead to slower

growth rates and delayed maturation (Lampl, Johnston, and

Malcolm 1978), increased adult mortality rates should lead

to the acceleration of growth rates and earlier maturation

(Migliano, Vinicius, and Lahr 2007; Walker et al. 2006).

What is the effect of adult mortality rates, rates of growth

and maturation, subsistence strategies, and variation in en-

vironmental settings on current human height diversity? We

analyze a cross-cultural database that includes 89 living hu-

man populations of foragers, small-scale farmers, horticul-

turalists, and pastoralists living in varying environments from

forests to deserts (see CA� online supplement A). We then

discuss the applications of our findings to the hominin fossil

record and propose testable hypotheses for explaining the

variation in body size observed in the genus Homo.

Material and Methods

We use a compiled database that includes information on thelife histories of 89 small-scale human populations. Part of thedata was obtained from the Comparative Human Life HistorySpreadsheet,1 with part of the data previously published inMigliano, Vinicius, and Lahr (2007) and Walker et al. (2006).We supplement this database with data from other ethno-graphic sources that provide information about average stat-ure, age at menarche, and survival to age 15 in traditionalsmall-scale societies. All populations are classified accordingto their primary environment using relevant ethnographicliterature (supplement A).

Most mortality data were obtained from Walker et al.(2006), which describes data quality. Populations were se-lected for inclusion on the basis of two specific criteria; onlytraditional small-scale societies were sampled, and popula-tions that had experienced recent significant changes in life-style were excluded.

Dietary variables were taken from Binford (2001), whodescribed the diet of hunter-gatherer populations in terms ofpercentage of food coming from hunting (percentage of re-liance on hunting), from fishing (percentage of reliance onfishing), and from gathering (percentage of reliance on gath-ering). We use these data to estimate the effects of relianceon animal protein (increased meat and fish in the diet) onhunter-gatherers’ size variation.

There are several limitations to this cross-cultural approach.First, the fact that different measurements have been taken bydifferent people at different times with variable sample sizespotentially introduces a number of sources of error. Second,the demographic indicators of mortality (survival to age 15,life expectancy at birth, and life expectancy at age 15) are de-rived primarily from retrospective interviews but in some casesare inferred from stable population models. Third, dietary per-centages rely on the quality of the data obtained for each pop-ulation (see Binford 2001 for a description of the subsistencedata). Nonetheless, if we are to understand how different eco-logical and demographic variables affect variation in humanbody size worldwide, it is necessary to rely on cross-culturalsamples. Here, we have done our best to ensure compatibilityof the data and present the results of this analysis as hypothesesto stimulate further work in this area.

To test data quality, we analyzed subsets of the data as wellas the entire data set. The results were very similar in allanalyses. For example, we regressed adult body size on prob-ability of survival to age 15 controlling for continent, sex, andenvironment for the total data set ( ) and only for then p 42hunter-gatherer sample ( ). In both samples survival atn p 29age 15 had a significant positive effect on adult height, andpeople in the savanna were significantly taller than people inthe forest (comparisons not shown).

1. http://dice.missouri.edu.

http://dice.missouri.edu

Migliano and Guillon Mortality and Height Variability S361

Table 1. Linear regression models using life expectancy at birth, life expectancy at age 15, and probability of survival toage 15 to predict adult height

Predictor

Whole model (including sex, continent,and one of the three mortality

predictors)

Last block (change)when includingone of the three

mortality predictors

Partial correlation and standardizedb coefficient (controlling for

sex and continent)

R R2 (SE) ANOVA (P) R2change Fchange (P) Partial correlation Standardized b (P)

Life expectancy at birth(n p 19) .787 .620 (.017) F p 4.24 (.001) .253 8.66 (.011) .632 .593 (.011)

Life expectancy at age 15(n p 19) .843 .711 (.015) F p 6.38 (.003) .336 15.06 (.002) .733 .649 (.002)

Probability of survival toage 15 (n p 42) .755 .570 (.017) F p 6.43 (!.001) .299 23.6 (!.001) .640 .592 (!.001)

Table 2. Linear regression model using probability of survival to age 15 to predict relative height at age 10

Predictor

Whole model (including sex, continent,and probability of survival to age 15)

Last block (change)when includingprobability of

survival to age 15


sex and continent)

R R2 (SE) ANOVA (P) R2 F (P) Partial correlation Standardized b (P)

Probability of survival toage 15 (n p 16) .878 .771 (.015) F p 9.24 (.002) .357 17.14 (.002) �.780 �.766 (.002)

Least squares multiple regression models are employed toassess the extent to which estimates of mortality (life expec-tancy at birth, life expectancy at age 15, and survival to age15) and diet predict human height. In order to include theeffect of the primary environment in the analysis, dummyvariables were created for each environment type and sub-sistence strategy. The primary environments are categorizedas forest, savanna, desert, and tundra. When dummy variablesare included in a regression model, stepwise methods becomeinappropriate, and so the enter method was used. To controlfor the nonindependence of human societies as statisticalpoints that can arise because of phylogenetic history, popu-lations were classified according to their continents, and thisvariable was used as a control in the regression analyses. Thiswas considered appropriate because our sample involves iso-lated populations that have a long tradition of life in theircurrent geographical settings. Bivariate correlation analysesare used to investigate relationships between life history var-iables and body size. We control for sex when males andfemales from different populations were entered in the sameanalyses and when the analyses included populations forwhich male and female variables had been calculated togetherin the original publications. We used log transformation whennormality transformations were required. All regression, cor-relation, and variance analyses were produced using PASW18 (SPSS, Chicago).

Results

Five separate sets of multiple regression analysis are carriedout. The first two focus on mortality rate and its relationship

to adult height variation and to growth and development. Thesecond two focus on environmental setting and its relation-ship to adult height variation and to both mortality rate andadult height variation. The last analysis looks at the relation-ship between subsistence strategy and adult height and lifehistory diversity. The major result emerging from these anal-yses is that measures of mortality (or survivorship) are theprime determinate in adult height in the sample of small-scale human populations, although environmental setting alsohas a significant influence.

Analysis I: Relationship between MortalityRate and Adult Height Variation

In separate multiple regression analyses including sex andcontinent, three different measures of mortality/survivorshipeach show a significant correlation with adult height, popu-lation average life expectancy at birth ( , ,2n p 19 R p 0.620

), life expectancy at age 15 ( , ,2P ! .001 n p 19 R p 0.711), and the probability of survival to age 15 ( ,P ! .001 n p 42

, ; table 1). The fact that life expectancy2R p 0.570 P ! .001at age 15 is strongly correlated with adult height suggests thatmortality rates in adulthood affect adult stature.

Controlling for sex and continent, life expectancy at birthis a strong predictor of adult height ( , ),b p 0.593 P p .011explaining an extra 25.3% of the variance in height in relationto what is explained only by continent and sex. When theprobability of survival to age 15 is used to predict adult height,the relationship is very similar to that with life expectancy atbirth ( , ), with an extra 29.9% of the varianceb p 0.592 P ! .001explained in relation to the basal model (which includes sex


Table 3. Linear regression model using survival to age 15 and life expectancy at age 15 to predict age at menarche

Predictor

Whole model (including continent andprobability of survival to age 15 or

life expectancy at age 15)

Last block (change)when includingprobability of

survival to age 15or life expectancy

at age 15

Partial correlation and standardizedb coefficient (controlling

for continent)


Probability of survival toage 15 (n p 16) .832 .693 (.033) F p 6.196 (.007) .073 2.62 (.134) .439 .324 (.134)

Life expectancy at age 15(n p 13) .864 .747 (.032) F p 5.915 (.016) .191 6.05 (.039) .656 .456 (.039)

Table 4. Linear regression model using environmental settings (forest, savanna [S], tundra [T], and desert [D]) to predictadult height

Predictor

Whole model (including sex, continent,and environment)

Last block (change)when including

environment

Partial correlation and standardizedb coefficient, foresta as

reference (controlling forsex and continent)


Environmental settings(n p 150) .789 .622 (.016) F p 22.8 (!.001) .167 20.5 (!.001) S, .539; D, .270;

T, .06S, .442 (!.001); D,

.181 (.001); T, .089(!.476)

a Environmental settings were entered as dummy variables because “environmental settings” is a categorical variable. As such, one of the categorieshas to be a reference (all other categories are measured against the reference).

and continent). Using life expectancy at age 15 as a predictorof adult height slightly increases correlations ( ,b p 0.642 P !

), and an extra 33.6% of the variance is explained in relation.001to the basal model. This brings the total variance explained bythe whole model (including continent, sex, and life expectancyat age 15) to 71% (table 1). Survival to age 15 and life expec-tancy at age 15 are highly correlated ( , ,n p 18 R p 558 P p

); however, the first expresses prereproductive survival,.016while the second expresses adult survival.

Analysis 2: Relationship between MortalityRates, Growth, and Development

This analysis investigates the relationship between probabilityof survival to age 15 and relative height at age 10 (percentageof adult height achieved by age 10), which is used as a proxyfor growth rate (the more growth completed by age 10, thefaster the growth rate). Probability of survival to age 15 isused here rather than life expectancy at age 15 because ofsample size constraints. The multiple linear regression modelcontrolling for sex and continent shows a negative associationbetween probability of survival to age 15 and percentage ofadult height achieved by age 10 ( , ,n p 16 b p �0.766 P p

), which implies that children exposed to higher mortality.002environments grow at faster rates than children in lower mor-tality environments (table 2).

In addition, the influence of mortality rate on age at men-

arche was analyzed. When probability of survival to age 15is used as the predictor variable to assess the effect of mortalityon age at menarche, the relationship is not significant becausethe inclusion of survival to age 15 as a predictor to the basalmodel (including continent) does not improve the model( , , , ; table 3).2n p 16 R p 0.073 F p 2.62 P p .134change change

However, when life expectancy at age 15 is used as the in-dependent variable, there is a significant correlation with ageat menarche when continent is controlled for ( ,b p 0.456

; table 3). Including life expectancy at age 15 as aP p .039predictor significantly improves the basal model ( ,n p 16

, ), explaining an extra 19.1% of theF p 6.05 P p .039change

variability in age at menarche in our samples and bringingthe total variability explained by the whole model to 74.7%.This is a positive relationship, meaning that populations wholive for longer have a later age at menarche (table 3).

Analysis 3: Relationship between EnvironmentalSettings and Adult Height Variation

To test whether environmental setting has an effect on bodysize, we ran a multiple regression analysis controlling for sexand continent, with adult height as the dependent variableand environmental settings (savanna, forest, tundra, and des-ert) as predictor variables. Dummy variables are used to enterthe environmental settings as predictors, as environmentalsettings are discrete. The inclusion of environmental settings


Table 5. Total frequency of cases in each continent and each environmental setting

Ecology Africa Asia Australia Europe North America South America

Desert 5 0 4 0 1 1Forest 17 53 3 2 2 18Savanna 14 8 1 0 2 6Tundra 0 0 0 2 10 1

Table 6. Linear regression model using environmental settings (forest, savanna [S], tundra [T], and desert [D]) and prob-ability of survival to age 15 to predict adult height

Predictor

Whole model (including sex, continent,environment, and probability of

survival to age 15)


survival toage 15

Partial correlation and standardized b

coefficient, foresta as reference(controlling for sex, continent,

and environment)


Environmental settingsand survival toage 15 (n p 42) .836 .700 (.015) F p 7.22 (!.001) .269 27.7 (!.001) S, .543; D, .187; T,

�.065; probabilityof survival, .687

S, .387 (.001); D,.115 (.296); T,�.106 (.717);probability ofsurvival 15,.569 (!.001)

a Environmental settings were entered as dummy variables because “environmental settings” is a categorical variable. As such, one of the categorieshas to be a reference (all other categories are measured against the reference).

in the last block significantly improves the basal model( , ), explaining an additional 16.7% ofF p 22.8 P ! .001change

the height variability. The complete model including sex, con-tinent, and environment is highly significant and explains62.2% of the variance in height ( , ,2n p 150 R p 0.622 P !

). Controlling for sex and continent, the savanna envi-.001ronment has a significantly positive effect on adult heightcompared with the forest environment ( , ,n p 150 b p 0.442

) and the desert environment ( , ,P ! .001 n p 150 b p 0.181; table 4). Populations are represented from all con-P p .001

tinents, and therefore phylogenetic relationships are unlikelyto explain why living in the savanna or desert is associatedwith greater stature than living in the forest (table 5).

Analysis 4: Relationship between Environmental Setting,Mortality Rates, and Adult Height Variation

When probability of survival to age 15 is included as a pre-dictor variable together with the environmental settings, thegeneral model improves ( , , ) in2n p 42 R p 0.700 P ! .001spite of the much smaller sample sizes ( vs. ).n p 42 n p 150The inclusion of probability of survival to age 15 togetherwith environmental settings in the last block of the regressionexplains an additional 26.9% of the height variability in oursample. Living in the savanna remains positively associatedwith height compared with living in the forest even when theeffect of survival to age 15 is taken into account ( ,n p 42

, ), but the contrast between living in theb p 0.387 P ! .001desert and living in the forest found previously is no longer

significant ( , , ). Probability of sur-n p 42 b p 0.115 P 1 .05vival to age 15 is the strongest predictor of height in the model( , , ; table 6). The reduction of then p 42 b p 0.569 P ! .001explanatory power of the environmental factors when theprobability of survival to age 15 is introduced as a predictorhas to be interpreted with caution because the sample size(and therefore the representation of all environmental settingsin each continent) is reduced when all variables are included(table 7).

Unfortunately, sample sizes and sample distributions areinsufficient to allow height, environmental variables, andother mortality indicator variables (life expectancy at birthand life expectancy at age 15 in the same analyses) to beanalyzed together. However, the results do show that the prob-ability of survival to age 15 is a strong predictor of adultheight and that living in the savanna has a more positiveinfluence on height than living in the forest independent ofthe effect of mortality.

Analysis 5: Relationship between Subsistence Strategies,Adult Height, and Life History Diversity

In order to assess the effect of subsistence strategy on adultbody size variation, we used data on diet composition inhunter-gatherer groups assembled by Binford (2001, table5.01). The percentages of reliance on fishing and hunting werecombined to calculate reliance on animal protein.

We use a multiple linear regression analysis with height asthe dependent variable to understand the role of the per-


Table 7. Frequency of cases in each continent and each en-vironmental setting when only individuals with data onsurvival to age 15 are considered

Ecology Africa Asia North America South America

Desert 3 0 1 1Forest 9 10 2 18Savanna 4 0 2 6Tundra 0 0 10 1

Table 8. Linear regression model using environmental settings (forest, savanna, tundra, and desert) and percentage of de-pendence on animal protein to predict adult height

Predictor

Whole model (including sex, continent,environment, and percent reliance

on animal protein)


percent reliance onanimal protein


sex and continent)


Percent relianceon animal protein(n p 52) .901 .811 (.0137) F p 15.6 (!.001) .008 1.73 (.196) .204 .196 (.196)

centage of reliance on animal protein in shaping adult heightvariability among hunter-gatherers, controlling for sex, con-tinent, and environmental variables. Although the total modelincluding sex, continent, environment, and percent relianceon animal protein is highly significant ( , ,2n p 52 R p 0.811

), percent reliance on animal protein is not a significantP ! .05factor in predicting height when the other variables are con-trolled ( , , ; table 8).n p 52 b p 0.196 P 1 .05

Discussion

These results show that variation in adult height is highlyinfluenced by mortality patterns in our worldwide sample ofsmall-scale human populations. Preadult mortality (proba-bility of survival to age 15) alone (when controlled by sexand continent) explains 29.9% of the variation in adult height,while adult survival (life expectancy at age 15) explains 33.6%.The model combining life expectancy at age 15 with sex andcontinent explains a great proportion of the variance (71.1%).

These results are in accordance with the predictions of lifehistory theory (Charnov 1992). Life history theory predictsthat high mortality rates should lead to relatively fast-pacedlife history strategies (Charnov 1992). Correspondingly, weshould expect that those human populations that face higherrates of adult mortality will have individuals who, on average,develop earlier. They will achieve full growth relatively early,reproduce at an earlier age than those with relatively lowmortality rates, and thereby reduce the chances of death be-fore reproduction (Migliano 2005; Migliano, Vinicius, andLahr 2007; Walker et al. 2006).

Our results also provide clues to the mechanisms throughwhich mortality rates affect adult body size. According to thetheory, populations under high risk of death should develop

faster, achieving adult body sizes sooner, and should also ma-ture earlier in order to start reproducing sooner. This is whatwe find in our analyses: populations living in higher mortalityenvironments have earlier menarche and grow at a faster rate(as proxied by the proportion of adult body size achieved atage 10) than populations living in lower mortality environ-ments (see tables 2, 3). This is the opposite effect observedwhen short stature is determined by malnourishment, wherepopulations should achieve short stature as a result of delayedsexual and growth maturation (Migliano, Vinicius, and Lahr2007).

As Kuzawa and Bragg (2012) have shown, developmentalplasticity exerts a strong influence on age at menarche andbody size (both weight and height). Populations experiencingovernutrition in the West or individuals recently adopted intoWestern culture have earlier menarche and earlier growthcessation and achieve larger body sizes as a consequence ofdecreasing nutritional constraints. The expression of this phe-notype should, therefore, be interpreted as a response to anunconstrained environment (where time and resources arevirtually unlimited, releasing trade-offs; fig. 1). However, thissituation is virtually nonexistent in tribal populations, whereboth time (survivorship) and resources (calories) are limitedto varying extents. Our results suggest that mortality rateshave a strong influence on adult height and maturation var-iability in natural environment populations; although the im-portance of nutrition should also be considered (fig. 1; seeKuzawa and Bragg 2012, fig. 3). Our analysis of the effectsof diet on height variation in a subset of the sample (thehunter-gatherers) for which data on reliance on gathering,fishing, and hunting were available did not reach significance.There was a positive correlation between a diet high in animalprotein and adult stature; however, a larger sample size andbetter data are needed.

Finally, our results indicate that living in the savanna hasa positive correlation with height compared with living in theforest irrespective of the differences observed in mortalityrates. This result is found even when controlling for continent(as a way to control for phylogenetic and statistical nonin-dependence between populations). Other effects, such as dif-ferences in temperature, humidity, or diet undoubtedly playa role in explaining interpopulation differences in stature. Tall


Figure 1. Interaction between nutrition and mortality rates in-fluencing adult height. In populations with no caloric restrictionsand low mortality rates, growth cessation happens relatively early(maximizing reproductive span) at larger body sizes (maximizingenergy budget to be invested into reproduction), such as in thepopulations of the United States (National Center for HealthStatistics [NCHS] data; white squares, dashed lines, NCHS 2002).A, In populations where nutritional stress is high and life ex-pectancy is relatively high, such as the Turkana pastoralists (datafrom Little, Galvin, and Mugambi 1983; black squares), growthcessation is delayed (maximizing body size); that is, Turkanapastoralists achieve the same adult height as the well-nourishedAmericans—however, much later. B, When high mortality ratesare combined with relatively poor nutrition, growth cessationcannot be delayed because reproductive life span is already im-paired; thus, populations in these conditions, such as the Aetafrom the Philippines (data from Migliano, Vinicius, and Lahr2007; gray squares), have an early growth cessation (as early asthe Americans) at much smaller sizes (2 SD below the Americanaverage).

stature and narrow elongated bodies have long been explainedas a thermoregulatory mechanism in hot savanna environ-ments (e.g., Ruff 1993). The association between tropical for-ests and short human stature has also been identified, anddifferent adaptive hypotheses have been suggested to explainit, from thermoregulatory adaptations proposing that shorterhumans produce less heat in a tropical humid environment,being more efficient at cooling down (Cavalli-Sforza 1986),to easier locomotion in closed forests (Turnbull 1986) andadaptation to poor carbohydrate availability in tropical forests(Bailey and Peacock 1988). Small body size in forested en-vironments is also seen in other mammals. For example, forestelephants have been estimated to be 35%–40% smaller instature than their savanna counterparts (Morgan and Lee2003), and the pygmy hippopotamus is found mainly in Li-beria and only in thick forests (Eltringham 1999). More dataare necessary to test whether these other variables explain partof the variance not captured by differences in mortality ratesand to understand how mortality rates interact (or are influ-enced) by these other pressures.

Applications to Hominin Evolutionary History

Body size increase is one of the main features distinguishingthe australopiths from early Homo and early Homo fromHomo erectus (Anton 2012; Holliday 2012; Pontzer 2012). Forearly Homo, height and weight can be approximately esti-mated for KNM-ER 3735, OH 62, KNM-ER 1472, and 1481.The height range is 118–149 cm (Anton 2012), and averageweight is 50–54 kg (Holliday 2012). This represents a 30%increase over the condition in Australopithecus (Anton 2012;Holliday 2012; Pontzer 2012). With the appearance of H.erectus there was an additional increase in body size of com-parable magnitude (Anton 2012).

When comparing the variability in body size in modernhumans to early Homo and H. erectus, it is clear that H. erectusheight overlaps with the recent human range, while the earlyHomo variation falls below recent human variation (fig. 2A).In contrast, inferred weight variation falls well within thehuman range for early and late Homo (fig. 2B). This patternindicates important stature/weight differences between earlyHomo and living humans, implying differences in body pro-portions (see Holliday 2012).

In H. erectus the range of body size variation (40–68 kg,146–185 cm) falls well within the interpopulation variationobserved in modern humans (fig. 2; Anton 2012). Moreover,Georgian specimens have ranges of size overlapping the in-trapopulation variation observed in Philippine Pygmies (fig.2). Although there is great variation in body size within H.erectus, the similarities with modern human ranges indicatethat perhaps the same patterns of diversification apply to thetwo groups. Therefore, we suggest that understanding thecauses of body size variation in living human populations isrelevant to interpreting the variation observed within Homoand in particular within H. erectus.


Figure 2. A, Distribution of height across 3,263 small-scale world population averages (gray) and within Aeta Pygmies from thePhilippines (black; ). The thick black line is the range of height variation in early Homo, while the thick dark gray line isn p 348the range of height variation in Homo erectus (Georgian specimens), and the thick light gray line is the range of height variationin H. erectus (African specimens). B, Distribution of weight across 249 small-scale world population averages (gray) and withinAeta Pygmies from the Philippines (black; ). The thick black line is the range of weight variation in early Homo, while then p 348thick dark gray line is the range of weight variation in H. erectus (Georgian specimens), and the thick light gray line is the rangeof weight variation in H. erectus (African specimens). Human variation from Kings Diversity Project (M. M. Lahr and R. Foley,unpublished data); Aeta Pygmy data (Philippines) from Migliano (2005); and hominin data from Anton (2012).

Our results indicate that two factors could have been as-sociated with the stature variation observed both between andwithin hominin taxa. The first is variation in preadult andadult mortality rates, and the second is occupation of differentenvironments. Although diet was not significant in our anal-yses, we know from studies of modern human populationsthat it also could be a relevant factor (e.g., Kuzawa and Bragg2012).

Testing the distinct influences of at least mortality rate andnutrition on observed body size in the fossil record may bepossible to achieve. The results of this study and of others(Kuzawa and Bragg 2012; Migliano, Vinicius, and Lahr 2007;Walker et al. 2006) indicate that poor nutrition and highmortality rates both lead to short stature but through differentmechanisms: poor nutrition should lead to delayed growthand maturation (due to the lack of resources), while highmortality rates should lead to early development and matu-ration (due to selection for early reproduction). Understand-ing how closely the pace of tooth development and the se-quence and timing of dental eruption match the pace of lifehistory in current and extinct populations (Dean 2006;Schwartz 2012) would help immensely with understandingthe causes behind body size variation in Homo. For example,if the short stature of Homo floresiensis (Brown et al. 2004)and H. erectus in Georgia (aka Homo georgicus; Lordkipanidzeet al. 2007) were a consequence of nutritional insufficiency,

we would expect lower rates of growth and development inrelation to larger H. erectus specimens. On the other hand, ifdifferences were due to differing mortality environments, wewould expect the smaller specimens to have a more rapid lifehistory pace (as proxied by their patterns of tooth develop-ment).

Although the data are not yet available to test the rela-tionship of nutrition and mortality rate on hominin statureand the relationship of these variables to the third compli-cating factor, environment, it is interesting to speculate onthe implications of these factors to the evolution of Homo. Ifmortality rate does prove to be a significant factor associatedwith hominin stature as our analyses suggest, it would implythat early Homo had a reduced mortality rate in relation tothe australopiths and that H. erectus had further reduced itsmortality rate in relation to early Homo. The question thenis how the hominins achieved reduced mortality rates in thecontext of an increasingly variable environment and expan-sion into a broader range of niches (Potts 2012). Potts arguesthat greater behavioral flexibility associated with a larger brainsize together with the capacity to extract more effectively pro-tein and fat resources and an increased capacity for avoidingpredation might have contributed to the reduction of mor-tality rates (see also Anton and Snodgrass 2012; Kaplan et al.2000).

Important factors undoubtedly include subsistence shifts


and predator avoidance but would also include other culturaladaptations as well as opportunities for cooperation, allo-parenting, parental investment, and increased child provi-sioning (Bribiescas, Ellison, and Gray 2012; Isler and vanSchaik 2012). The importance of cooperation in early Homoand H. erectus also would be in agreement with a number ofother lines of reasoning. For example, without an increase incooperative social behavior to help reduce interbirth intervals,energetic demands on H. erectus females would likely havebeen unmanageable (Aiello and Key 2002). The “gray ceilinghypothesis” also argues that the larger brain size found inearly Homo and H. erectus would not be possible withoutalloparenting and other forms of cooperative behavior (Islerand van Schaik 2012). Furthermore, Pontzer (2012) arguesthat the larger home-range size implied by the larger bodysizes characteristic of Homo would necessitate a high-through-put dietary strategy (increased daily energy expenditure) andgreater reproductive investment resulting in an increased life-time reproductive output. He suggests that this would ne-cessitate greater food availability perhaps facilitated throughfood sharing and increased cooperative behavior.

Our analyses suggest that the effect of survival on body sizevariation seems to be stronger than the influence of environ-mental settings or diet on adult stature within a species. Theimplication is that adapting to a particular climatic settingmight be a less important factor in relation to stature increasethan reducing mortality rates. Humans rely on behavioral andcultural plasticity to adapt to different climatic settings, andthis might also have been the case for Homo and especiallyH. erectus (Potts 2012). It is probable that certain groups ofearly Homo were more successful than others in bufferingenvironmental pressures, leading to differences in extrinsicmortality rates and consequently in body size.

Uncovering the particular reasons as well as estimating therelative influence of extrinsic mortality rates, diet, and en-vironmental setting in shaping this diversity will require aclose comparative look at behavioral adaptations, resourceexploration, predation, and competition in these groups. Itis a challenge for future research.

Acknowledgments

We are grateful to Leslie Aiello and Susan Anton, with assis-tance from Laurie Obbink, for organizing the Wenner-Grensymposium in Sintra, Portugal, and to all participants in thesymposium for insightful comments. We thank Lucio Vini-cius, Rodolph Schlaepfer, Ruth Mace, Tom Currie, and theHuman Evolutionary Ecology Research Group at UniversityCollege London for useful comments.

References CitedAdair, Linda S. 2007. Size at birth and growth trajectories to young adulthood.

American Journal of Human Biology 19(3):327–337.

Aiello, Leslie C., and Catherine Key. 2002. Energetic consequences of beinga Homo erectus female. American Journal of Human Biology 14(5):551–565.

Aiello, Leslie C., and Peter Wheeler. 1995. The expensive-tissue hypothesis:the brain and the digestive system in human and primate evolution. CurrentAnthropology 36:199–221.

Akachi, Yoko, and David Canning. 2007. The height of women in sub-SaharanAfrica: the role of health, nutrition, and income in childhood. Annals ofHuman Biology 34(4):397–410.


Anton, Susan C., and J. Josh Snodgrass. 2012. Origin and evolution of genusHomo: a new perspective. Current Anthropology 53(suppl. 6):S479–S496.

Ashton, Kyle G., Mark C. Tracy, and Alan de Queiroz. 2000. Is Bergmann’srule valid for mammals? American Naturalist 156:390–415.

Bailey, Robert C. 1991. The comparative growth of Efe Pygmies and Africanfarmers from birth to age 5 years. Annals of Human Biology 18:113–120.

Bailey, Robert C., and Nadine R. Peacock. 1988. Efe Pygmies of northeastZaire subsistence strategies in the Ituri Forest. In Coping with uncertaintyin food supply. Igor de Garine and Geoffrey Ainsworth Harrison, eds. Pp.88–117. Oxford: Oxford University Press.

Bergmann, Karl Georg Lucas Christian. 1847. Uber die Verhaltnisse der War-meokonomie der Thiere zuihrer Grosse. Gottinger Studien 3(1):595–708.

Binford, Lewis Roberts. 2001. Constructing frames of reference: an analyticalmethod for archaeological theory building using hunter-gatherer and environ-mental data sets. Berkeley: University of California Press.

Bribiescas, Richard G., Peter T. Ellison, and Peter B. Gray. 2012. Reproductiveeffort and paternal investment plasticity in men: implications for the evo-lution of Homo. Current Anthropology 53(suppl. 6):S424–S435.

Brown, Peter, Thomas Sutikna, Michael J. Morwood, Raden P. Soejono, Jat-miko, E. Wayhu Saptomo, and Rokus Awe Due. 2004. A new small-bodiedhominin from the Late Pleistocene of Flores, Indonesia. Nature 431(7012):1055–1061.

Cameron, Noel. 1991. Human growth, nutrition, and health-status in sub-Saharan Africa. Yearbook of Physical Anthropology 34:211–250.

Carmody, Rachel N., and Richard W. Wrangham. 2009. The energetic sig-nificance of cooking. Journal of Human Evolution 57:379–391.

Cavalli-Sforza, Luigi Lucca. 1986. African Pygmies. Orlando, FL: AcademicPress.

Charnov, Eric L. 1992. Life history invariants. Oxford: Oxford University Press.Danubio, Marica, and Emanuele Sanna. 2008. Secular changes in human

biological variables in Western countries: an updated review and synthesis.Journal of Anthropological Sciences 86:91–112.

Dean, M. Christopher. 2006. Tooth microstructure tracks the pace of humanlife-history evolution. Proceedings of the Royal Society B 273(1603):2799–2808.

Eltringham, S. Keith. 1999. The hippos: natural history and conservation. Lon-don: Academic Press.

Golden, Michael H. N. 1991. The nature of nutritional deficiency in relationto growth failure and poverty. Acta Paediatrica 80:95–110.

Harvey, Paul H., and Timothy H. Clutton-Brock. 1985. Life-history variationin primates. Evolution 39(3):559–581.

Harvey, Paul H., and Andy Purvis. 1999. Understanding the ecological andevolutionary reasons for life history variation: mammals as a case study. InAdvanced ecological theory: principles and applications. Jacqueline M.McGlade, ed. Pp. 232–248. Oxford: Blackwell Science.

Harvey, Paul H., and Richard M. Zammuto. 1985. Patterns of mortality andage at first reproduction in natural populations of mammals. Nature 315:319–320.



Kaplan, Hillard, Kim Hill, Jane Lancaster, and A. Magdalena Hurtado. 2000.A theory of human life history evolution: diet, intelligence, and longevity.Evolutionary Anthropology 9(4):156–185.


Lampl, Michelle, Francis E. Johnston, and Laurence A. Malcolm. 1978. Theeffects of protein supplementation on the growth and skeletal maturationof New Guinean school children. Annals of Human Biology 5:219–227.

Little, Michael A., Kathleen A. Galvin, and Mutuma Mugambi. 1983. Cross-


sectional growth of nomadic Turkana pastoralists. Human Biology 55:811–830.


Migliano, Andrea Bamberg. 2005. Why Pygmies are small: ontogenetic im-plications of life history evolution. PhD thesis, University of Cambridge.

Migliano, Andrea Bamberg, Lucio Vinicius, and Marta Mirazon Lahr. 2007.Life-history trade-offs explain the evolution of human Pygmies. Proceedingsof the National Academy of Sciences of the USA 104:20216–20219.

———. 2010. Why are Pygmies so short? a defence of Migliano’s hypothesis.Human Biology 82:109–113.

Morgan, Bethan J., and Phyllis C. Lee. 2003. Forest elephant (Loxodonta af-ricana cyclotis) stature in the Reserve de Faune du Petit Loango. GabonJournal of Zoology 259:337–344.

O’Connell, James F., Kristin Hawkes, and Nicholas G. Blurton Jones. 1999.Grandmothering and the evolution of Homo erectus. Journal of HumanEvolution 36:461–485.

Passey, Benjamin H., Naomi E. Levin, Thure E. Cerling, Francis H. Brown,and John M. Eiler. 2010. High-temperature environments of human evo-lution in East Africa based on bond ordering in paleosol carbonates. Pro-ceedings of the National Academy of Sciences of the USA 107:11245–11249.


Potts, Richard. 2012. An African environmental framework for investigatingthe evolution of early Homo. Current Anthropology 53(suppl. 6):S299–S317.

Promislow, Daniel E. L., and Paul H. Harvey. 1990. Living fast and dyingyoung: a comparative analysis of life-history variation among mammals.Journal of Zoology 220:417–437.

Roberts, Derek Frank. 1973. Climate and human variability. California: Cum-mings.

Ruff, Christopher B. 1993. Climatic adaptation and hominid evolution: thethermoregulatory imperative. Evolutionary Anthropology 2:53–60.


Silventoinen, Karri. 2003. Determinants of variation in adult body height.Journal of Biosocial Science 35(2):263–285.

Stearns, Stephen C. 1992. The evolution of life histories. New York: OxfordUniversity Press.

Stock, Jay T., and Andrea Bamberg Migliano. 2009. Stature, mortality, andlife history among indigenous populations of the Andaman Islands, 1871–1986. Current Anthropology 50(5):713–725.

Trauth, Martin H., Mark A. Maslin, Alan L. Deino, Annett Junginger, MosesLesoloyia, Eric O. Odada, Daniel O. Olago, Lydia A. Olaka, Manfred R.Strecker, and Ralph Tiedemann. 2010. Human evolution in a variable en-vironment: the amplifier lakes of eastern Africa. Quaternary Science Reviews29:2981–2988.

Turnbull, Colin M. 1986. Survival factors among Mbuti and other hunters ofthe equatorial African rain forest. In African Pygmies. Luigi L. Cavalli-Sforza,ed. Pp. 103–123. Orlando, FL: Academic Press.

Vrba, Elizabeth S. 1996. Climate, heterochrony, and human evolution. Journalof Anthropological Research 52(1):1–28.

Walker, Robert, Michael Gurven, Kim Hill, Andrea Migliano, Napoleon Cha-gnon, Roberta De Souza, Gradimir Djurovic, et al. 2006. Growth rates andlife histories in twenty-two small-scale societies. American Journal of HumanBiology 18(3):295–311.

Wheeler, Peter E. 1992. The thermoregulatory advantages of large body sizefor hominids foraging in savannah environments. Journal of Human Evo-lution 23:351–362.



Plasticity in Human Life History StrategyImplications for Contemporary Human Variation

and the Evolution of Genus Homo

by Christopher W. Kuzawa and Jared M. Bragg

The life history of Homo sapiens is characterized by a lengthy period of juvenile dependence that requires extensiveallocare, short interbirth intervals with concomitantly high fertility rates, and a life span much longer than that ofother extant great apes. Although recognized as species-defining, the traits that make up human life history are alsonotable for their extensive within- and between-population variation, which appears to trace largely to phenotypicand developmental plasticity. In this review, we first discuss the adaptive origins of plasticity in life history strategyand its influence on traits such as growth rate, maturational tempo, reproductive scheduling, and life span in modernhuman populations. Second, we consider the likely contributions of this plasticity to evolutionary diversificationand speciation within genus Homo. Contrary to traditional assumptions that plasticity slows the pace of geneticadaptation, current empirical work and theory point to the potential for plasticity-induced phenotypes to “lead theway” and accelerate subsequent genetic adaptation. Building from this work, we propose a “phenotype-first” modelof the evolution of human life history in which novel phenotypes were first generated by behaviorally or environ-mentally driven plasticity and were later gradually stabilized into species-defining traits through genetic accom-modation.

Introduction

Modern humans are characterized by a life history strategywith features that are distinct from other nonhuman primates,including slow childhood growth, early weaning followed bya long period of dependence, a shortened interbirth interval,and lengthy life spans (Hawkes et al. 1998; Hill 1993; Hilland Kaplan 1999; Robson, van Schaik, and Hawkes 2006).Extensive provisioning of dependents by kin and potentiallyunrelated individuals allows humans to “stack” offspring andspread the burden of provisioning across alloparents, thusfacilitating relatively high fertility despite the intensity of in-vestment and high survival of each offspring (Bogin 1999;Hawkes et al. 1998; Hill and Hurtado 1996; Kaplan et al.2000). Current interest in reconstructing the evolutionaryemergence of these characteristics in modern humans and thelife histories, behaviors, and biology of fossil hominins rep-resents a key intersection between human biology and paleo-anthropology.

Recent fossil discoveries have revealed extensive variation

Christopher W. Kuzawa is Associate Professor and Jared M. Braggis a Doctoral Student in Biological Anthropology in the Departmentof Anthropology, Northwestern University (1810 Hinman Avenue,Evanston, Illinois 60208, U.S.A. [[email protected]]). Thispaper was submitted 12 XII 11, accepted 18 VI 12, and electronicallypublished 27 VIII 12.

in the fossil signatures of life history variation in Homo erectus,calling into question previous assumptions regarding the evo-lution of size and shape in Homo and bringing variation perse to the fore as an important focus of analysis (e.g., Anton2003; Anton et al. 2007). These findings raise questions aboutthe evolutionary origins of this diversity, which might reflectdistinct species or locally adapted variants of the same species(Anton 2003). While the magnitude of this regional variationmay complicate taxonomic distinctions, it is also notable forits similarities to the variation observed across contemporaryhuman populations, among which there is extensive variationin life history traits such as growth rate, body size, repro-ductive scheduling, and even life span. Biological anthropol-ogists and others who study the life histories of contemporaryhumans have shown that much of this variation can be ex-plained as the outcome of phenotypic or developmental plas-ticity triggered in response to social, nutritional, demographic,and other environmental conditions (Chisholm 1993; Ellis etal. 2009; Kuzawa and Pike 2005; Walker et al. 2006a). Ifplasticity is an important contributor to contemporary humanvariation, it follows that it has likely also been important asan influence on the life histories of human ancestors. As such,considering the mechanisms and adaptive significance of phe-notypic plasticity in modern human life histories providesinsights into the origin and function of similar variation inearly Homo.

We have several interrelated goals in this review. We first



discuss the role of phenotypic and developmental plasticityas a primary means of population adjustment to environ-mental changes occurring on generational or multigenera-tional timescales. This background discussion culminates ina review of evidence for environmental influences on humanlife history variation, which highlights the effects that nutri-tion and cues of mortality risk have on growth rate, matu-rational tempo, and adult size. A review of the literature onmodern humans confirms that developmental plasticity is theprimary source of much human life history variation acrosscontemporary populations and points to the likely impor-tance of similar processes as contributors to regional andtemporal variation in the fossil record.

Because only those genotypes that are expressed pheno-typically are subjected to selection, environment-driven phe-notypic variation in human populations may also influencethe process, pace, and direction of evolutionary change (West-Eberhard 2003). From the perspective of the emerging syn-thesis of developmental and evolutionary biology, the gen-eration of regional variation through plasticity likely providedthe raw phenotypic variation that was then selectively retainedor pruned, leading the way for more gradual adaptation vianatural selection. We conclude by speculating that environ-ment-driven developmental plasticity may not only provideinsights into the origins of variation in contemporary andfossil hominin populations but also may have played a fun-damental role in the evolution of our species by facilitatingphenotypic adaptation that preceded more durable geneticchange and speciation.

The Unusual Human Life History

Organisms vary remarkably in the size and pace of life, whichis reflected in body size, growth rate, fertility rate, and lifespan. These traits help define a species’ life history, whichmay be viewed as a life-cycle strategy that optimizes expen-ditures in service of reproductive success (Stearns 1992). Clas-sically, organisms are viewed as being constrained by finiteenergetic resources that must be partitioned to growth, re-production, and maintenance functions (Gadgil and Bossert1970). Growth builds a larger body that is less prone to pre-dation and has an absolutely greater capacity to invest inreproduction. After growth ceases at maturity, energy previ-ously allocated to growth is then shunted into reproduction.In females this involves support of offspring growth in uteroand via breast milk, and in males the building and mainte-nance of sexually dimorphic, energetically costly traits andbehaviors (Charnov 1993; Kuzawa 2007). Consequently, or-ganisms face a fundamental trade-off in deciding how muchof their time and energy budget to invest in growth or re-production in the present or to processes that minimize orrepair cellular or tissue damage to extend life span. For in-stance, because resources are finite, it is assumed that organ-isms may invest heavily in productivity—growth and repro-

duction—or have long life spans, but not both (Stearns 1989;Williams 1966).

As the reproductive benefits from investing in a durable,long-lived body may only be realized in the future, the utilityof allocating scarce resources to maintenance activities islinked to the risk of unavoidable extrinsic mortality (e.g.,predation) that members of a population or species face,which is viewed as a primary driver of life history diversifi-cation and evolution (Charnov 1993; Promislow and Harvey1990). Species living in ecological contexts characterized byhigh mortality risk are less likely to live into the future andthus are predicted to allocate a larger fraction of their energybudget to current reproduction with comparably little devotedto maintenance (Kirkwood and Rose 1991). As a result, these“fast” life history species typically have shorter life spans andgive birth to many lower-quality offspring with relatively lowsurvival prospects. Conversely, when mortality risk is low,theory predicts a “slow” life history characterized by reducedexpenditure on growth or reproduction and greater invest-ment in life span–extending maintenance. Slow life historyspecies tend to give birth to fewer offspring, but they investmore intensively in each, enhancing survival (Charnov andBerrigan 1993).

The human life history strategy is unusual in that it is clearlya slow and investment-oriented strategy in most respects whilealso characterized by some of the demographic benefits typ-ically associated with fast life histories. Low mortality ratesenable slow childhood growth and delayed onset of physicalmaturity, nutritional independence, and reproduction well be-yond what is seen in other great apes (Hawkes et al. 1998;Kaplan et al. 2000; Walker et al. 2006b). Also consistent witha slow strategy, humans invest extensive amounts of time andenergy in offspring during this protracted period of depen-dence (Gurven and Walker 2006). Yet despite this, contem-porary foraging populations manage to reproduce nearly twiceas fast as other great apes and have higher completed fertility(Walker et al. 2008).

This unusual life history capacity to invest heavily in moreoffspring has been traced to the distinctively human practiceof weaning offspring early and the consequent shortening ofthe interbirth interval relative to other great apes (Galdikasand Wood 1990; Humphrey 2010; Knott 2001; Sellen 2006,2007). Early weaning is accompanied by a long transitionalperiod of providing weanlings specially prepared foods thatmay be acquired and provisioned flexibly within human socialunits (Bogin 1999; Bogin and Smith 1996; Hawkes et al. 1998;Kaplan et al. 2000; Lee 1996). This allows alloparents, suchas grandmothers or older siblings, to provide a substantialfraction of the energetic needs of each offspring, thus freeingmaternal metabolism to initiate new pregnancies. It is in-creasingly recognized that the remarkable demographic suc-cess and geographic range of our species hinges on the flex-ibility of these alloparental transfers, which allow us to “haveour cake and eat it too,” producing many high-quality and

Kuzawa and Bragg Plasticity and Human Life History Evolution S371

Figure 1. Timescales of human adaptability (modified after Ku-zawa 2005, 2008). Light gray p more rapidly responsive/lessdurable; black p slowest to respond/most durable. Black arrowsindicate the order with which novel phenotypes appear. Undermost circumstances, novel environments or behaviors first inducenovel phenotypes via reversible homeostatic processes that maybe replaced by more durably accommodated phenotypes via de-velopmental and intergenerational plasticity. If the changed con-ditions are stable for sufficient generations, natural selection maygradually fix the phenotype or reduce the costs associated withproducing it. In this way, highly plastic traits can “lead the way”and accelerate the pace of genetic change.

low-mortality offspring (Kramer 2010; Lee 2003; Wells andStock 2007).

The Timescales of Human Adaptation and theRole of Developmental Plasticity

While this strategy is characteristic of the human life historygenerally, nutritional and mortality conditions can vary widelyacross environments and through time. Although genetic ad-aptation by natural selection helps explain the durability ofspecies-level characteristics that differentiate us from othergreat apes, the transience of many of the ecological challengesthat populations confront may not be dealt with effectivelyby changes in gene frequencies, which require many gener-ations and hundreds if not thousands of years to accrue inthe gene pool. As such, population variation in life historyparameters is largely traceable to developmental plasticity.West-Eberhard (2003:33) defines “phenotypic plasticity” as“the ability of an organism to react to an internal or externalenvironmental input with a change in form, state, movementor rate of activity.” “Developmental plasticity” refers to thatsubset of plastic phenotypic responses that involve irreversiblemodifications to growth and development. Plasticity is un-derstood as being undergirded by a genetic architecture thatallows context-dependent trait expression in response to vary-ing environmental experiences and behaviors (McIntyre andKacerosky 2011; Stearns 1992; Stearns and Koella 1986).

Biological anthropologists have long emphasized nonge-netic means of adapting to environmental challenges, and thiswork has highlighted the importance of developmental plas-ticity as a key mode of human adaptation (fig. 1; Frisancho1977; Kuzawa 2005, 2008; Lasker 1969). The body copes withthe most rapid ecological fluctuations using rapid, self-cor-recting, and reversible homeostatic systems that respond tochanges or perturbations in a way that offsets, minimizes, orcorrects deviations from an initial state. Homeostatic systemscan be viewed as buffering the effects of environmental fluc-tuations to maintain approximate internal stability. A subsetof homeostatic processes (Sterling 2004) also have an antic-ipatory component and a capacity for gradual resetting oftarget state and regulatory set points described as “allostasis.”

When environmental trends are too chronic to be efficientlybuffered by homeostasis or allostasis and yet not chronicenough for substantial genetic change to consolidate around,the flexible capacities of such systems may be overrun. It iseasy to see how a sustained environmental change might over-load a homeostatic system. As an example, consider an in-dividual who moves to high altitude where oxygen saturationis low. Initial physiologic responses will include an elevatedheart rate, which increases the volume of blood and thus thenumber of oxygen-binding red blood cells that pass throughthe lungs. By engaging a homeostatic system—heart rate—the body has found a temporary fix. However, if heart rateis already high under resting conditions, there is less leewayto deal with new challenges that might require a further in-

crease in heart rate, such as running from a predator. Ho-meostatic changes may work as short-term solutions, but theyare a poor means of coping with a condition such as high-altitude hypoxia if this is the new baseline environmental state.

This is where the value of developmental plasticity becomesclear. Individuals raised at high altitude have a more efficientstrategy for coping with low oxygen availability, for they sim-ply grow larger lungs during childhood (Frisancho 1977). Thisis an example of how developmental plasticity allows organ-isms to adjust biological structure on timescales too rapid tobe dealt with through genetic change but too chronic to beefficiently buffered by homeostasis. Other classic examples ofexperience-driven plasticity include the development of theskeletal system (Pearson and Lieberman 2004), the centralnervous system (Edelman 1993), and the immune system(McDade 2003).

Intergenerational Phenotypic Inertia

There is a growing list of biological systems that are notmodified in response to the environment itself but to hor-monal or nutrient signals or cues of past environments asexperienced by ancestors, most typically the mother (Bateson2001; Gluckman and Hanson 2004; Kuzawa 2001, 2005; Wells2007; Worthman 1999). Brief “critical” or “sensitive” periodsin early development often overlap with ages of direct nu-trient, hormone, or behavioral dependence on the mother(e.g., via placenta, breast milk, or emotional attachment),which facilitate the transfer of integrated cues of past maternalor matrilineal experience (Kuzawa and Quinn 2009). Thesenewer examples of early life developmental plasticity are thusdistinct from conventional plasticity in the time depth and


Figure 2. Value of intergenerational averaging as a way to identifya trend in a noisy signal, in this case representing availability ofa hypothetical ecological resource. The two lines are runningaverages calculated across 10 time units (thin line) and 100 timeunits (dark line). As the window of averaging increases, an un-derlying long-term trend is uncovered. Transgenerational influ-ences of maternal and grandmaternal experience on fetal andinfant biology (inertia) may help achieve a similar feat. (FromKuzawa and Quinn 2009, with permission.)

stability of information to which the developing body re-sponds.

The tendency for plasticity to respond to parental nutrient,hormonal, or behavioral cues that integrate past environ-mental experience has been defined as “phenotypic inertia”(Kuzawa 2005, 2008). One possible explanation for the utilityof such effects is that offspring are calibrating growth andnutrient expenditure to the mother’s capacity and willingnessto invest in the present offspring, which reflect her own phe-notypically embodied nutritional history (Wells 2003, 2007).From an alternative perspective, these intergenerational effectscould calibrate offspring growth, metabolism, and physiolog-ical settings to a stable “running average” index of conditionsexperienced by recent matrilineal ancestors that serve as a“best guess” of conditions likely to be experienced in thefuture (fig. 2; Kuzawa 2005). By this reasoning, intergener-ational effects of matrilineal experience (and possibly alsopatrilineal experience via germ-line epigenetic inheritance; seeEisenberg, Hayes, and Kuzawa 2012; Pembrey 2010) couldextend the utility of developmental plasticity as a mode ofadaptation by allowing organisms to track more integratedand stable trends occurring across a multigenerational time-scale (for more, see Kuzawa 2005, 2008; Kuzawa and Quinn2009; Kuzawa and Thayer 2011).

Thus, the type of organismal response to ecological chal-lenge or novelty depends not only on the type of stressor orexperience but also its stability and duration. Selection hasshaped biology to respond to such experiences in a myriadof ways; as such, adaptation does not simply occur throughevolution by natural selection and alteration of gene fre-quencies but also through homeostasis, allostasis, phenotypic/developmental plasticity, and phenotypic inertia across gen-erations (fig. 1). As life history characteristics take manyforms, from behavioral to physiological to molecular (Hilland Hurtado 1996), we should expect a similar breadth ofmechanisms and timescales underlying life history evolutionand adaptation. As one example, mounting evidence suggeststhat many key life history traits are among those most stronglyshaped by early life experiences and that they may respondto signals conveyed by the mother across the placenta or viabreast milk (Kuzawa 2007; Kuzawa and Pike 2005; Kuzawaand Quinn 2009; Wells 2003). Continuing in this vein, weargue that adaptive phenotypic plasticity has a central placein explaining variation in life history traits among contem-porary human populations.

Variation in Modern Human Life HistoriesTraces Primarily to Environment-Driven Plasticity

We have reviewed the key derived characteristics of the humanlife history strategy. We have also considered the need for acapacity to modify priorities “on the fly” as an importantdimension of life history strategy for most organisms in lightof changing nutritional and demographic/mortality condi-

tions. Consistent with this expectation, most human life his-tory traits exhibit extensive sensitivity to ecological context(see table 1). Here we summarize the role of nutrition andcues of environmental risk as influences on the key life historyparameters of growth trajectory, adult size, and reproductivestrategy.

Somatic Growth and Adult Body Size

Growth during the postnatal period can be divided into sev-eral periods of distinct hormonal regulation that vary in sen-sitivity to environmental influence and that ultimately deter-mine age-specific contributions to adult size and sexualdimorphism (Karlberg 1989). Roughly the first two postnatalyears reflect a continuation of a growth regime begun in utero.At this age, production of insulin-like growth factors thatstimulate skeletal and somatic growth is insulin dependent,tying growth rate directly to nutritional intake and nutritionalsufficiency. Prenatal evidence for this effect comes from arecent large study showing that the level of a woman’s fastingglucose during pregnancy is a robust predictor of the size ofher offspring at birth, illustrating that fetal glucose supplydrives fetal growth (Metzger et al. 2009). After birth, exclusivebreast-feeding sustains the infant’s nutritional requirementsfor about the first 6 months of life, after which complementaryfoods must be introduced to avoid growth faltering (Sellen2006). While breast-fed, infants obtain balanced nutritionalresources to support growth and passive maternal immuneprotection, which minimizes the burden of energetically costlyinfections. The weaning transition often introduces nutri-tional stress as these resources are replaced with less balancedand less sterile complementary foods. As such, infancy is often

Table 1. Range of variation for key life history characteristics in modern human populations

Males Females

Characteristic Range Degree of plasticity Range Degree of plasticity Environmental influences

Height (cm):Birth (length) 49.0a–51.9b � 48.1a–52.2b � Mother’s nutrition before pregnancy increases; stress, infection decreaseWeaning (3 years) 82.0a–99.0b �� 81.0a–98.6b �� Exclusive breast-feeding until 6 months protects; introduction of complementary foods,

and infection lead to falteringMidchildhood (6 years) 97.0b–120.8b �� 96.8a–120.2b �� Deficits from early growth and poor environmental or nutritional milieu reduce statureAdult 144.4b–181.6a �� 135.8a–168.2b � Low birth weight and poor nutrition/growth before 3 years of age reduce stature

Weight (kg):Birth 2.4a–3.57a �� 2.50b–3.5a ��Weaning (3 years) 10.3a–16.1b �� 10.1a–16.1a ��Midchildhood (6 years) 14.5b–23.3b �� 13.8b–23.4b ��Adult 40.0a–88.1b �� 37.0a–87.4b ��

Age at maturity (years):Menarcheal age 12.1b–18.4a �� Poor growth/nutrition delay; prenatal stress/childhood psychosocial stress and abun-

dant nutrition accelerateAge at peak height velocityc 12.0–17.0 �� Fast postnatal/infancy growth acceleratesAge at first birthc 20.9–37.0 16.2–25.0 Stress or cues of extrinsic mortality acceleratesAge at menopaused 48.2–52.6 �� Higher parity, longer cycles delay menopause; some evidence that smoking, small size

at birth, poor early growth lead to earlier menopauseLife span:e

e0 21–37 21–37 Higher infant mortality and rates of infection and violence/accidents decreasee15 28.6–42.5 28.6–42.5 Reduced exposure to environmental and health stressors increasese45 13.7–24.2 13.7–24.2

a Eveleth and Tanner 1976.b Eveleth and Tanner 1990.c Walker et al. 2006a.d Leidy Sievert 2006.e Gurven and Kaplan 2007; ex is life expectancy at age x.


an age of nutritional stress with a high-mortality burden. This,combined with the need to buffer an unusually large andinflexible cerebral energy need, may help explain the heavyhuman investment in deposition of protective fat stores beforebirth and during the first 6 months of postnatal life (Kuzawa1998).

Because this difficult transition coincides with the age ofinsulin-dependent, nutrition-driven growth, weaning-relatedgrowth deficits can carry into adulthood to influence finalstature and body weight. Indeed, the magnitude of adultheight deficits relative to healthy reference data has beenshown to trace largely to growth faltering already present at2 or 3 years of age (Billewicz and McGregor 1981; Martorell1995), and much of the contemporary population variationin adult standing height is believed to reflect the effect ofnutrition and hygiene during infancy and early childhood(Eveleth and Tanner 1976, 1990; Habicht et al. 1974; Victoraet al. 2008).

Although growth rate remains sensitive to nutrition duringthe entire period of growth and development, long-term ef-fects of nutrition on adult size diminish after infancy andearly childhood as insulin-dependent growth is gradually re-placed by a growth-hormone-regulated growth regime (Karl-berg 1989). During childhood and puberty, nutritional deficitsprimarily slow the pace of maturity without affecting finalstature or body size. During the pubertal growth spurt (seebelow), onset of gonadal production of sex steroids increasesgrowth rate, especially in males. However, as with childhoodgrowth, there is little evidence for lasting effects of nutritionduring adolescence on final adult size. Generally, individualswho are better nourished or have more abundant fat storesduring childhood enter puberty earlier, experience a moreintense but briefer period of heightened pubertal growth, andattain maturity at a younger age (Tanner 1962). As nutritionalconditions deteriorate, onset of pubertal growth is delayed,and the spurt is also protracted such that growth velocitiesare slower but spread across a longer period. Collectively, thesefindings show that nutrition primarily influences adult sizeby influencing growth attainment during fetal life, infancy,and early childhood, when nutritional resources are derivedprimarily from the mother’s body, thus linking adult size inthe present generation with matrilineal nutritional history(Kuzawa 2005, 2007; Kuzawa and Quinn 2009; Wells 2007).

Although nutrition is clearly a powerful influence on hu-man variation in growth rate and adult size, it is worth notingthat important genetic contributions to population variationin stature are especially likely at the extremes. For instance,the shortest populations in the world are “Pygmy” popula-tions such as the Efe or Mbuti, whose atypically short staturelikely has at least a partial genetic component that reflectsconvergent genetic selection in response to ecological or mor-tality conditions common in rainforest environments (Perryand Dominy 2009; Pickrell et al. 2009; Walker et al. 2006a).Similarly, the tall mean stature of the tallest human groups,such as the Rift Valley pastoralist populations, may have a

partial genetic explanation (Gray, Wiebusch, and Akol 2004).However, these extremes aside, environmental factors, espe-cially differences in nutrition and pathogen burden duringthe first 2–3 years of life, are recognized as the primary driverof variation in mean adult size across populations (Evelethand Tanner 1976, 1990).

Nutrition as an Influence on Age at Reproductive Maturity

Age at maturity represents an important life history transitionbecause it marks the age at which the body shunts energypreviously allocated to somatic growth into reproduction(Charnov 1993; Kuzawa 2007). Although large adult size car-ries reproductive benefits, the ability to sustain the nutritionalrequirements of fast growth and the risk of preadult mortalitythat determines how long it is prudent to delay maturing areboth variable. Thus, it is expected that maturational tempowill follow a gradient of developmental plasticity (i.e., a re-action norm) that is sensitive to availability of nutritionalresources and also to cues reflecting the level of mortality risk(Coall and Chisholm 2003; Ellis et al. 2009; Stearns and Koella1986; Walker et al. 2006a). Consistent with this expectation,maturational tempo is among the most variable of humanlife history characteristics and exhibits sensitivity to both nu-tritional and psychosocial stressors.

The role of nutrition as a driver of childhood and adoles-cent growth helps explain the extreme environmental sensi-tivity of pubertal timing. Population means for menarchealage range from about 12 to 18 years (table 1). Rapid multi-generational secular trends clearly show that this variabilitylargely reflects environmentally driven plasticity in matura-tional tempo. As noted above, in some Western European andScandinavian countries, menarcheal age declined from around17 to 18 years in the mid-nineteenth century to present pop-ulation means of 12 or 13 years (Eveleth and Tanner 1976,1990). More recent studies in non-European populations alsodemonstrate rapid declines in menarcheal age. For instance,in South Korea, age at menarche declined from 17 years in1920 to 12–14 years in 1985, representing a rate of declineof 0.68 years/decade (Cho et al. 2009), while age at menarchedeclined at a similar rate of 0.65 years/decade from 1989 to2008 in a rural Gambian population (Prentice et al. 2010).

Striking evidence for the sensitivity of menarcheal timingto environmental influence is illustrated in growth studies ofgirls adopted from orphanages in India or Bangladesh intohigh-income Scandinavian households. These studies docu-ment relatively high rates of precocious puberty with adopteesentering puberty as early as 7 years of age (Proos, Hofvander,and Tuvemo 1991; Teilmann et al. 2006). Intriguingly, thedegree to which maturation was sped up in these girls de-pended on their age of adoption: girls adopted at older ages,who therefore spent more time in less favorable conditions,entered puberty earliest upon environmental improvement(Proos, Hofvander, and Tuvemo 1991). Such findings suggestthat an individual’s developmental response to environmental


factors such as nutrition may itself be contingent upon priordevelopmental conditions experienced during early life. Ad-ditional evidence for such “programming” effects of earlyexperiences comes from the finding that being born small—often a result of fetal nutritional deficit or maternal stressduring or before pregnancy—predicts earlier maturity espe-cially when small birth size is followed by rapid catch-upgrowth after birth (Adair 2001; Ibanez et al. 2000; Karaolis-Danckert et al. 2009; Ong et al. 2009).

The lack of an easily measured maturational marker inmales comparable to the onset of menses has constrainedunderstanding of both the extent of variability in male pu-bertal timing and sensitivities of male maturational tempo toenvironmental change. In the populations for which data areavailable from Walker et al.’s (2006a) tabulation of growthrates in small-scale societies, age at peak height velocity (aproxy for pubertal timing) ranges from 12 to 17 years (table1). However, few studies have investigated the environmentalor nutritional factors that predict variability in male matu-rational tempo. A recent study conducted in Germany foundthat in accordance with the effect of birth weight and earlygrowth among females, males who were born small and gainedweight rapidly from birth to 2 years experienced peak heightvelocity earlier (Karaolis-Danckert et al. 2009). Similarly, arecent study in a longitudinal birth cohort in the Philippinesreported that males who experienced rapid weight gain im-mediately after birth reached puberty earlier (Kuzawa et al.2010).

Nutrition, Developmental Plasticity, and the Originsof Sexual Dimorphism

In species marked by sexual size dimorphism, the greater sizeof males leads to correspondingly higher nutritional require-ments. This is believed to help explain why males tend toexhibit greater responses, both positive and negative, tochanges in nutritional conditions (Stinson 1985). As a resultof this differential sensitivity, the magnitude of sexual di-morphism in traits such as body size will tend to shift asprevailing nutritional conditions change. For instance, ba-boons that self-provisioned off of trash dumps were foundto weigh 50% more than wild-fed baboons (Altmann et al.1993). The magnitude of the weight gain was much greaterin the males than in the females, leading to an increase inadult size dimorphism (Altmann and Alberts 2005). In humanpopulations, adult size in males has similarly been shown tobe more sensitive to changes in socioeconomic condition ornutritional abundance (Stinson 1985).

Environmental experiences early in the life cycle may bekey to the establishment of these differences. Before birth andduring the first 6 months of postnatal life, testosterone pro-duction is temporarily high in males, which has long-term“organizational” effects on male reproductive biology, behav-ior, and body growth (Jost 1961; Phoenix et al. 1959). Recentevidence points to these early periods of hormonally driven

organizational effects as potential sources of plasticity in thepattern and degree of biological and behavioral differencebetween males and females. In a well-characterized birth co-hort in the Philippines, males who grew rapidly during theage of high postnatal testosterone but not at other early agesgained weight and height faster, matured earlier, and weretaller and more muscular as adults (Kuzawa et al. 2010). Theserelationships were greatly reduced or not present for mostoutcomes among same-aged females. Men who grew rapidlyafter birth also reported an earlier age at first sex, more life-time sex partners, and more recent sexual activity. Becausethese men also showed evidence for greater adult testicularsensitivity to luteinizing hormone and higher testosterone lev-els, the authors speculated that nutritional experiences duringthe early postnatal critical period might have lasting effectson the magnitude of physical and behavioral differences be-tween males and females. It is notable that male body sizeand related energetic costs were reduced in response to earlylife cues reflecting reduced nutrition, perhaps indicating acapacity to calibrate life history and energetic expenditure asprevailing nutritional conditions change. Sex-specific sensi-tivities of growth rate and developmental processes suggestthat any environmental changes that influence food avail-ability could have differential effects on males and femalesthat could thereby influence the pattern of sexual dimorphismwithin and between populations.

Psychosocial Stress, Maturational Tempo, andReproductive Scheduling

Although the public health and growth and development lit-eratures have traditionally focused on the role of nutritionand hygiene as influences on growth and adult size, morerecent work is showing that psychosocial stress can also in-fluence growth rate and maturational tempo. One link be-tween stress and growth stems from the energy burden of thestress response itself, which may compete with growth, leadingto a reduced growth rate (Nyberg et al., forthcoming).

In addition to such direct resource trade-offs, stress mayalso serve as a barometer of nonnutritional risks, such asunavoidable mortality, which is recognized as shaping theoptimal timing of maturity and reproductive scheduling inmodels of mammalian life history evolution (Charnov 1993;Promislow and Harvey 1990). Building from this premise, along-standing research tradition in developmental psychologyand anthropology has developed evolutionary explanationsfor the sensitivity of maturational tempo to parental or othersocial cues (Belsky, Steinberg, and Draper 1991; Chisholm1993; Draper and Harpending 1982). In these studies, theobservation that girls from harsh or unstable family environ-ments tend to mature earlier is interpreted as evidence thathuman maturational tempo is responsive to cues of extrinsicmortality risk as reflected in attachment quality and parentalinvestment (Chisholm 1993; Ellis 2004; Ellis et al. 2009). Nu-merous studies have tested this and related predictions (Ellis


and Garber 2000; Hulanicka, Gronkiewicz, and Koniarek2001; Pesonen et al. 2008; Tither and Ellis 2008). For instance,Chisholm et al. (2005) found that retrospectively reportedtotal life stress explained about 11% of the variance in me-narcheal age in a sample of college-aged women, with higherstress levels predicting earlier maturity. Quinlan (2003) sim-ilarly found that women whose parents separated before theywere 6 years old matured earlier than girls whose parents didnot separate. The effect size in these studies is often on theorder of 1–2 months (e.g., Belsky et al. 2010), which is quitesmall, especially when compared with the large multigener-ational trends in menarcheal age documented in associationwith nutritional improvements.

Cues of extrinsic mortality may better explain variation inreproductive scheduling and the intensity of parental invest-ment. Age at first reproduction varies widely in the data com-piled by Walker et al. (2006a; table 1), and there is a largeand growing literature showing that children exposed to highextrinsic mortality or low parental investment during earlylife not only mature earlier but also start reproducing at ayounger age (Burton 1990; Chisholm et al. 2005; Low et al.2008; Nettle, Coall, and Dickins 2010). Nettle (2010) foundthat across neighborhoods in England, women in the mostsocioeconomically deprived communities gave birth for thefirst time an average of 8 years earlier than women in themost affluent communities, paralleling Wilson and Daly’s(1997) finding of earlier and more intensive reproductivescheduling in Chicago neighborhoods with the highest ho-micide rates. In another study, Nettle, Coall, and Dickins(2011) found that prolonged maternal absence, low paternalinvestment, and many residential moves during childhoodwere independent predictors of earlier age at first birth. Eachof these stressors lowered the age at first reproduction byabout half a year, and their effects were additive.

There is also evidence that individuals who experiencestressful environments during childhood invest less in theirown offspring (Ellis et al. 2009; Hurtado et al. 2006). Forinstance, parental investment is reduced under conditions ofharsh, unavoidable stressors such as pathogen loads, famine,or warfare (Quinlan 2007). Such reduced investment neednot be solely behavioral: associations between low socioeco-nomic status and birth weight have been documented andinterpreted in terms of reduced investment in offspring (Coalland Chisholm 2003, 2010; Nettle 2010). Conversely, in high-opportunity/low-mortality populations, age at first reproduc-tion is typically delayed as a response to lowered mortalityrates and higher costs and future payoffs of investing in off-spring (Low et al. 2008). Indeed, in many European countries,age at first birth now routinely occurs in the 30s (ESHRECapri Workshop Group 2010) even as age at biological ma-turity has declined because of nutritional and energetic im-provements.

Anthropologists have criticized the focus on the nuclearfamily as the presumed unit of human child rearing in muchof this work (Chisholm et al. 2005; Hrdy 2009). Instead, hu-

mans can be described as a cooperatively breeding species inwhich reproducing females rely on flexible patterns of allo-parental care as a fundamental component of their repro-ductive strategy (Hill and Hurtado 2009; Hrdy 2005, 2009;Kramer and Ellison 2010). Moreover, complicating the tra-ditional emphasis on the presumed primary role of men ashunters and providers of calories, recent work highlights evi-dence for derived neuroendocrine adaptations specific to thehuman lineage that encourage direct male care of offspring(Gettler et al. 2011). In light of these revelations, it seemslikely that the developmental capacities for facultative ad-justment of life history have been shaped to be sensitive toa much broader range of social cues than previously theorized.For instance, while transplacental nutrients or maternal breastmilk are important sources of maternal ecological informationbefore birth and during infancy (Kuzawa and Quinn 2009),after 6 months of age, an increasing proportion of energeticneeds are met by complementary feeding of specially preparedfoods that may be provided by various relatives or groupmembers (Bogin and Smith 1996; Sellen 2006, 2007). To theextent that nutrition during infancy helps calibrate growthand reproductive expenditure, in humans these trajectoriesmay be as much reflective of availability of alloparental careas any direct maternal metabolic investment. This exampleillustrates how theories concerning the role of stress and socialcues in life history scheduling and resource allocation needto address the fact that human children are often highly relianton investment from individuals other than biological parents.

Summary: Phenotypic Responses to ChangingEcological Conditions

In figure 3, we summarize the pattern of life history traitsthat theory predicts in environments with different combi-nations of nutritional sufficiency and unavoidable mortalityrisk. At one extreme are the most-favorable environments inwhich nutritional sufficiency is high, allowing fast growth andthe attainment of a large adult size despite relatively earlymaturity (fig. 3A). In this example, early maturity is secondaryto favorable nutrition and a consequent fast growth rate.When populations experience energy sufficiency but cues ofhigh extrinsic risk, they are expected to grow quickly but alsomature early and thus are predicted to be slightly smaller asadults (fig. 3B). In contrast, low-nutrition and low-risk en-vironments are expected to lead to growth that is sufficientlyslow that despite a compensatory delay in maturity, individ-uals still attain a shorter adult size (fig. 3C). Finally, anothertheoretical extreme is represented by a combination of lownutritional resources and high extrinsic risk, which shouldlead to relatively early maturation at a small size (fig. 3D).

Mortality risk influences optimism about surviving into thefuture to reproduce, and as such there are predicted to betandem shifts in relative allocation between maintenance andreproduction and also in the level of investment in each off-spring. Theory predicts that as unavoidable mortality in-


Figure 3. Summary of developmentally plastic life history changes predicted by different combinations of nutritional sufficiencyand cues of threat or extrinsic mortality risk. Comparison group for depiction of birth weight, maturational tempo, and adultstature p high nutrition/low extrinsic mortality risk group (see text for discussion). Not drawn to scale.

creases, not only is maturity sped up but also age at first birth,which is accompanied by an increase in fertility rate relatedin part to a reduction in the interbirth interval and reductionsin offspring investment, size, and survival (Belsky, Steinberg,and Draper 1991; Chisholm 1993; Ellis et al. 2009; Nettle,Coall, and Dickins 2011). This plasticity in both the rate ofmetabolic expenditure and in the scheduling of developmentaland reproductive events helps explain much of the variationin growth rate, body size, and reproduction across contem-porary human populations.

Discussion: Summary and Implicationsof Life History Plasticity

Organisms must manage the costs of building a body and thebodies of offspring while calibrating the scheduling of de-velopmental and reproductive events in response to nutri-tional resources, the risk of mortality, and other changing orunpredictable ecological conditions (Charnov 1993; Prom-islow and Harvey 1990; Stearns 1992). It is this environmentalvariability that necessitates plasticity in growth, development,reproduction, and other components of a species’ life historystrategy. Consistent with this perspective, our review high-lights the overwhelming importance of environmentallydriven developmental plasticity as a source of contemporaryhuman life history variation.

Extrapolating from these findings in modern human pop-

ulations, developmental plasticity has likely been an importantinfluence on the phenotypes of earlier members of the hom-inin lineage. As the ancestors of contemporary human pop-ulations spread across Africa and eventually Eurasia, the localconditions that they faced would have varied in energetic anddemographic conditions. Regional or population differencesin such parameters as growth rate, maturational tempo, bodysize, and sexual dimorphism would have likely arisen via plas-ticity as local or regional populations were confronted by theseecological differences. While at present it is impossible toknow whether the same range of plasticity in life history re-sponses was present in ancestral hominins, it is likely thatplastic responses were at least similar in kind if not degree.For instance, it is probable that hominin populations thatexperienced nutritional abundance grew faster and maturedearlier at larger body sizes. Similarly, populations in contextswith relatively easy-to-exploit transitional foods may haveweaned earlier than other groups, leading to the achievementof higher fertility rates. It is also interesting to consider thatmales were likely more strongly influenced by improvementsin energetic conditions than were females, which could in-crease sexual size dimorphism in such environments. Con-versely, in energetically marginal circumstances, maturationwas likely delayed, dimorphism reduced, and body size andgrowth rates diminished. Building on this theme, we concludeby considering a final question with central importance toattempts to reconstruct hominin evolution and to interpret


the fossil record: what is the relationship between a trait’splasticity and its genetic evolution via natural selection?

Did Plasticity Lead the Way inHuman Evolution?

While biological anthropologists have long considered the roleof phenotypic and developmental plasticity as a means ofadaptation and a source of within-species variation, recenttheoretical and empirical work in evolutionary biology hasemphasized the role of plasticity as an influence on the paceof evolutionary change and speciation (Kuzawa 2012; West-Eberhard 2003). Classically, it was often assumed that plas-ticity decouples genotypes from the specifics of phenotypes,which could reduce the strength of selection on underlyinggenetic variants (see Ghalambor et al. 2007). Contrary to thisperspective, studies have shown that at least in some instances,the most phenotypically and developmentally plastic traits canevolve most rapidly (Stearns 1983; Wund et al. 2008). Forexample, after it was introduced to Hawaii, evolution of themosquito fish has been most rapid for traits that exhibitedthe greatest plasticity (Stearns 1983). There is similar evidencethat the diversification of fish species in the Canadian lakescreated after the retreat of the Laurentide Ice Sheet was fa-cilitated by plasticity (Robinson and Parsons 2002). Recentevolutionary theory is providing insights into how plasticitycan accelerate rather than dampen the pace of genetic ad-aptation (West-Eberhard 2003): when novel environmentsfirst induce phenotypes via developmental plasticity, plasticityserves as the source of raw phenotypic variability on whichnatural selection then acts to shape subsequent genetic ad-aptation.

For phenotypic plasticity to “lead the way” and facilitategenetic evolution, several steps must occur, with a typicalscenario involving the following: (1) an organism or popu-lation moves into a novel environment or experiences achange in an existing environment, (2) plasticity facilitatesaccommodation to the novel conditions by improving the“fit” between phenotype and environment, (3) the geneticarchitecture of this newly expressed phenotypic variation isthen modified by natural selection to improve on the initiallyplastic phenotype or to increase the efficiency with which thephenotype is produced.

A growing list of studies provides evidence that plasticitycan serve as an important source of phenotypic alternativesthat are then filtered by natural selection and stabilizedthrough genetic change. In an experiment designed to assessthe degree of developmental plasticity of jaw morphology(Aubret and Shine 2009), tiger snakes from populations re-cently introduced to island environments (where larger jawsizes are favored because of larger prey) showed a greatercapacity for plasticity to match head growth to prey size. Snakepopulations with longer histories on the island grew largerhead sizes regardless of the size of prey consumed duringgrowth and development, showing that the initial capacity for

plasticity was replaced by trait fixation after many generationsof consistent selection for larger jaws. Other studies illustratehow developmental mechanisms underlying such adaptivelyplastic phenotypes can provide the substrate for later speciesdivergence following longer periods of niche specialization.Among species of spadefoot toads, between-species diversitylargely traces to ancestral larval plasticity in response to cli-mactic conditions (Gomez-Mestre and Buchholz 2006). Sim-ilarly, a study of developmental morphology among three-spined sticklebacks found that the phenotypes of divergentfreshwater species mirrored patterns of environmentally in-duced developmental variation in the marine species, an ex-tant representative of the ancestor to the freshwater species,suggesting that plasticity provided the developmental alter-natives from which the various freshwater species diverged asthey moved into novel environments with different prey types(Wund et al. 2008).

Each of these examples illustrates how species diversity canoriginate from ancestrally shared patterns of plasticity amongpopulations exposed to distinct environmental conditions.This likely reflects gradual genetic improvement of the in-duced phenotype via natural selection, which can take theform of fixation (loss of plasticity), as a shift in the underlyinggenetic architecture of the reaction norm (Ghalambor et al.2007; Price, Qvarnstrom, and Irwin 2003), or as a compen-sation for some of the physiologic or other costs associatedwith the induced phenotype (Storz, Scott, and Cheviron2010).

By this reasoning, the various modes of adaptation (fig. 1)do not simply cover different timescales of variability but mayalso represent a sequence of evolutionary change with lessdurable nongenomic modes of biological adaptation allowingappropriate phenotypic adjustments to novel conditionswhich are eventually superseded by more durable and efficientgenetic accommodation (arrows, fig. 1; West-Eberhard 2003).Indeed, this idea that plasticity-induced phenotypic variantsallow organisms to cope with environmental and behavioralnovelty and that more durable genetic change might only laterand more gradually follow has a long history (Baldwin 1896;Waddington 1953) and has recently been the focus of rekin-dled attention among evolutionary biologists (Sarkar 1999;West-Eberhard 2003).

Some examples of plasticity “leading the way” are intuitive.For instance, the biomechanical sensitivity of skeletal devel-opment shows that behavioral change is accompanied bychanges in trabecular alignment and diaphyseal robustness(Pearson and Lieberman 2004). The capacity for more dra-matic plasticity-induced realignment of musculoskeletal ele-ments is illustrated by examples of quadrupedal animals bornwithout forelegs that facultatively adopt bipedal locomotion(West-Eberhard 2003:51–54). Similarly, Japanese macaquestrained to perform by walking upright exhibit a humanlikegait (Hirasaki et al. 2004), which is facilitated in part by similarchanges in bone morphology (Nakatsukasa and Hayama 1991,cited in Hirasaki et al. 2004). In light of this work, it seems


likely that the gradual adoption of locomotor changes amongearly hominins was first a behavioral innovation that led toplasticity-based developmental changes in skeletal morphol-ogy (Hirasaki et al. 2004) that were gradually and incremen-tally fixed by genetic evolution (West-Eberhard 2005).

We noted that developmental adaptation in lung volumeallows populations raised at high altitude to cope with hypoxiawithout having to mobilize homeostatic responses. While de-velopmental adaptation of this type is well documented (Fri-sancho 1977; Moore, Niermeyer, and Zamudio 1998), pop-ulations with long histories living at high altitude showevidence for genetic adaptation to hypoxia (Beall 2007). Someof these adjustments appear to compensate for some of thecosts associated with plasticity-induced phenotypes (Storz,Scott, and Cheviron 2010), demonstrating how developmentaladaptation is not only engaged before genetic change but maylead the way for gradual fixation of more durable geneticadaptations.

Extrapolating from these cases, we speculate that similarprinciples may apply to the evolution of human life historytraits. As one example, take the human Pygmy phenotypefound in populations inhabiting environments characterizedby low nutritional sufficiency (Shea and Bailey 1996) and highunavoidable mortality (Migliano and Guillon 2012; Migliano,Vinicius, and Lahr 2007; Walker et al. 2006a). Because a slowgrowth rate is a well-described response to low nutritionalavailability (Eveleth and Tanner 1990) while early maturitymay be driven by cues signaling high unavoidable mortalityrisk (Ellis et al. 2009), features of this phenotype were likelyinduced first by developmental plasticity within individualswho moved into these ecologies. Evidence for genetic con-tributions to short stature in these populations (Pickrell et al.2009) suggests that phenotypes induced by plasticity wereeventually accommodated by selection operating on novelgrowth-regulating mutations.

If some plastic traits have the potential to evolve mostrapidly under selection pressure (Stearns 1983; West-Eberhard2003), it is interesting to consider the role of development inthe evolution of the derived plastic features of the human lifehistory. Take, for instance, human longevity (Gurven andKaplan 2007; Hawkes et al. 1998; Kaplan et al. 2000). Thedelay in senescent processes that extend the human life spanby several decades beyond that of other extant great apes likelyrequired increasing maintenance expenditures (Hawkes 2003;Kaplan, Lancaster, and Robson 2003; Kaplan et al. 2000) thatalmost certainly required reduced expenditure in other do-mains, such as reproduction (Hawkes et al. 1998; Kirkwoodand Rose 1991). In humans, any reduction in maternal re-productive effort is likely facilitated by early weaning of de-pendent infants from maternal metabolic investment (breastmilk) and the early introduction of specially prepared foodsoften provided by alloparents (Sellen 2006, 2007).

Just as the impetus to adopt bipedal locomotion was almostsurely behavioral before genetic adaptation could graduallyoccur, so too was early weaning likely first a behavioral de-

cision. Early weaning facilitated by complementary feedinglikely “freed up” energy for increased maintenance expen-ditures, allowing investment in more durable and long-livedsomas. Consistent with this perspective, the longest humanlife spans tend to accompany reduced reproductive expen-diture as reflected in lower fertility rate or completed familysize (e.g., Doblhammer and Oeppen 2003; Gagnon et al. 2009;Jasienska 2009; but see Le Bourg 2007). By analogy, we spec-ulate that behaviorally driven decreases in reproductive effortfacilitated the initial expression of longer-lived phenotypes,made possible through increased maintenance effort, thatwere later genetically accommodated by natural selection asa species-defining trait over a longer, evolutionary time frame.This sequence of changes contrasts from scenarios positedpreviously in discussions of the adaptive evolution of humanlongevity, which often implicitly or explicitly assume that ge-netic changes that favor increased longevity would have ini-tiated this life history pattern (e.g., Hawkes et al. 1998).

The extensive developmental plasticity in human life his-tory traits reviewed here gains new importance in light ofevidence that plasticity can influence the direction and paceof evolutionary change. Many of the traits that differentiatethe human life history from that of other primates and greatapes—including slow growth rate, early weaning, delayed ma-turity, high fertility, and perhaps even long life span—dem-onstrate phenotypic variation that traces to developmentallyor behaviorally mediated plasticity in response to environ-mental factors such as nutrition and cues of unavoidable mor-tality. We have sketched some of the observations that leadus to hypothesize that this environmentally induced pheno-typic variation likely preceded and ultimately facilitated ge-netic adaptations that gradually stabilized the life history char-acteristics that help define our species. We hope that thisreview helps stimulate interest in the broader insights thatdevelopmental plasticity may provide into the diversificationand evolution of genus Homo, including the lineage that ledeventually to modern Homo sapiens.

Acknowledgments

We thank Leslie Aiello and Susan Anton for the invitation toparticipate in this stimulating conference and the Wenner-Gren Foundation for sponsoring it. J. M. Bragg was supportedby a National Science Foundation Graduate Research Fellow-ship.

References CitedAdair, Linda S. 2001. Size at birth predicts age at menarche. Pediatrics 107:

e59. http://pediatrics.aappublications.org/content/107/4/e59.Altmann, Jeanne, and Susan C. Alberts. 2005. Growth rates in a wild primate

population: ecological influences and maternal effects. Behavioral Ecologyand Sociobiology 57:490–501.

Altmann, Jeanne, Dale Schoeller, Stuart A. Altmann, Philip Muruthi, andRobert M. Sapolsky. 1993. Body size and fatness of free living baboonsreflect food availability and activity levels. American Journal of Primatology30:149–161.

http://pediatrics.aappublications.org/content/107/4/e59


Anton, Susan C. 2003. Natural history of Homo erectus. American Journal ofPhysical Anthropology 122:126–170.

Anton, Susan C., Fred Spoor, Connie D. Fellmann, and Carl C. Swisher III.2007. Defining Homo erectus: size considered. In Handbook of paleoanthro-pology, vol. 3. Winfried Henke, Ian Tattersall, and Thorolf Hardt, eds. Pp.1655–1693. New York: Springer.

Aubret, Fabien, and Richard Shine. 2009. Genetic assimilation and the post-colonization erosion of phenotypic plasticity in island tiger snakes. CurrentBiology 19:1932–1936.

Baldwin, James M. 1896. A new factor in evolution. American Naturalist 30:441–451.

Bateson, Patrick. 2001. Fetal experience and good adult design. InternationalJournal of Epidemiology 30:928–934.

Beall, Cynthia M. 2007. Two routes to functional adaptation: Tibetan andAndean high-altitude natives. Proceedings of the National Academy of Sciencesof the USA 104:8655–8660.

Belsky, Jay, Laurence Steinberg, and Patricia Draper. 1991. Childhood expe-rience, interpersonal development, and reproductive strategy: an evolu-tionary theory of socialization. Child Development 62:647–670.

Belsky, Jay, Laurence Steinberg, Renate M. Houts, and Bonnie L. Halpern-Felscher. 2010. The development of reproductive strategy in females. De-velopmental Psychology 46:120–128.

Billewicz, W. Z., and Ian A. McGregor. 1981. The demography of two WestAfrican (Gambian) villages, 1951–75. Journal of Biosocial Science 13:219–240.

Bogin, Barry. 1999. Evolutionary perspective on human growth. Annual Re-view of Anthropology 28:109–153.

Bogin, Barry, and B. Holly Smith. 1996. Evolution of the human life cycle.American Journal of Human Biology 8:703–716.

Burton, Linda. 1990. Teenage childbearing as an alternative life-course strategyin multigeneration black families. Human Nature 1:123–143.

Charnov, Eric L. 1993. Life history invariants: some explanations of symmetryin evolutionary ecology. Oxford: Oxford University Press.

Charnov, Eric L., and David Berrigan. 1993. Why do female primates havesuch long lifespans and so few babies? or life in the slow lane. EvolutionaryAnthropology: Issues, News, and Reviews 1:191–194.

Chisholm, James S. 1993. Death, hope, and sex: life-history theory and thedevelopment of reproductive strategies. Current Anthropology 34:1–12.

Chisholm, James S., Julie Quinlivan, Rodney Petersen, and David Coall. 2005.Early stress predicts age at menarche and first birth, adult attachment, andexpected lifespan. Human Nature 16:233–265.

Cho, Geum Joon, Hyun Tae Park, Jung Ho Shin, Jun Young Hur, Young TaeKim, Sun Haeng Kim, Kyu Wan Lee, and Tak Kim. 2009. Age at menarchein a Korean population: secular trends and influencing factors. EuropeanJournal of Pediatrics 169:89–94.

Coall, David A., and James S. Chisholm. 2003. Evolutionary perspectives onpregnancy: maternal age at menarche and infant birth weight. Social Scienceand Medicine 57:1771–1781.

———. 2010. Reproductive development and parental investment duringpregnancy: moderating influence of mother’s early environment. AmericanJournal of Human Biology 22:143–153.

Doblhammer, Gabriele, and Jim Oeppen. 2003. Reproduction and longevityamong the British peerage: the effect of frailty and health selection. Pro-ceedings of the Royal Society B 270:1541–1547.

Draper, Patricia, and Henry Harpending. 1982. Father absence and repro-ductive strategy: an evolutionary perspective. Journal of Anthropological Re-search 38:255–273.

Edelman, Gerald M. 1993. Neural Darwinism: selection and reentrant signalingin higher brain function. Neuron 10:115–125.

Eisenberg, Dan, M. Geoffrey Hayes, and Christopher W. Kuzawa. 2012. De-layed paternal age of reproduction in humans is associated with longertelomeres across two generations of descendants. Proceedings of the NationalAcademy of Sciences of the USA 109:10251–10256.

Ellis, Bruce J. 2004. Timing of pubertal maturation in girls: an integrated lifehistory approach. Psychological Bulletin 130:920–958.

Ellis, Bruce J., Aurelio Figueredo, Barbara Brumbach, and Gabriel Schlomer.2009. Fundamental dimensions of environmental risk. Human Nature 20:204–268.

Ellis, Bruce J., and Judy Garber. 2000. Psychosocial antecedents of variationin girls’ pubertal timing: maternal depression, stepfather presence, and mar-ital and family stress. Child Development 71:485–501.

ESHRE Capri Workshop Group. 2010. Europe the continent with the lowestfertility. Human Reproduction Update 16:590–602.

Eveleth, Phyllis B., and James M. Tanner. 1976. Worldwide variation in humangrowth. New York: Cambridge University Press.

———. 1990. Worldwide variation in human growth. 2nd edition. New York:Cambridge University Press.

Frisancho, A. Roberto. 1977. Developmental adaptation to high altitude hyp-oxia. International Journal of Biometeorology 21:135–146.

Gadgil, Madhav, and William H. Bossert. 1970. Life historical consequencesof natural selection. American Naturalist 104:1–24.

Gagnon, Alan, Ken R. Smith, Marc Tremblay, Helene Vezina, Paul-PhilippePare, and Bertrand Desjardins. 2009. Is there a trade-off between fertilityand longevity? a comparative study of women from three large historicaldatabases accounting for mortality selection. American Journal of HumanBiology 21:533–540.

Galdikas, Birute M. F., and James W. Wood. 1990. Birth spacing patterns inhumans and apes. American Journal of Physical Anthropology 83:185–191.

Gettler, Lee T., Thomas W. McDade, Alan B. Feranil, and Christopher W.Kuzawa. 2011. Longitudinal evidence that fatherhood decreases testosteronein human males. Proceedings of the National Academy of Sciences of the USA108:16194–16199.

Ghalambor, Cameron K., John K. McKay, Scott P. Carroll, and David N.Reznick. 2007. Adaptive versus non-adaptive phenotypic plasticity and thepotential for contemporary adaptation in new environments. FunctionalEcology 21:394–407.

Gluckman, Peter, and Mark A. Hanson. 2004. Living with the past: evolution,development, and patterns of disease. Science 305:1733–1736.

Gomez-Mestre, Ivan, and Daniel R. Buchholz. 2006. Developmental plasticitymirrors differences among taxa in spadefoot toads linking plasticity anddiversity. Proceedings of the National Academy of Sciences of the USA 103:19021–19026.

Gray, Sandra J., Brandi Wiebusch, and Helen A. Akol. 2004. Cross-sectionalgrowth of pastoralist Karimojong and Turkana children. American Journalof Physical Anthropology 125:193–202.

Gurven, Michael, and Hillard Kaplan. 2007. Longevity among hunter-gath-erers: a cross-cultural examination. Population and Development Review 33:321–365.

Gurven, Michael, and Robert Walker. 2006. Energetic demand of multipledependents and the evolution of slow human growth. Proceedings of theRoyal Society B 273:835–841.

Habicht, Jean-Pierre, Charles Yarbrough, Reynaldo Martorell, Robert M. Mal-ina, and Robert E. Klein. 1974. Height and weight standards for preschoolchildren: how relevant are ethnic differences in growth potential? Lancet303:611–615.

Hawkes, Kristen. 2003. Grandmothers and the evolution of human longevity.American Journal of Human Biology 15:380–400.


Hill, Kim. 1993. Life history theory and evolutionary anthropology. Evolu-tionary Anthropology 2:78–88.

Hill, Kim, and A. Magdalena Hurtado. 1996. Ache life history: the ecology anddemography of a foraging people. Hawthorne, NY: Aldine de Gruyter.

———. 2009. Cooperative breeding in South American hunter-gatherers.Proceedings of the Royal Society B 276:3863–3870.

Hill, Kim, and Hillard Kaplan. 1999. Life history traits in humans: theory andempirical studies. Annual Review of Anthropology 28:397–430.

Hirasaki, Eishi, Naomichi Ogihara, Yuzuru Hamada, Hiroo Kumakura, andMasato Nakatsukasa. 2004. Do highly trained monkeys walk like humans?a kinematic study of bipedal locomotion in bipedally trained Japanese ma-caques. Journal of Human Evolution 46:739–750.

Hrdy, Sarah. 2005. Comes the child before the man: how cooperative breedingand prolonged postweaning dependence shaped human potentials. In Hun-ter-gatherer childhoods: evolutionary, developmental, and cultural perspectives.Barry S. Hewlett and Michael E. Lamb, eds. Pp. 65–91. New Brunswick,NJ: Aldine Transaction.

———. 2009. Mothers and others: the evolutionary origins of mutual under-standing. Cambridge, MA: Harvard University Press.

Hulanicka, Barbara, Lidia Gronkiewicz, and Jan Koniarek. 2001. Effect offamilial distress on growth and maturation of girls: a longitudinal study.American Journal of Human Biology 13:771–776.

Humphrey, Louise T. 2010. Weaning behaviour in human evolution. Seminarsin Cell and Developmental Biology 21:453–461.

Hurtado, A. Magdalena, Carol A. Lambourne, Kim Hill, and Karen Kessler.


2006. The public health implications of maternal care trade-offs. HumanNature 17:129–154.

Ibanez, Lourdes, Angela Ferrer, Marcia Victoria Marcos, Francisco RodriguezHierro, and Francis de Zegher. 2000. Early puberty: rapid progression andreduced final height in girls with low birth weight. Pediatrics 106:e72.http://pediatrics.aappublications.org/content/106/5/e72.full.

Jasienska, Grazyna. 2009. Reproduction and lifespan: trade-offs, overall energybudgets, intergenerational costs, and costs neglected by research. AmericanJournal of Human Biology 21:524–532.

Jost, Alfred. 1961. The role of fetal hormones in prenatal development. HarveyLectures 55:201–226.

Kaplan, Hillard, Kim Hill, Jane Lancaster, and A. Magdalena Hurtado. 2000.A theory of human life history evolution: diet, intelligence, and longevity.Evolutionary Anthropology: Issues, News, and Reviews 9:156–185.

Kaplan, Hillard, Jane Lancaster, and Arthur Robson. 2003. Embodied capitaland the evolutionary economics of the human life span. Population andDevelopment Review 29:152–182.

Karaolis-Danckert, Nadina, Anette E. Buyken, Antje Sonntag, and Anja Kroke.2009. Birth and early life influences on the timing of puberty onset: resultsfrom the DONALD (Dortmund nutritional and anthropometric longitu-dinally designed) study. American Journal of Clinical Nutrition 90:1559–1565.

Karlberg, Johan. 1989. A biologically-oriented mathematical model (ICP) forhuman growth. Acta Paediatrica Scandinavica 350(suppl.):70–94.

Kirkwood, Thomas B. L., and Michael R. Rose. 1991. Evolution of senescence:late survival sacrificed for reproduction. Philosophical Transactions of theRoyal Society B 332:15–24.

Knott, Cheryl. 2001. Female reproductive ecology of the apes. In Reproductiveecology and human evolution. Peter T. Ellison, ed. Pp. 429–463. New Bruns-wick, NJ: Aldine Transaction.

Kramer, Karen L. 2010. Cooperative breeding and its significance to the de-mographic success of humans. Annual Review of Anthropology 39:417–436.

Kramer, Karen L., and Peter T. Ellison. 2010. Pooled energy budgets: resi-tuating human energy-allocation trade-offs. Evolutionary Anthropology: Is-sues, News, and Reviews 19:136–147.

Kuzawa, Christopher W. 1998. Adipose tissue in human infancy and child-hood: an evolutionary perspective. American Journal of Physical Anthro-pology 107:177–209.

———. 2001. Maternal nutrition, fetal growth, and cardiovascular risk inFilipino adolescents. PhD dissertation, Emory University, Atlanta.

———. 2005. Fetal origins of developmental plasticity: are fetal cues reliablepredictors of future nutritional environments? American Journal of HumanBiology 17:5–21.

———. 2007. Developmental origins of life history: growth, productivity, andreproduction. American Journal of Human Biology 19:654–661.

———. 2008. The developmental origins of adult health: intergenerationalinertia in adaptation and disease. In Evolution and health. Wenda Trevathan,E. O. Smith, and James J. McKenna, eds. Pp. 325–349. Oxford: OxfordUniversity Press.

———. 2012. Why evolution needs development, and medicine needs evo-lution. International Journal of Epidemiology 41:223–229.

Kuzawa, Christopher W., Thomas W. McDade, Linda S. Adair, and NanetteLee. 2010. Rapid weight gain after birth predicts life history and repro-ductive strategy in Filipino males. Proceedings of the National Academy ofSciences of the USA 107:16800–16805.

Kuzawa, Christopher W., and Ivy L. Pike. 2005. Introduction. American Journalof Human Biology 17:1–4.

Kuzawa, Christopher W., and Elizabeth A. Quinn. 2009. Developmental or-igins of adult function and health: evolutionary hypotheses. Annual Reviewof Anthropology 38:131–147.

Kuzawa, Christopher W., and Zaneta M. Thayer. 2011. Timescales of humanadaptation: the role of epigenetic processes. Epigenomics 3:221–234.

Lasker, Gabriel W. 1969. Human biological adaptability. Science 166:1480–1486.

Le Bourg, Eric. 2007. Does reproduction decrease longevity in human beings?Ageing Research Reviews 6:141–149.

Lee, Phyllis C. 1996. The meanings of weaning: growth, lactation, and lifehistory. Evolutionary Anthropology: Issues, News, and Reviews 5:87–98.

Lee, Ronald D. 2003. Rethinking the evolutionary theory of aging: transfers,not births, shape senescence in social species. Proceedings of the NationalAcademy of Sciences of the USA 100:9637–9642.

Leidy Sievert, Lynnette. 2006. Menopause: a biocultural perspective. New Bruns-wick, NJ: Rutgers University Press.

Low, Bobbi S., Ashley Hazel, Nicholas Parker, and Kathleen B. Welch. 2008.Influences on women’s reproductive lives: unexpected ecological under-pinnings. Cross-Cultural Research 42:201–219.

Martorell, Reynaldo. 1995. Results and implications of the INCAP follow-upstudy. Journal of Nutrition 125(suppl.):1127S–1138S.

McDade, Thomas W. 2003. Life history theory and the immune system: stepstoward a human ecological immunology. American Journal of Physical An-thropology 122:100–125.

McIntyre, Matthew H., and Pamela M. Kacerosky. 2011. Age and size atmaturity in women: a norm of reaction? American Journal of Human Biology23:305–312.

Metzger, Boyd E., Lynn P. Lowe, Alan R. Dyer, Elisabeth R. Trimble, BrianSheridan, Moshe Hod, Rony Chen, et al. 2009. Hyperglycemia and adversepregnancy outcome (HAPO) study: associations with neonatal anthropo-metrics. Diabetes 58:453–459.


Migliano, Andrea B., Lucio Vinicius, and Marta Mirazon Lahr. 2007. Lifehistory trade-offs explain the evolution of human Pygmies. Proceedings ofthe National Academy of Sciences of the USA 104:20216–20219.

Moore, Lorna, Susan Niermeyer, and Stacy Zamudio. 1998. Human adaptationto high altitude: regional and life-cycle perspectives. Yearbook of PhysicalAnthropology 107:25–64.

Nakatsukasa, Masato, and Shin-Ichi Hayama. 1991. Structural strength of thefemur of bipedally trained macaque. Journal of the Anthropological Societyof Nippon 99:289–296.

Nettle, Daniel. 2010. Dying young and living fast: variation in life historyacross English neighborhoods. Behavioral Ecology 21:387–395.

Nettle, Daniel, David A. Coall, and Thomas E. Dickins. 2010. Birthweight andpaternal involvement predict early reproduction in British women: evidencefrom the National Child Development Study. American Journal of HumanBiology 22:172–179.

———. 2011. Early-life conditions and age at first pregnancy in Britishwomen. Proceedings of the Royal Society B 278:1721–1727.

Nyberg, Colleen, William Leonard, Susan Tanner, Thomas McDade, TomasHuanca, Ricardo Godoy, and TAPS Bolivia Study Team. Forthcoming. Di-urnal cortisol rhythms and child growth: exploring the life history conse-quences of HPA activation among the Tsimane. American Journal of HumanBiology.

Ong, Ken K., Pauline Emmett, Kate Northstone, Jean Golding, Imogen Rogers,Andrew R. Ness, Jonathan C. Wells, and David B. Dunger. 2009. Infancyweight gain predicts childhood body fat and age at menarche in girls. Journalof Clinical Endocrinology and Metabolism 94:1527–1532.

Pearson, Osbjorn M., and Daniel E. Lieberman. 2004. The aging of Wolff’s“law”: ontogeny and responses to mechanical loading in cortical bone.American Journal of Physical Anthropology 125:63–99.

Pembrey, Marcus E. 2010. Male-line transgenerational responses in humans.Human Fertility 13:268–271.

Perry, George H., and Nathaniel J. Dominy. 2009. Evolution of the humanPygmy phenotype. Trends in Ecology & Evolution 24:218–225.

Pesonen, Anu-Katrina, Katri Raikkonen, Kati Heinonen, Eero Kajantie, TomForsen, and Johan G. Eriksson. 2008. Reproductive traits following a parent-child separation trauma during childhood: a natural experiment duringWorld War II. American Journal of Human Biology 20:345–351.

Phoenix, Charles, Robert Goy, Arnold Gerall, and William Young. 1959. Or-ganizing action of prenatally administered testosterone propionate on thetissues mediating mating behavior in the female guinea pig. Endocrinology65:369–382.

Pickrell, Joseph K., Graham Coop, John Novembre, Sridhar Kudaravalli, JunZ. Li, Devin Absher, Balaji S. Srinivasan, et al. 2009. Signals of recent positiveselection in a worldwide sample of human populations. Genome Research19:826–837.

Prentice, Sarah, Antony J. Fulford, Landing M. A. Jarjou, Gail R. Goldberg,and Ann Prentice. 2010. Evidence for a downward secular trend in age ofmenarche in a rural Gambian population. Annals of Human Biology 37:717–721.

Price, Trevor D., Anna Qvarnstrom, and Darren E. Irwin. 2003. The role ofphenotypic plasticity in driving genetic evolution. Proceedings of the RoyalSociety B 270:1433–1440.

Promislow, Daniel E. L., and Paul H. Harvey. 1990. Living fast and dyingyoung: a comparative analysis of life-history variation among mammals.Journal of Zoology 220:417–437.

http://pediatrics.aappublications.org/content/106/5/e72.full


Proos, Lemm A., Yngve Hofvander, and Torsten Tuvemo. 1991. Menarchealage and growth pattern of Indian girls adopted in Sweden. 1. Menarchealage. Acta Paediatrica Scandinavica 80:852–858.

Quinlan, Robert J. 2003. Father absence, parental care, and female repro-ductive development. Evolution and Human Behavior 24:376–390.

———. 2007. Human parental effort and environmental risk. Proceedings ofthe Royal Society B 274:121–125.

Robinson, Beren W., and Kevin J. Parsons. 2002. Changing times, spaces, andfaces: tests and implications of adaptive morphological plasticity in the fishesof northern postglacial lakes. Canadian Journal of Fisheries and AquaticSciences 59:1819–1833.

Robson, Shannen L., Carel P. van Schaik, and Kristen Hawkes. 2006. Thederived features of human life history. In The evolution of human life history.Kristen Hawkes and Richard R. Paine, eds. Pp. 17–44. Santa Fe, NM: Schoolof American Research Press.

Sarkar, Sahotra. 1999. From the Reaktionsnorm to the adaptive norm: thenorm of reaction, 1909–1960. Biology and Philosophy 14:235–252.

Sellen, Daniel W. 2006. Lactation, complementary feeding, and human lifehistory. In The evolution of human life history. Kristen Hawkes and RichardR. Paine, eds. Pp. 155–196. Santa Fe, NM: School of American ResearchPress.

———. 2007. Evolution of infant and young child feeding: implications forcontemporary public health. Annual Review of Nutrition 27:123–148.

Shea, Brian T., and Robert C. Bailey. 1996. Allometry and adaptation of bodyproportions and stature in African Pygmies. American Journal of PhysicalAnthropology 100:311–340.

Stearns, Stephen C. 1983. The evolution of life-history traits in mosquitofishsince their introduction to Hawaii in 1905: rates of evolution, heritabilities,and developmental plasticity. American Zoologist 23:65–75.

———. 1989. Trade-offs in life-history evolution. Functional Ecology 3:259–268.

———. 1992. The evolution of life histories. New York: Oxford UniversityPress.

Stearns, Stephen C., and Jacob C. Koella. 1986. The evolution of phenotypicplasticity in life-history traits: predictions of reaction norms for age andsize at maturity. Evolution 40:893–913.

Sterling, Peter. 2004. Principles of allostasis. In Allostasis, homeostasis, and thecosts of adaptation. Jay Shulkin, ed. Pp. 2–36. New York: Cambridge Uni-versity Press.

Stinson, Sara. 1985. Sex differences in environmental sensitivity during growthand development. American Journal of Physical Anthropology 28:123–147.

Storz, Jay F., Graham R. Scott, and Zachary A. Cheviron. 2010. Phenotypicplasticity and genetic adaptation to high-altitude hypoxia in vertebrates.Journal of Experimental Biology 213:4125–4136.

Tanner, James Mourilyan. 1962. Growth at adolescence, with a general consid-

eration of the effects of hereditary and environmental factors upon growth andmaturation from birth to maturity. Oxford: Blackwell.

Teilmann, Grete, Carsten B. Pedersen, Niels E. Skakkebaek, and Tina K. Jensen.2006. Increased risk of precocious puberty in internationally adopted chil-dren in Denmark. Pediatrics 118:391–399.

Tither, Jacqueline M., and Bruce J. Ellis. 2008. Impact of fathers on daughters’age at menarche: a genetically and environmentally controlled sibling study.Developmental Psychology 44:1409–1420.

Victora, Cesar G., Linda Adair, Caroline Fall, Pedro C. Hallal, Reynaldo Mar-torell, Linda Richter, and Harshpal Singh Sachdev. 2008. Maternal and childundernutrition: consequences for adult health and human capital. Lancet371:340–357.

Waddington, Conrad H. 1953. Genetic assimilation of an acquired character.Evolution 7:118–126.

Walker, Robert S., Michael Gurven, Oskar Burger, and Marcus J. Hamilton.2008. The trade-off between number and size of offspring in humans andother primates. Proceedings of the Royal Society B 275:827–834.

Walker, Robert S., Michael Gurven, Kim Hill, Andrea Migliano, NapoleanChagnon, Roberta D. Souza, Gradimir Djurovic, et al. 2006a. Growth ratesand life histories in twenty-two small-scale societies. American Journal ofHuman Biology 18:295–311.

Walker, Robert S., Kim Hill, Oskar Burger, and A. Magdalena Hurtado. 2006b.Life in the slow lane revisited: ontogenetic separation between chimpanzeesand humans. American Journal of Physical Anthropology 129:577–583.

Wells, Jonathan C. K. 2003. The thrifty phenotype hypothesis: thrifty offspringor thrifty mother? Journal of Theoretical Biology 221:143–161.

———. 2007. The thrifty phenotype as an adaptive maternal effect. BiologicalReviews 82:143–172.

Wells, Jonathan C. K., and Jay T. Stock. 2007. The biology of the colonizingape. Yearbook of Physical Anthropology 134:191–222.

West-Eberhard, Mary Jane. 2003. Developmental plasticity and evolution. NewYork: Oxford University Press.

———. 2005. Developmental plasticity and the origin of species differences.Proceedings of the National Academy of Sciences of the USA 102:6543–6549.

Williams, George C. 1966. Natural selection, the costs of reproduction, anda refinement of Lack’s principle. American Naturalist 100:687–690.

Wilson, Margo, and Martin Daly. 1997. Life expectancy, economic inequality,homicide, and reproductive timing in Chicago neighbourhoods. BritishMedical Journal 314:1271–1274.

Worthman, Carol M. 1999. Epidemiology of human development. In Hor-mones, health, and behavior. Catherine Panter-Brick and Carol M. Worth-man, eds. Pp. 47–104. Cambridge: Cambridge University Press.

Wund, Matthew A., John A. Baker, Brendan Clancy, Justin L. Golub, andSusan A. Foster. 2008. A test of the “flexible stem” model of evolution:ancestral plasticity, genetic accommodation, and morphological divergencein the threespine stickleback radiation. American Naturalist 172:449–462.



Conditions for Evolution of Small AdultBody Size in Southern Africa

by Susan Pfeiffer

Discoveries from diverse locales indicate that early Homo was sometimes petite. Small body size among fossil formsis difficult to explain because its existence in modern human populations is not fully understood. The history,ethnography, genetics, and bioarchaeology of KhoeSan peoples of southern Africa are reviewed in the context oftheir small adult body size. Since the Middle Stone Age, at least some southern African foragers were petite.Throughout the Later Stone Age (LSA; the Holocene), most groups followed a mobile, coastally oriented foragingstrategy that relied on small package size foodstuffs. Distinctive skeletal shape and allometry of LSA adult skeletonsprovide clues about selective factors. Neither dietary insufficiency nor heat dissipation models of selection apply inthe LSA context. Energetics and avoidance of serious accidents may be relevant factors. An aspect of life history—the timing of cessation of growth—has been assessed by comparing dental and skeletal development within juvenileskeletons. After a slow start, LSA child growth shows a tempo like that of modern children and no evidence ofearly maturation. Among fossil or recent forms, small body size should be assessed not only as possible evidenceof selection for smallness but also as evidence of the absence of selection for large body size.

Recent discoveries have drawn attention to the importanceof variability in human body size. The discovery of exceed-ingly small people of late Pleistocene Indonesia (Brown et al.2004; Jungers et al. 2009; Larson et al. 2009; Morwood andJungers 2009) is the best known example, but fossil elementsat the lower range of modern human size occur among newdiscoveries of putative Homo from east Africa and Georgia(Anton et al. 2007; Graves et al. 2010; Rightmire 2008; Ruff2010; Spoor et al. 2007). Unresolved topics include how muchof the size diversity should be attributed to sexual dimor-phism, how this diversity arose, and whether smallness wasa response to environmental variables (McHenry and Brown2008; Reno et al. 2010). These discoveries have drawn atten-tion to the need for more focused study of the ecological andbehavioral conditions under which isolated human popula-tions may develop small, gracile adult body size.

Theories about the selective value of small body size arenumerous, and opportunities to test competing hypothesesare rare. The population history of a more recent humanlineage can provide a case study, exploring possible selectiveprocesses. As well, the morphological characteristics of therecent group may help shape expectations about the mor-phology of small-bodied forms from earlier in the human

Susan Pfeiffer is Professor in the Department of Anthropology atthe University of Toronto (19 Russell Street, Toronto, Ontario M5S2S2, Canada [[email protected]]) and Research Associatein the Department of Archaeology at the University of Cape Town,South Africa. This paper was submitted 12 XII 11, accepted 21 VI12, and electronically published 18 IX 12.

lineage. The small, lightly built foragers of Holocene SouthAfrica are ancestral to the contemporary KhoeSan-speakingpeoples of southern Africa. The ancestors’ skeletons providea basis for a focused examination of one context of smallbody size. Genetic studies provide corroboration of the tem-poral and spatial framework of KhoeSan people, reinforcingarchaeological evidence of their antiquity and distinctiveness.From this base, information from bioarchaeology can be usedto test hypotheses that are important to evolutionary an-thropology.

The Southern African Context

The adult body size of KhoeSan speakers of southern Africa(often referred to as “Bushmen”) is quite small relative tothat of most human groups. With average adult statures doc-umented in historic times at 160.9 cm (men, ) andN p 79150.1 cm (women, ; Truswell and Hanson 1976), theN p 74hunter-gatherers of the Kalahari region may not reach thepygmoid category (defined as 155 cm; Cavalli-Sforza 1986),but male and female averages approximate the third percentilevalues of Centers for Disease Control 2000 adult statures(www.cdc.gov/growthcharts). As with many other small-bod-ied groups, there is a history of debate regarding whethertheir body size represents adaptation or pathology (cf. Lee1979; Wilmsen 1989). Their demographic and anthropomet-ric features, including the growth of KhoeSan children, figureinto the modeling of possible selection for smallness. Groupssuch as the !Kung (or Ju/’hoansi) and the G/wi have been


http://www.cdc.gov/growthcharts


Figure 1. Map of southernmost Africa, modified from Morris (1992). The area in which most burials are recovered is bounded bya curved line paralleling 34�S between the Cape Fold Mountains and the sea. KRM p Klasies River main site; Fy p fynbos biome;Fo p forest biome; Sa p savannah biome. Immediately north of these biomes are types of karoo.

documented in their Kalahari home ranges at latitudes of 20�

to 23�S. They are typically portrayed as representing a desert-or semidesert-adapted population (e.g., Walker et al. 2006).

Archaeological evidence (Deacon and Deacon 1999; Mitch-ell 2002) and genetic studies (Scheinfeldt, Soi, and Tishkoff2010; Schuster et al. 2010; Tishkoff et al. 2009) indicate thatcontemporary KhoeSan speakers are descendants of an an-cestral population that probably extended north at least tothe Kalahari and the Zambezi and south to the shores of theSouth Atlantic and Indian oceans. Physical evidence of theancestors includes not only sites and artifacts but also exten-sive rock art and open-air engraving sites, their creation span-ning the Holocene (Barham and Mitchell 2008). Naturalisticrepresentations of human figures in the rock art showKhoeSan morphological traits, and the hand prints are ofsmall hands of adult proportions (Manhire 1998; Parkington2003).

The greatest concentration of Later Stone Age (LSA) ar-chaeological material has been documented in southernmostcoastal and near-coastal regions of South Africa, with thehighest Holocene population numbers found around latitude34�S (fig. 1). While there are variations in rainfall, topography,and food resources throughout the region, the consistency ofthe Mediterranean climate, fynbos, and Afromontane forestplants (Cowling and Hilton-Taylor 1997; Schulze 1997) andmarine and terrestrial animals supported a broadly identifi-able cultural adaptation (Deacon and Deacon 1999; Mitchell2002; Pfeiffer and Sealy 2009). Modeling based on contem-porary genomic diversity indicates that LSA hunter-gathererpopulations expanded substantially during the past ca. 41,000years (Cox et al. 2009). Therefore, when trying to understandwhat factors might have acted on adult body size in thispopulation, evidence from South African LSA archaeology ismore pertinent than evidence from the Kalahari environmentin which relict populations live today.

Bioarchaeological evidence of the LSA in the coastal and

near-coastal South African Cape comes from primary inhu-mations of adults and juveniles found in rock shelters, shellmiddens, and sand dunes. Single bodies are normally hori-zontally positioned in a flexed posture with no clear preferencefor side of the body or direction of the head (Inskeep 1986).The Cape (including the Cape Fold Mountains) is distinctfrom the karoo, the eastern savannah and grasslands, in thefrequency of discovered burials. Whether because of culturaldifferences or because of substantial differences in populationdensity throughout the Holocene, very few burials of foragersare known from beyond the western and southern coasts andcoastal forelands regions (Morris 1992). All burials are con-sistent with the observation that Holocene Bushman ancestorswere consistently petite in stature (fig. 2; Kurki et al. 2008,2010; Sealy and Pfeiffer 2000). Adult body size fluctuatesthrough time and space as communities dealt with stressorsnot yet fully defined (Pfeiffer and Sealy 2006), but no groupfrom any time or place exceeded historic Bushman statures.A coefficient of variation of 6% for the plotted sample of 172maximum femoral lengths (mean p 407.8 mm, SD p 25.3mm) illustrates the temporal and spatial homogeneity of thepopulation.

Skeletal features can be used to extrapolate habitual be-haviors, which in turn can reflect morphology. The commonoccurrence of articular facet modification (Dewar and Pfeiffer2004) and acetabular beveling (Pfeiffer 2011) suggests thatLSA people habitually assumed deep squatting postures. Thiswould be possible only if subcutaneous fat deposits on thelower limbs were minimal. Hence, LSA people were both smalland lean. Skeletal dimensions confirm that at birth, infantswere not notably small (Harrington and Pfeiffer 2007). Fol-lowing a slight lag during the first year, the linear growth ofLSA children followed a normally shaped growth curve lead-ing to petite adult stature (Harrington and Pfeiffer 2008; Pfeif-fer and Harrington 2010). While some regional and temporalvariations in morphology have been documented (Ginter

Pfeiffer Small Body Size S385

Figure 2. Maximum femur length versus uncalibrated radiocarbon dates for 172 Later Stone Age adult skeletons: west (squares,), south (circles, ), east (triangles, ), north coast and interior (stars, ). A femur length of 460 mm isN p 76 N p 50 N p 43 N p 3

roughly equivalent to a stature of 170 cm, or 5 feet 7 inches. Pygmy femur length is ca. 380 mm.

2009; Pfeiffer and Sealy 2009; Stock and Pfeiffer 2004; Stynder,Ackermann, and Sealy 2007a, 2007b), there is also evidencefor a genetically modulated morphology with considerableantiquity. Genetic studies of descendant populations suggestthat the KhoeSan lineage is one of Africa’s oldest (Tishkoffet al. 1996, 2009). Craniometric research of archaeologicallyderived material indicates a homogeneous population untilat least 2000 BP (Stynder, Ackermann, and Sealy 2007a,2007b). Even dental dimensions are distinctively small (Black,Ackermann, and Sealy 2009). Thus, the factors required fornatural selection—namely time and isolation—were available.

It is not clear when adult small body size became establishedin southern Africa. Small adult bones are found among thehuman remains from some Middle Stone Age sites. Cranialand postcranial specimens from Klasies River main site, datingto about 115,000 years ago, have been observed to have ex-ternal dimensions comparable to LSA skeletons (Rightmireand Deacon 1991, 2001). This has been corroborated in anal-yses of individual skeletal elements (Churchill et al. 1996; Lam,Pearson, and Smith 1996; Pearson and Grine 1996, 1997;Pearson et al. 1998). However, there are other putatively Mid-dle Stone Age skeletal elements that are not particularly small,such as a fifth metatarsal from Klasies River main site (Right-mire et al. 2006), the late Pleistocene Hofmeyr skull (Grine

et al. 2007), and a humerus shaft from Border Cave (Pfeifferand Zehr 1996). There is also some scant evidence of largebody size in the early Holocene. One first metatarsal from asouthern Cape rock shelter (SAM-AP 4208B) is substantiallylarger than other LSA first metatarsals (Pfeiffer and Sealy2006). Dating to about 9,500 years ago, this bone is a reminderthat pre-Holocene and early Holocene skeletal remains areextremely rare, and we know very little about the populationat that time (fig. 3).

Early literature on southern African prehistory posited alarge, robust ancestral population known as the Boskop“race.” Before the era of radiocarbon dating, the size of askeleton was treated as an indication of its relative antiquity(Drennan 1938; FitzSimons 1926). The Boskop idea wassoundly refuted on cranial grounds (Brauer and Rosing 1989;Rightmire 1978, 1984; Singer 1958), but the temporal andspatial extent of the small-bodied population across the Af-rican landscape is open to interpretation (Morris 2002, 2003).There may have been morphological variability in the latePleistocene and the early Holocene for which evidence is notyet available. Marine-oriented adaptations along loweredcoastlines (relative to modern levels; Marean 2010) may ex-plain why late Pleistocene archaeological evidence is rare.Based on current evidence, it appears that adult smallness


Figure 3. Adult first metatarsals, plantar view, demonstrating size differences. In the upper left-hand corner is the Middle StoneAge metatarsal from Klasies River main site, estimated maximum length 56 mm, similar to the Later Stone Age (LSA) average. Allother metacarpals are LSA. In the upper right-hand corner is an LSA metatarsal (SAM-AP 4208B), radiocarbon dated to about9500 BP, maximum length 70 mm, well outside the LSA range. Photo by Susan Pfeiffer.

existed at the southern coast from 115,000 years ago, and itwas typical of the population by about 10,000 years ago, atleast in the coastal and near-coastal regions.

Some Morphological Correlates of Smallness

Pelvic Shape

Among modern KhoeSan people, ethnographic evidence in-dicates that newborns are not smaller than global standards.Howell recorded a mean birth weight of 3.08 kg (SD p 0.458)in 10 newborns, with just one infant falling below the WorldHealth Organization low birth weight criterion of 2,500 g(Howell 2000). To explore perinatal size in earlier millennia,the skeletons of 18 perinates from the LSA were measured.They show dimensions that are well within modern norms(Harrington and Pfeiffer 2008). To explore the relationshipbetween pelvic capacity and perinatal size, Kurki (2007) com-pared adult body size and pelvic dimensions in small-bodiedsouthern African Holocene foragers (women, ; men,N p 28

) with those from historic Portuguese of intermediateN p 31size (Coimbra) and larger-bodied adults (Hamann-Todd).Both males and females in the LSA sample show distinctivepelvic shapes. Despite narrow bi-iliac breadths, the small-bodied forager females have the relatively largest midplaneand outlet canal planes, suggesting adaptive allometric re-modeling in this small-bodied population. The LSA malesalso follow this pattern, although with smaller pelvic dimen-sions. The distinctive pelvic shape stays stable during the mid-Holocene period of body size variation, when some adultswere exceptionally small (Kurki, Stynder, and Pfeiffer 2012).The antiquity of KhoeSan skeletal morphology is reflected in

these distinctive aspects of pelvic shape, accommodating anadequate birth canal despite small external hip dimensions.

A recently described case study illustrates the narrow mar-gin of error within which this population operated. The sa-crum of a middle-aged LSA woman shows slight develop-mental asymmetry, with one wing (ala) about 5 mm narrowerthan the other. Probably subsequent to the strains of child-birth, this slightly asymmetrical and narrowed pelvis showsskeletal indications of joint instability to the point of ebur-nation on all pubic and auricular joint surfaces (Pfeiffer 2011).Had the obstetric canal been any further constrained, the birthevent might have been fatal to mother and child.

Physique and Proportions

Studies comparing skeletal variation in response to ecogeo-graphic factors often inappropriately categorize KhoeSan ma-terial as derived from 20� to 25�S, thereby positioning thepopulation as heat adapted. These comparative samples oftenrely on small collections of “Bushman/Kaffir/Hottentot” skel-etons that are held in Northern Hemisphere museums, ob-tained through purchase or exchange from South African cu-rators. Interest in the physical “type” fueled an active tradefor decades, with some exploitation of marked graves formaterial (Legassick and Rassool 2000). Such selective practicescould lead to morphological biases in those collections.

Kurki et al. (2008) used a well-characterized archaeologi-cally derived sample ( , ca. 34�S) to demonstrate thatN p 124limb and limb-trunk proportions of that set are the same asthose from samples used by earlier researchers. Small valuesfor dispersion around mean values reinforce the consistencyof proportionality within the sample. Comparative study


shows that both brachial and limb-to-trunk indices for theLSA sample are consistent not with low-latitude but with mid-latitude populations, such as those from North Africa. In thisrespect, the presumption of heat adaptation is not supported.On the other hand, the LSA sample shows some features thatare normally linked to heat adaptation, including distinctlynarrow bi-iliac breadth (BIB), low ratio of BIB to femurlength, and small stature. The authors demonstrate that small-bodied human groups show considerable diversity in skeletalshape and proportions. They conclude that in zones that arenot subject to climatic extremes, climatic factors may play aless important selective role and that life history or otherfactors may have affected body size and proportions. Sub-sequent research has demonstrated that when various statureand mass estimation methods are applied to the LSA sample,results can be quite anomalous (Kurki et al. 2010). Takentogether, the numerous well-preserved skeletons of the LSAprovide opportunities to explore morphological variation ata level of detail that is not possible with less well-proveniencedsamples of small-bodied modern humans.

Allometry of Head and Body Size

The scope of the LSA sample allows exploration of how dif-ferent skeletal components respond to transient stressors.Comparison of cranial centroid data (Stynder 2006) with pel-vic and femoral dimensions in a sample of 65 dated adultskeletons demonstrates the relationship between cranial andpostcranial components. During the mid-Holocene period,when variance increases through the presence of some evensmaller adults, reduced major axis regression of body size oncranial size indicates negative allometry between head andbody size. Femoral length, and to a lesser degree femoral headdiameter, decline more abruptly than cranial size; cranial sizeis more conserved when growth falters (Kurki, Stynder, andPfeiffer 2012).

Evolution of Small Body Size

With new evidence of size variability in fossil Homo, attentionhas been given to possible mechanisms of evolution of smallhuman body size (Becker et al. 2010; Bernstein 2010; Froment2001; Migliano, Vinicius, and Lahr 2007, 2010; Perry andDominy 2008; Walker and Hamilton 2008; Walker et al. 2006).The adaptive value of larger size is thought to reflect higherfecundity among larger females and greater access to matesamong larger males as a result of male-male competition.These positive features are then balanced against the fitnesscosts of growing and maintaining large body size (Blancken-horn 2000). Hypotheses about selection for smallness or theabsence of selection for largeness have been characterized astaking four approaches (Perry and Dominy 2008): thermo-regulation, food limitation, mobility, and life history. Pygmypopulations from humid, closed habitats such as tropical rain-forests may be adapted to long-term limitations on amount

and types of food available coupled with the thermoregulatoryadvantage of a high surface-to-volume ratio. Looking beyondPygmy groups, food limitations may occur in a variety ofhabitats. Small size can enhance mobility by reducing themetabolic costs of activity so that small bodies can be efficientif the necessary work does not require bursts of intense power.

A hypothesis of recent interest, potentially applicable to allhabitats, places body size in a life history framework (Charnov1993, 2001). It predicts a relationship between the pattern ofadult mortality risk and body size in which populations withconsistently high risk of young adult mortality will show earlyonset of reproduction, truncating growth to initiate repro-duction. Recent work by Migliano and colleagues has ex-panded the focus to include preadult mortality and evidencefor cessation of linear growth at around 13 years for Pygmywomen as additional evidence for selection in very small-bodied humans (Migliano 2005; Migliano and Guillon 2012;Migliano, Vinicius, and Lahr 2007). Methodological aspectsof the Migliano group’s approach have been debated (Beckeret al. 2010; Migliano, Vinicius, and Lahr 2010), but therecontinues to be interest in the idea that adult small size mayresult from indirect selection, and there is no question thatadaptive features of childhood can be crucial to a group’ssurvival.

Evolution of KhoeSan Small Body Size

In the context of the evolutionary pathways toward small bodysize that have been postulated, the LSA skeletal material canbe assessed. Given the absence of climatic extremes, the ther-moregulatory rationale is set aside.

Dietary Insufficiency

The earliest evidence of small adult body size comes fromKlasies River main site, where there is evidence for dietaryactivities that included roasting geophytes (corms and bulbs),hunting large game, and collecting shellfish (Barham andMitchell 2008). Evidence of harvesting of aquatic foods, es-pecially shellfish, is strongly associated with Middle Stone Agesites and continues to be important to coastal and near-coastalsubsistence throughout the LSA. Access to freshwater, in-cluding storage of freshwater in containers, is another themefrom the Middle Stone Age onward. Terrestrial game huntingremains important, with regional and temporal changes inthe species exploited and a general trend toward smaller“package size” as human population density increased duringthe LSA. Evidence of light-draw bows (and presumably poi-son-tipped arrows) is found around the end of the Pleistocene(Parkington 1998; Wadley 1998), although earlier inventionhas been proposed (Lombard and Phillipson 2010; Wadley,Hidgskiss, and Grant 2009). This technology continuedthroughout the LSA to the historic era.

If population density was focused along the coast, the ma-rine diet may be especially relevant if we assume that relatively


few people were foraging throughout the interior. While theidea of the “aquatic diet” being necessary for encephalizationhas been laid to rest (Carlson and Kingston 2007; Snodgrass,Leonard, and Robertson 2009), a diet high in marine proteindoes have unique features that could influence selective path-ways. Thanks to the research into dietary stable isotopes ofcarbon and nitrogen, chiefly from human bone collagen, re-gional dietary protein patterns are well characterized (Sealy2010; Sealy and van der Merwe 1988; Sealy et al. 1987). Highin nutrient density (the ratio of nutrient content to totalenergy content), marine foods and geophytes are easily gath-ered, if available. Shellfish and geophytes are relatively broadlydistributed and could be readily harvested by almost all groupmembers, and so it is difficult to restrict access to them. Readyaccess to reliable beds of shellfish could have sustained thenutrition of mothers during prolonged lactation (Clayton,Sealy, and Pfeiffer 2006) and formed part of an egalitariansubsistence framework. However, stable isotope evidencefrom the region of highest marine exploitation, the southcoast, indicates a time in the mid-Holocene when the land-scape was divided between foragers who had access to Capefur seals and other high trophic-level foods and neighboringgroups who lacked that access (Sealy 2006). This is a reminderthat access to coastal resources could be restricted, therebyshifting interpersonal competition from a scramble to a con-test. Another example of apparently stressful dietary condi-tions comes from the western coast during the “megamiddenperiod,” ca. 3000–2000 BP (Jerardino 2010). Evidence of largeopen sites plus intensification of harvesting of terrestrial andmarine foods is combined with a peak in numbers of humanburials and some evidence of interpersonal violence (Pfeiffer2013). It appears that increased human numbers outstrippedlocal resource availability, although it is difficult to identifywhat dietary elements were the limiting factors. Proteinsources do not appear to have been completely depleted, butperhaps the supply of carbohydrates (geophytes) was inade-quate.

In brief, there is no evidence for prolonged periods ofsignificant shortages either to specific dietary components orto food energy more generally. The topic of climate stabilityis also pertinent. Current reconstructions of late Pleistoceneand Holocene climate across southern Africa identify no par-ticularly rapid or dramatic changes. Southern African deep-sea cores document temperature fluctuations on a scale of 2�

to 4�C (Farmer, deMenocal, and Marchitto 2005) and a rel-atively humid period in the late Pleistocene followed by aperiod of gradually increasing aridity from 3800 cal BP tovery recent times (Chase et al. 2010; Quick et al. 2011; Scottet al. 2012).

Given the absence of pervasive skeletal indications of di-etary insufficiencies and no strong correlations between adultbody size and stable isotope values (Pfeiffer and Sealy 2006),it appears that nutrient levels were generally adequate.Whether food energy was adequate is a more difficult ques-tion. While there may be some merit to an argument that

the LSA diet was always nutrient dense, thereby being low infood energy, the evidence is not in place to test this idea withmuch rigor.

Energetics and Accident Avoidance

Much of the coastal and near-coastal terrain of southern Af-rica is characterized by hills and mountains of moderate tohigh relief (Kruger 1983; Schulze 1997). Foragers needed totraverse loose, sometimes wet rocky slopes daily. The LSAsubsistence strategy—relying on accessing freshwater, corms,and small protein packages—required persistent landscapescans, persistence hunting (Liebenberg 2006), and transportof parcels of variable size to camp sites that were frequentlymoved. Being mobile was probably more important than be-ing particularly strong. In similarly rugged landscapes else-where, skeletons of foraging groups show relatively high prev-alence of healed bone fractures (Kilgore, Jurmain, and vanGerven 1997). Breaks of arm bones and clavicles are associatedwith accidents rather than interpersonal violence and are oftenquantified to study trauma in past populations (Roberts andManchester 1995). Accidental fractures tend to be more com-mon among hunter-gatherers than among more sedentarygroups. An assessment of 1,353 major long bones from 152LSA adult skeletons showed 13 healed fractures, all of armbones except for one clavicle and one distal tibia (at the medialmalleolus). This is a comparable pattern of distribution tothat seen in other hunter-gatherers but a lower prevalence(Pfeiffer 2007; fig. 4). It is possible that a lighter, smaller bodyis less likely to sustain serious injury from slips and falls.Adult mortality has an effect on healed-fracture data becauseit is a cumulative measure. That is, a longer-lived group mayshow a higher prevalence of healed fractures. It can also beargued that hazards such as a broken ulna or clavicle are notlikely to affect fitness. Nevertheless, this evidence suggests thatthe sequelae of serious injury—including shock, blood loss,and immobility—may have been less common than wouldbe the case among groups with higher average adult bodymass.

Life History

There are several aspects to this theoretical approach. Oneaspect has been explored within the LSA context, that of ashorter childhood period to allow earlier onset of reproduc-tion. While cross-sectional growth data from skeletons cannotbe used to track the adolescent growth spurt (Sinclair andDangerfield 1998), cross-sectional plots of achieved lineargrowth coupled with the timing of epiphysis closure, can beuseful. Child growth during the South African LSA has beenstudied from various perspectives using a substantial sampleof immature skeletons with documented archaeological con-text, age at death, and (often) radiocarbon date. Juvenile re-mains have been described and analyzed in bioarchaeologicaldescriptions and analyses (Harrington 2010; Pfeiffer and van


Figure 4. Prevalence (%) of healed fractures for the most common accidentally broken bones, the radius and ulna. The first threecomparators are hunter-gatherer groups: Kulubnarti, Nubia (Kilgore, Jurmain, and van Gerven 1997); Atacama, Chile (Neves, Barros,and Costa 1999); and central California (Jurmain 1991). Numbers 1–6 are medieval British sites (Roberts and Manchester 1995).Number 5, Chichester, was a hospital site.

der Merwe 2004; Sealy et al. 2000). No instances of periostitisor other infectious processes have been found (Pfeiffer 2007).The rock shelter burial of one infant who had been sufferingfrom a metabolic disorder for a prolonged period before deathindicates that poor health before death did not preclude nor-mal burial practices (Pfeiffer and Crowder 2004). These ob-servations enhance confidence in the sample’s validity forhealth-related studies.

The growth of immature LSA skeletons has been ap-proached as a cross-sectional sample (Harrington and Pfeiffer2008). From initial scrutiny of 25 juvenile mandibular den-titions (15 of which were radiographed), we determined thatthe order and relative timing of deciduous and permanenttooth development were fully congruent with developmentalstandards (Matzke 2000). This congruence supports relianceon age estimates based on dental development (Smith 1991).While the teeth of southern Africans have been shown tocomplete crown formation at slightly earlier ages than teethof North Europeans (Reid and Dean 2006), that collection ofdiverse black African tribes is genetically distinct from theLSA population. Hence, if there is bias in the age estimates,its direction and magnitude remains unknown. It has beenreported that the error associated with dental age estimationmethods is about 1 year and that the tendency of these meth-ods is to slightly overage the younger children and underageolder children (Liversidge, Smith, and Maber 2010).

Focusing on juveniles for whom dental maturation can beused to estimate age at death, LSA data show that the lineargrowth of juveniles who failed to survive falls within 1 SD of

the proportion of adult growth that would be expected forliving healthy children (Harrington and Pfeiffer 2008). Theabsence of growth lag before death suggests that causes ofdeath were typically acute conditions rather than chronic con-ditions where death would be preceded by a slowing or ces-sation of growth. Assessment of long bone growth arrest lines(present in 25 of 53) and cribra orbitalia on juvenile frontalbones (present in 15 of 38) indicates that these signs of non-specific stress occur, but they are neither ubiquitous nor severe(Pfeiffer 2007).

Femur lengths of juveniles show that adult lengths are notachieved until dental ages of about 16 years (fig. 5), at whichtime skeletal maturation remains incomplete. Maturation andattainment of femur lengths representing probable adult stat-ure occur at ages that are compatible with a mean age at firstreproduction of 19.5 years, as noted among the !Kung byHowell (Howell 2000), because menarche occurs after thepeak in height velocity, and full sexual maturation occurs ayear or two after menarche (Sinclair and Dangerfield 1998).Among LSA populations, this suggests that child growth pro-ceeded at a normal tempo (assuming progress toward a smalladult end point) and that growth did not cease prematurely.

Discussion and Conclusions

Many societies tend to favor tallness as an indicator of ma-turity, strength, and even leadership potential (Ekwo et al.2005). Groups of small people or small individuals have beenalternately portrayed as anomalies of growth perturbation or


Figure 5. Percentage of adult linear growth (femur length measurements) attained among Later Stone Age juveniles with ageestimates (years) based on dental maturation. Shaded boxes summarize diaphysis measurements; the open box summarizes mea-surements of complete femora (diaphyses and epiphyses) of the oldest juveniles. Midlines indicate medians; boxes indicate first tothird quartiles; whiskers indicate the range excluding outliers, which are shown as circles. Figure prepared by L. Harrington.

the result of evolutionary adaptation (Schell and Magnus2007). In the case of KhoeSan-speaking people, bioarchaeo-logical evidence provides a framework for understanding thegroup’s origins. The locus of population density appears tohave been substantially south of the location where relict com-munities live today. This genetically unique populationevolved outside the range of climatic extremes, so their mor-phology is not likely to reflect a strong influence of extremetemperature or humidity. The long period of consistentlysmall adult stature, documented through study of archaeo-logically derived skeletons, is inconsistent with the idea thatsmallness reflects stressors within each individual’s lifetime.It is very improbable that for thousands of years, each in-dividual failed to achieve her/his body size potential in vir-tually the same way. The consistency of growth tempo andthe low coefficients of variation associated with linear mea-sures suggest canalization (Waddington 1957). The smallnessof LSA people and their KhoeSan descendants appears torepresent adaptation, but to what?

The LSA people had access to a wide range of food resourcesover thousands of years, although there may have been localesand periods in which shortages occurred. The framework forpostulating small stature as a response to pervasive food short-age—either specific nutrients or collective food energy—isnot apparent. Although much of the landscape is potentiallytreacherous to wide-ranging foragers, it seems improbablethat avoidance of serious injury would have been a strongfactor influencing survival and reproduction, although these

ideas could be explored further, especially if new informationand approaches arise.

The exploration of life history factors as the basis of aselective framework could be taken further. Adult propor-tionality is not temporally or spatially variable, suggestingconsistent canalization during development. Assessment ofjuvenile growth, comparing the completion of long bonegrowth against dental development, indicates no early trun-cation of growth. Assuming that puberty and long bonegrowth were linked in past peoples as they are today, thissuggests that onset of reproduction was not early. However,another perspective would be to argue that through slowedgrowth, LSA children were demanding less food energy,thereby leaving it for the reproducing adults. Future work toaccurately determine the adult ages at death is needed toexplore the possibility of early adult mortality. Questions re-main about the evolutionary value of postreproductive groupmembers (Hawkes 2010) and the adaptive value of equitablegendered roles (Gurven and Hill 2010). Given the strong rolesidentified for older, postreproductive members of KhoeSansocieties (Howell 2010), it seems unlikely that this class wasrare in past generations.

An alternate approach to the question of KhoeSan smallnesswould be to look at it as the product of the removal ofselection for large size coupled with isolation and stabilizingselective pressures favoring smallness. Sexual selection forlarge males can be mitigated by cultural practices, therebyremoving that directional pressure. Selection for larger females


through fecundity can be mitigated through shape adjust-ments that optimize obstetric canal volume. Assuming thatthe KhoeSan ancestors were the only human inhabitants ofa large region for at least 10,000 years, perhaps substantiallylonger, and assuming a small founding population at somepoint, smallness may have arisen by chance or in response toselective pressures at that earlier time. Those pressures mayhave related to energetics or diet or some other factor. Thereis some evidence to support this hypothesis. A distinctivepelvic shape accommodates relatively large infants, therebyreducing the selection for larger women through fecundity.A lightweight tool kit that relies on poison-tipped arrows andthe practice of persistence tracking may reduce or reverse theselection for larger men through muscular strength. Rock artportrays men as lean, muscular, and small relative to animals.Ethnographically, R. B. Lee documented that smaller hunterswere more successful than larger men (Lee 1979). GeorgeSilberbauer, when asked about sexual selection among theG/wi, said that tall men were viewed as “geeks,” that is, po-tentially less desirable mates (G. Silberbauer, personal com-munication, 1994).

Directional selective factors appear to have elicited smallbody size among some humans in this region during thePleistocene before ca. 110 ka. The evidence regarding whatsuch forces might have been is unavailable. Changes in sealevel (Carr et al. 2010) and some climatic shifts may be im-plicated. During the Holocene, the low variance in skeletalmeasures indicates the operation of stabilizing natural selec-tion in an environment where climate and food availabilitydo not appear to impose particular constraints on survival.A biocultural mix of dietary constraints and behavioral ad-aptations may be the most plausible explanation for the Ho-locene adaptive pattern. In the most recent millennium, con-tact with black Africans and subsequently with Europeansaltered the picture dramatically. From that point onward,small body size was one of several disadvantages borne by theBushmen.

The LSA story illustrates how bioarchaeology can be helpfulto evolutionary anthropology, providing a more completeframework for human adaptations than is available throughhistoric and ethnographic sources alone. Comparative studyof Holocene LSA skeletons and other small-bodied groupshas demonstrated that small-bodied humans are variablyshaped. Fragmentary fossil remains will generate attempts toestimate body mass, stature, and proportionality. These maybe even less precise when the skeletal components are fromvery small adults. With regard to possible ecogeographic pat-terning in our species, some of our expectations may be dis-torted by assumptions that a group’s modern locale (like theKalahari) represents the group’s geographic origin, as it mis-leads the search for selective factors that influenced charac-teristic morphology.

The consistency of small body size in southernmost Africasuggests that it may have been maintained under conditionsthat do not include climatic extremes or very high mortality.

Based on parallels with the LSA example, we can identifyecological factors that may have supported small body sizein some early Homo populations. Geographic isolation for aprolonged period is an important permissive variable. Thedynamics that are understood within island biogeography(Brown 1995) may apply to contiguous regions if there aresignificant physical barriers, such as mountains, and if pop-ulation density is well within the area’s carrying capacity.Dmanisi may be an example of this among known early Homosites. Possible selective variables include food resources withsmall package size dispersed so that interpersonal competitionis a scramble rather than a contest. Small body size could belinked to a social system with little emphasis on strengthcompetition among males so that larger males have little orno reproductive advantage. Uneven, irregular, or challengingtopography could also be a relevant variable insofar as seriouspersonal injury can reduce fitness. A subsistence system basedon small package resources can be fashioned within manyecological contexts and could be relatively flexible when en-vironmental conditions shift. As we learn more about thenatural history of small-bodied humans, we can weigh therelative importance of an array of conditions possibly asso-ciated with natural selection for that bauplan.

Acknowledgments

I thank the organizers of the Wenner-Gren 2011 spring sym-posium for including me in this very stimulating dialogue. Iam indebted to colleagues and to curators of South Africaninstitutions; they have guided me through study applicationsand details of archaeological context. They include JudithSealy, Chopi Jerardino, Helen Kurki, Lesley Harrington, AlanMorris, Johan Binneman, and James Brink. Jaime Ginterkindly provided unpublished data for figure 2. The SocialSciences and Humanities Research Council of Canada hassupported much of my work.

References CitedAnton, Susan C., Fred Spoor, Connie D. Fellmann, and Carl C. Swisher III.

2007. Defining Homo erectus: size considered. In Handbook of palaeoan-thropology, vol. 3. Wienfried Henke, Thorolf Rothe, and Ian Tattersall, eds.Pp. 1655–1693. Berlin: Springer.

Barham, Lawrence, and Peter Mitchell. 2008. The first Africans: African ar-chaeology from the earliest toolmakers to the most recent foragers. CambridgeWorld Archaeology. Cambridge: Cambridge University Press.

Becker, Noemie S. A., Paul Verdu, Barry S. Hewlett, and Samuel Pavard. 2010.Can life history trade-offs explain the evolution of short stature in humanpygmies? a response to Migliano et al. (2007). Human Biology 82:17–27.

Bernstein, Robin M. 2010. The big and the small of it: how body size evolves.Yearbook of Physical Anthropology 53:46–62.

Black, Wendy, Rebecca R. Ackermann, and Judith C. Sealy. 2009. Variationin Holocene KhoeSan dentition. American Journal of Physical Anthropology138(suppl. 48):92–93.

Blanckenhorn, Wolf U. 2000. The evolution of body size: what keeps organismssmall? Quarterly Review of Biology 75(4):385–407.

Brauer, Gunter, and Friedrich W. Rosing. 1989. Human biological history ofsouthern Africa. Rassengeschichte der Menscheit, Afrika II: Sudafrika. IlseSchwidezky, ed. Pp. 7–132. Munich: Oldenbourg.

Brown, James H. 1995. Macroecology. Chicago: University of Chicago Press.


Brown, Peter, Thomas Sutikna, Michael J. Morwood, Raden P. Soejono, Jat-miko, E. Wayhu Saptomo, and Rokus Awe Due. 2004. A new small-bodiedhominin from the late Pleistocene of Flores, Indonesia. Nature 431:1055–1061.

Carlson, Bryce A., and John D. Kingston. 2007. Docosahexaenoic acid, theaquatic diet, and hominin encephalization: difficulties in establishing evo-lutionary links. American Journal of Human Biology 19:132–141.

Carr, Andrew S., Mark D. Bateman, David L. Roberts, Colin V. Murray-Wallace, Zenobia Jacobs, and Peter J. Holmes. 2010. The last interglacialsea-level high stand on the southern Cape coastline of South Africa. Qua-ternary Research 73:351–363.

Cavalli-Sforza, Luigi Luca. 1986. African Pygmies. Orlando, FL: AcademicPress.

Charnov, Eric L. 1993. Life history invariants: some explorations of symmetryin evolutionary ecology. Oxford: Oxford University Press.

———. 2001. Evolution of mammal life histories. Evolutionary Ecology Re-search 3:521–535.

Chase, Brian M., Michael E. Meadows, Andrew S. Carr, and Paula J. Reimer.2010. Evidence for progressive Holocene aridification in southern Africarecorded in Namibian hyrax middens: implications for African monsoondynamics and the “African Humid Period.” Quaternary Research 74:36–45.

Churchill, Steven E., Osbjorn M. Pearson, Frederick E. Grine, Erik Trinkaus,and Trenton W. Holliday. 1996. Morphological affinities of the proximalulna from Klasies River main site: archaic or modern? Journal of HumanEvolution 31:213–237.

Clayton, Fiona H., Judith Sealy, and Susan Pfeiffer. 2006. Weaning age amongforagers at Matjes River Rock Shelter, South Africa, from stable nitrogenand carbon isotope analyses. American Journal of Physical Anthropology 129:311–317.

Cowling, Richard M., and Craig Hilton-Taylor. 1997. Phytogeography, floraand endemism. In Vegetation of southern Africa. Richard M. Cowling, DavidM. Richardson, and Shirley M. Pierce, eds. Pp. 43–61. Cambridge: Cam-bridge University Press.

Cox, Murray P., David A. Morales, August E. Woerner, Jesse Sozanski, JeffreyD. Wall, and Michael F. Hammer. 2009. Autosomal resequence data reveallate Stone Age signals of population expansion in sub-Saharan Africanforaging and farming populations. PLoS ONE 4:e6366.

Deacon, H. J., and Jeanette Deacon. 1999. Human beginnings in South Africa.Cape Town: Philip.

Dewar, Genevieve, and Susan Pfeiffer. 2004. Postural behavior of later StoneAge people in South Africa. South African Archaeological Bulletin 59:52–58.

Drennan, M. R. 1938. Archaeology of the Oakhurst shelter, George. 3. Thecave dwellers. Transactions of the Royal Society of South Africa 25:259–293.

Ekwo, Edem E., Carol Gosselink, Nancy Roizen, and Dana Brazdziunas. 2005.The effect of height on family income. American Journal of Human Biology3:181–188.

Farmer, E. Christa, Peter B. deMenocal, and Thomas M. Marchitto. 2005.Holocene and deglacial ocean temperature variability in the Benguela up-welling region: implications for low-latitude atmospheric circulation. Pale-oceanography 20, doi:10.1029/2004PA001049.

FitzSimons, Frederick W. 1926. Cliff dwellers of Zitzikama: results of recentexcavations. South African Journal of Science 23:813–817.

Froment, Alain. 2001. Evolutionary biology and health of hunter-gathererpopulations. In Hunter-gatherers: an interdisciplinary perspective. CatherinePanter-Brick, Robert H. Layton, and Peter Rowley-Conwy, eds. Pp. 239–266. Cambridge: Cambridge University Press.

Ginter, Jaime K. 2009. A bioarchaeological study of mid-Holocene commu-nities in the eastern Cape, South Africa: the interface between foraging andpastoralism. Azania: Archaeological Research in Africa 44:355, doi:10.1080/00672700903342428.

Graves, Ronda R., Robert C. McCarthy, Amy C. Lupo, Daniel J. Wescott, andDeborah L. Cunningham. 2010. New estimates of stature and body massfor KNM-WT 1500. American Journal of Physical Anthropology 50:115.

Grine, Frederick E., Richard M. Bailey, Katerina Harvati, R. P. Nathan, AlanG. Morris, Gideon M. Henderson, Isabelle Ribot, and Alistair W. G. Pike.2007. Late Pleistocene human skull from Hofmeyr, South Africa, and mod-ern human origins Science 315:226–229.

Gurven, Michael, and Kim Hill. 2010. Moving beyond stereotypes of men’sforaging goals: a reply to Hawkes, O’Connell, and Coxworth. Current An-thropology 51:265–267.

Harrington, Lesley. 2010. Ontogeny of post-cranial robusticity among Ho-locene hunter-gatherers of southernmost Africa. PhD dissertation, Univer-sity of Toronto.

Harrington, Lesley, and Susan Pfeiffer. 2007. Linear diaphyseal growth insouthern African forager children. American Journal of Physical Anthropology132(suppl. 44):125–126.

———. 2008. Juvenile mortality in southern African archaeological contexts.South African Archaeological Bulletin 63:95–101.

Hawkes, Kristen. 2010. How grandmother effects plus individual variation infrailty shape fertility and mortality: guidance from human-chimpanzeecomparisons. Proceedings of the National Academy of Sciences of the USA107:8977–8984.

Howell, Nancy. 2000. Demography of the Dobe !Kung. 2nd edition. New York:de Gruyter.

———. 2010. Life histories of the Dobe !Kung: food, fatness and well-being overthe lifespan. Berkeley: University of California Press.

Inskeep, Ray. 1986. A preliminary survey of burial practices in the Later StoneAge, from the Orange River to the Cape coast. In Variation, culture andevolution in African populations. Ronald Singer and John K. Lundy, eds. Pp.221–240. Johannesburg: Witwatersrand University Press.

Jerardino, Antonieta. 2010. Large shell middens in Lambert’s Bay, South Af-rica: a case of hunter-gatherer resource intensification. Journal of Archae-ological Science 37:2291–2302.

Jungers, William L., Susan G. Larson, William Harcourt-Smith, Michael J.Morwood, Thomas Sutikna, Rokus Awe Due, and Tony Djubiantono. 2009.Descriptions of the lower limb skeleton of Homo floresiensis. Journal ofHuman Evolution 57:538–554.

Jurmain, Robert D. 1991. Paleopathology of trauma in a prehistoric centralCalifornia population. In Human paleopathology: current syntheses and futureoptions. Donald J. Ortner and Arthur C. Aufderheide, eds. Pp. 241–248.Washington, DC: Smithsonian Institution.

Kilgore, Lynn, Robert D. Jurmain, and Dennis van Gerven. 1997. Palaeo-epidemiological patterns of trauma in a medieval Nubian skeletal popu-lation. International Journal of Osteoarchaeology 7:103–114.

Kruger, G. P. 1983. Terrain morphology map of southern Africa. Pretoria: Soiland Irrigation Research Institute. [In Afrikaans.]

Kurki, Helen K. 2007. Protection of obstetric dimensions in a small-bodiedhuman sample. American Journal of Physical Anthropology 133:1152–1165.

Kurki, Helen K., Jaime K. Ginter, Jay T. Stock, and Susan Pfeiffer. 2008. Adultproportionality in small-bodied foragers: a test of ecogeographic expecta-tions. American Journal of Physical Anthropology 136:28–38.

———. 2010. Body size estimation of small-bodied humans: applicability ofcurrent methods. American Journal of Physical Anthropology 141:169–180.

Kurki, Helen K., Deano D. Stynder, and Susan Pfeiffer. 2012. Allometry ofhead and body size in Holocene foragers of the South African Cape. Amer-ican Journal of Physical Anthropology 147:462–471.

Lam, Yin M., Osbjorn M. Pearson, and Cameron M. Smith. 1996. Chin mor-phology and sexual dimorphism in the fossil hominid mandible samplefrom Klasies River mouth. American Journal of Physical Anthropology 100:545–557.

Larson, Susan G., William L. Jungers, Matt W. Tocheri, Caley M. Orr, MichaelJ. Morwood, Thomas Sutikna, Rokus D. Awe, and Tony Djubiantono. 2009.Descriptions of the upper limb skeleton of Homo floresiensis. Journal ofHuman Evolution 57:555–570.

Lee, Richard B. 1979. The !Kung San: men, women and work in a foragingsociety. Cambridge: Cambridge University Press.

Legassick, Martin, and Ciraj Rassool. 2000. Skeletons in the cupboard: SouthAfrican museums and the trade in human remains, 1907–1917. Cape Town:South African Museum and McGregor Museum.

Liebenberg, Louis. 2006. Persistence hunting by modern hunter-gatherers.Current Anthropology 47:1017–1026.

Liversidge, Helen M., B. Holly Smith, and Melissa Maber. 2010. Bias andaccuracy of age estimation using developing teeth in 946 children. AmericanJournal of Physical Anthropology 143:545–554.

Lombard, Marlize, and Laurel Phillipson. 2010. Indications of bow and stone-tipped arrow use 64 000 years ago in KwaZulu-Natal, South Africa. Antiquity84:635–648.

Manhire, Anthony. 1998. The role of handprints in the rock art of the south-western Cape. South African Archaeological Bulletin 168:98–108.

Marean, Curtis W. 2010. When the sea saved humanity. Scientific American303:54–61.

Matzke, Lise. 2000. Dental development and maturation of juvenile dentitionsfrom the later Stone Age, South Africa. MSc major research paper, Uni-versity of Toronto.

McHenry, Henry M., and Cassandra C. Brown. 2008. Side steps: the erratic


pattern of hominin postcranial change through time. Journal of HumanEvolution 55:639–651.

Migliano, Andrea B. 2005. Why are Pygmies small? ontogenetic implicationsof life history evolution. PhD dissertation, Cambridge University.


Migliano, Andrea B., Lucio Vinicius, and Marta Mirazon Lahr. 2007. Lifehistory trade-offs explain the evolution of human pygmies. Proceedings ofthe National Academy of Sciences of the USA 104:20216–20219.

———. 2010. Why are Pygmies so short? a defense of Migliano’s hypothesis.Human Biology 82:109–113.

Mitchell, Peter. 2002. The archaeology of southern Africa. Cambridge: Cam-bridge University Press.

Morris, Alan G. 1992. A master catalogue: Holocene human skeletons from SouthAfrica. Johannesburg: Witwatersrand University Press.

———. 2002. Isolation and the origin of the Khoisan: late Pleistocene andearly Holocene human evolution at the southern end of Africa. HumanEvolution 17:231–240.

———. 2003. The myth of the East African Bushmen. South African Ar-chaeological Bulletin 58:85–90.

Morwood, Michael J., and William L. Jungers. 2009. Conclusions: implicationsof the Liang Bua excavations for hominin evolution and biogeography.Journal of Human Evolution 57:640–648.

Neves, Walter A., A. Marcus Barros, and Maria Antonietta Costa. 1999. In-cidence and distribution of post-cranial fractures in the prehistoric pop-ulation of San Pedro de Atacama, northern Chile. American Journal ofPhysical Anthropology 109:253–258.

Parkington, John. 1998. Resolving the past: gender in the archaeological recordof the western Cape. In Gender in African prehistory. Susan Kent, ed. Pp.25–38. Walnut Creek, CA: Altamira.

Parkington, John E. 2003. Follow the San. Cape Town: Krakadouw Trust.Pearson, Osbjorn M., Steven E. Churchill, Frederick E. Grine, Erik Trinkaus,

and Trenton W. Holliday. 1998. Multivariate analyses of the hominid ulnafrom Klasies River mouth. Journal of Human Evolution 34:653–656.

Pearson, Osbjorn M., and Frederick E. Grine. 1996. Morphology of the BorderCave hominid ulna and humerus. South African Journal of Science 92:231–236.

———. 1997. Re-analysis of the hominid radii from Cave of Hearths andKlasies River mouth, South Africa. Journal of Human Evolution 32:577–592.

Perry, George H., and Nathaniel J. Dominy. 2008. Evolution of the humanPygmy phenotype. Trends in Ecology & Evolution 24:218–225.

Pfeiffer, Susan. 2007. The health of foragers: people of the Later Stone Age,southern Africa. In Ancient health: skeletal indicators of agricultural andeconomic intensification. Mark N. Cohen and Gillian Crane-Kramer, eds.Pp. 223–236. Gainesville: University Press of Florida.

———. 2011. Pelvic stress injuries in a small-bodied forager. InternationalJournal of Osteoarchaeology 21:694–703.

———. 2013. Population dynamics in the southern African Holocene: humanburials from the west coast. In The archaeology of the west coast of SouthAfrica. Antonieta Jerardino, David Braun, and Antonia Malan, eds. Johan-nesburg: Witwatersrand University Press. Forthcoming.

Pfeiffer, Susan, and Christian Crowder. 2004. An ill child among mid-Holoceneforagers of southern Africa. American Journal of Physical Anthropology 123:23–29.

Pfeiffer, Susan, and Lesley Harrington. 2010. Child growth among southernAfrican foragers. In Human diet and nutrition in biocultural perspective. TinaMoffat and Tracey Prowse, eds. Pp. 35–56. Studies of the Biosocial Society,vol. 5. Oxford: Berghahn.

Pfeiffer, Susan, and Judith Sealy. 2006. Body size among Holocene foragersof the Cape ecozone, southern Africa. American Journal of Physical An-thropology 129:1–11.

———. 2009. Regional patterns among Holocene hunter-gatherers of south-ern Africa. American Journal of Physical Anthropology 138(suppl. 48):210.

Pfeiffer, Susan, and Nikolaas J. van der Merwe. 2004. Cranial injuries to LaterStone Age children from the Modder River mouth, southwestern Cape,South Africa. South African Archaeological Bulletin 59:59–65.

Pfeiffer, Susan, and Marie Zehr. 1996. A morphological and histological studyof the human humerus from Border Cave. Journal of Human Evolution 31:49–59.

Quick, Lynne J., Brian M. Chase, Michael E. Meadows, Louis Scott, and PaulaJ. Reimer. 2011. A 19.5 kyr vegetation history from the central Cederberg

Mountains, South Africa: palynological evidence from rock hyrax middens.Palaeogeography, Palaeoclimatology, Palaeoecology 309:253–270.

Reid, Donald J., and M. Christopher Dean. 2006. Variation in modern humanenamel formation times. Journal of Human Evolution 50:329–346.

Reno, Philip, Melanie McCollum, Richard Meindl, and C. Owen Lovejoy.2010. An enlarged postcranial sample confirms Australopithecus afarensisdimorphism was similar to modern humans. Philosophical Transactions ofthe Royal Society B 365:3355–3363.

Rightmire, G. Philip. 1978. Human skeletal remains from the Southern CapeProvince and their bearing on the Stone Age prehistory of South Africa.Quaternary Research 9:219–230.

———. 1984. The fossil evidence for hominid evolution in southern Africa.In Southern African prehistory and paleoenvironments. Richard G. Klein, ed.Pp. 147–168. Rotterdam: Balkema.

———. 2008. Homo in the middle Pleistocene: hypodigms, variation, andspecies recognition. Evolutionary Anthropology 17:8–21.

Rightmire, G. Philip, and Hilary J. Deacon. 1991. Comparative studies of latePleistocene human remains from Klasies River mouth, South Africa. Journalof Human Evolution 20:131–156.

———. 2001. New human teeth from Middle Stone Age deposits at KlasiesRiver, South Africa. Journal of Human Evolution 41:535–544.

Rightmire, G. Philip, Hilary J. Deacon, Jeffrey H. Schwartz, and Ian Tattersall.2006. Human foot bones from Klasies River main site, South Africa. Journalof Human Evolution 50:96–103.

Roberts, Charlotte, and Keith Manchester. 1995. The archaeology of disease.2nd edition. Ithaca, NY: Cornell University Press.

Ruff, Christopher. 2010. Body size and body shape in early hominins: im-plications of the Gona Pelvis. Journal of Human Evolution 58:166–178.

Scheinfeldt, Laura B., Sameer Soi, and Sarah A. Tishkoff. 2010. Workingtoward a synthesis of archaeological, linguistic and genetic data for inferringAfrican population history. Proceedings of the National Academy of Sciencesof the USA 107:8931–8938.

Schell, Lawrence M., and P. D. Magnus. 2007. Is there an elephant in theroom? addressing rival approaches to the interpretation of growth pertur-bations and small size. American Journal of Human Biology 19:606–614.

Schulze, Roland E. 1997. South African atlas of agrohydrology and climatology.Pretoria: Water Research Commission.

Schuster, Stephen C., Webb Miller, Aakrosh Ratan, Lynn P. Tomsho, BelindaGiardine, Lindsay R. Kasson, Robert S. Harris, et al. 2010. Complete Khoi-san and Bantu genomes from southern Africa. Nature 463:943–947.

Scott, Louis, Frank H. Neumann, George A. Brook, C. Britt Bousman, ElinNorstrom, and A. A. Metwally. 2012. Terrestrial fossil-pollen evidence ofclimate change during the last 26 thousand years in southern Africa. Qua-ternary Science Reviews 32:100–118.

Sealy, Judith. 2006. Diet, mobility and settlement pattern among Holocenehunter-gatherers in southernmost Africa. Current Anthropology 47:569–595.

———. 2010. Isotopic evidence for the antiquity of cattle-based pastoralismin southernmost Africa. Journal of African Archaeology 8:65–81.

Sealy, Judith, and Susan Pfeiffer. 2000. Diet, body size and landscape useamong Holocene peoples in the Southern Cape, South Africa. Current An-thropology 41:642–655.

Sealy, Judith, Susan Pfeiffer, Royden Yates, Kathy Willmore, Anthony Manhire,Tim Maggs, and John Lanham. 2000. Hunter-gatherer child burials fromthe Pakhuis Mountains, western Cape: growth, diet and burial practices inthe late Holocene. South African Archaeological Bulletin 55:32–43.

Sealy, Judith C., and Nikolaas J. van der Merwe. 1988. Social, spatial andchronological patterning in marine food use as determined by 13C mea-surements of Holocene human skeletons from the south-western Cape,South Africa. World Archaeology 20:87–102.

Sealy, Judith C., Nikolaas J. van der Merwe, Julia A. LeeThorp, and John L.Lanham. 1987. Nitrogen isotopic ecology in southern Africa: implicationsfor environmental and dietary tracing. Geochemica et Cosmochimica Acta51:2707–2717.

Sinclair, David, and Peter Dangerfield. 1998. Human growth after birth. 6thedition. Oxford: Oxford University Press.

Singer, Ronald. 1958. The Boskop “race” problem. Man 232:173–178.Smith, B. Holly. 1991. Standards of human tooth formation and dental age

assessment. In Advances in dental anthropology. Mark A. Kelley and ClarkSpencer Larsen, eds. Pp. 143–168. New York: Wiley-Liss.

Snodgrass, Joshua J., William R. Leonard, and Marcia L. Robertson. 2009.The energetics and encephalization in early hominids. In The evolution ofhominin diets: integrating approaches to the study of Paleolithic subsistence.


Jean-Jacques Hublin and Michael P. Richards, eds. Pp. 15–29. Dordrecht:Springer.

Spoor, Fred, Meave G. Leakey, Patrick N. Gathogo, Frank H. Brown, SusanC. Anton, Ian McDougall, Christopher Kiarie, Frederick K. Manthi, andLouise N. Leakey. 2007. Implications of new early Homo fossils from Ileret,east of Lake Turkana, Kenya. Nature 448:688–691.

Stock, Jay, and Susan K. Pfeiffer. 2004. Long bone robusticity and subsistencebehaviour among later Stone Age foragers of the forest and fynbos biomesof South Africa. Journal of Archaeological Science 31:999–1013.

Stynder, Deano D. 2006. A quantitative assessment of variation in HoloceneKhoeSan crania from South Africa’s western, south-western, southern andsouth-eastern coasts and coastal forelands. PhD dissertation, University ofCape Town.

Stynder, Deano D., Rebecca R. Ackermann, and Judith C. Sealy. 2007a. Cran-iofacial variation and population continuity during the South African Ho-locene. American Journal of Physical Anthropology 134:489–500.

———. 2007b. Early to mid-Holocene South African Later Stone Age humancrania exhibit a distinctly KhoeSan morphological pattern. South AfricanJournal of Science 103:349–352.

Tishkoff, Sarah A., Erin Dietzsch, William Speed, Andrew J. Pakstis, JudithR. Kidd, K. Cheung, and Batsheva Bonne-Tamir, et al. 1996. Global patternsof linkage disequilibrium at the CD4 locus and modern human origins.Science 271:1380–1387.

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.

Truswell, A. Stewart, and John D. L. Hanson. 1976. Medical research amongthe !Kung. In Kalahari hunter-gatherers. Richard B. Lee and Irven DeVore,eds. Pp. 166–194. Cambridge, MA: Harvard University Press.

Waddington, Conrad H. 1957. The strategy of the genes. London: Allen &Unwin.

Wadley, Lyn. 1998. The invisible meat providers: women in the Stone Age ofSouth Africa. In Gender in African prehistory. Susan Kent, ed. Pp. 69–82.Walnut Creek, CA: Altamira.

Wadley, Lyn, Tamaryn Hidgskiss, 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.

Walker, Robert S., Michael Gurven, Kim Hill, Andrea Migliano, NapoleonChagnon, Roberta De Souza, Gradimir Diurovic, et al. 2006. Growth ratesand life histories in twenty-two small-scale societies. American Journal ofHuman Biology 18:295–311.

Walker, Robert S., and Marcus J. Hamilton. 2008. Life-history consequencesof density dependence and the evolution of human body size. CurrentAnthropology 49:115–122.

Wilmsen, Edwin N. 1989. Land filled with flies: a political economy of theKalahari. Chicago: University of Chicago Press.



Growth, Development, and Life Historythroughout the Evolution of Homo

by Gary T. Schwartz

For over a century, paleoanthropologists have listed the presence of prolonged periods of gestation, growth, andmaturation, extremely short interbirth intervals, and early weaning among the key features that distinguish modernhumans from our extant ape cousins. Exactly when and how this particular scheduling of important developmentalmilestones—termed a “life history profile”—came to characterize Homo sapiens is not entirely clear. Researchershave suggested that the modern human life history profile appeared either at the base of the hominin radiation (ca.6 Ma), with the origins of the genus Homo (ca. 2.5 Ma), or much later in time, perhaps only with H. sapiens (ca.200–100 Ka). In this short review, evidence of the pace of growth and maturation in fossil australopiths and earlymembers of Homo is detailed to evaluate the merits of each of these scenarios. New data on the relationship betweendental development and life history in extant apes are synthesized within the context of life history theory anddevelopmental variation across modern human groups. Future directions, including new analytical tools for extractingmore refined life history parameters as well as integrative biomechanical and developmental models of facial growthare also discussed.

Introduction

Compared with our closest living relatives—the extant greatapes—the sole surviving member of the genus Homo possessesa suite of features that make us quite distinct, including un-usually large brains, obligate bipedality, a reliance on the pro-duction and use of tools, and a strikingly different life history.Generally, patterns of mammalian growth, development, andlife history are thought of as lying along a spectrum some-where between two end points that are colloquially referredto as “live fast, die young” and “live slow, die old.” Modernhuman life history incorporates elements of both schedules:long gestation periods, altricial offspring, enlarged brains,slow maturation rates, increased life span, and protracted pe-riods of offspring dependence are suggestive of a “live slow”strategy, whereas relatively early weaning, short interbirth in-tervals, and the ability to overlap births (resulting in the pres-ence of multiple offspring) are suggestive of a “live fast” sched-ule. Of interest to paleoanthropologists is whether this“modern human life history package” evolved as a single de-velopmental module or accumulated in a mosaic fashion(with different attributes appearing at different points intime). Furthermore, it is of great interest to understand when

Gary T. Schwartz is Associate Professor of Anthropology at theInstitute of Human Origins at Arizona State University (900 SouthCady Mall, Tempe, Arizona 85287-2402, U.S.A. [[email protected]]).This paper was submitted 12 XII 11, accepted 28 VI 12, andelectronically published 20 XI 12.

this transition occurred and thus determine whether the hu-man life history package appeared as part of a suite of fun-damental adaptations at the base of the hominin clade;whether it evolved somewhat later, perhaps tied to the re-organization of the cranium and postcranium that charac-terized the earliest members of the genus Homo; or whetherit appeared even later still, perhaps in the last hundred thou-sand years or so.

The last decade has been marked by a tremendous amountof research into reconstructing the pattern of growth, devel-opment, and life history of extant great apes, australopiths,and early to later members of the genus Homo. Novel ana-lytical techniques, imaging modalities, and hard-fought ob-servational data from naturalistic studies of great apes cannow be synthesized to paint a broad view of the evolution oflife history throughout the course of the human story. Thegoal of this paper is to review the current state of knowledgeof the evolution of human life history within the comparativecontext of what we know about these attributes in populationsof extant hominoids and fossil hominins. These data will beevaluated within the light of what is known about primatelife history, ecology, diet, and so forth, and will be used tohelp suggest future avenues of inquiry into studies of homininlife history.

What Is Life History?

By marrying the principles of organic evolution with thoseof theoretical population ecology, life history theory seeks tounderstand the general rules that account for the tremendous



variation in life cycles across all organisms (Stearns 1992). Inshort, life history theory views variation in the pattern, se-quence, and pace of growth as the outcome of how naturalselection operates on a series of trade-offs in the allocationof an organism’s energetic budget. Smith and Tompkins(1995:257) emphasized the importance of these trade-offs indefining life history as the allotment of an organism’s energy“towards, growth, maintenance, reproduction, raising off-spring to independence, and avoiding death.” Bogin (1988:154) suggested that the life history of a species can be viewedas a strategy that determines “when to be born, when to beweaned, how many and what type of prereproductive stagesof development to pass through, when to reproduce, andwhen to die.” Central to these and other definitions is thenotion that energy is a limiting commodity that is distributedtoward growth, maintenance (of tissues), and/or reproductionthroughout the lives of individuals (Bogin and Smith 2000;Roff 2002; Stearns 1992). In a strict Darwinian sense, selectionshould favor an apportionment of energy in ways that reducemortality and maximize fecundity. Thus, the application oflife history theory to ontogenetic studies seeks to understandhow the scheduling of key events in an organism’s life cycle(including but not limited to gestation length, age at weaning,interbirth interval, timing of maturation, age at first repro-duction, frequency of reproduction, fecundity, and life span)better enable some individuals of a species to minimize mor-tality risks more effectively and thus increase overall fitness.This scheduling or sequence of events can be thought of asa “life history profile” and is, effectively, the product of howdevelopmental variables (e.g., growth rates, age at skeletalmaturation) interact with demographic variables (e.g., sur-vival, reproduction, population growth) to influence individ-ual survival (Godfrey, Petto, and Sutherland 2002; Ross 1998).

Aside from being viewed as an energetic trade-off, a species’life history profile is directly linked to rates of extrinsic mor-tality, which is defined as the risk of death as a result ofenvironmental conditions such as predation, disease, acci-dents, and so forth (Stearns 1992). High extrinsic mortalityshould favor shorter life spans, while lower extrinsic mortalityresults in a larger proportion of the population surviving toolder ages. Compared with great apes, human populationsare able to ramp up reproductive success because our lifehistory profile is characterized by lower mortality rates, whichhas the effect of slowing down rates of maturation, delayingreproduction, and thus spreading it out across later years. Fora large primate, a strategy including a prolongation of growthand development comes with some risk, as evidenced by re-cent data on the inability of orangutan populations (with lateages of reproductive maturity and interbirth intervals of ∼7years) to recover from even slight reductions in populationdensity (Knott, Thompson, and Wich 2009). At a finer scaleof comparison (i.e., across populations of modern Homo sap-iens), differences in subsistence strategies, environments, andmortality rates comingle to produce tremendous variation inpatterns of human ontogeny (e.g., Migliano and Guillon

2012). For instance, in a survey of 22 small-scale societies,Walker et al. (2006) found that populations experiencing lowsurvivorship (i.e., high mortality) during the subadult yearswere characterized by an overall pattern of accelerated de-velopment and thus reached the important developmentalmilestones of puberty, menarche, and first reproduction atrelatively earlier ages. Interestingly, new evidence suggests thatadverse early life conditions in humans, ranging from thebiological (low birth weight for gestational age, breast-feedingduration) to the psychosocial (separation anxiety, family res-idential relocation, degree of parental involvement) mediatereproductive scheduling by accelerating the age at first preg-nancy (Nettle, Coall, and Dickins 2010 and referencestherein). A more detailed understanding of exactly how suchenvironmental constraints, broadly speaking, shape life his-tory variation across human populations would be extremelyinformative. Similarly, uncovering the role developmentalplasticity (the capacity of an individual to modify its ontogenyin response to shifting environmental conditions on a fairlyrapid timescale—days, months, or a few years) plays in gen-erating novel phenotypes that enhance the evolutionary po-tential, or “evolvability,” of developmental systems may helpilluminate the process(es) whereby selection assembled thetotal package of modern human life history attributes over aperiod of several hundred thousand years (e.g., see Kuzawaand Bragg 2012; West-Eberhard 2003).

Reconstructing Hominin Life History Profiles

When viewed in the light of life history theory, it might seemimpossible to infer such reproductive, physiological, demo-graphic, environmental, and behavioral parameters from fos-silized remains. However, the vast majority of the homininfossil record comprises isolated teeth and dentognathic re-mains, and many of the important life history variables dis-cussed above are tightly linked with aspects of developingdentitions. As a result, studies on the timing of particulardental developmental events figure prominently in paleoan-thropological investigations that aim to reconstruct the lifehistory profiles of fossil primates and hominins (e.g., Beynonet al. 1998; Bromage and Dean 1985; Conroy and Kuykendall1995; Conroy and Vannier 1991a, 1991b; Dean et al. 1993,2001; Godfrey et al. 2001; Mann 1975; Robson and Wood2008; Schwartz et al. 2005; Smith 1989, 1993; Smith, Crum-mett, and Brandt 1994; Smith, Gannon, and Smith 1995;Smith et al. 2010a, 2010b; Zihlman, Bolter, and Boesch 2004).

Ever since the pioneering studies on primate growth bySchultz (1935, 1940, 1960) and Sacher (1959, 1975, 1978;Sacher and Staffeldt 1974), evolutionary biologists and paleo-anthropologists have linked aspects of somatic and neuralgrowth rates to reproduction, metabolism, and life span inan attempt to reconstruct aspects of early hominin matura-tion. The first study to flesh out the relationship between thesequence and pace of dental development and aspects of lifehistory was by Schultz (1949), who observed that the per-

Schwartz Life History Evolution in Homo S397

manent replacement molars emerge into the oral cavity beforethe shedding of deciduous teeth in faster-growing primates.This phenomenon has since been dubbed “Schultz’s rule”(Smith 2000) and accurately relates dental emergence se-quences to maturation rates across primates as a whole, es-pecially anthropoids (see Godfrey et al. 2005). It also allowedpaleoanthropologists an opportunity to evaluate when themodern human developmental pattern (i.e., tooth crown de-velopment and emergence sequence) first appeared duringthe course of human evolution as it could be used to assessthe relative developmental status of juvenile hominins (Smith1986). Along with other studies examining the sequence ofdental development and emergence in hominins (e.g., Conroyand Kuykendall 1995; Conroy and Vannier 1991a, 1991b), itbecame increasingly clear that the earliest hominins, and in-deed even fossil species of the genus Homo, were characterizedby dental developmental patterns, and thus maturational pro-files, more similar to extant apes (at the time, meaning pre-dominantly Pan troglodytes). This work was extremely im-portant because the prevailing paradigm held that allhominins, including the earliest australopiths, possessed amodern humanlike maturation profile with the attendant pro-longed infant and childhood dependency that was so fun-damental to producing the social, cognitive, and cultural com-plexity that serves as the hallmark of our species (Mann 1975).

Given the intimate relationship between brain size, lifespan, and rates of maturation and the plasticity of certainreproductive parameters, Smith (1989) reasoned that the den-tition should be one of the more stable markers of growthgiven its high heritability and resistance to environmentalperturbations. She conducted a broad, interspecific compar-ison of dental maturation with various life history variablesand revealed the extremely high correlation between brainsize and the age at first molar (M1) emergence on the onehand and between M1 emergence age and various life historyvariables related to reproduction (gestation length, ages atweaning and first breeding, life span) on the other. This sem-inal study laid the groundwork for linking aspects of thetiming or pace of dental development to critical componentsof a species’ life history and, equally importantly, provided achronological marker for probing the maturation rates of fos-sils if information on the timing of key dental developmentevents (i.e., M1 emergence) could somehow be retrieved fromthe fossil record.

At around the same time, a group of paleoanthropologistsbegan to mine the rich vein of growth data contained withinteeth. Based on some foundational work in dental hard tissuebiology (Boyde 1963, 1964), a series of investigations beganto appear that documented how to retrieve information onthe absolute timing of dental development utilizing thegrowth record contained within enamel and dentine. The cellsthat secrete the dental tissues enamel and dentine (for reasonsof brevity, cementum is not discussed here) leave a record oftheir activity in the form of short- and long-period incre-mental growth lines. These include, respectively, daily cross

striations and Retzius lines in the enamel and the correspond-ing von Ebner and Andresen lines in dentine (Dean 1987,2006; Schwartz and Dean 2000; Smith 2006, 2008). Carefulcounts of these lines reveal the time taken to form a tooth,including the tooth crown and however much root hadformed at the time of death, thereby providing a detailedchronology of dental development that can, under the rightcircumstances, also yield precise ages for key events such asmolar emergence (e.g., Beynon, Dean, and Reid 1991; Deanet al. 2001; Dirks 1998; Kelley and Schwartz 2010; Smith,Reid, and Sirianni 2006; Smith et al. 2010a, 2010b).

Data on molar emergence ages in hominins, while rare, arebecoming more accessible, especially with the application ofnew, noninvasive imaging modalities that allow access to theinternal dental growth record (e.g., Smith et al. 2007b, 2007c,2010b). While it is critically important to establish a growingdatabase on molar emergence ages in key hominin taxa, it isalso necessary to chart more fully dental developmental var-iation in extant hominoids and to interpret that variation inlight of a particular species’ population ecology, demography,and life history. For now, good population data on molaremergence ages exist only for P. troglodytes (Anemone, Moo-ney, and Siegel 1996; Conroy and Mahoney 1991; Kuykendall,Mahoney, and Conroy 1992; Nissen and Reisen 1945, 1964;Reid et al. 1998; Schultz 1940; Smith et al. 2007a, 2010a;Zihlman, Bolter, and Boesch 2004; though see Dirks 2003 andDirks and Bowman 2007 for individual hylobatids; and Bey-non, Dean, and Reid 1991; Winkler, Schwartz, and Swindler1991; and Kelley and Schwartz 2010; and Willoughby 1978for individual gorillas and orangutans). There is also a grow-ing awareness that emergence ages for captive primate pop-ulations may be slightly advanced compared with wild pop-ulations (e.g., Kelley and Schwartz 2010; Smith and Boesch2011; Zihlman, Bolter, and Boesch 2004), suggesting at thevery least some caution in relying on databases derived ex-clusively from captive colonies. Expanding our knowledge ofdental developmental variation across natural fertility pop-ulations of modern humans is another key element (e.g., Liv-ersidge 2003) and will ultimately allow more fine-scale testsof how population-level variation in aspects of dental devel-opment (e.g., M1 emergence age) relates to that for variouslife history attributes.

Early Homo

It has generally been viewed that the origin of the genus Homowas characterized by a trend away from bipedal apelike formsto obligate terrestrial bipeds who were endowed with muchlarger brains and the capacity to manufacture and use stonetool technology and who also exhibited a shift in dietary andforaging adaptations. Fossil representatives of the genus Homowere first described by Leakey, Tobias, and Napier (1964) frommaterial derived from Bed I, Olduvai Gorge, Tanzania, andwere dated to 1.8–1.7 Ma. These authors viewed the materialas distinct from australopith material given its possession of


a larger, more globular and gracile cranium with a cranialcapacity 1600 cm3 and a concomitant reliance on the habitualproduction and use of lithic technology. Fossil evidence foreven earlier representatives of Homo includes mostly isolatedspecimens from South Africa (Sterkfontein, ca. 2.6 Ma),Kenya (Chemeron, 2.4 Ma), and Malawi (Uraha, ∼2.5–1.9Ma), with a maxilla from Hadar, Ethiopia (A.L. 666-1, 2.3Ma) being the best candidate for the earliest Homo (Kimbel,Johanson, and Rak 1997; see recent review in Kimbel 2009),thereby extending the origin of Homo further back in timeto at least 2.5 Ma, to a time when Africa was undergoing atransition toward cooler, drier conditions with an increase inmore open habitats (see also Potts 2012). Regardless of theprecise time and location of the origins of our own genus—though it is generally held to be within the critical intervalof 3–2.0 Ma—many paleoanthropologists support the usageof Leakey, Tobias, and Napier’s morphological-behavioralcomplex (increased brain size, stone tool production) as thesole criterion for inclusion within the genus and would there-fore recognize three securely attributed representative speciesof premodern early Homo in Africa: Homo habilis (1.9–1.4Ma), Homo rudolfensis (1.9–1.8 Ma), and Homo ergaster andHomo erectus (1.9–0.9 Ma; see Anton 2012; Kimbel 2009; andWood and Lonergan 2008 for recent reviews of the fossilevidence of Homo).

Not everyone would include the “transitional” species ofH. habilis and H. rudolfensis within the genus Homo. Woodand Collard (1999) maintained that all members of a genusshould occupy the same adaptive zone (and thus possess asimilar adaptive strategy) and proposed a set of criteria forthe inclusion of any species into the genus Homo. Based onits large body size, body proportions, reduced dentition, andcommitment to long-range bipedality, they recommend thatH. ergaster be the earliest hominin to satisfy their adaptivecriteria, thereby relegating both of the earlier “Homo” speciesto the genus Australopithecus. Recent metric analyses, how-ever, demonstrate similarities in some of these anatomicalcomplexes (limb lengths and proportions) and suggest thatreassigning these species may be unwarranted (Holliday2012).

Regardless of generic assignations, many of the homininsbefore 2.5 Ma can be broadly characterized as relatively small-brained, large-toothed, non–stone tool producing human an-cestors (though recent work suggests that some australopithtaxa may have been using stone tools; McPherron et al. 2010).As such, it may seem reasonable to postulate some sort ofgrade shift in life history during this critical interval. Alter-natively, species of early Homo (and Australopithecus for thatmatter) might each possess slightly different life histories, asthese profiles are closely calibrated to local environment andecologies, and the critical time interval for the evolution ofHomo (3–2.0 Ma) is characterized by magnified climate var-iability and thus variable adaptive settings (deMenocal 1995;Potts 1996, 2012).

What Do We Know about Life History in Early Homo?

In short, not as much as we would like to know. This is inpart because an organism’s “strategy” for parsing out energyfor purposes of growth, maintenance, and/or reproduction(i.e., their gestation length, age at weaning, interbirth interval,timing of maturation, frequency of reproduction, fecundity,and life span) is not a durable part of the fossil record. Im-portantly, there are now some good attempts at deriving ges-tation length, interbirth intervals, and weaning age from hardtissue remains for certain primate taxa (see Dirks et al. 2010;Humphrey et al. 2008a, 2008b; Schwartz et al. 2002), thoughhow fruitful they will be if applied to fossil human taxa iscurrently unknown.

For several decades, paleoanthropologists have viewed theevolution of modern human growth, development, and lifehistory through the lens of a comparative dichotomy betweenmodern Pan on the one hand and modern humans on theother. This is not unreasonable given that recent DNA analysessupport a Pan-Homo clade to the exclusion of all other hom-inoids (e.g., Bradley 2008; Ruvolo 1994). It is interesting,however, that several early hominin specimens exhibit strikinganatomical similarities with extant gorillas (Dean 2010). Forinstance, certain aspects of scapular morphology in the re-cently discovered Dikika skeleton and the mandibular mor-phology of other Australopithecus afarensis specimens bear aclose resemblance to the condition in extant Gorilla (Alem-seged et al. 2006; Rak, Ginzburg, and Geffen 2007). Unfor-tunately, not enough is currently known about developmentin Gorilla, or developmental variation among Gorilla spp., tospeculate on the importance of this for understanding howbest to model the mosaic pattern of great ape morphology/dental development evident within australopiths and perhapsearly Homo (Dean 2010).

What we have been able to reconstruct about life historyin early Homo, and indeed for several earlier and later hom-inins, is based on estimates of the pattern and pace of dentaldevelopment or on the tight association between brain sizeand life history (or some skeletal correlate of life history).Thus, a series of key questions related to understanding theontogeny of early Homo can now be asked. Is there evidencethat the earliest hominins (including the earliest members ofthe genus Homo) matured in a way—had a chronology ofdental development—that was similar to extant chimpanzees(and moreover, do all extant apes mature, dentally, in anidentical manner)? Do the early hominins differ in the chro-nology of tooth emergence in ways that might suggest de-velopmental heterogeneity in their reconstructed life historyprofiles? Do the earliest representatives of the genus Homoexhibit dental developmental chronologies that align themmore closely with the penecontemporaneous australopiths orwith geologically younger Neanderthals and archaic Homosapiens populations?

Inferences from extant ape development. We know more aboutchimpanzee growth, development, diet, and ecology than we


Table 1. Comparative life history and M1 emergence age (years) in extant great apes and hu-mans

Variable Gorilla Pan Pongo Homo

Age at first reproduction 10.1a (51.3) 14.3b (72.6) 15.7c (79.7) 19.7Interbirth interval 4.3 5.8b 6.9c 3.4d

Survivorshipe 20.6f (38.1) 29.7 (54.9) 43.0 (79.5) 54.1Age at M1 emergence 3.8 (65.5) 4.0b (69.0) 4.6c (79.3) 5.8Cranial capacity (cm3) 484 383 379 1,293

Note. Numbers in parentheses indicate the percentage relative to the values in modern humans. References for lifehistory, survivorship, and M1 emergence data are reported in Kelley and Schwartz (2010).a Value for mountain gorilla Gorilla gorilla beringei, which is likely to be earlier than in Gorilla gorilla gorilla.b Values for Pan troglodytes verus only (Taı Forest, Ivory Coast).c Values for Pongo pygmaeus pygmaeus (i.e., orangutans from Borneo) only.d Interbirth interval is anomalously low in modern humans compared with other anthropoid species. This lifehistory variable is included for between-ape comparisons only, and percentages of human values were not calculated.e Expected age at death at age 15 years based on empirically derived survivorship curves.f Average of female (24.8 years, or 45.8%) and male (16.4 years, or 30.3%) values. Courtesy of Anne Bronikowskiand the Dian Fossey Gorilla Fund International.

do for any other ape, and with the exception of baboons,perhaps for any other primate. Dean (2010) recently synthe-sized all that is known about dental development in Pan,revealing some interesting similarities with and differencesfrom modern humans. Overall, chimpanzee dental develop-ment occurs along an accelerated schedule, taking ∼12 yearscompared with 18 years in modern humans. This accelerationis reflected in advanced median molar emergence ages thatoccur well before those of humans. The first mean gingivalemergence ages for chimp M1s were reported to be 3.3 years(range of 2.6–3.8 years; Nissen and Reisen 1964). After threemore decades of studies on captive chimps, median emergenceages for M1s are reconstructed as quite close to that originalestimate, at 3.2 years (M1 range, 2.26–4.38 years; M1 range,2.14–3.99 years; Kuykendal, Mahoney, and Conroy 1992).Thus, for several decades, it was generally taken that M1emergence in Pan, and thus for African apes as a whole,occurred somewhere in the range of 3.0–3.5 years. By com-parison, the range for modern humans calculated acrossglobal populations averages 4.7–7.0 years (Liversidge 2003).

Given the presence of the so-called “wild effect” (sensuSmith and Boesch 2011; see also Hamada et al. 1996; Kimuraand Hamada 1996), it became critical to evaluate gingivalemergence ages in wild populations. To date, the only reliableestimate for age at M1 emergence in noncaptive apes is asingle individual of Pan troglodytes verus at approximately 4.1years for the maxillary M1. The likely age of emergence forthe mandibular M1 in this individual was approximately 3.8–3.9 years, resulting in a combined age of M1 emergence ofapproximately 4.0 years (Smith et al. 2010a; Zihlman, Bolter,and Boesch 2004). Interestingly, a recent analysis by Dean(2010) using data from the histological growth record ofcrowns and roots found that the predicted mean age of at-tainment for molar emergence in Pan M1 is 4.1 years, nearlycoincident with that reported in the wild P. troglodytes verusindividual.

To date, no reliable gingival emergence data exist for any

great ape species other than the common chimpanzee. Thisdeficiency may limit the accuracy and reliability of life historyreconstructions for fossil hominins, and so it is critical toobtain M1 emergence data for both African and Asian apesand to obtain these data from noncaptive animals. Recently,reliable ages at M1 emergence were reported for orangutan(Pongo pygmaeus pygmaeus, 4.6 years) and gorilla (Gorillagorilla gorilla, 3.8 years) obtained from wild-shot individualsin museum osteology collections (Kelley and Schwartz 2010).These data offer support for the likelihood of a later averageage at M1 emergence in free-living chimpanzees than in cap-tive animals. Although limited, they also suggest that the av-erage age at M1 emergence in noncaptive extant great apesranges from just younger than 4 years to just older than 4.5years, or approximately 1 year later than the conventionalrange reported as ∼3.0–3.5 years.

These new comparative data allow an evaluation of justhow consistent M1 emergence data are with the comparativelife histories of extant Asian and African apes and modernhumans. As can be seen from table 1, the new ape emergencedata fit well with expectations based on the comparative lifehistories of living hominoids both in relation to one anotherand in comparison with that of modern humans. Ages at M1emergence between 3.8 and 4.6 years for great apes represent∼65%–80% of modern human emergence at ∼6 years( years, mean � 1 SD), signaling a close fit between6.2 � 0.8ages at attainment (or duration) of some key life history eventsin great apes, as these life history attributes are also ∼60%–80% of the modern human value. As a side note, and bycomparison, gingival emergence ages of 3.0–3.5 in extant apeswould represent only ∼50%–60% of the average modern hu-man value. These new data reinforce earlier studies (Smith1989) that identified dental eruption as a reliable means bywhich to reconstruct life history profiles in extinct hominins.

Unfortunately, reliable ages at M1 emergence are availablefor a very small number of fossils. It is difficult to extractthese data given the relatively few specimens that died during


Table 2. Estimated ages (years) at death and M1 emergence in great apes, australopiths,and Homo

Species (specimen) Estimated age at death Age at M1 emergence

Great apes:Pongo pygmaeus pygmaeus 4.6Gorilla gorilla beringei 3.8Pan troglodytes verus 4.0

Australopiths:Australopithecus afarensis (LH 2) 3.25 2.9*Australopithecus africanus (Sts 24) 3.30 2.9*A. africanus (Taung 1) 3.73–3.93 3.3–3.5*Paranthropus robustus (SK 62) 3.35–3.48 3.8–3.9*P. robustus (SK 63) 3.15–4.23 2.9–3.2*Paranthropus boisei (KNM-ER 1820) 2.5–3.1 2.7–3.3*

Homo :Homo erectus (Sangiran S7-37) 4.4*Homo ergaster (KNM-WT 15000) 8.3–8.8 4.5*Homo neanderthalensis (La Chaise) 6.7*Homo sapiens (Global) 5.8 (4.7–7.1)

Note. An asterisk indicates estimated values.Sources. Data for australopiths compiled from Dean et al. (1993); Bromage and Dean (1985); Dean (1987);Beynon and Dean (1988); Conroy and Vannier (1991a, 1991b); Lacruz, Ramirez Rozzi, and Bromage (2005);and Kelley and Schwartz (2012). Data for species of Homo are combined from Dean et al. (2001); Liversidge(2003); and Macchierelli et al. (2006).

the eruptive process of M1 and the difficulty in arranging forthese specimens to be subjected to the latest in noninvasiveimaging technology. Rather, utilizing the incremental growthrecord contained within dental hard tissues, estimates of theage at death for juveniles with mixed dentitions (both decid-uous and permanent teeth present) near or subsequent to M1gingival emergence do exist for a handful of australopiths andearly and later species of the genus Homo. These age-at-deathestimates can thus be used as the basis for estimating ages atM1 emergence (table 2; Kelley and Schwartz 2012). All of theaustralopith emergence age estimates resemble extant Africanapes more than they do modern humans, and this also holdsfor the earliest members of the genus Homo (H. erectus, H.ergaster) for which there are reliable data.

An important way to evaluate these calculated ages at M1emergence is in the context of its relationship with cranialcapacity. As mentioned earlier, cranial capacity and age at M1emergence are strongly correlated in extant anthropoids(Godfrey et al. 2001; Smith 1989), and both exhibit a highcorrelation with aspects of life history. As is evident in figure1, all of the great apes fall on or above the regression line,with Gorilla having the relatively earliest age at M1 emergence,as might be expected given its more folivorous diet (see God-frey et al. 2001). Humans fall well below the regression line,but it is most reasonable to attribute this to the tremendousincrease in cranial capacity in human populations during thelater Pleistocene, which has seemingly forced a partial dis-sociation between cranial capacity and age at M1 emergence.Without this dissociation, M1 emergence (with a predictedmean age of 7.1 years using the regression model versus theactual modern human interpopulation mean of ∼5.8 years)as well as the subsequent emergence of the more posterior

molars would be delayed perhaps beyond ages that are fullycompatible with weaning and the food-processing require-ments of adolescent growth. All of the australopith and earlyHomo specimens (and thus, perhaps, species) also fall belowthe regression line, indicating that these species were alsocharacterized by a relatively advanced age at M1 emergencefor their brain size.

The total range of variation in modern human M1 emer-gence age is quite large, spanning almost 2.5 years (reviewedin Liversidge 2003). Despite that, reconstructed emergenceages for some of the early Homo species do not extend intothe range for modern humans. Importantly, we do not yetknow exactly how, if at all, certain life history attributes covarywith dental development within and among modern humanpeoples or how to integrate these sorts of important intra-specific data with the interspecific trends discussed here.

The earliest species of Homo: the “transitional” hominins H.habilis and H. rudolfensis. Based on detailed reconstructionsof dental chronologies, it seems that australopiths fall wellwithin the range of emergence ages known for captive andwild chimps, and none falls within the ranges known formodern humans. Thus, it is reasonable to conclude that earlyhominins possessed a life history profile more similar to mod-ern African apes than to modern humans. Dental maturationdata for the earliest members of Homo, however, are muchmore limited. To date, dental maturational data for H. habilisand H. rudolfensis consist of either reconstructed root for-mation times (OH 16, H. habilis) or crown formation times(KNM-ER 1805E, H. habilis; KNM-ER 1590, KNM-ER 1802,KNM-ER 1482B, H. rudolfensis; but see Anton 2012 as to


Figure 1. Bivariate plot of ln M1 emergence age in months (y)versus ln cranial capacity in cubic centimeters (x) for a sampleof anthropoids (taxa include Callithrix jacchus, Saguinas fusci-collis, Saguinas nigricollis, Cebus albifrons, Cebus apella, Saimirisciureus, Aotus trivirgatus, Trachypithecus cristata, Chlorocebus ae-thiops, Macaca fascicularis, Macaca mulatta, Macaca nemestrina,Macaca fuscata, Papio cynocephalus, Papio anubis, Pan troglodytes,Pongo pygmaeus, Gorilla gorilla, and Homo sapiens and are derivedprimarily from Smith [1989] with supplemental data from recentanalyses, especially on great ape molar emergence; see references).Summary statistics for the ordinary least squares regression areas follows: , 95% confidence interval (CI;y p 0.630x � 0.072slope): 0.555–0.704, ( ), and the 95% predic-2R p 0.949 P ! .001tion intervals are indicated by the shaded region. Reduced majoraxis regression: , 95% CI (slope): 0.572–0.721.y p 0.646x � 0.144Ceboids are indicated by x, cercopithecoids are filled circles, greatapes are filled squares, and H. sapiens is the open diamond.Australopith species are represented by triangles (from top left,clockwise: SK 48, DIK-1-1, Taung 1, OH 5) and Homo erectus(KNM-WT 15000 and Sangiran S7-37) by open squares.

species groups). Taken together, they suggest a dental devel-opmental profile that was more modern apelike than modernhumanlike, providing some tantalizing evidence that neitherspecies likely possessed an extended period of childhood de-pendence (Dean 1995; Dean et al. 2001).

Later species of Homo. Given the extraordinarily complete andwell-preserved juvenile specimen KNM-WT 15000, we per-haps know more about overall growth and development, andthus life history, in African H. erectus than any other earlyhominin taxon. Based on the lack of fusion of the distal elbowjoint and the acetabulum in particular, KNM-WT 15000 wasgiven an age of death at ca. 13 years, or the early part of theadolescent stage of growth (i.e., postpubertal; Ruff and Walker1993). Interestingly, the mostly completed dentition (26 per-manent teeth had emerged, all but the M3s and maxillarycanines) suggested an age at death of ∼10.2 years. This discordof almost 3 years suggests a somewhat unique developmentaltrajectory for H. erectus (Smith 2004). Equally interesting isthat based on a chimpanzee developmental standard, the state

of somatic development suggests an age at death of 7–7.5years, while the state of dental development suggests an ageestimate of 7 years (Smith 1993). Each scenario carries vastlydifferent implications for the life history profile of H. erectus,and dental developmental data hold the potential to resolvethe discrepancy.

Data on the pace of development derived from the histologyof enamel and dentine originally yielded an age-at-death es-timate of closer to 8 than to 12 years of age (Dean et al. 2001).A more extensive analysis, informed by several more years ofdata on crown and root development in larger samples ofhumans and fossil hominins and thus based on clearer esti-mates of certain dental growth parameters, has confirmed thisby suggesting that an age-at-death interval of 7.6–8.8 years ismost appropriate (Dean and Smith 2009). Furthermore, thereconstructed age at M1 emergence based on inferences fromincremental growth data is 4.5 years, just slightly outside theknown ranges for extant captive (2.1–4.0 years) and wild (3.8–3.9 years) Pan and slightly earlier than that for modern hu-mans (4.7–7.0 years). Given the relationship between brainsize and dental development (see fig. 1), a brain size estimatefor KNM-WT 15000 (810 cm3) generates a point predictionfor M1 emergence age of 5.2 years (95% prediction interval:3.2–8.7 years), suggesting that H. erectus likely possessed rapidmaturation and was more modern apelike in overall growthand development and “certainly closer to the expectation foran ape of comparable dental and skeletal maturity (ca. 7.5)than for a human (ca. 10–15)” (Dean and Smith 2009:114).Taken together, these data make it unlikely that all or evensome of the distinctive features of modern human life historywere present ca. 2.0–1.5 Ma.

Slower maturation translates into later ages for achievingcertain developmental milestones, such as the onset of pu-berty, adolescence, and so forth, and is intricately linked tothe ability of mothers to wean offspring earlier and shorteninterbirth intervals, thereby increasing fertility by having mul-tiple, overlapping offspring. This “stacking” phenomenon isonly possible because of the lower energetic requirements forfueling growth in slower maturing organisms compared withthe tremendous energetic burden mothers would face havingto subsidize the growth of fast-growing, multiple offspring(Dean and Smith 2009; Gurven and Walker 2006). Availabledata from modern hunter-gatherers suggest that humans fol-low the ecological risk aversion model (Janson and van Schaik1993) that posits slow growth and the maintenance of smallsizes for longer periods of time, reduces feeding competition,and translates into significant energetic savings. This energeticsavings is offset by a period of accelerated growth, known asthe adolescent growth spurt, which could be subsidized byolder individuals, highlighting the importance of older in-dividuals in contributing to the care and feeding of children(i.e., paternal care, “grandmothering,” etc.; e.g., Hawkes 2003;Hawkes et al. 1998; Kaplan et al. 2000). Thus, the humanstrategy can be seen as one where higher fertility is achievedby emphasizing more slow-growing children with a later


growth spurt than few faster-growing ones (e.g., Bogin 1988,1997; Gurven and Walker 2006; Leigh 2001; Leigh and Park1998). The combined lack of evidence for protracted growthand for an adolescent growth spurt (Anton and Leigh 2003;Smith 1993; Smith and Tompkins 1995) in H. erectus and H.ergaster supports the assertion that fully modern human lifehistories had yet to evolve by 1.5 Ma.

Recently, several investigations have been launched to as-certain whether the modern human pattern of growth, de-velopment, and life history characterized Neanderthal (Homosapiens neanderthalensis) populations (e.g., Bayle et al. 2009a,2009b, 2010; Coqueugniot and Hublin 2007; Guatelli-Stein-berg 2009; Guatelli-Steinberg et al. 2005; Macchiarelli et al.2006; Ponce de Leon et al. 2008; Ramirez-Rozzi and Bermudezde Castro 2004; Smith et al. 2007b, 2007c, 2010b). A con-vincing argument is mounting that dentally, Neanderthalsmay have experienced accelerated growth, which would inturn suggest that a growth profile that included prolongeddental development, and one that may have included mostor all of the other human life history attributes, did not evolveuntil the appearance of H. sapiens. Unfortunately, very littleis known about dental development in taxa postdating H.erectus and predating Neanderthals, but it would seem par-simonious to reconstruct dental development as being at leastas accelerated in middle Pleistocene taxa such as Homo an-tecessor and Homo heidelbergensis. Limited data are not in-consistent with this hypothesis: certain growth parameters ofanterior teeth in these species seem more similar to Nean-derthals than to modern humans (Ramirez-Rozzi and Ber-mudez de Castro 2004).

Reconstructing life history in early Homo. Given the availabledata, the full suite of modern human life history character-istics was most certainly not present at the base of the homininlineage, nor was it present at the emergence of the genusHomo, but it likely occurred at some time during the middleto late Pleistocene. That does not mean, however, that allhominins before the appearance of H. sapiens possessed a lifehistory that was completely modern apelike despite the factthat all of the included species of early Homo seem quiteaccelerated dentally. There are several possible interpretationsof the relatively early M1 emergence ages of the hominins(reviewed in detail in Kelley and Schwartz 2012). Aside fromscenarios regarding the manner in which these age estimateswere generated or how incremental growth is charted, thereare several ways to interpret these data.

Perhaps the ages at M1 emergence indicate the presence ofrelatively rapid life histories in australopiths and early Homoand thus are more similar to Gorilla than to Pan. This couldbe related to dietary differences: lower-quality food such asthat preferred by primary folivores such as Gorilla gorilla ber-ingei display more accelerated life history schedules and rel-atively precocious dental development than similar-size fru-givores (Breuer et al. 2009; McFarlin et al. 2009).

While diet type and nutritional quality are still debated for

many hominin taxa, it is clear that members of early Homowere committed terrestrial bipeds. Across primates, highlyterrestrial species possess more accelerated life history sched-ules than nonterrestrial species (Deaner, Barton, and vanSchaik 2003; Ross 1992, 1998), likely as a result of increasedextrinsic mortality in the form of predation. Early homininsclearly succumbed to predation with some regularity, and thesignal of relatively rapid life history profiles in australopithsand early Homo may be a direct outcome of selection oper-ating on low survivorship by accelerating overall developmentto reach sexual maturity at an earlier age. In that context,dental development may be linked, perhaps through a mech-anism such as pleiotropy, to overall somatic development andis therefore similarly accelerated.

Scenarios to explain early Homo life history need not relyon inferences based on diet or inferred mortality profilesalone. The combination of low nutritive value of ingestedfood with high rates of extrinsic mortality would have theresult of selecting for individuals within populations thatwould grow at a slow rate and mature early, producing adultsof small stature such as those found within contemporaryhunter-gatherers of the rainforests (Kuzawa and Bragg 2012).High mortality on its own would also select for faster growthand early maturation, though at “normal” adult sizes. In thiscontext, it is interesting to speculate that populations expe-riencing increases in nutritional quality along with high ex-trinsic mortality should grow faster to reach maturation ear-lier at larger adult sizes compared with ancestral populationswith low nutritional quality and high mortality. As shown byMcHenry (1992, 1994), Holliday (2012), and Pontzer (2012),there is good evidence for a general trend of an increase inbody mass from Australopithecus to Homo (also see Ruff 2002and references therein). If rates of extrinsic mortality wereheld constant, this could imply a transition from homininpopulations with low nutritional quality to those with highernutritional quality. This is not inconsistent with recent dietaryreconstructions of early Homo, wherein H. erectus diets werereconstructed as far more varied than in preceding H. habilis(Ungar 2012; Ungar et al. 2011), perhaps suggesting that abroadening of the resource base was an important contrib-uting factor to the evolution of larger body sizes and perhapsultimately to shifts in life history. Interestingly, researchershave speculated on whether similar conditions may have ledto selection for accelerated growth in Neanderthals and in-clude scenarios where they experienced serious nutritionalstress linked with elevated rates of young adult mortality (Og-livie, Curran, and Trinkaus 1989; Pettitt 2000; Trinkaus 1995;Trinkaus and Tompkins 1990). According to life history the-ory, both of these factors in combination would have theeffect of selecting for rapid and early maturation.

A second possible interpretation is that reconstructions ofthe pace of life history are indeed more accurately reflectedby brain size, and so the scheduling of at least some life historyattributes occurred at a slower pace than would be inferredfrom simply evaluating M1 emergence ages alone. In other


words, different life history parameters could be dissociatedfrom one another so that selection could act on them indi-vidually or as smaller developmental subsets. It is critical tobear in mind that dissociations among developmental sys-tems—the scheduling of dental events versus that of repro-ductive events, for instance—may not be as tightly linkedacross closely related species as they appear to be across hom-inoid genera. In fact, the clear associations between dentaldevelopment and life history variables as well as among lifehistory variables that exist when examined across primates asa whole are known to break down when examined acrossclosely related species. The general trend for primates holdsthat larger species take longer to grow and reach sexual ma-turity; however, data on hylobatids suggest that it is thesmaller-bodied Hylobates, not the larger Syndactylus that pos-sesses a later age at sexual maturity and a longer life span(Dirks and Bowman 2007). That same study also demon-strated that the age at molar emergence is not correlated withage at menarche or the age at first reproduction in exactlythe same way in both cercopithecoids and hylobatids. On aneven broader scale, certain strepsirrhines “buck” the primatetrend as a means of solving the problem of how to cope withhighly seasonal and unpredictable environments. Comparedwith lemurids, large-bodied indriids exhibit extreme dentalprecocity while maintaining slower rates of somatic growth,thus allowing for relatively earlier weaning as a strategy tohelp reduce the metabolic burden on mothers (e.g., Godfreyet al. 2001). These are just a few examples of how an un-derstanding of developmental dissociations, or “modularity”(sensu Leigh and Blomquist 2007), urges some caution indirectly linking dental development and the scheduling of lifehistory. An exciting avenue of future study would be to doc-ument the extent to which dental developmental profiles arecorrelated with life history attributes within populations ofmodern humans, an endeavor made easier these days by theever-expanding data on the chronology of developing teethin worldwide populations (see Liversidge 2003). This mayultimately yield clearer and more refined insights into thefiner details of life history evolution across hominin speciesand especially within more closely related and even conspecifichominin populations. At the same time, these analyses mayunveil the extent to which the human life history package isdissociable and allow us to begin to develop models of howshifting patterns of ecology, subsistence, demography, diet,and so forth, throughout the Homo lineage may have resultedin a more piecemeal acquisition of the fully modern humanlife history profile. Some new data suggest that this may bean interesting way forward. Very recently, DeSilva (2011) pos-ited that infant : mother mass ratios of ∼5% (generally ∼6%in modern humans; cf. 3% in extant chimpanzees) were al-ready present in early australopiths. That author suggests thatmore modern humanlike birthing strategies, the adoption ofalloparenting behavior, and so forth, may have been present13 Ma, well before the origin of Homo. On the other hand,other aspects of the human life history package such as life

span, for example, may be a more recent acquisition. Newdata on adult mortality patterns suggest similar populationdemographics for late archaic (Neanderthals) and early mod-ern (Middle and earlier Upper Paleolithic) humans, an ob-servation that weakens support for some sort of demographicadvantage related to enhanced longevity for early modernhumans (Trinkaus 2011).

Future Directions

Discoveries of new infant and subadult fossils along withadvances in noninvasive imaging and analytical methods areproviding opportunities to probe further the fossil record ofhuman growth and development. For instance, it is likely thatthe age at which important life history events, such as wean-ing, will be assessable directly from the fossil record. Thetiming of weaning is a key life event for both mother andoffspring, and analyses of life history variation as it relates toweaning across Primates provide one example of the selectivebasis of these sorts of developmental dissociations. WithinMalagasy prosimians, selection has acted to accelerate weaningand dental development but has delayed the age at first re-production (Godfrey et al. 2001; Richard et al. 2002; Schwartzet al. 2002). It has been suggested that some australopithsshow rapid deciduous tooth wear, which was taken as evidencesuggestive of relatively early weaning (Aiello, Montgomery,and Dean 1991; Dean 2006, 2010). It may now be possibleto retrieve direct evidence for reconstructing weaning age andshifts in energy provisioning for offspring through the evo-lution of Homo. Across Primates, weaning is closely tied intime to the emergence of M1. However, human life historyis characterized by relatively early weaning followed by a pro-longed period of postweaning dependency. The advancementof weaning age throughout human evolution coupled withrapid and early brain growth implies a shift in how the risingenergetic demands of offspring are met: initially, energeticcosts are subsidized completely by the mother but then bymembers of the social group through the provisioning ofweanlings (Humphrey 2010). This pattern of high maternalinvestment and alloparenting behavior is important becauseit is a clear determinant of birth spacing. Such a stratagemhas also been suggested to characterize the earliest membersof Homo (Aiello and Wells 2002). Recently, models have beenproposed to establish the precise age at which organisms wereweaned by accessing the isotopic record, in particular, stron-tium : calcium ratios (Sr/Ca) preserved within dental hardtissues (e.g., Humphrey et al. 2008a, 2008b). Charting shiftsin this ratio throughout the developmental period associatedwith early tooth tissue formation is one exciting way of re-constructing infant diet as well as tracking dietary transitionsthroughout early life. If early Homo and later-occurring ar-chaic Homo populations were indeed characterized by rela-tively early weaning, then analyses of the isotopic chemistrythroughout enamel development hold the potential to verify


Figure 2. Bivariate plot of bite point positions for all teeth and masticatory muscle positions relative to the temporomandibularjoint (TMJ; y-axis) versus age (x-axis) in a cross-sectional ontogenetic series of modern humans (Nubian archaeological samplehoused at University of Colorado, Boulder). Bite points are measured as the distance of each tooth from the TMJ, measured in theocclusal plane, and are illustrated by the two vertical arrows (left) indicating bite points for the dm1 and dm2. Masticatory muscleposition (for the superficial and deep masseters, temporalis, and medial pterygoid muscles) is defined as the point where eachmuscle’s resultant force crosses the occlusal plane relative to the position of the TMJ. Primary masticatory adductor position andorientation are based on a series of 2-D and 3-D linear and angular measurements and, taken together, capture what has beentermed “masticatory system configuration” (see Spencer 1995, 1999). Note the consistent position of the dm2 (small white sphere)and the permanent M1, M2, and M3 (large white spheres) anterior to the masticatory muscles (i.e., above the white dotted line) atthe time of emergence. Also note the differing rates of anterior growth of the dental arcade and masticatory muscles (as indicatedby the slopes of second-order polynomials); space for emerging molars is a product of these different growth rates. A color versionof this figure appears in the online edition of Current Anthropology.

this with direct evidence for the age at weaning from the fossilrecord itself.

While continued probing of the fossil record to establishmore precise demographic and maturational profiles holdsthe potential to yield key details in the evolution of humanlife history, another interesting way forward is to attempt tounderstand the processes that lie behind the slightly dissimilarpatterns in dental development, and thus inferred life historyprofiles, among hominins. A host of studies have advancedour understanding of how dental developmental variationintersects with primate life history variation, but surprisinglylittle is known about the precise mechanism that governs,modulates, regulates, constrains, and so forth, the timing ofmolar eruption and as such the underlying processes thatregulate these temporal events are largely unknown.

Recently, it was postulated that a set of biomechanical con-straints regulates masticatory system configuration through-out ontogeny and therefore modulates the position and ul-timately the timing of emerging molars within developingfaces (Spencer and Schwartz 2008). Based on an ontogeneticsample of modern human crania, it seems that successive

molar emergence events are predominantly a function of ratesof facial growth such that successive molars (deciduous andpermanent) emerge at a consistent position relative to themasticatory musculature (fig. 2). Moreover, there appears tobe a consistent position of newly erupted molars (deciduousand permanent) relative to the temporomandibular joint(TMJ) that in an archaeological sample of modern humansis ∼40 mm (fig. 3). This ontogenetic arrangement ensuresthat there is a biomechanically optimal location for molareruption anterior to the net vector of masticatory muscleeffort (note position of white dots relative to dotted whiteline, fig. 2) and that each successive molar erupts into thisoptimal position only at a point during ontogeny when it isvacated as a result of facial growth. This all suggests covari-ation in rates of facial growth (indicated by the slope of thefirst part of the curve for each molar, fig. 2), the position ofthe masticatory musculature, the spatial position of an erupt-ing molar, and the timing of molar emergence.

The validity of this biomechanical model for modulatingthe timing of molar emergence has not been fully established.Indeed, whether this constraint operates across hominoids is


Figure 3. Consistent position of newly erupted molars relativeto the temporomandibular joint (TMJ) as indicated by the lengthof the horizontal white line. The dot on the left of the line marksthe position of the TMJ in lateral view, while the dot on the rightmarks the position of the newly emerged molar. Note that theline is the same length for individual human crania at the timeof M1 emergence (top), M2 emergence (middle), and M3 emer-gence (bottom). This is illustrated graphically by the consistentwidth of the shaded rectangle. While the absolute distance isdifferent, the same pattern of spatial consistency in molar po-sition holds for an ontogenetic sample of wild-living Pan (seetext). A color version of this figure appears in the online edition.

not yet clear, but preliminary data on an ontogenetic seriesof western African chimpanzees (Pan troglodytes verus; n p

) are striking: like the modern human sample, no significant37differences are present between the position of each succes-sively emerging molar and the TMJ (Kruskal-Wallis, ,df p 2

) despite the fact that, unsurprisingly, the absoluteP p .6215distance is slightly greater in this sample (∼48 mm).

A fuller mapping of the influence of these biomechanicalconstraints onto variation in the ontogeny of masticatorymuscle position and explicitly testing hypotheses that inte-grate craniofacial architecture, muscle function, facial growth,and molar emergence across hominoids is currently under-way. These are critical comparative data because selection foraccelerated molar eruption, as seems to characterize earlyHomo relative to modern humans, should require a similaracceleration in facial growth. The delay in molar emergence

ages in modern humans may therefore result from reducedrates of facial growth and extreme orthognathy, perhaps incombination with a developmental delay in facial growth. Alater initiation of facial growth would result in a delay inclearance of a “biomechanically appropriate” space availablefor molar emergence.

In the absence of good ontogenetic data on craniofacialgrowth in early Homo, it is useful to evaluate how well thisbiomechanical model integrates craniofacial morphology withdata on developmental rates within other members of theHomo lineage. This model predicts the advanced molar emer-gence schedules of Neanderthals to be related to a combi-nation of their higher degree of midfacial prognathism andaccelerated cranial growth trajectories. Some evidence in sup-port of faster rates of craniofacial growth (Ponce de Leon andZollikofer 2001; Ponce de Leon et al. 2008) and dental de-velopment (Smith et al. 2007c, 2010b) exist. Thus, selectionmay have accelerated age at weaning and thus ecological in-dependence by advancing rates of cranial growth in a mannerthat permitted a more accelerated molar development/erup-tion schedule, which would represent an effective life historystrategy under conditions of high extrinsic mortality. Morethorough explorations of how the timing of molar develop-ment and emergence may result from the complex spatialinterplay between growing faces and expanding neurocraniahold tremendous potential for illuminating the underlyingmechanism that regulates molar emergence and, ultimately,for unlocking the linkages between dental development andlife history events.

Finally, it is generally agreed that the earliest species ofHomo evolved from Australopithecus, either in East or SouthAfrica. The cranium from Bouri, Ethiopia, at 2.5 Ma attrib-uted to Australopithecus garhi (Asfaw et al. 1999) and theassociated cranial and postcranial material for the newly an-nounced Australopithecus sediba from Malapa, South Africa,at 1.9 Ma (Berger et al. 2010) may therefore provide importantclues for helping to better understand the complex interplayamong morphological, ecological, reproductive, and behav-ioral adaptations that underlies the transition to and ultimatesuccess of the genus Homo.

Acknowledgments

I would like to thank Leslie Aiello and Susan Anton for theirinvitation to participate in the Wenner-Gren workshop, all ofthe conference participants for their stimulating and thought-ful discussions throughout the conference, and the two anon-ymous reviewers for their input and constructive commentson this manuscript. The ideas and work laid out here are theresults of many discussions and ongoing collaborations, es-pecially with Chris Dean, Jay Kelley, Tanya Smith, DebbieGuatelli-Steinberg, Laurie Godfrey, Wendy Dirks, Bill Kimbel,Mark Spencer, Terry Ritzman, Kierstin Catlett, and Halszka


Glowacka. I am grateful to the Institute of Human Originsat Arizona State University for their generous support.

References CitedAiello, Leslie C., Carey Montgomery, and M. Christopher Dean. 1991. The

natural history of deciduous tooth attrition in hominoids. Journal of HumanEvolution 21:397–412.

Aiello, Leslie C., and Jonathan C. K. Wells. 2002. Energetics and the evolutionof the genus Homo. Annual Review of Anthropology 31:323–338.

Alemseged, Zeresenay, Fred Spoor, William H. Kimbel, Rene Bobe, DenisGeraads, Denne Reed, and Jonathan G. Wynn. 2006. A juvenile early hom-inin skeleton from Dikika, Ethiopia. Nature 443:296–301.

Anemone, Robert L., Mark P. Mooney, and Michael I. Siegel. 1996. Longi-tudinal study of dental development in chimpanzees of known chronolog-ical age: implications for understanding the age at death of Plio-Pleistocenehominids. American Journal of Physical Anthropology 99:119–133.


Anton, Susan C., and Steven R. Leigh. 2003. Growth and life history in Homoerectus. In Patterns of growth and development in the genus Homo. JenniferL. Thompson, Gail E. Krovitz, and Andrew J. Nelson, eds. Pp. 219–245.Cambridge: Cambridge University Press.

Asfaw, Berhane, Tim White, C. Owen Lovejoy, Bruce Latimer, Scott Simpson,and Gen Suwa. 1999. Australopithecus garhi: a new species of early homininfrom Ethiopia. Science 284:629–635.

Bayle, Priscilla, Jose Braga, Arnaud Mazurier, and Roberto Macchiarelli. 2009a.Brief communication: high-resolution assessment of the dental develop-mental pattern and characterization of tooth tissue proportions in the lateUpper Paleolithic child from La Madeleine, France. American Journal ofPhysical Anthropology 138:493–498.

———. 2009b. Dental developmental pattern of the Neanderthal child fromRoc de Marsal: a high resolution 3D analysis. Journal of Human Evolution56:66–75.

Bayle, Priscilla, Roberto Macchiarelli, Erik Trinkaus, Cidalia Duarte, ArnaudMazurier, and Joao Zilhao. 2010. Dental maturational sequence and dentaltissue proportions in the early Upper Paleolithic child from Abrigo do LagarVelho, Portugal. Proceedings of the National Academy of Sciences of the USA107:1338–1342.


Beynon, A. David, C. Barry Clayton, Fernando V. Ramirez Rozzi, and DonaldJ. Reid. 1998. Radiographic and histological methodologies in estimatingthe chronology of crown development in some modern humans and greatapes: a review with some applications for studies on juvenile hominids.Journal of Human Evolution 35:351–370.

Beynon, A. David, and M. Christopher Dean. 1988. Distinct dental devel-opment patterns in early fossil hominids. Nature 335:509–514.

Beynon, A. David, M. Christopher Dean, and Donald J. Reid. 1991. Histo-logical study on the chronology of the developing dentition in gorilla andorangutan. American Journal of Physical Anthropology 86:189–203.

Bogin, Barry. 1988. Patterns of human growth. Cambridge: Cambridge Uni-versity Press.

———. 1997. Evolutionary hypotheses for human childhood. Yearbook ofPhysical Anthropology 104:63–90.

Bogin, Barry, and B. Holly Smith. 2000. Evolution of the human life cycle.In Human biology: an evolutionary and biocultural perspective. Sarah Stinson,Barry Bogin, Rebecca Huss-Ashmore, and Dennis O’Rourke, eds. Pp. 377–424. New York: Wiley-Liss.

Boyde, Alan. 1963. Estimation of age at death of young human skeletal remainsfrom incremental lines in dental enamel. In Third international meeting inforensic immunology, medicine, pathology and toxicology (London, April 16–24th, 1963): plenary session 2A. Pp. 36–46. Amsterdam: Exerpta Medica.

———. 1964. The structure and development of mammalian enamel. PhDthesis, University of London.

Bradley, Brenda J. 2008. Reconstructing phylogenies and phenotypes: a mo-lecular view of human evolution. Journal of Anatomy 212:337–353.

Breuer, Thomas, Mireille B. Hockemba, Claudia Olejniczak, Richard J. Parnell,and Emma J. Stokes. 2009. Physical maturation, life-history classes and age

estimates of free-ranging western gorillas: insights from Mbeli Bai, Republicof Congo. American Journal of Primatology 71:106–119.

Bromage, Timothy G., and M. Christopher Dean. 1985. Re-evaluation of theage at death of immature fossil hominids. Nature 317:525–527.

Conroy, Glenn C., and Kevin L. Kuykendall. 1995. Paleopediatrics: or whendid human infants really become human? American Journal of PhysicalAnthropology 98:121–131.

Conroy, Glenn C., and C. James Mahoney. 1991. Mixed longitudinal studyof dental emergence in the chimpanzee, Pan troglodytes (Primates, Pongi-dae). American Journal of Physical Anthropology 86:243–254

Conroy, Glenn C., and Michael W. Vannier. 1991a. Dental development inSouth African australopithecines. 1. Problems of pattern and chronology.American Journal of Physical Anthropology 86:121–136.

———. 1991b. Dental development in South African australopithecines. 2.Dental stage assessment. American Journal of Physical Anthropology 86:137–156.

Coqueugniot, Helene, and Jean-Jacques Hublin. 2007. Endocranial volumeand brain growth in immature Neanderthals. Periodicum Biologorum 109:379–385.

Dean, M. Christopher. 1987. The dental developmental status of six EastAfrican juvenile fossil hominids. Journal of Human Evolution 16:197–213.

———. 1995. The nature and periodicity of incremental lines in the primatedentition and their relationship to periradicular bands in OH 16 (Homohabilis). In Aspects of dental biology: paleontology, anthropology, and evolution.Jacopo Moggi-Cecchi, ed. Pp. 239–265. Florence: International Institute forthe Study of Man.

———. 2006. Tooth microstructure tracks the pace of human life history.Proceedings of the Royal Society B: Biological Sciences 273:2799–2802.

———. 2010. Retrieving chronological age from dental remains of early fossilhominins to reconstruct human growth in the past. Philosophical Trans-actions of the Royal Society B: Biological Sciences 365:3397–3410.

Dean, M. Christopher, A. David Beynon, J. Francis Thackeray, and GabrieleA. Macho. 1993. Histological reconstruction of dental developmental andage at death of a juvenile Paranthropus robustus specimen, SK 63, fromSwartkrans, South Africa. American Journal of Physical Anthropology 91:401–419.

Dean, M. Christopher, Meave G. Leakey, Donald J. Reid, Friedemann Schrenk,Gary T. Schwartz, Christopher Stringer, and Alan C. Walker. 2001. Growthprocesses in teeth distinguish modern humans from Homo erectus andearlier hominins. 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: originsand evolution of the genus Homo. Frederick E. Grine, John C. Fleagle, andRichard E. Leakey, eds. Pp. 101–120. New York: Springer.

Deaner, Robert O., Robert A. Barton, and Carel P. van Schaik. 2003. Primatebrains and life history: renewing the connection. In Primate life historiesand socioecology. Peter M. Kappeler and Michael E. Pereira, eds. Pp. 233–265. Chicago: University of Chicago Press.

deMenocal, Peter B. 1995. Plio-Pleistocene African climate. Science 270:53–59.

DeSilva, Jeremy M. 2011. A shift toward birthing relatively large infants earlyin human evolution. Proceedings of the National Academy of Sciences of theUSA 108:1022–1027.

Dirks, Wendy. 1998. Histological reconstruction of dental development andage of death in a juvenile gibbon (Hylobates lar). Journal of Human Evolution35:411–425.

———. 2003. Effect of diet on dental development in four catarrhine pri-mates. American Journal of Primatology 61:29–40.

Dirks, Wendy, and Jacqui E. Bowman. 2007. Life history theory and dentaldevelopment in four species of catarrhine primates. Journal of Human Evo-lution 53:309–320.

Dirks, Wendy, Louise T. Humphrey, M. Christopher Dean, and Teresa E.Jeffries. 2010. The relationship of accentuated lines in enamel to weaningstress in juvenile baboons (Papio hamadryas anubis). Folia Primatologica81:207–223.

Godfrey, Laurie R., Andrew J. Petto, and Michael R. Sutherland. 2002. Dentalontogeny and life-history strategies: the case of the giant extinct indrioidsof Madagascar. In Reconstructing behavior in the primate fossil record. J.Michael Plavcan, Richard F. Kay, William L. Jungers, and Carel P. van Schaik,eds. Pp. 113–157. New York: Kluwer.

Godfrey, Laurie R., Karen E. Samonds, William L. Jungers, and Michael R.Sutherland. 2001. Teeth, brains, and primate life histories. American Journalof Physical Anthropology 114:192–214.


Godfrey, Laurie R., Karen E. Samonds, Patricia C. Wright, and Stephen J.King. 2005. Schultz’s unruly rule: dental developmental sequences andschedules in small-bodied, folivorous lemurs. Folia Primatologica 6:77–99.

Guatelli-Steinberg, Debbie. 2009. Recent studies of dental development inNeanderthals: implications for Neanderthal 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 Neanderthals werecomparable to those of modern humans. Proceedings of the National Acad-emy of Sciences of the USA 102:14197–14202.

Gurven, Michael, and Robert Walker. 2006. Energetic demands of multipledependents and the evolution of slow human growth. Proceedings of theRoyal Society B: Biological Sciences 273:835–841.

Hamada, Yuzuru, Toshifumi Udono, Migaku Teramoto, and Tsutomu Suga-wara. 1996. The growth pattern of chimpanzees: somatic growth and re-productive maturation in Pan troglodytes. Primates 37:279–295.

Hawkes, Kristen. 2003. Grandmothers and the evolution of human longevity.American Journal of Human Biology 15:380–400.



Humphrey, Louise T. 2010. Weaning behavior in human evolution. Seminarsin Cellular and Developmental Biology 21:453–461.

Humphrey, Louise T., M. Christopher Dean, Teresa E. Jeffries, and MalcomPenn. 2008a. Unlocking evidence of early diet from tooth enamel. Pro-ceedings of the National Academy of Sciences of the USA 105:6834–6839.

Humphrey, Louise T., Wendy Dirks, M. Christopher Dean, and Teresa E.Jeffries. 2008b. Tracking dietary transitions in weanling baboons using stron-tium/calcium ratios in enamel. Folia Primatologica 79:197–212.

Janson, Charles H., and Carel P. van Schaik. 1993. Ecological risk aversion injuvenile primates: slow and steady wins the race. In Juvenile primates: lifehistory, development and behavior. Michael Pereira and Lynn Fairbanks, eds.Pp. 57–76. New York: Oxford University Press.

Kaplan, Hillard, Kim Hill, Jane Lancaster, and Ana Magdalena Hurtado. 2000.A theory of human life history evolution: diet, intelligence, and longevity.Evolutionary Anthropology 9:156–184.

Kelley, Jay, and Gary T. Schwartz. 2010. Dental development and life historyin living African and Asian apes. Proceedings of the National Academy ofSciences of the USA 107:1035–1040.

———. 2012. Life-history inference in the early hominins Australopithecusand Paranthropus. International Journal of Primatology 33:1332–1363.

Kimbel, William H. 2009. The origin of Homo. In The first humans: originand early evolution of the genus Homo. Frederick E. Grine, John G. Fleagle,and Richard E. Leakey, eds. Pp. 31–37. New York: Springer.

Kimbel, William H., Don C. Johanson, and Yoel Rak. 1997. Systematic as-sessment of a maxilla of Homo from Hadar, Ethiopia. American Journal ofPhysical Anthropology 103:235–262.

Kimura, Tasuku, and Yuzuru Hamada. 1996. Growth of wild and laboratoryborn chimpanzees. Primates 37:237–251.

Knott, Cheryl D., Melissa Emery Thompson, and Serge A. Wich. 2009. Theecology of reproduction in wild orangutans. In Orangutans: geographic var-iation in behavioral ecology and conservation. Serge A. Wich, S. Suci Utami-Atmoko, Tatang Mitra Setia, and Carel P. van Schaik, eds. Pp. 171–188.Oxford: Oxford University Press.

Kuykendall, Kevin L., C. James Mahoney, and Glenn C. Conroy. 1992. Probitand survival analysis of tooth emergence ages in a mixed-longitudinal sam-ple of chimpanzees (Pan troglodytes). American Journal of Physical Anthro-pology 89:379–399.


Lacruz, Rodrigo S., Fernando Ramirez Rozzi, and Timothy G. Bromage. 2005.Dental enamel hypoplasia, age at death, and weaning in the Taung child.South African Journal of Science 101:567–569.

Leakey, Louis S. B., Phillip V. Tobias, and John Napier. 1964. A new speciesof the genus Homo from Olduvai Gorge. Nature 202:7–9.

Leigh, Steven R. 2001. Evolution of human growth. Evolutionary Anthropology10:223–236.

Leigh, Steven R., and Gregory E. Blomquist. 2007. Life history. In Primatesin perspective. Christina J. Campbell, Agustın Fuentes, Katherine C.

MacKinnon, Melissa Panger, and Simon K. Bearder, eds. Pp. 396–407. NewYork: Oxford University Press.

Leigh, Steven R., and Paul B. Park. 1998. Evolution of human growth pro-longation. American Journal of Physical Anthropology 107:331–350.

Liversidge, Helen. 2003. Variation in modern human dental development. InPatterns of growth and development in the genus Homo. Jennifer L. Thomp-son, Gail E. Krovitz, and Andrew J. Nelson, eds. Pp. 73–113. Cambridge:Cambridge University Press.

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.

Mann, Alan E. 1975. Some paleodemographic aspects of the South African Aus-tralopithecines. Philadelphia: University of Pennsylvania Press.

McFarlin, Shannon C., Timothy G. Bromage, Alecia A. Lilly, Michael R. Cran-field, Stephen P. Nawrocki, Amandine Eriksen, David Hunt, AlbertNdacyayisenga, Celestin Kanimba Misago, and Anthony Mudakikwa. 2009.Recovery and preservation of a mountain gorilla skeletal resource inRwanda. American Journal of Physical Anthropology 138(suppl.):257–258.

McHenry, Henry M. 1992. Body size and proportions in early hominids.American Journal of Physical Anthropology 87:407–431.




Nettle, Daniel, David A. Coall, and Thomas E. Dickins. 2010. Early-life con-ditions and age at first pregnancy in British women. Proceedings of the RoyalSociety B: Biological Sciences 278:1721–1727.

Nissen, Henry W., and Austin H. Riesen. 1945. The deciduous dentition ofchimpanzee. Growth 9:265–274.

———. 1964. The eruption of the permanent dentition of chimpanzee. Amer-ican Journal of Physical Anthropology 22:285–294.

Ogilvie, Marsha D., Bryan K. Curran, and Erik Trinkaus. 1989. Prevalenceand patterning of dental enamel hypoplasia among the Neanderthals. Amer-ican Journal of Physical Anthropology 79:25–41.

Pettitt, Paul B. 2000. Neanderthal lifecycles: developmental and social phasesin the lives of the last archaics. World Archaeology 31:351–366.

Ponce de Leon, Marcia S., Lubov Golovanova, Vladimir Doronichev, GalinaRomanova, Takeru Akazawa, Osamu Kondo, Hajime Ishida, and ChristophP. E. Zollikofer. 2008. Neanderthal brain size at birth provides insights intothe evolution of human life history. Proceedings of the National Academy ofSciences of the USA 105:13764–13768.

Ponce de Leon, Marcia S., and Christoph P. Zollikofer. 2001. Neanderthalcranial ontogeny and its implications for late hominid diversity. Nature 412:534–538.


Potts, Richard. 1996. Evolution and climate variability. Science 273:922–923.———. 2012. Environmental and behavioral evidence pertaining to the evo-

lution of early Homo. Current Anthropology 53(suppl. 6):S299–S317.Rak, Yoel, Avishag Ginzburg, and Eli Geffen. 2007. Gorilla-like anatomy on

Australopithecus afarensis mandibles suggests Au. afarensis links to robustaustralopiths. Proceedings of the National Academy of Sciences of the USA104:6568–6572.

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

Reid, Donald J., Gary T. Schwartz, Malkanthi S. Chandrasekera, and M. Chris-topher Dean. 1998. A histological reconstruction of dental development inthe common chimpanzee, Pan troglodytes. Journal of Human Evolution 35:427–448.

Richard, Alison F., Robert E. Dewar, Marion Schwartz, and Joelisoa Rat-sirarson. 2002. Life in the slow lane? demography and life histories of maleand female sifaka (Propithecus verreauxi verreauxi). Journal of Zoology 256:421–436.

Robson, Shannen L., and Bernard A. Wood. 2008. Hominin life history: re-construction and evolution. Journal of Anatomy 212:394–425.

Roff, Derek A. 2002. Life history evolution. Sunderland, MA: Sinauer.Ross, Caroline. 1992. Environmental correlates of the intrinsic rate of natural

increase in primates. Oecologia 90:383–390.


———. 1998. Primate life histories. Evolutionary Anthropology 6:54–63.Ruff, Christopher B. 2002. Variation in human body size and shape. Annual

Review of Anthropology 31:211–232.Ruff, Christopher B., and Alan C. Walker. 1993. Body size and body shape.

In The Nariokotome Homo erectus skeleton. Richard E. F. Leakey and AlanC. Walker, eds. Pp. 234–265. Cambridge, MA: Harvard University Press.

Ruvolo, Maryellen. 1994. Molecular evolutionary processes and conflictinggene trees: the hominoid case. American Journal of Physical Anthropology94:89–113.

Sacher, George A. 1959. Relationship of lifespan to brain weight and bodyweight in mammals. In The lifespan of animals. G. E. W. Wolstenholme andMeave O’Connor, eds. Pp. 115–133. CIBA Foundation Colloquia on Aging,vol. 5. London: Churchill.

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

———. 1978. Longevity, aging, and death: an evolutionary perspective. Ger-ontology 18:112–119.

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 108:593–616.

Schultz, Adolph H. 1935. Eruption and decay of the permanent teeth inprimates. American Journal of Physical Anthropology 19:489–581.

———. 1940. Growth and development of the chimpanzee. Contributions toEmbryology 28:1–63.

———. 1949. Ontogenetic specializations of man. Archiv der Julius Klaus-Stiftung fur Vererbungsforschung, Sozialanthropologie und Rassenhygiene 24:197–216.

———. 1960. Age changes in primates and their modification in man. InHuman growth. James M. Tanner, ed. Pp. 1–20. Oxford: Pergamon.

Schwartz, Gary T., and M. Christopher Dean. 2000. Interpreting the hominiddentition: ontogenetic and phylogenetic aspects. In Development, growthand evolution: implications for the study of the hominid skeleton. PaulO’Higgins and Martin J. Cohen, eds. Pp. 207–233. London: Academic Press.

Schwartz, Gary T., Patrick Mahoney, Laurie R. Godfrey, Frank P. Cuozzo,William L. Jungers, and Gisele F. N. Randria. 2005. Dental development inMegaladapis edwardsi (Primates, Lemuriformes): implications for under-standing life history variation in subfossil lemurs. Journal of Human Evo-lution 49:702–721.

Schwartz, Gary T., Karen E. Samonds, Laurie R. Godfrey, William L. Jungers,and Elwyn L. Simons. 2002. Dental microstructure and life history in sub-fossil Malagasy lemurs. Proceedings of the National Academy of Sciences ofthe USA 99:6124–6129.

Smith, B. Holly. 1986. Dental development in Australopithecus and early Homo.Nature 323:327–330.

———. 1989. Dental development as a measure of life history variation inprimates. Evolution 43:683–688.

———. 1993. The physiological age of KNM-WT 15000. In The NariokotomeHomo erectus skeleton. Richard E. F. Leakey and Alan C. Walker, eds. Pp.195–220. Cambridge, MA: Harvard University Press.

———. 2000. “Schultz’s rule” and the evolution of tooth emergence andreplacement patterns in primates and ungulates. In Development, functionand evolution of teeth. Mark F. Teaford, Moya Meredith Smith, and MarkW. J. Ferguson, eds. Pp. 212–227. Cambridge: Cambridge University Press.

Smith, B. Holly, and Christophe Boesch. 2011. Mortality and the magnitudeof the “wild effect” in chimpanzee tooth emergence. Journal of HumanEvolution 60:34–46.

Smith, B. Holly, Tracey L. Crummett, and Kari L. Brandt. 1994. Ages oferuption of primate teeth: a compendium for aging individuals and com-paring life histories. Yearbook of Physical Anthropology 37:177–231.

Smith, B. Holly, and Robert L. Tompkins. 1995. Toward a life history of theHominidae. Annual Review of Anthropology 24:257–279.

Smith, Richard J., Patrick J. Gannon, and B. Holly Smith. 1995. Ontogeny ofaustralopithecines and early Homo: evidence from cranial capacity and den-tal eruption. Journal of Human Evolution 29:155–168.

Smith, Shelley. 2004. Skeletal age, dental age, and the maturation of KNM-WT 15000. American Journal of Physical Anthropology 125:105–120.

Smith, Tanya M. 2006. Experimental determination of the periodicity of in-cremental features in enamel. Journal of Anatomy 208:99–114.

———. 2008. Incremental dental development: methods and applications inhominoid evolutionary studies. Journal of Human Evolution 54:205–224.

Smith, Tanya M., Donald J. Reid, M. Christopher Dean, Anthony J. Olejniczak,and Lawrence B. Martin. 2007a. Molar development in common chim-panzees (Pan troglodytes). Journal of Human Evolution 52:201–216.

Smith, Tanya M., Donald J. Reid, and Joyce E. Sirianni. 2006. The accuracyof histological assessments of dental development and age at death. Journalof Anatomy 208:125–138.

Smith, Tanya M., B. Holly Smith, Donald J. Reid, Heike Siedel, Linda Vigilant,Jean-Jacques Hublin, and Christophe Boesch. 2010a. Dental developmentof the Taı Forest chimpanzees revisited. Journal of Human Evolution 58:363–373.

Smith, Tanya M., Paul Tafforeau, Donald J. Reid, Rainer Grun, Stephen Eggins,Mohamed Boutakiout, and Jean-Jacques Hublin. 2007b. Earliest evidenceof modern human life history in North African early Homo sapiens. Pro-ceedings of the 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. 2010b. Dentalevidence for ontogenetic differences between modern humans and Nean-derthals. Proceedings of the National Academy of Sciences of the USA 107:20923–20928.

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

Spencer, Mark A. 1995. Masticatory system configuration and diet in an-thropoid primates. PhD dissertation, State University of New York, StonyBrook.

———. 1999. Constraints on masticatory system evolution in anthropoidprimates. American Journal of Physical Anthropology 108:483–506.

Spencer, Mark A., and Gary T. Schwartz. 2008. The ontogeny of masticatorysystem configuration in humans and its influence on the timing of molareruption. American Journal of Physical Anthropology 135(suppl.):299.

Stearns, Stephen. 1992. The evolution of life histories. Oxford: Oxford UniversityPress.

Trinkaus, Erik. 1995. Neanderthal mortality patterns. Journal of ArchaeologicalScience 22:121–142.

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

Trinkaus, Erik, and Robert L. Tompkins. 1990. The Neanderthal 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.New York: Wiley-Liss.


Ungar, Peter S., Frederick E. Grine, Mark F. Teaford, and Sireen El Zaatari.2011. Dental microwear and diets of African early Homo. Journal of HumanEvolution 50:78–95.

Walker, Robert S., Michael Gurven, Kim Hill, Andrea Migliano, NapoleanChagnon, Roberta De Souza, Gradimir Djurovic, et al. 2006. Growth ratesand life histories in twenty-two small-scale societies. American Journal ofHuman Biology 18:295–311.

West-Eberhard, Mary Jane. 2003. Developmental plasticity and evolution. NewYork: Oxford University Press.

Willoughby, David P. 1978. All about gorillas. South Brunswick, NJ: Barnes.Winkler, Linda A., Jeffrey H. Schwartz, and Daris R. Swindler. 1991. Aspects

of dental development in the orangutan prior to the eruption of the per-manent dentition. American Journal of Physical Anthropology 86:255–272.

Wood, Bernard A., and Mark Collard. 1999. The human genus. Science 284:65–71.

Wood, Bernard A., and Nicholas Lonergan. 2008. The hominin fossil record:taxa, grades and clades. Journal of Anatomy 212:354–376.

Zihlman, Adrienne, Debra Bolter, and Christophe Boesch. 2004. Wild chim-panzee dentition and its implications for assessing life history in immaturehominin fossils. Proceedings of the National Academy of Sciences of the USA101:10541–10543.



Body Size, Size Variation, and Sexual SizeDimorphism in Early Homo

by J. Michael Plavcan

Size variation provides important clues about the taxonomy, morphology, behavior, and life history of extinct species.Body size variation in living species is commonly attributed to Bergmann’s rule, resource availability, nutrition,local selection pressures, and sexual size dimorphism. While our understanding of the mechanisms producing sizevariation in living species has grown more sophisticated in recent years, our ability to apply this knowledge to thefossil record is limited by the quality of the available fossil and extant comparative samples. New discoveries offossil Homo have expanded the known range of size variation and provide hints of geographic and temporal variationin size within and between named taxa and possible strong sexual size dimorphism. Even so, the range of sizevariation in Homo habilis/rudolfensis and Homo erectus matches or even is less than that seen in geographicallyrestricted samples of living anthropoid primates. These observations dictate caution in interpreting the meaning ofvariation in early Homo but also underscore the critical necessity of improving comparisons of size among fossilsand establishing an adequate comparative database of living species that allows us to discriminate between the effectsof epigenetic and selective factors on the expression of variation.

The meaning of size variation has long been a topic of debateconcerning the origins and evolution of Homo, playing a keyrole in discussions about species recognition, sexual dimor-phism, life history, and behavior (Aiello and Key 2002; Anton2003; Kappelman 1996; McHenry 1992, 1994; Wood 1993).Discoveries of new specimens of early Homo and Homo erectushave expanded the apparent range of size variation, raisedawareness of the biological implications of geographic andtemporal variation in early Homo, and altered our perceptionof potential size dimorphism in early Homo and H. erectus(Aiello and Key 2002; Anton 2012; Anton and Snodgrass 2012;Kuzawa and Bragg 2012; Rightmire, Van Arsdale, and Lord-kipanidze 2008; Ruff 2002; Spoor et al. 2007). To understandthe implications of this variation, it is critical to parse outthe sources and causes of size variation in living animals andlook for ways to identify the causes of variation in the fossilrecord.

Discussions of variation in fossil Homo naturally turn tomodern humans as a model of patterns of intra- and inter-specific size variation and sexual dimorphism (e.g., Anton2003; Kramer et al. 1995; Lieberman, Pilbeam, and Wood1988; Miller 2000; Rightmire, Van Arsdale, and Lordkipanidze2008; Skinner, Gordon, and Collard 2006). Modern Homoshows substantial size variation that is attributed to phenom-

J. Michael Plavcan is Professor in the Department of Anthropology,University of Arkansas (Fayetteville, Arkansas 72701, U.S.A.[[email protected]]). This paper was submitted 12 XII 11,accepted 3 VII 12, and electronically published 15 X 12.

ena such as Allen’s and Bergmann’s rules, secular trends indiet and nutrition, selection for smaller or larger size in as-sociation with environmental or behavioral factors, and sexualdimorphism (Ruff 2002). The role of developmental plasticityin response to environmental and nutritional variation as acause of size variation is also receiving greater attention (An-ton 2012; Kuzawa and Bragg 2012).

At the same time, great apes, and particularly Pongo andGorilla, are among the most sexually dimorphic primates(Smith and Jungers 1997). Australopithecus too shows evi-dence of strong sexual size dimorphism (Gordon, Green, andRichmond 2008; Lockwood et al. 1996), raising the questionof whether size variation in early Homo could reflect strongsexual dimorphism as opposed to either taxonomic variationor intraspecific variation due to other factors. This issue iscritical, because strong size dimorphism has been linked tosexual selection, and changes in size dimorphism imply sig-nificant changes in behavior and life history (Gordon 2006a,2006b; Leigh 1992; Lindenfors 2002; Lovejoy 2009; Martin,Willner, and Dettling 1994; Plavcan 2001; Plavcan and vanSchaik 1997a, 1997b; Reno et al. 2003, 2010).

Most analyses of size variation in the fossil record focuson adult variation. Ontogenetic variation is informative aboutlife history adaptations during growth and development(Leigh 1992, 1995) and about species life history and adap-tations in general (Leigh and Blomquist 2007). Ontogeneticvariation is relatively easy to recognize as a separate sourceof variation because juvenile mammals leave indications ofimmaturity, including unfused epiphyses and cranial suturesand unerupted teeth. Variation in patterns of growth and



development themselves constitute important evidence abouta species biology, morphology, and life history (e.g., studiesof KNM-WT 15000; Schwartz 2012) and so are outside thescope of this review. But it is also important to recognize thatadult variation is the product of ontogenetic variation. Hence,factors that may affect growth and development—such asnutrition, disease, or other intrinsic factors—may ultimatelygenerate adult size variation (Kuzawa and Bragg 2012). Pat-terns of adult morphological variation also are altered byvariation rates of growth and maturation of organ systemsand skeletal components (e.g., Isler et al. 2007; Leuteneggerand Masterson 1989a, 1989b; Lockwood 1999; Plavcan 2003),leading to natural interspecific variation in the patterns oftrait variation.

Outside of ontogenetic variation, we can think of variationwithin species as falling into three classes (simplifying thescheme of Albrecht and Miller 1993): intrinsic character var-iation within an interbreeding population, intersexual varia-tion indicative of separate selective forces acting on males andfemales or sexual differences in response to environmentalfactors (Altmann and Alberts 2005; Fernandez-Duque 2011;Kuzawa and Bragg 2012; Wells 2007, 2012), and interpopu-lational variation due to variable local adaptations or temporalshifts in selective or environmental factors. A final source ofvariation in fossil samples is the mixing of multiple taxa. Thiscan be thought of as confounding variation because it bothinterferes with the recognition of taxonomic diversity andobscures levels of intraspecific variation. The rest of this articlewill first review what is known about the pattern and causesof these sources of variation in living species as it might applyto the fossil record generally and then briefly consider vari-ation in fossil Homo in the light of variation in living primates.

Intrinsic Population Variation

Variation is a natural feature of any morphological characterand is classically separated into genetic and environmentalcomponents (Lande 1980). Identifying the separate geneticand nongenetic components of variation in living species isdifficult and requires information on pedigrees and, ideally,all of the environmental variables that might influence theexpression of a trait from growth in the womb to diet anddisease process in an adult. In practice, this sort of analysisof the causes of normal interindividual population variationcannot be carried out for fossil samples. However, if there arelimits to what might be thought of as “normal” variationwithin populations, then at least a rough baseline of com-parison can be established to help recognize variation due toenvironmental, geographic, temporal, or sexual differences.This logic forms the foundation of most comparative studiesof variation in fossils that employ living species as a baselineof comparison for the fossils (Albrecht and Miller 1993; Copeand Lacy 1992; Fleagle, Kay, and Simons 1980; Kelley and

Plavcan 1998; Kramer et al. 1995; Lockwood et al. 1996; Plav-can and Cope 2002).

Critically, variation itself is size dependent, but in a regular,mathematically predictable way. In absolute terms, gorillabody size (average female mass 71,500 g in western lowlandgorillas; Smith and Jungers 1997) is more variable than thatof pygmy marmosets (average female body mass 122 g; Smithand Jungers 1997). A 5% weight difference from the averagegorilla female is 3,575 g, more than 29 times the average massof the pygmy marmosets. Thus the standard deviation (SD)of any trait in a gorilla will be enormously larger than thatof the same trait in a marmoset.

Consequently, any study of size variation in living or fossilspecies must compare measures of relative size (Plavcan andCope 2002; Sokal and Braumann 1980; Yablokov 1974). Thereare two simple methods: the coefficient of variation (CV) andthe simple log transformation of raw data (one should not,however, compare CVs of log-transformed data, as this willreintroduce a correlation of size with variation). CVs are com-monly used, being simply the SD divided by the mean of atrait multiplied by 100 to represent the value as a percentage(Sokal and Braumann 1980). A correction factor is usuallyadded for small sample sizes. Log transformation is moreconvenient, however, for statistical comparison of the data(Plavcan and Cope 2002), though nonparametric tools suchas the Fligner-Killeen test (Kramer et al. 1995) are also pow-erful.

Traits differ in the magnitude of variation, but most, ad-justing for the correlation between size and variation, fallwithin a modest range (Yablakov 1974). As a rule of thumb,molar teeth are the least variable dental traits, with within-sex CVs ranging from about 1.5% to 8% across primates forrelatively geographically restricted samples (Gingerich andSchoeninger 1979; Plavcan 1993). Osteological traits likewisetend to show low within-sex variation. Among 40 craniofacialdimensions for 135 species of primates, within-sex CVs formost average less than 10% (fig. 1). Postcranial variation issimilarly low (Gordon 2004). Within-sex CVs for body massitself range from about 12% to 15% (Smith and Jungers 1997)for a sample of 19 primates, which, when accounting fordimensionality, is comparable to values for dental, cranial,and skeletal characters. Thus, within-sex variation is generallylimited across primates. It should be emphasized that theseassessments of variation refer to living species sampled fromrestricted geographic ranges or at least subspecies of livingspecies.

Most anthropoid primates, including humans, are sexuallydimorphic to some degree. This means that variation needsto be considered within the sexes separately to control for theeffect of dimorphism on population variation. It has beensuggested that males are more variable than females in pri-mates (Leutenegger and Cheverud 1982, 1985) and that suchvariability is causally related to the evolution of sexual di-morphism. Controlling for the correlation between size andvariation, males and females on average do not differ from

Plavcan Variation in Early Homo S411

Figure 1. Profile of average coefficients of variation (CVs) for 40 linear craniometric dimensions from females in 135 primate taxa.Male profiles are nearly identical. For most characters, CVs do not exceed a value of 8.

one another across species (Plavcan 2000b; Plavcan and Kay1988; Smith and Jungers 1997). However, males of stronglydimorphic species do tend to be more variable than femalesin raw data space. Interestingly, males of Pongo may showunusual size variation in association with a “waiting room”mating strategy (Utami Atmoko and van Hooff 2004). Adultmale orangutans are dimorphic in that only resident domi-nant males achieve full body size with the development ofstrong secondary sexual features, including cresting and ro-busticity of the skull. Other males retain a subadult body formeven though the dentition has completely erupted and thesemales are capable of reproduction (Utami Atmoko and vanHooff 2004). The result is that while male teeth are not nec-essarily more variable than those of females, skeletal and cra-nial features are (Leutenegger and Masterson 1989a, 1989b).This pattern of variation is exceptional among primates, buta similar pattern has been associated with variation in Par-anthropus robustus (Lockwood et al. 2007).

Although in general, trait variation tends to be limited inpopulations, some traits show intrinsically higher or lowerdegrees of variation (fig. 1). Those interested in assessing thecause of variation in fossils intentionally focus on charactersthat show inherently low degrees of variation. For example,the first molar teeth tend to be the least variable and showthe lowest degrees of sexual dimorphism in the dentitionregardless of size dimorphism (Gingerich and Schoeninger1979). Therefore, tests of whether fossil dental samples showinflated variation usually focus on the molar teeth. Articularsurfaces of joints are strongly correlated with body size andtend to show less variation than, for example, diaphysealbreadths (Lieberman, Delvin, and Pearson 2001; Ruff 2002;Trinkaus, Churchill, and Ruff 1994). Wood and Lieberman(2001) evaluated whether craniodental traits that might beunder greater or lesser mechanical loads or that show differentdegrees of heritability might be expected to show more or

less population variation. Wood and Lieberman (2001) pro-vide evidence that while heritability is not systematically as-sociated with greater or lesser variability, craniofacial traitssubjected to low strains (e.g., from the basicranium and neu-rocranium) tend to show less variation than those subjectedto higher strains and therefore are preferable for assessingwhether populations show unusual variation.

Sexual Dimorphism

Sexual dimorphism has played a key role in debates aboutthe biological and taxonomic meaning of variation in earlyHomo, especially Homo habilis, Homo rudolfensis, and Homoerectus (e.g., Anton 2003; Miller 2000; Rightmire, Van Arsdale,and Lordkipanidze 2008; Skinner, Gordon, and Collard 2006;Wood 1993). In most anthropoid primates, males are largerthan females. Females can be larger than males in callitrichids,while strepsirrhines show minimal dimorphism (Kappeler1991; Smith and Jungers 1997). Gorillas show the greatestdegree of body mass dimorphism among hominoid primates,though Mandrillus sphinx appears to be the most size-dimorphic anthropoid (Gordon 2004). Modern humans showan unusual pattern of relatively modest body mass dimor-phism (about 15%; Smith and Jungers 1997) but moderate“lean” body mass dimorphism (about 44%; Wang et al. 2001).This latter value reflects a proportionally greater amount offat in human females (Plavcan 2012; Wang et al. 2001; Wells2007, 2012). Skeletal dimorphism in modern humans is mod-erate and slightly greater than that of chimpanzees (Gordon,Green, and Richmond 2008), but when corrected for dimen-sionality it is proportional to lean body mass dimorphism inmagnitude (Plavcan 2012). Cranial dimorphism in modernhumans tends to be modest, being intermediate between bodymass dimorphism and lean body mass dimorphism (Plavcan2012).


Because most cranial, dental, and postcranial traits are nor-mally or at least unimodally distributed in males and females,the effect of dimorphism is to simultaneously increase thetotal sample variance and generate bimodality (Godfrey, Lyon,and Sutherland 1993; Plavcan 1994, 2000b; Plavcan and Kay1988). For fossils, dimorphism is commonly estimated usingone of several techniques that correlate sample variation withdimorphism (Gordon, Green, and Richmond 2008; Joseph-son, Juell, and Rogers 1996; Plavcan 1994; Plavcan and Cope2002) largely because the sex of individual specimens cannotbe reliably determined. It is common practice to assume thatlarge specimens are male and small specimens are female (e.g.,Lockwood 1999; McHenry 1992). However, this practice willtend to overestimate the magnitude of sexual dimorphism (aswill all methods) by ignoring overlap in male and femaledistributions, especially when true dimorphism is slight.

Importantly, while dimorphism in some skeletal featuressuch as pelvic size and shape can reflect functional or devel-opmental differences between males and females, dimorphismin most skeletal features is often assumed to reflect overallbody size dimorphism. However, patterns of dimorphism dif-fer throughout the dentition, skull, mandible, and skeleton,and such patterns differ between species (Gordon, Green, andRichmond 2008; Plavcan 2001, 2002, 2003). For example,within M. sphinx, possibly the most size-dimorphic primatealive (ratio of male to female body mass 2.69; Gordon 2004),dimorphism in skull dimensions ranges from a low of 1.06(male mean divided by female mean) for postorbital breadthto a high of 2.5 for zygomatic arch thickness (Plavcan 2002).Orbital dimensions differ dramatically in dimorphism in thisspecies (as in most other primates), with orbital height di-morphism of 1.07 and biorbital breadth dimorphism of 1.73.The implication is that one cannot simply compare variationacross traits and assume that they will yield similar signals ofsize dimorphism for any particular species (Plavcan 2002,2003).

The issue of estimating the magnitude of dimorphism isimportant because dimorphism constitutes critical evidenceof behavior and life history in extinct species and thus hasweighed heavily in discussions of the evolution of homininbehavior (e.g., DeSilva 2011; Gordon 2006a; Lovejoy 1981,2009; Martin, Willner, and Dettling 1994; McHenry 1994;Moore 1996; Plavcan and van Schaik 1997a). Dimorphism isone of the only anatomical traits that is directly causally re-lated to social behavior that is preserved in the fossil record(Plavcan 2004a). But dimorphism is a complex phenomenon,and a simple one-to-one correspondence between behaviorand the magnitude of dimorphism does not exist (Plavcan2000a). Dimorphism reflects separate causal factors influenc-ing male and female traits whose expression is potentiallylimited by the genetic correlation between males and females(Gordon 2006a, 2006b; Greenfield 1992; Lande 1980; Leigh1992; Lindenfors 2002; Martin, Willner, and Dettling 1994;Plavcan 2011; Plavcan, van Schaik, and Kappeler 1995).Therefore, in order to understand the biological implications

of dimorphism in the fossil record, we need to understandthe factors that affect both male and female size.

Body size dimorphism should represent a balance of thecosts and benefits of altering body size from a theoreticaloptimum (Fairbairn 1997; Gordon 2006a; Lande 1980; Leighand Blomquist 2007). In all nonhuman primates, males andfemales occupy similar niches, living together, eating the samefoods, occupying the same substrates and territories, and sus-ceptible to the same forces of ecological variation and pre-dation. While the phenomenon of niche dimorphism is com-mon in, for example, monogamous birds (Emlen and Oring1977; Selander 1972), it has never been demonstrated in pri-mates (Plavcan 2001). Rather, those differences in diet, lo-comotion, and substrate use that do occur in primates aremore likely a consequence of dimorphism than a cause of it(Clutton-Brock, Harvey, and Rudder 1977). The result is thatin the absence of selective factors uniquely targeting size inone sex, both sexes should achieve approximately the samebody size. Corroborating this is the observation that mono-morphism has evolved independently in association with anabsence of agonistic male competition and sexual selectionfor male size (Lindenfors and Tullberg 1998).

The male contribution to dimorphism in primates is gen-erally thought to be straightforward: agonistic male-malecompetition for mates as predicted by sexual selection theory(Clutton-Brock, Harvey, and Rudder 1977; Ely and Kurland1989; Ford 1994; Gaulin and Sailer 1984; Gordon 2006b; Kayet al. 1988; Lindenfors 2002; Lindenfors and Tullberg 1998;Martin, Willner, and Dettling 1994; Mitani, Gros-Louis, andRichards 1996; Plavcan 2001, 2004b; Plavcan and van Schaik1992, 1997b). Where males can monopolize access to receptivefemales to the exclusion of other males, competition resultingin male reproductive skew will ensue. Because body size helpsmales win contests and is heritable, selection should favorlarge male size. Multiple comparative studies have corrobo-rated the sexual selection hypothesis using surrogate measuresof sexual selection and male competition, including compe-tition levels (the potential frequency and intensity of malecompetition), breeding system (monogamous/polyandrous,multimale, single male), socionomic sex ratio (the numberof adult males to adult females per group), and operationalsex ratio (the number of adult males to the number of re-ceptive adult females in a group; Cheverud, Dow, and Leu-tenegger 1985; Ely and Kurland 1989; Ford 1994; Gaulin andSailer 1984; Gordon 2006a, 2006b; Greenfield 1992; Leuten-egger and Kelly 1977; Lindenfors and Tullberg 1998; Mitani,Gros-Louis, and Richards 1996; Plavcan 2004b; Plavcan andvan Schaik 1992; Plavcan, van Schaik, and Kappeler 1995).

The above analyses have been used to support inferencesof polygyny or specific mating systems in hominins and otherextinct taxa (e.g., Gordon, Green, and Richmond 2008; Kap-pelman 1996; Lockwood 1999; McHenry 1994). However, inno study is the magnitude of size dimorphism uniquely as-sociated with one or the other breeding system, competitionlevel, sex ratio, or operational sex ratio (OSR; Plavcan 2000a).


While very strong dimorphism is invariably associated withpolygyny, skewed OSRs, and intense male competition, mono-morphism is not uniquely associated with any particular mat-ing system. This means that a lack of dimorphism alone ina fossil sample cannot be used as evidence for monogamy, ahumanlike mating system, polyandry, or any other matingsystem (Plavcan 2004a).

Apart from sexual selection acting on males, no other factor(predation, diet, life history variables, etc.) has been shownin comparative analysis to be consistently associated with di-morphism, at least as far as the male contribution to dimor-phism (Ford 1994; Gordon 2006a; Plavcan 2001). This doesnot mean that male size is only affected by sexual selectionin every species. Rather, across species this is the only factorto have a consistently detectable association with dimorphism.

Recent years have seen greater attention to the female con-tribution to body size dimorphism (Bribiescas, Ellison, andGray 2012; Clutton-Brock, Harvey, and Rudder 1977; Gordon2006a, 2006b; Kuzawa and Bragg 2012; Lindenfors 2002; Mar-tin, Willner, and Dettling 1994; Plavcan 2004b; Wells 2007,2012). Several models posit the importance of changes infemale body size in the evolution of hominin dimorphism(Cartmill and Smith 2009; Gordon 2006a; Lovejoy 2009; Mar-tin, Willner, and Dettling 1994). There is clear evidence thatdimorphism changes through changes in female size withinspecies, even in monogamous species. For example, Fenandez-Duquez (2011) demonstrates that dimorphism in Aotus variesas a result of changes in female size correlated with longitude.Similarly, intraspecific variation in dimorphism associatedwith changes in female trait size has been documented forhowler monkeys (Jones et al. 2000), vervet monkeys (Turner,Anapol, and Jolly 1997), baboons (Dunbar 1990), macaques(Macaca fascicularis [Fooden 1995]; Macaca nemestrina [Al-brecht 1980]), and guerezas (Hayes, Freedman, and Oxnard1995).

Females of most primates compete for resources rather thanaccess to mates (Sterck, Watts, and van Schaik 1997). Selectionfor larger or smaller female size centers on the balance be-tween resource abundance and reliability, risks to infant andmaternal mortality, and maximizing reproductive success(Gordon 2006a; Kuzawa and Bragg 2012; Leigh 1995; Lin-denfors 2002; Ralls 1976; Turner, Anapol, and Jolly 1997; Wells2012). Factors hypothesized to favor larger females are re-source competition (Gordon 2006a; Leigh and Shea 1996;Lindenfors 2002; Plavcan 2011) and the fact that larger fe-males can produce larger offspring that suffer lower mortalityrisk, can produce more or better milk (thereby allowing off-spring to grow faster), or may be better at defending andcarrying offspring (Gordon 2006a; Kuzawa and Bragg 2012;Ralls 1976; Smith et al. 2012; Wells 2012). Factors hypothe-sized to favor smaller females are selection for early cessationof growth (favoring earlier breeding and increased reproduc-tive output) and size reduction to decrease absolute metabolicdemand when individuals may face periods of limited re-

sources (Gordon 2006a; Leigh 1995; Leigh and Shea 1996;Lindenfors 2002; Martin, Willner, and Dettling 1994).

Within humans, some evidence suggests that small femalebody size is favored where resources are scarce during lac-tation (Gordon 2006b; Ralls 1976). Kuzawa and Bragg (2012)note that larger female body size may be an epiphenomenalresponse associated with better maternal nutrition. At thesame time, human males appear to respond more rapidly tochanges in nutrition than human females (Anton and Snod-grass 2012; Bribiescas, Ellison, and Gray 2012).

Each of the above hypotheses has some data to support it.Still open to question is how much dimorphism can be at-tributed uniquely to changes in female size in any given spe-cies. There is no species of primate that shows strong sizedimorphism without male competition and polygyny. Twocercopithecoid taxa were thought to show a combination ofstrong size dimorphism and monogamy (Cercopithecus ne-glectus and Simias concolor; Leutennegger and Lubach 1987).Further study of both of these species demonstrated that po-lygyny is present in other parts of the species range and thathunting pressure explained the presence of monogamouspairs at least in S. concolor (Brennan 1985; Watanabe 1981).On the other hand, there are no monogamous or polyandrousanthropoids that show strong size dimorphism, though sev-eral monogamous species do show modest levels of size di-morphism (Plavcan 2000a, 2001). If selection to alter femalesize—either larger or smaller—is common independent ofchanges in male size and contributes significantly to inter-specific variation in size dimorphism, then we should expectto see greater variation in dimorphism among monogamousand polyandrous species. That this is not the case suggeststhat large male size in dimorphic species is maintained inspite of pressure to reduce male size to that of females andthat most substantial dimorphism in primates is a functionof changes in male and not female size.

On the other hand, that variation in dimorphism withinspecies has been documented and tied to changes in bothmale and female size (Albrecht 1980; Altmann and Alberts2005; Fernandez-Duque 2011; Gordon 2006b; Plavcan, vanSchaik, and McGraw 2005; Turner, Anapol, and Jolly 1997)raises the question of how to reconcile broadly based inter-specific comparative analyses with studies of changes withinspecies. The answer probably lies partly in the scale of com-parison. Broadly based comparative analyses offer statisticaltests of factors that should have a general effect on dimor-phism, but they mask considerable variation. Even the stron-gest behavioral correlate of dimorphism in anthropoid pri-mates explains less than half of the interspecific variation indimorphism among species. Beyond this, there are few dataallowing a systematic, broadly based comparative analysis thatcan effectively quantify variation in the factors that affectfemale size and hence the general contribution of changes infemale size to dimorphism.

The implications of this for the fossil record and for modelsof human evolution are important. It appears that large


Figure 2. Box plots showing variation in skull size in Pongo andCebus olivaceus. Skull size is represented by a geometric mean ofnine dimensions.

Figure 3. Box plots illustrating strong geographic variationamong baboons and Hanuman langurs. Note that changes infemale size are consistently accompanied by parallel changes inmale size.

changes in dimorphism are not likely to be either gained orlost through changes in female size alone. Rather, a loss orgain of dimorphism accompanied by substantial shifts in fe-male size should signal both selection to alter female size anda change in selection on male size, most likely because ofchanges in female distribution or behavior that alter the mo-nopolization potential of females (Plavcan, van Schaik, andMcGraw 2005).

Beyond this, two points need to be emphasized. First, thecorrespondence between dimorphism and categorical esti-mates of breeding system or male competition commonlyused to support inferences of mating system or male com-petition in hominins (e.g., Kappelman 1996; Kimbel and De-lezene 2009; Lovejoy 1981, 2009; McHenry 1994; Moore 1996;Plavcan 2000a, 2001; Plavcan and van Schaik 1997a) is toocrude to provide anything but the most general information—substantial size dimorphism is associated with male compe-tition and polygyny (Plavcan 2000a). A single numerical es-timate of dimorphism tells us little else. However, as ourknowledge of the factors that change male and female sizeincreases, the greatest information about changes in behaviorand life history in the fossil record and in human evolutionshould come through the study of temporal changes in maleand female size relative to each other. But before this can bedone, much more study is required of the forces that influencefemale size independently of male size.

Temporal and Geographic Variation

Geographic variation is common in primates and can be sub-stantial. Well-documented examples of strong size variationinclude savanna baboons, various macaques, and Hanumanlangurs (Albrecht and Miller 1993; Dunbar 1990; Jolly 2001).Because geographic variation represents population-specificsize changes within a species, it can be used to model temporalvariation (Albrecht and Miller 1993; Plavcan 1993; Shea,Leigh, and Groves 1993). At the same time, geographic and

temporal variation have been cited as factors confoundingattempts to recognize and estimate dimorphism when the sexof individuals cannot be determined, and these factors haveplayed significant roles in debates concerning how much var-iation should be expected within fossil species especially if thefossils show a wide geographic or temporal distribution (Al-brecht and Miller 1993; Gordon, Green, and Richmond 2008;Plavcan, van Schaik, and McGraw 2005; Reno et al. 2003,2010; Shea, Leigh, and Groves 1993).

The magnitude of size differences between closely relatedspecies or subspecies can vary widely. Figure 2 shows variationin skull size in closely related species or subspecies of Pongoand Cebus olivaceus. Size differences among the groups ineach set are comparatively minimal. In contrast, figure 3shows similar plots illustrating strong geographic variationamong subspecies of Papio and Semnopithecus.

While numerous studies have documented geographic var-iation in primate size, comparatively few have tested hypoth-eses about the specific causes of size differences among pop-ulations (Albrecht and Miller 1993). Those have found thatbody size tends to follow Bergmann’s rule, with size increasingwith increasing latitude, and that populations living in moreproductive habitats with higher rainfall and less seasonalitytend to be bigger (Albrecht and Miller 1993; Dunbar 1990;Fernandez-Duque 2011; Fooden 1995; Katzmarzyk and Leon-ard 1998; Plavcan, van Schaik, and McGraw 2005; Ruff 2002;Turner, Anapol, and Jolly 1997). However, Albrecht and Miller(1993) caution that some species fail to show this relationshipbetween productivity and size (toque macaques), while othersactually reverse the trend (Sundaic pigtail macaques). Plavcan,van Schaik, and McGraw (2005) find that while intraspecificsize variation is associated with variation in latitude and pro-ductivity, interspecific size variation is not correlated withmeasures of rainfall or seasonality.

Male and female size have been shown to vary amongpopulations independently within species but in association


Figure 4. Box plots of skull bizygomatic breadth data from theHowells (1973) data set for 10 populations of humans comparedwith skull lengths for Gorilla gorilla gorilla. The dashed verticalline separates the Gorilla samples from the Homo samples. Sam-ples were selected from the Howells (1973) data set to representthe maximum range of variation in female skull size. Bizygomaticbreadth was chosen for comparability to the Gorilla sample.Other variables show a similar pattern. Note that in spite of thepattern of clear differences among human populations, overallhuman variation is limited. Note also that for all populations,changes in female size are accompanied by parallel changes inmale size. Females are indicated by gray-shaded boxes, males byunshaded boxes.

Table 1. Coefficients of variation (CV) of bizygomaticbreadth for samples of humans, apes, and monkeys

Sample CV N

Howells data set:All 6.0 2,412Average 4.9 26a

Range 4.1–5.6Papio spp.:b

All 12.9 162Average 9.6 4a

Range 7.2–12.3Semnopithecus entellus:

All 12.4 116Average 6.9 6a

Range 6.3–8.4Gorilla gorilla 11.0 79Pan pansicus 5.4 39Pan troglodytes troglodytes 6.4 47Pan troglodytes schweinfurthii 5.8 31Pan troglodytes combined 6.4 78Hylobates lar 4.4 53

a Average CV based on 26 samples in the Howells (1973) data set,four for Papio, and six for Semnopithecus. For the two monkeys,CVs for subgroups were only calculated for groups with N of nineor more and both sexes present in the sample.b Nonhuman primate data collected by the author.

with different factors. For example, Turner, Anapol, and Jolly(1997) demonstrate that Cercopithecus aethiops female sizevariation changes with resource abundance—where popula-tions have reliable and abundant resources and with increas-ing rainfall females increase in size with a concomitant re-duction in size dimorphism. In contrast to this, Altmann andAlberts (2005) demonstrate that compared with baboons liv-ing on an entirely wild diet, “resource enhanced” baboonsshow greater size dimorphism through a differential increasein male body mass relative to that of females. Dunbar (1990)also presents evidence that Papio size and size dimorphismincreases with increasing rainfall specifically through changesin male size. Albrecht (1980) demonstrates a similar phe-nomenon for Macaca nemestrina, noting that both male andfemale skull size increase with increasing distance from theequator but that male skull size increases relatively faster.

Variation in body size is well documented for extant Homo(Albrecht and Miller 1993; Katzmarzyk and Leonard 1998;Kuzawa and Bragg 2012; Ruff 2002) and tends to follow sim-

ilar patterns to those seen in nonhuman primates. Humanstend to be larger in more productive habitats, and body sizetends to increase with increasing distance from the equator.Human body size is also thought to vary with nutrition anddisease (Ruff 2002). Local adaptation to variation in climate,resource availability, and other factors is also thought to con-tribute to variation in human body size. Kuzawa and Bragg(2012) specifically note that human body size tends to increaserelatively faster in males with increasing resource abundance.

With regard to the fossil record, it is of great interest toknow how modern human interpopulation variation in cra-nial and skeletal size compares with that of extant primates.Figure 4 shows human skull bizygomatic breadth data fromHowells’s (1973) data set compared with data for a sampleof extant Gorilla gorilla gorilla. The human samples illustratedin the figure were selected to represent the greatest range ofvariation among the populations available in Howells’s (1973)data set, yet still, considered together, they show considerablyless variation than seen between the sexes of the extant Gorillasample. The CV for the entire Howells (1973) data set( samples with both males and females represented)N p 26is 6.0, which is approximately half that of the Papio, Sem-nopithecus, and Gorilla samples shown in figures 3 and 4 andis also less than that of a single subspecies of chimpanzees(table 1). Geographic variation is clearly present in the humansample, but combining even the most disparate populationsof humans does not increase the total sample variation sub-stantially (the average CV for individual human samples inthe Howells [1973] data set is 4.9). Note that in spite of theoverall size differences among the human samples, sexual di-


Table 2. Coefficients of variation (CV) of femoral headdiameter for samples of humans, apes, and monkeys

Sample CV N

Homo sapiens:All 9.2 1,519Average 7.7 11a

Range 6.4–9.1Hylobates lar 4.4 113Pan paniscus 5.5 18Pan troglodytes 6.5 109Gorilla gorilla 12.0 122Pongo pygmaeus 13.2 14

Sources. Human data from Auerbach and Ruff (2004); nonhuman pri-mate data from Gordon (2004).a Number of samples.

Figure 5. Box plots showing variation in femoral head diameterfor 11 human populations from the Goldman data set (Auerbachand Ruff 2004) and Gorilla gorilla gorilla (from Gordon 2004).Samples of humans were selected on the basis of sample size;only samples with more than 12 males and 12 females wereincluded. Note that as for the craniometric variable in figure 4,male and female distributions tend to parallel one another. Atthe same time, the overall variation among the humans is moresimilar to that of the Gorilla sample. However, also note that inhumans, male and female distributions consistently overlap oneanother, reflecting lower degrees of dimorphism than the Gorillasample, in which the male and female distributions do not over-lap. Females are indicated by gray-shaded boxes, males by un-shaded boxes.

morphism remains relatively stable. Male versus female bi-zygomatic breadth is highly correlated in the Howells (1973)data set ( , samples, ). Thus, evenr p 0.971 N p 26 P ! .001though dimorphism might be influenced by shifts in male orfemale size, the overall magnitude of the effect is minimal bycomparison with overall human variation and especially bycomparison with the magnitude of variation typically seen insingle species and even subspecies of extant primates.

Geographic variation in human femoral head size, whichis often viewed as one of the best proxies for body size, showsthe same pattern as seen for human bizygomatic breadth (fig.5). The CV of femoral head diameter for all human samplescombined is 9.2, while that for the Gorilla sample is 12.1(table 2). Within-population CVs for the human sample rangefrom a low of 6.4 to a high of 9.1 with an average of 7.7,suggesting, as for the bizygomatic breadth data, that geo-graphic variation, while clearly present, does not inflate com-bined-population variation to a great degree. Even so, thetotal range of femoral head size across human populations is

comparable with that of the Gorilla sample. By contrast, theCV for the same dimension for a single subspecies of chim-panzees (Pan troglodytes troglodytes) is 6.5. It is notable thatvariation is similar for cranial and postcranial dimensions ingreat apes while the human postcranial variation is greaterthan that of the skull dimension. This reflects the fact thatthe magnitude of human postcranial dimorphism is similarto that of the lean body mass dimorphism while that of skulldimorphism is similar to that of the total body mass dimor-phism (Plavcan 2012). Finally, as for bizygomatic breadth,male and female changes in femoral head size change in par-allel to one another, with for male versus femaler p 0.934femoral head diameter ( , for all samples ofN p 42 P ! .001Homo in the Goldman postcranial data set). Thus, while di-morphism varies across populations (fig. 6), it is actually min-imally influenced by changes in the size of either sex by com-parison with interspecific variation in the magnitude ofdimorphism.

This observation is potentially important for translatinginferences about epigenetic changes in modern human sizeand dimorphism to the fossil record (Anton and Snodgrass2012; Kuzawa and Bragg 2012). While it is clear that changesin diet and resource availability can and do alter male orfemale growth and consequent size dimorphism in modernprimates and humans (Altmann and Alberts 2005; Gordon2006a; Turner, Anapol, and Jolly 1997), what we see in thehuman data is that at least for the craniometric data andfemoral head size available here, geographic variation in mod-ern humans is relatively limited by comparison with intra-sexual variation in a single geographically restricted sampleof Gorilla, and variation in the magnitude of dimorphismamong regional human samples is comparatively restricted.

Temporal Variation, Sexual Dimorphism,and Fossil Homo

Recent discoveries of early Homo highlight the importance ofvariation in the fossil record. Fossils from Dmanisi, Olorge-salie, Gona, as well as a well-preserved skull of early Homo


Figure 6. Sexual dimorphism (male average divided by femaleaverage) for samples of Homo from the Howells (1973) data setand samples of extant great apes. Each circle represents a valueof dimorphism for a single sample. The samples for Homo andGorilla are the same as were used in figure 5.

from Kenya (KNM-ER 42700) in particular suggest not onlyan increase in size variation across time and space (Anton2012; Anton and Snodgrass 2012; Rightmire, Van Arsdale,and Lordkipanidze 2008) but also that sexual size dimorphismmay have been substantial in early Homo (Spoor et al. 2007).It has long been suggested that there has been an increase inbody size in Homo erectus over earlier hominins and that areduction in size dimorphism in the human lineage occurredthrough a relative increase in female body size, yet recentdiscoveries underscore the uncertainty in these claims (Aielloand Key 2002; McHenry 1992, 1994).

Size variation and dimorphism for early Homo and H.erectus has largely been assessed using craniometric data andestimates of body mass based on postcranial variables. Butany assessment of size variation in Homo needs to be qualifiedby the fact that the fossil record is sparse and yields only cluesabout size. Assessments of size in living species usually focuson body mass as an estimate of size, and comparisons of thefossil record should understand that skeletal size variation isnot uniform and isometric within or between species (Gor-don, Green, and Richmond 2008), making comparisons ofestimated mass in fossils to actual mass of living species proneto significant error (Smith 1996). As noted above, dimor-phism among skeletal features with a single species can varysubstantially. Our inability to accurately identify sex withavailable remains adds considerable uncertainty to estimatingthe magnitude of dimorphism, particularly when dimorphism

is modest. Most postcranial remains identified as early Homoare not associated with cranial remains, making taxonomicattributions uncertain. Finally, small sample sizes introducesubstantial error into any estimate of size variation even forliving species. In spite of the bleak uncertainty of the fossilrecord, we can still estimate size and size variation and withthose estimates provide limited assessments and establish atthe very least a framework for establishing hypotheses andplace limitations on what can and cannot be said with theremains that we do have available.

Estimates of body size based on postcranial size, summa-rized in Pontzer (2012), are illustrated in figure 7 comparedwith a series of extant primates for which body mass datafrom wild-shot specimens are available (Isler et al. 2007). Alldata are ln transformed to standardize variation across thesize range. For the fossils, presumed sex is indicated accordingto Pontzer (2012) and Anton (2012). Several features of thisgraph immediately stand out. First, sexual dimorphism inboth Australopithecus, and in Homo habilis/rudolfensis appearsstrong with no overlap between sexes. This is because sex inthese specimens is assigned on the basis of size itself. All threeof these samples appear to show bimodality, which if the sexassignments are correct would imply strong size dimorphism.Importantly, the H. habilis/rudolfensis sample is likely com-posed of two species (Anton 2012), making any assessmentof dimorphism highly uncertain and contingent on the tax-onomic assignment of specimens. However, even with theconfounding effect of taxonomic mixing in at least one sam-ple, the range of variation in all of these samples is comparablewith that of Hylobates lar, which shows only slight dimor-phism, suggesting caution. More interestingly, the total rangeof variation of all four fossil samples falls within the rangeof all extant species, including the gibbons. However, onlythe H. erectus sample is not significantly more variable thanthe gibbon sample. In fact, for H. erectus, in terms of relativevariation, the total range of estimated body mass falls withinthat of a single-sex sample of Hylobates and Pongo abelli.Notably, the Dmanisi specimens, which as a group are indeedsmaller on average than other H. erectus (Anton and Snod-grass 2012), do not in fact alter the range of estimated bodymasses. Assuming that the sex assignments are correct, theDmanisi specimens do increase the range of variation in eithersex beyond what is seen within geographically restricted sin-gle-sex samples of extant species. The impression left in thiscomparison is that variation within the entire H. erectus sam-ple is unremarkable by comparison with that found in typicalliving primates.

McHenry (1992, 1994) suggested that H. erectus may havelost sexual size dimorphism through a disproportionate in-crease in female size. Close examination of figure 7 supportsthis notion, albeit very weakly. Homo habilis/rudolfensis“males” are about the same size as those of H. erectus, whileH. habilis/rudolfensis “females” are visibly smaller even thoughthe H. erectus sample illustrated here includes the Gona andDmanisi specimens. As noted by Anton (2012) and Anton


Figure 7. Estimated body mass (Pontzer 2012) for hominins based on postcrania compared with body mass data for selected extanttaxa (Isler et al. 2007). Sex assignations for hominins are according to Pontzer (2012). Arrows for Homo erectus indicate Dmanisispecimens.

and Snodgrass (2012), the sample sizes are small, sex assig-nation is uncertain, and there are almost certainly two taxain the H. habilis/rudolfensis sample, meaning that support forthe hypothesis must be considered weak, and could changewith the addition of a single specimen to the early Homosample.

Figure 8 suggests that there is an effect of geographic var-iation on cranial size in H. erectus. Spoor et al. (2007) analyzeskull length in samples of Pan, Gorilla, and H. erectus, sug-gesting that variation within the fossil sample may indicatesubstantial size dimorphism, especially given the diminutivesize of KNM ER 42700. Figure 8A shows skull lengths of H.erectus plotted against age along with a sample from theHowells (1973) data set. The combined-sex CV for samplesof Gorilla gorilla and Gorilla beringei for the same measure-ment is 12.1 and 10.1 respectively (Plavcan 2003), and thecombined-sex CV for the extant humans shown in figure 8Ais 3.4 ( , Norse sample, Howells [1973] data set). TheN p 110CV for the entire H. erectus sample is 7.9 ( ). However,N p 22the H. erectus sample is correlated with age ( ,r p 0.683

, ). Evaluating the standard error (SE) of re-N p 22 P ! .001siduals of the fossils and living species, variation of H. erectusdrops when controlled for geological age as expected giventhe significant correlation (table 3). However, breaking thesample into African, Javan, and Chinese samples shows thatthe African sample, which is older (and includes the Dmanisispecimens), shows almost twice the residual variation of thelater samples, which themselves show variation comparableto modern Homo and Pan paniscus, and less than Gorilla andPan troglodytes. Figure 8A shows that most of the excessive

variation in the African sample can be attributed to OH 9and KNM-ER 42700. If the large African variation is due tosexual dimorphism, than the magnitude would lie betweenthat of chimpanzees and gorillas. However, the specimens areseparated geographically and by 320,000 years in time, a mag-nitude easily unappreciated in a graph.

Figure 8B shows the sample comparison using biasterionicbreadth. For the ape samples, bimastoid breadth is substi-tuted, which should be roughly comparable for the purposesof this comparison. A similar pattern to that seen for skulllength is found in terms of a correlation with age and greatervariation in the African versus the Javan and Chinese samples.However, the Chinese sample is clearly smaller in this di-mension than contemporaneous Javan specimens (Anton2002; Kidder and Durband 2004), and while numerically thesize-adjusted residual variation of the early African specimenslies between that of a chimpanzee and a gorilla (table 3), OH9 and KNM-ER 42700 do not stand out as extreme outliers.

Discussion

While these comparisons should not be considered the de-finitive assessment of variation in early Homo, they do servethe purpose of illustrating the difficulty of assigning specificcauses to variation in fossil samples. Ideally, we need fossilsfrom the same time and locality in order to quantify variationand sexual dimorphism and to test hypotheses about thecauses of such variation. Indeed, even studies of living speciesare strongly limited by the quality of samples available. Mostmuseum osteological collections of humans and primates


Figure 8. Skull length (glabella-opisthocranion) and biasterionic breadth plotted by age for Homo erectus compared with a singlesample of Homo sapiens (Norse) from the Howells (1973) data set. Black circles are African H. erectus, diamonds are Javan, andtriangles are Chinese specimens. Gray circles are H. sapiens. Data for H. erectus from Spoor et al. (2007) and Anton (2003). Whereages were given as a range, the approximate midpoint was used.

were not systematically collected and cannot be used to testhypotheses about the influence of changes in resource abun-dance, dietary quality, disease, and other factors on intra-specific variation in a controlled, comparative context. Thus,the baseline of comparison for testing hypotheses about thecauses of variation in fossils is limited by the quality of dataavailable for extant taxa, adding to the problems inherent tothe fossil record itself.

That said, while new discoveries of Homo erectus and earlyHomo (Homo habilis/rudolfensis) have expanded our appre-ciation of size variation in the fossil record, the current fossilrecord provides little information about geographic, temporal,or sexual variation in any species. For example, although Aus-tralopithecus afarensis appears to be strongly sexually size di-morphic in most analyses (Gordon, Green, and Richmond2008; Harmon 2006; Kimbel and Delezene 2009; Lockwoodet al. 1996; McHenry 1992; Scott and Stroik 2006), this viewis not universally accepted (Reno et al. 2010). In contrast tothis, estimated body mass for H. erectus, inclusive of Dmanisiand Gona specimens, is not significantly different from asingle sex of a single subspecies of living gibbons even thoughthe combined sample of H. erectus appears to show at leasta temporal shift in size. Therefore, even though we know thatfemale and male body mass can change through selective andepigenetic effects associated with variation in ecology, diet,disease, and other factors in humans and other primates (andthat such effects can potentially be substantial), any signalfrom such effects on the current fossil sample of H. erectuscannot be distinguished from normal intrapopulation varia-tion in a single living species.

Another illustration is provided by the shift in body sizebetween H. habilis/rudolfensis and H. erectus. As shown in

figure 7, there is an apparent shift in female size that is notaccompanied by a shift in male size, thereby giving the im-pression that sexual dimorphism is reduced from earliestHomo to H. erectus. Within H. erectus, early African samplesare clearly more variable in skull dimensions than later sam-ples, and this variation might reflect sexual dimorphism (thetemporal trend does not appear to explain the variation inthese early specimens). However, the individual sex of the H.habilis/rudolfensis specimens can only be inferred on size, andthe sample itself is so small that even the introduction of afew new specimens could completely change our image ofthis species. Furthermore, the impression of elevated sexualdimorphism in African H. erectus is essentially based on twospecimens separated in time and space whose size differencecould reflect temporal or geographic variation in overall sizeunrelated to dimorphism (see also Anton 2012). This leavesus with only a tantalizing apparent signal that is suggestiveof changes in size and dimorphism but nothing more. Cer-tainly the data are suggestive that changes in female size playedan important role in changes in human size dimorphism. Asillustrated above, though, the underlying causes for shifts indimorphism among primates are complex, and it is difficultto support any single cause directly on the basis of currentlyavailable evidence. While we can formulate reasonable modelsfor the evolution of the modern pattern of human dimor-phism, we need much better samples of fossils in order tosupport or test such models with any reasonable certainty.

These comparisons underscore several problems that needto be addressed if we are to improve our ability to interpretthe causes of variation in the fossil record. First, body sizeestimates are based on disparate parts and limited remains.A significant portion of the fossil record is omitted from


Table 3. Comparisons of variation and relative variation inskull length and biasterionic breadth in Homo erectus, liv-ing Homo, and a series of extant ape samples

Sample Skull lengtha Biasterionic breadthb

Homo erectus:All:

Mean (ln data) 5.246 4.770SD .083 .084SD age adjustedc .060 .073CV (raw data) 7.9 8.2

African:d

Mean (ln data) 5.176 4.708SD .085 .084SD age adjusted .083 .081CV (raw) 8.3 8.2

Java:Mean (ln data) 5.297 4.835SD .058 .036SD age adjusted .052 .034CV (raw) 5.5 3.5

China:Mean (ln data) 5.266 4.716SD .021 .049SD age adjusted .021 .049CV (raw) 2.0 4.6

Homo (Norse):Mean (ln data) 5.206 4.691SD .034 .046CV (raw data) 3.4 4.6

Pan paniscus:Mean (ln data) 5.109 4.677SD .024 .041CV (raw data) 3.0 4.1

Pan troglodytes troglodytes:Mean (ln data) 5.281 4.817SD .047 .060CV (raw data) 4.0 5.9

Pan troglodytes schweinfurthii:Mean (ln data) 5.264 4.783SD .044 .060CV (raw data) 4.4 6.0

Gorilla gorilla beringei:Mean (ln data) 5.646 5.030SD .126 .114CV (raw data) 10.1 11.1

Gorilla gorilla gorilla:Mean (ln data) 5.577 4.994SD .138 .112CV (raw data) 12.1 11.0

Note. SD p standard deviation; CV p coefficient of variation.a Skull length from Anton (2012) for H. erectus. Skull length for extantapes is the same measurement (glabella to inion) listed as neurocraniallength in Plavcan (2002, 2003).b Biasterionic breadth from Anton (2012) for H. erectus. For the apes,bimastoid breadth (Plavcan 2002, 2003) is used, which is based on land-marks that are close enough that the comparisons should not be affected.c Standard errors of residuals from a least squares regression of ln-trans-formed skull length against geological age as listed in Anton (2012).d Divisions of H. erectus following Anton (2012).

analysis simply because it is difficult to estimate size from theremains. As emphasized by Smith (1996), we need more sys-tematic comparative analysis of the relationship between skel-etal size and body size in living species in order to moreeffectively utilize the remains we have. Furthermore, in orderto understand the relevance of modern studies of life historyand size as either responses to selection or epiphenomenalchanges to the fossil record, we need to document the changesin both body mass and skeletal measures of size in livingspecies that are preserved in the fossil record.

Comparative studies of the selective factors that affect sizeare relatively uncommon. Early Homo undoubtedly differedfrom modern Homo, and while studies within humans arecritical for understanding human evolution, comparativestudies offer an understanding of the breadth of factors thatmight affect variation. Thus, for example, increased resourceabundance leads to larger male size in baboons and modernhumans but relatively larger female size in vervet monkeys.It cannot automatically be assumed that patterns seen in mod-ern humans will of necessity be found in early hominins, andthe only way to understand possible variation in causes is toincrease the database of comparative studies.

We need much more refined studies of geographic variationin living primates. Most of the reference samples we havefrom museums are not systematically collected, and manystudies using museum collections not only omit informationabout the geographic distribution of specimens but also oftencombine specimens of different subspecies and even speciesinto a single sample. The effect of seasonality and resourcevariation cannot be assessed in such samples (Plavcan, vanSchaik, and McGraw 2005). In order to test hypotheses thatvariation in fossil Homo might reflect epiphenomenal varia-tion in response to variation in resource abundance or re-sponses to selection and local adaptation, we need targetedstudies that carefully measure these effects and how theymight be recognized in fossil samples.

This does not mean that the study of fossils is futile, how-ever. The samples that we do have provide us with hypothesesand directions for research in living species. For example, theappearance of a shift in size from early to later Homo im-mediately raises the question of why and underscores theimportance of improving samples of fossils and generatingbetter ways of estimating size with the scattered remains thatwe have. At the same time, our improved understanding ofthe causes of variation in living species gives us a much betterframework for interpreting the data that we have in the fossilrecord. For example, we now understand and expect that sizechanges in species may be epigenetic and reflect regional ortemporal variation in resource availability and not necessarilyselection. Our understanding of the causes of dimorphism inprimates argues strongly against inferring particular matingsystems associated with some degree of sexual size dimor-phism and instead suggests that we should understand long-term changes in size dimorphism as indicative of changes inthe selective factors affecting males and females separately.


While on the one hand this perhaps underscores how littlewe actually know about the causes of variation in early Homo,on the other hand it also gives us a clearer understanding ofthe questions we can answer and the types of data necessaryto answer them.

Acknowledgments

I thank Susan Anton and Leslie Aiello for the generous in-vitation to participate in this symposium as well as for theirencouragement, support, and insightful comments. I am es-pecially indebted to all of the conference participants for theircritiques, novel ideas, new perspectives, and challenges to as-sumptions and cherished ideas. I thank one anonymous re-viewer for pointed and helpful comments. This work wassupported by grants from the National Science Foundation,the Leakey Foundation, and the Wenner-Gren Foundation.

References CitedAiello, Leslie C., and Cathy Key. 2002. Energetic consequences of being a

Homo erectus female. American Journal of Human Biology 14:551–565.Albrecht, Gene H. 1980. Latitudinal, taxonomic, sexual, and insular deter-

minants of size variation in pigtail macaques, Macaca nemestrina. Inter-national Journal of Primatology 1:141–152.

Albrecht, Gene H., and Joseph M. A. Miller. 1993. Geographic variation inprimates: a review with implications for interpreting fossils. In Species,species concepts, and primate evolution. William H. Kimbel and LawrenceB. Martin, eds. Pp. 211–237. New York: Plenum.

Altmann, Jeanne, and Susan C. Alberts. 2005. Growth rates in a wild primatepopulation: ecological influences and maternal effects. Behavioral Ecologyand Sociobiology 57:490–501.

Anton, Susan C. 2002. Evolutionary significance of cranial variation in AsianHomo erectus. American Journal of Physical Anthropology 118:301–323.

———. 2003. Natural history of Homo erectus. Yearbook of Physical Anthro-pology 46:126–170.

———. 2012. Early Homo: who, when, and, where. Current Anthropology53(suppl. 6):S278–S298.


Auerbach, Benjamin M., and Christopher B. Ruff. 2004. Human body massestimation: a comparison of “morphometric” and “mechanical” methods.American Journal of Physical Anthropology 125:331–332.

Brennan, E. Jean. 1985. De Brazza’s monkeys (Cercopithecus neglectus) inKenya: census, distribution and conservation. American Journal of Prim-atology 8:269–277.


Cartmill, Matt, and Fred H. Smith. 2009. The human lineage. Hoboken, NJ:Wiley.

Cheverud, James M., Malcolm M. Dow, and Walter Leutenegger. 1985. Thequantitative assessment of phylogenetic constraints in comparative analysis:sexual dimorphism in body weight among primates. Evolution 38:1335–1351.

Clutton-Brock, Timothy H., Paul H. Harvey, and B. Rudder. 1977. Sexualdimorphism, socionomic sex ratio and body weight in primates. Nature269:191–195.

Cope, Dana A., and Michael G. Lacy. 1992. Falsification of a single specieshypothesis using the coefficient of variation: a simulation approach. Amer-ican Journal of Physical Anthropology 89:359–378.


Dunbar, Robin I. M. 1990. Environmental determinants of intraspecific var-

iation in body weight in baboons (Papio spp.). Journal of Zoology (London)220:157–169.

Ely, John, and Jeffrey A. Kurland. 1989. Spatial autocorrelation, phylogeneticconstraints, and the causes of sexual dimorphism in primates. InternationalJournal of Primatology 10:151–171.

Emlen, Stephen T., and Lewis W. Oring. 1977. Ecology, sexual selection, andthe evolution of mating systems. Science 191:215–233.

Fairbairn, Daphne J. 1997. Allometry for sexual size dimorphism: pattern andprocess in the coevolution of body size in males and females. Annual Reviewof Ecology and Systematics 28:659–687.

Fernandez-Duque, Eduardo. 2011. Rensch’s rule, Bergmann’s effect, and adultsexual dimorphism in wild monogamous owl monkeys (Aotus azarai) ofArgentina. American Journal of Physical Anthropology 146:38–48.

Fleagle, John G., Richard F. Kay, and Elwyn L. Simons. 1980. Sexual dimor-phism in early anthropoids. Nature 287:328–330.

Fooden, Jack. 1995. Systematic review of southeast Asian longtail macaques,Macaca fascicularis (Raffles, [1821]). Fieldiana, Zoology, new series, 81.Chicago: Field Museum of Natural History.

Ford, Susan M. 1994. Evolution of sexual dimorphism in body weight inplatyrrhines. American Journal of Primatology 34:221–224.

Gaulin, Steven J. C., and Lee Douglas Sailer. 1984. Sexual dimorphism inweight among primates: the relative impact of allometry and sexual selec-tion. International Journal of Primatology 5:515–535.

Gingerich, Philip D., and Margaret J. Schoeninger. 1979. Patterns of toothsize variability in the dentition of primates. American Journal of PhysicalAnthropology 51:457–466.

Godfrey, Laurie R., Sarah K. Lyon, and Michael R. Sutherland. 1993. Sexualdimorphism in large-bodied primates: the case of the subfossil lemurs.American Journal of Physical Anthropology 90:315–334.

Gordon, Adam D. 2004. Evolution of body size and sexual size dimorphismin the order Primates: Rench’s rule, quantitative genetics, and phylogeneticeffects. PhD dissertation, University of Texas, Austin.

———. 2006a. Scaling of size and dimorphism in primates. 1. Microevolution.International Journal of Primatology 27:27–61.

———. 2006b. Scaling of size and dimorphism in primates. 2. Macroevo-lution. International Journal of Primatology 27:63–105.

Gordon, Adam D., David J. Green, and Brian G. Richmond. 2008. Strongpostcranial size dimorphism in Australopithecus afarensis: results from twonew resampling methods from multivariate data sets with missing data.American Journal of Physical Anthropology 135:311–328.

Greenfield, Leonard O. 1992. Origin of the human canine: a new solution toan old enigma. Yearbook of Physical Anthropology 35:153–185.

Harmon, Elizabeth H. 2006. Size and shape variation in Australopithecus afar-ensis proximal femora. Journal of Human Evolution 51:217–227.

Hayes, Vanessa J., Len Freedman, and Charles E. Oxnard. 1995. The differentialexpression of dental sexual dimorphism in subspecies of Colobus guereza.International Journal of Primatology 16:971–996.

Howells, William W. 1973. Cranial variation in man: a study by multivariateanalysis of patterns of differences among recent human populations. Papersof the Peabody Museum of Archeology and Ethnology, vol. 67. Cambridge,MA: Peabody Museum of Archeology and Ethnology, Harvard University.

Isler, Karin, E. Christopher Kirk, Joseph M. A. Miller, Gene A. Albrecht, BruceR. Gelvin, and Robert D. Martin. 2007. Endocranial volume of primatespecies: scaling analyses using a comprehensive and reliable data set. Journalof Human Evolution 55:967–978.

Jolly, Clifford J. 2001. A proper study for mankind: analogies from the pa-pionin monkeys and their implications for human evolution. Yearbook ofPhysical Anthropology 44:177–204.

Jones, Andrea L., David DeGusta, Scott P. Turner, Christina J. Campbell, andKatharine Milton. 2000. Craniometric variation in a population of mantledhowler monkeys (Alouatta palliata): evidence of size selection in femalesand growth in dentally mature males. American Journal of Physical An-thropology 113:411–434.

Josephson, Steven C., Kenneth E. Juell, and Alan R. Rogers. 1996. Estimatingsexual dimorphism by method-of-moments. American Journal of PhysicalAnthropology 100:191–206.

Kappeler, Peter M. 1991. Patterns of sexual dimorphism in body weight amongprosimian primates. Folia Primatologica 57:132–146.

Kappelman, John. 1996. The evolution of body mass and relative brain sizein fossil Homo. Journal of Human Evolution 30:243–276.

Katzmarzyk, Peter T., and William R. Leonard. 1998. Climatic influences onhuman body size and proportions: ecological adaptations and secular trends.American Journal of Physical Anthropology 106:483–503.


Kay, Richard F., J. Michael Plavcan, Kenneth E. Glander, and Patricia C.Wright. 1988. Sexual selection and canine dimorphism in New World mon-keys. American Journal of Physical Anthropology 77:385–397.

Kelley, Jay, and J. Michael Plavcan. 1998. A simulation test of hominoid speciesnumber at Lufeng, China: implications for the use of the coefficient ofvariation in paleotaxonomy. Journal of Human Evolution 35:577–596.

Kidder, James H., and Arthur C. Durband. 2004. A re-evaluation of the metricdiversity within Homo erectus. Journal of Human Evolution 46:297–313.

Kimbel, William H., and Lucas K. Delezene. 2009. “Lucy” redux: a review ofresearch on Australopithecus afarensis. Yearbook of Physical Anthropology 52:2–48.

Kramer, Andrew, Steven M. Donnelly, James H. Kidder, Stephen D. Ousley,and Stephan M. Olah. 1995. Craniometric variation in large-bodied hom-inoids: testing the single-species hypothesis for Homo habilis. Journal ofHuman Evolution 29:443–462.


Lande, Russell. 1980. Sexual dimorphism, sexual selection, and adaptation inpolygenic characters. Evolution 33:292–305.

Leigh, Steven R. 1992. Patterns of variation in the ontogeny of primate bodysize dimorphism. Journal of Human Evolution 23:27–50.

———. 1995. Socioecology and the ontogeny of sexual size dimorphism inanthropoid primates. American Journal of Physical Anthropology 97:339–356.

Leigh, Steven R., and Gregory E. Blomquist. 2007. Life history. In Primatesin perspective. Christina J. Campbell, Augustin Fuentes, Katherine C.MacKinnon, Melissa Panger, and Simon K. Bearder, eds. Pp. 396–407. NewYork: Oxford University Press.

Leigh, Steven R., and Brian T. Shea. 1996. Ontogeny of body size variationin African apes. American Journal of Physical Anthropology 99:43–65.

Leutenegger, Walter, and James M. Cheverud. 1982. Correlates of sexual di-morphism in primates: ecological and size variables. International Journalof Primatology 3:387–402.

———. 1985. Sexual dimorphism in primates: the effects of size. In Size andscaling in primate biology. William L. Jungers, ed. Pp. 33–50. New York:Plenum.

Leutenegger, Walter, and James T. Kelly. 1977. Relationship of sexual dimor-phism in canine size and body size to social, behavioral and ecologicalcorrelates in anthropoid primates. Primates 18:117–136.

Leutenegger, Walter, and Gabriele Lubach. 1987. Sexual dimorphism, matingsystem, and the effect of phylogeny in De Brazza’s monkey (Cercopithecusneglectus). American Journal of Primatology 13:171–179.

Leutenegger, Walter, and Thomas J. Masterson. 1989a. The ontogeny of sexualdimorphism in the cranium of Bornean orang-utans (Pongo pygmaeus pyg-maeus). 1. Univariate analysis. Zeitschrift fur Morphologie und Anthropologie78:1–14.

———. 1989b. The ontogeny of sexual dimorphism in the cranium ofBornean orang-utans (Pongo pygmaeus pygmaeus). 2. Allometry and het-erochrony. Zeitschrift fur Morphologie und Anthropologie 78:15–24.

Lieberman, Daniel E., Maureen J. Delvin, and Osbjorn M. Pearson. 2001.Articular area responses to mechanical loading: effects of exercise, age, andskeletal location. American Journal of Physical Anthropology 116:266–277.

Lieberman, Daniel E., David R. Pilbeam, and Bernard A. Wood. 1988. Aprobabilistic approach to the problem of sexual dimorphism in Homo ha-bilis: a comparison of KNM-ER 1470 and KNM-ER 1813. Journal of HumanEvolution 17:503–511.

Lindenfors, Patrik. 2002. Sexually antagonistic selection on primate size. Jour-nal of Evolutionary Biology 15:595–607.

Lindenfors, Patrik, and Birgitta S. Tullberg. 1998. Phylogenetic analysis ofprimate size evolution: the consequences of sexual selection. Biological Jour-nal of the Linnean Society 64:413–447.

Lockwood, Charles A. 1999. Sexual dimorphism in the face of Australopithecusafricanus. Journal of Human Evolution 31:537–548.

Lockwood, Charles A., Colin G. Menter, Jacopo Moggi-Cecchi, and Andre W.Keyser. 2007. Extended male growth in a fossil sample hominin species.Science 318:1443–1446.

Lockwood, Charles A., Brian G. Richmond, William L. Jungers, and WilliamH. Kimbel. 1996. Randomization procedures and sexual dimorphism inAustralopithecus afarensis. Journal of Human Evolution 31:537–548.

Lovejoy, C. Owen. 1981. The origin of man. Science 211:341–350.———. 2009. Reexamining human origins in light of Ardipithecus ramidus.

Science 326:74e1–74e8.

Martin, Robert D., Lesley A. Willner, and Andrea Dettling. 1994. The evolutionof sexual size dimorphism in primates. In The differences between the sexes.Roger V. Short and Evan Balaban, eds. Pp. 159–200. Cambridge: CambridgeUniversity Press.

McHenry, Henry M. 1992. Body size and proportions in early hominids.American Journal of Physical Anthropology 87:407–431.


Miller, Joseph M. A. 2000. Craniofacial variation in Homo habilis: an analysisof the evidence for multiple species. American Journal of Physical Anthro-pology 112:103–128.

Mitani, John, Julie Gros-Louis, and Andrew F. Richards. 1996. Sexual di-morphism, the operational sex ratio, and the intensity of male competitionin polygynous primates. American Naturalist 147:966–980.

Moore, Jim. 1996. Savanna chimpanzees, referential models and the last com-mon ancestor. In Great ape societies. William C. McGrew, Linda F. Marchant,and Toshisada Nichida, eds. Pp. 275–292. Cambridge: Cambridge UniversityPress.

Plavcan, J. Michael. 1993. Catarrhine dental variability and species recognitionin fossils. In Species, species concepts, and primate evolution. William H.Kimbel and Lawrence B. Martin, eds. Pp. 239–263. New York: Plenum.

———. 1994. A comparison of four simple methods for estimating sexualdimorphism in fossils. American Journal of Physical Anthropology 94:465–476.

———. 2000a. Inferring social behavior from sexual dimorphism in the fossilrecord. Journal of Human Evolution 39:327–344.

———. 2000b. Variance dimorphism in primates. American Journal of PhysicalAnthropology 30(suppl.):251.

———. 2001. Sexual dimorphism in primate evolution. Yearbook of PhysicalAnthropology 44:25–53.

———. 2002. Taxonomic variation in the patterns of craniofacial dimorphismin primates. Journal of Human Evolution 42:579–608.

———. 2003. Scaling relationships between craniofacial sexual dimorphismand body mass dimorphism in primates: implications for the fossil record.American Journal of Physical Anthropology 120:38–60.

———. 2004a. Evidence for early anthropoid social behavior. In Anthropoidorigins: new visions. Callum Ross and Richard F. Kay, eds. Pp. 383–412.New York: Kluwer Academic/Plenum.

———. 2004b. Sexual selection, measures of sexual selection, and dimorphismin primates. In Sexual selection in primates: new and comparative perspectives.Peter Kappeler and Carel P. van Schaik, eds. Pp. 230–252. Cambridge:Cambridge University Press.

———. 2011. Understanding dimorphism as a function of changes in maleand female traits. Evolutionary Anthropology 20:143–155.

———. 2012. Sexual size dimorphism, canine dimorphism, and male-malecompetition in primates: where do humans fit in? Human Nature 23:45–67.

Plavcan, J. Michael, and Dana Cope. 2002. Metric variation and species rec-ognition in the fossil record. Evolutionary Anthropology 10:204–222.

Plavcan, J. Michael, and Richard F. Kay. 1988. Sexual dimorphism and dentalvariability in platyrrhine primates. International Journal of Primatology 9:169–178.

Plavcan, J. Michael, and Carel P. van Schaik. 1992. Intrasexual competitionand canine dimorphism in anthropoid primates. American Journal of Phys-ical Anthropology 87:461–477.

———. 1997a. Interpreting hominid behavior on the basis of sexual dimor-phism. Journal of Human Evolution 32:345–374.

———. 1997b. Intrasexual competition and body weight dimorphism in an-thropoid primates. American Journal of Physical Anthropology 103:37–68.

Plavcan, J. Michael, Carel P. van Schaik, and Peter M. Kappeler. 1995. Com-petition, coalitions and canine size in primates. Journal of Human Evolution28:245–276.

Plavcan, J. Michael, Carel P. van Schaik, and W. Scott McGraw. 2005. Sea-sonality, social organization, and sexual dimorphism in primates. In Sea-sonality in primate evolution. Diane K. Brockman and Carel P. van Schaik,eds. Pp. 401–441. Cambridge: Cambridge University Press.


Ralls, Kathrine. 1976. Mammals in which females are larger than males. Quar-terly Review of Biology 51:245–276.

Reno, Philip L., Melanie A. McCollum, Richard S. Meindl, and C. OwenLovejoy. 2010. An enlarged postcranial sample confirms Australopithecus


afarensis dimorphism was similar to modern humans. Philosophical Trans-actions of the Royal Society B 365:3355–3363.

Reno, Philip L., Richard S. Meindl, Melanie A. McCollum, and C. OwenLovejoy. 2003. Sexual dimorphism in Australopithecus afarensis was similarto that of modern humans. Proceedings of the National Academy of Sciencesof the USA 100:9404–9409.

Rightmire, G. Philip, Adam Van Arsdale, and David Lordkipanidze. 2008.Variation in the mandibles from Dmanisi, Georgia. Journal of Human Evo-lution 54:904–908.

Ruff, Christopher B. 2002. Variation in human body size and shape. AnnualReview of Anthropology 31:211–232.


Scott, Jeremiah E., and Laura K. Stroik. 2006. Bootstrap tests of significanceand the case for humanlike skeletal-size dimorphism in Australopithecusafarensis. Journal of Human Evolution 51:422–428.

Selander, Robert K. 1972. Sexual selection and dimorphism in birds. In Sexualselection and the descent of man, 1871–1971. Bernard Campbell, ed. Pp. 180–230. Chicago: Aldine.

Shea, Brian T., Steven R. Leigh, and Colin P. Groves. 1993. Multivariate cra-niometric variation in chimpanzees: implications for species identificationin paleoanthropology. In Species, species concepts, and primate evolution.William H. Kimbel and Lawrence B. Martin, eds. Pp. 265–296. New York:Plenum.

Skinner, Matthew, Adam D. Gordon, and Nicole J. Collard. 2006. Mandibularsize and shape variation in the hominins at Dmanisi, Republic of Georgia.Journal of Human Evolution 51:36–49.

Smith, Jennifer E., Eli M. Swanson, Daphna Reed, and Kay E. Holekamp.2012. Evolution of cooperation among mammalian carnivores and its rel-evance to hominid evolution. Current Anthropology 53(suppl. 6):S436–S452.

Smith, Richard J. 1996. Biology and body size in human evolution: statisticalinference misapplied. Current Anthropology 37:451–481.

Smith, Richard J., and William L. Jungers. 1997. Body mass in comparativeprimatology. Journal of Human Evolution 32:523–559.

Sokal, Robert R., and Carlos A. Braumann. 1980. Significance tests for co-efficients of variation and variability profiles. Systematic Zoology 29:50–63.

Spoor, Fred, Meave G. Leakey, Patrick N. Gathogo, Frank H. Brown, Susan

C. Anton, Ian McDougall, Christopher Kiare, Fredrick K. Manthi, andLouise N. Leakey. 2007. Implications of new early Homo fossils from Ileret,east of Lake Turkana, Kenya. Nature 448:688–691.

Sterck, Elisabeth H. M., David P. Watts, and Carel P. van Schaik. 1997. Theevolution of female social relationships in nonhuman primates. BehavioralEcology and Sociobiology 41:291–309.

Trinkaus, Erik, Steven E. Churchill, and Christopher B. Ruff. 1994. Postcranialrobusticity in Homo. 2. Humeral bilateral asymmetry and bone plasticity.American Journal of Physical Anthropology 93:1–34.

Turner, Trudy R., Fred Anapol, and Clifford J. Jolly. 1997. Growth, devel-opment, and sexual dimorphism in vervet monkeys (Cercopithecus aethiops)at four sites in Kenya. American Journal of Physical Anthropology 103:19–35.

Utami Atmoko, Suci, and Jan A. R. A. M. van Hooff. 2004. Alternative malereproductive strategies: male bimaturism in orangutans. In Sexual selectionin primates: new and comparative perspectives. Peter M. Kappeler and CarelP. van Schaik, eds. Pp. 196–207. Cambridge: Cambridge University Press.

Wang, Zimian, Moonseong Heo, Robert C. Lee, Donald P. Kotler, Robert T.Withers, and Steven D. Heymsfield. 2001. Muscularity in adult humans:proportion of adipose tissue-free body mass as skeletal muscle. AmericanJournal of Human Biology 13:612–619.

Watanabe, Kunio. 1981. Variation in group composition and population den-sity of the two sympatric Mentawian leaf-monkeys. Primates 22:145–160.

Wells, Jonathan C. K. 2007. Sexual dimorphism of body fat composition. BestPractice in Research: Clinical Endocrinology and Metabolism 21:415–430.


Wood, Bernard. 1993. Early Homo: how many species? In Species, speciesconcepts, and primate evolution. William H. Kimbel and Lawrence B. Martin,eds. Pp. 485–522. New York: Plenum.

Wood, Bernard, and Daniel Lieberman. 2001. Craniodental variation in Par-anthropus boisei: a developmental and functional perspective. AmericanJournal of Physical Anthropology 116:13–25.

Yablokov, Alexey V. 1974. Variability in mammals. Springfield, VA: NationalTechnical Information Service, U.S. Department of Commerce.



Male Life History, Reproductive Effort, andthe Evolution of the Genus Homo

New Directions and Perspectives

by Richard G. Bribiescas, Peter T. Ellison, and Peter B. Gray

The evolution of male life history traits was central to the emergence of the genus Homo. Compared with earlierhominins, changes in the behavioral and physiological mechanics of growth, survivorship, reproductive effort, andsenescence all likely contributed to shifts in how males contributed to the evolution of our genus. For example, therange of paternal investment in modern Homo sapiens is unusual compared with most mammals and primates, allbut certainly contributing to the evolution of the suite of life history traits that define Homo, including high fertility,large bodies, altricial offspring, large brains, and long lives. Moreover, the extensive range of phenotypic and behavioralvariation in somatic and behavioral reflections of male reproductive effort across modern H. sapiens is especiallynoteworthy. We propose that selection for a broad range of variation in traits reflective of male reproductive effortwas important to the evolution of Homo. We examine factors that contribute to this variation, proposing thatselection for paternal and somatic investment plasticity across the entire male life span was important for theevolution of Homo. Potential research strategies and directions for new research for exploring these issues in thefossil record are also discussed.

The evolution of the genus Homo was marked by the emer-gence of life history traits that were distinct from previoushominins. Larger brains and bodies, more gracile and gen-eralized dentition, and a general decrease in cranial robusticityindicated a novel evolutionary path that lead to our ownspecies (Wood and Collard 1999). While some early Homotraits were not identical to modern humans, such as accel-erated dental growth (Dean et al. 2001; Schwartz 2012) andperhaps slower infant brain development (Coqueugniot et al.2004), neonatal brain size, as extrapolated from obstetric mea-surements of a Homo erectus female pelvis, appears similar tomodern Homo sapiens (Simpson et al. 2008), intimating fetalgrowth and maternal energetic investment that was not verydissimilar from ourselves. Enhanced somatic growth throughincreases in the pace of development, life span, or both com-pared with earlier hominins were necessary to achieve theyoung adult size exhibited by the most complete H. erectusspecimen, KNM-WT 15000 (Brown et al. 1985), although

Richard G. Bribiescas is Professor in the Department ofAnthropology at Yale University (New Haven, Connecticut 06511,U.S.A. [[email protected]]). Peter T. Ellison is Professorin the Department of Human Evolutionary Biology at HarvardUniversity (Cambridge, Massachusetts 02138, U.S.A.). Peter B. Grayis Assistant Professor in the Department of Anthropology at theUniversity of Nevada, Las Vegas (Las Vegas, Nevada 89154, U.S.A.).This paper was submitted 12 XII 11, accepted 25 VI 12, andelectronically published 27 IX 12.

estimates of adult size for this individual are still the subjectof debate (Graves et al. 2010), and considerable size variationis now known across the species (Anton and Snodgrass 2012).Moreover, the energetic demands of a larger body size, largerbrains, and greater investment in altricial offspring surely re-sulted in the evolution of shifts in the way Homo stored,allocated, and managed energetic resources compared withearlier hominins (Aiello and Wells 2002).

Relative changes in juvenile and adult mortality due to theattenuation of predation, conspecific threats, food stress, orecological stochasticity all likely shaped Homo life historytraits (Michod 1979; Stearns and Koella 1986). Greater avail-ability and optimization of energy resource allocation alsowas needed to support greater female fertility (Aiello andWells 2002; Knott 2001). What factors changed comparedwith earlier hominins that resulted in selection for these lifehistory traits, ultimately leading to emergence and prolifer-ation of Homo? We propose that the variability and plasticityof male life histories was a potent force in the developmentof the defining life history traits in Homo.

The high metabolic costs of female reproduction as well asparental care requirements of Homo offspring (Aiello and Key2002) suggest that supplementary support provided a selectiveadvantage to produce and raise offspring to reproductive ma-turity compared with mothers who did so independently.Aside from child self support (Kramer 2005), kin, alloparents,and fathers are the sole sources of extramaternal care. Indeed,compelling arguments highlight the role of kin and nonkin


Bribiescas, Ellison, and Gray Male Life History and the Emergence of Homo S425

support (Bentley and Mace 2009; Hill et al. 2011). Comparedwith earlier hominins, Homo males had to have exerted agreater positive effect on offspring and mates through theaccentuation of food availability and quality as well as de-creasing conspecific and predation risk and immunologicalchallenges, just to name a few factors.

This is not a new proposal. Gettler (2010) argues that theevolution of male paternal investment in our genus, mostlyin the form of offspring carrying, allowed for the liberationof energetic resources in females that was thereby allocatedtoward shorter interbirth intervals and higher female fertility.Key and Aiello (2000) suggest that paternal investment mayhave evolved in response to a “prisoner’s dilemma,” proposingthat when male reproductive costs are less than female re-productive costs, males should cooperate with females evenwhen females do not reciprocate. Serving to complementthese perspectives, our approach is more male-centric, at-tending to the somatic and behavioral trade-offs within malelife histories that may have primed these positive effects onoffspring and mates. Because male investment in offspringand mates is not without cost and requires a reallocation ofresources between the somatic and behavioral demands ofgrowth (Bribiescas 2001; Ellison 2003), the physiology centralto male life histories had to evolve as well. Here we introducethese male trade-offs, unique aspects of male reproductiveinvestment in Homo, and discuss possible ways in which theseideas might be applied and tested within the hominin fossilrecord.

Reproductive Effort and Male Trade-Offs

Reproductive effort is defined as the allocation of limitingresources, principally energy and time, to reproduction. Whileelegant in its simplicity, this definition requires elaborationin order to be useful, because in actuality life itself can bedefined as the capture of energy in the service of the pro-duction of new organisms, and thus all of life’s processes areultimately reproductive effort. Survival only contributes tofitness to the extent that it enhances reproduction. The con-cept of reproductive effort as it is deployed in life historytheory is usually narrowed to the allocation of resources toreproduction at the expense of other, nonoverlapping cate-gories of allocation, such as survival. A conventional divisionof allocation domains into growth, reproduction, and main-tenance was introduced by Gadgil and Bossert (1970) andwidely adopted by others as a heuristic framework. The im-portant point that all such frameworks share, however, is thepremise that investment of limiting resources in one categorycomes at the expense of investment in other categories. This“trade-off” principle represents a constraint that allows formodeling optimal allocations in terms of different end points.For evolutionary analysis, the most relevant end point is Dar-winian fitness, and based on certain sets of assumptions (e.g.,stable population theory), the allocation of resources among

competing categories that maximizes fitness can be deter-mined. When analyzed over the entire life span of an organ-ism, such a pattern of resource allocation is often referred toas a life history or reproductive “strategy.”

The key concept in life history strategies is “trade-off.”Important trade-offs occur not only between competing cat-egories of allocation at any point in time, which might crudelybe reduced to the trade-off between investing in offspringversus investing in self, but also between reproduction nowand reproduction later, or investing in offspring already bornversus new offspring. Finer distinctions in reproductive ef-fort—such as “mating effort,” “parenting effort,” “gestationeffort,” “lactation effort,” and so forth—are often useful indistinguishing variations in reproductive strategies (Stearns1989). One can also distinguish between “behavioral effort”and “physiological effort.” The field of behavioral ecology hasflourished by analyzing behavioral strategies in terms of theireffects on fitness. But the physiological modulation of energyallocation to all categories, including reproduction, is alsofundamental. While behavioral strategies are most effective inresponding to short-term conditions and the shifting land-scape of conspecific behavior, physiological strategies are usu-ally tuned to the longer-term features of an organism’s ecologyand the more predictable unfolding of its life stages.

Male versus Female Reproductive Effort

Trivers (1972), drawing on Bateman (1948), noted that theasymmetry in male and female reproductive strategies inmany animal species can be traced to the asymmetry in gametesizes, itself a reflection of asymmetrical investment of re-sources in individual gametes. Because females provide most(in fact, all, in most cases) of the metabolic resources thatare required in very early embryogenesis, they produce fewergametes from a given supply of energy, males many more.From this fundamental asymmetry arises a general dimor-phism in reproductive strategy whereby females invest moreeffort in the somatic growth of offspring (“parenting effort”in Trivers’s [1972] terms) and males invest more effort inopportunities to father those offspring (“mating effort”).

Female parenting effort is generally considered to be heavilyweighted toward physiological investment in the offspringsoma through gestation, egg laying, lactation, and so on. Es-timates of female reproductive effort are therefore often madefrom measurements of total clutch mass, birth weight, orlactation performance in relation to maternal body weight(Calow 1979). Male mating effort is generally understood asprimarily behavioral. Estimates of male reproductive effortare therefore often made by measuring time or energy spentin territory defense, mate guarding, competitive interactionswith other males, and so forth. Alternatively, reproductiveeffort in either sex can be estimated in terms of the costimposed on other allocation categories, such as survival andmaintenance, through measures such as body weight lost or


mortality suffered during the mating season (Clutton-Brock1984).

The dichotomization of male and female reproductive ef-fort into physiological and behavioral is not absolute, ofcourse, although it may be a useful first approximation. Fe-males invest considerable behavioral effort in reproduction,including such obvious behaviors as incubation of eggs andprovisioning of young. Males in turn make physiological in-vestments in the somatic tissues and metabolic processes thatunderpin their competitive ability (Bribiescas 1996).

The importance of physiological investment in somatic tis-sues in males is evident, for example, in sex differences inprocesses that contribute to the adolescent transition to sexualmaturity. While females must deal with the energetic andpelvic skeletal demands of passing a large-brained infant—hence the relationship between skeletal growth, the produc-tion of ovarian hormones, and age of menarche (Ellison1982)—male sexual maturation is less constrained by the neg-ligible metabolic demands of spermatogenesis and instead iscontingent on the need to increase testosterone levels andgrow sexually dimorphic muscle tissue as well as its sup-porting skeletal structure (Bribiescas 2001).

But neither is the dichotomization of male and female re-productive effort into mating and parenting effort absolute.Females of many species compete behaviorally and physio-logically with other females to attract the highest-qualitymales, and males of many species participate in the provi-sioning, carrying, and direct care of offspring (Gross 1996).In species that have been selected for long-term mate fidelity,male and female reproductive success become more closelyaligned, and levels and patterns of reproductive effort con-verge, although phylogenetic constraints on basic patterns ofreproductive physiology make it difficult to erase all asym-metries.

Human Male Reproductive Effort

The aspect of male reproductive effort most often recognizedand studied is behavioral, expressed in competition with othermales for reproductive access to females. The investment isnot only behavioral, however. Physiological and/or somaticinvestment is usually involved as well. Investment in bodymass, particularly muscle mass, is substantial in Gorilla, Pongo,and Pan (Leonard and Robertson 1997). Pan also makes aconsiderable metabolic investment in gamete production, pre-sumably as a consequence of sperm competition (Harcourtet al. 1981). Males in these three genera invest little if anyparenting effort. Male siamangs, in contrast, invest less inmating effort, either behaviorally or physiologically, than thegreat apes but much more in parenting effort (Lappan 2008).Although males do not incur the metabolic costs of femalereproduction such as pregnancy and lactation, sexually di-morphic mass and body composition results in investmentin tissue that is reflective of reproductive effort in males (Bri-biescas 2001). Indeed, over a lifetime, the cumulative meta-

bolic cost of sexual dimorphism in primate males is com-parable to the energetic costs of several offspring born tofemales (Key and Ross 1999).

Dimorphic somatic strategies between male and female hu-mans reflect differences in physiological reproductive effort.Human males invest a substantial amount in the productionand maintenance of a larger overall body size and in musclemass in particular, presumably a reflection of an evolutionaryheritage of male mating competition (Puts 2010). The degreeof somatic dimorphism is not as great, however, as in othergreat apes or as in other hominin species. In concert withthis relative de-emphasis on somatic investment, humanmales demonstrate much more in the way of parenting effortthan other great apes, suggesting a legacy of relatively greaterlong-term mate fidelity. The observed pattern among extantprimates suggests that sexual dimorphism in body size in thefossil record can be interpreted as a reflection of the relativeimportance of mating and parenting effort in the reproductivestrategies of extinct hominin species (McHenry 1994).

Reproductive effort in the form of paternal investment isespecially germane to the evolution of Homo. Although sexualdimorphism in early Homo remains unclear, gracile mor-phology and lower sexual dimorphism in modern humanscompared with australopithecines and paranthropines suggesta greater role for paternal investment in early Homo, althoughcaution is merited because of the large degree of error inestimating sexual dimorphism in Homo and other hominins(Anton and Snodgrass 2012). However, paternal investmentis not unique among primates, with about 40% of all speciesexhibiting some form of paternal investment (Kleiman andMalcolm 1981). Therefore, paternal investment alone prob-ably does not account for the emergence of Homo life historytraits. The unique nature of paternal investment in modernhumans is evident not only in its presence but also in itsvariability, both in intensity of time and energy investmentand form, in terms of offspring care, provisioning, and othersocial benefits (Gray and Anderson 2010; Hewlett 1992). Tounderstand paternal investment in early Homo, we turn tomodern humans since our species is a more appropriate com-parative model than are other extant apes that exhibit littleor no paternal investment and greater degrees of sexual di-morphism.

Paternal Investment in Homo

A meta-analysis including 22 societies found that the deathof a father had a measurable effect on child mortality in onlyseven societies (Sear and Mace 2008), leading some to con-clude that paternal investment is of little importance. How-ever, this ignores influences of paternal care on other aspectsof maternal and child morbidity and mortality. In particular,fathers may exert influences on their offspring’s social successand in turn on their reproductive prospects, something thatis not captured in the preceding analyses. Further, a majorinfluence of paternal care appears to be in shortening human


birth intervals, helping women have more children than theyotherwise could (Marlowe 2010).

Forms of paternal care can be distinguished in several ways.Kleiman and Malcolm (1981) distinguished between directand indirect care. Direct care includes holding, carrying, andmaintaining proximity to offspring, whereas indirect care en-compasses more distal forms such as protection and resourcedefense. Forms of direct care may require steeper trade-offswith male mating effort than indirect care, and indeed, muchof the discussion concerning whether male behavior repre-sents mating or parenting effort depends on how one inter-prets resource acquisition efforts (e.g., do men hunt in orderto provision family members or to gain status and repro-ductive benefits; Hawkes 1991).

Within contemporary hunter-gatherer societies, the mostsalient model for early Homo socioecology, most male re-source provisioning focuses on meat and honey. Considerabledebate has focused on whether big game hunting by foragermen represents mating effort or parenting effort. Becauselarge game are widely shared in a forager camp and successfulhunters are often rewarded with reproductive benefits suchas additional wives or extramarital affairs (Kaplan and Hill1985), some scholars suggest that male hunting representsmating effort (Hawkes 1991). However, fathers often targetresources such as smaller game and honey that are prefer-entially shared with their families; in some cases a larger shareof big game animals may reach a successful hunter’s family;and by several measures biological fathers tend to providemore care of their offspring than do stepfathers (Gurven andHill 2009). Accordingly, among foragers, male care likely rep-resents a mix of mating and parenting effort and a mix thatcan shift depending on context (e.g., Marlowe [1999] foundless paternal care among Hadza living in larger camps).

Paternity and paternity certainty are associated with pa-ternal care, both between and within societies. Theory sug-gests that males seek to channel their reproductive effort inways that maximize their own fitness; accordingly, men areexpected to preferentially bias their care toward their ownoffspring. Across societies, men provide more indirect care inthose societies where they have higher paternity certainty(Hartung 1985). More specifically, societies with matrilinealinheritance tend to have lower paternity certainty becausefathers are often less involved in day-to-day family life,whether because of engagement in long-distance warfare orsubsistence activities drawing them away from mothers andchildren. Conversely, societies with patrilineal inheritance,where fathers pass resources to their children, are character-ized by higher paternity certainty in part because a father’srelatives can aid observation of his wife’s fidelity. Geneticevidence of patrilineality has been reported in Neanderthalsas has isotopic evidence for smaller and more localized homeranges, which can be cautiously interpreted as patrilinealityin australopiths and paranthropines (Copeland et al. 2011;Lalueza-Fox et al. 2011). Similar molecular and isotopic meth-ods could be deployed in fossil Homo specimens in the future.

Within societies, various studies have found that biologicalfathers provide more direct or indirect care of their offspringcompared with stepfathers. This has held for men’s invest-ments in their children’s college educational expenses (An-derson, Kaplan, and Lancaster 1999), men’s financial expen-ditures on their high school aged children (Anderson et al.1999), the time spent by Hadza forager fathers in direct careof children (Marlowe 1999), and the time spent by rural Trin-idad fathers with children (Flinn 1988). Other lines of evi-dence are consistent with differential attachment and care ofoffspring depending on paternity. Most prominently, Daly andWilson (1999) suggest that the greater risk of child abuse andinfanticide observed among stepchildren compared with bi-ological children in Canada, the United States, and elsewhereis an outcome of differential male concern over these children.Biological fathers are less likely to inflict such costly behavioron their children than are stepfathers.

Mothers are situated at the intersection between male careand offspring. The nature of a man’s relationship to an off-spring’s mother thus serves as an important factor associatedwith paternal care. More affiliative husband-wife relationshipsare associated with greater direct care across populations(Whiting and Whiting 1975). They are also associated withgreater paternal care within societies such as the United States(Parke 1996). Especially in the case of caring for young chil-dren, mothers may serve as “gatekeepers,” playing an im-portant role determining who has access to their offspring,and this may include paternal access. This concept applies topostdivorce paternal care in the United States: poorer-qualityrelationships between a custodial mother and a child’s fatherare associated with lower paternal care. However, these re-lationships matter less as children, especially sons, grow older,and in societies where fathers retain child custody in the eventof divorce.

Mothers structure male reproductive effort in another way:males may provide care to mothers and their offspring inorder to garner mating opportunities with mothers. The con-cept that male care, in this context, represents mating effortrather than parenting effort was initially highlighted throughfield research with baboons (Smuts and Gubernick 1992). Thesame concept may apply to human male care. The direct andindirect care that stepfathers provide to stepchildren can beconceptualized as mating effort—care designed to maintainreproductive access to the children’s mother. Given thatmothers also value traits in potential mates associated withlong-term emotional stability and capacity for providing in-direct care (Buss 1989), even within long-term reproductiverelationships, men’s care devoted to mothers and their off-spring can be conceived in part as mating effort.

Paternal investment has primarily been addressed throughinvestigations of between- and within-societal variation inpatterns of direct and indirect paternal care. Paternal carevaries according to a host of factors: cultural transmission,mode of subsistence, marital system, rates of between-societalaggression, sexual division of labor, demographic patterns,


Figure 1. Summary of the observed diversity of male direct andindirect reproductive investment behaviors in modern humans.

availability of allomothers, paternity and paternity certainty,and the nature of a man’s relationship to a child’s mother.This diverse array of factors highlights the relevance of so-cioecological context to the broad expression of human pa-ternal care, suggesting that male reproductive effort can beadjusted in facultative ways likely conducive to men’s repro-ductive success. Such facultative adjustment indicates signif-icant plasticity in Homo male reproductive strategies com-pared with other great apes.

Plasticity in Male Reproductive Effort

Plasticity is the range of variation and responsiveness in aphenotypic or behavioral trait in response to an environ-mental challenge. The obvious advantage of plasticity is theability to adjust appropriately to environmental stochasticityin a manner that maintains or increases an organism’s fitness.An organism with infinite plasticity would be tremendouslysuccessful. However, plasticity has costs resulting in a rangeof phenotypic and behavioral variation that is managed byproximate mechanisms that are subject to natural selection.In essence, the range of trait plasticity is itself an importantadaptation (Pigliucci 2001). Compared with other great apes,phenotypic and behavioral plasticity are among the most im-portant traits in Homo as demonstrated by the geographicexpansion out of Africa into novel ecological settings (Fin-layson 2005). The broad geographic and ecological range ofHomo signals an extraordinary ability to adjust to environ-mental and social variability. While some reflections of re-productive effort, such as sexual dimorphism, inextricablycovary with female reproductive effort (e.g., body size), phe-notypic and behavioral plasticity reflecting adjustments to re-productive strategies are evident in many male mammals,including humans. Indeed, because body size sexual dimor-phism and its accompanying metabolic cost differences arelikely to be smaller than initially thought in early Homo(Graves et al. 2010), more plastic aspects of male reproductiveeffort, such as the extent and form of paternal care, probablytook on greater importance.

Male plasticity provides the ability to respond to an ever-changing environmental and social landscape. However, as-sociated costs include the maintenance of detection mecha-nisms of change, possible misinformation from cues, and theneed to reallocate, refurbish, and rebuild somatic resourcessuch as muscle and fat more frequently in response to en-vironmental shifts (Relyea 2002). Arguably the most impor-tant cost of male reproductive effort plasticity is the misas-sessment of paternity certainty and cuckoldry, a hurdle thatimpinges on the evolution of paternal care and mate invest-ment. Accurately determining paternity is challenging becauseof internal fertilization, with numerous behavioral strategiesevolving such as mate guarding, sperm plugs, the developmentof affiliations with mothers, and infanticide, just to name afew (Busse 1985). Males would either have had to decreasethe probability of cuckoldry or adjust their mate and paternal

investment in association with paternity uncertainty risk. Inaddition to the anticuckoldry strategies deployed by otherprimates, Homo males evolved a unique suite of abilities toadjust their investment in more subtle ways in response topaternity uncertainty conditions. This is evident in the di-versity of paternal investment behaviors under varied con-ditions of paternity certainty in modern Homo (fig. 1).

Unlike other primates, even those who invest in paternalinvestment, modern Homo sapiens males exhibit significantpopulation variation in direct care of offspring. Hunter-gath-erer fathers tend to spend more time in proximity and directlycaring for their young offspring than men living in small-scale horticultural, agricultural, and especially pastoralist so-cieties (Marlowe 2000a). Several quantitative studies of for-ager paternal care suggest that fathers spend around 5% ofwaking hours in direct care of infants and toddlers (Fouts2008). The Aka, where fathers spend around 20% of their dayhours in camp holding children, is an outlier among foragersbut is likely explained by husband-wife subsistence activitiescentered on net hunting together (Hewlett 1991). Across so-cieties including hunter-gatherers, horticulturalists, agricul-turalists, and pastoralists, several factors underlie this varia-tion in direct paternal care across subsistence mode. Thesexual division of labor in hunter-gatherers may allow fathersand mothers to spend more time near each other in relaxedways, fostering paternal care. Furthermore, hunter-gathererstend to have lower rates of polygyny and often have lowerrates of between-group aggression, both factors that draw menaway from direct care in favor of investment in male-malerelationships and additional mates (Gray and Anderson 2010;Marlowe 2000a). These contexts that are similar to early Homosocioecology suggest that wider socioecological factors shap-


Table 1. Occurrence of significant paternal investment inapes under various conditions of paternity certainty

Genus No paternityPossiblepaternity

Multiplepaternity

Likelypaternity

Homo X X X XPan 0Gorilla 0Pongo 0Hylobates 0 X

Note. “Significant” is defined as exhibiting enough direct offspring careand/or provisioning to be indicative of a defining trait in a species. Xsindicate significant occurrence of paternal investment. Zeros indicatenonhuman multiple paternity; this is culturally defined and cannot beassessed outside of modern Homo.

ing subsistence, between-group aggression, and marriage pat-terns also influence the specifics of direct paternal care.

Men also exhibit population variation in indirect care ofoffspring. Highlighting a role for resource provisioning (e.g.,food, money), indirect care also varies across subsistencemode (Marlowe 2000a). Men in pastoralist societies tend toprovide the most resource provisioning, followed by men inpastoralist and hunter-gatherer societies, with men in horti-cultural societies providing the least. Here, the livestock suchas cattle and camels provided by a father plays importantroles in subsistence (e.g., meat or blood from animals) andin his offspring’s—and often sons in particular—reproductiveprospects. The livestock a father provides as bridewealth tohis male relatives may make the difference between their abil-ity to marry or not. But what of paternal care when fatherhoodis accepted by the male? Paternal care exhibits tremendousvariation, suggesting that it is not obligatory but rather bestviewed as the product of adaptive male plasticity in repro-ductive effort interacting with female (e.g., mate) and off-spring (e.g., parent-offspring conflict) life history agendas inspecific socioecological contexts.

Across larger-scale societies, including contemporarynation-states, direct paternal care varies considerably. Roop-narine (cited in Gray and Anderson 2010) has summarizedmuch of this quantitative variation, finding that fathers inJapan spend about 20 minutes daily with their children,whereas fathers in India spend around 3–5 hours daily withtheir young children. Many fathers spend no time engagedin direct care. To try understanding the variation in directpaternal care across this larger-scale swath of societies, in-cluding the United States and European countries today, var-ious factors are likely at play. The sexual division of laborvaries, with some cases favoring male focus on providingindirect care rather than direct care, as has been the case forurban Japan (Schwalb et al. 2010). As more and more womenhave entered the global competitive labor market in recentdecades, in the United States and elsewhere, fertility has de-clined, and men have increased the time devoted to directpaternal care within a changing sexual division of labor (Gau-thier, Smeeding, and Furstenberg 2004). The availability ofallomothers such as grandmothers or siblings affects directpaternal care. Where such caregivers are available, fathers mayprovide less direct paternal care; indeed, among the Hadza,Bofi foragers, and Khasi of northern India, male care is neg-atively associated with grandmaternal involvement (Fouts2008). Further, the availability of mates may influence directpaternal care. In demographic contexts with female-biasedsex ratios or high variance in male mate value, males mayallocate their reproductive effort toward mating effort ratherthan direct child care.

Another intriguing example of the extreme malleability ofpaternal organization in Homo is the possibility of partitioningpaternity between different males. The high value of paternalinvestment, long lives, and the importance of navigating asocial landscape that is shaped by complex culture, language,

and ecology seems to have resulted in the potential for theevolution of “contingency strategies” in response to the riskof nonpaternity. Partible paternity may be an example of sucha contingency strategy. In some societies, different men areacknowledged to be the father of the same child. Walker, Flinn,and Hill (2010) have discussed the role of partible paternityand the possible fitness benefits in several native South Amer-ican populations. Among many of these populations, a com-munity consensus on paternity is determined in a number ofways that involve belief systems of how children are conceived.The result is a partitioning of paternal investment in the faceof high levels of paternity uncertainty. It is impossible todetermine whether such a social arrangement had any impacton the evolution of our genus, but it does introduce thepossibility of paternity becoming a social currency that canbe exchanged and bartered for benefits such as greater male-male affiliation and future mate access. It also illustrates therange of plasticity that can reconcile paternity uncertainty andthe care and provisioning needs of high-maintenance altricialoffspring (table 1).

Older Fathers: A Unique Developmentin Homo?

The potential for significant fitness later in life is an importantlife history trait in modern humans (Bribiescas 2006, 2010).Although spermatogenesis and the capacity to fertilize ova issomewhat compromised compared with younger men (de LaRochebrochard et al. 2006), reproductive hormone functionremains largely intact (Bribiescas 2005; Ellison et al. 2002),with cross cultural demographic assessments indicating sig-nificant male fertility after the age of 50 (Tuljapurkar, Pules-ton, and Gurven 2007). Fitness at later life obviously requiresan increase in longevity and the involvement of younger, pre-menopausal females. The evolution of longevity in Homo hascentered on the selection for genes that favor larger bodiesand longer lives. However, what selection factors would haveallowed and cultivated fertility at older ages in males?

There is no evidence of early Homo cultivating or raisinganimals or crops or any other resource that could be accu-


Figure 2. Fertility distributions (in age-specific fertility [ASF] rates as a fraction of total fertility [TF] rate) for women (dashed line)and men (solid black line) for the hunter-gatherer Dobe !Kung of Botswana, the forest-living Ache, the Amazonian forager-horticulturalist Yanomamo, the Bolivian forager-horticulturalists Tsimane, agricultural Gambian villagers, and modern Canadians.The shaded area represents realized male fertility after the age of last female reproduction (Tuljapurkar, Puleston, and Gurven 2007).A color version of this figure is available in the online edition of Current Anthropology.

mulated and sequestered in the same manner that supportsthe accumulation of wealth and polygyny in modern Homosapiens (Cronk 1991). Therefore, if males were fathering chil-dren at older ages, it was the result of some other resourceor ability that benefited female fertility and/or child survi-vorship. One possibility is the leveraging of age-derived skilland experience. Among modern forager groups, older mentend to have higher caloric returns from hunting comparedwith younger men who are stronger or more physically fit,suggesting that skill and experience can supersede the physicalability and strength of younger men (Walker and Hill 2003;Walker et al. 2002). Older men can also provide child carethat is especially valuable for extremely altricial offspring.

Leveraging age, experience, and skill for greater fitnessshould result in greater mating opportunities and significantfertility in older men. However, the common understandingbased on classic female-based demography is that human ag-ing outpaces reproduction in a manner that is contrary towhat would be expected by natural selection (Hamilton 1966).But in contrast to common acceptance of low male fertilityat older ages, one that basically mirrors menopause in in-dustrial societies, Marlowe (2000b) and Charlesworth (2001)suggested fertility in older males (150 years old) to be higherin nonindustrialized populations. A comparison of male fer-tility in several populations by Tuljapurkar, Puleston, andGurven (2007) supports their assertion, arguing that fertilityat older ages provides the selective advantage that results ingreater life span, with genes associated with longevity being

passed to daughters. Therefore, if older males can procuremating opportunities because of their ability to secure re-sources through skill and mental capacity as opposed to phys-ical prowess, this would result in significant male fertility atolder ages, a trait unique to Homo (fig. 2).

This also has implications for the evolution of human lon-gevity, menopause, and the role of grandmothers. For ex-ample, Eisenberg (2011) suggests that telomere length is fa-vored when older men father offspring. Offspring of oldermen tend to have longer telomeres, a possible reflection ofincreased investment in somatic maintenance. Marlowe ar-gues further that increases in male fertility at older ages andselection for genes that increase longevity can be passed onto daughters who then live past the viability of their ova.Consequently, the evolution of increased longevity via greatermale fertility at older ages may have contributed to the emer-gence of female longevity and postreproductive female re-productive effort through grandmother-based provisioningand child care (Hawkes et al. 1998; Tuljapurkar, Puleston, andGurven 2007).

Proximate Mechanisms and Male Life History

As with most other vertebrates, reproductive effort in Homois under considerable hormonal control, regulating energeticresources between investment in mating effort, survivorship,and paternal care (Bribiescas 2001). Understanding the role


of hormones in hominin evolution has potential but must bedone cautiously because we are limited to using extant greatape models of endocrine function that may have differed insignificant ways from other extinct hominins (Crews and Ger-ber 2003). Nonetheless, early Homo life history and morpho-logical traits are more in line with modern human traits com-pared with other great apes. Therefore, analyses of hormonalaspects of reproductive strategies in modern human malescan be assumed to be at least a modest reflection of earlyHomo.

As in other species, human male reproductive effort islargely influenced by testosterone, which governs somatic andbehavioral investment allocations (Bribiescas 1996, 2001).Testosterone also has a negative effect on immune functionand allocation trade-offs between maintenance and repro-ductive effort (Bribiescas and Ellison 2007; Muehlenbein andBribiescas 2005). In general, humans exhibit a 10-fold rangeof variation in testosterone levels under daily conditions (Zitz-mann and Nieschlag 2001), although nonindustrial popula-tions exhibit lower levels and less variation, which are likelyto be more indicative of early Homo endocrine function com-pared with more modern ecological settings (Bribiescas 1996,2001). Age can also attenuate testosterone (Harman et al.2001), although the degree and pattern change with age isless in nonindustrial societies (Bribiescas and Hill 2010; El-lison et al. 2002).

After controlling for energetics, age, and other extraneousfactors, men who traditionally spend more time with theirchildren exhibit lower testosterone levels than men who donot (Muller et al. 2009). A longitudinal investigation of pair-bonded Filipino men before and after becoming fathers re-vealed a 26% and 34% decrease in median levels of morningand evening testosterone, respectively, compared with singlenonfathers (Gettler et al. 2011). The selective factors involvedwith these testosterone differences in association with pater-nity are better understood and documented in other verte-brates, such as lower offspring mortality due to paternal in-vestment and provisioning (Ketterson et al. 1992); however,they are less clear in Homo.

The physiology of human reproductive behavior is alsomodulated by variation in neuropeptide hormones such asoxytocin, vasopressin, and prolactin, reflecting facultative ad-justment of paternal care (Gray and Anderson 2010). Thesehormones commonly affect paternal investment in manymammals and all but certainly affected the behavioral rep-ertoire of early Homo (Wynne-Edwards 2001). The ranges oftestosterone within and between individuals as well as withinand between populations illustrate the range of male Homoplasticity.

Deriving Evidence of Reproductive Strategiesfrom the Homo Fossil Record

Viewing the hominin fossil record through the lens of lifehistory theory, energetics, and reproductive ecology can pro-

vide useful insights (Catlett et al. 2010). However, testinghypotheses regarding hominin behavior using the fossil recordis a daunting task. Sexual dimorphism in the mammalianfossil record is among the most salient reflections of behaviorthat can be discerned from extinct species; however, the fossilrecord of Homo is relatively sparse compared with other mam-mals, resulting in vigorous debate around body size estimationand sexual dimorphism (Anton and Snodgrass 2012; Graveset al. 2010; Plavcan 2012; Pontzer 2012). Nonetheless, thepossibility of discerning cues of male reproductive effort inthe fossil record using specific areas of morphology that arerelated to reproduction would be extremely useful. For ex-ample, baculum length relative to body size has been suggestedto be associated with mating strategies, with the smallest bac-ulums being observed in more pair-bonded primate species(Dixson 1987; Martin 2007). Unfortunately, only one primatebaculum from an unidentified Eocene adapiform has beenfound in the fossil record (von Koenigswald 1979). The lackof baculum in modern humans is difficult to interpret becausethe absence of the baculum does not necessarily mean thatas a species, we fall at the absolute lower end of the range ofprimate size variation. Nonetheless, the discovery of a hom-inin baculum (if present) would be informative.

The examination of other morphological characters thatare under the influence of sex hormones and indicators ofreproductive behaviors is a palpable possibility. For example,sex-specific differences in the pelvis reflect variation betweenmales and females in the action and circulating levels of es-trogens. Similar research might be deployed to explore tes-tosterone-sensitive areas of male skeletal morphology, such ascraniofacial features, and cautiously testing for associationswith aspects of reproductive effort, such as behavior, attrac-tiveness, and competitiveness. Obviously there are tremen-dous hurdles to consider, not the least of which is sorting outsources of inter- and intraspecific variation. However, giventhe paucity of information available from the Homo fossilrecord and the compelling information that has been derivedregarding craniofacial morphology, reproductive behavior,and hormones in modern humans, this is certainly worthyof consideration.

Associations between fossil teeth and sex hormones wouldbe of great value. However, sex hormones do not exhibit anysignificant relationship with variation in tooth morphology,such as crown size in modern humans (Guatelli-Steinberg,Sciulli, and Betsinger 2008). Craniofacial sexual dimorphismis a viable alternative because it is not isometrically relatedwith general body size dimorphism in primates and manyother organisms, with significant taxonomic differences in theassociation between craniofacial and body size sexual dimor-phism (Plavcan 2003). Craniofacial sexual dimorphism seemsto have been significantly influenced by selection factors thatare largely independent from those affecting sex differencesin overall body size. Moreover, associations between behaviorand sexually dimorphic features such as canine size and bodysize in primates are subject to significant error ranges and


Figure 3. Enhanced craniofacial bone growth (arrows) betweenpredefined points in response to low-dose testosterone treatmentin delayed-puberty boys (114 years old, ) compared withn p 7untreated height-matched controls ( ; Verdonck et al. 1999).n p 7

should be interpreted with caution (Plavcan and van Schaik1997). While investment in larger body size and perhaps ca-nine size indicates metabolic investment in physical compet-itive ability, evidence from modern humans suggests thatcraniofacial characteristics serve as cues of attractiveness, matequality, paternal investment, and competitiveness (Perrett,May, and Yoshikawa 1994; Pound, Penton-Voak, and Surridge2009; Roney et al. 2006; Waynforth, Delwadia, and Camm2005). Many if not all of these craniofacial features are theresult of testosterone variation during development (Verdoncket al. 1999).

The identification of testosterone receptors in bone is wellestablished (Colvard et al. 1989). Surprisingly, much of theevidence for the effects of testosterone on craniofacial featuresis indirect, emerging from growth observations during humanmale puberty or from receptor mapping in rodent models(Lin et al. 2004; Pirinen 1995) with no direct mapping ofcraniofacial androgen receptors in any primate model. Com-parisons of orchidectomized and sham-operated newbornmale mice indicate that testosterone presence does not affectoverall skull size but does result in differential craniofacialgrowth patterns. Parallel comparison of ovarectomized andsham-operated newborn mice indicates only modest estrogeninfluences, thereby supporting the dominant role of testos-terone in sex differences in craniofacial morphology (Fujitaet al. 2004). Dahinten and Pucciarelli (1986) examined the

effects of castration, testosterone supplementation, and en-

ergetic stress on skull sexual dimorphism in rats. Their find-

ings showed that castration, testosterone supplementation,

and energetic stress, individually and in combination, had a

significant influence on skull sexual dimorphism as well as

variation in androgen-sensitive craniofacial features. Estrogen

supplementation of ovarectomized rats had no effect.

So what testosterone-sensitive craniofacial features should

be assessed in Homo or any other primate model? Such a

question requires a much greater and more detailed discussion

than what can be provided now. However, there are some

encouraging road signs that may provide some initial guid-

ance. Low-dose testosterone treatment of boys with delayed

puberty revealed significant changes in craniofacial features

compared with untreated boys. Significant distance-length in-

creases between predefined landmarks in testosterone-treated

boys were in mandibular ramus length, upper anterior face

height, and total cranial base length (Verdonck et al. 1999;

fig. 3). Because the focus of this study was on orthodontics,

unfortunately no measurements were made on other andro-

gen-sensitive sites such as the brow ridge.

Using this very preliminary guidance, the following strategy

might be deployed with the fossil record. (1) Determine the

presence, distribution, and density of androgen receptors on

craniofacial features in a primate model. Modern humans

would seem to be the most phylogenetically appropriate spe-

cies. (2) Determine and control for allometric effects between

craniofacial features and body size. (3) Conduct within-genus

(Homo) comparisons of the size, shape, landmark distances,

and bone thickness of fossil craniofacial regions that are as-

sociated with high concentrations of testosterone receptors.

(4) Compare variation in testosterone-sensitive craniofacial

regions in Homo fossil specimens with modern human male

reproductive behaviors and associations (i.e., attractiveness).

Drawing on modern human information, changes or dif-

ferences in androgen-sensitive craniofacial regions in the fossil

record could then be used (cautiously) as an indication of

the evolution of male features that may be sensitive to female

choice or indicators of intrasexual signaling. Obviously pleio-

tropic and other factors would need to be identified and

explored.

Beyond craniofacial and hormone associations, recent ad-

vances in sequencing the genome of Neanderthals and other

prehistoric organisms raise the possibility of employing sim-

ilar methods in early Homo specimens (Asara et al. 2007;

Green et al. 2010), although significant hurdles remain (Aus-

tin, Smith, and Thomas 1997), especially given the tropical

distribution of most early Homo specimens. If these hurdles

can be overcome, exploring gene variants that regulate hor-

mones associated with paternal behavior would be of great

utility toward understanding the evolutionary role of male

Homo behavior.


Conclusion

Understanding male life history and forging new and excitingmethods for testing hypotheses that emerge from life historytheory regarding the emergence of the genus Homo shouldbe a top research priority for anthropologists. As the fossilrecord of early Homo grows, it would be ideal to have methodssuch as those outlined at the ready to expand our under-standing of the emergence and evolution of our genus.

Acknowledgments

We thank Brenda Bradley, Erin Burke, Jessamy Doman, An-drew Hill, Marcia Inhorn, Grazyna Jasienska, Robert Walker,and Tim Webster for their insights into genes, fossils, and allthings male and paternal. Susan Anton, Leslie Aiello, and theWenner-Gren symposium participants made invaluable sug-gestions that vastly improved this paper.


Homo erectus female. American Journal of Human Biology 14:551–565.Aiello, Leslie C., and Jonathan C. K. Wells. 2002. Energetics and the evolution

of the genus Homo. Annual Review of Anthropology 31:323–338.Anderson, Kermyt G., Hillard Kaplan, David Lam, and Jane Lancaster. 1999.

Paternal care by genetic fathers and stepfathers. 2. Reports by Xhosa highschool students. Evolution and Human Behavior 20:433–451.

Anderson, Kermyt G., Hillard Kaplan, and Jane Lancaster. 1999. Paternal careby genetic fathers and stepfathers. 1. Reports from Albuquerque men. Evo-lution and Human Behavior 20:405–431.


Asara, John M., Mary H. Schweitzer, Lisa M. Freimark, Matthew Phillips, andLewis C. Cantley. 2007. Protein sequences from mastodon and Tyranno-saurus rex revealed by mass spectrometry. Science 316:280–285.

Austin, Jeremy J., Andrew B. Smith, and Richard H. Thomas. 1997. Palaeon-tology in a molecular world: the search for authentic ancient DNA. Trendsin Ecology & Evolution 12:303–306.

Bateman, Angus J. 1948. Intra-sexual selection in Drosophila. Heredity 2:349–368.

Bentley, Gillian R., and Ruth Mace. 2009. Substitute parents: biological andsocial perspective on alloparenting across human societies. Studies of the Bio-social Society, vol. 3. New York: Berghahn.

Bribiescas, Richard G. 1996. Testosterone levels among Ache hunter/gatherermen: a functional interpretation of population variation among adult males.Human Nature 7:163–188.

———. 2001. Reproductive ecology and life history of the human male.Yearbook of Physical Anthropology 44:148–176.

———. 2005. Age-related differences in serum gonadotropin (FSH and LH),salivary testosterone, and 17-b estradiol levels among Ache Amerindianmales of Paraguay. American Journal of Physical Anthropology 127:114–121.

———. 2006. On the evolution of human male reproductive senescence:proximate mechanisms and life history strategies. Evolutionary Anthropology15:132–141.

———. 2010. An evolutionary and life history perspective on human malereproductive senescence. Annals of the New York Academy of Sciences 1204:54–64.

Bribiescas, Richard G., and Peter T. Ellison. 2007. How hormones mediatetrade-offs in human health and disease. In Evolution in health and disease.Stephen C. Stearns and Jacob C. Koella, eds. Pp. 77–93. Oxford: OxfordUniversity Press.

Bribiescas, Richard G., and Kim R. Hill. 2010. Circadian variation in salivarytestosterone across age classes in Ache Amerindian males of Paraguay. Amer-ican Journal of Human Biology 22:216–220.

Brown, Frank, John Harris, Richard Leakey, and Alan Walker. 1985. EarlyHomo erectus skeleton from west Lake Turkana, Kenya. Nature 316:788–792.

Buss, David M. 1989. Sex differences in human mate preferences: evolutionaryhypotheses tested in 37 cultures. Behavioural and Brain Sciences 12:1–49.

Busse, Curt D. 1985. Paternity recognition in multi-male primate groups.American Zoologist 25:873–881.

Calow, Peter. 1979. The cost of reproduction: a physiological approach. Bi-ological Reviews 54:23–40.

Catlett, Kierstin K., Gary T. Schwartz, Laurie R. Godfrey, and William L.Jungers. 2010. “Life history space”: a multivariate analysis of life historyvariation in extant and extinct Malagasy lemurs. American Journal of PhysicalAnthropology 142:391–404.

Charlesworth, Brian. 2001. Patterns of age-specific means and genetic vari-ances of mortality rates predicted by the mutation-accumulation theory ofageing. Journal of Theoretical Biology 210:47–65.

Clutton-Brock, Timothy H. 1984. Reproductive effort and terminal investmentin iteroparous animals. American Naturalist 123:212–229.

Colvard, Douglas S., Erik F. Eriksen, Philip E. Keeting, Elizabeth M. Wilson,Dennis B. Lubahn, Frank S. French, B. Lawrence Riggs, and Thomas C.Spelsberg. 1989. Identification of androgen receptors in normal humanosteoblast-like cells. Proceedings of the National Academy of Sciences of theUSA 86:854–857.

Copeland, Sandi R., Matt Sponheimer, Darryl J. de Ruiter, Julia A. Lee-Thorp,Daryl Codron, Petrus J. le Roux, Vaughn Grimes, and Michael P. Richards.2011. Strontium isotope evidence for landscape use by early hominins.Nature 474:76–78.

Coqueugniot, Helene, Jean-Jacques Hublin, Francis Veillon, Francis Houet,and Teuku Jacob. 2004. Early brain growth in Homo erectus and implicationsfor cognitive ability. Nature 431:299–302.

Crews, Douglas E., and Linda M. Gerber. 2003. Reconstructing life history ofhominids and humans. Collegium Antropologicum 27:7–22.

Cronk, Lee. 1991. Wealth, status, and reproductive success among the Mu-kogodo of Kenya. American Anthropologist 93:345–360.

Dahinten, Silvia L., and Hector M. Pucciarelli. 1986. Variations in sexualdimorphism in the skulls of rats subjected to malnutrition, castration, andtreatment with gonadal hormones. American Journal of Physical Anthro-pology 71:63–67.

Daly, Martin, and Margo Wilson. 1999. The truth about Cinderella: a Dar-winian view of parental love. Darwinism Today. New Haven, CT: Yale Uni-versity 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.

de La Rochebrochard, Elise, Jacques de Mouzon, Francois Thepot, and PatrickThonneau. 2006. Fathers over 40 and increased failure to conceive: thelessons of in vitro fertilization in France. Fertility and Sterility 85:1420–1424.

Dixson, Alan F. 1987. Baculum length and copulatory behavior in primates.American Journal of Primatology 13:51–60.

Eisenberg, Dan T. 2011. An evolutionary review of human telomere biology:the thrifty telomere hypothesis and notes on potential adaptive paternaleffects. American Journal of Human Biology 23:149–167.

Ellison, Peter T. 1982. Skeletal growth, fatness and menarcheal age: a com-parison of two hypotheses. Human Biology 54:269–281.

———. 2003. Energetics and reproductive effort. American Journal of HumanBiology 15:342–351.

Ellison, Peter T., Richard G. Bribiescas, Gillian R. Bentley, Benjamin C. Camp-bell, Susan F. Lipson, Catherine Panter-Brick, and Kim Hill. 2002. Popu-lation variation in age-related decline in male salivary testosterone. HumanReproduction 17:3251–3253.

Finlayson, Clive. 2005. Biogeography and evolution of the genus Homo. Trendsin Ecology & Evolution 20:457–463.

Flinn, Mark V. 1988. Stepparent and genetic parent offspring relationships ina Caribbean village. Ethology and Sociobiology 9:335–369.

Fouts, Hillary N. 2008. Father involvement with young children among theAka and Bofi foragers. Cross-Cultural Research 42:290–312.

Fujita, Takako, J. Ohtani, M. Shigekawa, T. Kawata, M. Kaku, S. Kohno, K.Tsutsui, et al. 2004. Effects of sex hormone disturbances on craniofacialgrowth in newborn mice. Journal of Dental Research 83:250–254.

Gadgil, Madhav, and William Bossert. 1970. Life history consequences ofnatural selection. American Naturalist 104:1–24.


Gauthier, Anne H., Timothy M. Smeeding, and Frank F. Furstenberg. 2004.Are parents investing less time in children? trends in selected industrializedcountries. Population and Development Review 30:647–671.

Gettler, Lee T. 2010. Direct male care and hominin evolution: why male-childinteraction is more than a nice social idea. American Anthropologist 112:7–21.

Gettler, Lee T., Thomas W. McDade, Alan B. Feranil, and Christopher W.Kuzawa. 2011. Longitudinal evidence that fatherhood decreases testosteronein human males. Proceedings of the National Academy of Sciences of the USA108(39):16194–16199, doi:10.1073/pnas.1105403108.


Gray, Peter B., and Kermyt G. Anderson. 2010. Fatherhood: evolution andhuman paternal behavior. Cambridge, MA: Harvard University Press.

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.

Gross, Mart R. 1996. Alternative reproductive strategies and tactics: diversitywithin sexes. Trends in Ecology & Evolution 11:92–98.

Guatelli-Steinberg, Debbie, Paul W. Sciulli, and Tracy K. Betsinger. 2008.Dental crown size and sex hormone concentrations: another look at thedevelopment of sexual dimorphism. American Journal of Physical Anthro-pology 137:324–333.

Gurven, Michael D., and Kim Hill. 2009. Why do men hunt? a reevaluationof “man the hunter” and the sexual division of labor. Current Anthropology50:51–74.

Hamilton, William. 1966. The moulding of senescence by natural selection.Journal of Theoretical Biology 12:12–45.

Harcourt, Alexander H., Paul H. Harvey, Susan G. Larson, and R. V. Short.1981. Testis weight, body weight and breeding system in primates. Nature293:55–57.

Harman, S. Mitchell, E. Jeffrey Metter, Jordan D. Tobin, Jay Pearson, andMarc R. Blackman. 2001. Longitudinal effects of aging on serum total andfree testosterone levels in healthy men: Baltimore Longitudinal Study ofAging. Journal of Clinical Endocrinology and Metabolism 86:724–731.

Hartung, John. 1985. Matrilineal inheritance: new theory and analysis. Be-havioral and Brain Sciences 8:661–670.

Hawkes, Kristen. 1991. Showing off: tests of an hypothesis about men’s for-aging goals. Ethology and Sociobiology 12:29–54.

Hawkes, Kristen, John F. O’Connell, Nicholas G. B. Jones, Helen Alvarez, andEric L. Charnov. 1998. Grandmothering, menopause, and the evolution ofhuman life histories. Proceedings of the National Academy of Sciences of theUSA 95:1336–1339.

Hewlett, Barry S. 1991. Intimate fathers: the nature and context of Aka Pygmypaternal infant care. Ann Arbor: University of Michigan Press.

———, ed. 1992. Father-child relations: cultural and biosocial contexts. Foun-dations of Human Behavior. Hawthorne, NY: de Gruyter.

Hill, Kim R., Robert S. Walker, Miran Bozicevic, James Eder, Thomas Head-land, Barry Hewlett, A. Magdalena Hurtado, Frank Marlowe, Polly Wiess-ner, and Brian Wood. 2011. Co-residence patterns in hunter-gatherer so-cieties show unique human social structure. Science 331:1286–1289.

Kaplan, Hillard, and Kim Hill. 1985. Hunting ability and reproductive successamong male Ache foragers: preliminary results. Current Anthropology 26:131–133.

Ketterson, Ellen D., Van Nolan Jr., Licia Wolf, and Charles Ziegenfus. 1992.Testosterone and avian life histories: effects of experimentally elevated tes-tosterone on behavior and correlates of fitness in the dark-eyed junco (Juncohyemalis). American Naturalist 140:980–999.

Key, Catherine, and Leslie C. Aiello. 2000. A prisoner’s dilemma model ofthe evolution of paternal care. Folia Primatologica (Basel) 71:77–92.

Key, Catherine, and Caroline Ross. 1999. Sex differences in energy expenditurein non-human primates. Proceedings of the Royal Society B 266:2479–2485.

Kleiman, Devra G., and James R. Malcolm. 1981. The evolution of maleparental investment in primates. In Parental care in mammals. David J.Gubernick and Peter H. Klopfer, eds. Pp. 347–387. New York: Plenum.

Knott, Cheryl D. 2001. Female reproductive ecology of the apes: implicationsfor human evolution. In Reproductive ecology and human evolution. PeterT. Ellison, ed. Pp. 429–463. New York: de Gruyter.

Kramer, Karen L. 2005. Children’s help and the pace of reproduction: co-operative breeding in humans. Evolutionary Anthropology 14:224–237.

Lalueza-Fox, Charles, Antonio Rosas, Almudena Estalrrich, Elena Gigli, PaulaF. Campos, Antonio Garcia-Tabernero, Samuel Garcia-Vargas, et al. 2011.

Genetic evidence for patrilocal mating behavior among Neandertal groups.Proceedings of the National Academy of Sciences of the USA 108:250–253.

Lappan, Susan. 2008. Male care of infants in a siamang (Symphalangus syn-dactylus) population including socially monogamous and polyandrousgroups. Behavioral Ecology and Sociobiology 62:1307–1317.


Lin, Ines C., Alison E. Slemp, Catherine Hwang, Jagajan Karmacharya, AshleyD. Gordon, and Richard E. Kirschner. 2004. Immunolocalization of an-drogen receptor in the developing craniofacial skeleton. Journal of Cranio-facial Surgery 15:922–927.

Marlowe, Frank. 1999. Male care and mating effort among Hadza foragers.Behavioral Ecology and Sociobiology 46:46–57.

———. 2000a. Paternal investment and the human mating system. Behav-ioural Processes 51:45–61.

———. 2000b. The patriarch hypothesis: an alternative explanation of men-opause. Human Nature 11:27–42.

———. 2010. The Hadza: hunter-gatherers of Tanzania. Origins of HumanBehavior and Culture, vol. 3. Berkeley: University of California Press.

Martin, Robert D. 2007. The evolution of human reproduction: a primato-logical perspective. American Journal of Physical Anthropology 134(suppl.45):59–84.

McHenry, Henry M. 1994. Behavioral ecological implications of early hominidbody size. Journal of Human Evolution 27:77–87.

Michod, Richard E. 1979. Evolution of life histories in response to age-specificmortality factors. American Naturalist 113:531–550.

Muehlenbein, Michael P., and Richard G. Bribiescas. 2005. Testosterone-mediated immune functions and male life histories. American Journal ofHuman Biology 17:527–558.

Muller, Martin N., Frank W. Marlow, Revocatus Bugumba, and Peter T. El-lison. 2009. Testosterone and paternal care in East African foragers andpastoralists. Proceedings of the Royal Society B 276:347–354.

Parke, Ross D. 1996. Fatherhood. Developing Child Series. Cambridge, MA:Harvard University Press.

Perrett, David I., Keith A. May, and Sakiko Yoshikawa. 1994. Facial shape andjudgements of female attractiveness. Nature 368:239–242.

Pigliucci, Massimo. 2001. Phenotypic plasticity: beyond nature and nurture.Baltimore: Johns Hopkins University Press.

Pirinen, Sinikka. 1995. Endocrine regulation of craniofacial growth. ActaOdontologica Scandinavica 53:179–185.

Plavcan, J. Michael. 2003. Scaling relationships between craniofacial sexualdimorphism and body mass dimorphism in primates: implications for thefossil record. American Journal of Physical Anthropology 120:38–60.

———. 2012. Body size, size variation, and sexual size dimorphism. CurrentAnthropology 53(suppl. 6):S409–S423.

Plavcan, J. Michael, and Carel P. van Schaik. 1997. Interpreting hominidbehavior on the basis of sexual dimorphism. Journal of Human Evolution32:345–374.

Pontzer, Herman. 2012. Ecological energetics in early Homo. Current An-thropology 53(suppl. 6):S346-S358.

Pound, Nicholas, Ian S. Penton-Voak, and Alison K. Surridge. 2009. Testos-terone responses to competition in men are related to facial masculinity.Proceedings of the Royal Society B 276:153–159.

Puts, David A. 2010. Beauty and the beast: mechanisms of sexual selectionin humans. Evolution and Human Behavior 31:157–175.

Relyea, Rick A. 2002. Costs of phenotypic plasticity. American Naturalist 159:272–282.

Roney, James R., Katherine N. Hanson, Kristina M. Durante, and Dario Maes-tripieri. 2006. Reading men’s faces: women’s mate attractiveness judgmentstrack men’s testosterone and interest in infants. Proceedings of the RoyalSociety B 273:2169–2175.

Schwalb, David W., Jun Nakazawa, Toshiya Yamamoto, and Jung-Hwan Hyun.2010. Fathering in Japan, China, and Korea: changing contexts, images,and roles. In The role of the father in child development. Michael Lamb, ed.Pp. 341–387. Hoboken, NJ: Wiley.


Sear, Rebecca, and Ruth Mace. 2008. Who keeps children alive? a review ofthe effects of kin on child survival. Evolution and Human Behavior 29:1–18.

Simpson, Scott W., Jay Quade, Naomi E. Levin, Robert Butler, Guillaume


Dupont-Nivet, Melanie Everett, and Sileshi Semaw. 2008. A female Homoerectus pelvis from Gona, Ethiopia. Science 322:1089–1092.

Smuts, Barbara, and David J. Gubernick. 1992. Male-infant relationships innon-human primates: paternal investment or mating effort? In Father-childrelations: cultural and biosocial contexts. Barry S. Hewlett, ed. Pp. 1–30.Foundations of Human Behavior. Hawthorne, NY: de Gruyter.

Stearns, Stephen C. 1989. Trade-offs in life history evolution. Functional Ecol-ogy 3:259–268.

Stearns, Stephen C., and Jacob C. Koella. 1986. The evolution of phenotypicplasticity in life-history traits: predictions of reaction norms for age andsize at maturity. Evolution 40:893–913.

Trivers, Robert L. 1972. Parental investment and sexual selection. In Sexualselection and the descent of man. Bernard Campbell, ed. Pp. 52–97. Chicago:Aldine.

Tuljapurkar, Shripad D., Cedric O. Puleston, and Michael D. Gurven. 2007.Why men matter: mating patterns drive evolution of human lifespan. PLoSONE 2:e785.

Verdonck, Anna, M. Gaethofs, Carine Carels, and Francis de Zegher. 1999.Effect of low-dose testosterone treatment on craniofacial growth in boyswith delayed puberty. European Journal of Orthodontics 21:137–143.

von Koenigswald, Wighart. 1979. Ein Lemurenrest aus dem eozanen Olschieferder Grube Messel bei Darmstadt. Palaontologische Zeitschrift 53:63–76.

Walker, Robert, and Kim Hill. 2003. Modeling growth and senescence inphysical performance among the Ache of eastern Paraguay. American Jour-nal of Human Biology 15:196–208.

Walker, Robert, Kim Hill, Hillard Kaplan, and Garnett McMillan. 2002. Age-dependency in hunting ability among the Ache of eastern Paraguay. Journalof Human Evolution 42:639–657.

Walker, Robert S., Mark V. Flinn, and Kim R. Hill. 2010. Evolutionary historyof partible paternity in lowland South America. Proceedings of the NationalAcademy of Sciences of the USA 107:19195–19200.

Waynforth, David, Sonia Delwadia, and Miriam Camm. 2005. The influenceof women’s mating strategies on preference for masculine facial architecture.Evolution and Human Behavior 26:409–416.

Whiting, John, and Beatrice Whiting. 1975. Aloofness and intimacy of hus-bands and wives: a cross-cultural study. Ethos 3:183–207.


Wynne-Edwards, Katherine E. 2001. Hormonal changes in mammalian fa-thers. Hormones and Behavior 40:139–145.

Zitzmann, Michael, and Eberhard Nieschlag. 2001. Testosterone levels inhealthy men and the relation to behavioural and physical characteristics:facts and constructs. European Journal of Endocrinology 144:183–197.



Evolution of Cooperation among MammalianCarnivores and Its Relevance

to Hominin Evolution

by Jennifer E. Smith, Eli M. Swanson, Daphna Reed,and Kay E. Holekamp

CA� Online-Only Material: Supplement A with PDF

Anthropological theory suggests direct links between the origins of cooperation in hominins and a shift toward anenergy-rich diet. Although the degree to which early hominins ate meat remains controversial, here we reevaluatethe notion, originally suggested by Schaller and Lowther in 1969, that mammalian carnivores can shed light onhuman origins. Precisely when cooperation evolved in hominins or carnivores is unknown, but species from bothgroups cooperatively hunt large game, defend resources, guard against predators, and rear young. We present alarge-scale comparative analysis of extant carnivore species, quantifying anatomical, ecological, and behavioral cor-relates of cooperation to determine whether metabolic rate, body and relative brain size, life history traits, and socialcohesion coevolved with cooperation. We focus heavily on spotted hyenas, which live in more complex societiesthan other carnivores. Hyenas regularly join forces with kin and nonkin to hunt large antelope and to defendresources during intergroup conflicts and disputes with lions. Our synthesis highlights reduced sexual dimorphism,increased reproductive investment, high population density, fission-fusion dynamics, endurance hunting of big gamein open habitats, and large brains as important correlates of cooperation among carnivores. We discuss the relevanceof our findings to understanding the origins of cooperation in hominins.

The evolutionary trajectory from hominin to humanity,from small-brained australopithecine to encephalisedHomo erectus, began 2.6 Ma with an interest in meat.(Bunn 2006:205)

The evolutionary origins and maintenance of cooperationpose an evolutionary puzzle for anthropologists and biologists(reviewed by Clutton-Brock 2009a; Dugatkin 2002; Melis and

Jennifer E. Smith is an Assistant Professor of Biology at Mills College(5000 MacArthur Boulevard, Oakland, California 94613, U.S.A.[[email protected]]). Eli M. Swanson is a PhD candidate in theZoology Department and Ecology, Evolutionary Biology, andBehavior (EEBB) Program (203 Natural Science Building, MichiganState University, East Lansing, Michigan 48824, U.S.A.). DaphnaReed is an undergraduate in the Institute for Society and Genetics,University of California, Los Angeles (Box 957221, 1323 Rolfe Hall,Los Angeles, California 90095-7221, U.S.A.). Kay E. Holekamp is aDistinguished University Professor of Zoology and Director of theEEBB Program (203 Natural Science Building, Michigan StateUniversity, East Lansing, Michigan 48824, U.S.A.). This paper wassubmitted 12 XII 11, accepted 6 VII 12, and electronically published13 XI 12.

Semmann 2010; Noe 2006; Nowak 2006; Queller 1985; West,El Mouden, and Gardner 2011; West, Griffin, and Gardner2007). Nevertheless, solving this mystery remains central tounderstanding the unprecedented capacity for human rangeexpansion into an extraordinary diversity of habitats acrossthe globe (Bingham 1999).

Although there is much ongoing debate about the precisecomposition of the diet of early hominins (e.g., degree towhich it included nuts, tubers, and protein and fat from an-imals), most current anthropological scenarios suggest a directlink between a transition toward eating a high-quality dietand the evolution of cooperation in hominins (e.g., Aielloand Wheeler 1995; Bramble and Lieberman 2004; O’Connell,Hawkes, and Blurton Jones 1999; O’Connell et al. 2002;Pontzer 2012; Wrangham et al. 1999). Although the precisetiming of events remains elusive, the suite of cooperative be-haviors that have been suggested to coevolve with a rich hom-inin diet includes group hunting, defense of food and space,protection from predators, and care of young (e.g., Bunn2006; Bunn and Kroll 1986; Byrne 1995; Grove 2010; Hartand Sussman 2009; Milton 1999; Ungar 2012; Wrangham etal. 1999).

Three major ecological models proposed to explain hom-


Smith et al. Carnivore Cooperation S437

inin evolution are currently prevalent in the literature. Thefirst two models focus on the importance of eating a calorie-rich diet but differ in their emphasis on the nature of thecalorie-rich diet and how foraging shaped the origins of co-operation (reviewed by Pontzer 2012). The “hunting-scav-enging” model emphasizes the importance of energy-richmeat and bone marrow in the hominin diet. Some variantsof this model propose that natural selection favored coop-eration among individuals that hunted large game using per-sistence running (also called “cursorial hunting”), while oth-ers emphasize cooperation and food sharing with younger,less able or less successful kin (Aiello and Wheeler 1995;Bramble and Lieberman 2004; Kaplan et al. 2000). The secondmodel is the “underground storage organ” (USO) model thatemphasizes gathering USOs (such as tubers), food sharing,and sometimes cooking (O’Connell, Hawkes, and BlurtonJones 1999; Wrangham et al. 1999). It suggests that a shifttoward an arid climate favored exploitation of these calorie-rich foods and emphasizes the role of elders in provisioningyounger kin. Where these models posit increased energeticbenefits from intergenerational cooperation, arguments canbe made for the coevolution of increases in the length oflactation, gestation, and longevity as well as increases in ne-onate and adult body mass and daily energy expenditure buta decrease in sexual dimorphism. In contrast to these twodiet-related hypotheses, the “predator protection” hypothesissuggests that cooperation evolved because of intense preda-tion by large-bodied carnivores and that selection favored anincrease in relative brain size to permit complex forms ofcooperative defense required to outwit predators (Hart andSussman 2009).

Relevance of Extant Mammalian Carnivores toTesting of Ecological Models

Foundational inquiries about the social lives of early homininsfocused primarily on nonhuman primates (e.g., Reynolds1966; Washburn and Devore 1961), and primate research con-tinues to be an import source of inference today (e.g., Fuentes,Wyczalkowski, and MacKinnon 2010). Nonetheless, Schallerand Lowther (1969) provided the transformative insight thatthe ecologies of modern hunter-gathers—and by inferencethose of early hominins—might closely resemble those ofextant carnivores. Schaller and Lowther (1969:308) proposedthat “it might be more productive to compare hominids [nowhominins] with animals which are ecologically but not nec-essarily phylogenetically similar, such as the social carnivores.”This landmark paper was followed by several pioneering stud-ies on wild carnivores (e.g., Kruuk 1972; Mech 1970; Schaller1972). Ongoing research continues to reveal new complexitiesabout the social lives of carnivores. Because the behavioraltraits proposed to be important for hominin evolution arealso salient features in the lives of social carnivores, study ofthis taxonomic group might indeed offer important insights.

Here we use the term “carnivore” to refer to those extantspecies of mammals belonging to the order Carnivora re-gardless of their dietary niche. The order Carnivora aroseduring the late Paleocene from a radiation of mammals whosediet was comprised primarily of meat (Wilson and Mitter-meier 2009). Importantly, however, extant carnivores occupya vast range of habitats and many ecological niches; speciesin this order belong to dietary classes that include herbivores,insectivores, omnivores, piscivores, and carnivores (Wilsonand Mittermeier 2009). Although the extent to which earlyhominins ate meat is the subject of ongoing debate, there isgrowing evidence that additional aspects of hominin socialitymight resemble extant mammalian carnivores. For instance,as is the case for many mammalian carnivores (reviewed byPalomares and Caro 1999), early hominins also were likelyhunted or otherwise killed by carnivores (Hart and Sussman2009). The evolution of cooperative defense of food and space,breeding, and protection from predators as well as fission-fusion dynamics (see below) in the order Carnivora suggeststhat this taxonomic group continues to offer underexploitedopportunities for testing hypotheses relevant to the evolutionof hominins regardless of the precise diets of early hominins.

Evidence that early hominins included some meat in theirdiet from the Pleistocene onward and perhaps earlier becomesincreasingly common as we move forward in the archaeo-logical record. Although stone tools may have been used pri-marily for processing vegetative foods (Grine and Fleagle2009), discoveries of fossilized bones with butchery markssuggest early hominins might have used stones to removeenergy-rich flesh from the carcasses of large mammals by 2.5–3.4 Ma (McPherron et al. 2010; but see Domınguez-Rodrigo,Pickering, and Bunn 2010, 2011). Moreover, sharp-edged cut-ting tools and cut-marked animal bones at Gona, Ethiopia(Semaw et al. 1997), and stone caches at Olduvai, Tanzania(Bunn and Kroll 1986; Potts and Shipman 1981), suggesthominin butchery and consumption of skeletal muscle andtissues by 2.6 Ma (Potts 2012). By 2.5–2.0 Ma, stone tooltransport (e.g., carrying stones hundreds of meters) for ex-tractive foraging permitted access to new resources (Potts2012). Molar morphology and microwear data indicate die-tary expansion to include items with mechanical propertiesconsistent with those of animal fat and protein around 2.0–1.8Ma (Ungar 2012). Fauna in archeological contexts suggestpersistent foraging on game by 2.0–1.5 Ma (Potts 2012).

Nonetheless, the debate continues regarding how earlyhominins acquired prey animals (Bunn and Kroll 1986; Pottsand Shipman 1981). Were prey run to exhaustion, ambushedat short range, or passively scavenged from carcasses aban-doned by mammalian carnivores (Bramble and Lieberman2004; Bunn and Pickering 2010; Domınguez-Rodrigo andPickering 2003; Shipman 1986)? Which physical, ecological,and social factors facilitated the evolutionary leap from Aus-tralopithecus to early Homo? To answer these questions, wesearch here for convergences between early hominin foragers


and extant carnivores (e.g., Finarelli 2010; Finarelli and Flynn2009).

Our overarching goal is to provide an updated assessmentregarding whether the behavior of extant mammalian car-nivores is indeed relevant to understanding early homininsirrespective of the degree to which the hominins may havebeen carnivorous. We draw on a unique combination of newknowledge and new computational tools unavailable toSchaler and Lowther in the 1960s to reevaluate this notion.We first review the current literature on cooperation in mam-malian carnivores, focusing primarily on new data on onehighly cooperative species, the spotted hyena Crocuta crocuta,to update the original framework presented by Schaller andLowther (1969). Then we perform a large-scale comparativeanalysis on extant species from Carnivora to identify variablespermitting and constraining cooperation. Finally, we considerthese findings in light of current ecological models proposedto explain the evolutionary origins for cooperation in thegenus Homo.

Feeding Competition and Fission-FusionDynamics in Extant Social Carnivores

Most (85%–90%) terrestrial mammalian carnivores are sol-itary, interacting exclusively with their mates and offspring oralien conspecifics at territorial boundaries (reviewed by Hole-kamp, Boydston, and Smale 2000). Sociality appears to havearisen as a derived trait because the ancestral condition withinmost carnivore families is to live solitarily (Dalerum 2007).Group life permits individuals to detect or evade predatorsor to improve their ability to acquire or defend resources(Johnson et al. 2002). In some species of carnivores, gregar-iousness itself may have been favored by improved energyintake (Creel and Macdonald 1995; Dalerum 2007). Some ofthe best-studied cooperative hunters include spotted hyenas(Holekamp et al. 1997; Kruuk 1972; Smith et al. 2008), lions(Panthera leo; Packer and Ruttan 1988; Packer, Scheel, andPusey 1990; Scheel and Packer 1991), African wild dogs (Ly-caon pictus; Creel 1997; Creel and Creel 1995), and wolves(Canis spp.; Mech 1970). For these species, prey animals rep-resent large, ephemeral packets of energy-rich food that occurunpredictably in space and time. Often several hunters maybe required to secure a single prey animal, each of which mayweigh hundreds of kilograms. Although lions assume specificroles when hunting in groups (Heinsohn and Packer 1995;Stander 1992), there is no evidence that lions or other car-nivores rely on advanced planning to capture prey (Hole-kamp, Boydston, and Smale 2000). For example, even thoughspotted hyenas are efficient hunters and directly kill 60%–95% of the food they eat, these carnivores appear to followsimple rules when hunting together, such as “move whereveryou need to in order to keep the selected prey animal betweenyou and another hunter.”

Given the nature and distribution of their food resources,contest competition at ungulate carcasses is often intense

among carnivores, and access to food is critical to individualfitness (e.g., Carbone et al. 2005; Courchamp, Rasmussen,and Macdonald 2002; Holekamp, Smale, and Szykman 1996).Even within the most stable groups of carnivores, intensecompetition disrupts grouping behavior (reviewed by Aureliet al. 2008; Holekamp, Boydston, and Smale 2000; Smith etal. 2008). Whereas individual African wild dogs living in co-hesive groups maximize per capita energy gain when huntingin large packs (Creel and Creel 1995), individual spotted hy-enas, especially low-ranking ones, accrue the greatest energeticbenefits when they hunt alone (fig. 1; Smith et al. 2008).Hunting in groups larger than an optimal size is also costlyto lions (Packer, Scheel, and Pusey 1990).

Intense competition leads most social carnivores to live ingroups structured by fission-fusion dynamics in which in-dividuals regularly break up into small foraging parties whenfood is scarce and gather again when food is abundant (re-viewed by Smith et al. 2008). For example, dholes (Cuonalpinus; Venkataraman, Arumugam, and Sukumar 1995),white-nosed coatis (Nasua narica; Gompper 1996), Europeanbadgers (Meles meles; Kruuk and Parish 1982), and kinkajous(Potos flavus; Kays and Gittleman 2001), as well as spotted,brown (Hyaena brunnea), and striped (Hyaena hyaena; Kruuk1976; Mills 1990; Smith et al. 2008; Wagner 2006) hyenastemporarily leave (fission from) their companions to avoidcompetitors when feeding. Members of each of these speciesalso regularly meet up again and spend time with (fusionwith) conspecifics when the benefits of group living are high(reviewed by Aureli et al. 2008; Creel and Macdonald 1995).

Maternal Capital, Tolerance, and Coalitions inHyenas and Other Carnivores

Social Complexity of Spotted Hyenas

Spotted hyenas represent a well-studied species; these hyenaslive in societies that are considerably more complex than thoseof other gregarious carnivores (Drea and Frank 2003; Hole-kamp, Sakai, and Lundrigan 2007). In fact, their social livesare strikingly similar to those of many species of Old Worldmonkeys. Spotted hyenas are an interesting species in whichto investigate the evolution and mechanisms promoting co-operation in carnivores. Most social carnivores—includingwolves, social mongooses such as meerkats (Suricata suri-catta), lions, and wild dogs—live in small groups in whichadult members of each sex are closely related to one another(Clutton-Brock 2002; Creel and Creel 1991). In contrast, spot-ted hyenas reside in large permanent social groups called“clans” (Kruuk 1972), consisting of up to 90 or more indi-viduals with low mean relatedness (Holekamp et al. 2012).

Hyena clans are strikingly similar in their size, composition,and hierarchical organization to troops of Old World mon-keys. Like troops of macaques, baboons, and vervet monkeys,hyena clans contain multiple adult males and multiple ma-trilines of adult female kin and their offspring (Frank 1986).


Figure 1. Per capita energy gain as a function of foraging groupsize among adult (A) African wild dogs Lycaon pictus (reprintedfrom Creel 1997 with permission from Elsevier) and (B) spottedhyenas Crocuta crocuta (reprinted from Smith 2008 with per-mission from Elsevier). Points in A represent mean numbers ofwild dogs, with point size proportional to the number of ob-servations, and the dashed line represents the linear regression;points in B represent individual hyenas found in subgroups ofvarious sizes.

As in these species of monkeys, individual hyenas within eachclan can be ranked in a linear dominance hierarchy based onoutcomes of agonistic interactions (Frank 1986; Kruuk 1972;Smith et al. 2008; Tilson and Hamilton 1984). Dominancerelationships are extremely stable across years and ecologicalcontexts (Frank 1986; Smith et al. 2011), but rank itself isnot correlated with size or fighting ability (Engh et al. 2000).Instead, as in many monkeys (e.g., Chapais 1992; Cheney1977; Horrocks and Hunte 1983; Walters 1980), coalition for-mation plays an important role in acquisition and mainte-nance of social rank among spotted hyenas (Engh et al. 2000;Holekamp and Smale 1993; Smale, Frank, and Holekamp1993; Smith et al. 2010; Zabel et al. 1992).

During an early stage of ontogeny, each hyena comes tounderstand its own position in its clan’s dominance hierarchy(Holekamp and Smale 1993; Smale, Frank, and Holekamp1993). This process requires a type of associative learningcalled “maternal rank inheritance” in which the mediatingmechanisms are virtually identical to those operating in cer-copithecine primates (Engh et al. 2000; Holekamp and Smale1991). In fact, because of their aptitude for social learning,spotted hyenas are capable of solving cooperation problemsin captivity (Drea and Carter 2009).

Maternal Capital Influences Reproductive Successof Spotted Hyenas

Recent data suggest that maternal capital, in terms of bothsocial allies and energy reserves, has important life historyconsequences for mammals (Bribiescas, Ellison, and Gray2012; Isler and van Schaik 2012; Wells 2012). Social capitalof mothers clearly is important for nonhuman primates suchas in baboons Papio spp. (Silk, Alberts, and Altmann 2003;Silk et al. 2009). Similarly, data from spotted hyenas living in“baboon-like” societies suggest that maternal phenotype andmaternal rank in particular have profound effects on femalereproductive success (Holekamp et al. 2012). Support net-works, especially among maternal kin, endure across the lifespan despite ecological constraints imposed on them by fluc-tuating availability of resources (Holekamp et al. 2012). Be-cause social status determines priority of access to kills, highrank has enormous effects on hyenas’ net energy gain (Hoferand East 2003; Smith et al. 2008). Birthrates and survivorshipare so much greater for high- than for low-ranking hyenas(Watts et al. 2009) that dominants tend to have many moresurviving kin in the population than do subordinates (Hole-kamp et al. 2012). As a result of these large networks of allies,high-ranking hyenas have the most social capital (e.g., Smithet al. 2010; Van Horn et al. 2004).

Compared with most other carnivores, spotted hyenas havea prolonged period of fetal development, are of unusuallylarge mass at birth, and are remarkably precocial; they areborn with open eyes and fully erupted canine and incisorteeth (Drea and Frank 2003). Nevertheless, cubs undergo anexceptionally long period of nutritional dependence afterbirth, and dependence on the mother continues long aftercubs are weaned because their feeding apparatus develops veryslowly, and this handicaps their feeding (Watts et al. 2009).Maternal capital is extremely influential in determining thepace of development and reproduction (Holekamp et al.2012). Cubs usually nurse for over a year and may nurse forup to 2 years. Because it takes juveniles years of practice tobecome proficient at hunting and at cracking through boneto access marrow (Holekamp et al. 1997; Mills 1990; Tanneret al. 2010), mothers use coalitionary aggression to help theiroffspring gain access to ungulate kills long after weaning.

Among nonkin, adult female hyenas preferentially toleratefeeding at shared kills by those nonkin with which they as-


sociate most often (Smith, Memenis, and Holekamp 2007).Dominants engage in reciprocal trading for services providedby subordinates, such as help hunting and defense of territoryboundaries. In exchange, dominants withhold aggressionfrom those unrelated hyenas with which they maintain thestrongest relationships. In this respect, spotted hyenas differfrom most carnivores, because most species of mammaliancarnivores allow only their kin to feed at kills. For example,lions and African wild dogs (Creel and Creel 1995; Packer,Pusey, and Eberly 2001) only hunt cooperatively and sharemeat with relatives or mates (reviewed by Clutton-Brock2009b). Individual differences in the extent to which spottedhyenas tolerate one another in feeding contexts suggest somedegree of meat sharing among unrelated individuals (Smith,Memenis, and Holekamp 2007).

Although all adult female hyenas breed, rates of reproduc-tion vary based on social rank and prey density (Frank, Hole-kamp, and Smale 1995; Hofer and East 2003; Holekamp,Smale, and Szykman 1996). Rank-related variation in females’ability to access food (e.g., Frank 1986; Smith et al. 2008) hasstriking effects on the growth rates of their cubs; high-rankingcubs grow faster than their low-ranking peers (Hofer and East1996, 2003). Dominant females start breeding at younger agesthan do subordinates (Frank, Holekamp, and Smale 1995;Holekamp, Smale, and Szykman 1996; Watts et al. 2009).Interestingly, above and beyond the effects of rank, largerfemale hyenas produce more offspring over their lifetimesthan do smaller females, suggesting that large body size con-fers an evolutionary advantage (Swanson, Dworkin, and Hole-kamp 2011).

Evolutionary Forces and Mechanisms Promoting Intragroupand Intergroup Coalitions

As in most monkeys and great apes, spotted hyenas bias theirsocial support toward kin during intragroup disputes (re-viewed by Smith et al. 2010). We recently elucidated the evo-lutionary forces favoring intragroup coalitions among adultfemale spotted hyenas (Smith et al. 2010). First, we tested theprediction from kin selection theory (Hamilton 1964a, 1964b)that individuals should bias helpful behavior toward relativesand harmful behavior away from relatives if doing so providesinclusive fitness benefits. Second, we examined the hypothesisthat natural selection might favor interventions on behalf ofnonkin via reciprocal altruism if the projected future benefitsto the donor outweigh the immediate costs (Trivers 1971).Finally, we asked whether females gain direct benefits fromcooperative acts through better access to food or by rein-forcing the status quo (Brown 1983; Connor 1995; West-Eberhard 1975). As predicted by kin selection theory, femalespotted hyenas support close maternal and paternal kin mostoften, and the density of cooperation networks increases withgenetic relatedness. As is the case in most animal societies(reviewed by Clutton-Brock 2009a), we found no evidenceof reciprocal altruism (e.g., enduring alliances based on re-

ciprocal support among nonkin). Instead, hyenas gained di-rect benefits from joining forces to attack subordinates andmonitored the number of dominant bystanders in the “au-dience” at fights to minimize costs to themselves. Taken to-gether, the combined evolutionary forces of kin selection anddirect benefits appear to favor flexible decisions regardingwhether or not adult female hyenas intervene in fights (Smithet al. 2010).

Local ecology determines whether hyena clans maintainwell-defined and vigorously defended territories (e.g., MasaiMara, Kenya; Boydston, Morelli, and Holekamp 2001) ormore permeable territorial boundaries with considerablerange overlap and tolerance of outsiders (e.g., the KalahariDesert; Mills 1990). Populations that engage in territory de-fense form large intergroup coalitions against members ofother social groups. These intergroup disputes over territoryboundaries or kills may involve up to 56 group mates joiningforces against a common enemy (fig. 2; Smith et al. 2008).Individuals within hyena clans are on average more closelyrelated to one another than to individuals belonging to neigh-boring clans, but relatedness among hyena clan members islow (Van Horn et al. 2004). Thus, as in chimpanzees (e.g.,Pan troglodytes; Goldberg and Wrangham 1997), spotted hy-enas on average derive large net direct fitness benefits fromjoining forces with large numbers of nonrelatives during in-tergroup conflicts despite the risk of serious injury or death(Boydston, Morelli, and Holekamp 2001; Henschel and Skin-ner 1991; Hofer and East 1993; Kruuk 1972; Mills 1990).Several other social carnivores, including white-nosed coatis(Nasua narica; Gompper, Gittleman, and Wayne 1998) andgray wolves (Canis lupus; Lehman et al. 1992), also engage inlethal intergroup conflicts at territorial borders. However, incontrast to most social carnivores who only join forces withgenetic relatives during intergroup conflicts (reviewed byClutton-Brock 2009b), spotted hyenas do so with large num-bers of unrelated group mates.

Whereas both cognitive and noncognitive (emotional andtemperamental) factors promote cooperation and tolerancein living chimpanzees and humans (Hrdy 2009; Melis andSemmann 2010; Tomasello et al. 2005), all available evidenceto date suggests that cooperation among extant carnivores isfacilitated by noncognitive mechanisms. For example, greetingceremonies facilitate intra- and intergroup cooperationamong hyenas, helping potential allies to reach the same mo-tivational state before cooperating (Smith et al. 2011). Greet-ings occur when two hyenas stand parallel to one anotherand sniff each others’ anogenital regions (East, Hofer, andWickler 1993; Kruuk 1972). These signals allow hyenas toquickly confirm relationship status in a society in which groupmembers spend much of their time apart (Smith et al. 2011).Similarly, in African wild dogs (Creel 1997; Creel and Creel2002) and gray wolves in North America (Mech 1970), greet-ings function to promote group hunting.

Because cooperation is vulnerable to cheaters, theory pre-dicts that punishment and threats should evolve, but there is


Figure 2. Mean � SE subgroup size (left vertical axis, histogram bars) and proportion of observations in which spotted hyenaswere found in subgroups containing more than one individual (right vertical axis, circles) in each of the following contexts: (1)hunting: one or more resident hyenas chased a prey animal for at least 50 m, regardless of the outcome; (2) natal den: one or moreresident hyenas observed at an isolated den used by only one mother for shelter of a single litter until her cubs are 2–5 weeks old;(3) kills: one or more resident hyenas observed feeding on at least one fresh ungulate carcass; (4) courtship/mating: immigrantmale(s) direct mating tactics toward a sexually mature female; (5) communal den: one or more resident hyenas observed at a denused concurrently by several litters; (6) border patrols: residents engaged in high rates of scent marking and defecation alongterritory boundaries; (7) clan wars (intergroup conflicts, called “clan wars” by Kruuk 1972): agonistic interactions between residentand alien hyenas at territory boundaries; (8) conflict with lions: agonistic interactions observed between resident hyenas and atleast one lion; (9) other: none of the contexts above applied. Sample sizes, shown below each bar, represent numbers of observationsessions assigned to each context. Different letters indicate statistically significant differences between contexts after correcting formultiple testing. The shaded bar represents the baseline value of subgroup sizes occurring in “other” sessions against which othergroups were compared. Figure adopted with permission from Smith et al. (2008).

limited evidence of effective threats against “free riders”among carnivores (reviewed by Cant 2011). Spotted hyenasuse unprovoked aggression to reinforce dominance status, yetthere is no evidence that these attacks promote coalition for-mation (Engh et al. 2005; Smith et al. 2010). Lions similarlyfail to punish cheaters in the context of group hunting(Packer, Pusey, and Eberly 2001) despite their ability to rec-ognize laggards (Heinsohn and Packer 1995). Finally, al-though banded mongooses Mungos mungo and meerkats doenforce cooperative breeding, the threat of eviction is inef-fective at preventing cheating in the first place (Cant et al.2010; Clutton-Brock, Hodge, and Flower 2008).

Testing Ecological Models Using Data fromExtant Mammalian Carnivores

We next used phylogenetic comparative methods to evaluatethe extent to which each of the competing ecological modelsof human evolution explains cooperation among extant mam-

malian carnivores. To test competing hypotheses, we assessedthe extent to which anatomical, life history, ecological, orbehavioral variables coevolved with cooperation among allspecies ( ) of carnivores for which all of the salientN p 87predictor variables of interest were currently available. Thisdata set (see CA� online supplement A) captures the diversesuite of social and ecological traits exhibited by members ofthe order Carnivora. Whenever possible, we extracted all datafor a single variable (e.g., diet or forms of cooperation ex-hibited) from the same source.

Forms of Cooperation Exhibited by Carnivores

We based the current analysis on reports from the literatureto calculate a composite score of cooperation by counting andsumming the number of different forms of cooperation ex-hibited by each species from among the following possibilities:(1) alloparental care, (2) group hunting, (3) intragroup co-alition formation, (4) coalition formation during intergroup


contests or warfare among conspecifics, and (5) cooperativeprotection from predators. Most cooperation data were ex-tracted from Creel and Macdonald (1995). Intragroup coa-lition data were from Smith et al. (2010).

Each of our five cooperation types was assigned a value of0 if it did not occur and 1 if it did occur in a particularspecies. The only exception to this rule was alloparenting;following Creel and Creel (1991), intermediate species wereassigned a score of 0.5 for alloparenting if they shared a com-munal den (e.g., home base at which young born to morethan one mother are raised) but did not engage in true allo-parental care such as allonursing or provisioning of the off-spring of others. Although this composite “cooperation score”assumes values ranging from 0 to 5, in some analyses, wesimply asked whether or not a species engaged in any of thesefive forms of cooperation.

Alloparental care is defined here as all aspects of care inwhich individuals guard, groom, carry, play with, feed, ornurse the offspring of others (Creel and Creel 1991). We definegroup hunting as concurrent attack by more than one con-specific directed toward a selected prey item regardless of itsoutcome or fitness consequences (Holekamp et al. 1997). Co-alition formation occurs when two or more individuals joinforces to direct aggression toward the same target(s). Intra-group coalitions are directed toward group mates, whereasintergroup coalitions are directed toward conspecifics be-longing to a different social group. Cooperative defenseagainst predators includes mobbing of predators (e.g., groupmembers collectively fend off potential predators by attackingthem) or cooperative vigilance, defined here as any behavioraladjustments that reduce the risk of predation for membersof the group.

Ecological Predictors of Cooperation

We defined two binary variables based on diet. First, we cat-egorized species as “meat eaters” using an absolute definitionbased on whether their diet was comprised mainly of anyform of meat, including small prey (e.g., rodents, birds).“Meat eaters” excluded primarily insectivorous, omnivorous,piscivorous, or herbivorous species. Second, we inquiredwhether species with diets comprised mainly of large verte-brate prey (110 kg; Gittleman 1989) had the highest coop-eration scores. Habitat types increased in vegetative coverfrom open (e.g., savannah) to mixed (e.g., woodland savan-nah) to closed (e.g., forest) habitats. Population density wasthe number of individuals per square kilometer.

Behavioral Predictors of Cooperation

Home range sizes were average values for adults of both sexes.We also assigned a binary variable based on whether or nota species hunted cursorially regardless of hunting group size.Cursorial hunters were defined as carnivores that primarilyused endurance to exhaust targets by chasing them for long

distances before capturing them. Noncursors were those spe-cies that relied on stealth while stalking and capturing preyat short distances or that displayed no hunting whatsoever.Finally, we assigned a cohesion index to reflect the increasingdegree of sociality for each focal species ranging from (1)solitary (only with conspecifics for mating), (2) pair living(stable bond between adult male and female of the samespecies), (3) fission-fusion dynamics, and (4) obligately gre-garious (always found in close proximity to conspecifics).

Anatomical and Physiological Predictors of Cooperation

Sexual dimorphism was measured as the body mass ratio ofadult males to adult females. We corrected brain volume(mm3) for body size by taking the residuals from phyloge-netically corrected regressions of log-transformed brain vol-ume on log-transformed mass. We used basal metabolic ratescorrected for body mass as a measure of energy expenditurebecause existing field metabolic data are rarely available forcarnivores.

Life History Predictors of Cooperation

Longevity was the maximum life span (in years) in the wildfor each species. Gestation length was the number of daysbetween conception and birth. Litter size, a measure of re-productive investment, was the mean number of offspringborn in a single litter. Neonate mass was offspring weight atbirth (kg) corrected for adult mass. Lactation duration wasthe mean number of days between birth and cessation oflactation in nursing females, and did not include subsequentperiods of offspring dependence.

Comparative Methods and Phylogenetic GeneralizedLeast Squares Regression

Phylogenetic comparative methods represent powerful toolsto study adaptation because they account for potential au-tocorrelation due to shared evolutionary history (e.g., Clut-ton-Brock and Harvey 1977; Felsenstein 1985; Gittleman andLuh 1992; Swanson et al. 2006). We considered shared an-cestry of these 87 species by building a phylogeny primarilybased on Bininda-Emonds et al. (2007) and adjusting somerelationships based on updated phylogenies or phylogeniesfocused on smaller taxonomic subgroups (Finarelli 2008;Flynn et al. 2005; Gottelli et al. 1994; Johnson et al. 2006;Koepfli et al. 2006, 2007, 2008; Patoua et al. 2009; Perini,Russo, and Schrago 2009; Sato et al. 2009; Yoder et al. 2003).We resolved polytomies wherever possible, assigning branchlengths of million years ago when branch lengths�61 # 10were not estimated in the source. The only unknown rela-tionships for our phylogeny were those of gray wolves, Ethi-opian wolves Canis simensis, and coyotes Canis latrans; thispolytomy was retained for regressions and randomly resolved.The resulting branch lengths were set to 0 for estimation ofphylogenetic signal and ancestor reconstruction. Unless stated


otherwise, all analyses were carried out in R version 2.12.1(R Foundation for Statistical Computing 2010).

We estimated the strength of the phylogenetic signal usingBlomberg’s K and tested whether it differed from that pre-dicted by a null model specifying no effect of phylogeny.Specifically, we asked whether K estimated for the actual tiparrangement differed from that generated based on 100,000randomized tip arrangements (Blomberg, Garland, and Ives2003) using the “picante” package in R (Kembel et al. 2010).Our estimate of K revealed a moderate phylogenetic auto-correlation that was significantly different from 0 (K p

, , ).0.195 Z p �1.58 P p .032Phylogenetic generalized least squares regression (PGLS)

allows for simultaneous consideration and estimation of thedegree of phylogenetic nonindependence using Pagel’s lambda(l). Allowing character evolution through modes other thanBrownian motion is one of the main advantages of PGLS overphylogenetically independent contrasts (PIC), another com-mon approach (Felsenstein 1985). Lambda represents a con-tinuous variable for which 0 describes a trait that displays nophylogenetic signal and 1 describes a trait that has evolvedunder Brownian motion. PGLS is equivalent to PIC for char-acters evolving under Brownian motion with completely re-solved phylogenies (Rohlf 2001).

We built general linear models in a PGLS framework usingthe “nlme” package in R with the composite cooperation scoreas the response variable and behavioral, ecological, morpho-logical, and life history variables as predictors. We allowed l

to take its Maximum Likelihood Estimate (MLE) in eachmodel. We sequentially entered and dropped all potential ex-planatory terms, including two-way interactions predicted byour hypotheses. We deemed the candidate model with thesmallest Akaike Information Criterion corrected (AICc) forsmall sample size to be the best while retaining any modelswith dAICc (difference between the AICc of the best modeland the model being considered) values of less than 2 asessentially equivalent (Burnham and Anderson 2002). We ob-tained statistics for terms removed from our best model byindividually adding each term to the minimal model. We onlyreport interaction terms that improved the fit of our bestmodel.

We also used phylogenetic generalized estimating equations(Paradis, Claude, and Strimmer 2004) with a binary responsevariable to estimate the effect of each predictor on the fivebinary cooperation variables, including the alloparenting var-iable for which animals were cooperative if they either allo-parented or shared a communal den. Because some MLEestimates of models failed to converge, we estimated univar-iate models for each predictor variable.

Factors Predicting Composite CooperationScores in Carnivora

The MLE for the strength of phylogenetic signal for the finalmodel (l) was 0.060, suggesting that more closely related

species display weakly similar residual errors. Our data setwas limited to and species, respectively, forN p 57 N p 46which data on longevity in the wild and mass-corrected basalmetabolic rates were available. We first inquired whether eachof these variables predicted our cooperation score. After cor-recting for phylogeny, neither longevity (b � SE p .139 �

, , ) nor mass-corrected basal metabolic.564 t p .246 P p .806rate ( , , ) significantlyb � SE p .098 � .521 t p .187 P p .852predicted cooperation across Carnivora. Therefore, we re-moved these predictors from our final analysis because doingso permitted us to test the remaining variables using the sta-tistical power of our full data set ( species).N p 87

We used AICc to recover our best model explaining vari-ation in the cooperation scores among carnivores (table 1).After correcting for the effects of phylogeny, the results fromour best model suggest that the greatest number of cooper-ative behaviors occurs among cursorial hunters in specieslacking strong male-biased size dimorphism that have largelitters and engage to some extent in hunting big game (table1). In addition, we have some evidence that species that aretall for their size (e.g., have large relative shoulder heightscompared with their mass) display a greater number of co-operative behaviors but only in species that exhibit cursorialhunting (table 1). The only additional model that is statis-tically indistinguishable from our best model (within 2 dAICc;Burnham and Anderson 2002) further suggests that speciesliving in open habitats engage in a greater number of co-operative behaviors than do those in dense habitats.

In general, the degree of social cohesion was an importantdeterminant of the cooperation score assigned to each species(table 2). We excluded solitary species from this analysis be-cause these animals had no group members with which tocooperate. Even after correcting for multiple testing (Storeyand Tibshirani 2003), fission-fusion and obligately gregariousspecies had significantly greater cooperation scores than didspecies living in pairs (table 2). Interestingly, our finding thatspecies with fission-fusion lifestyles were just as cooperativeas those species always found in cohesive social groups sug-gests that fission-fusion dynamics permit species to avoidcostly competition without sacrificing benefits accruing fromcooperation with group mates.

The overall tendency for carnivore species to cooperate wasgenerally low (cooperation score: ,mean � SE p 0.76 � 0.14range 0–5, species), but our results indicate strongN p 87variation both within and among families regarding coop-erative tendencies (fig. 3). The proportion of species thatengaged in each form of cooperation contributing to the co-operation score was also variable. A higher proportion ofspecies participated in alloparental care ( ), inter-0.20 � 0.04group contests ( ), and group hunting (0.18 � 0.04 0.18 �

) than in cooperative protection from predators0.04( ) or intragroup coalitions ( ; CA�0.13 � 0.04 0.07 � 0.03supplement A).

We found no evidence of cooperation among Ailuridae,Ursidae, or Viverridae (fig. 3). In contrast, most members of


Table 1. Best candidate model and the only model !2 dAICc of the best model explaining the composite cooperationscore across Carnivora

b SE t P AICc dAICc

Best candidate model:Intercept 1.719 .790 2.176 .032 247.2Relative shoulder height �.736 .515 �1.429 .157 0Cursorial hunting 1.691 .484 3.492 .001**Litter size (log) .488 .233 2.092 .040**Sexual dimorphism (log) �2.162 .935 �2.312 .023**Hunting of big game .462 .297 1.557 .123Relative shoulder height: cursorial hunting 4.340 2.414 1.798 .076**

Only model !2 dAICc of the best model:Intercept 1.441 .766 1.881 .064 249.1Relative shoulder height �.718 .478 �1.502 .137 1.9Cursorial hunting 1.701 .481 3.535 .001**Litter size (log) .419 .220 1.910 .060**Open vs. mixed habitat .265 .238 1.115 .268Open vs. closed habitat .763 .334 2.282 .025**Mixed vs. closed habitat .498 .303 1.641 .105Sexual dimorphism (log) �2.061 .916 �2.249 .027**Hunting of big game .420 .284 1.475 .144Relative shoulder height: cursorial hunting 4.237 2.425 1.747 .085**

Note. dAICc p difference between the AICc of the best model and the model being considered. Addition of these variables failed to improve thefit of our best candidate model: log neonate mass corrected for log adult mass; log brain size corrected for log adult mass, log mass, log durationof lactation, log home range size, log gestation length, and diet. In addition, none of these variables were significant at when added to thea ! .10best model. The addition of log population density to the best model did not significantly improve the model, resulting in a model with a

, but the variable was statistically significant when added ( , , ). Results of ANOVA for overall habitatdAICc 1 2 b p .125 � .061 t p 2.07 P p .042for model B are , .F p 3.05 P p .08462,80

* .a ! .10** (in bold).a ! .05

Table 2. Effects of sociality on composite cooperationscores across nonsocial species of Carnivora

Effect b SE t P

Intercept .759 .518 1.466 .154Pair bonding vs. fission-fusion 1.788 .597 2.995 .015**Pair bonding vs. obligately social 1.869 .679 2.754 .015**Fission-fusion vs. obligately social .082 .717 .114 .910

Note. Solitary species were excluded from this analysis because thesespecies had no group members with which to cooperate. P values arepresented in their corrected form to account for multiple comparisons.** (in bold).a ! .05

Canidae and Herpestidae investigated here were highly co-operative, engaging on average in at least one form of co-operation (fig. 3). In particular, African wild dogs engaged inall five types of cooperation. Lions, cheetahs, and snow leop-ards (Panthera uncia) represent interesting outliers becausethey are the only cooperative species in the family Felidae.Interestingly, although Hyaenidae had a low mean coopera-tion score ( , species), spotted hyenas were1.25 � 1.09 N p 4far more cooperative than any other members of their family(score p 4.5). Members of Eupleridae, Mephitidae, and Pro-cyonidae had low mean cooperation scores and high variationin the degree of cooperation among family members.

Factors Predicting Each Form of Cooperation in Carnivora

Next, we used univariate tests to consider the effects of eachof the predictor variables on each unique form of cooperation.Perhaps because cooperation among carnivores is generallyrare, some models failed to converge. Thus, those variablesnot explicitly stated here as either having a significant ornonsignificant effect on each form of cooperation failed tosuccessfully converge (table 3). Overall, the results of thesetests resembled the general patterns found for our compositecooperation scores; the tendency to cooperate generally in-creased with cursorial hunting of big game (table 3).

This analysis also revealed coevolution of particular traitswith each specific form of cooperation. For example, coop-

erative hunters produced larger litters and were less sexuallydimorphic than noncooperatively hunting species (table 3).Additional meat gained from cooperative hunting, therefore,appears to permit mothers to increase their investment incurrent reproductive effort by increasing offspring numberand also increasing their own body size relative to that ofmales. Similarly, mothers of species that cooperatively de-fended themselves from predators also invested more in cur-rent reproduction, weaning offspring at later ages than didmothers of noncooperative species. Species living in densepopulations were also the most likely to cooperatively defendthemselves from predators, presumably because cooperativedefense is mainly a numbers game, requiring a large numberof individuals to detect and cooperatively mob predators.


Figure 3. Cooperation scores among and within families in the order Carnivora. Dark horizontal lines in box plots represent medians,with boxes spanning the middle 50% of the data for each family group. Whiskers stretch to any values that are outside boxes butwithin 2.5 quartiles from the median. Families are ranked by mean cooperation scores from least cooperative to most cooperative(left to right). Outliers within each family are those values greater than 2.5 quartiles away from the median and are indicated bycircles.

Moreover, because predators often aggregate in areas of highprey density (known as the “pantry effect”; Ford and Pitelka1984), prey in dense populations are under strong selectionto evolve effective antipredator behavior. Interestingly, ourdata suggest that increased brain size might have coevolvedwith alloparenting, an important finding consistent with com-parative data on primates suggesting that alloparenting en-hances energy required to sustain expensive neural tissue (Islerand van Schaik 2012).

Evolutionary Losses and Gains of Cooperation in Carnivora

We next used each binary cooperation variable in an ancestorreconstruction using MLE to ask whether the data support agreater likelihood of cooperation arising or being lost acrossevolutionary transitions within Carnivora. Using the “ape”package in R, we estimated ancestral values for the commonancestors of the species present in our phylogeny. For everyancestral node, each signifying the ancestor of a pair of taxa,we estimated the probability that the last common ancestorat that node was cooperative.

Our ancestor reconstruction suggests that extant species ofcarnivores probably evolved from noncooperative ancestors(fig. 4). Furthermore, the same is generally true at the familylevel, with the most recent common ancestors of most familiesdisplaying very low probabilities for any cooperation with theexceptions of Canidae, Herpestidae, Hyaenidae, and Mephi-tidae. In particular, Felidae, Ursidae, Eupleridae, and Viver-ridae exhibit almost no suggestion of cooperative behavior inthe family’s ancestral state. Interestingly, cooperative speciesappear in every extant family except for Ailuridae, Ursidae,and Viverridae.

Insights into the Evolutionary Origins ofCooperation in Homo

Multiple Factors Coevolve with the Emergence of Cooperation

Overall, the results of comparative analysis revealed that mul-tiple factors are important correlates of cooperation in mam-malian carnivores. These data therefore suggest that mean-ingful links exist between cooperation and changes in


Table 3. Factors influencing the tendency to engage in eachof the five forms of cooperation across Carnivora

Factor b SE t P

Cooperative hunting:a

Hunting of big game 1.74 .63 2.75 .013**Litter size (log) 1.34 .59 2.26 .037**Sexual dimorphism �5.02 2.74 �1.83 .084*

Intragroup coalitions:b

Cursorial hunting 4.23 .99 4.26 .001**Hunting of big game 2.77 1.14 2.43 .026**Litter size (log) 1.82 .91 2.00 .060*Flesh diet 2.02 1.13 1.78 .092*

Territory defense:c

Cursorial hunting 1.57 .79 1.98 .064*Predator protection:d

Hunting of big game 1.87 .67 2.80 .012**Cursorial hunting 1.89 .80 2.40 .027**Population density (log) .33 .15 2.13 .047**Weaning age (log) .96 .55 1.74 .099*

Alloparenting (none vs.any form):e

Flesh diet 1.34 .50 2.68 .015**Hunting of big game 1.20 .53 2.27 .036**Relative brain size 1.55 .89 1.74 .099*

Note. Results from generalized estimating equations with each cooper-ation variable as a binary response for significant variables that converged.Each test given is from a univariate model.a for relative brain volume, flesh diet.P 1 .10b for sexual dimorphism, log gestation length, log weaning age,P 1 .10neonate mass corrected for adult mass, relative brain volume.c for log home range, sexual dimorphism, log litter size, logP 1 .10population density, log gestation length, relative shoulder height, neonatemass corrected for adult mass, relative brain volume, big game.d for log home range, sexual dimorphism, log litter size, logP 1 .10gestation length, relative shoulder height, neonate mass corrected foradult mass, relative brain volume, flesh diet.e for log-log home range, sexual dimorphism, log populationP 1 .10density, log gestation length, log weaning age, neonate mass correctedfor adult mass, cursorial hunting.* .a ! .10** (in bold).a ≤ .05

morphology (e.g., relative height at shoulder for body mass,reduced sexual dimorphism, increase in relative brain size),foraging and ranging behaviors (e.g., endurance hunting ofbig game in open landscapes, fission-fusion sociality), and lifehistory traits (e.g., increased age of weaning, increased re-productive output), all of which are relevant to early homi-nins. These findings extend earlier studies suggesting a pos-itive relationship between hunting large game in open habitatsand the emergence of sociality in carnivores (Gittleman andHarvey 1982; Packer, Scheel, and Pusey 1990). Early Homoapparently exhibited larger body and brain sizes and perhapsslower growth rates than did Australopithecus (Anton 2012).Our data suggest that meaningful inferences about social evo-lution are possible based on these morphological shifts.

Whereas each of the factors elucidated here for extant car-nivores likely played some role in hominin evolution, onlysome of them depend on meat eating, and none of them are

fully predicted by a singular referential model evaluated here.Our analysis underscores the need for paleoanthropologiststo consider a multitude of factors (rather than a single orrelatively small number of factors) simultaneously when at-tempting to explain the evolution of cooperation in hominins.This notion of a multifaceted approach to hominin evolutionhas been suggested previously (e.g., Potts 1994), but here weprovide new lines of evidence to support it. Briefly, as pre-dicted by the hunting-scavenging model involving persistencerunning, we found that carnivores engaging in cursorial hunt-ing of large-bodied prey are most likely to cooperate. More-over, morphological traits important for hunting, such as largerelative shoulder height, and those theorized to result fromreduced male-male competition (e.g., Plavcan 2012), such asa reduction in sexual dimorphism, were most common inhighly cooperative species. These morphological changes areof particular relevance because such factors may be evaluatedin the fossil remains of early Homo. Nevertheless, those speciesof carnivores that were most cooperative in breeding generallyhad larger relative brain sizes and greater reproductive in-vestment than did noncooperative carnivores. Relative brainsize failed to predict cooperative defense against predators,which does not support the prediction of the predator pro-tection model. Species that weaned offspring at the oldestages (enhanced reproductive investment) and lived at thehighest population densities (presumably in areas withclumped resources) were most likely to cooperate in defenseagainst predators.

Relevance of Extant Carnivores for UnderstandingBehavioral Shifts in Hominins

Large mammalian carnivores might have coexisted with Homoby as early as 2.6 Ma. However, tooth marks on bones suggestthat carnivores only started to regularly visit butchery sitesby roughly around 2.0 Ma (Domınguez-Rodrigo and Mar-tinez-Navarro 2012). Despite the surprisingly high numberof tooth marks on animal bones around this time, extrinsicmortality of early Homo apparently declined, a pattern thathas been attributed to either increased cooperative defenseagainst predators (Hart and Sussman 2009) or cooperativebreeding (Hrdy 2009). Both forms of cooperation occur inextant carnivores reviewed here. Evidence from carnivoresmore broadly indicates that both the emergence of cooperativedefense against predators and the pace of reproduction canrespond in a flexible fashion to variation in the availabilityand acquisition of energy-rich foods. Thus, it is possible thatsimilar flexibility influenced shifts in reproductive investmentand rates of reproduction among early hominins. Interest-ingly, just as flexibility in female reproduction among spottedhyenas exceeds that of most extant carnivores, the plasticreproductive responses typical of early Homo appear to sur-pass those of chimpanzees or gorillas (Bribiescas, Ellison, andGray 2012; Isler and van Schaik 2012; Wells 2012). Thus,maternal capital might have played a central role in hominin

Figure 4. Phylogeny of the carnivore species used in our study. Tip labels display species names. At tips of the phylogeny are circlesshaded to varying degrees indicating for each cooperation variable whether or not the species exhibits the trait. Black indicates traitpresence, whereas white indicates that the species does not exhibit that trait. For alloparenting, gray indicates communal denning.Circles representing composite cooperation scores (a continuous measure of cooperation) are shaded, with dark shading representingthe highest cooperation score. Pie charts are given for each binary variable for select taxonomic groups indicating the probabilitiesthat common ancestors were uncooperative (white) or cooperative (black). Superscripts for each taxonomic grouping refer to labelednodes on the phylogeny.


evolution of large brains and slow life histories, although theprecise timing of each remains enigmatic.

Around 2.3–1.7 Ma, a major shift in a suite of behaviorsbecame persistent in the fossil record of early Homo (Potts1998a, 2012); these included stone transport, tool making,and access to large animals. In the Turkana and Olduvai ba-sins, stone transport distances increased, artifacts were dis-tributed more widely, and processing of animal tissues inten-sified, including the extraction of meat and marrow from largeanimals (e.g., Blumenschine 1995; Bunn and Kroll 1986; Potts1988). Although no evidence currently exists of tool useamong extant carnivores, most carnivores are well endowedwith massive jaws that permit them to capture prey and gainaccess to animal tissues without tool use. Nonetheless, mostmammalian carnivores are unable to capture prey exceeding10 kg (Gittleman 1989). Those that capture large prey huntcooperatively when doing so, but there is no evidence ofadvanced planning before hunts by carnivores. In contrast,modern hunts by the Ache of Paraguay (Kaplan and Hill 1985)and subsistence whale hunters of Lamalera, Indonesia (Alvardand Nolin 2002; O’Connell, Hawkes, and Blurton Jones 1988),regularly capture prey of much larger than 10 kg in huntsrequiring advanced planning. Recent comparative data sug-gest that cumulative culture, the summation of innovationsover time, may indeed be unique to modern humans (Deanet al. 2012). Altogether, these data suggest that cooperativecapture of some prey items might have been possible beforethe evolution of large brains, but complex forms of coop-erative hunting requiring advanced planning probablyemerged later in hominin evolution.

Around 1.9–1.5 Ma, landscape instability likely promotedcarrying of stones and meat over greater distances (e.g., 2–13 km; Potts 1998a). Selection favoring other behavioral traits,including sociality, of early hominins was also likely drivenby intense variation in ecological and climatic conditions(Potts 1998b, 2012). Similarly, spotted hyenas, the most abun-dant large carnivore in sub-Saharan Africa, may have alsobeen subject to strong selection for their behavioral flexibilityto cope with demands of life in a socially and ecologicallydynamic landscape (Holekamp and Dloniak 2010; Holekampet al. 2012). The ecological dominance of spotted hyenas overother carnivores in Africa may in large part be attributed tothe behavioral flexibility that their impressive morphologyaffords them (Holekamp and Dloniak 2010). Adult spottedhyenas are efficient hunters (Holekamp et al. 1997) and ex-tractive foragers (Tanner et al. 2010) capable of fully exploitinga wide array of foods ranging from termites to large ungulateprey. These hyenas effectively crack through bones as large asgiraffe leg bones to access marrow, allowing them to efficientlyconsume entire carcasses. In the face of burgeoning humanpopulations across Africa, their behavioral flexibility has per-mitted these animals to persist at high densities despite theenergetic demands of being a top predator (Boydston et al.2003). It might be this behavioral flexibility with respect toforaging that permits spotted hyenas to cope with ecological

insults more effectively than can most other large carnivoreswith more restricted dispersal abilities or less versatile mor-phological adaptations (Holekamp et al. 2012). In this respect,data on spotted hyenas are consistent with the notion thatnatural selection favors species living in variable habitats thatare best able to cope with and respond to changing environ-mental conditions.

Fission-fusion sociality may have played an underappre-ciated role in reducing the costs confronted by early homininsin changing environments. That is, hominin communitiesmight have retained their ability to cooperate with groupmates despite their splitting into more and more complexlevels of temporary subgroups as the total area required pergroup increased (e.g., from those occupied by Australopithecusand Homo habilis to that occupied by Homo erectus; reviewedby Grove, Pearce, and Dunbar 2012). Our comparative dataare consistent with the notion that fission-fusion socialitypermits the maintenance of cooperation; extant carnivoresliving in societies structured by fission-fusion dynamics en-gaged in just as many forms of cooperation as species in highlycohesive groups and were more cooperative than those re-stricted to living in pairs. Despite spending much of theirtime apart from group members, most gregarious carnivoresregularly meet up with conspecifics for protection from pred-ators, reinforcement of social bonds, and sharing of spoilsfrom hunts. Evolution of an increasingly complex multilevelfission-fusion society beyond that observed in any extantmammalian carnivore may have similarly helped early hom-inins cope with increased foraging demands attributed toranging over large areas as they migrated toward high latitudes(Grove, Pearce, and Dunbar 2012). Taken together, evidencefrom extant mammalian carnivores reviewed here as well asthat from nonhuman primates (e.g., Wrangham, Gittleman,and Chapman 1993) and modern small-scale hunter-gathers(e.g., Marlowe 2005) suggests the most parsimonious inter-pretation of the fossil record is that development toward anincreasingly complex multilevel fission-fusion society allowedfor both cooperation and range expansion at key transitionalstages of hominin evolution.

In conclusion, it has long been recognized that understand-ing the evolution and mechanisms of cooperation amongmammalian carnivores might shed light on the factors shapinghominin evolution (Hill 1982; Kaplan and Hill 1985; Schallerand Lowther 1969). Our study confirms that meaningful linksare possible between measurable morphological traits andseemingly elusive behavioral traits. New analyses on coop-erative breeding and predator protection for these species,some but not all of which also happen to eat meat, make thistaxonomic group relevant today even if Homo was not sub-stantially carnivorous. Our results also move the field forwardby emphasizing the need for the consideration of multiplefactors rather than predictions from a single existing modelwhen explaining social evolution of early hominins. New in-sights suggesting convergent evolution between extant mam-malian carnivores and early hominins will undoubtedly


emerge as we continue to learn more about the biology ofextant mammalian carnivores.

Acknowledgments

We thank Leslie Aiello and Susan Anton for their invitationto participate in the symposium that led to this paper and tothe other participants for useful conversations. Commentsfrom Rick Potts were particularly instrumental in this process.We are grateful to Laurie Obbink for helping to organize thismeeting. We thank John Finarelli for assisting us with ac-quiring an accurate phylogeny and for his helpful suggestionsfor pruning it for our use here. We are also grateful to ourcolleagues at the University of California, Los Angeles (UCLA)Center for Behavior, Evolution, and Culture and those par-ticipating in the Modeling Social Complexity InvestigativeWorkshop at the National Institute for Mathematical and Bi-ological Synthesis (NIMBios; sponsored through National Sci-ence Foundation (NSF) Award EF-08325858 and the Uni-versity of Tennessee, Knoxville) for rich discussions on thistopic. This research was funded by postdoctoral fellowshipsfrom the American Association of University Women(AAUW) and the UCLA Institute for Society and Genetics toJ. E. Smith; an NSF Predoctoral Fellowship and UniversityDistinguished Fellowship from Michigan State University toE. M. Swanson; and NSF grants IOB0618022, IOS0819437,and IOS1121474 to K. E. Holekamp. The material in thispaper is also based in part on work supported by the NSFunder cooperative agreement DBI-0939454.

References CitedAiello, Leslie C., and Peter Wheeler. 1995. The expensive tissue hypothesis:


Alvard, Michael S., and David A. Nolin. 2002. Rousseau’s whale hunt? co-ordination among big-game hunters. Current Anthropology 43:533–559.


Aureli, Filippo, Colleen M. Schaffner, Christophe Boesch, Simon K. Bearder,Josep Call, Colin A. Chapman, Richard Connor, et al. 2008. Fission-fusiondynamics: new research frameworks. Current Anthropology 49:627–654.

Bingham, Paul M. 1999. Human uniqueness: a general theory. Quarterly Re-view of Biology 74:133–169.

Bininda-Emonds, Olaf R. P., Marcel Cardillo, Kate E. Jones, Ross D. E.MacPhee, Robin M. D. Beck, Richard Grenyer, Samantha A. Price, RutgerA. Vos, John L. Gittleman, and Andy Purvis. 2007. The delayed rise ofpresent-day mammals. Nature 446:507–512.

Blomberg, Simon P., Theodore Garland Jr., and Anthony R. Ives. 2003. Testingfor phylogenetic signal in comparative data: behavioral traits are more labile.Evolution 57:717–745.

Blumenschine, Robert J. 1995. Percussion marks, tooth marks, and experi-mental determinations of the timing of hominid and carnivore access tolong bones at FLK Zinjanthropus, Olduvai Gorge, Tanzania. Journal ofHuman Evolution 29:21–51.

Boydston, Erin E., Karen M. Kapheim, Heather E. Watts, Micaela Szykman,and Kay E. Holekamp. 2003. Altered behaviour in spotted hyenas associatedwith increased human activity. Animal Conservation 6:207–219.

Boydston, Erin E., Toni Lyn Morelli, and Kay E. Holekamp. 2001. Sex dif-ferences in territorial behavior exhibited by the spotted hyena (Hyaenidae,Crocuta crocuta). Ethology 107:369–385.



Brown, Jerram L. 1983. Cooperation: a biologist’s dilemma. Advances in theStudy of Behavior 13:1–37.

Bunn, Henry T. 2006. Meat made us human. In Evolution of the human diet:the known, the unknown, and the unknowable. Peter Ungar, ed. Pp. 191–211. Oxford: Oxford University Press.

Bunn, Henry T., and Ellen M. Kroll. 1986. Systematic butchery by Plio-Pleistocene hominids at Olduvai Gorge, Tanzania. Current Anthropology 27:431–452.

Bunn, Henry T., and Travis R. Pickering. 2010. Bovid mortality profiles inpaleoecological context falsify hypotheses of endurance running-huntingand passive scavenging by early Pleistocene hominins. Quaternary Research74:395–404.

Burnham, Kenneth P., and David R. Anderson. 2002. Model selection andinference: a practical information-theoretic approach. 2nd edition. New York:Springer.

Byrne, Richard R. 1995. The thinking ape: evolutionary origins of intelligence.Oxford: Oxford University Press.

Cant, Michael A. 2011. The role of threats in animal cooperation. Proceedingsof the Royal Society B: Biological Sciences 278:170–178.

Cant, Michael A., Sarah J. Hodge, Matthew B. V. Bell, Jason S. Gilchrist, andHazel J. Nichols. 2010. Reproductive control via eviction (but not the threatof eviction) in banded mongooses. Proceedings of the Royal Society B: Bi-ological Sciences 277:2219–2226.

Carbone, Chris, Lory Frame, George Frame, James R. Malcolm, John H.Fanshawe, Clare D. FitzGibbon, George B. Schaller, Iain J. Gordon, J.Marcus Rowcliffe, and Johan T. du Toit. 2005. Feeding success of Africanwild dogs (Lycaon pictus) in the Serengeti: the effects of group size andkleptoparasitism. Journal of Zoology 266:153–161.

Chapais, Bernard. 1992. The role of alliances in the social inheritance of rankamong female primates. In Coalitions and alliances in humans and otheranimals. Alexander H. Harcourt and Frans B. M. de Waal, eds. Pp. 29–60.Oxford: Oxford Science.

Cheney, Dorothy L. 1977. The acquisition of rank and the development ofreciprocal alliances among free-ranging immature baboons. Behavioral Ecol-ogy and Sociobiology 2:303–318.

Clutton-Brock, Timothy H. 2002. Breeding together: kin selection and mu-tualism in cooperative vertebrates. Science 296:69–72.

———. 2009a. Cooperation between non-kin in animal societies. Nature 462:51–57.

———. 2009b. Structure and function in mammalian societies. PhilosophicalTransactions of the Royal Society B: Biological Sciences 364:3229–3242.

Clutton-Brock, Timothy H., and Paul H. Harvey. 1977. Primate ecology andsocial organization. Journal of Zoology 183:1–39.

Clutton-Brock, Timothy H., Sarah J. Hodge, and T. P. Flower. 2008. Groupsize and the suppression of subordinate reproduction in Kalahari meerkats.Animal Behaviour 76:689–700.

Connor, Richard C. 1995. The benefits of mutualism: a conceptual framework.Biological Reviews of the Cambridge Philosophical Society 70:427–457.

Courchamp, Franck, Gregory S. A. Rasmussen, and David W. Macdonald.2002. Small pack size imposes a trade-off between hunting and pup-guard-ing in the painted hunting dog Lycaon pictus. Behavioral Ecology 13:20–27.

Creel, Scott. 1997. Cooperative hunting and group size: assumptions andcurrencies. Animal Behaviour 54:1319–1324.

Creel, Scott, and Nancy M. Creel. 1991. Energetics, reproductive suppressionand obligate communal breeding in carnivores. Behavioral Ecology andSociobiology 28:263–270.

———. 1995. Communal hunting and pack size in African wild dogs, Lycaonpictus. Animal Behaviour 50:1325–1339.

———. 2002. The African wild dog: behavior, ecology, and conservation. Mono-graphs in Behavior and Ecology. Princeton, NJ: Princeton University Press.

Creel, Scott, and David Macdonald. 1995. Sociality, group size, and repro-ductive suppression among carnivores. Advances in the Study of Behavior24:203–257.

Dalerum, Fredrik. 2007. Phylogenetic reconstruction of carnivore social or-ganizations. Journal of Zoology 273:90–97.

Dean, Lewis G., Rachel L. Kendal, Steven J. Schapiro, Bernard Thierry, andKevin N. Laland. 2012. Identification of the social and cognitive processesunderlying human cumulative culture. Science 335:1114–1118.


Domınguez-Rodrigo, Manuel, and Bienvenido Martinez-Navarro. 2012. Taph-onomic analysis of the early Pleistocene (2.4 Ma) faunal assemblage fromA.L. 894 (Hadar, Ethiopia). Journal of Human Evolution 62:315–327.

Domınguez-Rodrigo, Manuel, and Travis R. Pickering. 2003. Early hominidhunting and scavenging: a zooarcheological review. Evolutionary Anthro-pology 12:275–282.

Domınguez-Rodrigo, Manuel, Travis R. Pickering, and Henry T. Bunn. 2010.Configurational approach to identifying the earliest hominin butchers. Pro-ceedings of the National Academy of Sciences of the USA 107:20929–20934.

———. 2011. Reply to McPherron et al.: doubting Dikika is about data, notparadigms. Proceedings of the National Academy of Sciences of the USA 108:E117.

Drea, Christine M., and Allisa N. Carter. 2009. Cooperative problem solvingin a social carnivore. Animal Behaviour 78:967–977.

Drea, Christine M., and Laurence G. Frank. 2003. The social complexity ofspotted hyenas. In Animal social complexity: intelligence, culture, and indi-vidualized societies. Frans B. M. de Waal and Peter L. Tyack, eds. Pp. 121–148. Cambridge, MA: Harvard University Press.

Dugatkin, Lee A. 2002. Animal cooperation among unrelated individuals.Naturwissenschaften 89:533–541.

East, Marion L., Heribert Hofer, and Wolfgang Wickler. 1993. The erect“penis” as a flag of submission in a female-dominated society: greetings inSerengeti spotted hyenas. Behavioral Ecology and Sociobiology 33:355–370.

Engh, Anne L., Katrina Esch, Laura Smale, and Kay E. Holekamp. 2000.Mechanisms of maternal rank “inheritance” in the spotted hyaena, Crocutacrocuta. Animal Behaviour 60:323–332.

Engh, Anne L., Erin R. Siebert, David A. Greenberg, and Kay E. Holekamp.2005. Patterns of alliance formation and postconflict aggression indicatespotted hyaenas recognize third-party relationships. Animal Behaviour 69:209–217.

Felsenstein, Joseph. 1985. Phylogenies and the comparative method. AmericanNaturalist 125:1–15.

Finarelli, John A. 2008. A total evidence phylogeny of the Arctoidea (Car-nivora: Mammalia): relationships among basal taxa. Journal of MammalianEvolution 15:231–259.

———. 2010. Does encephalization correlate with life history or metabolicrate in Carnivora? Biology Letters 6:350–353.

Finarelli, John A., and John J. Flynn. 2009. Brain-size evolution and socialityin Carnivora. Proceedings of the National Academy of Sciences of the USA106:9345–9349.

Flynn, John J., John A. Finarelli, Sarah Zehr, Johnny Hsu, and Michael A.Nedbal. 2005. Molecular phylogeny of the Carnivora (Mammalia): assessingthe impact of increased sampling on resolving enigmatic relationships. Sys-tematic Biology 54:317–337.

Ford, R. Glenn, and Frank A. Pitelka. 1984. Resource limitation in populationsof the California vole. Ecology 65:122–136.

Frank, Laurence G. 1986. Social organization of the spotted hyaena (Crocutacrocuta). 2. Dominance and reproduction. Animal Behaviour 34:1510–1527.

Frank, Laurence G., Kay E. Holekamp, and Laura Smale. 1995. Dominance,demography, and reproductive success of female spotted hyenas. In SerengetiII: dynamics, management, and conservation of an ecosystem. Anthony R. E.Sinclair and Peter Arcese, eds. Pp. 364–384. Chicago: University of ChicagoPress.

Fuentes, Agustın, Matthew A. Wyczalkowski, and Katherine C. MacKinnon.2010. Niche construction through cooperation: a nonlinear dynamics con-tribution to modeling resilience and evolutionary history in the genusHomo. Current Anthropology 51(3):435–444.

Gittleman, John L. 1989. Carnivore behavior, ecology and evolution. Ithaca, NY:Cornell University Press.

Gittleman, John L., and Paul H. Harvey. 1982. Carnivore home-range size,metabolic needs and ecology. Behavioral Ecology and Sociobiology 10:57–63.

Gittleman, John L., and Hang-Kwang Luh. 1992. On comparing comparativemethods. Annual Review of Ecology and Systematics 23:383–404.

Goldberg, Tony L., and Richard W. Wrangham. 1997. Genetic correlates ofsocial behaviour in wild chimpanzees: evidence from mitochondrial DNA.Animal Behaviour 54:559–570.

Gompper, Matthew E. 1996. Sociality and asociality in white-nosed coatis(Nasua narica): foraging costs and benefits. Behavioral Ecology 7:254–263.

Gompper, Matthew E., John L. Gittleman, and Robert K. Wayne. 1998. Dis-persal, philopatry, and genetic relatedness in a social carnivore: comparingmales and females. Molecular Ecology 7:157–163.

Gottelli, Dada, Claudia Sillero-Zubiri, Geoffrey D. Applebaum, Michael S.Roy, Derek J. Girman, Jesus Garcia-Moreno, Elaine A. Ostrander, and Rob-

ert K. Wayne. 1994. Molecular genetics of the most endangered canid: theEthiopian wolf Canis simensis. Molecular Ecology 3:301–312.

Grine, Frederick, and John G. Fleagle. 2009. The first humans: a summaryperspective on the origin and early evolution of the genus Homo. In Thefirst humans: origin and early evolution of the genus Homo. Fred E. Grine,John G. Fleagle, and Richard E. Leakey, eds. Pp. 197–207. New York:Springer.

Grove, Matt. 2010. Logistical mobility reduces subsistence risk in huntingeconomies. Journal of Archaeological Science 37:1913–1921.

Grove, Matt, Eiluned Pearce, and Robin I. M. Dunbar. 2012. Fission-fusionand the evolution of hominin social systems. Journal of Human Evolution62:191–200.

Hamilton, William D. 1964a. The genetical evolution of social behavior. 1.Journal of Theoretical Biology 7:1–16.

———. 1964b. The genetical evolution of social behavior. 2. Journal of The-oretical Biology 7:17–52.

Hart, Donna, and Robert W. Sussman. 2009. Man the hunted: primates, pred-ators, and human evolution. Expanded edition. Boulder, CO: Westview.

Heinsohn, Robert, and Craig Packer. 1995. Complex cooperative strategies ingroup-territorial African lions. Science 269:1260–1262.

Henschel, Johannes R., and John D. Skinner. 1991. Territorial behaviour bya clan of spotted hyaenas Crocuta crocuta. Ethology 88:223–235.

Hill, Kim. 1982. Hunting and human evolution. Journal of Human Evolution11:521–544.

Hofer, Heribert, and Marion L. East. 1993. The commuting system of Serengetispotted hyaenas: how a predator copes with migratory prey. 2. Intrusionpressure and commuter’s space use. Animal Behaviour 46:559–574.

———. 1996. The components of parental care and their fitness consequences:a life history perspective. Verhandlungen der Deutschen Zoologischen Ge-sellscaft 89:149–164.

———. 2003. Behavioral processes and costs of co-existence in female spottedhyenas: a life history perspective. Evolutionary Ecology 17:315–331.

Holekamp, Kay E., Erin E. Boydston, and Laura Smale. 2000. Group travelin social carnivores. In On the move: how and why animals travel in groups.Sue Boinski and Paul A. Garber, eds. Pp. 587–627. Chicago: University ofChicago Press.

Holekamp, Kay E., and Stephanie M. Dloniak. 2010. Intraspecific variationin the behavioral ecology of a tropical carnivore, the spotted hyena. Advancesin the Study of Behavior 42:189–229.

Holekamp, Kay E., Sharleen T. Sakai, and Barbara L. Lundrigan. 2007. Socialintelligence in the spotted hyena (Crocuta crocuta). Philosophical Transac-tions of the Royal Society B: Biological Sciences 362:523–538.

Holekamp, Kay E., and Laura Smale. 1991. Dominance acquisition duringmammalian social development: the “inheritance” of maternal rank. Amer-ican Zoologist 31:306–317.

———. 1993. Ontogeny of dominance in free-living spotted hyaenas: juvenilerank relations with other immature individuals. Animal Behaviour 46:451–466.

Holekamp, Kay E., Laura Smale, Richard J. Berg, and Susan M. Cooper. 1997.Hunting rates and hunting success in the spotted hyena (Crocuta crocuta).Journal of Zoology (London) 242:1–15.

Holekamp, Kay E., Laura Smale, and Micaela Szykman. 1996. Rank and re-production in the female spotted hyaena. Journal of Reproduction and Fer-tility 108:229–237.

Holekamp, Kay E., Jennifer E. Smith, Christopher C. Strelioff, Russell C. VanHorn, and Heather E. Watts. 2012. Society, demography and genetic struc-ture in the spotted hyena. Molecular Ecology 21:613–632.

Horrocks, Julia, and Wayne Hunte. 1983. Maternal rank and offspring rankin vervet monkeys: an appraisal of the mechanisms of rank acquisition.Animal Behaviour 31:772–782.

Hrdy, Sarah B. 2009. Mothers and others: the evolutionary origins of mutualunderstanding. Cambridge, MA: Harvard University Press.


Johnson, Dominic D. P., Roland Kays, Paul G. Blackwell, and David W. Mac-donald. 2002. Does the resource dispersion hypothesis explain group living?Trends in Ecology and Evolution 17:563–570.

Johnson, Warren E., Eduardo Eizirik, Jill Pecon-Slattery, William J. Murphy,Agostinho Antunes, Emma Teeling, and Stephen J. O’Brien. 2006. The LateMiocene radiation of modern Felidae: a genetic assessment. Science 311:73–77.


Kaplan, Hillard, and Kim Hill. 1985. Food sharing among Ache foragers: testsof explanatory hypotheses. Current Anthropology 26:223–246.

Kaplan, Hillard, Kim Hill, Jane Lancaster, and A. Magdalena Hurtado. 2000.A theory of human life history evolution: diet, intelligence, and longevity.Evolutionary Anthropology 9:156–185.

Kays, Roland W., and John L. Gittleman. 2001. The social organization of thekinkajou Potos flavus (Procyonidae). Journal of Zoology (London) 253:491–504.

Kembel, Steven W., Peter D. Cowan, Matthew R. Helmus, William K. Corn-well, Helene Morlon, David D. Ackerly, Simon P. Blomberg, and CampbellO. Webb. 2010. Picante: R tools for integrating phylogenies and ecology.Bioinformatics 26:1463–1464.

Koepfli, Klaus-Peter, Kerry A. Deere, Graham J. Slater, Colleen Begg, KeithBegg, Lon Grassman, Mauro Lucherini, Geraldine Veron, and Robert K.Wayne. 2008. Multigene phylogeny of the Mustelidae: resolving relation-ships, tempo and biogeographic history of a mammalian adaptive radiation.BMC Biology 6:10.

Koepfli, Klaus-Peter, Matthew E. Gompper, Eduardo Eizirik, Cheuk-ChungHo, Leif Linden, Jesus E. Maldonado, and Robert K. Wayne. 2007. Phy-logeny of the Procyonidae (Mammalia: Carnivora): molecules, morphologyand the Great American Interchange. Molecular Phylogenetics and Evolution43:1076–1095.

Koepfli, Klaus-Peter, Susan M. Jenks, Eduardo Eizirik, Tannaz Zahirpour,Blaire Van Valkenburgh, and Robert K. Wayne. 2006. Molecular systematicsof the Hyaenidae: relationships of a relictual lineage resolved by a molecularsupermatrix. Molecular Phylogenetics and Evolution 38:603–620.

Kruuk, Hans. 1972. The spotted hyena: a study of predation and social behavior.Chicago: University of Chicago Press.

———. 1976. Feeding and social behaviour of the striped hyaena (Hyaenavulgaris Desmarest). East African Wildlife Journal 14:91–111.

Kruuk, Hans, and Tim Parish. 1982. Factors affecting population density,group size and territory size of the European badger, Meles meles. Journalof Zoology 196:31–39.

Lehman, Niles, Peter Clarkson, L. David Mech, Thomas J. Meier, and RobertK. Wayne. 1992. A study of the genetic relationships within and amongwolf packs using DNA fingerprinting and mitochondrial DNA. BehavioralEcology and Sociobiology 30:83–94.

Marlowe, Frank W. 2005. Hunter-gatherers and human evolution. Evolution-ary Anthropology 14:54–67.


Mech, L. David. 1970. The wolf: ecology and behavior of an endangered species.Minneapolis: University of Minnesota Press.

Melis, Alicia P., and Dirk Semmann. 2010. How is human cooperation dif-ferent? Philosophical Transactions of the Royal Society B: Biological Sciences365:2663–2674.

Mills, Gus L. 1990. Kalahari hyaenas: comparative behavioral ecology of twospecies. London: Unwin Hyman.

Milton, Katharine. 1999. A hypothesis to explain the role of meat-eating inhuman evolution. Evolutionary Anthropology 8:11–21.

Noe, Roland. 2006. Cooperation experiments: coordination through com-munication versus acting apart together. Animal Behaviour 71:1–18.

Nowak, Martin A. 2006. Five rules for the evolution of cooperation. Science314:1560–1563.

O’Connell, James F., Kristen Hawkes, and Nicholas Blurton Jones. 1988. Hadzascavenging: implications for Plio/Pleistocene hominid subsistence. CurrentAnthropology 29:356–363.

———. 1999. Grandmothering and the evolution of Homo erectus. Journalof Human Evolution 36:461–485.

O’Connell, James F., Kristen Hawkes, Karen D. Lupo, and Nicholas BlurtonJones. 2002. Male strategies and Plio-Pleistocene archaeology. Journal ofHuman Evolution 43:831–872.

Packer, Craig, Anne E. Pusey, and Lynn E. Eberly. 2001. Egalitarianism infemale African lions. Science 293:690–693.

Packer, Craig, and Lore Ruttan. 1988. The evolution of cooperative hunting.American Naturalist 132:159–198.

Packer, Craig, David Scheel, and Anne E. Pusey. 1990. Why lions form groups:food is not enough. American Naturalist 136:1–19.

Palomares, Francisco, and Tim M. Caro. 1999. Interspecific killing amongmammalian carnivores. American Naturalist 153:492–507.

Paradis, Emmanuel, Julien Claude, and Krbinian Strimmer. 2004. APE: anal-

yses of phylogenetics and evolution in R language. Bioinformatics 20:289–290.

Patoua, Marie-Lilith, Patricia A. Mclenachan, Craig G. Morley, Arnaud Cou-loux, Andrew P. Jennings, and Geraldine Verona. 2009. Molecular phylogenyof the Herpestidae (Mammalia, Carnivora) with a special emphasis on theAsian Herpestes. Molecular Phylogenetics and Evolution 53:69–80.

Perini, Fernando A., Claudia M. Russo, and Carlos G. Schrago. 2009. Theevolution of South American endemic canids: a history of rapid diversifi-cation and morphological parallelism. Journal of Evolutionary Biology 23:311–322.



Potts, Richard. 1988. Early hominid activities at Olduvai. New York: Aldine.———. 1994. Variables versus models of early Pleistocene hominid land-use.

Journal of Human Evolution 27:7–24.———. 1998a. Environmental hypotheses of hominin evolution. Yearbook of

Physical Anthropology 41:93–136.———. 1998b. Variability selection in hominid evolution. Evolutionary An-

thropology 7:81–96.———. 2012. Environmental and behavioral evidence pertaining to the evo-

lution of early Homo. Current Anthropology 53(suppl. 6):S299–S317.Potts, Richard, and Pat Shipman. 1981. Cutmarks made by stone tools on

bones from Olduvai Gorge, Tanzania. Nature 291:577–580.Queller, David C. 1985. Kinship, reciprocity and synergism in the evolution

of social behaviour. Nature 318:366–367.R Development Core Team. 2010. R: a language and environment for statistical

computing. Vienna: R Foundation for Statistical Computing.Reynolds, Vernon. 1966. Open groups in hominid evolution. Man, n.s., 1:

441–452.Rohlf, F. James. 2001. Comparative methods for the analysis of continuous

variables: geometric interpretations. Evolution 55:2143–2160.Sato, Jun J., Mieczyslaw Wolsan, Shinji Minami, Tetsuji Hosoda, Martua H.

Sinaga, Kozue Hiyama, Yasunori Yamaguchi, and Hitoshi Suzuki. 2009.Deciphering and dating the red panda’s ancestry and early adaptive radi-ation of Musteloidea. Molecular Phylogenetics and Evolution 53:907–922.

Schaller, George B. 1972. The Serengeti lion. Chicago: University of ChicagoPress.

Schaller, George B., and Gordon R. Lowther. 1969. Relevance of carnivorebehavior to study of early hominids. Southwestern Journal of Anthropology25:307–341.

Scheel, David, and Craig Packer. 1991. Group hunting behaviour of lions: asearch for cooperation. Animal Behaviour 41:697–709.

Semaw, Sileshi, Paul Renne, John W. K. Harris, Craig S. Feibel, Raymond L.Bernor, N. Fesseha, and Kenneth Mowbray. 1997. 2.5-million-year-old stonetools from Gona, Ethiopia. Nature 385:333–336.

Shipman, Pat. 1986. Scavenging or hunting in early hominids: theoreticalframework and tests. American Anthropologist 88:27–43.

Silk, Joan B., Susan C. Alberts, and Jeanne Altmann. 2003. Social bonds offemale baboons enhance infant survival. Science 302:1231–1234.

Silk, Joan B., Jacinta C. Beehner, Thore J. Bergman, Catherine Crockford,Anne L. Engh, Liza R. Moscovice, Roman M. Wittig, Robert M. Seyfarth,and Dorothy L. Cheney. 2009. The benefits of social capital: close socialbonds among female baboons enhance offspring survival. Proceedings of theRoyal Society B: Biological Sciences 276:3099–3104.

Smale, Laura, Laurence G. Frank, and Kay E. Holekamp. 1993. Ontogeny ofdominance in free-living spotted hyaenas: juvenile rank relations with adultfemales and immigrant males. Animal Behaviour 46:467–477.

Smith, Jennifer E., Joseph M. Kolowski, Katharine E. Graham, Stephanie E.Dawes, and Kay E. Holekamp. 2008. Social and ecological determinants offission-fusion dynamics in the spotted hyaena. Animal Behaviour 76:619–636.

Smith, Jennifer E., Sandra K. Memenis, and Kay E. Holekamp. 2007. Rank-related partner choice in the fission-fusion society of spotted hyenas (Cro-cuta crocuta). Behavioral Ecology and Sociobiology 61:753–765.

Smith, Jennifer E., Katherine S. Powning, Stephanie E. Dawes, Jillian R. Es-trada, Adrienne L. Hopper, Stacey L. Piotrowski, and Kay Holekamp. 2011.Greetings promote cooperation and reinforce social bonds among spottedhyaenas. Animal Behaviour 81:401–415.

Smith, Jennifer E., Russell C. Van Horn, Katherine S. Powning, Alison R.Cole, Katharine E. Graham, Sandra K. Memenis, and Kay E. Holekamp.


2010. Evolutionary forces favoring intragroup coalitions among spottedhyenas and other animals. Behavioral Ecology 21:284–303.

Stander, Philip E. 1992. Cooperative hunting in lions: the role of the individual.Behavioral Ecology and Sociobiology 29:445–454.

Storey, John D., and Robert Tibshirani. 2003. Statistical significance for ge-nomewide studies. Proceedings of the National Academy of Sciences of theUSA 100:9440–9445.

Swanson, Brook O., Todd A. Blackledge, Adam P. Summers, and Cheryl Y.Hayashi. 2006. Spider dragline silk: correlated and mosaic evolution in high-performance biological materials. Evolution 60:2539–2551.

Swanson, Eli M., Ian Dworkin, and Kay E. Holekamp. 2011. Lifetime selectionon a hypoallometric size trait in the spotted hyena. Proceedings of the RoyalSociety B: Biological Sciences 278:3277–3285.

Tanner, Jaime B., Miriam L. Zelditch, Barbara L. Lundrigan, and Kay E.Holekamp. 2010. Ontogenetic change in skull morphology and mechanicaladvantage in the spotted hyena (Crocuta crocuta). Journal of Morphology271:353–365.

Tilson, Ronald T., and William J. I. Hamilton. 1984. Social dominance andfeeding patterns of spotted hyaenas. Animal Behaviour 32:715–724.

Tomasello, Michael, Malinda Carpenter, Josep Call, Tanya Behne, and HenrikeMoll. 2005. Understanding and sharing intentions: the origins of culturalcognition. Behavioral and Brain Sciences 28:675–735.

Trivers, Robert L. 1971. The evolution of reciprocal altruism. Quarterly Reviewof Biology 46:35–57.


Van Horn, Russell C., Anne L. Engh, Kim T. Scribner, Stephan M. Funk, andKay E. Holekamp. 2004. Behavioural structuring of relatedness in the spot-ted hyena (Crocuta crocuta) suggests direct fitness benefits of clan-levelcooperation. Molecular Ecology 13:449–458.

Venkataraman, Arun B., R. Arumugam, and R. Sukumar. 1995. The foragingecology of dhole (Cuon alpinus) in Mudumalai Sanctuary, southern India.Journal of Zoology (London) 237:543–561.

Wagner, Aaron P. 2006. Behavioral ecology of the striped hyaena (Hyaenahyaena). PhD dissertation, Montana State University.

Walters, Jeffrey. 1980. Interventions and the development of dominance re-lationships in female baboons. Folia Primatologica 34:61–89.

Washburn, Sherwood L., and Irven Devore. 1961. Social behavior of baboonsand early man. In Social life of early man. Sherwood Washburn, ed. Pp.91–105. Viking Fund Publications in Anthropology, no. 31. London: Me-thuen.

Watts, Heather E., Jaime B. Tanner, Barbara L. Lundrigan, and Kay E. Hole-kamp. 2009. Post-weaning maternal effects and the evolution of femaledominance in the spotted hyena. Proceedings of the Royal Society B: BiologicalSciences 276:2291–2298.

Wells, Jonathan C. K. 2012. The capital economy in hominin evolution: howadipose tissue and social relationships confer phenotypic flexibility andresilience in stochastic environments. Current Anthropology 53(suppl. 6):S466–S478.

West, Stuart A., Claire El Mouden, and Andy Gardner. 2011. Sixteen commonmisconceptions about the evolution of cooperation in humans. Evolutionand Human Behavior 32:231–262.

West, Stuart A., Ashleigh S. Griffin, and Andy Gardner. 2007. Evolutionaryexplanations for cooperation. Current Biology 17:R661–R672.

West-Eberhard, Mary Jane. 1975. The evolution of social behavior by kinselection. Quarterly Review of Biology 50:1–33.

Wilson, Don E., and Russell A. Mittermeier. 2009. Handbook of mammals ofthe world. Madrid: Lynx.

Wrangham, Richard W., John L. Gittleman, and Colin A. Chapman. 1993.Constraints on group size in primates and carnivores: population densityand day-range as assays of exploitation competition. Behavioral Ecology andSociobiology 32:199–209.

Wrangham, Richard W., James H. Jones, Greg Laden, David Pilbeam, andNancyLou Conklin-Brittain. 1999. The raw and the stolen: cooking and theecology of human origins. Current Anthropology 40:567–594.

Yoder, Anne D., Melissa M. Burns, Sarah Zehr, Thomas Delefosse, GeraldineVeron, Steven M. Goodman, and John J. Flynn. 2003. Single origin ofMalagasy Carnivora from an African ancestor. Nature 421:734–737.

Zabel, Cynthia, J., Stephen E. Glickman, Laurence G. Frank, Katya B. Wood-mansee, and Geoffrey Keppel. 1992. Coalition formation in a colony ofprepubertal spotted hyenas. In Coalitions and alliances in humans and otheranimals. Alexander H. Harcourt and Frans B. M. de Waal, eds. Pp. 112–135. Oxford: Oxford University Press.



How Our Ancestors Broke throughthe Gray Ceiling

Comparative Evidence for Cooperative Breeding in Early Homo

by Karin Isler and Carel P. van Schaik

The “expensive brain” framework proposes that the costs of an increase in brain size can be met by any combinationof increasing the total energy turnover or reducing energy allocation to other expensive functions, such as maintenance(digestion), locomotion, or production (growth and reproduction). Here, we explore its implications for humanevolution. Using both comparative data on extant mammals and life-table simulations from wild extant apes, weshow that primates with a hominoid lifestyle face a gray ceiling that limits their brain size, with larger values leadingto demographic nonviability. We argue that cooperative care provides the most plausible exaptation for the increasein brain size in the Homo lineage.

For a change in any character to be adaptive, it must bringa net fitness benefit relative to the ancestral state. To explainthe evolution of larger brains, many hypotheses have beendevised that focus on the adaptive benefits without consid-ering the costs (e.g., Dunbar 1998). Here, following the earlyproponents of an energetic viewpoint (e.g., Aiello andWheeler 1995; Martin 1981), we argue that the high costs ofbrain tissue relative to those of other organs (Rolfe and Brown1997) should also be considered because they may limit thenet benefits to those situations where the survival benefits oflarger brains outweigh the demographic consequences of theincreased allocation of energy. Indeed, given that absolutebrain size is tightly correlated with overall cognitive perfor-mance (Deaner et al. 2007; Reader, Hager, and Laland 2011),most lineages would be able to derive a great variety of cog-nitive benefits from larger brains (e.g., Shettleworth 2010),suggesting the possibility that the ability to overcome the costsmay in fact be limiting and thus may explain most of thebrain-size variation in homeothermic vertebrates.

The “expensive brain” framework notes that evolutionaryincreases in brain size can be paid for in two complementarybut nonexclusive ways (fig. 1): (i) by increasing energy turn-over or (ii) by reducing allocation to other targets, such asmaintenance, locomotion, and production (Isler and vanSchaik 2009a). This framework can be applied to homininevolution. Early Homo is associated with the first increase inbrain size among hominins outside the range of brain sizes

Karin Isler is Senior Lecturer and Carel P. van Schaik is Professorand Director of Museum at the Anthropological Institute andMuseum, University of Zurich (Winterthurerstrasse 190, CH-8057Zurich, Switzerland [[email protected]]). This paper was submitted12 XII 11, accepted 3 VII 12, and electronically published 17 X 12.

found among great apes (Schoenemann 2006). This increasein brain size has no doubt brought various cognitive benefits,perhaps to do with tool use or cooperative hunting or otherforms of cooperation. The question pursued here, however,is how the increasing encephalization could be afforded (Ai-ello and Key 2002; Aiello and Wells 2002; Leonard et al. 2003).

Thus, following the first pathway, a part of the brain-sizeincrease in early Homo may be attributed to an increase inmetabolic turnover. Supportive evidence comes from the find-ing that the positive correlation between basal metabolic rate(BMR) and brain size is most pronounced in primates (Islerand van Schaik 2006b). Although the BMRs of humans andchimpanzees are similar and near the value predicted fromthe Kleiber line for their respective body mass (Kleiber 1961),humans exhibit a higher percentage of body fat comparedwith most primates (reviewed in Wells 2006), and thus BMRrelative to lean body mass is likely to be higher than in chim-panzees (Aiello and Wells 2002). In addition, there is growingevidence for a pronounced difference in daily energy expen-diture between humans and great apes (Pontzer 2012; Pontzeret al. 2010).

Environmental conditions should affect the potential re-action space for stabilizing the energy throughput on a higherlevel. Increased metabolic turnover may only be possible inhabitats that allow for a continuous food supply. Thus, whenperiods of unavoidable food scarcity recur, we expect mostspecies to be forced to evolve smaller brains than their sistertaxa in less seasonal environments. Indeed, we found thatseasonality in food (and hence energy) intake is negativelycorrelated with brain size in strepsirrhine and catarrhine pri-mates (van Woerden, van Schaik, and Isler 2010; van Woerdenet al. 2012) as predicted by the expensive brain framework.Work on birds, however, had earlier suggested that habitat

mailto:



Figure 1. Expensive brain framework. From an ultimate per-spective, any increase in brain tissue must be paid for either byany combination of increased energy turnover or by reducedenergy allocation to other expensive body functions.

Table 1. Phylogenetic regressions of life history traits versusfemale brain and body mass data in nonhuman primates( species)N p 86

Femalebrain mass

Femalebody mass

Life history parameter l P Effect P Effect

Neonate body mass .959 !.0001 .674 .014 .209Gestation .996 .0005 .071 .095 �.084Lactation .737 !.0001 .811 .396 �.12Interbirth interval .925 .013 .461 .688 �.054Litter size .999 .017 �.242 .166 .097Annual fertility .955 .002 �.702 .355 .148Age at first reproduction .848 .0009 .573 .111 �.198Maximum life span .864 .0004 .425 .051 �.167Maximum reproductive life span .815 .001 .412 .064 �.171rmax .950 .0004 �.688 .172 .186

Source. Primate life history and female brain and body mass data aretaken from the compilation described in van Schaik and Isler (2012).Note. Phylogenetic least squares regressions were calculated withpglm.est in the R-package CAIC (Orme et al. 2010; Purvis and Rambaut1995; R Development Core Team 2010). A l value close to 1 indicatesa strong phylogenetic influence on the respective parameters (Garland,Harvey, and Ives 1992).

seasonality imposes selection on increased brain size (e.g., Sol2009), a view known as the “cognitive buffer” hypothesis.This effect was also found among catarrhine primates in thatrelatively large-brained species show a larger difference be-tween the seasonality of their habitat and the annual variationin food intake (van Woerden et al. 2012). Nevertheless, therelationship between relative brain size and habitat seasonalityis neutral, indicating that the cognitive buffering may at bestlevel out the energetic constraint (van Woerden et al. 2012).

The habitats invaded by early Homo were clearly more sea-sonal than the gallery forests, lacustrine edges, and woodlandsinhabited by their ancestors (Potts 1998; Reed 1997). Fromthe comparative evidence, we tentatively conclude that theincreasing habitat seasonality was an important selective forcein the early Homo lineage, although the primate data suggestthat at this point it had not yet led to an increase in brainsize. Rather, increasing habitat seasonality may have shapedthe unique human combination of storing body fat in com-bination with cognitive solutions to survive irregular star-vation periods (Navarrete, van Schaik, and Isler 2011; see alsoKuzawa 1998; Wells 2010).

Turning to the second pathway, are there trade-offs betweenthe brain and other expensive body functions that may explainearly human encephalization? In a classic study, Leslie Aielloand coworkers proposed that energetic effects on human brainsize were mainly linked to reduced allocation to intestinaltissues because of increased meat eating (Aiello and Key 2002;Aiello and Wheeler 1995). However, comparative support forthis “expensive tissue” hypothesis is limited. Early studies hadfound no evidence for it in bats or birds (Isler and van Schaik2006a; Jones and MacLarnon 2004). A recent study of a largesample of mammals, including 23 species of primates, withmatching brain and organ mass data (Navarrete, van Schaik,and Isler 2011) also failed to support it. These results put thegeneral validity of this hypothesis in doubt. Moreover, for thespecific case of humans, we argue that the currently availabledata on great-ape digestive tract anatomy (Chivers and Hladik1980) are not sufficiently clear to claim reduction of the gutin the human lineage (Hladik, Chivers, and Pasquet 1999).

Another trade-off, that between the energy used for lo-comotion and for the brain as shown in birds (Isler and van

Schaik 2006a), may also have played a role in human evo-lution when in early Homo an energetically less efficient, aus-tralopithecine-like form of bipedalism evolved into a modernstriding gait. The abandonment of the energetically very ex-pensive climbing also freed these hominins from the anatom-ical compromise between climbing and walking (Isler and vanSchaik 2006a). Apart from reducing costs of locomotion, thischange in the locomotor habits may also have induced areduction of maintenance costs during rest, as humans arereported to have relatively less muscle mass than great apes(Leonard et al. 2003; Snodgrass, Leonard, and Robertson2009). However, this seeming difference could arise becauseof the higher amount of fat stores in humans. At present,hypotheses explaining increased encephalization in the hu-man lineage with metabolic trade-offs through a shift in bodycomposition are only weakly supported by empirical data(Muchlinski, Snodgrass, and Terranova 2012).

In this paper, we explore the trade-off between brain sizeand production, which includes growth and reproduction.This effect is well established among birds (e.g., Iwaniuk andNelson 2003) and mammals (Isler and van Schaik 2009a,2009b). Here, we will use correlations between life historycharacteristics and thus reproductive capacity and brain sizein extant primates to shed light on the evolutionary historyof early hominins. Briefly, we will argue that great apes havebrain sizes that are close to the maximum achievable withtheir lifestyle and that our hominin ancestors could only breakthrough this so-called gray ceiling after they had adoptedcooperative breeding.

Isler and van Schaik Cooperative Breeding in Early Homo S455

Figure 2. Residuals of life history traits versus residuals of endocranial volume in primates ( species; Homo sapiens wasN p 86excluded while calculating the regressions). Residuals were obtained from least squares regressions of the respective trait versusfemale body mass. A color version of this figure is available in the online edition of Current Anthropology.

Brain Size and Life History Traits:How Do Humans Differ?

From the expensive brain framework it follows that an in-crease in relative brain size could be paid for by reducedinvestment in production (i.e., slowing down growth, reduc-

ing reproduction, or both). We have shown that relatively

large-brained precocial mammals exhibit a reduced fertility

rate by producing much larger offspring after longer interbirth

intervals (Isler and van Schaik 2009a). This is probably be-

cause relatively large-brained immatures are highly vulnerable

to temporary shortfalls in energy supply (the “brain mal-


Table 2. Life history parameters of humans and other great apes

Parameter Gorilla gorilla Pan troglodytes Pan paniscus Pongo pygmaeus Pongo abelii Human mean 14

Female body mass (kg) 71.5 40.4 33.2 36.9 41.1 45.26Female brain size (cm3) 434 357 326 337 346 1,213Gestation length (m) 8.45 7.73 7.6 8.22 8 8.9Neonate body mass (g) 2,124 1,846 1,447 1,968 1,969 3,319Twinning rate 1/100 2.8/100 ? ? ? 1/100Interbirth interval (years) 5 5.43 4.8 7.35 9.3 3.331Weaning age (years) 3.5 4 3 5.3 5.5? 2.83Female age at first reproduction (years) 10.2 13.25 14.2 15.7 15.4 18.84Maximum life span 55 59.4 54.5 56.3 59 85

Sources. Values are taken from van Schaik and Isler (2012), from Ely et al. (2006) for chimpanzee twinning rate, from Walker et al. (2006) for themean of 14 human subsistence populations, and from Barrickman et al. (2008) for human brain size.

Figure 3. Maximum population growth rate rmax as a function of (A) endocranial volume (ECV) and (B) body mass in nonhumanprimates ( species; Homo is shown for comparison but is not included in the calculation). A color version of this figure isN p 85available in the online edition of Current Anthropology.

nutrition risk” hypothesis of Deaner, Barton, and van Schaik2003), so that a relatively large neonatal body mass is neededto buffer this risk. In addition to larger newborns, the reducedallocation to production slows down development and delaysthe age at first reproduction in relatively large-brained pre-cocial mammals and especially in primates (table 1). Indeed,a recent analysis for a carefully compiled data set of wildprimates showed that brain size is the best predictor of theduration of all stages of developmental life history except the(poorly delineated) lactational period and that taking bodysize into account does not improve the fit (Barrickman et al.2008).

To assess to what extent this effect of brain size on lifehistory also characterizes humans, we should look at humanlife history traits in relation to relative brain size. Althoughsome have questioned whether extant human foragers rep-resent the “natural” condition for our species, they are cer-tainly situated at the lower end of the spectrum of humanreproductive capacity and can thus serve as a conservativeestimate for comparison with extant ape species.

If we plot life history traits versus relative brain size innonhuman primates (fig. 2) and assess the values of humanforagers and horticulturalists based on those expected for

other primates, the main human characteristic is a distinctlyshortened period to weaning, and thus an increased annualfertility rate, for its brain size. On the other hand, humansexhibit considerably smaller neonates but only a slight de-crease in gestation length and a perfectly normal age at firstreproduction for their brain size.

Thus, the main deviation from expectation is that humansmanage to have much higher investment in reproduction(both pre- and postnatally) than expected for their brain andbody size. The same conclusion is reached when we comparethe life history of human foragers directly with that of extantnonhuman hominoids (table 2). This difference points tomajor changes in lifestyles adopted by hominins, which willbe explored after we determine that a given lineage has amaximum brain size it can achieve.

Brain Size and Maximum PopulationGrowth Rates

In large-brained mammals and primates, the developmentalslowdown and reduced reproductive rate are accompanied byan increased adult life span (Isler and van Schaik 2009a), butthe question arises whether the increased life span can con-


Table 3. Maximum population growth rates of humans and other great apes

Rate Gorilla gorilla Pan troglodytes Pan paniscus Pongo pygmaeus Pongo abelii Human mean 14

Interbirth interval (years) 5 5.43 4.8 7.35 9.3 3.331Female age at first reproduction (years) 10.2 13.25 14.2 15.7 15.4 18.84Maximum life span 55 59.4 54.5 56.3 59 85rmax .054 .049 .047 .031 .025 *DTmin (years) 12.8 14.1 14.7 22.4 27.7 *Maximum age at last birth 45? 45 45 45? 45 47rmax** (using maximum age at last birth) .051 .044 .043 .025 .017 .047DTmin** (years) 13.6 15.8 16.1 27.7 40.8 14.7

Sources. Maximum age at last birth for apes (Emery Thompson et al. 2007; Wich et al. 2004); for humans (Hill and Hurtado 1996; Howell 1979).Note. The human rmax and DTmin values calculated from maximum life span would be artificially high (*). Because of midlife menopause in humans,rmax and DTmin are more realistically calculated using maximum age at last reproduction instead of maximum life span (**). Then, however, thesame rationale must be followed for the other apes. These values should not be compared with those of other primates or mammals. DTmin pminimum time to double population size.

Table 4. Multiple regression of variables affecting rmax

simultaneously in nonhuman primates ( species,N p 85)2r p 0.869

Variable Estimate t ratio P

Intercept .083 .22 .829ln female endocranial volume �.663 �5.24 !.0001ln female body mass .067 .69 .495Terrestriality .168 3.91 .0002Nocturnality �.131 �2.40 .019Hominoidea vs. others �.232 �3.53 .0007

Note. Parametrization of the covariates was chosen empirically in orderto explain as much variation of rmax as possible as follows: terrestrialityand nocturnality were coded as binary variables (none or !5% vs. 15%of terrestriality; nocturnal vs. diurnal or cathemeral).

tinue to fully compensate the reduced production per unittime as brain size increases. On average, females of everyspecies leave roughly two viable adult offspring per lifetime,but species vary dramatically in their maximum reproductivecapacity under ideal conditions. We need a measure of re-productive capacity that represents maximum possible life-time reproductive success. The net reproductive rate (R(0)) ofextant populations is derived from life tables and will hardlyever represent optimal conditions. A far better estimate ofmaximum reproductive capacity is maximum populationgrowth rate (rmax). Additionally, in contrast to a rough productof average fertility and maximum fertile life span, rmax takesgeneration time into account. To give a stark example, if twootherwise identical species differed because in one, femalesstart to reproduce at age 1 and die at age 21, whereas in theother, females start to breed at age 21 and die at age 41, thefirst species would soon outcompete the second (Lewontin1978). The value of rmax is defined as

�r �ra �r(w�1)1 p e � be � be ,

where a p age at first reproduction, w p age at last repro-duction, or maximum life span, and b p birth rate (of femaleoffspring) per year (Cole 1954). We can calculate rmax fromage at first reproduction, maximum life span, and annualfertility rates by solving Cole’s (1954) equation numerically(Ross 1988, 1992). Enough reliable data for its calculationexist for many extant primate species. From rmax, the mini-mum number of years needed to double population size(DTmin) is calculated as

ln (2)DT p .min rmax

We have shown previously (Isler and van Schaik 2009b)that this rmax shows a very strong negative correlation withbrain mass in mammals and precocial birds, and indeed, thatbrain mass is a better predictor of rmax than is body mass. Thesame is found within primates as a group (fig. 3) and if wecontrol for phylogenetic nonindependence (table 1). Thisfinding is not a statistical artifact because brain mass might

be a better estimate of body size than body mass itself bybeing less prone to error variance (Economos 1980) and be-cause the relationship is found only for brain mass and notfor the mass of other organs, which also show a low degreeof variation (see Isler and van Schaik 2009b, app.). What isespecially striking is that great apes, in particular orangutans,show the lowest possible rmax, quite possibly close to what isminimally viable demographically.

Similarly, although rmax is based on an average annual fer-tility rate, we may expect that using a maximum fertility ratewould only strengthen the observed relationship, as small-brained species probably exhibit a higher plasticity of repro-duction in response to ecological conditions. In this case, rmax

would underestimate the maximum reproductive capacitymostly in small-brained species, yielding an even strongernegative correlation between maximum reproductive capacityand brain size. Using rmax is therefore a conservative approachfor our purpose.

The Gray Ceiling in Primates

The negative relationship between rmax and brain mass, con-trolling for body mass, indicates that as brain size increases,the increase in life span is increasingly unable to fully com-pensate for the costs incurred by long developmental periods


Figure 4. Relationship between maximum population growthrate rmax versus endocranial volume (ECV) in nonhuman pri-mates as affected by (A) terrestriality and (B) nocturnality. Toillustrate the magnitude of differences, slopes of the regressionlines were forced to be identical in both groups. Symbols as infigure 3; multivariate statistics in table 5. A color version of thisfigure is available in the online edition of Current Anthropology.

and lower reproductive rates. The most likely reason is thatthere is a realistic minimum mortality rate set by freak ac-cidents and freak environmental events (droughts, floods,fires, epidemics, lightning strikes, etc.) that are truly unavoid-able regardless of niche or behavior. Thus, as this minimummortality level is approached, further increases in brain sizewill of course continue to yield lower production but willinevitably lead to only a modest improvement in survival andthus maximum life span. As a result, rmax declines.

It is likely that such a low reproductive potential as foundin great apes compromises demographic viability for two rea-sons. First, where survival must be near perfect just to main-tain population stability, there is virtually no room for selec-tive mortality. This means that drastic changes in theenvironment must be met with phenotypically plastic re-sponses (including individual learning and innovativeness andsocially learned innovations, i.e., culture) rather than selectivemortality and that populations are almost certainly at higher

risk of local extinction in such conditions. Indeed, a popu-lation’s maximum reproductive capacity directly affects themaximum rate of environmental change that it can adapt towithout going extinct (Lynch and Lande 1993). Second, lowreproductive potential even under perfect conditions also im-plies a limited ability of a species to recover from populationcrashes and thus a species that is less likely to build up enoughindividuals to colonize new areas or habitats until the nextcrisis period. We can therefore use this reproductive potentialas an estimate of the ability to stave off population or speciesextinction.

A major consequence of this rule is that ever-lower rmax

with increasing brain size should lead us to expect a particularmaximum brain size, which we call the gray ceiling. As brainsexceed this size, population extinction becomes increasinglylikely, leading eventually to the extinction of the populationor species whenever major changes in habitat (e.g., due toclimate change) take place. Given that among primates, greatapes are at the minimum of demographic viability, we mustconclude that in this lineage no major increase in brain sizeshould be possible. Nonetheless, humans, of course, achievedexactly this, raising the question how this was possible.

Calculating rmax of extant humans is complicated by theexistence of midlife menopause, which is unique among pri-mates. If instead of maximum life span we use maximumobserved age at last reproduction (for females) and do thesame thing for great apes, the rmax of humans lies betweenthe values of gorillas and chimpanzees (table 3) instead of farlower, as one would expect based on the brain-size effect onrmax. Notice that Sumatran orangutans have a potential DTmin

of over 25 years, which may well be the lowest value observedfor all extant mammals (the actual value itself is not to betaken too seriously because it refers to a theoretical construct;it is only meant to be used for comparative purposes). In themore seasonal African environments, such a value may notbe realistic, and the observed values of the African great apes(between 13.6 and 16.1 years) suggest a realistic value of thepotential DTmin of around 20 years.

Predicting Human rmax

In comparison with other hominoids, humans exhibit a muchlarger rmax than expected for our extremely large brain size(fig. 3A). But what value of rmax would be predicted for atypical hominoid of humanlike brain and body mass? To an-swer this question, we must consider possible correlates ofeither brain size or rmax to construct a multivariate linearmodel that explains as much variation in primate rmax as pos-sible.

In a multivariate analysis within nonhuman primates( species; table 4), rmax is affected by arboreality (spe-N p 85cies that are at least partly terrestrial have a higher rmax; fig.4A) and by nocturnality (nocturnal species have a lower rmax

than diurnal or cathemeral species; fig. 4B) but not by diet(percentage of leaves or fruit or animal matter in the diet).


Table 5. Hypothetical life history traits of Homo sapiens predicted from primate and hominoid trends

Trait Model A: primate Model B: hominoid Actual values: mean 14

Litter size !1 !1 1.011Neonate mass (g) 7,377 6,865 3,319Gestation length (months) 10.2 10.9 8.9Lactation length (years) 5.46 7.57 2.83Interbirth interval (years) 5.89 7.89 3.33Age at first reproduction (years) 17.3 22.6 18.8Maximum life span (years) 68.7 79.1 85rmax .027 .022 .047DTmin (years) 25.4 32.3 14.5

Source. For comparison, the actual mean values of 14 extant human populations are taken from Walker et al. (2006).Note. Model A includes terrestriality, nocturnality, and female endocranial volume and body mass, whereas model B additionallytakes membership to Hominoidea into account. Using the predicted values for interbirth interval and age at first reproductionand setting litter size to 1 and maximum age at last reproduction to 47 years, rmax values for Homo sapiens can also be calculateddirectly, yielding .031 (model A) and .014 (model B).

Figure 5. Maximum population growth rate rmax versus brainsize (endocranial volume [ECV]) in nonhuman primates for spe-cies that exhibit cooperative breeding (Homo sapiens is excludedfrom the calculation); species that show at least some amountof allomaternal care such as paternal care, communal nursing,or babysitting; and species that show no allomaternal care at all.To illustrate the magnitude of differences, slopes of the regressionlines were forced to be identical in the three groups. Symbols asin figure 3; multivariate statistics are given in table 7. A colorversion of this figure is available in the online edition of CurrentAnthropology.

But even if these covariates are controlled for, hominoid spe-cies exhibit a lower rmax than other primates ( in aP p .0007multiple regression; table 4).

If humans followed the general primate trend, their rmax

would be estimated as 0.027 (predicted from a multivariatemodel including terrestriality, nocturnality, and female bodymass and endocranial volume [ECV]; table 5). If we take intoconsideration that we are hominoids too, the predicted rmax

would be even lower, about 0.022. This means that the DTmin

under optimum conditions would be around 30 years, whichwould almost certainly not lead to demographically viable

populations under the unstable African conditions in whichhumans evolved.

These hypothetical human rmax values, assuming a lifestylelike that of other primates, are lower than those found forany extant mammalian species. The lowest observed rmax val-ues are found in species that experience very low adult mor-tality rates (i.e., live in extremely stable habitats and hardlysuffer from predation), such as orangutans: 0.025 (Pongo abe-lii) and 0.031 (Pongo pygmaeus); killer whales: 0.028 (Pseu-dorca crassidens) and 0.046 (Orcinus orca); chimpanzees: 0.049(Pan troglodytes) and 0.047 (Pan paniscus); gorillas and Af-rican elephants: 0.054; and dugongs: 0.058. In conclusion,regardless of which model we use, a species with human brainand body mass would not be able to survive if it otherwiseadheres to a primate or hominoid lifestyle let alone whetherit was not completely arboreal and living in African woodlandor savanna.

Why Could Humans Break throughthe Gray Ceiling?

Up to this point, we have shown that a human brain–bodysize relationship would not be demographically feasible in aprimate following a typical hominoid lifestyle even if we takedifferences in diet and locomotor patterns into account. Themain distinction affecting interbirth intervals and weaningage is our system of cooperative care for infants and mothers(Burkart, Hrdy, and van Schaik 2009; Burkart and van Schaik2010; Hrdy 2005). Callitrichids are the only other primatesthat exhibit cooperative breeding to a similar extent. Indeed,the maximum reproductive rate of callitrichines is, becauseof twinning, on roughly the same grade as Homo sapiens (fig.5).

A multivariate regression yields a clear additional effect ofthis very rough measure of the extent of allomaternal care innonhuman primates taking into account the known covariatessuch as terrestriality, nocturnality, and diet (table 6). A morequantitative measurement of the extent and dimensions of


Table 6. Multiple regression of variables affecting rmax

simultaneously in nonhuman primates including a roughmeasure of allomaternal care ( species, )2N p 72 r p 0.897

Variable Estimate t ratio P

Intercept �.580 �1.39 .168ln female endocranial volume �.540 �4.32 !.0001ln female body mass .100 1.01 .317Terrestriality .106 2.40 .019Nocturnality �.001 �.01 .989Hominoidea vs. others �.316 �4.86 !.0001Cooperative breeding:

Cooperative vs. some allomaternal care .385 4.07 .0001Some vs. no allomaternal care .301 4.41 !.0001

Note. Allomaternal care was assigned to three categories: “cooperativecare”: cooperatively breeding species (callitrichines); “some allomaternalcare”: species in which at least a modest amount of help for the motheris provided through paternal care, babysitting, allonursing, or passivefood sharing; “no allomaternal care”: the remaining species. For the othercovariates, see table 4. If the variable “Hominoidea vs. others” is excluded,the effect of allomaternal care on rmax is still significant. In comparisonto the model in table 4, nocturnality does not affect rmax in this model.This indicates that the difference in rmax between nocturnal and diurnalprimates is better explained by the differences in the breeding systemthan by their activity pattern.

Table 7. Hypothetical life history traits of Homo sapiens predicted from a primate trend includ-ing cooperative breeding

TraitModel C:

primate and helpModel D:

hominoid and helpActual values:

mean 14

Litter size !1 1.036 1.011Neonate mass (g) 6,824 6,476 3,319Gestation length (months) 11.1 11.8 8.9Lactation length (years) 3.78 5.28 2.83Interbirth interval (years) 3.16 4.41 3.33Age at first reproduction (years) 16.6 20.9 18.8Maximum life span (years) 67.5 76.7 85rmax .057 .042 .047DTmin (years) 12.2 16.5 14.5

Source. For comparison, the actual mean values of 14 extant human populations are taken from Walker et al. (2006).Note. Model C includes female brain and body mass, terrestriality, and the allomaternal care category, whereas modelD additionally takes membership to Hominoidea into account. Using the predicted values for litter size, interbirthinterval, age at first reproduction, and maximum age at last reproduction set to 47 years, rmax values for Homo sapienscan also be calculated directly, yielding .054 (model C) and .035 (model D).

allomaternal help confirms this relationship (van Schaik andIsler 2012).

The inclusion of allomaternal care in the model to predicthypothetical human life history traits yields values that aremuch closer to the actual values of extant human subsistencepopulations (table 7). It is not clear a priori which of the twomodels (general primate [C] or hominoid [D] in table 7)provides the most accurate answer.

Southeast Asian hominoids (gibbons, orangutans) live inregions that were at least in part affected less by the series ofPleistocene glaciations than Africa (Whitmore 1984). Thisrelative stability may have allowed for slower viable rmax (per-haps in part achieved through lower BMRs; Pontzer et al.

2010), and they may pull the estimates for humans down.On the other hand, the heavily terrestrial gorillas may biasthe estimates in the opposite direction, and the chimpanzeevalues are actually predicted quite well by the hominoidmodel. For now, therefore, we present both sets of results andexpect that the true values may be intermediate.

Table 7 shows that the predicted age at first reproduction,fertility rates, interbirth intervals, and rmax are fairly accuratein both models. Note that in both models C and D, interbirthintervals are anomalously shorter than lactation periods,which is due to the result that in nonhuman primates, allo-maternal care reduces interbirth intervals more than it short-ens lactation periods. We are thus confident that cooperativecare is indeed responsible for the observed differences betweenhuman and ape life history traits. This interpretation is sup-ported by another result in table 7. Human life span is some-what longer than predicted, which may be linked to our ten-dency to support the sick and injured, which should improvesurvival relative to the baseline situation of no support, as ingreat apes, and thus over time maximum life span.

There is one major discrepancy between model and ob-servation that may therefore reflect another effect than co-operative breeding. Neonate mass is much smaller and ges-tation length somewhat shorter than the very large valuespredicted (largely due to our very large brain size). This dis-crepancy may be linked to the obstetrical dilemma, causedby the narrowing of the pelvic canal as a result of bipedalism(Montagu 1961; Trevathan 1987; Washburn 1960), which atsome point has become limiting for the size of the humanneonate. It is certainly consistent with the secondary altri-ciality of human neonates. Note, however, that a more altricialstate at birth can explain only this one minor difference be-tween the life histories of humans and great apes, whereasthe overall difference can be attributed to the extensive allo-maternal care in humans.


Figure 6. Minimum population doubling time versus endocranial volume (ECV) in nonhuman primates (Homo sapiens is excludedfrom the calculation). The vertical line represents an ECV of 655 cm3. Note that values are not log transformed here. A colorversion of this figure is available in the online edition of Current Anthropology.

When Did Humans Break throughthe Gray Ceiling?

In this section, we aim to predict maximum potential pop-ulation growth rates of extinct hominins from the primatemodel to find out when they would have reached the regionof demographic nonviability without a change in the breedingsystem. In a first attempt, we plot DTmin of nonhuman pri-mates versus their ECVs (fig. 6). To get a reasonable estimateof a threshold value, we conservatively assume that a doublingtime beyond 30 years ( ) would not yield viabler p 0.023max

populations. This is a very conservative estimate, as no otherliving mammal exhibits such a low maximum reproductiverate. From the relationship between DTmin and ECV, we con-clude that this value would be reached with an ECV of about650 cm3. If terrestriality is included in the model, which israther likely for all early hominins (remember we do notrequire a high percentage of terrestrial locomotion here), thethreshold would be even lower, about 610 cm3. This crudefirst attempt suggests that the first species to break throughthe gray ceiling was early Homo, which must therefore havehad extensive allomaternal care. Using the more sophisticatedmodel D—which was specific for the hominoids and includednot only brain size but also body mass, terrestriality, and thelevel of allomaternal care—the effect of a change in breedingsystem can be specified in greater detail.

Table 8 lists predicted interbirth intervals and the corre-sponding DTmin for fossil hominin taxa groups. If we assumeno allomaternal help, the predicted age at first reproduction(AFR) ranges from 12.6 years in Australopithecus afarensis to26.1 years in the very large-brained Qafzeh Homo sapiens, andthe predicted interbirth interval (IBI) is from 6 to 8.4 years

(table 8). If we assume cooperative breeding, the predicted

AFR is between 10.9 and 22.6 years, while the predicted IBI

is 3.4 for A. afarensis and 4.7 for Qafzeh H. sapiens. Estimating

twinning rate from our models is not feasible because the

twinning callitrichines introduce a strong body-mass depen-

dency of twinning rates. To estimate population growth rates,

we therefore set litter size to 1.01; that is, twinning occurs in

1% of births. The results of the model are illustrated in figure

7.We assume that a DTmin of somewhere between 15 years

(extant chimpanzees) and 20 years would still be feasible. It

is apparent that no help for mothers (as in orangutans) results

in a very steep relationship between population doubling

times and brain size. Species are included in the category of

“some help” even if they exhibit minimal helping behaviors,

such as passive food sharing or babysitting, with only minimal

frequency. Extant African apes (gorillas and chimpanzees) are

at this lower end of the spectrum. From our model, it seems

that an ECV of more than 700 cm3 would not yield sustainable

populations with such an intermediate system of allomaternal

care. Only with full cooperative breeding (as in extant humans

or callitrichines) would fossil hominins have been able to

provide sufficient energy for a sustainable population growth

rate and support a brain that is larger than 700 cm3.

In conclusion, a gradual change in lifestyle toward a sub-

stantial increase in allomaternal help (including provisioning

of mothers and weaned offspring) may have evolved early in,

or even before, the genus Homo. (For the challenge of allo-

cating the earliest Homo fossils to meaningful clusters, see

Anton 2012.)


Table 8. Predicted life history and demographic parameters of early hominins

Allomaternal careduring predicted

interbirth interval(years)

Allomaternal careduring predicted

DTmin (years)

Species and sample Time (Ma)Female endocranial

volume Female body mass No Some CB No Some CB

Australopithecus afarensis:A.L. 333-105 3.2 343 29.3 6.04 4.59 3.43 17.7 14.0 10.3A.L. 444-2 3.2 550 51.3* 6.88 5.24 3.91 23.4 17.7 12.6

Australopithecus africanus:STS 71 2.75 428 26.6 6.21 4.73 3.53 20.1 15.6 11.4STW 505 2.5 560 46.8 6.85 5.21 3.89 23.8 18.0 12.7

Australopithecus boisei:KNM-ER 732 female 1.7 500 32.0 6.48 4.93 3.68 22.1 17.0 12.2OH 5 male 1.8 530 57.6 6.91 5.26 3.93 22.8 17.3 12.4

Early Homo:KNM-ER 1813 1.89 509 34.9 6.56 4.99 3.72 22.4 17.1 12.3KNM-ER 1805 1.89 580 30.3* 6.61 5.03 3.76 24.6 18.6 13.1KNM-ER 1470 1.89 752 45.6 7.18 5.46 4.07 30.0 21.8 14.8

Homo erectus:Africa:

KNM-ER 42700 (Ileret) 1.55 690 45* 7.06 5.38 4.01 27.9 20.5 14.2KNM-ER 3733 1.8 850 59.2 7.50 5.71 4.26 33.4 23.7 15.7

Georgia:D3444 1.77 638 47* 7.00 5.33 3.97 26.2 19.5 13.6D2280 1.77 775 52.6* 7.31 5.56 4.15 30.7 22.2 15.0

Asia:Zhoukoudian XI female .42 1,015 51.8 7.64 5.81 4.34 41.2 27.9 17.6Zhoukoudian X male .42 1,225 65.6* 8.06 6.13 4.57 53.2 33.3 19.8

Archaic Homo sapiens:Steinheim .25 1,110 60.5 7.86 5.98 4.47 45.9 30.1 18.6Jebel Irhoud .09 1,305 80.5 8.3 6.31 4.71 58.0 35.2 20.5

Homo neanderthalensis:Saccopastore female .12 1,245 66.6 8.09 6.16 4.59 54.6 33.9 20.0Le Moustier male .041 1,565 81.2 8.56 6.52 4.86 89.3 45.8 23.8

H. sapiens:Zhoukoudian 102 female .015 1,380 43.2 7.91 6.02 4.49 74.5 41.7 22.5Qafzeh 9 female .1 1,531 64.6 8.35 6.36 4.74 89.5 46.1 23.8Extant females 0 1,213 45.3 7.79 5.93 4.42 55.1 34.3 20.1

Sources. Endocranial volumes (cm3) and body mass (kg) estimates of fossils are taken from Gabunia et al. (2000), Kappelman (1996), Spocter andManger (2007), Spoor et al. (2007), and Ruff (2010).Note. As sex determination is notoriously difficult for early hominins, we list both a small and a large morph from reasonably complete crania.Body mass estimates denoted with an asterisk do not correspond to the same fossil as the endocranial volume. The body mass for a small AfricanH. erectus (45 kg) is very roughly estimated from comparing other estimates with the size of the Ileret cranium. Interbirth intervals and age at firstreproduction are estimated using model D from table 6 (excluding the effect of nocturnality). For calculating minimum population doubling time(DTmin), maximum age of reproduction is set to 47 years and twinning rate to 1/100. Values !20 years are highlighted in boldface. CB p cooperativebreeding. The predictions for 1.9 Ma Australopithecus sediba would be very similar to A. africanus min. (endocranial volume of 420 cm3 in a juvenilemale, body mass of the adult female estimated at 27 kg; Berger et al. 2010).

Were Early Homo Cooperative Breeders?

We believe that Homo erectus (p ergaster), as it emerged ataround 1.8 Ma, was a good candidate for having extensiveallomaternal care for two major reasons. First, they were likelythe first systematic hunters of large game (Foley and Lee 1991;Pobiner et al. 2008). Large-game hunting requires cooperationduring the hunt, cooperative defense against other dangerouscarnivores, extremely high tolerance around kills, and fre-quent food sharing, perhaps even to the point of provisioning.These features are all more likely among cooperative breeders

(van Schaik and Burkart 2010). Indeed, among mammals,carnivores are more likely to be cooperative breeders (Smith

et al. 2012; Solomon and French 1997; Spencer-Booth 1970).

Second, the weaned juveniles were less likely to make a living

on their own and would have strongly benefited from allo-

maternal support. They lived on the savanna, where resources

harvested as efficiently by juveniles as adults, such as soft fruits,

are much scarcer than in forests (Hawkes et al. 1998), leading

to reduced juvenile foraging efficiency. The latter is especially

likely if they had already acquired a great reliance on meat


Figure 7. Minimum population doubling time (DTmin; years) of fossil hominins predicted from model D (individual values listedin table 8). We assume that a DTmin of somewhere between 15 years (extant chimpanzees) and 20 years would still be feasible(shaded bar). The lack of smoothness results from the inclusion of body mass in the model. A color version of this figure is availablein the online edition of Current Anthropology.

(Domınguez-Rodrigo and Pickering 2003), because the diffi-culty of learning how to hunt means that provisioning meathas strong positive effects on the fitness prospects of the young.More seasonal habitats are more likely to contain cooperativebreeders (Hatchwell 2007; Rubenstein and Lovette 2007). Theargument is further supported by H. erectus (p ergaster) oc-cupying a much larger geographic range than earlier homininspecies. Hrdy (2005, 2009) has argued convincingly that colo-nizing hostile new habitats is facilitated by cooperative breeding.

Identifying the source of extensive allomaternal care in earlyHomo is difficult, as the defining feature of human caretakingseems to be its large flexibility (Hrdy 2009). In present-day hu-man societies, grandmothers and males, but also not directlyrelated adults (Hill and Hurtado 2009), play a major role. Asmidlife menopause is extremely rare in mammals (Packer, Tatar,and Collins 1998), we cannot apply comparative evidence to theevolution of grandmothering. However, males were almost cer-tainly involved in meat sharing and thus allomaternal care assoon as confrontational scavenging or hunting of large game waspresent (Marlowe 2007). In sum, while the brain size of H. erectusand various other indicators suggest that females of this speciesreceived much allomaternal care, we assume that male-femalepair bonds accompanied by selective food sharing were sourcesof this care, but we can make no conclusions about the role ofgrandmothers.

Discussion

The analyses reported here suggest that the inability of survivalto keep up with reduced production as brain size increases leads

to a reduction in rmax in larger-brained organisms. There comesa point where no further increases in brain size are possiblebecause the long-term viability of populations is severely com-promised. This point we call the “gray ceiling.” For great apesliving a great-ape lifestyle, we put this conservatively at 600–700 cm3. This explains why extant great apes and extinctaustralopithecines seem to have converged on similar brainsizes, but it makes the “escape” from great-ape level brainsizes by Homo even more striking. Assigning a distinct bound-ary to a highly fragmentary fossil record is tricky, but Homorudolfensis (i.e., KNM-ER 1470) is a likely candidate for sucha change in lifestyle. The first well-documented hominin toshow brains that exceed this size was Homo erectus (p erg-aster), which arose in Africa at around 1.8 Ma, occupiedsavanna habitats, hunted large game, and rather quickly hadmoved into other geographic regions.

Allomaternal care tends to lead to higher female repro-ductive output in both primates (Mitani and Watts 1997; Rossand MacLarnon 2000) and carnivores (Isler and van Schaik2009a). We propose that as in other mammals and birds, theadoption of cooperative breeding (Hrdy 2005, 2009) had al-lowed H. erectus (p ergaster) to increase its rmax, which, givenits value near the gray ceiling, made possible an expansion ofits brain size. As a conservative estimate, our gray ceiling valueof 600–700 cm3 provides an upper boundary to brain size ifa species is adhering to an apelike lifestyle. Of course, wecannot exclude the possibility that cooperative breeding pre-dated a pronounced increase of encephalization by severalmillion years, as suggested in Lovejoy’s (2009) scenario for


the adaptive suite of characters assigned to Ardipithecus ram-idus. In this case, however, another explanation would beneeded for the long time lag between the onset of provisioningand increase in brain size. Australopithecines were adept bi-peds without sectorial canine complexes, but there is no evi-dence for a shift in life history traits and developmental tra-jectories before Homo (Dean 2006; Dean and Lucas 2009;Schwartz 2012; but see DeSilva 2011).

In conclusion, if we rely on estimating the effect of evo-lutionary processes known to operate in primates or in ver-tebrates in general, there is evidence for several factors thatallowed for brain-size expansion throughout the evolutionaryhistory of the human lineage. A more seasonal environment,a change in diet toward higher-quality food sources, and moreefficient locomotion all may have played a role (Potts 2011).Instead of a comprehensive but unique “adaptive suite” ofhuman traits (Lovejoy 2009), however, we find broad com-parative support for a decisive role of cooperative breedingas the initial trigger of many subsequent changes in humanbiology (Burkart, Hrdy, and van Schaik 2009; Burkart andvan Schaik 2010). As such a redistribution of energy towardmothers and infants is possible without changing the overallenergy budget, it may have facilitated subsequent changes thatled to the relatively high energetic throughput of modernhumans as compared with extant apes (Pontzer 2012; Pontzeret al. 2010).

Acknowledgments

We thank Susan Anton and Leslie Aiello for inviting us tothe Wenner-Gren spring symposium 2011 and for editing thisvolume. Many people have contributed data to our compi-lations, which we gratefully acknowledge. Financial supportfor this study was provided through Swiss National ScienceFoundation grant 3100A0-117789, the A. H. Schultz Foun-dation, and the University of Zurich.


Homo erectus female. American Journal of Human Biology 14(5):551–565.Aiello, Leslie C., and Jonathan C. K. Wells. 2002. Energetics and the evolution

of the genus Homo. Annual Review of Anthropology 31:323–338.Aiello, Leslie C., and Peter Wheeler. 1995. The expensive-tissue hypothesis:

the brain and the digestive system in human and primate evolution. CurrentAnthropology 36(2):199–221.


Barrickman, Nancy L., Meredith L. Bastian, Karin Isler, and Carel P. vanSchaik. 2008. Life history costs and benefits of encephalization: a compar-ative test using data from long-term studies of primates in the wild. Journalof Human Evolution 54(5):568–590.


Burkart, Judith M., Sarah B. Hrdy, and Carel P. van Schaik. 2009. Cooperativebreeding and human cognitive evolution. Evolutionary Anthropology 18:175–186.

Burkart, Judith M., and Carel P. van Schaik. 2010. Cognitive consequencesof cooperative breeding in primates? Animal Cognition 13(1):1–19.

Chivers, David J., and Claude M. Hladik. 1980. Morphology of the gastro-intestinal tract in primates: comparisons with other mammals in relationto diet. Journal of Morphology 166:337–386.

Cole, Lamon C. 1954. The population consequences of life-history phenom-ena. Quarterly Review of Biology 29(2):103–137.

Dean, M. Christopher. 2006. Tooth microstructure tracks the pace of humanlife-history evolution. Proceedings of the Royal Society B 273:2799–2808.

Dean, M. Christopher, and Victoria S. Lucas. 2009. Dental and skeletal growthin early fossil hominins. Annals of Human Biology 36(5):545–561.

Deaner, Robert O., Rob Barton, and Carel P. van Schaik. 2003. Primate brainsand life histories: renewing the connection. In Primate life histories andsocioecology. Peter Kappeler and Michael E. Pereira, eds. Pp. 233–265. Chi-cago: University of Chicago Press.

Deaner, Robert O., Karin Isler, Judith M. Burkart, and Carel P. van Schaik.2007. Overall brain size, and not encephalization quotient, best predictscognitive ability across non-human primates. Brain, Behavior and Evolution70:115–124.

DeSilva, Jeremy M. 2011. A shift toward birthing relatively large infants earlyin human evolution. Proceedings of the National Academy of Sciences, U.S.A.108(3):1022–1027.

Domınguez-Rodrigo, Manuel, and Travis R. Pickering. 2003. Early hominidhunting and scavenging: a zooarcheological review. Evolutionary Anthro-pology 12(6):275–282.

Dunbar, Robin I. M. 1998. The social brain hypothesis. Evolutionary Anthro-pology 6(5):178–190.

Economos, Angelos C. 1980. Brain–life span conjecture: a re-evaluation ofthe evidence. Gerontology 26:82–89.

Ely, John J., William I. Frels, Sue Howell, M. Kay Izard, Michale E. Keeling,and D. Rick Lee. 2006. Twinning and heteropaternity in chimpanzees (Pantroglodytes). American Journal of Physical Anthropology 130(1):96–102.

Emery Thompson, Melissa, James H. Jones, Anne E. Pusey, Stella Brewer-Marsden, Jane Goodall, David Marsden, Tetsuro Matsuzawa, et al. 2007.Aging and fertility patterns in wild chimpanzees provide insights into theevolution of menopause. Current Biology 17(24):2150–2156.

Foley, Robert A., and Phyllis C. Lee. 1991. Ecology and energetics of en-cephalization in hominid evolution. Philosophical Transactions of the RoyalSociety B 334:223–232.

Gabunia, Leo, Abesolom Vekua, David Lordkipanidze, Carl C. Swisher III,Reid Ferring, Antje Justus, Medea Nioradze, et al. 2000. Earliest Pleistocenehominid cranial remains from Dmanisi, Republic of Georgia: taxonomy,geological setting, and age. Science 288(5468):1019–1025.

Garland, Theodore, Paul H. Harvey, and Anthony R. Ives. 1992. Proceduresfor the analysis of comparative data using phylogenetically independentcontrasts. Systematic Biology 41(1):18–32.

Hatchwell, Ben J. 2007. Avian reproduction: role of ecology in the evolutionof cooperative breeding. Current Biology 17(19):R845–R847.

Hawkes, Kristen, James F. O’Connell, Nicholas G. Blurton Jones, Helen Al-varez, and Eric L. Charnov. 1998. Grandmothering, menopause, and theevolution of human life histories. Proceedings of the National Academy ofSciences, U.S.A. 95(3):1336–1339.

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

———. 2009. Cooperative breeding in South American hunter-gatherers.Proceedings of the Royal Society B 276(1674):3863–3870.

Hladik, Claude M., David J. Chivers, and Patrick Pasquet. 1999. On diet andgut size in non-human primates and humans: is there a relationship tobrain size? discussion and criticism. Current Anthropology 40:695–697.

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

Hrdy, Sarah B. 2005. Evolutionary context of human development: the co-operative breeding model. In Attachment and bonding: a new synthesis. C.Sue Carter, Lieselotte Ahnert, Klaus E. Grossmann, Sarah B. Hrdy, MichaelE. Lamb, Stephen W. Porges, and Norbert Sachser, eds. Pp. 9–32. Cam-bridge, MA: MIT Press.

———. 2009. Mothers and others: the evolutionary origins of mutual under-standing. Cambridge, MA: Belknap.

Isler, Karin, and Carel P. van Schaik. 2006a. Costs of encephalisation: theenergy trade-off hypothesis tested on birds. Journal of Human Evolution51(3):228–243.

———. 2006b. Metabolic costs of brain size evolution. Biology Letters 2:557–560.


———. 2009a. The expensive brain: a framework for explaining evolutionarychanges in brain size. Journal of Human Evolution 57(4):392–400.

———. 2009b. Why are there so few smart mammals (but so many smartbirds)? Biology Letters 5(1):125–129.

Iwaniuk, Andrew N., and John E. Nelson. 2003. Developmental differencesare correlated with relative brain size in birds: a comparative analysis. Ca-nadian Journal of Zoology 81:1913–1928.

Jones, Kate E., and Ann M. MacLarnon. 2004. Affording larger brains: testinghypotheses of mammalian brain evolution on bats. American Naturalist164(1):E20–E31.

Kappelman, John. 1996. The evolution of body mass and relative brain sizein fossil hominids. Journal of Human Evolution 30(3):243–276.

Kleiber, Max. 1961. The fire of life: an introduction to animal energetics. NewYork: Wiley.


Leonard, William R., Marcia L. Robertson, J. Josh Snodgrass, and ChristopherW. Kuzawa. 2003. Metabolic correlates of hominid brain evolution. Com-parative Biochemistry and Physiology A 136(1):5–15.

Lewontin, Richard C. 1978. Adaptation. Scientific American 239(3):156–169.Lovejoy, C. Owen. 2009. Reexamining human origins in light of Ardipithecus

ramidus. Science 326(5949):74e71–74e78.Lynch, Michael, and Russell Lande. 1993. Evolution and extinction in response

to environmental change. In Biotic interactions and global change. Peter M.Kareiva, Joel G. Kingsolver, and Raymond B. Huey, eds. Pp. 234–250. Sun-derland, MA: Sinauer.

Marlowe, Frank W. 2007. Hunting and gathering: the human sexual divisionof foraging labor. Cross-Cultural Research 41(2):170–195.

Martin, Robert D. 1981. Relative brain size and basal metabolic rate in ter-restrial vertebrates. Nature 293(5824):57–60.

Mitani, John C., and David Watts. 1997. The evolution of non-maternalcaretaking among anthropoid primates: do helpers help? Behavioral Ecologyand Sociobiology 40(4):213–220.

Montagu, Ashley. 1961. Neonatal and infant immaturity in man. Journal ofthe American Medical Association 178(1):56–57.

Muchlinski, Magdalena N., J. Josh Snodgrass, and Carl J. Terranova. 2012.Muscle mass scaling in primates: an energetic and ecological perspective.American Journal of Primatology 74:395–407.

Navarrete, Ana F., Carel P. van Schaik, and Karin Isler. 2011. Energetics andthe evolution of human brain size. Nature 480:91–93.

Orme, David, Rob P. Freckleton, Gavin Thomas, Thomas Petzoldt, and Su-sanne Fritz. 2010. CAIC: comparative analyses using independent contrasts.https://r-forge.r-project.org/projects/caic/.

Packer, Craig, Marc Tatar, and Anthony Collins. 1998. Reproductive cessationin female mammals. Nature 392(6678):807–811.

Pobiner, Briana L., Michael J. Rogers, Christopher M. Monahan, and JohnW. K. Harris. 2008. New evidence for hominin carcass processing strategiesat 1.5 Ma, Koobi Fora, Kenya. Journal of Human Evolution 55(1):103–130.


Pontzer, Herman, David A. Raichlen, Robert W. Shumaker, Cara Ocobock,and Serge A. Wich. 2010. Metabolic adaptation for low energy throughputin orangutans. Proceedings of the National Academy of Sciences, U.S.A.107(32):14048–14052.

Potts, Richard. 1998. Environmental hypotheses of hominin evolution. Year-book of Physical Anthropology 41:93–136.

———. 2011. Evolution: big brains explained. Nature 480:43–44.Purvis, Andy, and Andrew Rambaut. 1995. Comparative analysis by inde-

pendent contrasts (CAIC): an Apple Macintosh application for analysingcomparative data. Computer Applications in the Biosciences 11(3):247–251.

R Development Core Team. 2010. R: a language and environment for statisticalcomputing. Vienna, Austria: R Foundation for Statistical Computing.

Reader, Simon M., Yfke Hager, and Kevin N. Laland. 2011. The evolution ofprimate general and cultural intelligence. Philosophical Transactions of theRoyal Society B 366:1017–1027.

Reed, Kaye E. 1997. Early hominid evolution and ecological change throughthe African Plio-Pleistocene. Journal of Human Evolution 32(2/3):289–322.

Rolfe, David F. S., and Guy C. Brown. 1997. Cellular energy utilization andmolecular origin of standard metabolic rate in mammals. Physiological Re-views 77(3):731–758.

Ross, Caroline. 1988. The intrinsic rate of natural increase and reproductiveeffort in primates. Journal of Zoology (London) 214(2):199–219.

———. 1992. Environmental correlates of the intrinsic rate of natural increasein primates. Oecologia 90(3):383–390.

Ross, Caroline, and Ann M. MacLarnon. 2000. The evolution of non-maternalcare in anthropoid primates: a test of the hypotheses. Folia Primatologica71(1/2):93–113.

Rubenstein, Dustin R., and Irby J. Lovette. 2007. Temporal environmentalvariability drives the evolution of cooperative breeding in birds. CurrentBiology 17:1414–1419.

Ruff, Christopher B. 2010. Body size and body shape in early hominins:implications of the Gona Pelvis. Journal of Human Evolution 58(2):166–178.

Schoenemann, P. Thomas. 2006. Evolution of the size and functional areasof the human brain. Annual Review of Anthropology 35:379–406.


Shettleworth, Sara J. 2010. Clever animals and killjoy explanations in com-parative psychology. Trends in Cognitive Sciences 14(11):477–481.

Smith, Jennifer E., Eli M. Swanson, Daphna Reed, and Kay E. Holekamp.2012. Evolution of cooperation among mammalian carnivores and its rel-evance to hominid evolution. Current Anthropology 53(suppl. 6):S436–S452.

Snodgrass, J. Josh, William R. Leonard, and Marcia L. Robertson. 2009. Theenergetics of encephalization in early hominids. In The evolution of hominindiets. Jean-Jacques Hublin and Michael P. Richards, eds. Pp. 15–29. Ver-tebrate Paleobiology and Paleoanthropology Series. Berlin: Springer.

Sol, Daniel. 2009. Revisiting the cognitive buffer hypothesis for the evolutionof large brains. Biology Letters 5(1):130–133.

Solomon, Nancy G., and Jeffrey A. French, eds. 1997. Cooperative breeding inmammals. Cambridge: Cambridge University Press.

Spencer-Booth, Yvette. 1970. The relationships between mammalian youngand conspecifics other than mothers and peers: a review. Advances in theStudy of Behavior 3:119–194.

Spocter, Muhammad A., and Paul R. Manger. 2007. The use of cranial variablesfor the estimation of body mass in fossil hominins. American Journal ofPhysical Anthropology 134(1):92–105.

Spoor, Fred, Meave G. Leakey, Patrick N. Gathogo, Francis H. Brown, SusanC. Anton, Ian McDougall, Christopher Kiarie, Fredrick K. Manthi, andLouise N. Leakey. 2007. Implications of new early Homo fossils from Ileret,east of Lake Turkana, Kenya. Nature 448(7154):688–691.

Trevathan, Wenda. 1987. Human birth: an evolutionary perspective. New York:Aldine.

van Schaik, Carel P., and Judith M. Burkart. 2010. Mind the gap: cooperativebreeding and the evolution of our unique features. In Mind the gap: tracingthe origins of human universals. Peter M. Kappeler and Joan Silk, eds. Pp.477–496. Berlin: Springer.

van Schaik, Carel P., and Karin Isler. 2012. Life history evolution in primates.In The evolution of primate societies. Josep Call, Peter M. Kappeler, JohnMitani, Ryne A. Palombit, and Joan Silk, eds. Pp. 220–244. Chicago: Uni-versity of Chicago Press.

van Woerden, Janneke T., Carel P. van Schaik, and Karin Isler. 2010. Effectsof seasonality on brain size evolution: evidence from strepsirrhine primates.American Naturalist 176(6):758–767.

van Woerden, Janneke T., Erik P. Willems, Carel P. van Schaik, and KarinIsler. 2012. Large brains buffer energetic effects of seasonal habitats incatarrhine primates. Evolution 66:191–199.

Walker, Robert, Michael Gurven, Kim Hill, Andrea Migliano, Napoleon Chag-non, Roberta De Souza, Gradimir Djurovic, et al. 2006. Growth rates andlife histories in twenty-two small-scale societies. American Journal of HumanBiology 18(3):295–311.

Washburn, Sherwood L. 1960. Tools and human evolution. Scientific American203(3):63–75.

Wells, Jonathan C. K. 2006. The evolution of human fatness and susceptibilityto obesity: an ethological approach. Biological Reviews 81:183–205.

———. 2010. The evolutionary biology of human body fatness. Cambridge:Cambridge University Press.

Whitmore, Timothy C. 1984. Tropical rain forests of the Far East. Oxford:Clarendon.

Wich, Serge A., S. Suci Utami-Atmoko, Tatang Mitra Setia, Herman D. Rijksen,Chris L. Schurmann, Jan A. R. A. M. van Hooff, and Carel P. van Schaik.2004. Life history of wild Sumatran orangutans (Pongo abelii). Journal ofHuman Evolution 47(6):385–398.

https://r-forge.r-project.org/projects/caic/



The Capital Economy in Hominin EvolutionHow Adipose Tissue and Social Relationships Confer Phenotypic

Flexibility and Resilience in Stochastic Environments

by Jonathan C. K. Wells

The global distribution of our species indicates a biology capable of adapting to an extraordinary range of ecosystems,generating interest in how such a biology evolved. Whereas much attention has been directed to genetic adaptationand developmental plasticity as adaptive strategies, ecological stochasticity within the life course may be addressedby additional strategies such as bet hedging and phenotypic flexibility. Both social relationships and adipose tissuemay be considered as “energy capital” conferring reversible phenotypic flexibility across the life course. Evidencefrom primates and contemporary humans demonstrates the value of such energy capital in accommodating ecologicaluncertainty. The fact that Homo sapiens is characterized by high levels of both cooperative sociality and adipositycompared with extant apes suggests that ecological stochasticity may have been a key ecological stress in the evolutionof our genus. The benefits of phenotypic flexibility for ecological risk management may have preceded and enabledthe emergence of traits such as carnivory, encephalization, colonizing, and the maintenance of a single breedingspecies across diverse environments.

To understand the particular characteristics of extant humans,much work has focused on physical traits readily discerniblein the fossil record, such as cranial capacity, dentition, andpostcranial anatomy. Such investigation enables the recon-struction of trends in functional traits such as encephalization,locomotion, and reproduction (e.g., Anton 2003; Bramble andLieberman 2004; Bruner, Manzi, and Arsuaga 2003; DeSilva2011; Rosenberg and Trevathan 2002). Evolutionary changesin such traits are typically considered physical or physiological“solutions” to ecological stresses, such as climate trends orshifts in the availability of resources (Klein and Edgar 2002;Vrba 1985), although neutral evolution and genetic drift mayalso contribute to phenotypic trends. Much attention has beendirected, for example, to understanding what might havedriven the trend toward bipedal locomotion in earlier hom-inins (e.g., Crompton, Sellers, and Thorpe 2010) or progres-sive encephalization in the genus Homo (e.g., Bruner, Manzi,and Arsuaga 2003). Dental analysis offers the capacity to re-construct how hominins responded to changes in the avail-ability of food sources (Organ et al. 2011), indicating, forexample, a postaustralopithecine shift to a broader diet base(Ungar 2012). Although the emergence of the genus Homo

Jonathan C. K. Wells is Professor of Anthropology and PediatricNutrition at the Childhood Nutrition Research Centre, UniversityCollege London Institute of Child Health (30 Guilford Street, LondonWC1N 1EH, United Kingdom [[email protected]]). Thispaper was submitted 12 XII 11, accepted 3 VII 12, and electronicallypublished 23 X 12.

has generally been considered an integrated shift in many ofthese trends, this perspective is now undergoing reappraisalin the light of more comprehensive skeletal evidence (Antonand Snodgrass 2012).

The study of ecogeographical and temporal variability withincontemporary humans plays a key role in these reconstructions.For example, substantial research has been conducted on theassociation between climate and morphology in humans (Hier-naux and Froment 1976; Katzmarzyk and Leonard 1998; Rob-erts 1953; Wells 2012b). Such work then informs understandingof past variability in hominin morphology (Perry and Dominy2009; Ruff 2002). Similarly, variability in contemporary or re-cent human dentition can be used to infer past dietary stressesand to evaluate evolutionary patterns (Anton, Carter-Menn,and DeLeon 2011; Kaifu et al. 2003).

Although not addressed explicitly by all authors, nonpath-ological anatomical variation in the fossil record has tradi-tionally tended to receive a genetic interpretation (e.g., Hol-liday 1997; though see Ruff 2002). First, phylogenetic analysestypically show good agreement between molecular and mor-phological data (Gibbs, Collard, and Wood 2000; Gilbert andRossie 2007; Pilbeam 2000), potentially implying a similarscenario for within-species variability. Second, the long-stand-ing attention directed to long-term ecological trends as keystresses (e.g., the “savannah hypothesis”; Dart 1925; Klein andEdgar 2002) would also match with a genetic model of ad-aptation. Consistent with such a perspective, climate has in-deed been associated with genetic variability among contem-porary humans (Hancock et al. 2010, 2011).


Wells The Capital Economy in Hominin Evolution S467

More recently, however, the contribution of developmentalplasticity to phenotypic variability has been recognized inextant primates and humans, suggesting that a similar ap-proach is required for the fossil record (e.g., Ruff 2002). Ex-perimental work on primates has demonstrated rapid trans-generational shifts in body proportions in populationsexposed to novel thermal environments (Paterson 1996) orto shifts in the availability of energy (Price, Hyde, and Coe1999), clearly demonstrating a contribution of plasticity tophenotypic change across generations. Similar work in hu-mans has shown that both nutritional experience and psy-chosocial stress in early life can exert long-term effects onmultiple aspects of phenotype (e.g., Bogin et al. 2002; Ellis etal. 2009). A more comprehensive approach must thereforeassume that variability in the hominin fossil record reflectsthe influences of both genotype and developmental plasticity(Kuzawa and Bragg 2012; Ruff 2002; Wells and Stock 2007).

The fundamental connection between contemporary hu-man variability and past evolutionary change may also applyto phenotypic traits that leave, at best, weak signals in thefossil record. Recently, for example, much interest has beendirected to why and when the unusual life history profile ofhumans (lengthy life spans and developmental periods butshort interbirth intervals) emerged in our evolutionary history(Aiello and Key 2002; Anton 2003; Hawkes 2006; Hrdy 2009).Life history variability among contemporary humans is as-sumed to incorporate adaptation to stresses such as energyavailability and mortality risk, though the associations arecomplex (Ellis et al. 2009; Migliano and Guillon 2012; Perryand Dominy 2009; Walker and Hamilton 2008; Walker et al.2006). In turn, humans per se have an unusually lengthyperiod of development, suggesting that the Homo niche fa-vored a progressive reorganization of the tempo of devel-opment in response to diverse ecological pressures, some ofthem potentially compounded by hominin traits such as in-creasing brain size (Isler and van Schaik 2012; Robson, vanSchaik, and Hawkes 2006; Schwartz 2012).

Information on some life history traits can be gleaned fromthe fossil record, though others, such as interbirth intervaland total fertility rate, may leave no material signal. For ex-ample, dental eruption patterns represent markers of matu-ration rate across species, though interpretation is complex(Schwartz 2012), while age at weaning can potentially be es-timated from isotopic enrichment ratios (e.g., Herring, Saun-ders, and Katzenberg 1998). Furthermore, plasticity in suchdevelopmental traits may be revealed by traits such as bodyproportions; for example, leg length appears particularly vul-nerable to poor childhood nutrition, and short leg length maytherefore act as a skeletal marker for slowed development (e.g.,Li, Dangour, and Power 2007), though other ecologicalstresses such as cold climates may generate similar effects(Katzmarzyk and Leonard 1998).

This overall approach is thus proving very valuable forunderstanding both when and why human traits emerged andalso how they have come to vary across ecological conditions

since the origin of our species. However, the interpretationof such phenotypic variation in the fossil record as adaptationis possible only because each adaptation of genotype and de-velopmental plasticity may generate relatively irreversible phe-notypic effects within the life course.

Natural selection acts not on traits but on strategies (Hous-ton and McNamara 1999), for example, not on tall height,but on growing faster or for longer. In contrast to the ex-amples described above, some adaptive strategies are relativelyreversible, which in turn tells us about the kind of ecologicalstresses that drive them. Such plasticity may relate either tothe instability of phenotype within individuals or to pheno-typic relationships between individuals. In this article, I firstdescribe how stochastic environments favor such phenotypicflexibility, which may emerge in different formats. I then re-view how both primates and humans benefit from such flex-ible traits, focusing in particular on two generalized stores ofenergy. Finally, I consider what role such energy stores mayhave played in human evolution.

Modes of Adaptation

Phenotypic plasticity embraces many forms that can vary bothin their relative speed of response to ecological stresses and intheir degree of reversibility. At one extreme, developmentalplasticity tends to be relatively irreversible, reflecting interac-tions between genotype and environment in early life, whereasat the other extreme, some aspects of metabolism and behaviorcan alter on a second-by-second basis (Gabriel et al. 2005;Piersma and Drent 2003). In turn, the time lag of plastic re-sponse following the environmental cue may also vary (Gabrielet al. 2005). Most work has focused on the two extremes, usingquantitative genetic models to address the norms of reactionunderlying developmental plasticity or optimality theory to ad-dress behavioral ecology (Gabriel et al. 2005).

Such variability in the form of plasticity reflects variabilityin the frequency and regularity of ecological stresses relativeto the life span of the organism (Piersma and Drent 2003).Although attention has traditionally been directed to long-term ecological shifts affecting hominin evolution, there isincreasing interest in the diversity of ecological stresses, someof which are short-term but frequent and unpredictable (Bobeand Behrensmeyer 2004; Potts 1996, 2012).

Genetic adaptations can develop over variable time spans.There is increasing recognition that specific traits underlyingthe complex life history profile of humans have a polygeneticbasis (Elks et al. 2010; Hirschhorn and Lettre 2009; Yang et al.2010); hence, phenotypic variability in such traits may derivein large part from variability in gene frequencies within andbetween populations (Hancock et al. 2010). Although oftenassumed to require hundreds of generations, genetic changecan occur rapidly during periods of rapid population growth,and hence it may have been particularly important since theNeolithic revolution (Cochran and Harpending 2009).


Developmental plasticity represents a faster mode of phe-notypic change, theoretically capable of fine-tuning pheno-type to more recent or local conditions (Stearns 1992; West-Eberhard 2003). A key mechanism underlying such plasticityis epigenetic variability in gene expression in response to en-vironmental factors occurring during early development(Youngson and Whitelaw 2008). Such epigenetic marks arethen broadly maintained through the life course, though withprogressive attrition over time (Fraga et al. 2005). It is tempt-ing to assume that developmental plasticity is adaptive; how-ever, in long-lived species there may be a systematic shift inecological conditions between the life course periods of de-velopment and reproduction, potentially resulting in costlymaladaptation.

Many ecological stresses are not systematic, however, andoccur irregularly through the life course while still exertingpowerful effects on biology. Potts (2012) has argued that stud-ies using single markers of past ecological change have tendedto bias interpretation toward particular climatic trends (e.g.,temperature or aridity) and that a more integrative approachemphasizes temporal fluctuations in the overall level of eco-logical variability and hence shifts attention from “ecologicalstress” to “ecological uncertainty.”

As the frequency of stochasticity accelerates relative to theduration of the life course and hence the level of uncertaintyincreases, systematic adaptations of the kind described aboveare neither especially valuable nor indeed easy to develop. Atthe simplest level, the environment may “flip” between twoextremes, generating a dilemma for the organism: adapt toone, the other, or neither, in which case it will be maladaptedin both situations. In unstable environments, therefore, “com-mitment” to any simple adaptive strategy is problematic, anda more sophisticated strategy is required.

Already in early hominin evolution, australopithecines wereable to tolerate substantial environmental variability between3.4 and 2.9 mya, during which temperature and rainfall fluc-tuated, and they appear to have done so without favoring oneecosystem over another (Bonnefille et al. 2004). The greatergeographical range of early Homo suggests even greater ca-pacity to tolerate diverse environments. What kinds of adap-tive strategy could have enabled such versatility?

Evolutionary Economics

To investigate the challenge of ecological uncertainty, it isuseful to draw on concepts from the discipline of economics.Economic ideas have long been widely used in biology, givenimplicit or explicit recognition of common ground between“rational decision making” and the process of natural selec-tion (Vermeij 2004). Classic work addressed parental invest-ment (Trivers 1972), evolutionary game theory (MaynardSmith 1982), and optimal foraging theory (MacArthur andPianka 1966), but a “second wave” of evolutionary economicsis now seeing a range of new ideas tested on biological phe-nomena (Hammerstein and Hagen 2005). Many of these ideas

explicitly address the concepts of risk and uncertainty, whichto variable degrees challenge both organisms and financialinstitutions. Such uncertainty requires “decisions” to be madeon the basis of imperfect information. For biological organ-isms, such decisions are shaped by the forces of selection intoa coherent “adaptive strategy” (Laubichler, Hagen, and Ham-merstein 2005).

Consider an environment that flips irregularly between twostates of contrasting quality, such that the phenotypic traitsthat maximize fitness in one environment are very differentto those that maximize fitness in the other. A number ofresponses are possible, depending on the timescale of changeand the availability of cues.

If the rate of flipping is low, each generation of organismscan search for cues of environmental quality and use theseto guide ontogenetic development. Here, risk is addressed byacquiring up-to-date information. Such developmental plas-ticity is certainly important in promoting phenotypic vari-ability, some of it potentially adaptive, among contemporaryhuman populations, such as brain sparing in response to fetalundernutrition (Hales and Barker 1992). An enhanced levelof such plasticity may plausibly have emerged during recenthominin evolution. This hypothesis remains untested, how-ever, because of inadequate information on plasticity in non-human apes.

If the rate of environmental flipping is high, another po-tential response of the organism is “stochastic phenotypeswitching” across generations, more commonly known as “bethedging” (Beaumont et al. 2009; Philippi and Seger 1989).Here, no information is sought; rather, environmental ran-domness is addressed directly by phenotypic randomness. Bethedging may be further subdivided into a conservative strat-egy (avoiding phenotypic extremes and opting for an inter-mediate phenotype) versus a diversified strategy (producingmultiple phenotypes; Philippi and Seger 1989). In each case,the variance in parental fitness is decreased by increasing thelikelihood that at least a subset of offspring will prove adaptedto whatever environment is encountered. Consistent withsuch theory, recent experiments in bacteria have demon-strated the emergence of bet-hedging phenotypes in popu-lations exposed repeatedly to contrasting conditions (Beau-mont et al. 2009).

The success of developmental plasticity and bet hedging isinfluenced by the rate of generation time relative to the rateof environmental flipping. In any ape or hominin past orpresent, requiring at least a decade to grow to reproductivematurity, neither developmental plasticity nor bet hedging aresatisfactory solutions to high levels of ecological stochasticity.Information accessed in early life loses value because it cannotrefer to the multiple ecological states that may occur in suc-cession, while each of the bet-hedged variants may be selectedout when in the “wrong” environment. This is not to suggestthat bet hedging and developmental plasticity do not con-tribute to overall adaptive strategy in humans but simply that


they do not comprise an adequate response to increasingecological stochasticity.

An additional strategy is therefore required by organismswith lengthy developmental periods if they are to tolerate en-vironments with high levels of stochasticity. A key feature ofthis alternative strategy is reversibility; in other words, a re-duction in life course “commitment” to any specific phenotypicstate (Wells and Stock 2012). Phenotypic flexibility meets thisrequirement by increasing the breadth of environmental tol-erance, that is, the range of viable ecological conditions (Gabrielet al. 2005). In the following section, I discuss the generic roleof “capital,” another concept derived from evolutionary eco-nomics, in conferring such phenotypic flexibility.

The Evolutionary Economics of Capital

In seminal work, Kaplan and colleagues (Kaplan, Lancaster,and Robson 2003; Kaplan et al. 2000) proposed that devel-opment be considered as “a process in which individuals andtheir parents invest in a stock of embodied capital” (Kaplanet al. 2000:164). They (Kaplan et al. 2000:164) proposed that“in a physical sense, embodied capital is organized somatictissue. In a functional sense, embodied capital includesstrength, immune function, coordination, skill, knowledge,and social networks, all of which affect the profitability ofallocating time and other resources to alternative activitiessuch as resource acquisition, defense from predators and par-asites, mating competition, parenting and social dominance.”

This broad approach is instrumental in integrating a varietyof raw inputs and phenotypic outputs relevant to life historyvariability. I propose to build on this approach here by em-phasizing two components of embodied capital representingenergy stores. Previously, I differentiated “illiquid” and “liq-uid” capital specifically to distinguish between componentsof embodied capital that represent commitment to a givenphenotype and those that offer reversibility (Wells 2010b).Energy represents liquid capital that may be gained and lostfrom two primary stores, social capital and adipose tissue.Each of these stores allows the conversion of diverse rawsubstrates (e.g., discrete food parcels, behavioral interactions)into generalized energy currencies. While Kaplan and col-leagues (Kaplan, Lancaster, and Robson 2003; Kaplan et al.2000) discussed social networks as an example of embodiedcapital, they paid minimal attention to adipose tissue as afunctional tissue in the same context.

The capacity to store energy buffers strongly against fluc-tuations in energy supply and demand (Pond 1998). Classi-cally, life history strategy is assumed to comprise a series oftrade-offs, most importantly between the functions of main-tenance, growth, reproduction, and immune function, butalso between current and future reproduction (Hill 1993; Zeraand Harshman 2001). The traditional model of resource al-location assumes that investment in one function is at thecost of investment in another because of limited overall re-sources (van Noordwijk and de Jong 1986). Energy stores can

negate such negative correlations on an immediate basis,breaking down the simple association between phenotype andselection. By introducing a time lag between energy incomeand energy utilization, life history decisions may be detachedfrom the immediacy of ecological stochasticity. It is the buf-fering of life history “investment decisions” from selectivepressures that is key to the value of energy capital. Whiledevelopmental plasticity primarily affects the rates of growthand maturation, phenotypic flexibility is of particular im-portance for maintaining flexibility in the trade-off betweencurrent and future reproduction in adult life.

At the level of the cell, energy needs are relatively constantalthough still subject to influences such as physical activityand infection. Cellular needs are maintained by regulatoryprocesses that can balance shortfalls against reserves andhence maintain function even as the environment varies. Clas-sically, for example, ecologists distinguish between “incomebreeders,” which capture the energy needed for reproductiveeffort from the environment on a daily basis, and “capitalbreeders,” which accumulate energy stores before breeding(Jonsson 1997; Stearns 1992). The key difference is that capitalbreeders can breed in a range of ecological circumstancesbecause they are buffered from uncertainty in energy avail-ability (Pond 1984; Stearns 1992). Most attention has focusedon adiposity as the primary source of capital for fundingreproduction, but using the concept of the extended phe-notype (Dawkins 1982), social networks likewise provide en-ergy capital, as in the case of cooperative breeding (Hrdy2009).

Storing energy physically in the body and behaviorally inthe social group represents complementary strategies with dif-ferent implications for addressing uncertainty. Social capitalstored in networks allows energy to be differentially distrib-uted across a group of individuals. At any given time, a singleindividual may fail to achieve daily energy balance withoutexperiencing penalties because of the role of others in sub-sidizing total energy requirements. For periods of high energydemand such as reproduction, this social distribution of en-ergy supply reduces the need of any individual to capturedirectly from the environment or to store within the bodythe total energy required. However, such social stores alsomake possible social competition over access to energy. Anassumed relationship of reciprocity might fail to materializewhen most needed. Equally, an adverse ecological event mightsubstantially reduce the energy content of an entire socialnetwork, placing all individuals at risk. Obligations or prom-ises of producing energy might therefore become relativelyworthless in a low-productivity environment.

Adipose tissue offers greater individual control of energyby maintaining it in physical form but at the cost of havingto transport it as additional body weight. This could increasethe risk of predation as demonstrated empirically in somespecies and hypothesized for hominins (Speakman 2007).Lipid stores allow some functions, or tissues, to fail to achievedaily energy balance, their requirements being met instead


through lipolysis in adipose tissue (Pond 1998). Somatic en-ergy stores are essentially ring-fenced for individual use,though that does not preclude targeting them at particularsocial relationships, for example, investing in offspring. In-creasing internal energy storage is favored when the energyneeds of a social group are heterogeneous, so that the valueof energy likewise differs between individuals.

Generically, organisms occupying more stochastic environ-ments are predicted to benefit from such energy capital, lead-ing to species differences in traits such as adiposity and so-ciality (Pond 1998). On the one hand, energy capital is ahighly effective form of phenotypic flexibility, allowing capitalto be stored or invested in good conditions and used to bufferpoor conditions. This reversibility reduces the need for ex-pensive permanent traits that are only suited to particularecological conditions. On the other hand, the magnitude ofcapital in any given individual may guide life history strategy,as demonstrated in other species (e.g., Liao et al. 2011; Zeraand Zhao 2003). Energy capital may, for example, affect de-velopmental plasticity but may further enable energy parti-tioning between different life history functions during adult-hood, such as meeting the costs of immune function duringinfection and funding reproduction when infection is absent(Wells 2010a). As with other traits reviewed in the introduc-tion, we can therefore consider how capital stores vary withinand between contemporary populations and attempt to re-construct the evolutionary history of capital stores in pastpopulations.

Variability in “Commitment” across Species

The concepts of phenotypic reversibility versus irreversibilitymatch closely with the more established terminology of plas-ticity and canalization. These latter traits are integrally related(Flatt 2005) precisely because the plasticity of some traits inthe face of ecological stresses underlies the relatively invariantphenotype of others. In mammals, for example, growth ofthe brain is relatively protected from environmental stresses,and other traits are more plastic in response to nutritional orclimatic stress (e.g., Barbiro-Michaely et al. 2007). Organismsare therefore more committed to brain growth than they areto other components of growth. Beyond such within-organ-ism contrasts, species also vary among themselves in the ex-tent of genetic specialization versus phenotypic plasticity, thatis, in their commitment to specific niches.

The lack of commitment in the human genome is alreadyevident from comparisons with the genomes of other apes.Although subject to continued gene flow between popula-tions, chimpanzees, gorillas, and orangutans have been con-sidered to comprise two different species or subspecies despiteoccupying relatively small geographic ranges compared withhumans (Becquet and Przeworski 2007; Hey 2010; Warren etal. 2001), and furthermore, the degree of genetic variabilityin each such species is greater than in the entire human ge-nome (Gagneux et al. 1999; Kaessmann et al. 2001). Chim-

panzees have also been subject to more positive selection thanhumans (Bakewell, Shi, and Zhang 2007). While there is in-creasing evidence for multiple recent species of Homo (Homoneanderthalensis, Homo floriensis; Brown et al. 2004) and anew type recently reported from Siberia (Krause et al. 2010;Reich et al. 2010), only one species exists in the contemporaryworld. Although a reduced level of genetic variation is pre-dicted in modern humans given our relatively recent speci-ation (Cann, Stoneking, and Wilson 1987) and potentiallydue to culturally mediated migration patterns (Premo andHublin 2009), a proportion of the genetic variability of otherHomo species is also assumed to have passed into our ownspecies (Garrigan and Hammer 2006; Green et al. 2010; Reichet al. 2010).

Regardless of the timescale and mode through which ge-netic variability might have accumulated, what is arguablymost notable in humans is our relatively low level of geneticdifferentiation in proportion to the extraordinary diversity ofhabitats we now occupy. This contrasts markedly with muchgreater genetic differentiation in apes despite their inhabitingrelatively homogeneous environments (Gagneux et al. 1999).The relative genetic unity in Homo sapiens implies continuedinterbreeding between regional populations even as they oc-cupied diverse habitats and ecological niches. Indeed, somestudies also suggest a degree of recent interbreeding with othercontemporaneous Homo species (Green et al. 2010; Reich etal. 2010; though see Eriksson and Manica 2012).

It is clear that some adaptive strategies have become in-corporated into the human genome, such as the elongationof growth and its division into several distinct developmentalperiods (Bogin and Smith 1996), that therefore representcommitment to a generic human niche. Beyond this com-mitment, the within-population genetic variability that is evi-dent in any human life history trait (e.g., stature, birth weight,age at menarche, longevity; see Wells and Stock 2011) maybe considered to represent diversified bet hedging—that is,the maintenance of phenotypic diversity among offspring toincrease parental fitness (Ellis et al. 2009; Wells 2009)—whileeach of these life history traits also shows substantial plasticity(Wells and Stock 2011). However, my argument here is thatthe limited genetic commitment of humans is complementednot only by developmental plasticity but also by extensive useof energy stores that may be allocated flexibly between com-peting phenotypic traits and between individuals and hencealleviate ecological stochasticity. The next section considersthe contribution of energy capital to phenotypic flexibility innonhuman primates.

Capital and Flexibility in Nonhuman Primates

Primates have high levels of sociality, and their tendency tosolve ecological problems in social ways is well established(Hrdy 2009; Lee 1999; Sussman, Garber, and Cheverud 2005).A review of primate activity budgets suggested the proportionof time directed to sociality ranged from 1% to 8% in pro-


simians, from1% to 22% in New World monkeys, from 1%to 28% in Old World monkeys, and from 2% to 25% in apes(Sussman, Garber, and Cheverud 2005). Importantly, the ma-jority of such social behavior was cooperative and affiliativerather than antagonistic.

Individuals may invest in social relationships and accu-mulate social capital that may be converted back into bene-ficial physical resources (Isler and van Schaik 2009). For ex-ample, female chacma baboons who formed stronger andmore stable social bonds with other female baboons hadgreater longevity (Silk et al. 2010) and also produced offspringwho themselves had greater longevity through increasing therates of offspring survival (Silk et al. 2009). Similarly, socialrelationships may benefit male fitness, as demonstrated inAssamese macaques, where strong bonds between males werelinked to the formation of coalitions, dominance rank, andthe number of offspring sired (Schulke et al. 2010).

Such social investments make their maximal contributionto flexibility when they enable cooperative breeding, a be-havior evident in some primate species as well as other mam-mals and birds (Isler and van Schaik 2009) and involving acontinuum of behavior ranging from carrying or protectingoffspring to “babysitting,” provisioning, or suckling them(Hrdy 2004, 2009). Here, social capital contributes directly tothe capacity of reproducing females to meet the total energydemands of their offspring. In Hanuman langurs, infants wereobserved to be carried by females other than the mother forup to half of daylight hours, which has sufficient effect onthe maternal energy budget to reduce the time to the nextconception (Hrdy 2004). Among marmosets and tamarins,males may carry offspring for the majority of the time andalso supplement maternal milk in the offspring diet withsmall-prey items (Garber 1997). Collectively, maternal fitnessin these species is positively correlated with the number ofmale helpers, and in cotton-top tamarins, such help appearsobligatory, with mothers lacking assistance tending to aban-don their young (Hrdy 2004). Such cooperative behavior notonly increases the flexibility in provisioning offspring but alsopermits flexibility in mating arrangements (Hrdy 2004).

Such social interactions reduce the strength with whichenergy turnover is a function of a single organism’s body sizeand instead increase the strength with which it reflects socialrelationships (Kramer and Ellison 2010; Reiches et al. 2009).Recent exploration of such communal energy budgets hasemphasized “activity subsidies” (e.g., babysitting) as well asredistributions of food and multiple and changing contri-butions from ascendant and descendant kin across the lifecourse (Kramer and Ellison 2010; Reiches et al. 2009). Byenabling energy dynamics to be temporarily invested extra-corporeally, social relationships promote maximal flexibilityin reproductive energetics.

Such energy transfers need not necessarily be maintainedin social relationships and may instead be converted to phys-ical capital, either extracorporeal, in the form of food hoards,or physiological, in the form of adipose tissue. Extracorporeal

food stores are known in some bird and mammal species,while adiposity is a widespread somatic adaptation to eco-logical uncertainty (Pond 1984, 1998).

In primates, the use of adipose tissue to resolve seasonalfluctuations in energy availability was elegantly demonstratedin a study of orangutans where urinary ketones indicated themetabolism of fat during periods of negative energy balance(Knott 1998). The highest levels of ketones were observed inpregnant or lactating females, highlighting the value of adi-pose tissue in buffering reproduction from ecological sto-chasticity. Greater adiposity or leptin levels have been reportedin females compared with males in baboons (Rutenberg etal. 1987), macaques (Schwartz and Kemnitz 1992), chimpan-zees (Bribiescas and Anestis 2010), and orangutans (Morbeckand Zihlman 1988), but information on adiposity and itsregulatory effects remains surprisingly scarce in the primateliterature considering how important the trait appears inhominin evolution (Wells 2010a). Intriguingly, the males ofsome marmoset and tamarin species have been shown to gainweight during their mate’s pregnancy, a strategy suggested asadaptive given their key alloparental role after delivery whenthey may carry offspring equivalent to 20% of their own bodyweight (Ziegler et al. 2006).

Importantly, a rapidly growing literature has demonstratedthat adipose tissue is not only a fuel store but also the sourceof numerous signaling molecules that coordinate energytrade-offs between competing biological functions such asgrowth, immune function, and reproduction (Wells 2010a).The role of adipose tissue in such functions has been dem-onstrated in a variety of species (Bartness, Demas, and Song2002; Demas 2004; Pond 1998). Among the hormones se-creted by adipose tissue is leptin, signaling the availability ofenergy to the brain, and studies in a variety of species haveshown that it helps regulate the allocation of energy betweencompeting functions (Demas and Sakaria 2005; Drazen, De-mas, and Nelson 2001; Schneider 2004). Adipose tissue alsoprovides both the energy and the molecular precursors forimmune function, and it does so in tissue-specific fashion,indicating differential preparation of specific tissues for dif-ferent types of pathogen (Pond 2003). The emergence of ad-ipose tissue has been proposed to have aided control over thedynamics of energy supply and demand between tissues, forexample, to allow high energy processes such as lactation tooccur during periods of minimal energy intake in primitivemammals (Pond 1984).

These data from primates and other mammals illustratefor both sexes how accumulating energy capital can conferdiverse fitness benefits while increasing resilience to ecologicalfactors operating at the level of the individual. However, howthese strategic decisions are enacted at the molecular level indiverse species remains to be established in detail. Impor-tantly, the use of energy capital by marmosets and tamarinsaffects not only immediate reproduction but also demogra-phy. These primates have a powerful ability to increase pop-ulation size and colonize new niches (Hrdy 2009), indicating


that the adaptive strategy of phenotypic flexibility that canbuffer individuals from energy stress in harsh conditions canfurthermore drive population growth beyond that feasible inindividual income breeders in good conditions.

Capital Investment and Humans

Recent work has highlighted a variety of ways in which thestorage, mobilization, and transfer of capital within and be-tween individuals confer significant flexibility in contempo-rary humans. Adipose tissue can supply energy for growth,reproduction, and immune function, and a number of hor-mones have been shown to contribute to this strategic par-titioning of energy. Paradoxically, the evidence that leptinlevels correlate with energy stores in humans and hence or-chestrate such trade-offs is currently inconsistent and bettersupported in females than in males (Bribiescas 2001; Kuzawa,Quinn, and Adair 2007; Sharrock et al. 2008). It is likely thatleptin makes up only one of a range of molecular signalsinvolved in the regulation of biological functions (Schneider2004); for example, insulin is increasingly recognized to playa key role in coordinating life history trade-offs (Harshmanand Zera 2007; Watve and Yajnik 2007).

The life course trajectory of human body composition in-dicates the changing value of fat during development, breed-ing, and aging. High neonatal adiposity of humans comparedwith other mammals has been hypothesized to buffer theobligatory energy demands of the large Homo brain (Kuzawa1998). However, this need not represent the only or even theprimary function of such adiposity. Unpublished data fromEthiopia ( ) show that over the first 6 months, infantsn p 255first increase disproportionately in adiposity and then accretedisproportionately lean mass (Andersen, Friis, and Wells, un-published data). The magnitude of infant energy stores maytherefore guide the early accretion of lean tissue, which hasless capacity to oscillate in response to energy fluctuations.Adipose tissue may be important in supporting, regulating,and protecting each of somatic growth, immune function,and brain development during infancy when mortality risk ishighest (Kelly 1995) while furthermore accommodating theinitiation of sex differences in life history strategy. Becauseearly energy supply derives primarily from the maternal bud-get, such regulatory effects of adipose tissue are ultimately atransgenerational maternal effect (Ong et al. 2007; Wells2010b), as discussed below.

During development, the two sexes show contrasting so-matic capital accumulation, with adult males achieving greaterheight and relative lean mass and less fat mass than females(Wells 2007b). These differences in adiposity are particularlyevident for fat distribution, with the enhanced peripheral fatof females both signaling reproductive capacity and meetingthe energy costs of lactation (Wells 2010a). Humans are capitalbreeders and hence can breed across a range of ecologicalconditions and seasons, though in good conditions they usedaily “energy income” rather than drawing on energy stores.

Through adulthood, gender differences in fat distributionslowly decrease, so that by old age the genders converge inbody shape, and central adiposity in both sexes is increasedrelative to peripheral adiposity (Wells, Cole, and Treleaven2008). Both sexes lose lean mass such that in later life thebody becomes more “economical” (Zafon 2007). These shiftsin fat distribution have been interpreted as life history strategyin which increased prioritization of central adiposity withincreasing age promotes immune function at the expense ofreproduction (Wells, Griffin, and Treleaven 2010).

The ratio of fat to lean tissue therefore emerges as a keyadaptive strategy in which growth can be tailored to ecologicalconditions under the transducing effect of maternal pheno-type. The magnitude of lean mass has long-term implicationsfor energy requirements, while the magnitude of adipositycan respond by providing appropriate “risk management”(Wells 2012c). For example, the low birth weight characteristicof contemporary Indian populations, poorly nourished formany generations, incorporates a low level of investment inlean mass balanced by an increased investment in adiposity(Yajnik et al. 2003). Collectively, these traits represent an eco-nomical and resilient phenotype capable of tolerating highuncertainty in infant energy supply.

More broadly, many environmental factors generate vari-ability in the fat-lean ratio, indicating that the strategy foraccumulating and using energy capital is sensitive to multiplelife course and ecological signals, including physical (e.g., cli-mate), ecological (e.g., diet, disease), social (e.g., maternalrank), and demographic (e.g., parity, interbirth interval) fac-tors (Wells 2011, 2012b). The magnitude of sexual dimor-phism in adiposity has likewise further been shown to varyin relation to ecological factors (Wells 2012d). Capital ac-quisition also reflects a multitude of dynamic social inter-actions in which energy is passed to and from individualsthrough a complex set of relationships. One might considerthe high adiposity of infants as a temporary accumulation ofcapital favored by the sum total of maternal and alloparentalprovisioning during pregnancy, a source of income less guar-anteed as the offspring matures. Thus, energy capital may berapidly moved between individuals not only to boost thestores of some individuals but also as “loans” physically pro-tected from those with competing interests.

Research on diverse human populations has demonstratedan extraordinary degree of variability in the nature and structureof relationships that underlie the distribution of energy capital(Kaplan, Lancaster, and Robson 2003; Sear and Mace 2008).In 10 foraging societies, on average men acquired ∼68% of thecalories and ∼88% of the protein and hence were by far thedominant supplier of calories to weaned offspring as well asprotein and fat to women, though there was also variabilitybetween populations in these contributions (Kaplan et al. 2000).This highlights the sexual division of labor as another key com-ponent of human flexibility in capital acquisition. Hrdy (2009)argued that while male provisioning is critical for cooperativebreeding, the specific contribution of the father cannot be guar-


Figure 1. Alternative and complementary strategies for accom-modating ecological stochasticity. The brain integrates social andbehavioral information, influencing physiological regulatory sys-tems. Adipose tissue integrates ecological information and influ-ences on neuroendocrine regulatory systems. Both of these organsprovide “risk management” of life history strategy in the face ofecological stochasticity.

anteed, resulting in a range of strategies for mothers to acquire“additional fathers” and hence bet-hedge across multiple maleproviders. Similarly, in relation to females, maternal grand-mothers have been shown in several societies to be importantalloparental provisioners of grandoffspring and hence to pro-mote their survival (Hawkes, O’Connell, and Blurton Jones1989; Sear and Mace 2008).

Cooperative breeding and pooled energy budgets thereforeoffer unprecedented flexibility in humans as they impose min-imal constraints on the social relationships generating energysupply at the individual level (Hrdy 2009; Reiches et al. 2009).More broadly, social capital may be distributed across multiplesocial relationships, which may be accumulated through thelife course. No single social relationship may prove a reliablesource of support in every circumstance; hence, bet hedgingacross multiple relationships may again prove the optimalstrategy. Because a group of humans must inevitably competefor the sum total of available energy in the local environment,networks of social support can never be entirely benign butmust rather contain a variety of different relationships, somebased on kinship and others on political alliances. For ex-ample, studies of the !Kung bushmen of southern Africa showthat relationships of reciprocity are key to the managementof nutritional risk (Howell 2010), a scenario also observed inmany other nonindustrialized societies (Couper-Johnston2000). Such social capital distributes risk across rather thanwithin individual bodies and hence buffers uncertainty inindividual foraging returns.

The combination of variable dynamics of cooperativebreeding and life course flexibility in adipose tissue storesmeans that humans are uniquely undercommitted to any spe-cific niche, which helps us to understand our unusual capacityto colonize. Such undercommitment is backed up by othergeneralized subsistence strategies such as meat eating andcooking (Foley 2001; Sullivan, Hagen, and Hammerstein2008), each of which reduces the need to adapt physiologicallyto specific diets. Social capital likewise underpins a wide va-riety of human relationships, including coalitions and alli-ances and other forms of cooperative behavior.

In the context of ecological uncertainty, one could arguethat energy capital is not “for” anything in particular. Itspurpose might be seen instead as a way of avoiding com-mitting energy to any specific function at any specific time,instead providing sophisticated “risk management” of life his-tory strategy by integrating a cumulative array of individualsignals from the physical and social environment. The brainand adipose tissues both play a part in such integrative reg-ulation and hence emerge as important “control centers” forlife history strategy (Wells 2010a, 2012c). Intriguingly, eachof these organs acts strongly on the other (fig. 1).

With respect to the brain, recent research on the neuroen-docrine basis of social behavior has demonstrated the effectsof a range of hormones; for example, oxytocin plays a keyrole in cooperative behavior, while neuropeptide Y plays anequally fundamental role in the stress response (Hirsch and

Zukowska 2012; Soares et al. 2010). Each of these hormonesin turn affects adiposity and metabolic traits such as appetite(Hirsch and Zukowska 2012; Maejima et al. 2011). Thus, thebrain contributes to the regulation of life history strategy anddoes so in response to social and ecological stresses.

With respect to adipose tissue, hormones such as leptinand insulin signal the level and consistency of lipid stores tothe brain (Benoit et al. 2004) and in doing so affect manylife history functions, including reproductive behavior (Cun-ningham, Clifton, and Steiner 1999; Watve and Yajnik 2007).Preliminary evidence from rodents suggests that levels of ad-iposity may influence oxytocin (Flak et al. 2011) and hencepotentially social behavior. Thus, while further research isrequired, it is already apparent that adipose tissue likewisecontributes to the regulation of life history strategy and againdoes so in response to diverse ecological stresses.

Clearly these two risk-management systems differ in theirreceptivity to specific signals, their timescale of response, theirdegree of flexibility, and their functional targets. Together,however, they represent an important link between the levelof capital and life history strategy and thereby highlight theimportance of capital from an evolutionary perspective.

Capital in Hominins

Reconstructing energy capital dynamics in hominins is atpresent difficult because of the potential transiency of energycapital per se and the lack of clear markers in the fossil record.Phenotypic flexibility to some extent must remain hiddenfrom preservation.

Intriguingly, however, both adipose tissue and bone are


Figure 2. Positive feedback cycle in which the capacity to toleratestochastic environments promotes the capacity to colonize newenvironments. Both social and physical forms of energy capitalmay have provided the necessary phenotypic flexibility.

increasingly being reinterpreted as dynamic rather than inerttissues, and growing awareness of endocrine links betweenthese tissues may aid novel interpretation of the fossil recordin due course. For example, bone is now known to contributeto the regulation of energy metabolism (Fernandez-Real andRicart 2011), while leptin mediates skeletal turnover (Ducyet al. 2000). Osteocalcin, which can be recovered from somefossilized material (Nielsen-Marsh et al. 2002), shows an in-verse correlation with adiposity in living humans (Fernandez-Real and Ricart 2011). Further work in this area may generatemore robust findings. Given that both adiposity and socialcapital vary substantially within and between contemporaryhuman populations, we can assume that they representeddynamic adaptive strategies in Homo evolution, but the detailsremain to be established (see Wells 2010a for a preliminaryestimation of adiposity in prehuman hominins).

Both adiposity and social capital enable variation in themode and breadth of environmental tolerance; that is, theycan vary in relation to average environmental conditions (e.g.,hot vs. cold climates) but also provide phenotypic flexibilityto buffer variability in any given environment. Recently, Na-varrete, van Schaik, and Isler (2011) reported an inverse cor-relation across ∼100 mammal species between brain and ad-ipose tissue masses. These authors suggested that storing in-formation in the brain or energy in the body represent twoalternative strategies for accommodating uncertainty in en-ergy availability. The fact that our own species is a positiveoutlier for both traits offers a powerful indication that eco-logical stochasticity has been a key stress in the evolution ofthe genus Homo.

In turn, the phenotypic flexibility that emerged under suchselective pressures appears to have conferred on Homo anextraordinary capacity to colonize new niches (Wells andStock 2007). The energy capital that provides resilience againsttough conditions also favors successful probing of novelniches and demographic expansion as demonstrated by rapidrecovery from population crashes (Hill and Hurtado 1996).Through positive feedback loops, colonizing, with its popu-lation booms and busts, can become its own selective pressure,as shown in figure 2. Thus, withstanding “ecological feast andfamine” may have made possible “reproductive feast and fam-ine.”

Cooperative breeding is also notable in many social car-nivores (Smith et al. 2012), which likewise benefit from asocial means of buffering individuals’ risk of failure in cap-turing prey. According to this perspective, the emergence ofenergy capital as an adaptive response to ecological uncer-tainty in early Homo may have opened up carnivory as a newsubsistence niche, making use of two types of bet hedging:across individuals (e.g., buffering variance in productivity be-tween individual hunters or gatherers) and across the genders(diversifying the resources harvested by the whole group).Similarly, cooperative breeding may have made possible thefunding of larger brains (Hrdy 2009; Isler and van Schaik2012).

Capital and Transgenerational Effects

As discussed above, capital reduces exposure to ecologicalstresses and hence to selection, which aids in understandingour species’ limited local genetic commitment. However, theimplications of energy capital in Homo go much further be-cause of the particular sensitivity of developmental trajectoryto environmental factors operating in early life.

As in most mammals, hominin fetuses must initially ex-perience ecological pressures under the transducing effect ofmaternal phenotype (Wells 2003, 2007c). Most perspectiveson developmental plasticity emphasize how developing off-spring adapt to the ecological environment (e.g., Bateson2001; Kuzawa 2005). Some argue that offspring use environ-mental cues adaptively by anticipating the future environmentin which breeding will occur (Gluckman and Hanson 2004;Gluckman, Hanson, and Beedle 2007). This model fits poorlywith the observation that human mothers substantially buffersuch cues through both physiological and social mechanisms(Wells 2003), and the anticipatory model has been criticizedby several authors (Bogin, Varela Silva, and Rios 2007; Rickardand Lummaa 2007; Wells 2007a, 2011, 2012a).

I have argued that given ecological stochasticity, offspringshould respond not to short-term external cues but to ma-ternal phenotype, which represents the cumulative effect of“an extensive period of time, with short-term fluctuationssmoothed out to provide a more reliable rating of environ-mental quality” (Wells 2003:152). This smoothing effect wassubsequently labeled “intergenerational inertia” by Kuzawa(2005), but Kuzawa’s approach, like that of Gluckman andHanson, does not address the dynamics deriving from parent-offspring conflicts of interest (Haig 1993, 1997; Trivers 1974),which these authors reject (Gluckman et al. 2009).

Clearly the offspring requires information in order to selecta developmental trajectory, but in the absence of direct cuesfrom the external environment, the offspring must use thosederiving from maternal phenotype. In accepting such cues,however, the offspring must also submit to maternal strategy(Wells 2003, 2010a), and this strategy is particularly important


in human life history because it initially falls to the motherto meet the high costs of offspring brain growth (Aiello andKey 2002; Martin 1983, 1989) and to allocate resources be-tween several competing offspring during much of the re-productive career (Wells 2003). Closure of the critical windowof physiological sensitivity in early infancy, despite the ex-tended period of growth, results in the canalization of off-spring growth trajectory in the second year of life (Smith etal. 1976) and allows the mother to “fix” her own investmentstrategy in each offspring at the time of weaning (Wells 2003).

The cues received by developing offspring before weaningtherefore refer not to ecological conditions per se but to themagnitude of maternal capital (Wells 2010a), incorporatingboth social (i.e., extended phenotype) and physical dimen-sions. Receiving a given investment of capital need not implyany anticipation of future ecological conditions; it is simplywhat is available according to maternal resources and in-vestment strategy. If developing offspring adapt to maternalphenotype rather than to ecological conditions, this furtheremphasizes a lack of long-term commitment in adaptation.Lineages, especially matrilines, may accumulate capital andtransfer it across generations (see Smith et al. 2010 regardingtransgenerational transmission of body weight in foragers)using the flexibility of capital stores to buffer short-term eco-logical perturbations (Wells 2003, 2010a). Evolutionary mod-els of transgenerational information transfer based entirely onexternal ecological signals offer insufficient explanation forthe existence of phenotypic variability within a population atany given point in time and provide a simplistic model ofthe process of adaptation.

In stochastic environments, developmental strategies se-lected in early life have a high risk of being maladaptive duringsome or all of adulthood. These potential costs may be al-leviated in part by maternal buffering in early life and in partby reversible phenotypic plasticity subsequently conferred byadiposity and social strategies. Energy capital plays a key rolein both processes.

Conclusions

Accumulating social and physiological capital represents twokey modes of phenotypic flexibility in which life history de-cisions may be buffered from ecological stochasticity. The twostrategies also contrast in an intriguing way, with one per-taining to the brain and, over evolutionary time, to increas-ingly purposive strategy formation and the other to physi-ology, in which energy stores themselves enact strategicdecisions across competing biological function. These traitsthat proved adaptive in resolving ecological uncertainty ap-pear to have given rise subsequently to large brains and thecapacity to colonize novel habitats, now recognized as twoquintessential human traits. It is arguably such phenotypicflexibility and the lack of genetic commitment that maintainedmodern humans as a single species across diverse ecologies.

Acknowledgments

I very much appreciate the invitation from Susan Anton andLeslie Aiello to participate in this workshop and their criticalcomments on the manuscript as well as the comprehensivediscussions with the other participants.


Homo erectus female. American Journal of Human Biology 14:551–565.Anton, Susan C. 2003. Natural history of Homo erectus. Yearbook of Physical

Anthropology 37:126–170.Anton, Susan C., Hannah Carter-Menn, and Valerie B. DeLeon. 2011. Modern

human origins: continuity, replacement, and masticatory robusticity in Aus-tralasia. Journal of Human Evolution 60:70–82.


Bakewell, Margaret A., Peng Shi, and Jianzhi Zhang. 2007. More genes un-derwent positive selection in chimpanzee evolution than in human evo-lution. Proceedings of the National Academy of Sciences of the USA 104:7489–7494.

Barbiro-Michaely, Efrat, Michael Tolmasov, Shunit Rinkevich-Shop, JudithSonn, and Avraham Mayevsky. 2007. Can the “brain-sparing effect” bedetected in a small-animal model? Medical Science Monitor 13:BR211–BR219.

Bartness, Timothy J., Gregory E. Demas, and C. Kay Song. 2002. Seasonalchanges in adiposity: the roles of the photoperiod, melatonin and otherhormones, and sympathetic nervous system. Experimental Biology and Med-icine 227:363–376.

Bateson, Patrick. 2001. Fetal experience and good adult design. InternationalJournal of Epidemiology 30:928–934.

Beaumont, Hubertus J., Jenna Gallie, Christian Kost, Gayle C. Ferguson, andPaul B. Rainey. 2009. Experimental evolution of bet hedging. Nature 462:90–93.

Becquet, Celine, and Molly Przeworski. 2007. A new approach to estimateparameters of speciation models with application to apes. Genome Research17:1505–1519.

Benoit, Stephen C., Deborah J. Clegg, Randy J. Seeley, and Stephen C. Woods.2004. Insulin and leptin as adiposity signals. Recent Progress in HormoneResearch 59:267–285.

Bobe, Rene, and Anna K. Behrensmeyer. 2004. The expansion of grasslandecosystems in Africa in relation to mammalian evolution and the origin ofthe genus Homo. Palaeogeography, Palaeoclimatology, Palaeoecology 207:399–420.


Bogin, Barry, Patricia Smith, Alicia Bibiana Orden, Maria I. Varela Silva, andJames Loucky. 2002. Rapid change in height and body proportions of MayaAmerican children. American Journal of Human Biology 14:753–761.

Bogin, Barry, Maria I. Varela Silva, and Luis Rios. 2007. Life history trade-offs in human growth: adaptation or pathology? American Journal of HumanBiology 19:631–642.

Bonnefille, Raymonde, Richard Potts, Francoise Chalie, Dominique Jolly, andOdile Peyron. 2004. High-resolution vegetation and climate change asso-ciated with Pliocene Australopithecus afarensis. Proceedings of the NationalAcademy of Sciences of the USA 101:12125–12129.


Bribiescas, Richard G. 2001. Serum leptin levels and anthropometric correlatesin Ache Amerindians of eastern Paraguay. American Journal of PhysicalAnthropology 115:297–303.

Bribiescas, Richard G., and Stephanie F. Anestis. 2010. Leptin associationswith age, weight, and sex among chimpanzees (Pan troglodytes). Journal ofMedical Primatology 39:347–355.

Brown, Peter, Thomas Sutikna, Michael J. Morwood, Raden P. Soejono, Jatmiko,E. Wayhu Saptomo, and Rokus Awe Due. 2004. A new small-bodied homininfrom the Late Pleistocene of Flores, Indonesia. Nature 431:1055–1061.

Bruner, Emiliano, Giorgio Manzi, and Juan Luis Arsuaga. 2003. Encephali-zation and allometric trajectories in the genus Homo: evidence from the


Neandertal and modern lineages. Proceedings of the National Academy ofSciences of the USA 100:15335–15340.

Cann, Rebecca L., Mark Stoneking, and Allan C. Wilson. 1987. MitochondrialDNA and human evolution. Nature 325:31–36.

Cochran, Gregory, and Henry Harpending. 2009. The 10,000 year explosion:how civilization accelerated human evolution. New York: Basic.

Couper-Johnston, Ross. 2000. El Nino: the weather phenomenon that changedthe world. London: Hodder & Stoughton.

Crompton, Robin H., William I. Sellers, and Susannah K. Thorpe. 2010.Arboreality, terrestriality and bipedalism. Philosophical Transactions of theRoyal Society B 365:3301–3314.

Cunningham, Matthew J., Donald K. Clifton, and Robert A. Steiner. 1999.Leptin’s actions on the reproductive axis: perspectives and mechanisms.Biology of Reproduction 60:216–222.

Dart, Raymond. 1925. Australopithecus africanus: the man-ape from SouthAfrica. Nature 115:195–199.

Dawkins, Richard. 1982. The extended phenotype. Oxford: Oxford UniversityPress.

Demas, Gregory E. 2004. The energetics of immunity: a neuroendocrine linkbetween energy balance and immune function. Hormones and Behavior 45:173–180.

Demas, Gregory E., and Sangeeta Sakaria. 2005. Leptin regulates energetictradeoffs between body fat and humoural immunity. Proceedings of the RoyalSociety B 272:1845–1850.


Drazen, Deborah L., Gregory E. Demas, and Randy J. Nelson. 2001. Leptineffects on immune function and energy balance are photoperiod dependentin Siberian hamsters (Phodopus sungorus). Endocrinology 142:2768–2775.

Ducy, Patricia, Michael Amling, Shu Takeda, Matthias Priemel, Arndt F. Schil-ling, Frank T. Beil, Jianhe Shen, Charles Vinson, Johannes M. Rueger, andGerard Karsenty. 2000. Leptin inhibits bone formation through a hypo-thalamic relay: a central control of bone mass. Cell 100:197–207.

Elks, Cathy E., John R. Perry, Patrick Sulem, Daniel I. Chasman, Nora Fran-ceschini, Chunyan He, Kathryn L. Lunetta, et al. 2010. Thirty new loci forage at menarche identified by a meta-analysis of genome-wide associationstudies. Nature Genetics 42(12):1077–1085.

Ellis, Bruce J., Aurelio J. Figueredo, Barbara H. Brumbach, and Gabriel L.Scholmer. 2009. Fundamental dimensions of environmental risk: the impactof harsh versus unpredictable environments on the evolution and devel-opment of life history strategies. Human Nature 20:204–268.

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(35):13956–13960.

Fernandez-Real, Jose M., and Wifredo Ricart. 2011. Osteocalcin: a new linkbetween bone and energy metabolism. Some evolutionary clues. CurrentOpinion in Clinical Nutrition and Metabolic Care 14:360–366.

Flak, Jonathan N., Ryan Jankord, Matia B. Solomon, Eric G. Krause, andJames P. Herman. 2011. Opposing effects of chronic stress and weightrestriction on cardiovascular, neuroendocrine and metabolic function. Phys-iology and Behavior 104:228–234.

Flatt, Thomas. 2005. The evolutionary genetics of canalization. Quarterly Re-view of Biology 80:287–316.

Foley, Robert. 2001. The evolutionary consequences of increased carnivory inhumans. In Meat-eating and human evolution. Craig B. Stanford and HenryT. Bunn, eds. Pp. 305–331. Oxford: Oxford University Press.

Fraga, Mario F., Esteban Ballestar, Maria F. Paz, Santiago Ropero, FernandoSetien, Maria L. Ballestar, Damia Heine-Suner, et al. 2005. Epigenetic dif-ferences arise during the lifetime of monozygotic twins. Proceedings of theNational Academy of Sciences of the USA 102:10604–10609.

Gabriel, Wilfried, Barney Luttbeg, Andrew Sih, and Ralph Tollrian. 2005.Environmental tolerance, heterogeneity, and the evolution of reversible plas-tic responses. American Naturalist 166:339–353.

Gagneux, Pascal, Christopher Wills, Ulrike Gerloff, Diethard Tautz, Phillip A.Morin, Christophe Boesch, Barbara Fruth, Gottfried Hohmann, Olivia A.Ryder, and David S. Woodruff. 1999. Mitochondrial sequences show diverseevolutionary histories of African hominoids. Proceedings of the NationalAcademy of Sciences of the USA 96:5077–5082.

Garber, Paul A. 1997. One for all and breeding for one: cooperation andcompetition as a tamarin reproductive strategy. Evolutionary Anthropology5:187–199.

Garrigan, Daniel, and Michael F. Hammer. 2006. Reconstructing human or-igins in the genomic era. Nature Reviews Genetics 7:669–680.

Gibbs, Sally, Mark Collard, and Bernard Wood. 2000. Soft-tissue charactersin higher primate phylogenetics. Proceedings of the National Academy ofSciences of the USA 97:11130–11132.

Gilbert, Christopher C., and James B. Rossie. 2007. Congruence of moleculesand morphology using a narrow allometric approach. Proceedings of theNational Academy of Sciences of the USA 104:11910–11914.

Gluckman, Peter, Keith Godfrey, Mark Hanson, and Christopher Kuzawa.2009. Developmental aspects of diseases of modern environments. Evolutionand Medicine Reviews, October 28. http://evmedreview.com/?pp166.

Gluckman, Peter D., and Mark A. Hanson. 2004. The developmental originsof the metabolic syndrome. Trends in Endocrinology and Metabolism 15:183–187.

Gluckman, Peter D., Mark A. Hanson, and Alan S. Beedle. 2007. Early lifeevents and their consequences for later disease: a life history and evolu-tionary perspective. American Journal of Human Biology 19:1–19.

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.

Haig, David. 1993. Genetic conflicts in human pregnancy. Quarterly Reviewof Biology 68:495–532.

———. 1997. Parental antagonism, relatedness asymmetries, and genomicimprinting. Proceedings of the Royal Society B 264:1657–1662.

Hales, C. Nicolas, and David J. Barker. 1992. Type 2 (non-insulin-dependent)diabetes mellitus: the thrifty phenotype hypothesis. Diabetologia 35:595–601.

Hammerstein, Peter, and Edward H. Hagen. 2005. The second wave of evo-lutionary economics in biology. Trends in Ecology & Evolution 20:604–609.

Hancock, Angela M., Gorka Alkorta-Aranburu, David B. Witonsky, and AnnaDi Rienzo. 2010. Adaptations to new environments in humans: the role ofsubtle allele frequency shifts. Philosophical Transactions of the Royal SocietyB 365:2459–2468.

Hancock, Angela M., David B. Witonsky, Gorka Alkorta-Aranburu, CynthiaM. Beall, Amha Gebremedhin, Rem Sukernik, Gerd Utermann, JonathanK. Pritchard, Graham Coop, and Anna Di Rienzo. 2011. Adaptations toclimate-mediated selective pressures in humans. PLoS Genetics 7:e1001375.

Harshman, Lawrence G., and Anthony J. Zera. 2007. The cost of reproduction:the devil in the details. Trends in Ecology & Evolution 22:80–86.

Hawkes, Kristen. 2006. Slow life histories and human evolution. In The evo-lution of human life history. Kristen Hawkes and Richard R. Paine, eds. Pp.95–126. Oxford: Currey.

Hawkes, Kristen, James F. O’Connell, and Nicholas G. Blurton Jones. 1989.Hardworking Hadza grandmothers. Comparative Socioecology of Mammalsand Man 8:341–366.

Herring, D. Ann, Shelley R. Saunders, and M. Anne Katzenberg. 1998. In-vestigating the weaning process in past populations. American Journal ofPhysical Anthropology 105:425–439.

Hey, Jody. 2010. The divergence of chimpanzee species and subspecies asrevealed in multipopulation isolation-with-migration analyses. MolecularBiology and Evolution 27:921–933.

Hiernaux, Jean, and Alain Froment. 1976. The correlations between anthro-pobiological and climatic variables in sub-Saharan Africa: revised estimates.Human Biology 48:757–767.

Hill, Kim. 1993. Life history theory and evolutionary anthropology. Evolu-tionary Anthropology 2:78–88.

Hill, Kim, and Magdalena Hurtado. 1996. Ache life history: the ecology anddemography of a foraging people. New York: Aldine.

Hirsch, Dalay, and Zofia Zukowska. 2012. NPY and stress 30 years later: theperipheral view. Cellular and Molecular Neurobiology 32:645–659.

Hirschhorn, Joel N., and Guillaume Lettre. 2009. Progress in genome-wideassociation studies of human height. Hormone Research 71(suppl. 2):5–13.

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

Houston, Alisdair I., and John M. McNamara. 1999. Models of adaptive be-haviour: an approach based on state. Cambridge: Cambridge University Press.

Howell, Nancy. 2010. Life histories of the Dobe !Kung: food, fatness, and well-being over the life-span. Berkeley: University of California Press.

Hrdy, Sarah B. 2004. Comes the child before the man: how cooperative breed-ing and prolonged postweaning dependence shaped human potentials. InHunter-gatherer childhoods: evolutionary, developmental, and cultural per-spectives. Barry S. Hewlett and Michael E. Lamb, eds. Pp. 65–91. Somerset,NJ: Transaction/Aldine.

http://evmedreview.com/?p=166


———. 2009. Mothers and others: the evolutionary origins of mutual under-standing. Cambridge, MA: Belknap.

Isler, Karin, and Carel P. van Schaik. 2009. Why are there so few smartmammals (but so many smart birds)? Biology Letters 5:125–129.

———. 2012. How our ancestors broke through the gray ceiling: comparativeevidence for cooperative breeding in early Homo. Current Anthropology53(suppl. 6:S453–S465.

Jonsson, K. Ingemar. 1997. Capital and income breeding as alternative tacticsof resource use in reproduction. Oikos 78:57–66.

Kaessmann, Henrik, Victor Wiebe, Gunter Weiss, and Svante Paabo. 2001.Great ape DNA sequences reveal a reduced diversity and an expansion inhumans. Nature Genetics 27:155–156.

Kaifu, Yousuke, Kazutaka Kasai, Grant C. Townsend, and Lindsay C. Richards.2003. Tooth wear and the “design” of the human dentition: a perspectivefrom evolutionary medicine. Yearbook of Physical Anthropology 122:47–61.

Kaplan, Hillard, Kim Hill, Jane Lancaster, and Magdalena A. Hurtado. 2000.A theory of human life history evolution: diet, intelligence, and longevity.Evolutionary Anthropology 2000:156–185.

Kaplan, Hillard, Jane Lancaster, and Arthur Robson. 2003. Embodied capitaland the evolutionary economics of the human life span. In Life span: evo-lutionary, ecological, and demographic perspectives. James R. Carey and Shri-pad Tuljapurkar, eds. Pp. 152–182. New York: Population Council.


Kelly, Robert L. 1995. The foraging spectrum. Washington, DC: SmithsonianInstitution.

Klein, Richard G., and Blake Edgar. 2002. The dawn of human culture. NewYork: Wiley.

Knott, Cheryl D. 1998. Changes in orangutan caloric intake, energy balance,and ketones in response to fluctuating fruit availability. International Journalof Primatology 19:1061–1079.

Kramer, Karen L., and Peter T. Ellison. 2010. Pooled energy budgets: re-situating human energy allocation trade-offs. Evolutionary Anthropology 19:136–147.

Krause, Johannes, Qiaomei Fu, Jeffrey M. Good, Bence Viola, Michael V.Shunkov, Anatoli P. Derevianko, and Svante Paabo. 2010. The completemitochondrial DNA genome of an unknown hominin from southern Si-beria. Nature 464:894–897.

Kuzawa, Christopher W. 1998. Adipose tissue in human infancy and child-hood: an evolutionary perspective. American Journal of Physical Anthro-pology 27(suppl.):177–209.

———. 2005. Fetal origins of developmental plasticity: are fetal cues reliablepredictors of future nutritional environments? American Journal of HumanBiology 17:5–21.


Kuzawa, Christopher W., Elizabeth A. Quinn, and Linda Adair. 2007. Leptinin a lean population of Filipino adolescents. American Journal of PhysicalAnthropology 132:642–649.

Laubichler, Manfred D., Edward H. Hagen, and Peter Hammerstein. 2005.The strategy concept and John Maynard Smith’s influence on theoreticalbiology. Biology and Philosophy 20:1041–1050.

Lee, Phyllis C. 1999. Comparative primate socioecology. Cambridge: CambridgeUniversity Press.

Li, Leah, Alan D. Dangour, and Chris Power. 2007. Early life influences onadult leg and trunk length in the 1958 British birth cohort. American Journalof Human Biology 19:836–843.

Liao, Chen-Yu, Brad A. Rikke, Thomas E. Johnson, Jonathan A. L. Gelfond,Vivian Diaz, and James F. Nelson. 2011. Fat maintenance is a predictor ofthe murine lifespan response to dietary restriction. Aging Cell 10:629–639.

MacArthur, Robert H., and Eric R. Pianka. 1966. On optimal use of a patchyenvironment. American Naturalist 100:603–609.

Maejima, Yuko, Yusaku Iwasaki, Yui Yamahara, Misato Kodaira, Udval Sed-bazar, and Toshihiko Yada. 2011. Peripheral oxytocin treatment amelioratesobesity by reducing food intake and visceral fat mass. Aging 12:1169–1177.

Martin, Robert D. 1983. Human brain evolution in an ecological context. JamesArthur Lecture on the Evolution of the Human Brain, no. 52, 1982. NewYork: American Museum of Natural History.

———. 1989. Evolution of the brain in early hominids. Ossa 14:49–62.Maynard Smith, John. 1982. Evolution and the theory of games. Cambridge:

Cambridge University Press.


Morbeck, Mary E., and Adrienne L. Zihlman. 1988. Body composition andlimb proportions. In Orang-utan biology. Jeffrey H. Schwartz, ed. Pp. 285–296. Oxford: Oxford University Press.

Navarrete, Ana, Carel P. van Schaik, and Karin Isler. 2011. Energetics and theevolution of human brain size. Nature 480:91–93.

Nielsen-Marsh, Christina M., Peggy H. Ostrom, Hasand Gandhi, Beth Shapiro,Alan Cooper, Peter V. Hauschka, and Matthew J. Collins. 2002. Sequencepreservation of osteocalcin protein and mitochondrial DNA in bison bonesolder than 55 ka. Geology 30:1099–1102.

Ong, Ken K., Kate Northstone, Jonathan C. Wells, Chandra Rubin, AndrewR. Ness, Jean Golding, and David B. Dunger. 2007. Earlier mother’s age atmenarche predicts rapid infancy growth and childhood obesity. PLoS Med-icine 4(4):e132.

Organ, Chris, Charles L. Nunn, Zarin Machand, and Richard W. Wrangham.2011. Phylogenetic rate shifts in feeding time during the evolution of Homo.Proceedings of the National Academy of Sciences of the USA 108:14555–14559.

Paterson, James D. 1996. Coming to America: acclimation in macaque bodystructures and Bergmann’s rule. International Journal of Primatology 17:585–612.

Perry, George H., and Nathaniel J. Dominy. 2009. Evolution of the humanpygmy phenotype. Trends in Ecology & Evolution 24:218–225.

Philipi, Tom, and Jon H. Seger. 1989. Hedging evolutionary bets, revisited.Trends in Ecology & Evolution 4:41–44.

Piersma, Theunis, and Jan Drent. 2003. Phenotypic flexibility and the evo-lution of organismal design. Trends in Ecology & Evolution 19:228–233.

Pilbeam, David. 2000. Hominoid systematics: the soft evidence. Proceedingsof the National Academy of Sciences of the USA 97:10684–10686.

Pond, Caroline M. 1984. Physiological and ecological importance of energystorage in the evolution of lactation: evidence for a common pattern ofanatomical organisation of adipose tissue in mammals. Symposia of theZoological Society of London 51:1–32.

———. 1998. The fats of life. Cambridge: Cambridge University Press.———. 2003. Paracrine interactions of mammalian adipose tissue. Journal of

Experimental Zoology A 295(1):99–110.Potts, Richard. 1996. Humanity’s descent: the consequences of ecological insta-

bility. New York: Morrow.———. 2012. Environmental and behavioral evidence pertaining to the evo-

lution of early Homo. Current Anthropology 53(suppl. 6):S299–S317.Premo, Luke S., and Jean-Jacques Hublin. 2009. Culture, population structure,

and low genetic diversity in Pleistocene hominins. Proceedings of the Na-tional Academy of Sciences of the USA 106:33–37.

Price, Kimberly C., Janet S. Hyde, and Christopher L. Coe. 1999. Matrilinealtransmission of birth weight in the rhesus monkey (Macaca mulatta) acrossseveral generations. Obstetrics and Gynecology 94:128–134.

Reich, David, Richard E. Green, Martin Kircher, Johannes Krause, Nick Patter-son, Eric Y. Durand, Bence Viola, et al. 2010. Genetic history of an archaichominin group from Denisova Cave in Siberia. Nature 468:1053–1060.

Reiches, Meredith W., Peter T. Ellison, Susan F. Lipson, Katherine C. Sharrock,Eliza Gardiner, and Laura G. Duncan. 2009. Pooled energy budget andhuman life history. American Journal of Human Biology 21(4):421–429.

Rickard, Ian J., and Virpi Lummaa. 2007. The predictive adaptive responseand metabolic syndrome: challenges for the hypothesis. Trends in Endocri-nology and Metabolism 18:94–99.

Roberts, Derek F. 1953. Body weight, race, and climate. American Journal ofPhysical Anthropology 11:533–558.

Robson, Shannen L., Carel P. van Schaik, and Kristen Hawkes. 2006. Thederived features of human life history. In The evolution of human life history.Kristen Hawkes and Richard R. Paine, eds. Pp. 17–44. Oxford: Currey.

Rosenberg, Karen, and Wenda Trevathan. 2002. Birth, obstetrics and humanevolution. British Journal of Obstetrics and Gynecology 109:1199–1206.

Ruff, Christopher. 2002. Variation in human body size and shape. AnnualReview of Anthropology 31:211–232.

Rutenberg, Gary W., Anthony M. Coelho, Douglas S. Lewis, K. Dee Carey,and Henry C. McGill. 1987. Body composition in baboons: evaluating amorphometric method. American Journal of Primatology 12:275–285.

Schneider, Jill E. 2004. Energy balance and reproduction. Physiology and Be-havior 81(2):289–317.

Schulke, Oliver, Jyotsna Bhagavatula, Linda Vigilant, and Julia Ostner. 2010.Social bonds enhance reproductive success in male macaques. Current Bi-ology 20(24):2207–2210.



Schwartz, Susan M., and Joseph W. Kemnitz. 1992. Age- and gender-relatedchanges in body size, adiposity, and endocrine and metabolic parametersin free-ranging rhesus macaques. American Journal of Physical Anthropology89(1):109–121.

Sear, Rebecca, and Ruth Mace. 2008. Who keeps children alive? a review of theeffects of kin on child survival. Evolution and Human Behavior 29:1–18.

Sharrock, Katherine C., Christopher W. Kuzawa, William R. Leonard, SusanTanner, Victoria E. Reyes-Garcia, Vincent Vadez, Tomas Huanca, andThomas W. McDade. 2008. Developmental changes in the relationship be-tween leptin and adiposity among Tsimane children and adolescents. Amer-ican Journal of Human Biology 20(4):392–398.

Silk, Joan B., Jacinta C. Beehner, Thore J. Bergman, Catherine Crockford,Anne L. Engh, Liza R. Moscovice, Roman M. Wittig, Robert M. Seyfarth,and Dorothy L. Cheney. 2009. The benefits of social capital: close socialbonds among female baboons enhance offspring survival. Proceedings of theRoyal Society B 276(1670):3099–3104.

———. 2010. Strong and consistent social bonds enhance the longevity offemale baboons. Current Biology 20(15):1359–1361.

Smith, David W., William Truog, James E. Rogers, Larry J. Greitzer, AlfredL. Skinner, John J. McCann, and Mary A. S. Harvey. 1976. Shifting lineargrowth during infancy: illustration of genetic factors in growth from fetallife through infancy. Journal of Pediatrics 89(2):225–230.

Smith, Eric A., Kim Hill, Frank W. Marlowe, David Nolin, Polly Wiessner,Michael Gurven, Samuel Bowles, Monique Borgerhoff Mulder, Tom Hertz,and Adrian Bell. 2010. Wealth transmission and inequality among hunter-gatherers. Current Anthropology 51:19–34.

Smith, Jennifer E., Eli M. Swanson, Daphna Reed, and Kay E. Holekamp. 2012.Evolution of cooperation among mammalian carnivores and its relevance tohominid evolution. Current Anthropology 53(suppl. 6):S436–S452.

Soares, Marta C., Redouan Bshary, Leonida Fusani, Wolfgang Goymann, Mi-chaela Hau, Katharina Hirschenhauser, and Rui F. Oliveira. 2010. Hormonalmechanisms of cooperative behaviour. Philosophical Transactions of the RoyalSociety B 365:2737–2750.

Speakman, John R. 2007. A nonadaptive scenario explaining the genetic pre-disposition to obesity: the “predation release” hypothesis. Cell Metabolism6:5–12.

Stearns, Stephen C. 1992. The evolution of life histories. New York: OxfordUniversity Press.

Sullivan, Roger J., Edward H. Hagen, and Peter Hammerstein. 2008. Revealingthe paradox of drug reward in human evolution. Proceedings of the RoyalSociety B 275(1640):1231–1241.

Sussman, Robert W., Paul A. Garber, and Jim M. Cheverud. 2005. Importanceof cooperation and affiliation in the evolution of primate sociality. AmericanJournal of Physical Anthropology 128(1):84–97.

Trivers, Robert L. 1972. Parental investment and sexual selection. In Sexualselection and the descent of man. Bernard Campbell, ed. Pp. 136–179. Chi-cago: Aldine.

———. 1974. Parent-offspring conflict. American Zoologist 14:249–264.Ungar, Peter S. 2012. Dental evidence for the reconstruction of diet in African

early Homo. Current Anthropology 53(suppl. 6):S318–S329.van Noordwijk, Arie J., and Gerdien de Jong. 1986. Acquisition and allocation

of resources: their influence on variation in life history tactics. AmericanNaturalist 128:137–142.

Vermeij, Geerat J. 2004. Nature: an economic history. Princeton, NJ: PrincetonUniversity Press.

Vrba, Elizabeth. 1985. Ecological and adaptive changes associated with earlyhominid evolution. In Ancestors: the hard evidence. Eric Delson, ed. Pp. 63–71. New York: Liss.

Walker, Robert, Michael Gurven, Kim Hill, Andrea Migliano, Napoleon Chag-non, Roberta De Souza, Gradimir Djurovic, et al. 2006. Growth rates andlife histories in twenty-two small-scale societies. American Journal of HumanBiology 18:295–311.

Walker, Robert S., and Marcus J. Hamilton. 2008. Life-history consequencesof density dependence and the evolution of human body size. CurrentAnthropology 49:115–122.

Warren, Kristin S., Ernst J. Verschoor, Susan Langenhuijzen, Heriyanto, RalphA. Swan, Linda Vigilant, and Jonathan L. Heeney. 2001. Speciation and

intrasubspecific variation of Bornean orangutans, Pongo pygmaeus pyg-maeus. Molecular Biology and Evolution 18:472–480.

Watve, Milind G., and Chitaranjan S. Yajnik. 2007. Evolutionary origins of insulinresistance: a behavioral switch hypothesis. BMC Evolutionary Biology 7:61.

Wells, Jonathan C. K. 2003. The thrifty phenotype hypothesis: thrifty offspringor thrifty mother? Journal of Theoretical Biology 221:143–161.

———. 2007a. Flaws in the theory of predictive adaptive responses. Trendsin Endocrinology and Metabolism 18:331–337.

———. 2007b. Sexual dimorphism of body composition. Best Practice inResearch: Clinical Endocrinology and Metabolism 21:415–430.

———. 2007c. The thrifty phenotype as an adaptive maternal effect. BiologicalReviews of the Cambridge Philosophical Society 82:143–172.

———. 2009. Thrift: a guide to thrifty genes, thrifty phenotypes and thriftynorms. International Journal of Obesity 33:1331–1338.

———. 2010a. The evolutionary biology of human body fat: thrift and control.Cambridge: Cambridge University Press.

———. 2010b. Maternal capital and the metabolic ghetto: an evolutionaryperspective on the transgenerational basis of health inequalities. AmericanJournal of Human Biology 22:1–17.

———. 2011. The thrifty phenotype: an adaptation in growth or metabolism?American Journal of Human Biology 23:65–75.

———. 2012a. A critical appraisal of the predictive adaptive response hy-pothesis. International Journal of Epidemiology 41:229–235.

———. 2012b. Ecogeographical associations between climate and humanbody composition: analyses based on anthropometry and skinfolds. Amer-ican Journal of Physical Anthropology 147:169–186.

———. 2012c. The evolution of human adiposity and obesity: where did itall go wrong? Disease Models and Mechanisms 5(5):595–607.

———. 2012d. Sexual dimorphism in body composition across human pop-ulations: associations with climate and proxies for short- and long-termenergy supply. American Journal of Human Biology 24:411–419.

Wells, Jonathan C. K., Tim J. Cole, and Philip Treleaven. 2008. Age-variabilityin body shape associated with excess weight: the UK National Sizing Survey.Obesity 16:435–441.

Wells, Jonathan C. K., Lewis Griffin, and Philip Treleaven. 2010. Independentchanges in female body shape with parity and age: a life-history approachto female adiposity. American Journal of Human Biology 22:456–462.

Wells, Jonathan C., and Jay T. Stock. 2007. The biology of the colonizing ape.American Journal of Physical Anthropology 134(suppl. 45):191–222.

———. 2011. Re-examining heritability: genetics, life history and plasticity.Trends in Endocrinology and Metabolism 22:421–428.

———. 2012. The biology of human migration: the ape that won’t commit?In Causes and consequences of human migration: an evolutionary perspective.Michael H. Crawford and Benjamin C. Campbell, eds. Cambridge: Cam-bridge University Press. Forthcoming.

West-Eberhard, Mary J. 2003. Developmental plasticity and evolution. NewYork: Oxford University Press.

Yajnik, Chittaranjan S., Caroline H. Fall, Kurus J. Coyaji, S. S. Hirve, ShobhaRao, David J. Barker, Charu Joglekar, and Samantha Kellingray. 2003. Neo-natal anthropometry: the thin-fat Indian baby: the Pune Maternal NutritionStudy. International Journal of Obesity 27:173–180.

Yang, Jian, Beben Benyamin, Brian P. McEvoy, Scott Gordon, Anjali K. Hen-ders, Dale R. Nyholt, Pamela A. Madden, et al. 2010. Common SNPs explaina large proportion of the heritability for human height. Nature Genetics 42:565–569.

Youngson, Neil A., and Emma Whitelaw. 2008. Transgenerational epigeneticeffects. Annual Review of Genomics and Human Genetics 9:233–257.

Zafon, Carles. 2007. Oscillations in total body fat content through life: anevolutionary perspective. Obesity Reviews 8:525–530.

Zera, Anthony J., and Lawrence G. Harshman. 2001. The physiology of lifehistory trade-offs in animals. Annual Review of Ecological Systems 32:95–126.

Zera, Anthony J., and Zhangwu Zhao. 2003. Life-history evolution and themicroevolution of intermediary metabolism: activities of lipid-metabolizingenzymes in life-history morphs of a wing-dimorphic cricket. Evolution 573:586–596.

Ziegler, Toni E., Shelley L. Prudom, Nancy J. Schultz-Darken, Aimee V. Kurian,and Charles T. Snowdon. 2006. Pregnancy weight gain: marmoset and tam-arin dads show it too. Biology Letters 2:181–183.



Origins and Evolution of Genus HomoNew Perspectives

by Susan C. Anton and J. Josh Snodgrass

Recent fossil and archaeological finds have complicated our interpretation of the origin and early evolution of genusHomo. Using an integrated data set from the fossil record and contemporary human and nonhuman primate biology,we provide a fresh perspective on three important shifts in human evolutionary history: (1) the emergence of Homo,(2) the transition between non-erectus early Homo and Homo erectus, and (3) the appearance of regional variationin H. erectus. The shift from Australopithecus to Homo was marked by body and brain size increases, a dietary shift,and an increase in total daily energy expenditure. These shifts became more pronounced in H. erectus, but thetransformation was not as radical as previously envisioned. Many aspects of the human life history package, includingreduced dimorphism, likely occured later in evolution. The extant data suggest that the origin and evolution ofHomo was characterized by a positive feedback loop that drove life history evolution. Critical to this process wereprobably cooperative breeding and changes in diet, body composition, and extrinsic mortality risk. Multisystemevaluations of the behavior, physiology, and anatomy of extant groups explicitly designed to be closely proxied inthe fossil record provide explicit hypotheses to be tested on future fossil finds.

Recent fossil and archaeological finds have complicated ourinterpretation of the origin and early evolution of genusHomo. It now appears overly simplistic to view the origin ofHomo erectus as a punctuated event characterized by a radicalshift in biology and behavior (Aiello and Anton 2012; Anton2012; Holliday 2012; Pontzer 2012; Schwartz 2012; Ungar2012). Several of the key morphological, behavioral, and lifehistory characteristics thought to first emerge with H. erectus(e.g., narrow bi-iliac breadth, relatively long legs, and a more“modern” pattern of growth) seem instead to have arisen atdifferent times and in different species. Further, accumulatingdata from Africa and beyond document regional morpho-logical variation in early H. erectus and expand the range ofvariation in this species. These new finds also make the dif-ferences between H. erectus (s.l.) and Homo habilis (s.l.) lessstark and suggest that regional variation in the former mayreflect local adaptive pressures that result from inhabitingdiverse environments in Africa and Eurasia. The mosaic na-ture of these acquisitions and the greater range of intraspecificvariation, especially in H. erectus, call into question previousinferences regarding the selective factors behind the early evo-lution of our genus and its eventual dispersal from Africa.

Susan C. Anton is Professor, Department of Anthropology, New YorkUniversity (25 Waverly Place, New York, New York 10003, U.S.A.[[email protected]]). J. Josh Snodgrass is Associate Professor,Department of Anthropology, University of Oregon (1321 KincaidStreet, Eugene, Oregon 97403, U.S.A.). The authors contributedequally to this work. This paper was submitted 4 V 12, accepted 8VII 12, and electronically published 28 XI 12.

They also raise questions about when a modern pattern oflife history might have emerged and what role, if any, it playedin our early evolution.

Modern humans have diverged in numerous ways from thelife history patterns seen in other primates, and this “humanpackage” seems linked to our ability to support larger brainsand to disperse widely. Our unique suite of life history traitsincludes altricial birth, a large energy-expensive brain, longjuvenile dependency with relatively late reproduction, shortinterbirth intervals (IBIs) with high fertility, and a long post-reproductive life span (Bogin 1999; Flinn 2010; Hill and Hur-tado 1996; Kaplan et al. 2000; Leigh 2001). With this packagewe appear to have been able to circumvent several of the keyconstraints that affect other species. Many of the life historytraits that define modern humans serve to decrease age-spe-cific reproductive value (i.e., the contribution to the growthof the population) early in life and greatly increase the costsof reproduction and somatic maintenance. What is most strik-ing about contemporary human biology is that we are ableto produce numerous high-quality offspring that experiencerelatively low mortality, grow slowly, and live long lives. Inessence, we are able to “have our cake and eat it too” byavoiding some of the life history trade-offs seen in othermammals and having a life history pattern that is both “fast”and “slow” and that emphasizes quantity and quality (Kuzawaand Bragg 2012).

This life history shift in humans was almost certainlyfacilitated by substantial behavioral and cultural shifts, in-cluding (1) cooperation in foraging (e.g., hunting/division oflabor), which maximizes the ability to obtain a stable, high-



quality diet; and (2) cooperation in reproduction (e.g., allo-parenting and midwifery), which allows the compression ofthe IBI and the consequent stacking of offspring as well asthe care for and provisioning of the secondarily altricial off-spring necessitated by our unique obstetrical dilemma (Tre-vathan 1987). Several key questions about these behavioralshifts remain unanswered, including when these traitsemerged, whether they evolved together as a package or piece-meal in different hominin species, and the particular selectivepressures that drove their evolution.

To address these distinct data sets, we bring together ideasraised at the Wenner-Gren workshop “Human Biology andthe Origins of Homo” in Sintra, Portugal, 2011. To the paperspresented in this special issue we add new data and perspec-tives, summarize the fossil and archaeological records (tables1, 2), and consider what research on contemporary primatelife history trade-offs, developmental plasticity, and regionaladaptive patterns can help us infer about behavioral and cul-tural changes in early Homo (tables 3–5). These data give usa fresh perspective on three important shifts in human evo-lutionary history: (1) the emergence of genus Homo, (2) thetransition between non-erectus early Homo and H. erectus,and (3) the appearance of regional morphological variationin H. erectus (including Homo ergaster). Using this integrateddata set, we consider the implications for understanding thechanging selective pressures that led to the transition to andevolution of early Homo.

How What We Now Know from the HardEvidence Differs from What We ThoughtWe Knew

Over the past several decades, a consensus had emerged thatthe shift to humanlike patterns of body size and shape—andat least some of the behavioral parts of the “human pack-age”—occurred with the origin of Homo erectus (e.g., Anton2003; Shipman and Walker 1989). This was seen by manyresearchers as a radical transformation reflecting a sharp andfundamental shift in niche occupation, and it emphasized adistinct division between H. erectus on the one hand and non-erectus early Homo and Australopithecus on the other.1 EarliestHomo and Australopithecus were reconstructed as essentiallybipedal apes, whereas H. erectus had many of the anatomicaland life history hallmarks seen in modern humans. To some,the gap between these groups suggested that earlier speciessuch as Homo habilis should be excluded from Homo (Collardand Wood 2007; Wood and Collard 1999).

Recent fossil discoveries paint a picture that is substantiallymore complicated. These discoveries include new fossils of

1. While it is recognized that Australopithecus may be paraphyletic,for the purposes of the comparisons in this paper, the genus is consideredto exclude Paranthropus species but to include the best-represented spe-cies commonly assigned to Australopithecus, i.e., A. anamensis, A. afa-rensis, A. garhi, A. africanus, and A. sediba. When the data for specificcomparisons come from a single species, that species is indicated by name.

H. erectus that reveal great variation in the species, includingsmall-bodied members from both Africa and Georgia (Ga-bunia et al. 2000; Potts et al. 2004; Simpson et al. 2008; Spooret al. 2007), and suggest a previous overreliance on the Na-riokotome skeleton (KNM-WT-15000) in reconstructions ofH. erectus. Additionally, reassessments of the Nariokotomematerial have concluded that he would have been considerablyshorter than previous estimates (∼163 cm [5 feet 4 inches],not 185 cm [6 feet 1 inch]; Graves et al. 2010), younger atdeath (∼8 years old, not 11–13 years old; Dean and Smith2009), and with a life history pattern distinct from modernhumans (Dean and Smith 2009; Dean et al. 2001; Thompsonand Nelson 2011), although we note that there is substantialvariation in the modern human pattern of development (Se-selj 2011). Further, the recent discovery of a nearly completeadult female H. erectus pelvis from Gona, Ethiopia, which isbroad and has a relatively large birth canal, raises questionsabout the narrow-hipped, Nariokotome-based pelvic recon-struction and whether H. erectus infants were secondarily al-tricial (Graves et al. 2010; Simpson et al. 2008).2

In addition to recent changes in our understanding of H.erectus, new discoveries and reanalyses have complicated thepicture of earliest Homo by documenting its diversity andemphasizing underappreciated differences and similaritieswith H. erectus (Blumenschine et al. 2003; Spoor et al. 2007).Finally, a new view of Australopithecus has begun to emergein which it shares many postcranial characteristics with Homo,including a somewhat large body and relatively long legs(Haile-Selassie et al. 2010; Holliday 2012; Leakey et al. 2012;Pontzer 2012). These results suggest a previous overrelianceon the very small “Lucy” (A.L.288-1) skeleton to characterizethat species/genus.

Brains, Bodies, and Sexual Dimorphism

Although recent discoveries reveal a larger Australopithecus afa-rensis and a smaller, more variable H. erectus than previouslyknown, there still appear to be important differences betweenthe species. Even when including the largest of the new Aus-tralopithecus fossils and the smallest of the new early Homofossils, estimates suggest an average increase in body mass of33% from A. afarensis to early Homo (in this case H. habilis �Homo rudolfensis � early H. erectus; Holliday 2012; Pontzer2012). The difference is more modest—on the order of 10%—when comparing A. afarensis to only non-erectus early Homo(table 1). The fossil record also suggests a body mass increaseof ∼25% between early non-erectus Homo in East Africa andearly H. erectus (Africa � Georgia). This expanding fossil recorddocuments marked regional variation, with early African H.erectus being ∼17%–24% larger on average than Georgian H.

2. We note that there is some disagreement regarding the specific statusof the Gona pelvis, including suggestions that it may not be Homo (Ruff2010). Nonetheless, other reconstructions of the KNM-WT 15000 pelviswere narrower than the original, suggesting that its breadth may not bea strong anchor point for neonate head size.

Anton and Snodgrass Origins and Evolution of Genus Homo S481

erectus of approximately the same geological age (table 1; Anton2012).

Recent fossil evidence and reinterpretation of known speci-mens also documents a more mosaic pattern of evolving limbproportions, which has implications for locomotor reconstruc-tions. New work shows that despite absolute size differences andcontrary to conventional wisdom, relative hind-limb length doesnot differ from Australopithecus to Homo or among Homo (Hol-liday 2012; Holliday and Franciscus 2009; Pontzer 2012). Theforelimb, however, is relatively stronger and slightly longer inboth Australopithecus and non-erectus early Homo than it is inH. erectus (Ruff 2009). Further, the Georgian forelimb is slightlyshorter than in early African H. erectus, which may reflect atemporal, climatic, or even secular shift (Holliday 2012; Pontzer2012).

Cranial capacities show an increase of 130% from A. afa-rensis (mean p 478) to non-erectus early Homo (i.e., 1813 �1470 groups; mean p 629 cm3). This marks the first time thathominin cranial capacity expands beyond the range of variationseen among great apes (Schoenemann 2006). Also, althoughthe ranges overlap, average cranial capacity increases by ∼25%from early non-erectus Homo to early H. erectus in Africa andGeorgia (combined mean p 810 cm3) or by 130% when com-pared with just early African H. erectus (mean p 863 cm3).Among regional samples, both early African H. erectus and earlyIndonesian H. erectus are ∼25% larger on average than GeorgianH. erectus of about the same geological age, a similar differenceas for body size (tables 1, 2).

Despite the problems of assigning sex to individual fossils,preliminary patterns of sexual dimorphism can be consideredfor different species using brain and body size estimates (An-ton 2012; Plavcan 2012). The ratio of male to female meanvalues for brain and body size suggests that H. erectus ismodestly less dimorphic than is A. afarensis. However, sex ishard to estimate for fossils, and the degree of dimorphisminferred depends on the particular variable considered, themeans of comparison, and the specimens included in thesample (table 1; Plavcan 2012). For example, A. afarensis andearly H. erectus show no difference in size variation (CVs) forbody mass or endocranial capacity (table 1; and see table 3in Anton 2012). By other measures, H. habilis (exclusive of1470) is more dimorphic in body mass estimates than Aus-tralopithecus but less dimorphic in brain size (table 1; Plavcan2012). And H. erectus is more dimorphic than H. habilis inbrain size but less dimorphic in body size. Unfortunately, H.habilis values are particularly suspect given the small samplesand uncertainty regarding numbers of included species. Thesedata are equivocal as to the degree of dimorphism presentbut do not provide strong support for decreasing dimorphismin H. erectus (see Plavcan 2012).

Teeth, Development, and Diet

Examination of dental evidence such as tooth size, microwear,and developmental pattern can provide a window onto key

transitions in early Homo. As has been well documented, pos-terior teeth decrease in average size and increase in occlusalrelief from Australopithecus to Homo (Ungar 2012). The trendis somewhat more pronounced in H. erectus, which showssubstantial third molar reduction (Gabunia et al. 2000; In-driati and Anton 2008; Spoor et al. 2007). There is, however,substantial size overlap in jaw and tooth size among all earlyHomo (Anton 2008). In contrast, preliminary evidence sug-gests that incisor row length may be larger in non-erectusearly Homo than in Australopithecus and intermediate in sizein H. erectus (Ungar 2012). This may suggest dietary differ-ences relating to incisal preparation.

Dental topography and microwear for all early Homo aremore complex than in Australopithecus. Although early Homolikely ate a fairly generalized diet, this signal suggests theyalso consumed less brittle foods (Ungar and Scott 2009; Ungaret al. 2012). Homo erectus shows more variation and moresmall features than non-erectus early Homo, indicating greaterdietary breadth in the former (Ungar and Sponheimer 2011).The signal is similar across regional samples of H. erectus.Thus, dental morphology suggests consumption of a gener-alized diet in early Homo but with a modestly increased dietarybreadth compared with Australopithecus.

Although sample sizes are extremely small, there is someevidence that the emergence of the first permanent molar(M1), a variable that correlates with many life history traits,occurs about a year later in H. erectus than in A. afarensis(Dean et al. 2001; Schwartz 2012). This finding is consistentwith a recent analysis that documents relatively minor growthand life history differences in H. erectus compared with earlierhominins and living African apes (Thompson and Nelson2011). However, the pattern of skeletal and dental develop-ment in Nariokotome is not much outside the range of “tall”modern human children (Seselj 2011), hinting perhaps at themodularity (i.e., independence) of developmental systems.Unfortunately, there are no data for M1 emergence for non-erectus early Homo, and we caution that M1 development canbe decoupled from somatic growth rates (Dirks and Bowman2007; Godfrey et al. 2003). Thus, life history reconstructionssuggest a pattern of growth modestly different from Australo-pithecus yet distinct from later Homo species such as Nean-derthals and modern humans.

Climate and Environment

Although populations of early Homo likely lived in a varietyof specific environments, Potts (2012) reviews how multipleindependent paleoclimatic records show an increase in theamplitude of the climate shifts and an increasing unpredict-ability in their timing during the origin and early evolutionof Homo. He suggests that this inherent variation in climateplaced a premium on developmental plasticity—the capacityfor developing individuals to respond phenotypically to en-vironmental conditions (Lasker 1969; Wells 2012)—and likelybehavioral plasticity as well. The result of developmental plas-

S482

Table 1. Differences that may relate to life history inferences compared between Australopithecus and Homo and within early Homo species based on hard evidence

Australopithecus afarensis vs. early Homoa A. afarensis vs. Homo erectus/Homo ergaster b

African H. erectus vs. Homo habilis/Homo rudolfensis c

Average brain size (cm3; Anton 2012) Homo larger: 629 vs. 478 H. erectus/H. ergaster larger: 810/863 vs. 478 H. erectus larger: 863 vs. 629Average body mass (kg; Holliday 2012; Pontzer 2012) Homo larger: 44 vs. 40 H. erectus/H. ergaster larger: 52/55 vs. 40 H. erectus larger: 52/55 vs. 44BMR (kcal/day)d Homo larger: 1,191.7 vs. 1,134 H. erectus/H. ergaster larger: 1,308–1,351 vs. 1,134 H. erectus larger: 1,308–1,351 vs. 1,192TDEE (kcal/day)d Homo larger: 2,026–2,264 vs. 1,927–2,153 H. erectus/H. ergaster larger: 2,224–2,568 vs.

1,927–2,153H. erectus larger: 2,224–2,568 vs. 2,026–2,264

Humerofemoral strength proportions (Ruff 2008,2009)

Both similar to Pan Relatively less strong humerus in H. erectus/H.ergaster

Relatively less strong humerus in H. erectus

Humerofemoral length proportions (Holliday 2012;Pontzer 2012)

Same Same Same

Hind-limb length relative to body mass (Holliday2012; Pontzer 2012)

Same Same Same

Sexual dimorphism:Brains:

Male/female average brain size (sex designationsas per Anton 2012)

Homo less dimorphic: 1.05 (�625;�590) vs. 1.3 (�507; �400)

H. erectus/H. ergaster less dimorphic: 1.15/1.2(�840/924; �730/770) vs. 1.3 (�507; �400)

Homo erectus more dimorphic: 1.2 (�924;�770) vs. 1.05 (�625; �590)

CVs (Anton 2012) Homo less dimorphic: 12.2 vs. 15.7 No difference: 17.8/15.9 vs. 15.9 H. erectus more dimorphic: 17.8/15.9 vs.12.2

Bodies:Male/female mean:

Using associated skeletons (Anton 2012) Homo more dimorphic: 1.39 (�46; �33)vs. 1.32 (�39; �29.5)

H. erectus/H. ergaster less dimorphic: 1.06/1.0(�50/51; �47/51) vs. 1.32 (�39; �29.5)

H. erectus less dimorphic: 1.06/1.0 (�50/51;�47/51) vs. 1.39 (�46; �33)

Using sex estimates of Pontzer (2012) Homo more dimorphic: 1.77 (�56.7;�31.9) vs. 1.32 (�39; �29.5)

H. erectus/H. ergaster less dimorphic: 1.20/1.25(�55.8/60.4; �46.2/48.2) vs. 1.32 (�39;�29.5)

H. erectus less dimorphic: 1.20/1.25 (�55.8/60.4; �46.2/48.2) vs. 1.77 (�56.7; �31.9)

CVs:Of body mass data (Anton 2012) Homo more dimorphic: 33 vs. 20.2 No difference: 19.3/18.5 vs. 20.2 H. erectus less dimorphic: 19.3/18.5 vs. 33Of femur length data (Anton 2012) Homo more dimorphic: 13 vs. 16.3 H. erectus/H. ergaster less dimorphic: 8.7/5.8 vs.

16.3H. erectus less dimorphic: 8.7/5.8 vs. p 13

S483

Tooth size and shape (Ungar 2012) Larger I’s than A. afarensis; less buno-dont M’s with more occlusal reliefthan A. afarensis; overlapping rangesbut smaller average M’s than A. afar-ensis

Larger I’s than A. afarensis; less bunodont M’swith more occlusal relief and thinner enamelthan A. afarensis; smaller M’s than A. afarensis

H. erectus I’s intermediate between H. habi-lis and A. afarensis; thinner? enamel thanH. habilis; smaller M’s than H. habiliswith third molar reduction

Dental microwear (Ungar 2012; Ungar et al. 2012) Unremarkable M surface complexity Unremarkable M surface complexity with sub-stantial variation and more small features

Unremarkable M surface complexity withsubstantial variation and more small fea-tures

Age at M1 eruption (Schwartz 2012) H. habilis unknown; A. afarensis p 2.9–3.6 years

Later in H. erectus: 4.4–4.5 vs. 2.9–3.6 years H. habilis unknown; H. erectus p 4.4–4.5years

Site distribution (area/home range; Swisher et al.1994)

Similar? H. erectus across Old World by 1.6 Ma H. erectus across Old World by 1.6 Ma

Stone transit distances? (Braun et al. 2008; Potts2012)

H. habilis 10s to 100s of m from 2.5 to2.3 Ma and possibly farther after 1.95Ma; A. afarensis probably does notmove stone

H. erectus and possibly H. habilis after 1.95 Matransport rock 12–13 km; A. afarensis does not

Similar? H. erectus perhaps transports rockfarther?

Cut-marked/percussion-marked bone (Potts 2012) Ubiquitous after 2.5 Ma; one possibleoccurrence with A. afarensis

Ubiquitous after 2.5 Ma Similar? Ubiquitous after 2.5 Ma

Tool technologies (Lepre et al. 2011; Semaw et al.2003)

None before 2.6 Ma when Oldowan ap-pears

Oldowan � Acheulean after 1.76 Ma Similar tools or divided by taxon?

a Australopithecus afarensis is used for comparative purposes because its cranial capacities (Holloway and Yuan 2004) and body mass (Pontzer 2012) values are greater than those for A. africanusand thus provide a conservative comparison for differences in size. Cranial capacity of KNM-ER 1470 is excluded.b Given the small size of the Georgian remains, where available, H. erectus values are presented as the combined means for Georgian and early African H. erectus as a conservative comparison withA. afarensis followed by the early African H. erectus–only values. Age at M1 eruption is available only for the early African remains.c Homo rudolfensis is not included in brain size estimates, but postcrania assigned to the Homo sp. that may be H. rudolfensis are included in body size estimates.d Basal metabolic rate (BMR) is calculated by using the Oxford equations for prime adults (18–30 years) and the average body weight of each species. The average of male (16 # weight � 545)and female (13.1 # weight � 558) equations is reported. Total daily energy expenditure (TDEE) range is calculated as TDEE p BMR # physical activity level (PAL). A range of PALs from apelike(1.7; Pontzer and Kamilar 2009) to humanlike (1.9, being the mean of male, 1.98, and female, 1.82, averages for subsistence populations; Snodgrass 2012:368) are used. Lower mean values for Panhave been reported (1.5; Schroepfer, Hare, and Pontzer 2012), but given the high range of variation, we opt for the more conservative values previously published.


Table 2. Regional differences between early Homo erectus samples related to important variables of life history

African Homo erectus/Homo ergaster(1.8–1.5 Ma)a

Georgian H. erectus/H. ergaster (1.8–1.7 Ma)b Asian H. erectus (11.5 Ma)c

Average brain size (cm3) X p 863 (n p 5) X p 686 (n p 3) 908 (n p 1)Average body size (kg) X p 57 (n p 4); X p 54 (n p 5) X p 46 (n p 3) ?BMR (kcal/day)d 1,352 1,221 ?TDEE (kcal/day)d 2,298–2,568 2,075–2,319 ?Sexual dimorphism:

Brains (male/female mean values) ?1.2 (� p 924, � p 770) ?1.07 (� p 700, � p 655) ? (� p 908)Bodies (male/female mean values) 1.0/1.25 (� p 51/60.4, � p 51/48.2) ?1.21 (� p 48.8 [1], � p

40.2 [1])?

Age at M1 eruption (years) 4.4 (KNM-WT 15000) ? 4.5 (n p 1)Forelimb to hind-limb length pro-

portionsSimilar to earlier hominins or a

little shorterGeorgian has slightly

shorter forelimb?

Forelimb and strength proportions Less strong relative to hind limbthan in H. habilis

? ?

Tooth size/shape(Anton 2008; Indriati and An-ton 2008; Ungar 2012)

H. erectus/H. ergaster I’s intermediatebetween Homo habilis and Austra-lopithecus; smaller average M’s thanH. habilis or Australopithecus andwith M3 reduction

?; largest of the H. erectus/H. ergaster teeth, smallerthan H. habilis and withM3 reduction

?; larger than African, slightlysmaller than Georgian, withM3 reduction

Tooth microwear(Ungar 2012; Ungar et al.2012)

Unremarkable M surface complexitywith substantial variation and moresmall features

Unremarkable M surfacecomplexity with substan-tial variation and moresmall features

?

Transit distances? 12–13 km ? ?

a Cranial capacities for KNM-ER 3733, 3883, 42700, KNM-WT 15000, OH 9; body mass values for KNM-ER 736, 737 1808, KNM-WT 15000( ) and BSN49/P27 ( ). Sexes are unknown; however, KNM-ER 1808, 3733, 42700, and BSN 49/P27 are presumed females for this table;n p 4 n p 5KNM-ER 736 and 737 are not assigned to sex. Two sex dimorphism estimates are provided for body size: the first calculates body mass for maleskeleton KNM-WT15000 and female skeletons KNM-ER 1808 and BSN 49/P27; the second follows Pontzer’s sex designations for postcranial elements,includes more specimens, and moves KNM-ER 1808 to male. South African H. erectus do not preserve endocranial capacity. Body mass data forSouth African H. erectus are not included, but the few that are available are comparable to East African H. erectus and would not change the resultshere (see Anton 2012).b Cranial capacities for D2280, 2282, 3444; body mass values for large and small adult and D2021. Sexes are unknown; however, D2282 and thesmall adult are presumed females for this table; D2021 is unsexed.c Statements reflect Asian H. erectus older than 1.5 Ma only. Cranial capacity for Sangiran 4; dental dimensions for Sangiran 4 and S27; M1 emergencefrom Dean et al. (2001). While some postcranial size estimates have been made for mid-Pleistocene Asian H. erectus (Anton 2003), no postcranialfossils are available from the early Pleistocene.d BMR (basal metabolic rate) and TDEE (total daily energy expenditure) calculated as in table 1, using an average African H. erectus weight of 55kg as per table 1.

ticity is seen in recent secular trends in size in humans (e.g.,Boas 1912; Bogin 1999; Kaplan 1954; Shapiro 1939; Stinson2012) and is a critical means by which humans balance thehigh costs of growing large-brained offspring while adjustingto environmental change at the generational or multigener-ational timescale (Kuzawa and Bragg 2012; Walker et al. 2006;Wells 2012). If developmental pattern, particularly plasticity,is the target of selection (Kuzawa and Bragg 2012), a meansof assessing how to visualize this pattern in the skeletal recordof extant taxa is needed to lay a foundation for doing so inthe fossil record. A similar means is needed for identifyingbehavioral plasticity from the archaeological record.

Material Culture

The archaeological record provides evidence of several keybehaviors—including changes in dietary niche, ranging, andcognition—that are often associated with the rise of genusHomo. The manufacture and use of stone tools has long been

thought to signal a foraging shift and to be associated withthe origin of Homo (Leakey, Tobias, and Napier 1964). Thefirst unambiguous tools appear at 2.6 Ma, with cut-markedanimal bone ubiquitous in sites after this time (Potts 2012);however, one occurrence of cut-marked bone has been arguedto occur before the emergence of Homo (McPherron et al.2010, 2011; but see Domınguez-Rodrigo, Pickering, and Bunn2010, 2011). Although the Oldowan is linked to carcass pro-cessing, other uses related to plant food processing are im-portant (Roche, Blumenschine, and Shea 2009). This emerg-ing picture is consistent with dental evidence and supports amodest dietary shift to more carnivory in Homo and increaseddietary breadth compared with Australopithecus.

A second noteworthy change occurs at approximately 1.95Ma with an increase in stone transport distances that suggeststhe movement of rock over ∼12 km intervals (Braun et al.2008; Potts 2012). Further, by 1.76 Ma, Acheulean tools ap-pear in the record (Lepre et al. 2011). These changes are often


attributed to H. erectus and are used to suggest increasedrange, although it is worth noting that this temporal asso-ciation may be coincidental and that increased transit dis-tances may be characteristic of all post-2.0-Ma Homo. Cer-tainly after 1.6 Ma, H. erectus, but not other Homo, isdistributed across the Old World, suggesting even greaterranging.

What Changes in Fossil Homo May Mean for Energetics

Changes in brain and body size and ranging have importantimplications for daily energy expenditures that must in turnbe balanced by shifts in energy input (i.e., dietary quantityor quality) and/or shifts in allocation to somatic functions.Total daily energy expenditure (TDEE), or an individual’s totalmetabolic cost per day, encompasses the energy required forbasic bodily survival and maintenance (thermoregulation, im-mune function, physical activity, etc.) and that required forgrowth and reproduction. If basal metabolic rate (BMR) andphysical activity level (PAL) are known, TDEE can be esti-mated (TDEE p PAL # BMR). BMR has a strong correlationto body weight, and average PALs have been measured forsubsistence populations of humans and some great apes(Pontzer et al. 2010; Schroepfer, Hare, and Pontzer 2012;Snodgrass 2012). Thus, we can calculate a range of TDEEsfor each fossil hominin species by using alternately an ape(1.7) or human (1.9) subsistence average for PAL and humanequations for BMR (tables 1, 2).

When data from contemporary humans and other primatesare used to estimate key energy parameters for fossil species,TDEE increases in all early Homo over the condition in Aus-tralopithecus because of body size increases. If we assume thatdifferent species and genera shared similar PALs (i.e., are ei-ther all apelike or all humanlike), then H. habilis TDEE in-creases only modestly (5%) over the condition in A. afarensis.Homo erectus increases by 15% over A. afarensis. African H.erectus TDEE estimates are 10% greater than those for Geor-gian H. erectus. Alternatively, if suggestions of increased rang-ing in H. erectus (or early Homo) are considered to indicatethat Homo species can be attributed more humanlike PALscompared with A. afarensis, then the differences between thegenera would be greater. In either case, Homo appears to haverequired more energy input than Australopithecus or perhapsa shift to a higher throughput system (i.e., more caloriesconsumed and expended per day) than Australopithecus suchas is seen in humans versus great apes (see Pontzer 2012).

Summary of Fossil Changes in Early Homo

The suite of morphological and behavioral traits that char-acterize modern humans does not first appear with the originof H. erectus, at least not to the extent previously believed.Some critical changes such as hind-limb elongation occur atthe base of the hominin lineage (i.e., well before the originof genus Homo). Other traits, including modest brain and

body size increases and dietary differences, occur with theorigin of Homo. Still other changes, such as pelvic narrowingand marked encephalization, occur considerably later in timethan previously believed, with several of these traits not ap-pearing until the origin of modern humans.

While the nature of the fossil record makes any interpre-tation preliminary, current evidence is consistent with theview that there was not a radical shift in the biology andbehavior of H. erectus but instead that the full suite of mor-phological and life history traits that characterize our ownspecies first emerged in modern humans. The shift from Aus-tralopithecus to Homo was marked by body and brain sizeincrease, dental and other indicators of a dietary shift, andchanges in ranging behavior that imply increased TDEE.These shifts became more pronounced in H. erectus, but sub-stantial intraspecific variation exists. It also appears that thedevelopmental shift to the modern human condition occurredpiecemeal. Homo erectus development (based on the timingof M1 eruption) was later relative to Australopithecus but wasquicker than that seen in later Homo. This delay may havebeen present in non-erectus early Homo as well. An importantpoint that has emerged especially from Schwartz’s (2012)work is that there were diverse life history patterns amongfossil hominins, and an approach to human life history evo-lution that considers only “ape” versus “human” or “slow”versus “fast” is overly simplistic (see also Leigh and Blomquist2007, 2011; Robson and Wood 2008).

The increasing variability of climate over time suggests thatboth developmental and behavioral flexibility may have beenprized and that the apparent variation seen in the past needsto be carefully compared and parsed against extant variation.These data imply that the extant record should be plumbedin new ways for evidence of how the skeletons of living hu-mans and nonhuman primates reflect their environments, lifehistories, and behaviors. These analyses require the devel-opment of data sets in which the extant and fossil recordscan be more fully integrated.

Human Biology and the Origins of Homo:Implications for Understanding theFossil Record

Here we integrate recent advances in the study of contem-porary human and primate biology with the fossil record tobetter interpret the evidence discussed above (tables 3, 4). Weconcentrate on inferences regarding (1) the emergence of ge-nus Homo, (2) the transition between non-erectus early Homoand Homo erectus, and (3) the appearance of regional mor-phological variation in H. erectus. We outline predictions thatwe hope will help guide future research and suggest areas inwhich additional data from extant taxa would be particularlyuseful.


Table 3. Inferences regarding behavioral/cultural differences between Australopithecus and Homo

Australopithecus vs. early HomoAustralopithecus vs. Homo erectus/

Homo ergasterH. erectus vs Homo habilis/

Homo rudolfensis

Energetic requirements:Brains Homo larger on average H. erectus/H. ergaster larger H. erectus larger on averageBodies Homo larger on average H. erectus/H. ergaster larger H. erectus larger on average

Developmental rate:Brains ? ? ?Teeth (Schwartz 2012) ? H. erectus/H. ergaster slower than

Australopithecus but still fastcompared with Homo sapiens?

?

Bodies (Dean et al. 2001;Graves et al. 2010)

? H. erectus/H. ergaster body relativelyfaster than teeth intermediate be-tween Pan and H. sapiens

?

Diet (from teeth; Ungar2012; Ungar et al. 2012)

Tougher, less brittle food itemsin Homo; more incisal preparationin Homo?

Tougher, less brittle food items inH. erectus/H. ergaster; greater dietbreadth in H. erectus/H. ergasterthan Australopithecus

Tougher, less brittle food itemsin H. erectus; greater dietbreadth in H. erectus thanH. habilis/H. rudolfensis

Nutritional environment/diet:From brains/bodies Homo somewhat higher-quality diet H. erectus/H. ergaster higher-quality

dietH. erectus probably higher-

quality dietFrom archaeology Homo greater use of animal prod-

ucts?H. erectus/H. ergaster more signifi-

cant use of animal productsH. erectus likely greater use of

animal than H. habilis/H.rudolfensis

Locomotor repertoire Both have significant arboreal com-ponent

H. erectus/H. ergaster strongly ter-restrial

H. erectus more terrestrial

Home range (HR):Bodies Somewhat larger because of larger

body size?H. erectus/H. ergaster larger because

of body sizeH. erectus larger because of

body sizeSite distribution Similar? H. erectus/H. ergaster larger HR H. erectus larger HRStone transport ? H. erectus/H. ergaster larger HR H. erectus larger HR

Note. Based on hard-evidence differences in table 1.

The Emergence of Early Homo

The fossil record for earliest Homo is especially sparse, andinferences from it must be made cautiously. Nonetheless,available fossil evidence suggests that non-erectus early Homospecies were somewhat larger in average brain and body sizeand had slower developmental patterns than Australopithecus(Anton 2012; Holliday 2012; Pontzer 2012; Schwartz 2012).If confirmed, the extant record indicates that this brain andbody size increase was most likely to result from an increasein food availability and dietary quality and a reduction inextrinsic mortality risk.

Considerable evidence exists that improved diet quality andnutrient availability during growth influences adult body size(Kuzawa and Bragg 2012). In contemporary human popu-lations, secular trends to larger body size and earlier repro-ductive maturation occur quickly via developmental shiftsthat alter energy allocation during improved environmentalconditions such as higher-quality and more stable food re-sources and reduced infectious disease exposure (Boas 1912;Bogin 1999; Kaplan 1954; Shapiro 1939; Stinson 2012). Forexample, in a single generation, Mayan children growing upin the United States experienced a 10-cm population-levelincrease in stature compared with those in Guatemala (Boginand Rios 2003). Conversely, under stable yet extremely poorenvironmental conditions, there is evidence for a reduced

plasticity that leads to early maturation and small adult bodysize (see Migliano and Guillon 2012). While we focus mainlyon body size, we note that cranial characteristics and brainsize are subject to similar developmental plasticity (e.g., Boas1912), and we note that nondietary variables also contributeto growth and adult outcomes.

Recent work has provided extensive evidence that extrinsicmortality risk is a primary contributor to life history variationboth within and between species, with faster growth and ear-lier reproduction in environments of high (especially juvenile)mortality (Charnov 1993; Kuzawa and Bragg 2012; Stearns1992; Walker et al. 2006). For example, arboreal nonhumanprimates tend to have relatively protracted life histories thatappear to result from the relatively low predation risk andmortality they experience (Borries et al. 2011). And in hu-mans, extremely high mortality environments with pro-nounced juvenile and adult risk may help explain the fastdevelopmental life history pattern and small adult body sizeof “pygmy” populations such as the Aeta and Batak of thePhilippines (Migliano 2005; Migliano and Guillon 2012; Mig-liano, Vinicius, and Lahr 2007).

Thus, proximate environment-related shifts in life historycan influence morphology and are potentially identifiable inthe fossil record. Further, these developmental shifts may pro-vide a foundation for longer-term population-level adaptation


Table 4. Regional behavioral/cultural differences inferred between early Homo erectus samples

African Homo erectus/Homoergaster (1.8–1.5 Ma)

Georgian H. erectus/H. ergaster(1.8–1.7 Ma) Asian H. erectus (11.5 Ma)

Inferred energetic requirements:Brains (TDEE) Increased contribution of brain size

to metabolismIncreased contribution of brain size

to metabolism over conditionin H. habilis

Increased contribution of brainsize to metabolism

TDEE Higher in Africa because of bodysize differences

Lower in Georgia, with seasonalupregulation of metabolicexpenditures?

?

Inferred developmental rate Same as Asia ? Same as AfricaInferred diet (teeth) Tougher, less brittle food items in

H. erectus/H. ergasterTougher, less brittle food items in

H. erectus/H. ergaster?

Greater diet breadth than Homohabilis/Homo rudolfensis

Greater diet breadth than Homohabilis/Homo rudolfensis

Inferred nutritional environment:Anatomy High quality Nutritionally less sufficient during

growth given small sizeHigh quality given brain size

Archaeology High quality Perhaps more seasonal? ?Transit distances? 12–13 km ? ?Extrinsic mortality? Lower than Australopithecus based

on body and brain sizeLower than Australopithecus but

possibly higher than other H.erectus

Lower than Australopithecusbased on brain size

Note. Inferred from primary data in table 2. TDEE p total daily energy expenditure.

through natural selection (Kuzawa and Bragg 2012). Thecomplicated web of interactions means that shifts in bodysize, for example, may result from a variety of different inputsworking together or at cross-purposes (Kuzawa and Bragg2012; Migliano and Guillon 2012). It is nonetheless possibleto begin to make some predictions regarding the expectedoutcomes that various kinds of changes to extrinsic mortalityand other proximate factors might have on skeletal size andshape. In particular, increases in overall body size might resultfrom several different decreases in, for example, extrinsic mor-tality. These could include reduced susceptibility to predationor decreased infectious disease or parasite burden, for ex-ample. Future research should develop means of assessingfrom separate records (i.e., archaeological, paleontological,geological, and contemporary biological) the presence andrate of these various sources of mortality.

If decreases in extrinsic mortality and increases in energyavailability and dietary quality are driving factors in the originof Homo, then we can predict that if further evidence ofmultiple early non-erectus Homo taxa is found, each will belarger in average body size than Australopithecus. However,these multiple Homo species, while showing anatomical evi-dence of niche partitioning, may or may not differ from oneanother in body size.

The fossil record also suggests that non-erectus early Homowas smaller and developed more quickly than H. erectus, al-though again the early Homo record is quite sparse. Whenadditional fossils of early Homo are available, we predict thatnon-erectus early Homo will be found to have had a life historypattern intermediate between Australopithecus and H. erectuswith a modestly extended growth period, including the pres-ence of short childhood and adolescent periods. We base our

prediction on archaeological and paleontological evidence fordietary change in earliest Homo as well as modest body sizeincrease over the condition in Australopithecus. Additionalstudies of dental macro- and microstructure will help lay afoundation in extant taxa for understanding the relationshipof tooth form (especially molars) to diet. While we acknowl-edge that proxies for life history patterns are more compli-cated to reconstruct than proxies for other types of shifts,studies that consider how dental developmental profiles andtheir variation are correlated with life history attributes withinpopulations of living human and nonhuman primates will bean important means of contextualizing fossil data.

Another hypothesized contributor to the emergence of earlyHomo is related to the influence of increasing climatic vari-ability on biology. The geological record indicates increasedclimatic variability during the rise of early Homo, which, basedon extant human and primate biology, hints at the possibilitythat greater developmental plasticity than in Australopithecusmay have facilitated adjustments to short-term environmentalchange and initiated a cascade of events leading to greatercapacity for phenotypic plasticity as well as increased dispersalcapability. Given the paucity of the Homo record from 2.5 to1.5 Ma, part of the primary research agenda should be anemphasis on exploring sediments from this time period witha particular focus on differentiating between early non-erectusand early H. erectus lifeways. However, additional work onthe substantial fossil record of Australopithecus afarensis,which shows distinct temporal changes in morphology (Lock-wood, Kimbel, and Johanson 2000), might be useful in pro-viding a comparative hominin data set for testing the idea ofincreased developmental plasticity in Homo. Studies that focuson comparing variation potentially related to developmental


Table 5. Tertiary inferences regarding life history and behavior between Australopithecus and Homo

Australopithecus vs.early Homo

Australopithecus vs. Homo erectus/Homo ergaster

H. erectus vs. Homo habilis/Homo rudolfensis

Extrinsic mortality Possibly lower in Homo givenbody size

Lower in H. erectus Lower in H. erectus

Developmental plasticity ? Greater in H. erectus Greater in H. erectusBody composition Larger brains in Homo but

similar adiposityLarger brains in H. erectus/H. ergaster

and greater adiposityLarger brains in H. erectus/H. ergaster

and greater adiposityCooperative breeding

(alloparenting; Islerand van Schaik 2012)

Possibly more cooperativebreeding in Homo

H. erectus/H. ergaster more coopera-tive breeding necessitated by largeraverage brain size

H. erectus/H. ergaster more coopera-tive breeding necessitated by largeraverage brain size

Cooperative hunting Possibly greater in Homo Likely greater cooperative huntingbased on diet shift in H. erectus/H.ergaster

Likely greater cooperative huntingbased on diet shift in H. erectus/H.ergaster

Note. Based on hard-evidence differences in table 1.

plasticity (e.g., variation in body size or dimorphism at a giventime) among time packets of this taxon and with early Homowould facilitate the identification of genus-level differences inbiology. If increased developmental plasticity is present inHomo, one should find greater variation in the genus at anygiven time than in other well-represented genera. While non-erectus Homo samples are currently insufficient for such com-parisons, H. erectus provides more opportunities for suchinvestigations.

The Transition between Non-erectus Early Homo andHomo erectus

Even though recent fossil and archaeological discoveries chal-lenge the idea that the origin of H. erectus involved a punc-tuated transformation of biology and behavior, present evi-dence suggests that this species did diverge from otherhominins in several important ways. The life history patternof H. erectus appears to have been more protracted than thatof Australopithecus and Paranthropus and possibly non-erectusearly Homo. Despite this, when compared with modern hu-mans and later Homo (e.g., Neanderthals), H. erectus appearsto have had a more rapid life history, with less pronouncedsecondary altriciality, an earlier maturation, and a less pro-nounced adolescent growth spurt (Dean and Smith 2009;Graves et al. 2010; Guatelli-Steinberg 2009; Thompson andNelson 2011).

Present evidence suggests that a childhood phase of de-velopment (i.e., “early childhood”), with offspring beingweaned yet still dependent for food and experiencing rapidbrain growth but slow somatic growth, was in place by thetime of H. erectus (Bogin 2006; Thompson and Nelson 2011).Contemporary human biology suggests that this life historyshift in H. erectus most likely would have involved a short-ening of infancy with earlier weaning and probably alsoshorter IBIs. Importantly, this pattern would have resulted inhigher fertility and greater potential for population increase.Extending childhood by even a year would allow more timefor cognitive development, including the development of eco-

logical skills such as in foraging as well as the refinement ofsocial behaviors (Bogin 1999).

In order to assess, interpret, and characterize a species’ lifehistory pattern, we need to study multiple somatic systemssimultaneously (Leigh and Blomquist 2007, 2011; Seselj 2011).While multisystem studies have only begun to be applied tothe fossil record, in large part because of a dearth of associatedskeletal remains, they point to the need for extensive researchon extant taxa for which somatic and physiological data areknowable. So far, these integrative studies have focussed onthe hard tissues of the extant and fossil record (Clegg andAiello 1999; Guatelli-Steinberg 2009; Seselj 2011), and manymore such studies are needed. Further, there is a critical needfor studies that reach across both living and skeletal popu-lations to combine hard-tissue parameters (e.g., age, sex, andsize proxies), soft-tissue measures, and physiological data inliving humans, nonhuman primates, and other mammals (seehttp://bonesandbehavior.org; Smith et al. 2012). Work thatlinks conditions of nutritional stress to variation in both skel-etal maturation (e.g., Frisancho, Garn, and Ascoli 1970) anddental emergence patterns (e.g., Gaur and Kumar 2012) sug-gests that multiple modalities are influenced by this devel-opmental process. A key step forward will be to define datasets in extant taxa that are explicitly designed to be collectedand/or closely proxied in the fossil record in order that phys-ical or archaeological clues can be identified as signals fordevelopment of behavioral or physiological shifts in deep time(see http://bonesandbehavior.org).

The other significant life history shift that arguably emergedin early H. erectus is the extended time to maturity throughan elongated adolescence coupled with the development of apronounced late adolescent growth spurt (Bogin and Smith1996). The extant record indicates that this most likely wouldhave involved a reduction in extrinsic mortality risk andgreater nutritional access and stability than in early homininspecies (e.g., Robson and Wood 2008). Fossil and archaeo-logical evidence is consistent with increased access to higher-quality foods (i.e., those with relatively high energy and nu-

http://bonesandbehavior.org

http://bonesandbehavior.org


trient density; Potts 2012; Ungar 2012), resulting in potentiallyfewer periods of nutritional inadequacy that would have re-duced associated declines in immune function.

Beyond diet, the extant record strongly implicates reducedextrinsic mortality as a means of increasing size and delayingdevelopment; however, it remains unclear just how mortalityrisk might have been lowered for H. erectus. An importantclue may come from the extensive system of cooperative be-havior and breeding seen in modern humans. Cooperativebehavior, defined here as behaviors that provide a benefit toanother individual and may or may not have a cost to theactor, occurs widely in the natural world, yet the degree ofcooperation between unrelated individuals is unique to hu-mans (Clutton-Brock 2009; Melis and Semmann 2010). Incooperative breeders, allocare (including paternal care) allowsthe mother to channel resources to her own somatic main-tenance and reproduction; thus, allocare should generally befavored evolutionarily when the risk to the offspring is nottoo high (Lappan 2009; Ross and MacLarnon 2000). Amongmammalian species, those with greater allocare exhibit rela-tively rapid infant growth with earlier weaning and fasterreproductive (birth) rates, although these infants are notlarger at birth (Borries et al. 2011; Isler and van Schaik 2009;Mitani and Watts 1997; Ross and MacLarnon 2000; Smith etal. 2012). The well-developed system of cooperation in hu-mans plays a critical role in supporting the high costs ofencephalization that must be paid during pregnancy and lac-tation (Ellison 2008; Kramer 2010; Wells 2012) and is a majorfactor in enabling early weaning, relatively low extrinsic mor-tality, extended subadult dependence, and high fertility(Gurven and Hill 2009; Hill and Hurtado 2009; Kaplan et al.2000; Lancaster and Lancaster 1983). Thus, cooperative breed-ing was almost certainly a critical contributor to brain sizeincrease in the Homo lineage, although the timing of thisoccurrence is elusive.

An important observation related to the likely presence ofcooperative breeding in H. erectus is the link between de-mographic viability and encephalization. Isler and van Schaik(2012) suggest that demographic viability in primates is un-tenable at average cranial capacities over 700 cm3 (i.e., a “grayceiling”) because of low fertility related to a protracted sub-adult period characterized by rapid brain growth but slowsomatic growth. Smith (2012) and colleagues also found sup-port from the carnivores for the idea of the co-occurrence ofbrain size expansion and cooperative breeding. Cooperativebreeding in the form of direct care and the provisioning ofjuveniles with high-quality resources (e.g., animal fat and pro-tein) would have enabled early H. erectus to circumvent thisdemographic constraint and evolve a relatively large brainwhile also having a life history pattern with early weaningand short IBI that led to greater fertility and facilitated pop-ulation growth. Thus, a system of cooperative breeding, al-though not as well developed as in modern humans, seemslikely to have been in place by the time of H. erectus if not

in non-erectus Homo (Bribiescas, Ellison, and Gray 2012; Gett-ler 2010; Key and Aiello 2000; Swedell and Plummer 2012).

Additionally, once cooperative breeding is present, we ex-pect a fundamental shift in social organization that may bevisible in the archaeological record (Potts 2012; Smith et al.2012; Swedell and Plummer 2012). This may be reflected inevidence for greater or more complicated extractive foraging(Swedell and Plummer 2012) or in the aggregation of multipleindividuals (Potts 2012; Smith et al. 2012). Future archaeo-logical endeavors should aim to identify material cultural sig-natures reflecting these shifts. This research could potentiallybe coupled with stable isotope studies, which have showngreat potential for identifying signatures of population move-ment during the lifetime of an individual (e.g., Copeland etal. 2011). Finally, we suggest that future studies focus on otheraspects of extrinsic mortality relevant to shaping body sizeand shape, including predation rates as well as contributorsto intrinsic mortality rates such as dietary breadth, quality,and availability.

Regional Variation, Climatic Adaptation, and Dispersal inHomo erectus

By the time of H. erectus, the trend toward greater rangingthat may have started at the base of the genus had blossomedinto long-range dispersals into a variety of different climaticcontexts (e.g., the Republic of Georgia and tropical southeastAsia; Anton and Swisher 2004). Widely dispersed living mam-mals face a number of similar challenges and tend to sharea number of attributes including behavioral plasticity, soci-ality, and relatively high rates of reproduction (i.e., high in-trinsic rates of natural increase; Anton, Leonard, and Rob-ertson 2002). In addition, contemporary humans add greateradiposity that buffers individuals in shifting environments andalso allows maintenance of brain metabolic requirements,both of which are critical to successful dispersal (Kuzawa1998; Leonard et al. 2003; Wells 2010). And humans exhibitgreat developmental plasticity that preserves flexibility in theface of short-term environmental changes (Walker et al.2006). As such, it seems likely, based on what we know ofthe extant record, that H. erectus (1) had a different bodycomposition than earlier hominins, with higher levels of ad-iposity; (2) possessed a level of developmental plasticity sim-ilar to that seen in modern humans, which may help explainthe long existence of this species; and (3) may have had greaterbehavioral plasticity, which would have favored their successover less versatile members of genus Homo (see Smith et al.2012).

To see how body composition might have changed, we lookto humans who differ from other mammals (including non-human primates) by having particularly high levels of fat,large brains, small guts, and low muscularity (Aiello andWheeler 1995; Leonard et al. 2003; Wells 2010). These dif-ferences in body composition structure variation in energydemands because of marked differences in organ-specific met-


abolic rates. While most internal organs—such as the heart,lungs, kidneys, liver, and spleen—appear to be tightly scaledwith body mass (Calder 1984; Stahl 1965), the brain, gut,skeletal muscle, and adipose tissue vary according to func-tional demands (Aiello and Wheeler 1995; Calder 1984;Muchlinski, Snodgrass, and Terranova 2012; Schmidt-Nielsen1984; Wells 2010). Nonhuman primates are “undermuscled”when compared with other mammals, which likely reflectsthe arboreal heritage of the order. Humans, and especiallyhuman females, appear to be even less muscular (Muchlinski,Snodgrass, and Terranova 2012; Snodgrass, Leonard, and Rob-ertson 2009). Although this could be an adaptation to reduceenergetic costs associated with bipedal locomotion, it is morelikely a reflection of our high levels of adipose tissue for aprimate of our size.

Thus, besides the brain, the single most important com-ponent of body composition for understanding human evo-lution is arguably adipose tissue (Wells 2010, 2012). This tissueis closely linked to brain development and immune function,likely underpins the exceptional dispersal abilities of our ge-nus, and helps explain our ability to withstand seasonal andperiodic fluctuations in food availability (Kuzawa 1998; Wells2010, 2012). Humans are exceptional in having fat storesconsiderably larger than most free-living primates and ter-restrial tropically living mammals, and this is true for non-Western human populations as well (body fat levels average25% for adult females and 13% for adult males; Pond 1998;Wells 2006, 2010). Humans are extremely fat at birth (∼15%fat) and during infancy (peaking at ∼25%–30% fat), whichcontrasts markedly with wild primates (baboons, 3%), do-mesticated species (pigs, 1.3%), and even seals (harp seals,10.4%; Kuzawa 1998). Adipose tissue in humans serves pri-marily as a nutritional buffer against long-term (e.g., seasonalor periodic) decreases in energy availability, and fat is animportant adaptation for preserving cerebral metabolism inthe face of the high and obligate metabolic demands of thelarge human brain (Kuzawa 1998; Leonard et al. 2003). Fur-ther, human sex differences in adiposity are shaped by dif-ferences in reproductive strategies—in particular, the enor-mous energetic costs of pregnancy and lactation borne byfemales (Snodgrass 2012; Valeggia and Ellison 2001). This shiftin body composition and concomitant increased energeticbuffering (i.e., “somatic capital” of Kaplan et al. 2000) mayhave played a central role in the ability of H. erectus to suc-cessfully disperse into new environments, especially those withseasonal and periodic variation in climate and food avail-ability.

Our ability to identify shifts in body composition in thefossil record is, of course, limited, and we are further con-strained by the surprisingly little body composition data avail-able for living primates and other mammals. Refining ourunderstanding of body composition in extant species will helpto identify which aspects of body composition in humans arederived and to outline adaptive scenarios related to their di-vergence (Wells 2012). It may also allow more nuanced pre-

dictions of body composition in fossil hominins as well astests as to when in our lineage adiposity and sex-specificpatterns of adiposity arose.

Plavcan (2012) notes that because of the difference betweentotal body mass and lean body mass in humans and the dif-ferential distribution of fat in human females, degrees of cra-nial and postcranial skeletal variation differ in humans butnot other apes. As in all other primates, postcranial variation(as reflected in CVs of linear dimensions) is similar to thatof lean body mass variation (excluding adipose tissue) inhumans. But unlike other primates, these CVs are greater thancranial CVs. Cranial variation in humans is similar to that oftotal body mass variation (including adipose tissue, whichmay indicate the importance of adipose tissue to brain main-tenance). Presumably this reflects increased adiposity in hu-mans and differential fat distribution in human females versusmales. Thus, finding the point at which measures of cranialand postcranial skeletal dimorphism diverge may provide apreliminary clue as to when greater (or at least differential)adiposity arose in the lineage. At present, endocranial andfemoral length CVs are similar to one another within earlynon-erectus Homo, early H. erectus, and A. afarensis. Thus, atleast the differential fat distribution seen in human males andfemales had yet to develop by the time of early H. erectus,although we currently have no window on to whether in-creased adiposity was present in both sexes (table 1; and seetable 5 in Anton 2012).

Our inference of greater developmental plasticity in H.erectus is supported by the variation seen in size across re-gional samples but is also an insight that requires that we usecaution in interpreting the meaning of morphological differ-ences among samples in body proportions, size, and sexualdimorphism. Caution is required for several reasons. First,some of the size variation shows a temporal trend in H. erectus(see Plavcan 2012). Second, total variation in H. erectus is notparticularly remarkable relative to extant primates (Plavcan2012). Third, it is well established that developmental plas-ticity can shift these signals rapidly in extant human andnonhuman primates (e.g., Bogin and Rios 2003) and for avariety of different reasons (Kuzawa and Bragg 2012).

To test the extent to which developmental plasticity waspresent in H. erectus and how similar it was to the humanform, we need to understand how such plasticity is reflectedin the skeletons of humans and nonhuman primates. Sur-prisingly, the extent of variation in developmental plasticityamong primates is not well studied, and the lack of these datais a barrier to interpreting variation in the human fossil rec-ord. One study of baboons demonstrates the potential fordramatic shifts in growth, reproduction, and body size withaltered environmental conditions (Altmann and Alberts2005). Garbage-foraging baboons Papio cynocephalus showfaster maturation and larger body size than other savannahbaboons as a result of better food availability during ontogeny.Given their generalized ecologies and broad geographic dis-tributions, papionin monkeys (baboons, mandrills, and ma-


caques) and modern humans are arguably the best analoguesfor early Homo, especially H. erectus (Jolly 2001; Swedell andPlummer 2012), and should prove a fruitful area of focus forfuture studies.

A related issue with importance for interpreting the hom-inin fossil record is the influence of developmental plasticityon sexual dimorphism (see Bribiescas, Ellison, and Gray2012). Although both sexes experience developmental plas-ticity, males are disproportionately able to capitalize on high-quality environments, whereas females are more environ-mentally buffered and less negatively influenced by poorenvironments (Altmann and Alberts 2005; Kuzawa 2007; Stin-son 1985). Thus, sexual dimorphism can shift rapidly overtime with reduced sexual dimorphism in bad times and ac-centuated sexual dimorphism under more optimal environ-mental conditions (Stini 1972, 1975). This may hint at thecause of the apparent reduction in dimorphism in GeorgianH. erectus (see table 2). Additionally, environmental condi-tions experienced during development influence testosteroneand thus shape sexually dimorphic traits, including stature,bone growth, and muscle mass (Bribiescas, Ellison, and Gray2012; Kuzawa et al. 2010). This topic has not been system-atically studied across primates, but we believe it should formthe basis for future investigations as it has important impli-cations for making inferences from morphological variation.

Although we recognize that other, longer-term forces suchas mate competition are critical to shaping differences in sex-ual dimorphism across taxa (see Plavcan 2012), a more sys-tematic understanding of intraspecific variation across geo-graphic, environmental, and nutritional contexts in primatesis critical to contextualizing variation in the fossil record.Evaluations of specific skeletal responses (e.g., brow devel-opment) to environmental signals in extant taxa may helpelucidate the meaning of sexual dimorphism in the fossil rec-ord (Bribiescas, Ellison, and Gray 2012). More systematicstudies of geographic variation in nonhuman primate sam-ples, both skeletal and living, that pay particular attention tohow the adult form of skeletal traits (including overall sizebut also secondary sex characteristics such as robusticity) isaffected by developmental plasticity and how the sexes aredifferentially affected in different environments will greatlyimprove our ability to differentiate adaptation from epiphe-nomenal variation in the fossil record (Plavcan 2012; see Fer-nandez-Duque 2011 for an example of the use of skeletalproxies from living animals).

Another means of differentiating among hypotheses forintraspecific variation in body size and shape considers dif-ferences in proportions due to the timing of growth disrup-tions. Poor growth related to environmental conditions typ-ically occurs during infancy when growth rate is rapid andthe body is uniquely vulnerable to insult. In humans, thisheightened vulnerability is associated with the introductionof supplemental, often low-quality foods, which usually beginsat ∼4–6 months of age. These foods may also inadvertentlyintroduce pathogens (Sellen 2001; Snodgrass, Leonard, and

Robertson 2009). For this reason, poor growth resulting fromenvironmental conditions disproportionately affects limbsand their distal segments, which grow at a more rapid ratethan the trunk during infancy. At ∼2–3 years of age, declininggrowth rates and more developed immune and digestive sys-tems reduce the risks of permanent growth disruptions (Bogin1999; Kuzawa 1998). As a result, the secular increase in heightexperienced by most human populations in the twentiethcentury was associated with disproportionate gains in limblength, particularly distal segments (Stinson 2012). In fact,relatively short legs are interpreted as reflecting an adverseearly developmental environment (Bogin and Varela-Silva2010). Although genetic factors related to ultimate causes suchas climatic adaptation (Katzmarzyk and Leonard 1998; Rob-erts 1978) are important contributors to body proportions,proximate factors such as nutrition during developmentclearly play an important role in humans (Bogin and Rios2003; Eveleth and Tanner 1990; Stinson 2012). Studies of howthe skeleton is affected by nutritional insufficiency during thelonger weaning period of great apes will be important toconsidering the applications to the fossil record.

Further, exposure to persistent and ubiquitous stressorsthat are not effectively buffered by cultural/behavioral mech-anisms will lead to adjustments initially through develop-mental plasticity and later, if experienced at a population levelover multiple generations, by genetic changes resulting frompolygenic adaptation (Kuzawa and Bragg 2012). Thus, studiesthat combine dental, cranial, and postcranial analysis can po-tentially expand our ability to interpret variation in body sizeand proportions seen among regional samples of H. erectus.While the state of the fossil record is currently quite far fromadequate for such purposes, regional samples of H. erectusmay begin to be probed using integrative studies, and addi-tional research on extant taxa will help provide the compar-ative foundation for this work.

We could hypothesize, for example, that the smaller overallsize of Georgian H. erectus is due to decreased nutritionalsufficiency during development or increased extrinsic mor-tality (due to predation or disease; Anton 2012; Migliano andGuillon 2012). Or, we might predict it results from small-packet resources that are widely dispersed in a topographicallychallenging area with little selective pressure for large bodysize—an explanation offered for the small adult body size ofLate Stone Age humans of southern Africa (Pfeiffer 2012).Or, the apparently shortened arms of the Dmanisi group mayperhaps reflect climatic adaptation (Pontzer 2012).

We could test these hypotheses by considering the specificanatomical and archaeological signatures each implies. Forexample, nutritional stress-related small adult body size likelywould be accompanied by shortened distal limb segments andmarked enamel hypoplasias, the latter of which provide apermanent record of systemic physiological stress, whereasclimatic adaptation might result in shortened arms but notnecessarily differentially short distal limb segments in botharms and legs. Alternatively, increased extrinsic mortality such


Figure 1. A positive feedback loop between cooperative behavior (initially in breeding), diet quality and stability, cognitive abilities(brain size), and extrinsic mortality risk drove life history evolution and contributed to cultural change in genus Homo. Gradual,self-reinforcing shifts in these central elements had consequences for life history traits including extending the developmental period,increased fertility, and larger body size; body composition including increased adiposity, reduced gut size, and reduced muscularity;communication including eventually the development of language; and cultural change including more complex extractive foraging.Early Homo showed only modest increases in the central elements. The fully modern package of life history and other consequencesmay not have emerged until recent humans.

as predation and parasites should lead to differences in neo-natal size (Kuzawa and Bragg 2012) relative to nonstressedgroups but not necessarily to indicators of nutritional stressand may also yield archaeological signals. Evidence of packetsize may come from archaeological evidence. Surprisingly,research of this nature in contemporary humans and non-human primates is fairly limited. However, a study amongAboriginal Australians—which found that permanent stunt-ing was seen only in individuals with enamel defects that wereearly (within the first 18 months of life), severe (enough toproduce paired enamel defects), and repeated (during infancyand childhood; Floyd and Littleton 2006)—shows the poten-tial of multisystem studies. Such an integrative approach mayhelp us interpret body size and proportion variation in Homoand to differentiate adaptive variation from responses to prox-imate environmental factors such as diet and disease.

Finally, we suggest that from other records, local environ-mental signals should also be plumbed and developed to un-derstand the specific as well as the regional and global contextof fossil groups. If extrinsic mortality has such importantconsequences for size and shape variation, then additionalmeans of assessing extrinsic mortality must be pursued. Asmentioned earlier, these include archaeological means for as-sessing predation and diet as well as geochemical means forreconstructing plausible climates and diets. Thus, we advocatea multipronged approach to future research agendas that (ob-viously) includes collection of new fossil hominins and a focusin extant mammals on skeletal end results of environmentaland physiological parameters, especially in widely dispersedtaxa.

A Model for the Origins and Evolution ofGenus Homo

The integration of paleoanthropological data with informa-tion from primatology and human biology leads us to theconclusion that the origin and evolution of early Homo wascharacterized by a positive feedback loop that drove life his-tory evolution and contributed to cultural change. The centralelements of this model are cooperative behavior, diet, cog-nitive abilities, and extrinsic mortality risk (fig. 1). The modelpostulates gradual self-reinforcing shifts in these central ele-ments with consequences for life history traits (e.g., extendeddevelopmental period, increased fertility, and larger bodysize), body composition (e.g., adiposity, gut size, and mus-cularity), communication abilities (the development of lan-guage), and cultural change (tool use). The model expandson Hrdy’s (2009) cooperative breeding hypothesis, which pos-tulates that beginning with the rise of the genus Homo, allo-maternal care and provisioning drove life history evolution,and it recognizes, as does Kaplan et al. (2000), that reducingmortality rates, investing in embodied capital (fat), and in-creasing cooperation are in a positive feedback loop with brainsize. However, it does not rest on a particular kind of foodresource or social structure but recognizes that increasing dietquality and/or throughput and cooperation remain critical togrowing big brains and large bodies.

At present, it is impossible to identify the initial evolu-tionary change or changes, but it seems most likely that be-havioral changes related to diet and perhaps cooperation wereearly additions. In contrast, encephalization would likely have


been a secondary change, because comparative studies suggestthat alterations in diet quality and body composition werenecessary preconditions of hominin brain expansion. Further,reductions in mortality risk (both intrinsic and extrinsic) mostlikely would have been substantially influenced by dietaryshifts and increased cooperative behavior and thus wouldlikely have been downstream changes.

Present fossil and archaeological evidence suggests sub-stantial changes in diet occurred initially with non-erectusearly Homo and were followed by marked dietary change inHomo erectus. In particular, earliest Homo likely consumed asubstantially higher-quality diet than Australopithecus andParanthropus, as the result of the consumption of high-qualityplant foods (e.g., underground storage organs) as well as an-imal source foods. Homo erectus appears to have occupied anew ecological position for hominins that almost certainlyinvolved a considerable increase in access to animal foods.This dietary shift to more energy- and nutrient-dense foodswould potentially have allowed for an increase in brain sizeby removing constraints on brain growth; in addition, thisdietary change may have selected for increased brain size andcognitive capacity related to increased foraging, extraction,and processing abilities associated with higher-quality diets.The reliance on high-quality foods may have also selected forcooperative social systems that would have increased the abil-ity to hunt and process foods. A variable and flexible systemof cooperative breeding would have reduced extrinsic mor-tality risk even further, especially for juveniles, through directcare and provisioning, and it would have contributed ener-getically to reproductive-aged females. Cooperative breedingwould have contributed to the ability of hominins to supportthe growth and high maintenance costs of large brains amongjuveniles through care and provisioning and would also haveselected for enhanced social cognitive processes that may haveled to further increases in brain size.

While we do not suggest that a fully human pattern of lifehistory traits (e.g., extended developmental period, increasedfertility, and larger body size), body composition, commu-nication abilities, and cultural change was present in earlyHomo or H. erectus, by the time of H. erectus the archaeo-logical record of dispersal provides evidence of sufficient plas-ticity and perhaps adiposity to colonize various environments.The best evidence for developmental plasticity in H. erectuscomes from the degree of morphological variation in size inthe species both within and outside of Africa, which providesnot only evidence of long-term trends but also short-termvariability at all times and in all places. Our expectation isthat regional morphs of H. erectus were established fairlyquickly but that significant population divergence was miti-gated by these same short-term developmental parameters.

The greatest advances in understanding the evolution ofthe early genus Homo will be guided by multipronged researchagendas that pay careful attention to determining the localenvironmental conditions (broadly understood) of fossilgroups and coordinate this work with multisystem evaluations

of the behavior, physiology, and anatomy of extant groups.These data sets must be explicitly designed to be measurableor closely proxied in the fossil record.

Acknowledgments

We benefited greatly from the lively discussions at the Wenner-Gren Foundation workshop “Human Biology and the Originsof Homo” held in Sintra, Portugal, in March 2011. We thankLeslie Aiello, who co-organized the workshop with S. C. An-ton, our fellow participants, and Laurie Obbink for makingthe Sintra meeting such a productive, stimulating, and con-vivial week. Following Sintra, Chloe, Bill Leonard, and Her-man Pontzer engaged in critical discussions of energetics andorgan-specific metabolism, Leslie Aiello and Tom Schoene-mann provided important input on the early conception ofthis paper, Chris Kuzawa provided critical feedback on a lateriteration, and Leslie Aiello and the reviewers greatly improvedthe final paper. Emily Middleton and the Current Anthro-pology staff provided expert editorial advice.

References CitedAiello, Leslie C., and Susan C. Anton. 2012. Human biology and the origins

of Homo: an introduction to supplement 6. Current Anthropology 53(suppl.6):S269–S277.

Aiello, Leslie C., and Peter Wheeler. 1995. The expensive-tissue hypothesis:the brain and the digestive system in human and primate evolution. CurrentAnthropology 36:199–221.

Altmann, Jeanne, and Susan C. Alberts. 2005. Growth rates in a wild primatepopulation: ecological influences and maternal effects. Behavioral Ecologyand Sociobiology 57:490–501.

Anton, Susan C. 2003. A natural history of Homo erectus. Yearbook of PhysicalAnthropology 46:126–170.

———. 2008. Framing the question: diet and evolution in early Homo. InPrimate craniofacial function and biology. Christopher J. Vinyard, MatthewJ. Ravosa, and Christine E. Wall, eds. Pp. 443–482. New York: Springer.



Anton, Susan C., and Carl C. Swisher III. 2004. Early dispersals of Homo fromAfrica. Annual Reviews of Anthropology 33:271–296.


Boas, Franz. 1912. Changes in bodily form of descendants of immigrants. NewYork: Columbia University Press.

Bogin, Barry B. 1999. Patterns of human growth. 2nd edition. Cambridge:Cambridge University Press.

———. 2006. Modern human life history: the evolution of human childhoodand fertility. In The evolution of human life history. Kristen Hawkes andRichard R. Paine, eds. Pp. 197–230. Santa Fe, NM: School of AmericanResearch.

Bogin, Barry, and Luis Rios. 2003. Rapid morphological change in livinghumans: implications for modern human origins. Comparative Biochemistryand Physiology 136A:71–84.


Bogin, Barry, and Maria Ines Varela-Silva. 2010. Leg length, body proportions,and health: a review with a note on beauty. International Journal of Envi-ronmental Research and Public Health 7:1047–1075.

Borries, Carola, Amy Lu, Kerry Ossi-Lupo, Eileen Larney, and Andreas Koenig.


2011. Primate life histories and dietary adaptations: a comparison of Asiancolobines and macaques. American Journal of Physical Anthropology 144:286–299.



Calder, William A. 1984. Size, function, and life history. Cambridge, MA: Har-vard University Press.

Charnov, Eric L. 1993. Life history invariants: some explanations of symmetryin evolutionary ecology. Oxford: Oxford University Press.

Clegg, Margaret, and Leslie C. Aiello. 1999. A comparison of the NariokotomeHomo erectus with juveniles from a modern human population. AmericanJournal of Physical Anthropology 110:81–93.

Clutton-Brock, Tim. 2009. Cooperation between non-kin in animal societies.Nature 462:51–57.

Collard, Mark, and Bernard Wood. 2007. Defining the genus Homo. In Phy-logeny of hominids, vol. 3. of Handbook of paleoanthropology. Winfried Henkeand Ian Tattersall, eds. Pp. 1575–1611. Berlin: Springer.

Copeland, Sandy R., Matthew Sponheimer, Darryl J. de Ruiter, Julia A. Lee-Thorp, Daryl Codron, Petrus J. le Roux, Vaughan Grimes, and Michael P.Richards. 2011. Strontium isotope evidence for landscape use by early hom-inins. Nature 474:76–79.

Dean, Christopher, Meave G. Leakey, Don 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, Christopher, and B. Holly Smith. 2009. Growth and development ofthe Nariokotome youth, KNM-WT 15000. In The first humans: origin andearly evolution of the genus Homo. Fred E. Grine, John G. Fleagle, andRichard E. Leakey, eds. Pp. 101–120. New York: Springer.

Dirks, Wendy, and Jacqui E. Bowman. 2007. Life history theory and dentaldevelopment in four species of catarrhine primates. Journal of Human Evo-lution 53:309–320.

Domınguez-Rodrigo, Manuel, Travis R. Pickering, and Henry T. Bunn. 2010.Configurational approach to identifying the earliest hominin butchers. Pro-ceedings of the National Academy of Sciences of the USA 107(49):20929–20934.

———. 2011. Reply to McPherron et al.: doubting Dikika is about data, notparadigms. Proceedings of the National Academy of Sciences of the USA108(21):E117.

Ellison, Peter T. 2008. Energetics, reproductive ecology, and human evolution.Paleoanthropology 2008:172–200.

Eveleth, Phyllis B., and James M. Tanner. 1990. Worldwide variation in humangrowth. 2nd edition. Cambridge: Cambridge University Press.

Fernandez-Duque, Eduardo. 2011. Rensch’s rule, Bergman’s effect and adultsexual dimorphism in wild monogamous owl monkeys of Argentina. Amer-ican Journal of Physical Anthropology 146:38–48.

Flinn, Mark V. 2010. Evolutionary biology of hormonal responses to socialchallenges in the human child. In Human evolutionary biology. Michael P.Muehlenbein, ed. Pp. 405–424. Cambridge: Cambridge University Press.

Floyd, Bruce, and Judith Littleton. 2006. Linear enamel hypoplasia and growthin an Australian Aboriginal community: not so small, but not so healthyeither. Annals of Human Biology 33:424–443.

Frisancho, A. Roberto, Stanley M. Garn, and Werner Ascoli. 1970. Unequalinfluence of low dietary intakes on skeletal maturation during childhoodand adolescence. American Journal of Clinical Nutrition 23:1220–1227.


Gaur, Rajan, and Pawan Kumar. 2012. Effect of undernutrition on deciduoustooth emergence among Rajput children of Shimla District of HimachalPradesh, India. American Journal of Physical Anthropology 148:54–61.

Gettler, Lee T. 2010. Direct male care and hominin evolution: why male-childinteraction is more than a nice social idea. American Anthropologist 112:7–21.

Godfrey, Laurie R., Karen E. Samonds, William L. Jungers, and Michael R.Sutherland. 2003. Dental development and primate life histories. In Primate

life histories and socioecology. Peter M. Kappeler and Michael E. Pereira,eds. Pp. 177–203. Chicago: University of Chicago Press.


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

Gurven, Michael, and Kim Hill. 2009. Why do men hunt? a reevaluation of“man the hunter” and the sexual division of labor. Current Anthropology5:51–74.


Hill, Kim, and A. Magdalena Hurtado. 1996. Ache life history. New York:Aldine.

———. 2009. Cooperative breeding in South American hunter-gatherers.Proceedings of the Royal Society B: Biological Sciences 276:3863–3870.


Holliday, Trenton W., and Robert G. Franciscus. 2009. Body size and itsconsequences: allometry and the lower limb length of Liang Bua 1 (Homofloresiensis). Journal of Human Evolution 57:223–228.

Holloway, Ralph L., and Michael S. Yuan. 2004. Endocranial morphology ofA.L. 444-2. In The skull of Australopithecus afarensis. William H. Kimbel,Yoel Rak, and Donald C. Johanson, eds. Pp. 123–135. New York: OxfordUniversity Press.

Hrdy, Sarah B. 2009. Mothers and others: the evolutionary origins of mutualunderstanding. Cambridge, MA: Harvard University Press.

Indriati, Etty, and Susan C. Anton. 2008. Earliest Indonesian facial and dentalremains from Sangiran, Java: a description of Sangiran 27. AnthropologicalScience 116:219–229.

Isler, Karin, and Carel P. van Schaik. 2009. The expensive brain: a frameworkfor explaining evolutionary changes in brain size. Journal of Human Evo-lution 57:392–400.


Jolly, Clifford J. 2001. A proper study for mankind: analogies from the pa-pionin monkeys and their implications for human evolution. Yearbook ofPhysical Anthropology 44:177–204.

Kaplan, Bernice A. 1954. Environment and human plasticity. American An-thropologist 56:780–800.

Kaplan, Hillard S., Kim Hill, Jane Lancaster, and A. Magdalena Hurtado.2000. A theory of human life history evolution: diet, intelligence, and lon-gevity. Evolutionary Anthropology 9:156–183.


Key, Catherine, and Leslie C. Aiello. 2000. A prisoner’s dilemma model ofthe evolution of paternal care. Folia Primatologica (Basel) 71:77–92.

Kramer, Karen L. 2010. Cooperative breeding and its significance to the de-mographic success of humans. Annual Review of Anthropology 39:417–436.


———. 2007. Developmental origins of life history: growth, productivity, andreproduction. American Journal of Human Biology 19:654–661.


Kuzawa, Christopher W., Thomas W. McDade, Linda S. Adair, and NanetteLee. 2010. Rapid weight gain after birth predicts life history and repro-ductive strategy in Filipino males. Proceedings of the National Academy ofSciences of the USA 107:16800–16805.

Lancaster, Jane B., and Chet S. Lancaster. 1983. Parental investment: thehominid adaptation. In How humans adapt: a biocultural odyssey. DonaldJ. Ortner, ed. Pp. 33–69. Washington, DC: Smithsonian Institution.

Lappan, Susan. 2009. The effects of lactation and infant care on the adultenergy budgets in wild siamangs (Symphalangus syndactylus). AmericanJournal of Physical Anthropology 140:290–301.


Lasker, Gabriel W. 1969. Human biological adaptability. Science 166:1480–1486.

Leakey, Louis S. B., Phillip V. Tobias, and John R. Napier. 1964. A new speciesof the genus Homo from Olduvai Gorge. Nature 202:7–9.

Leakey, Meave G., Fred Spoor, M. Christopher Dean, Craig S. Feibel, SusanC. Anton, Christopher Kiarie, and Louise N. Leakey. 2012. New early Homofossils from Koobi Fora, northern Kenya, highlight facial diversity. Nature488:201–204.

Leigh, Stephen R. 2001. Evolution of human growth. Evolutionary Anthro-pology 10:223–236.

Leigh, Stephen R., and Gregory E. Blomquist. 2007. Life history. In Primatesin perspective. Christina J. Campbell, Agustın Fuentes, Katherine C.MacKinnon, Melissa Panger, and Simon K. Bearder, eds. Pp. 396–407. NewYork: Oxford University Press.

———. 2011. Life history. In Primates in perspective. 2nd edition. ChristinaJ. Campbell, Agustın Fuentes, Katherine C. MacKinnon, Simon K. Bearder,and Rebecca M. Stumpf, eds. Pp. 418–428. New York: Oxford UniversityPress.

Leonard, William R., Marcia L. Robertson, J. Josh Snodgrass, and ChristopherW. Kuzawa. 2003. Metabolic correlates of hominid brain expansion. Com-parative Biochemistry and Physiology 136A:5–15.

Lepre, Christopher J., Helene Roche, Dennis V. Kent, Sonia Harmand, RhondaL. Quinn, Jean-Philippe Brugal, Pierre-Jean Texier, Arnaud Lenoble, andCraig S. Feibel. 2011. An earlier origin for the Acheulian. Nature 477:82–85.

Lockwood, Charles A., William H. Kimbel, and Donald C. Johanson. 2000.Temporal trends and metric variation in the mandibles and dentition ofAustralopithecus afarensis. Journal of Human Evolution 39:23–55.


———. 2011. Tool-marked bones from before the Oldowan change the par-adigm. Proceedings of the National Academy of Sciences of the USA 108(21):E116.

Melis, Alicia P., and Dirk Semmann. 2010. How is human cooperation dif-ferent? Philosophical Transactions of the Royal Society B: Biological Sciences365:2663–2674.

Migliano, Andrea Bamberg. 2005. Why pygmies are small: ontogenetic im-plications of life history evolution. PhD dissertation, University of Cam-bridge.


Migliano, Andrea Bamberg, Lucio Vinicius, and Marta Mirazon Lahr. 2007.Life history trade-offs explain the evolution of human pygmies. Proceedingsof the National Academy of Sciences of the USA 104:20216–20219.

Mitani, John C., and David Watts. 1997. The evolution of non-maternal care-taking among anthropoid primates: do helpers help? Behavioral Ecology andSociobiology 40:213–220.

Muchlinski, Magdalena N., J. Josh Snodgrass, and Carl J. Terranova. 2012.Muscle mass scaling in primates: an energetic and ecological perspective.American Journal of Primatology 74:395–407.

Pfeiffer, Susan. 2012. Conditions for evolution of small adult body size insouthern Africa. Current Anthropology 53(suppl. 6):S383–S394.


Pond, Caroline M. 1998. The fats of life. Cambridge: Cambridge UniversityPress.


Pontzer, Herman, and Jason M. Kamilar. 2009. Great ranging associated withgreater reproductive investment in mammals. Proceedings of the NationalAcademy of Sciences of the USA 106(1):192–196.

Pontzer, Herman, David A. Raichlen, Robert W. Shumaker, Cara Ocobock,and Serge A. Wich. 2010. Metabolic adaptation for low energy throughputin orangutans. Proceedings of the National Academy of Sciences of the USA107(32):14048–14052.

Potts, Richard. 2012. Environmental and behavioral evidence pertaining tothe evolution of early Homo. Current Anthropology 53(suppl. 6):S299–S317.


Roberts, Derek F. 1978. Climate and human variability. 2nd edition. MenloPark, CA: Cummings.

Robson, Shannen L., and Bernard Wood. 2008. Hominin life history: recon-struction and evolution. Journal of Anatomy 212:394–425.

Roche, Helene, Robert J. Blumenschine, and John J. Shea. 2009. Origins andadaptations of early Homo: what archaeology tells us. In The first humans:origin and early evolution of the genus Homo. Fred E. Grine, John G. Fleagle,and Richard E. Leakey, eds. Pp. 197–207. New York: Springer.

Ross, Caroline, and Ann MacLarnon. 2000. The evolution of non-maternalcare in anthropoid primates: a test of the hypotheses. Folia Primatologica71:93–113.

Ruff, Christopher B. 2008. Femoral/humeral strength in early African Homoerectus. Journal of Human Evolution 54:383–390.

———. 2009. Relative limb strength and locomotion in Homo habilis. Amer-ican Journal of Physical Anthropology 138:90–100.

———. 2010. Body size and body shape in early hominins: implications ofthe Gona pelvis. Journal of Human Evolution 58:166–178.

Schmidt-Nielsen, Knut. 1984. Scaling: why is animal size so important? Cam-bridge: Cambridge University Press.

Schoenemann, P. Thomas. 2006. Evolution of the size and functional areasof the human brain. Annual Review of Anthropology 35:379–406.

Schroepfer, Kara K., Brian Hare, and Herman Pontzer. 2012. Energy expen-diture in semi free-ranging chimpanzees measured using doubly labeledwater. American Journal of Physical Anthropology 147(suppl. 54):263.


Sellen, Daniel W. 2001. Comparison of infant feeding patterns reported fornonindustrial populations with current recommendations. Journal of Nu-trition 131:2707–2715.

Semaw, Sileshi, Michael J. Rogers, Jay Quade, Paul R. Renne, Robert F. Butler,Dietrich Stout, Manuel Dominguez-Rodrigo, William Hart, Travis Pick-ering, and Scott Simpson. 2003. 2.6-million-year-old stone tools and as-sociated bones from OGS-6 and OGS-7, Gona, Afar, Ethiopia. Journal ofHuman Evolution 45:169–177.

Seselj, Maja. 2011. Human growth evolving: integrating dental and skeletalgrowth proxies to understand ontogeny in Pleistocene Homo. PhD disser-tation, New York University.

Shapiro, Harry L. 1939. Migration and environment: a study of the physicalcharacteristics of the Japanese immigrants to Hawaii and the effects of envi-ronment on their descendants. London: Oxford University Press.



Smith, Jennifer E., Eli M. Swanson, Daphna Reed, and Kay E. Holekamp.2012. Evolution of cooperation among mammalian carnivores and its rel-evance to hominin evolution. Current Anthropology 53(suppl. 6):S436–S452.

Snodgrass, J. Josh. 2012. Human energetics. In Human biology: an evolutionaryand biocultural approach. 2nd edition. Sara Stinson, Barry Bogin, and DennisO’Rourke, eds. Pp. 327–386. New York: Wiley.

Snodgrass, J. Josh, William R. Leonard, and Marcia L. Robertson. 2009. Theenergetics of encephalization in early hominids. In Evolution of hominiddiets: integrating approaches to the study of Palaeolithic subsistence. Jean-Jacques Hublin and Michael Richards, eds. Pp. 15–29. Dordrecht: Springer.

Spoor, Fred, Meave G. Leakey, Patrick N. Gathogo, Frank H. Brown, SusanC. Anton, Ian McDougall, Christopher Kiarie, Frederick K. Manthi, andLouise N. Leakey. 2007. Implications of new early Homo fossils from Ileret,east of Lake Turkana, Kenya. Nature 448:688–691.

Stahl, Walter R. 1965. Organ weights in primates and other mammals. Science150:1039–1042.

Stearns, Stephen C. 1992. The evolution of life histories. Oxford: Oxford Uni-versity Press.

Stini, William A. 1972. Reduced sexual dimorphism in upper arm musclecircumference associated with a protein-deficient diet in a South Americanpopulation. American Journal of Physical Anthropology 36:341–352.

———. 1975. Adaptive strategies of human populations under nutritionalstress. In Biosocial interrelations in population adaptation. Elizabeth S. Watts,Frances E. Johnson, and Gabriel W. Lasker, eds. Pp. 19–41. The Hague:Mouton.

Stinson, Sara. 1985. Sex differences in environmental sensitivity during growthand development. Yearbook of Physical Anthropology 28:125–147.

———. 2012. Growth variation: biological and cultural factors. In Human


biology: an evolutionary and biocultural approach. 2nd edition. Sara Stinson,Barry Bogin, and Dennis O’Rourke, eds. Pp. 587–635. New York: Wiley.

Swedell, Larissa, and Thomas Plummer. 2012. A papionin multilevel societyas a model for hominin social evolution. International Journal of Prima-tology, doi:10.1007/s10764-012-9600-9.

Swisher, Carl C., III, Garniss H. Curtis, Teuku Jacob, Ann G. Getty, AgusSuprijo, and Widiasmoro. 1994. Age of the earliest known hominids inJava, Indonesia. Science 263:1118–1121.

Thompson, Jennifer L., and Andrew J. Nelson. 2011. Middle childhood andmodern human origins. Human Nature 22:249–280.

Trevathan, Wenda R. 1987. Human birth: an evolutionary perspective. Haw-thorne, NY: Aldine de Gruyter.


Ungar, Peter S., Kristin L. Krueger, Robert J. Blumenschine, Jackson Njau,and Robert S. Scott. 2012. Dental microwear texture analysis of homininsrecovered by the Olduvai Landscape Paleoanthropology Project, 1995–2007.Journal of Human Evolution 63:429–437.

Ungar, Peter S., and Robert S. Scott. 2009. Dental evidence for diets of earlyHomo. In The first humans: origin and early evolution of the genus Homo.

Fred E. Grine, John G. Fleagle, and Richard E. Leakey, eds. Pp. 121–134.New York: Springer.

Ungar, Peter S., and Matt Sponheimer. 2011. The diets of early hominins.Science 334:190–193.

Valeggia, Claudia R., and Peter T. Ellison. 2001. Lactation, energetics andpostpartum fecundity. In Reproductive ecology and human evolution. PeterT. Ellison, ed. Pp. 85–105. New York: Aldine.

Walker, Robert, Michael Gurven, Kim Hill, Andrea Migliano, Napolean Chag-non, Roberta De Souza, Gradimir Djurovic, et al. 2006. Growth rates andlife histories in twenty-two small-scale societies. American Journal of HumanBiology 18:295–311.

Wells, Jonathan C. K. 2006. The evolution of human fatness and susceptibilityto obesity: an ethological approach. Biological Reviews 81:183–205.

———. 2010. The evolutionary biology of human body fatness: thrift and control.Cambridge: Cambridge University Press.



CS5 CAv53nS6 Cover CAsupplement PMS8641 · 2015-11-07 · Ecological Energetics in Early Homo...

Documents

Transcript of CS5 CAv53nS6 Cover CAsupplement PMS8641 · 2015-11-07 · Ecological Energetics in Early Homo...