48 Evolutionary Anthropology ARTICLES bipedalism. In Corruccini RS, Cio- 39 Steudel K, Beattie J (n.d.) Does limb length An allometric approach. Am J Physiol chon RL (eds), Integrative Paths to the Past: predict the energetic cost of locomotion in 247:R806-R815. Paleontological Advances in Honor of E Clark mammals? J Zoo1 (London), in press. 48 Cowan IR, Farquhar GD (1977) Stomata1 Howell, pp 269-284. Englewood Cliffs: Pren- 40 Cavanagh PR, Williams KR (1982) The ef- function in relation to leaf metabolism and tice Hall. fect of stride length variation on oxygen up- environment. Symp SOCExp Biol31:471-505. 30 Sutherland DH, Olshen R, Cooper L, Woo take during distance running. Med Sci Sports SLY (1980) The development of mature gait. J Exer 14:30-35. 49 McHenry HM (1991) Sexual dimorphism Bone Joint Surg 62A:336-353. 41 Cotes JE, Meade F (1960) The energy ex- in Australopithecus afarensis. J Hum Evol 31 Steudel K (n.d.)Limb morphology, bipedal penditure and mechanical energy demand in 20:2 1-32. gait and the energetics of hominid locomo- walking. Ergonomics 3:97-120. 50 Post DG (1980) Sexual dimorphism in the tion. Am J Phys Anthropol, in press. 42 Harris M, Steudel K (1993) Hindlimb anthropoid primates: Some thoughts on 32 Cavagna GA, Heglund NC, Taylor CR length correlates in the Carnivora. Am Zoo1 causes, correlates, and the relationship to (1977) Mechanical work in terrestrial locomo- 33:74A. body size, unpublished manuscript. tion: Two basic mechanisms for minimizing 43 Garland T Jr, Janis CM (1993) Does meta- energy expenditure. Am J Physiol 233:R243- 51 Day MH (1977) Locomotor adaptations in tarsal/femur ratio predict maximal running man. Biol Hum Affairs 42:149-15 1. R261. speed in cursorial mammals? J Zoo1 (London) 33 Lovejoy CO (1981) The origin of man. Sci- 229:133-15 1. 52 Ravey M (1978) Bipedalism: An earlywarn- ence 21 1:2441-3450. 44 Tsuji JS, Huey RB, Van Berkum FH, Gar- ing system for Miocene hominoids. Science 34 Stern JT Jr, Susman RL (1983) The locomo- land T Jr, Shaw RG (1989) Locomotor per- 199:372. tor anatomy of Australopithecus afarensis. Am formance of hatchling fence lizards 53 Wheeler PE (1984) The evolution of J Phys Anthropol60:279-317. (Sceloporus occidentalis): Quantitative genet- bipdedality and loss of functional body hair in 35 Janis CM, Wilhelm PB (1993) Were there ics and morphometric correlates. Evol Ecol hominids. J Hum Evol 13:91-98. mammalian uursuit oredators in the Tertiarv? 3:240-252. Dances with‘ wolf ivatars. J Mammal E;ol 45 Garland T Jr (1985) Ontogenetic and indi- 54 Du Brul EL (1 962) The general phenome- I:103-125. vidual variation in size, shape and speed in the non of bipedalism. Am Zoo1 2:205-208. 36 Klein DR, Meldgaard M, Fancy SG (I 987) Australian agamid lizard arnphibolurus 55 Jolly CJ (1970) The seed-eaters: A new Factors determining leg length in Rangifer nuchalis. J Zoo1 (London) 207:425-439. model of hominid differentiation based on a tarandus. J Mammal 68:642-655. 46 Losos JB (1990) The evolution of form and baboon analogy. Man 5:l-26. 37 Jungers WL (1982) Lucy’s 1imbs:Skeletal function: Morphology and locomotor per- allometry and locomotion in Australopithecus formance in West Indian Anolis lizards. Evo- 56 Sinclair ARE, Leakey MD, Norton-Grif- afarensis. Nature 297676-678. lution 44:1189-1203. fiths M (1987) Migration and hominid bipedalism. Nature 325:307-308. 38Wolpoff MH(1983)Lucy’slittlelegs.JHum 47 Garland T Jr (1984) Physiological corre- Evol 12:443453. lates of locomotory performance in a lizard: 0 1994 Wiley-Liss, Inc

Multiple Dispersals and Modern Human Origins

MARTA MIRAZON LAHR and ROBERT FOLEY

Despite a massive endeavour, the problem of modern human origins not only ing regional ones. Weidenreich,’ who remains unresolved, but is usually reduced to “Out of Africa” versus multiregional first proposed the theory of multire- evolution. Not all would agree, but evidence for a single recent origin is accumulat- gional evolution, explained regional ing. Here, we want to go beyond this debate and explore within the “Out of Africa” differences in morphology between framework an issue that has not been fully addressed: the mechanism by which modem groups like Asians and Austra- modern human diversity has developed. We believe there is no clear rubicon of lians as resulting from relatively inde- modern Homo sapiens, and that multiple dispersals occurred from a morphologically pendent evolution from Sinanthyopus variable population in Africa. Pre-existing African diversity is thus crucial to the way and Pithecanthropus. The early mul- human diversity developed outside Africa. The pattern of diversity-behavioural, tiregional models suffered from the linguistic, morphological and genetic-can be interpreted as the result of dispersals, lack of a mechanism for the mainte- colonisation, differentiation and subsequent dispersals overlaid on former popula- nance of worldwide parallelisms.*,2 tion ranges. The first dispersals would have originated in Africa from where two different geographical routes were possible, one through EthiopidArabia towards Marta Mirazon Lahr is a fellow of Clare South Asia, and one through North AfricdMiddle East towards Eurasia. College, Cambridge. Her research focuses on modern human origins and subsequent differentiation. Robert Foley is Director of the Duckworth Laboratory at the University A model of multiregional evolution and subsequent regional di- of Cambridge and a Fellow of King’s was the first comprehensive theory of versity as resulting from the transfor- College. He has carried out research into the evolutionary biology of fossil hominids, the evolution of modern humans from mation of archaic hominid groups and he is the author of Another Unique their hominid ancestors. Multire- into modern populations in each part Species as well as editor of several books. gional evolution in the Pleistocene ex- of the world. Modern human features Key words: Hominid, colonization, evolution, plains both the origins of modern have been superimposed on pre-exist- population diversity ARTICLE5 Evolutionary Anthropology 49

Recently however, Wolpoff and others tradition (the Howieson’s Poort) be- the technique. Indeed, it apparently is 3-5 have proposed a modified version tween 85 and 60 ky in South impossible at the moment to prove of this earlier theory in which gene The first archeological assemblages statistically the branching pattern of flow takes a major role. Accordingly, that present blade tools are found mtDNA lineages. However, contrary the multiregional model proposes that around the Mediterranean.25In North to what has been claimed,43 these each modern human regional popula- Africa, there is evidence of a pre- problems do not completely discredit tion arose from archaic regional in- tradition,26 while early the genetic evidence. The great diver- habitants, and that a balance between upper assemblages have sity of African mtDNA lineages re- gene flow and isolation allowed re- been identified in Boker Tachtit, Is- mains unchallenged.44 gional differentiation without specia- rael, and Ksar Akil, , between Recently, Rogers and Jorde chal- tion and the maintenance of grade 47 and 38 ky ag~.~~,~~The Aurignacian lenged the notion that greatest diver- similarities worldwide. seems to have spread rapidly through sity equates with greatest age and The “Out of Africa” model is more Europe as seen at 43 kya in provided an explanation in terms of recent. It is based on fossil evidence (Bacho Kir~)~~and at 40 kya in paleodemography.45 They c onc 1u d e for an earlier appearance of modern (EAbreda, El Castil10.~~~~Associated that the mtDNA diversity patterns re- humans in Africa than elsewhere. with this geographical expansion flect the fact that Africa held a larger Howells proposed the idea of a single process of Aurignacian peoples population than other regions and recent origin as the “Noah’s Ark through Europe, the terminal Nean- throughout the period. Taking these model. This hypothesis has been derthal industries, including the demographic parameters into ac- elaborated in the last few years by sev- Chatelperronean, Szelettian, and U1- count, Harpending finds, through eral researcher~.~-I~This model high- luzzian, have been interpreted as the pairwise comparisons (mismatch dis- lights the discontinuity in the fossil result of an acculturation process.33~3~ tributions), that there is evidence of a record, suggesting a recent localized To the fossil and archeological re- leading wave signal in African sam- origin in Africa, followed by geo- cord, the molecular evidence should ples, suggesting that Africa may have graphical expansion and replacement be added. In 1987,Cann and co-work- been the source of dispersal of mod- of archaic populations. ers obtained a phylogenetic tree based ern humans.46Althoughit is clear that on mtDNA in which one branch led the mitochondria1 data cannot be in- THE EVIDENCE FOR THE ORIGINS solely to Africans and the other terpreted as tightly as the original OF MODERN HUMANS branch to Afncans and other popula- “Eve hypothesis” proposed, the nu- In recent years, the application of tions.36.Thistree reflected the fact that clear DNA evidence is increasingly ro- new dating techniques like electron Africans present the greatest diversity bust and the genetic evidence, overall, spin resonance and thermolumines- of mtDNA lineages. This could be ex- strongly supports a recent African ori- cence to Upper Pleistocene fossils has plained either by a faster rate of mu- gin of modern people.47 had a revolutionary effect on late tations in Africans than in the rest of The evolutionary interpretations of hominid chronology.11-’3These tech- the world or by a longer time during the genetic evidence have been in such niques, which date beyond the range which mutational differences accu- extraordinary agreement with the of I4C, have proven three significant mulated in this group. Given the lack “Out of Africa” model that the two points: that hominids with a modem of evidence from other sources of a concepts have been interlaced. It is morphology occurred in the Middle faster mutation rate in Africans, the often mistakenly assumed that the East around 100 ky ago;l4-l6that rela- authors deduced that the evolution of main evidence for a single origin of tively gracile moderns lived in Afnca modern humans started in Africa and modern humans is genetic and not around 70 ky and that Neander- then expanded to the rest of the world. morphological or chronological. The thal remains in Europe and the Mid- Since these results were first pub- conclusion that the available evidence dle East date to 60 to 40 ky ago, lished, further research has both rein- strongly supports a recent, single Afri- postdating early modern forms.I1J8 forced and undermined this can origin of modern people must fol- Furthermore, the remains from conclusion. On the one hand, research low if one takes into account the Klasies River Mouth (KRM) in South on the mtDNA diversity of localized following points: the earliest modern Africa, presenting a variable but mod- groups like the San and Papuans has people are found in Africa or the Mid- ern morph~logy,’~are firmly associ- thrown light on the levels of diversifi- dle East some 60,000years before they ated with early last interglacial levels, cation of recent people.36~37In addi- appear in other regions;s different and therefore are 120/100 ky old.20,21 tion, mounting nuclear genetic hominids overlap in time and space in This new chronology has also af- evidence points strongly to an African the , Europe, and, prob- fected the archeological record. Now origin of all modern gr0ups.~8~0On ably, East Asia;’ the archeological evi- it is known that technologies based on the other hand, Maddison41and Tem- dence in Europe points to a distinct the production of blades appear rela- plet~n~~have shown that the statisti- replacement of local traditions;34.48 tively late in the record, some 60,000 cal procedures for rooting the mtDNA the nuclear and mtDNA evidence indi- years after the appearance of morpho- phylogenetic tree and the statistical cates an African ancestry of all mod- logically modern people, although significance of the single African ern humans;38 and there is an there is some evidence of more com- branch were incorrect. These techni- apparently strong correlation be- plex behavior in a Middle Stone Age cal problems are a major drawback in tween recent linguistic differentiation 50 EvolutionaryAnthropology ARTICLES and genetic differentiati0n.4~In spite traits actually occurred in the geo- a multiregional model of modern hu- of all this, however, if morphological graphical area conventionally associ- man origins. continuity from regional archaic ated with them. In other words, the Two other lines of evidence related hominids to modern regional popula- regional traits fail to characterize East to the mechanisms of multiregional tions can indeed be observed in the Asians and, in Australians, do so only evolution also refute the basis of the fossil record, there will still be evi- in terms of robusti~ity.~~,~~Further- multiregional model. One, a survey of dence that multiregional evolution more, this work showed that the devel- the fossil evidence of subspeciation in took place. opment of characteristics such as animals with a wide geographical dis- pronounced tori and ridges occurs re- tribution, has shown that multire- gional evolution as a mechanism is TESTING THE MODELS gardless of geographical region among both modern and prehistoric popula- undocumented except, possibly, in the tions with large cranial and dental di- Javanese rhinoceros, and that the The Multiregional Hypothesis mensions. Another finding, that the common pattern observed is one of in- The basis for the Multiregional development of a number of facial fea- terspecific or intraspecific replace- model is that we can observe unique tures depends on the presence of large ment.54Second, recent assessments by regional patterns of morphological supraorbital ridges,5*,S3is particularly Harpending and co-workers of the continuity across the transition from ar- relevant in the light of Habgoods con- demographic density of Homo erectiis chaic to modern forms of hominids.4~~clusions regarding the combined oc- populations suggest that these popula- This interpretation of continuity makes tions never achieved the critical size to two assumptions: that such features maintain the levels of gene flow neces- are indeed regional markers; and that sary for multiregional evolution to oc- they are not functionally determined. cur.45f46 Three recent studies have inde- Besides the chronology of the fossils pendently tackled the problem of re- A consistent mechanism and the genetic evidence for a single gional morphological contin~ity.~~-~~for the world expansion origin of modern people, these studies Two of these dealt with the regional of modern humans from show that the multiregional hypothe- distribution of features identified as sis is based on incorrect premises of presenting continuity through time their original African morphological continuity and demo- among worldwide archaic popula- source has not been graphic patterns. In addition, the re- tions, and therefore, tested their proposed. In this regard, cent dating of some of the H. erectus uniquely Asian and Southeast Asian fossils of Javas6 takes the original ex- character in the pa~t.~~,~’The authors the “Out of Africa” pansion of Homo erectus (and, accord- of both studies found that these fea- model lacks sufficient ing to the multiregional model, the tures were common in Homo erectus origin of regional differentiation) to and “archaic” Homo sapiens fossils specificity to account for around 1.8 MA. This early date is far throughout the world, but reached the regional patterns of outside of even the most generous con- somewhat different conclusions about modern human diversity fidence limits for the origin of modern their role in proving continuity. On the mtDNA lineages. basis of their plesiomorphic character, and the specifics of Groves concluded that these features both morphological The “Out of Africa” Model should not be used as evidence of phy- and behavioral logenetic relationship^.^^ On the other If the basis of the Multiregional hand, Habgood considered that the evolution over the model can be discredited, can the “Out combined occurrence of features like last 100yOOOyears. of Africa” model be taken as the best supraorbital tori and zygomaxillary explanation for the origins of modern tuberosities in Javanese H. erectus and humans? The lines of evidence in sup- Australian aborigines can be used as port of that model, in terms of the con- evidence of morphological continu- tinuity in form from archaic to ity.5’ In the third study, Lahr dealt with currence of features of robusticity, modern fossils in Africa,lothe discrep- the regional distribution of features of which Lahr showed instead to be cor- ancy between the dates in AfricanLe- continuity in recent populations and related with each other. The implica- vantine modern humans and their relationship to metrical parame- tion is that traits that have been elsewhere, and the genetic data, are all compatible. However, there still are ters of the skull, testing their uniquely claimed to show links between, for ex- Asian and Australian character in the problems with this hypothesis. Those ample, Javanese Homo erectus and present and the independence of that problems relate less to contradictoiy Australian aborigines, are, in fact, evi- character from developmental proc- evidence than to the lack of specificity esses.52 Lahr found that of the thirty dence of the link between modern Aus- in both the model and the data. The so-called Asiatic regional traits, tralians and a robust modern ancestor model has three components, which, twenty-one did indeed have a signifi- anywhere in the world. The only con- to some extent, can be treated inde- cant incidence in a particular region of clusion to be drawn is that the mor- pendently: a single origin in Africa; a the world. However, only ten of these phological evidence does not support pattern of total replacement involving ARTICLES Evolutionary Anthropology 51

EUROPE I MIDDLE I AFRICA SOUTHEAST , , EAST ASIA I EAST I : ASIA 1 I

Figure 1, Spatial and temporal distribution of important human fossils in the Upper Pleistocene, together with a representation of present regional cranial variation. Continuous lines represent prehistoric occupation: diagonal double lines represent taxonomic distinctions within a region. no admixture with other hominid THE EVIDENCE FOR THE ORIGINS subdivision of gene pools must have populations; and a mechanism of dis- OF HUMAN DIVERSITY occurred in order for populations to persal across the world. Although cur- have acquired and established their Human diversity refers to the bio- rent evidence supports a single differences. logical and technological differences African origin, the problem of replace- In order to investigate the origins of among modern populations today and ment versus admixture remains a ma- diversity, three points should be taken in the recent past. Although many peo- jor issue. Moreover, a consistent into account: first, that only scant evi- ple consider the biological diversity of mechanism for the world expansion of dence is available for the first part of present-day humans to be vast, ge- the period second, that at the point at modern humans from their original netic studies show that it is very lim- which there is a record of modern peo- African source has not been proposed. ited when compared to that of In this regard, the “Out of Africa” ple in various parts of the world, these chimpanzees .56-59 Modern hum ans model lacks sufficient specificity to are in fact an extraordinarily homoge- populations are already different from account for the regional patterns of neous species.45Nevertheless, differ- each other; and third, that diversity in- modern human diversity and the spe- ences in their morphology, genetics, creases with time, i.e., people become cifics of both morphological and be- and archeology are apparent for as increasingly more different between havioral evolution over the last long as there is evidence of modern wide geographical distances. There- 100,000 years. people. It is the evolutionary origin of fore, two sets of evidence are impor- These problems highlight the need these differences from a recent com- tant, the patterns of diversity among to develop more precise ideas about mon ancestor that we seek to explore. the first occupants of each region for the origins of diversity. A theory of It is clear, and must be stressed, that which there are records and the sub- modern human origins has to be able there has been interbreeding between sequent pattern of differentiation of to explain not only the appearance of modern groups throughout the pe- each of these groups. Moreover, any modern people, but the origin of the riod, and that gene flow was one theoretical model that attempts to ex- diversity of modern populations. mechanism of change, although some plain the evolutionary process that 52 Evolutionary Anthropology ARTICLES created this diversity has to encom- North Africa/Middle East (Skhul, of Middle Stone Age traditions until pass a mechanism that would explain Qaf~eh),6~and South Afri~a.~~,~~In around 40 ky, after which date the area how modern people appeared in vari- North Africa, the early moderns may seems to have been essentially unin- ous regions, how these peoples ac- have remained restricted to the area habited until the quired their early differentiation, and for a long period. There is no evidence around 20 ky.48 Genetic mtDNA and how they expanded and differentiated supporting further expansion from Y-chromosome studies have shown to produce modern levels of diversity. the Middle East at this stage. In addi- that the San and Pygmies are distinct, Clearly, the available data are not suf- tion, biogeographic data show a move- which suggests their early differentia- ficient to answer all the relevant ques- ment of Palearctic in the tion from other pop~lations.36,~~,~5 tions, especially those pertaining to Levant between Stages 5 and 4 (70-60 The main characteristic of these differentiation. However, enough evi- k~).6~It is possible that descendants of groups was a trend toward extreme dence has accumulated from diverse this early modern population devel- gracilization. It is possible that this disciplines and groups of scholars to oped the Aterian Middle Palaeolithic small body size adaptation allowed give us insight into how this process in North Africa, which shows later colonization of the rainforests of might have occurred. certain derived characteristic^.^^,^^ western Africa. The close linguistic and genetic affinities of east and The Modern Regional southern African hunter-gatherer Populations population^^^ may be the result of sub- sequent movements. Within Africa, The point of origin of modern hu- and superimposed on the early popu- mans within Africa is unclear, but a It is not yet possible to lation patterns, are the recent expan- case can be made for East Africa on determine whether the sions of farming communities, the grounds that the earliest transi- speakers of the Niger-Kordofanian tional forms (Omo) are found there.'O Southeast Asian languages, and from western Asia to Furthermore, East Africa, with its di- population around 40 ky North Africa, speakers of the Afro-Asi- verse habitat and potential for isola- ago is the result of atic languages. tion, has been a major source of evolutionary novelty.60Geographical long-term differentiation expansion probably led to the early of people present in the Australia and Island Melanesia differentiation of populations, as is area for 20 to 30 ky, The first record of people in Austra- suggested by the diversity of Middle lia is archeological. Roberts and col- Stone Age traditions.61,62The early represents part of the leagues recently obtained artifacts modern fossils of Skhul and Qafzeh in widespread expansion showing a relatively unsophisticated the Middle East should be considered stone tool technology based on flakes as an extension of North Afncan popu- of peoples from North dated to around 60 ky.78.79 This tech- lations, for they are accompanied by Africa or the Middle East nology, together with elaborate bone African faunas, at least in Qaf~eh,~~ after 45,000 years ago, tools and art, characterized most of and do not seem to have expanded any the subsequent Australian prehis- further at ths time.64Outside Africa, or a mixture of both. The tory.80Around 40 ky ago, the number modern humans appeared at different evidence suggests that and geographical distribution of ar- regions at different times,I3 first in cheological sites in Australia in- Australia and Asia, then later in Europe. the first hypothesis is creased sharply. At this moment, it is Although this may be partly an artifact correct. not known whether there was con- of a poor fossil record, some of these tinuous occupation from 60 ky with a temporal patterns have remained rela- demographic expansion around 40 ky, tively stable as new dates have been or whether the large number of sites obtained in the last few years. at 40 ky reflects a flow of people into The evidence from different re- There is ~limatic,7~,~'faunal, and ar- Australia from outside. The archeo- gions of the world after the appear- che~logical~~evidence of biogeographic logical data suggest that Australian ance of modern humans is highly movements across the Sahara around populations remained relatively iso- variable, both spatially and tempo- 50 ky ago. The lack of substantial fossil lated until very recent times, when mi- rally, but vast (Fig. 1). We do not pre- material from this time in this area crolithic tools and the dingo were sume to cover it here, but we will precludes interpretations of the effect introduced.80 The first fossil evidence outline certain chronological, mor- of gene flow from sub-Saharan Africa in Australia was found in the south- phological, and archeological aspects on North African populations, which east, at the sites of Mungo and the Wil- that are relevant to the process of di- eventually developed into the robust landra Lakes system.81f82These fossils versification of peoples. Mesolithic groups of Afalou, , show remarkable variation in relation and Mechta.73 In sub-Saharan Africa, to the level of robusticity they pre- Africa the evidence after the first modern fos- sent, the crania ranging from very About 100 ky ago, there were mod- sils, (Omo, KRM, Border ) is scant. gracile to more robusticity than most ern humans in East Africa (Omo)? In , there is continuity modern ones, past or present.83 They ARTICLES Evolutionary Anthropology 53

also show morphological features that It is not known how the Southeast . However, nuclear genetic stud- clearly link them to recent Australian Asian population represented by fos- ies by Ca~alli-Sforza~~~show that aborigine~.8~>8~Most craniometric sils like Niah and Tabon relates to the Northeast Asian (Japanese, Korean, studies show the Australo-Melanesian earliest Southeast Asian population Mongolian) and Amerindian popula- population as that most closely related from which Australians and New tions are closer to Eurasians than to to African Therefore, a Guineans derive. It is not yet possible southern Chinese and Southeast pattern of early colonization followed to determine whether the Southeast Asians. A possible explanation for by relative isolation may be repre- Asian population around 40 ky ago is these differences is that the nuclear ge- sented in Melanesia by the level of dif- the result of long-term differentiation netic patterns reflect invasions into ferences in mtDNA lineages among of people present in the area for 20 to northeastern Asia by Siberian peoples Papuan tribes36 and of the compara- 30 ky, represents part of the wide- of Eurasian origin, either in the late tively high incidence of plesiomorphic spread expansion of peoples from Pleistocene, as indicated by archeologi- arche~logical~~and skeletalj2 traits. North Africa or the Middle East after cal evidence,”J-lIo or recently by peo- 45,000 years ago, or a mixture of both. ples speaking the Altaic languages.77 There is also the question of Amer- Eastern Asia indian affinities. If the Americas were In eastern Asia, the archeological colonized early (40 to 20 kya111-113), and paleontological record for the the first inhabitants would have been first half of the Upper Pleistocene is derived from the less specialized Asian very poor. The first fossil evidence of We propose a model to populations, whereas if they were modern people in Southeast Asia explain the diversity and colonized late (15 to 10 kya114), the dates to around 40 ky ago (Niah, Wad- disparity of the first inhabitants would have been de- jak, Tab~n).~OThe Southeast Asian rived from relatively specialized East population of the late Pleistocene and paleoanthropological Asian groups. Evidence of a more ro- its modern descendants can be identi- data in the Upper bust and less specialized Asian mor- fied craniallyg1and dentally (Sun- Pleistocene based on the phology within the Americas has been dad~nts).~*The range of this suggested for Holocene remains1lj,116 population reached beyond Southeast concept of geographical and for marginal native American Asia into southern , (then expansions and groups (Fueguians, Patagonians).Il7 connected to the mainland), and even- On the other hand, Turner’s finding tually P01ynesia.~~~~~In eastern Asia, dispersals. This model is that all Amerindian remains present a the best known early modern fossils based on a single-origin homogeneous Sinodont dental pat- are those of Upper Cave. hypothesis, followed by tern96supports a recent migration into These fossils were first described as the New World (although his sample showing such variation as to resemble multiple dispersals out of does not include southernmost South three different populations, Chinese, Africa through two rather American groups). This hypothesis Eskimo, and Melanesian.95 Recent independent routes, then has been supported by genetic and lin- northeastern Asians show a derived guistic data. l4 However, other genetic morphology, represented by the subsequent expansions evidence shows that the timing and “Mongoloid features of facial flatness and dispersals from source of Amerindian differentiation and Sinodon ty.939496-98Turner has is still unclear.’18 recognized Sinodont characteristics secondary geographical in the Upper Cave Zhoukoudian re- sources. main~.~~However, other authors have Northern, Southern, and found that these fossils are not closely Western Asia related to recent Asians,Ioowhich im- This area is not a unit in bio- plies a relatively late appearance of the geographical terms. The area of the The evidence suggests that the first hy- typical Mongoloid morphology. Holo- Middle East was at times an extension pothesis is correct: current Southeast cene human remains with Mongoloid of African faunal distributions and at Asian populations are closer, both ge- features are found in a wide area, from other times an extension of Eurasian Eastern Siberia (where Turner identi- neti~ally~~and archeologically,89 to ones.64The population history of the fied a sharp east-west boundary in the Australo-Melanesians than either is to Middle East is complex. After the first incidence of Sinodonty in the region Eurasians and Indians. Turner has appearance of modern humans in the of Lake Baika199),, Korea, proposed that East Asians derive from last interglacial, the area was occupied Japan, and the Americas. In Japan, the a southeast Asian Cra- by an archaic population with Euro- prehistoric Jomonese and recent nia1,84.85.91,96-98,’01-103 dental,92 archeo- pean affinities (Kebara, Ainu, who show S~ndadonty~~and are logical,89 and mtDNA104evidence Amud).lI9 Technologically, there is no cranially isolated from recent Japa- further suggest that Southeast, East, clear distinction between the early nese and Chinese,91,96,97must be survi- and Northeast Asians are closely re- modern and archaic populations; both vors of the population before the lated, presenting a tight Mongoloid are accompanied by Middle Paleo- Sinodont expansion. complex independent of Eurasia and lithic industries. Between 50-40 ky, 54 Evolutionary Anthropology ARTICLES

Multiple Dispersals as a Model for the Origins of Human Diversity These patterns of differentiation suggest that modern populations changed at different rates during the w Upper Pleistocene, depending on lev- i els of gene flow and demographic pressures, and that geographical disper- : sal and expansion were main compo- nents of the process of differentiation. The "Out of Africa" model implies that it is dispersal beyond Africa that is wandering, irruption or migration critical. This ignores two things: the ttleneck first is that because Afnca itself is more than one third of the habitable Old World, dispersals within Africa 4- 4- GEOGRAPHIC RANGE -b are equally important; the second is that dispersal and divergence within Figure 2 Schematic representation of the geographical expansion of a species through time. Africa would lead to variable popula- showing localized origins. unsuccessful events, successful dispersal events, population isolation through the development of barners, and superimposition of later dispersal events on the range tions leaving Africa at different times of previous expansions Adapted from Tchernov 130 and possibly by different routes. This means that the levels of differentiation in the populations colonizing the the first technologies recognized as the technology is Chatelperr~nian,'~~other continents were probably al- early are found at a industry with ele- ready high. Boker Tachtit and Ksar Akil, the latter ments similar to the Upper Paleolithic. Beyond the original routes out of associated with a modern human Once considered as evidence of tech- Africa, the data also point to a com- skeleton.27J8 The subsequent pattern nological continuity between Nean- plex process of population differentia- in the Middle East is one of continuity, derthals and modern humans in tion involving the incomplete from a robust modern Upper Paleo- Europe, the Chatelperronian, being superimposition of dispersing popu- lithic population (Qafzeh I and 11, contemporaneous with the earliest lations on previously existing ones. In Ohalo), to the Natufian pre-agricultu- Aurignacian sites, is currently inter- morphological terms, the temporal ralists, to the first Neolithic farmers. preted as resulting from accultura- and spatial variation in the presence The paleoanthropology of the last tion.33J4Similar interpretations apply of common features suggests that 100 ky in the Indian subcontinent is to industries like the Szelettian of modern humans differentiated, ac- virtually unknown. The first evidence Eastern Europe or the Ulluzzian of It- quiring and losing traits in a stepwise of modern humans is found relatively aly. The first evidence of modern hu- manner, reducing at each step the late (28 ky) in ,zO and most mans in Europe is mainly communality of modern cranial probably is not representative of the archeological, and is related to the traits?' The archeological remains first population that occupied the spread of Aurignacian sites. The ma- also show uneven development of area. This skeleton is accompanied by jority of cases of Upper Paleolithic fos- technologies during the Upper Pleis- a stone tool industry similar to that of sils date to later periods and are tocene and the appearance of techno- the European Upper Paleolithic. accompanied by subsequent industries. logical innovations at particular In Northern Asian and Siberia, the These populations were robust and had temporal and geographic points, sug- archeological record indicates a date larger crar~ial2~than more recent peo- gesting a generally localized process of first occupation between 35 and 20 ple, and, in many cases, were distinct of differentiation. Further, the genetic ky by modern people manufacturing from recent Europeans.126,127 data available from contemporaneous Upper Paleolithic-like stone tool~.~*J~~Superimposed on these Paleolithic populations show that degrees of dif- populations are the dispersals of agri- ferentiation and admixture vary Europe culture-related peoples in the early markedly, some populations being the During most of the Upper Pleisto- Holocene.128Modern Europeans are result of very recent expansions and cene, Europe was occupied by a Nean- cranially the most homogeneous of re- differentiation and others being the derthal population. The last known gional human p0pulations.5~If the result of previous expansions that in Europe are those of model proposed by Renfrew is correct, may or may not have come into recent St. Cesaire in (36 ky)122and Za- and these agriculturalists brought contact with other populations.4" farraya in Spain.I23 European Nean- with them a branch of the Indo-Euro- Therefore, the evolutionary origin of derthals were typically associated pean family of language^,'^ then the present populations may be extraordi- with ind~stries.~~In St. Basques and Lapps must represent SUT- narily varied. At this point, however, Cesaire and other similarly late sites, viving Palaeolithic groups.12* the evidence is only tentative. In east- ARTICLES Evolutionary Anthropology 55

em Asian and Australia, for example, we may be sampling three populations that resulted from three distinct dis- persals. A population like the Austra- lians may be the descendant of a group who left Africa between 100 and 60 ky, and hence present a high degree of morphological continuity with early and more robust moderns. A popula- tion like the Southeast Asians may be the result of long-term tropical differ- entiation in the area and geographical expansion. Differing amounts of gene W 3 - Late Upper Pleistocene 15 - 0 kyr flow, together with the effects of the db break-up of the Sunda land mass, could account for the varying levels of distinctness of groups like the An- daman Islanders, Philippinos, and others. A population like the present east Asians may be the result of a rela- tively recent adaptation, having un- dergone its own geographical expansion that included the Americas. Therefore, it is possible to see differen- tiation followed by population growth and dispersal as the mechanism of ex- pansion of modern humans out of Africa and as the mechanism of development of subsequent regional populations. 2 - Mid Upper Pleistocene: 50 - 15 kyr

THE MULTIPLE DISPERSALS MODEL We have seen that the morphologi- cal basis of the multiregional model is incorrect. That fact, together with the chronological, archeological, and ge- netic evidence, indicates that the alter- native single-origin explanation is more compatible with the available data. However, we have also seen that although the “Out of Africa” hypothe- sis explains the origins of modern hu- IJ mans from an archaic source, the 1 - Early Upper Pleistocene: 100 - 50 kyr model lacks a mechanism to explain ydb v the origins of human diversity. We propose a model to explain the diver- sity and disparity of the paleoanthro- Figure 3. Two views of the pattern of modern human divergence and dispersal according to the pological data in the Upper multiple dispersals model. (a) Possible Pleistocene dispersals: 1) In the early Upper Pleistocene Pleistocene based on the concept of (100-50 ky) there are dispersals within and out of Africa of early robust forms of modern humans; geographical expansions and disper- 2) The mid-Upper Pleistocene (5G15 ky) is the scene of world glaciations and the dispersals within southeast Asia and of Eurasian Upper Palaeolithic popuiations; and 3)In the late Upper Pleistocene sals. This model is based on a single-ori- (15-0 ky), we see the recent dispersals, some associated with agricultural expansions, ihat have gin hypothesis, followed by multiple been superimposed on the Palaeolithic human distribution. dispersals out of Africa through two rather independent routes, then sub- sequent expansions and dispersals from ary novelty tends to occur in small ar- cal ones involving humans and other secondary geographical sources. eas and that successful populations species. l3l-m The process begins with can expand explosively, or at least rela- dispersals into new regions and if A Mechanism: Dispersals and tively rapidly, from these centers of colonization is successful, the avail- Evolutionary Change origins 129,130 (Fig. 2). This seems to be able habitats of the colonized region Biogeographical comparisons sug- the case with well-documented pa- would be occupied. Range expansion gest that the appearance of evolution- leontological events, as well as histori- then takes place. Through time, this 56 Evolutionary Anthropology ARTICLES

ing populations) to complete, either directly, through factors like disease, or indirectly, through the competition for resources and differential repro- duction or demography. If such dis- persal events occurred frequently during the later Pleistocene and were the primary mechanisms by which the 0 human population diversified, then the mosaic of modern human diversity can be seen as the product of several 10,000 events of differing geographical extent occurring over 100,000 years. This process would account for the pattern 20,000 of modern human variability as a con- sequence of differential ancestral morphology occurring in successive 30,000 dispersals followed by local adapta- tion. According to this view, Africa is unique only as the original source of 40,000 .. <:.. .: i :...... ,: populations, whereas the diversifica- .. : ; ___...... '. t I..- tion process involved varied geo- j _.....' graphical foci, both African and 50,000 non-African. Where modern popula- tions were already present, dispersals 60,000 .. would have acted as a primary mecha- .,.. : ..... L--- ...... : ._...... nism of gene flow...... -. ,...... - ...... --,___...... -.. ..._... .-. 70,000 ,,.." ...' southern route A Pattern: Routes of Dispersal and levels of Differentiation

80,000 Traditionally, it has been assumed that expansions out of Africa occurred through the narrow corridor of north- 90,000 eastern Africa and the Middle East, a lorthern route northern route. This assumption car- ries two implications. First, the direc- 1oo,ooa tion of movement was across a major desert towards the Middle East, an ?rsals n Africa area that certainly was populated at the time. Therefore, movement would have been strongly constrained by cli- matic conditions and competition with other hominids. Second, that there was morphological and genetic unity in the expanding population at Figure 3. (b) The implied time depth of the main populations of living Homo sapiens. any one time. However, another route of dispersal, through the Horn of Af- would be followed by differentiation thernselve~,*3~they can be spectacu- rica towards the Arabian Peninsula, a as the population breaks up due to the larly successful. The mechanisms in- southern route, has been used by ani- mals in the past. Use of this route by appearance of barriers. Further dis- volved in the displacement of local expanding early moderns has also persals would be superimposed on this populations by invading groups may been suggested 49.60 (Fig. 3). pattern, resulting in a complex pal- be diverse, ranging from adaptive su- How would the existence of two dif- impsest of relic and recent popula- periority, disease, habitat disturbance ferent routes out of Afnca affect the tions. Range expansion, migrations, and destruction, and differential re- subsequent diversity of modern hu- dispersals, colonizations, and differ- productive rates.'34 The processes of mans? Three main aspects of the ex- ential survival of populations are the displacement ensuing from the expan- pansion process would be affected. norms of evolutionary biogeogra- sion of modern human populations First, the climatic conditions neces- ~hy.*30-'3~Furthermore, even if only a may have ranged from nonexistent (to- sary for northward expansion are very small proportion of invaders establish tal interbreeding of local and incom- strict-faunal dispersals across the ARTICLES Evolutionary Anthropology 57

Sahara occurred during short epi- tions and the relatively early dates for cant magnitude. This conclusion is sodes of fast deglaciation, during the occupation of Australia. further supported by recent research which wet conditions prevailed in Is the hypothesis of differentiation by Waddle using matrix correlation of most of northern Africa. These strict in Africa of populations ancestral to Eurafrican fossils.142 climatic constraints would not have regional groups consistent with the Another way of explaining the origi- acted on populations expanding east- available evidence? One way of inter- nal diversity of regional populations is ward toward Asia from East Africa. preting the differences between re- that these populations were already Second, the circum-Mediterranean gional modern populations when they different at the time they left Africa. In area was certainly occupied by homi- appear in the archeological record is this case, a certain amount of prior nid populations during the last 200 ky. that these differences, in fact, reflect differentiation and population subdi- Hence, any movement into this region various levels of admixture between vision would have taken place prior to would imply competition with other modern people and local archaic expansion. There is little fossil data to groups. Although hominids were pre- populations. Although it is commonly support or refute this idea. Between sent throughout southern and eastern stated that there are two models of 130 and 60 ky ago in Africa, there are Asia in the late Pleistocene, the sizes modern human origins, others have fossils from south, east, and north Af- and densities of populations may have been proposed, such as the Afro-Euro- rica, but these are very few, localized, been highly variable. The third and, pean Sapiens Hypothe~isl~8J3~and the and early The groups in these three perhaps, the most important aspect is Assimilation Hypothe~is.l~~,l~~Both of regions do show marked population that as hominids took different routes these hypotheses are largely con- differences, and indicate an even ear- at different times, it is likely that the cerned with the level of genetic admix- lier date for the first appearance of a African source populations also dif- ture between dispersing African common morphology. They are also fered from each other. populations and indigenous popula- technologically diverse, although all How does the available data fit the tions, especially European Neander- present what are called Middle Stone hypothesis of different routes of ex- thals. They differ, however, in the Age or Middle Paleolithic stone tools. pansion? A northern faunal route of amount of gene flow considered to Therefore, both the fossil and archeo- dispersal into the Nile Valley and have taken place The Afro-European logical evidence tentatively suggests across northern Africa has been inter- Sapiens model sees archaic genes per- African diversification at this time. mittently used since the Miocene, and sisting in a modern gene pool, whereas However, genetic evidence obtained the was alterna- the Assimilation Hypothesis sees by Harpending and co-workers tively occupied by Afro-Arabian or modern genes imposed on an archaic strongly supports the idea of diversifi- cation of populations prior to geo- Palearctic elements." As Tchern0v6~3" gene pool. The difference is therefore graphical expansion, a model they has argued, the paleoanthropological one of relative contribution. However, have called the Weak Garden of evidence suggests that an early mod- the absence of consistently transi- Eden.45,46,143Their interpretation of ern population took an inland route to tional fossils throughout the world10 mtDNA diversity through pairwise north Africa and the Middle East dur- argues against gene flow as the main distributions indicates that modern ing the hypsithermal phase of the last process in the geographical expansion humans underwent a bottleneck intergla~ial.~OJ3~-l~~However, this of modern morphology, malng the around 100 ky, that this bottleneck population, represented by the Skhul Assimilation Hypothesis as unlikely as was followed by population subdivi- and Qafzeh fossils, seems to have the multiregional model. Although sion and relative isolation of these di- faced a competitive or geographical some hybridization between modern verse modern groups, and that barrier to further expansion at that humans and archaic populations demographic geographical expansion stage. The main expansion of modern might have occurred, it is not clear occurred between 70 and 40 ky, the humans out of Africa through the Le- that the morphological features con- youngest being that of Europeans. vantine corridor occurred around 45 sidered to reflect continuity between Therefore, molecular data show that a ky,48 but this event postdates the first Neanderthals and early Upper Paleo- large amount of population differen- occupation in A~stralia.~~,~~Further- lithic Europeans (generally measures tiation occurred prior to expansion. more, the morphological and archeo- of robusticity, as is the case for the logical features of Middle Eastern and claimed similarities between Austra- European population of 40 ky ago pre- lians and Homo erectus), are relevant CONCLUSIONS cludes them from the ancestry of phylogenetic markers. Furthermore, Although we believe that a recent many Australian fossils. If, however, the mtDNA data show that no widely single origin is the evolutionarily cor- another population, already geneti- divergent lineage that could be attrib- rect explanation for the appearance of cally separated from other modern uted to Neanderthal descent has ever Homo sapiens, we suggest that the groups, expanded out of East Africa been sampled. The possibility of find- "Out of Africa" model does not explain toward the Indian subcontinent at any ing such a lineage remains, at least the temporal and geographical pat- time between 100 and 50 kya, this theoretically, until all people have terns of diversification observed in the could explain some of the marked dif- been sampled, but its absence in the Upper Pleistocene. The theoretical ex- ferences in morphology and techno- thousands of individuals studied so far planation of human diversity that best logical traditions at the point of the indicates that even if such interbreed- accounts for data on the dates of first first appearance of regional popula- ing took place, it did not have signifi- occupation, morphological variation, 58 Evolutionary Anthropology ARTICLES and technological innovations in- version of it. Thanks also to R. Klein 16 Stringer CB, Grun R, Schwarcz HP, Gold- berg P (1989) ESR dates for the hominid bur- volves multiple dispersal events. These and six anonymous referees for their ial site of Skhul in . Nature338:756-758, dispersals would have taken different reviews. All errors that remain are our 17 Grun R, Beaumont PB, Stringer CB (1990) routes out of Africa, as well as differ- own. Special thanks to B. Arensburg, ESR dating evidence for early modern hu- mans at in South Africa. Nature ent routes and directions from other H. Harpending, R. Haydenblit, J. 334:537-539. subsequent non-African sources. The Nichols, and K. Robson Brown for 18 Valladas H, Joron JL, Valladas G, varying order and geographical extent stimulating discussions on the subject Arensburg B, Bar-Yosef 0, Belfer-Cohen A, Goldberg P, Laville H, Meignen L, Rak Y, of the early dispersals from Africa are of modern human differentiation and Tchernov E, Tillier AM, Vandermeersch B proposed to explain the different to F. Lahr for his help with the figures. (1987) Thermoluminescence dates for the Ne- times at which various regions of the Financial support to MML was pro- anderthal burial site at Kebara in Israel. Na- ture 330:159-160. world were first occupied. The incom- vided by CNPq, (), the Leaky 19 Rightmire P, Deacon HJ (1991) Compara- plete geographical extent of sub- Trust, and the CARE Foundation. tive studies of late Pleistocene human remains sequent dispersals is proposed to from Klasies River Mouth, South Africa. J REFERENCES Hum Evol20:131-156. explain the persistence of relic popu- 20 Deacon HJ (1989) Late Pleistocene pa- lations as new populations were su- 1 Weidenreich F (1943) The skull of Sinan- laeoecology and archaeology in the southern perimposed on the range of earlier thropus pekinensis: A comparative study of a Cape, South Africa. In Mellars P, Stringer CB primitive hominid skull. Palarontol Sinica N. (eds) The Human Revolution, pp 547-564. Ed- groups. Finally, high levels of isola- S., No. 10. inburgh: Edinburgh University Press. tion, especially in tropical areas, are 2 Coon CS (1962) The Origin of Races. New 21 Singer R, Wymer J (1982) The Middle Stone proposed to explain the differing rates York: Knopf. Age at Klasies River Mouth in Soutk Africa. Chi- 3 Thorne A, Wolpoff MH (1981) Regional con- cago: University of Chicago Press. of morphological, genetic, and tech- tinuity in Australasian Pleistocene hominid 22 Brooks AS, Yellen JE (1989) An archae- nological change through time be- evolution. Am J Phys Anthropol55:337-349. ological perspective on the African origins 01 tween long-standing adaptations and 4 Wolpoff MH, Wu XZ, Thorne A (1984) Mod- modernhumans. Am JPhysAnthropol78:197. ern Homo sapiens origins: A general theory of 23 Clark JD (1982) The transition from Lower very recent colonizations. hominid evolution involving the fossil evi- to Middle Palaeolithc in the African continent. Modern humans originated from a dence from East Asia. In Smith FH, Spencer In Ronen A (ed) The Transition from Lower to recent single evolutionary event, F (eds),The OriginofModem Humans:A World Middle Palaeolithic and the Origin of Modem Surveyofthe FossilEvidence, pp411483. New Man, pp 235-255. Oxford: BAR Int Series 15 1. whereas modern human diversity is York: Alan R. Liss. 24 Deacon HJ, Schuurman R (1992) The ori- the result of multiple evolutionary 5 Frayer DW, Wolpoff MH, Thorne AG, Smith gins of modern people: The evidence from events brought about by multiple geo- FH, Pope GG (1993) Theories of modern hu- Klasies River. In Brauer G, Smith FH (eds) man origins: The palaeontological test. Am Continuity or Replacement? Controversies in graphical dispersals. Much remains to Anthropol 95: 14-50. Homo sapiens Evolution, pp 121-129. Rotter- be done to document these various 6 Howells WW (1976) Explaining modern dam: Balkema. dispersals and their morphological, man: Evolutionists versus migrationists. J 25 Gowlett JAJ (1987) The coming of modern Hum Evol 5:477495. man. Antiquity 61:210-219. archeological, and genetic correlates. 7 Stringer CB (1989) The origin of modern hu- 26 McBurney CBM (1 967) and the However, we also need to investigate mans: A comparison of the European andnon- Stone Age of the Southeast Mediterranean. further the possibility that one princi- European evidence. In Mellars P, Stringer CB Cambridge: Cambridge University Press. (eds) The Human Revolution, pp 232-244. Ed- 27 Marks AE (1990) The Middle and Upper Pa- pal difference between modern hu- inburgh: Edinburgh University Press. laeolithic oftheNeareast and theNi1eValley:The mans and archaic hominids lay not in 8 Stringer CB (1993) Reconstructing recent problem of cultural transformations. In Mel- any major cognitive difference or any human evolution. In Aitken MJ, Stringer CB, lars P (ed) The Emergence ofModern Humans, single edge of advantage, but simply in Mellars PA (eds) The Origin ofModem Humans pp 56-80. Ithaca: Cornell University Press. and the Impact of Chronometric Dating. pp 28 Bergman CA, Stringer CB (1989) Fifty the social and ecological potential of 179-195. Princeton: Princeton University years after: Egbert, an early Upper Palaeo- modern humans to disperse at higher Press. lithic juvenile from Ksar Akil, Lebanon. Paleo- 9 Stringer CB, Hublin JJ, Vandermeersch B rient 15:99-11 1. rates than could archaic hominids. (1984) The origin of anatomically modern hu- 29 Smith FH (1984) Fossil hominids from the We are not naive or optimistic mans in Western Europe. In Smith FH, Upper Pleistocene of central Europe and the enough to suppose that this paper will Spencer F (eds) The Origin ofModern Humans: origin of modern Europeans. In Smith FH, A World Survey of the Fossil Evidence, pp 51- Spencer F (eds) The Origin ofModem Humans; end the debate between the multire- 135. New York: Alan R. Liss. A World Survey of the Fossil Evidence, pp 137- gional and “Out of Africa” theories. We 10Stringer CB,Andrews P(1988) Geneticand 209. New York: Alan R. Liss. hope nonetheless, that by more fully fossil evidence for the origin of modem hu- 30 Bischoff JL, Soler N, Maroto J, Julia R mans. Science 239: 1263-1268. (1 989) Abrupt Mousterian/Aurignacian developing the mechanisms for the 11 Grun R, Stringer CB (1 99 1) Electron spin boundary at c 40 ka bp: Accelerator 14C dates evolution of modern human diversity resonance dating and the evolution of modern from 1’Abreda Cave (Catalunia, Spain). J and by emphasizing the importance of humans. Archaeometry 33:153-199. Archeol Sci 16:563-576. 12 Aitken MJ, Stringer CB, Mellars PA (eds) 31 Cabrera V, Bischoff JL (1989) Accelerator pre-existing variability at the origins (1993) The Origin of Modem Humans and the 14C dates for early Upper Palaeolithic (Basal of modern humans, albeit within Af- Impact of Chronometric Dating. Princeton: Aurignacian) at El Castillo Cave (Spain). J rica, rather than worldwide, we have Princeton University Press. Archeol Sci 16577-584. 13 Foley RA, Lahr MM (1992) Beyond “Out of 32 Straus LG (1 989) Age of the modern Euro- at least brought something new to this Africa”: Reassessing the origins of Homo sapi- peans. Nature 342:476477. discussion. ens. J Hum Evol22:523-529. 33 Mellars P (1991) Cognitive changes and 14 Valladas H, Reyss JL, Valladas G, Bar-Yosef the emergence of modern humans in Europe. 0, Vandermeersch B (1988) Thermolumines- J Cambridge Archeol 1:63-76. ACKNOWLEDGMENTS cence dating of Mousterian “Proto-Cro-Mag- 34 Mellars P (1993) Archaeology and the non” remains from Israel and the origin of population-dispersal hypothesis of modern We thank J. Fleagle for the invita- modern man. Nature 331:614-616. human origins in Europe. In Aitken MJ, tion to write this paper, as well as C. 15 Schwarcz HP, Grun R, Vandermeersch B, Stringer CB, Mellars PA (eds) The Origin of Bar-Yosef 0, Valladas H, Tchernov E (1988) Modern Humansand the ImpactofChronomet- Gamble, W.W. Howells, F. Lahr, and C. ESR dates for the hominid burial site of ric Dating, pp 196-216. Princeton: Princeton Stringer for commenting on an earlier Qafzeh in Israel. J Hum Evol 17:733-737. University Press. ARTKLES Evolutionary Anthropology 59

35 Cann RL, Stoneking M, Wilson AC (I 987) AG, Suprijo A, Widiasmoro (1994) Age of the 73 Brauer G, Rimbach KW (1990) Late ar- Mitochondrial DNA and human evolution. earliest known hominids in Java, . chaic and modern Homo sapiens from Nature 325:31-36. Science 263: 11 18-1 121. Europe, Africa and southwest Asia: Cra- 36 Vigilant L, Stoneking M, Harpending H, 56 Li WH, Sadler LA (1991) Low nucleotide niometric comparisons and phylogenetic im- Hawkes K, Wilson A (1989) African popula- diversity in man. Genetics 129513-523. plications. J Hum Evol 19:789-807. tions and the evolution of human mtDNA. Sci- 57 Ferris S, Brown W, Davidson W, Wilson A 74 Lucotte G (1992) African pygmies have the ence 253:1503-1507. (1981) Extensive polymorphism in the mito- more ancestral gene pool in studies for Y-chro- 37 Stoneking M, Jorde LB, Bhatia K, Wilson chondrial DNAof apes. Proc Nat Acad Sci USA mosome DNA haplotypes. In Brauer G, Smith A1 (1990) Geographic variation in human mi- 78:63 19-6323. FH (eds) Continuity or Replacement? Contro- tochondrial DNA from . 58 Kocher T, WilsonA (1991) Sequenceevolu- versies in Homo sapiens Evolution, pp 75-81. Genetics 124:717-733. tion of mitochondrial DNA in humans and Rotterdam: Balkema. 38 Wainscoat JS, Hill AVS, Boyce AL, Flint J, chimpanzees: Control region and a protein- 75 Stine OC, Dover GH, Zhu D, Smith KD Hernandez M, Thein SL, Old JM, Lynch JR, coding region. In Osawa s,Honjo T (eds) Evo- (1992) The evolution of two west African Falusi AG, Weatherall DJ, Clegg JB (1986) lution of Life: Fossils, Molecules and Culture, populations. J Mol Evol34:336-344. Evolutionary relationships of human popula- pp 391413. New York Springer-Verlag. 76 Ruhlen M (1990) An overview of genetic tions from an analysis of nuclear DNA poly- 59 Wilson A, Cann R. Carr S, George M, classification. In Hawkins JA, Gell-Mann M morphisms. Nature 319: 491-493. Gyllensten U, Helm-Bychowski K, Higuchi R, (eds) The Evolution of Human Languages, pp 39 Mountain J, Lin AA, Bowcock M, Cavalli- Palumbi S, Prager E, Sage R, Stoneking M 1-27. Addison-Wesley. Sforza LL (1993) Evolution of modern hu- (1985) Mitochondrial DNA and two perspec- 77 Renfrew C (1991) Before Babel: Specula- mans: Evidence from nuclear DNA tives on evolutionary genetics. Biol J Linnean tion on the origins of linguistic diversity. Cam- polymorphisms. In Aitken MJ, Stringer CB, Soc 26:375-400. bridge Archeol J 1:3-23. Mellars PA (eds) The Origin of Modern Hu- 60 Kingdon J (1993) Self-Made Man and His 78 Roberts RG, Jones R, Smith MA (1990) mans and the Impact of Chronometric Dating. Undoing. London: Simon & Schuster. Thermoluminescence dating of a 50,000 year pp 69-83. Princeton: Princeton University 61 Masao FT (1992) The Middle Stone Age old human occupation site in northern Aus- Press. with reference to . In Brauer G, tralia. Nature 345:153-156. 40 Tishkoff SA, Kidd KK (1994) Evidence for Smith FH (eds) Continuity or Replacement? 79 Roberts RG, Spooner N (1992) Lumines- stability of a highly variable short tandem re- Controversies in Homo sapiens Evolution, pp cence dating of early occupation sites in peat polymorphism during recent human evo- 99-109. Rotterdam: Balkema. northern Australia. Paper presented at the lution. Am J Phys Anthropol Suppl 18:195. 62 Clark JD (1993) African and Asian perspec- conference “Australia Day” University of Cam- 41 Templeton AR (1 992) Humans origins and tive on the origins of modern humans. In Ait- bridge. analysis of mitochondrial DNA sequences. ken MJ, Stringer CB, Mellars P (eds) The 80 Kirk RL (1981) Aboriginal Man Adapting. Science 255:737. Origin of Modem Humans and the Impact of Oxford: Oxford University Press. 42 Maddison DR (1991) African origin of hu- Chronometric Data, pp 201-215. Princeton: Princeton University Press. 81 Thome AG (1977) Separation or reconcili- man mitochondrial DNA reexamined. Syst ation? Biological clues to the development of Zoo1 40:355-363. 63 Tchernov E (1992) Biochronology, palem- Australian society. In Allen J (ed) Sunda and 43 Wolpoff MH, Thorne A (1991) The case cology and dispersal events of hominids in the Sahul: Prehistoric Studies in Southeast Asia, against Eve. New Scientist, 22 June:37-4 1. southern Levant. In Akazawa T, Aoki K, Melanesia, and Australia, pp 187-204. London: 44 Stoneking M, Sherry ST, Redd AJ, Vigilant Kimura T (eds) The Evolution and Dispersal of Academic Press. L (1993) New approaches to dating suggest a Modem Humans in Asia, pp 149-188. Tokyo: Hokusen-sha. 82 Thome AG (1 984) Australia’s human ori- recent age for the human mtDNA ancestor. In gins: How many sources? Am J Phys Anthro- Aitken MJ, Stringer CB, Mellars PA (eds) The 64 Tchernov E (1992) Eurasian-African biotic pol 63:227. Origin of Modern Humans and the Impact of exchanges through the Levantine corridor during the Neogene and Quaternary. Cour 83 Habgood PJ (1985) The origin of the Aus- Chronometric Dating, pp 167-175. Princeton: tralian Aborigines: An alternative approach Princeton University Press. Fors Senck 153:103-123. 65 Vandermeersch B (1989) The evolution of and view. In Tobias PV (ed) Human Evolution: 45 Rogers AR, Jorde LB (n.d.) Genetic Evi- Past, Present and Future, pp 367-380. New dence on Modern Human Origins. modern humans: Recent evidence from Southwest Asia. In Mellars P, Stringer CB York: Alan R. Liss. 46 Harpending H, Sherry ST, Rogers AR, 84 Pietrusewsky M (1984) Metric and Non-Met- Stoneking M (1993) The genetic structure of (eds) The HumanRevolution, pp 155-164. Ed- inburgh: Edinburgh University Press. tic Cranial Variation in Australian Populations ancient human populations. Curr Anthropol Compared with Populations fiom the Pacific and 34:483-496. 66 Deacon HJ (1993) Southern Africa and modem human origins. In Aitken MJ, Stringer Asia. Canberra: Australian Inst. Aboriginal 47 Stoneking M (1993) DNA and recent hu- Studies. man evolution. Evol Anthropol2:60-73. CB, Mellars P (eds) The Origin of Modern Hu- mans and the Impact of Chronometric Dating. 85 Howells WW (1989) Skull shapes and the 48 Klein RG (1992) The archaeology of mod- pp 177-1 83. Princeton: Princeton University map. Papers Peabody Museum, Arch Ethnol, em human origins. Evol Anthropol 1:5-14. Press. vol79, Cambridge: Harvard University Press. 49 Cavalli-Sforza LL, Piazza A, Menozzi P, 67 Rightmire GP, Deacon HJ (1991) Compara- 86 Hedges SB, Kumar S, Tamura K, Stoneking Mountain J (1988) Reconstruction of human tive studies of late Pleistocene human remains M (1991) Comment on “Human origins and evolution: Bringing together genetic, archae- from Klasies River Mouth, South Africa. J analysis of mitochondrial DNA sequences” by ological and linguistic data. Proc Nat Acad Sci Hum Evol20:131-156. JA Templeton. Science 255:737-739. 85:6002-6006. 68 Close AE, Wendorf F (1990) North Africa 87 Vigilant L, Stoneking M, Harpending H, 50 Groves CP (1989) A regional approach to at 18,000 BP. In Gamble C, Soffer 0 (eds) The Hawkes K, Wilson A (1991) African popula- the problem of the origin of modern humans World at 18,000 Be Vol.2, pp 41-57. London: tions and the evolution of human mitochon- in Australasia. In Mellars P, Stringer CB (eds) Unwin. drial DNA. Science 253:1503-1507. The Human Revolution, pp 274-285. Edin- 69 Debenath A, Raynal JP, Roche J, Texier JP, 88 Wright RVS (1992) Correlation between burgh: Edinburgh University Press. Ferembach D (1986) Stratigraphie, habitat, cranial form and geography in Homo sapiens: 5 1 Habgood PJ (1 989) The origin of anatomi- typologie et devenir de l’Aterian marocain: CRANID. A computer program for forensic cally modem humans in Australasia. In Mel- Donnes recentes. L‘Anthropol90:233-246. and other applications. Perspect Hum Biol lars P, Stringer CB (eds) The Human 70 Petit-Mane N (1989) Interglacial environ- 2/Arch Oceania 27:128-134. Revolution, pp 245-273. Edinburgh: Edin- ments in presently hyperarid Sahara: Paleocli- 89 Pope GG (1988) Recent advances in Far burgh University Press. matic implications. In Leinen M, Sarntheim M Eastern palaeoanthropology. Ann Rev Anthro- 52 Lahr MM (1992) The origins of modem (eds) Paleoclimatology and Paleometeorology: pol 17:43-77. humans: A test of the Multiregional hypothe- Modern and Past Patterns of Global Atmospheric 90 Jones R (1 989) East of Wallace’s Line: Is- sis. PhD dissertation, University of Cam- Transport,pp 637-661. London: KluwerAcPress. sues and problems in the colonization of the bridge. 71 Fontes JC, Gasse F (1989) On the ages of Australian continent. In Mellars P, Stringer CB 53 Lahr MM (1994) The Multiregional Model humid and late Pleistocene phases in North (eds) The Human Revolution, pp 743-782. Ed- of modem human origins: A reassessment of Africa: Remarks on “Late Quaternary climatic inburgh: Edinburgh University Press. its morphological basis. J Hum Evol26:23-56. reconstruction for the Maghreb (North Af- 91 Brace CL, Brace ML, Dodo Y, Leonard WR, 54 Groves CP (1992) How old are subspecies? rica)’’ by P Rognon. Pal Pal Pal 70:393-398. Li Y, Shao XQ, Sangvichien S, Zhang Z (1990) A tiger’s eye-view of human evolution. Perjpect 72 Wendorf E Schild R, Siad R, Haynes CV, Gau- Micronesians, Asians, Thais and relations: A Hum Biol2/ Archaeol Oceania 27:153-160. tier A, Kobusiewicz M (1976) The prehistory of craniofacial and odontometric perspective. 55 Swisher CC 111, Curtis GH, Jacob T, Getty the Egyptian Sahara. Science 193:103-114. Micronesia, Suppl2:323-348. 60 Evolutionary Anthropology ARTICLES

92 Turner CG I1 (1987) Late Pleistocene and 108 Jia L, Huang W (1985) The late Palaeo- Upper Palaeolithic crania. Fysisch-Antropol Holocene population history of East Asian lithic of China. In Wu R, Olsen JW (eds) Pa- Mededelingen 2:7-15. based on dental variation. Am J Phys Anthro- laeoanthropologyand Palaeolithic Archaeology 127 van Vark GN, Bilsborough A, Henke W pol 73:305-321. in the Peoplek Republic of China, pp 21 1-223. (1992) Affinities of European Upper Palaeo- 93 Turner CG I1 (1990) The major features of Orlando: Academic Press. lithic Homo sapiens and later human evolu- Sundadonty and Sinodonty, including sugges- 109 Olsen JW (1987) Recent developments in tion. J Hum Evol 23:40 1417. tions about East Asian microevolution, popu- the Upper Pleistocene prehistory of China. In 128 Cavalli-Sforza LL (1991) Genes, people lation history and late Pleistocene Soffer 0 (ed) The Pleistocene Old World: Re- and languages. Sci Am, November. relationships with Australian Aborigines. Am gional Perspectives, pp 135-146. New York: J Phys Anthropol82:295-317. Plenum. 129 Foley RA (1989) The ecological condi- 94 Turner CG I1 (1990) Origin and affinity of 110 Reynolds TEG, Kaner SC (1990) Japan tions of speciation: A comparative approach the people of : A dental anthropological and Korea at 18,000 BP. In Soffer 0, Gamble to the origins of anatomically modern hu- mans. In Mellars P, Stringer CB (eds) assessment. Micronesia, Suppl2:403416. C (eds) The World at 18,000 BF: Vol 1,pp 296- The Hu- 31 1. London: Unwin Hyman. man Revolution, pp 298-320. Edinburgh: 95 Weidenreich F (1939) On the earliest rep- Edinburgh University Press. resentatives of modern mankind recovered on 111 Dillehay TD, Collins MB (1988) Early cul- the soil of East Asia. Peking Nat Hist Bull tural evidence from MonteVerde in . Na- 130 Tchernov E (1992) Dispersal: A sugges- 13:205-224. ture 332:150-152. tion for a common usage of this term. Cour 96 Brace CL, Brace ML, Leonard WR (1989) 112 Gruhn R (1 987) On the settlement of the Fors Senck 153:21-25. Reflections on the face of Japan: A multivari- Americas: South American evidence for an ex- 131 Groves RH, Burdon JJ (1986) Ecology of ate and odontometric perspective. Am J Phys panded time frame. Curr Anthropol 28:363- Biological Invasions, Cambridge: Cambridge Anthropol 78:93-113. 365. University Press. 97 Howells WW (1 984) Prehistoric human re- 113 Delibrias G, Guidon N (1986) L‘abri Toca 132 Gray AJ, Crawley MJ, Edwards PJ (eds) mains from China. In Schwidetzky I (ed) Ras- do Boqueirao da sitio da Pedra Furada. L‘An- (1987) Colonization, Succession and Stability. sengeschichte der Menschheit, pp 29-38. thropol 90:307-3 16. Oxford: Blackwell Scientific. Munich: R.O. Verlag. 114 Greenberg J, Turner CG 11, Zegura S 133 Hengeveld R (1989) Dynamics of Biologi- 98 Howells WW (1986) Physical anthropology (1986) The settlement of the Americas: A com- cal Invarions. London: Chapman & Hall. of the prehistoric Japanese. In Pearson RJ (ed) parison of the linguistics, dental and genetic Windows on the Japanese Past: Studies in Ar- evidence. Curr Anthropol27:477497. 134 Lodge DM (1993) Biological invasions: Lessons for ecology. TREE 8:133-137. chaeology and Prehistory,pp 85-99. Ann Arbor: 115 Steele DG, Powell JF (1992) Peopling of Michigan Center for Japanese Studies. the Americas: Paleobiological evidence. Hum 135 Causse C, Conrad G, Fontes JC, Gasse F, 99 Turner CG I1 (1985) The dental search for Biol 64:303-336. Gibert E, Kassir A (1988) Ledernier“Humide” Native American origins. In Kirk R, Szath- 116 Neves WA, Pucciarelli HM (1991) Mor- Pleistocene du Sahara nord-occidental dat- mary E (eds) Out ofAsia: Peopling ofthe Ameui- phological affinities of the first Americans: An erait de 80-100,000 ans. C R Acad Sci Paris, t cas and the Pacific, pp 31-78. Canberra: exploratory analysis based on early South 306, Serie II:1459-1464. Journal of Pacific History. American human remains. J Hum Evol 136 Causse C, Coque R, Fontes JC, Gasse F, 100 Kamminga J, Wright RVS (1989) The Up- 21 :261-273. Gibert E, Ben Ouezdou H, Zouari K (1989) per Cave at Zhoukoudian and the origins of 117 Lahr MM (1993) Patterns of modern hu- Two high levels of continental waters in the the Mongoloids. J Hum Evol 17:739-767. man diversification: Implications for Amerin- southern Tunisian chotts at about 90 and 150 101 Hanihara K (1985) Origins and affinities of dian origins. Paper presented at the XI11 ka. Geology 17:922-925. the Japanese as viewed from cranial measure- ICAEJ, , Aug. 1993. 137 Gaven C, Hillaire-Marcel C, Petit-Marie N ments.InKirkR, SzathmaryE(eds)OutofA.sia. 118 Szathmary EJE (1993) Genetics of abo- (1981) A Pleistocene lacustrine episode in Pwplingofthe Americas and the Pacific,pp 105- riginal North Americans. Evol Anthropol southeastern . Nature 290:131-133. 112. Canberra: Journal of Pacific History. 1:202-220. 138 Brauer G (1984) The “Afro-European sa- 102 Pietrusewsky M (1990) Craniofacial vari- 119 Bar-Yosef 0 (1993) The role of western piens hypothesis” and hominid evolution in ation in Australasian and Pacific populations. Asian in modern human origins. In Aitken MJ, East Asia during the late Middle and Upper Am J Phys Anthropol82:319-340. Stringer CB, Mellars PA (eds) The Origin of Pleistocene. Cour Fors Senck 69:145-165. 103 Pietrusewsky M, Li Y, Shao X, Nguyen QQ Modern Humans and the Impact ofChronomet- 139 Brauer G (1992) L‘hypotheseAfricainede (1992) Modern and near modern populations ric Dating, pp 132-147. Princeton: Princeton of Asia and the Pacific: A multivariate cra- l’origine des hommes modernes. In Hublin JJ, University Press. Tillier AM (eds) AmOrigines d’Homo sapiens, niometric interpretation. In Akazawa T, Aoki 120 Kennedy KAR, Deraniyagala SU (1989) K, Kimura T (eds) The Evolution and Dkpersal pp 181-215. Paris: Presses UniversitC de Fossil remains of 28,000 year old hominids France. ofModern Humans in Asia, pp 53 1-558. Tokyo: from Sri Lanka. Curr Anthropol30:394-399. Hokusen-sha. 140 Smith FH (1992) The role of continuity in 121 Morlan RE (1987) The Pleistocene ar- 104 Ballinger SW, Schurr TG, Torroni A, Can modern human origins. In Brauer G, Smith chaeology of Beringia. In Nitecki MH, Nitecki FH (eds) Continuity or Replacement? Contro- YY, Hodge JA, Hassan K, Chen KH, Wallace DV (eds) The Evolution ofHuman Hunting, pp DC (1992) Southeast Asian mitochondria1 versies in Homo sapiens Evolution, pp 145- 267-307. New York: Plenum. DNA analysis reveals genetic continuity of an- 156. Rotterdam: Balkema. 122 Mercier N, Valladas H, Joron JL, Reyss JL, cient Mongoloid migrations. Genetics 141 Smith FH, Simek JF, Harrill MS (1989) 130:139-152. Leveque F, Vandermeersch B (1991) Ther- moluminescence dating of the late Neander- Geographical variation in supraorbital torus 105 Cavalli-Sforza LL, Menozzi P, Piazza A thal remains from Saint-Cesaire. Nature reduction during the later Pleistocene (c (1992) Genetic history and geography of Asia. 351 :737-739. 80,000-1 5,000 B.P.). In Mellars P, Stringer CB In Akazawa T, Aoki K, KimuraT (eds) The Evo- (eds) The Human Revolution. pp 172-193. Ed- lution and Dispersal of Modern Humans in 123 Hublin JJ (pers. comm.) inburgh: Edinburgh University Press. Asia, pp 61 3-623. Tokyo: Hokusen-sha Pub1 124 Leveque F, Vandermeersch B (1981) Le co. neanderthalien de Saint-Cesaire. Recherche 142 Waddle DM (1994) Matrix correlation 12:242-244. tests support a single origin for modern hu- 106 Bellwood P (1990) From late Pleistocene mans. Nature 368:452454. to early Holocene in Sundaland. In Gamble G, 125 Kidder JH, Jantz RL, Smith FH (1992) Soffer 0 (eds) The World at 18,000 BP, Vol2, Defining modern humans: A multivariate ap- 143 Sherry ST, Rogers AR, Harpending H, pp 155-163. London: Unwin Hyman. proach. In Brauer G, Smith FH (eds) Conrinu- Soodyall H, Jenkins T, Stoneking M (n.d.) Mis- ity or Replacement? Controversies in Homo match distributions of mtDNA reveal recent 107 Chen C, Olsen JW (1990) China at the last human population expansions. Hum Biol. in glacial maximum. In Soffer 0, Gamble C (eds) sapiens Evolution, pp 157-1 77. Rotterdam: Balkema. press. The Worldat 18,000 Be Vol 1~ pp 276-295. Lon- don: Unwin Hyman. 126 van Vark GN (1 990) A study of European 0 1994 Wiley-Liss, Inc.