Edited by JENSLORENZFRANZEN

Forschungsinstitut Senckenberg Frankfurt am Main

Frankfurt am Main 1994

~ Courier Fo!'o ungs-InstitutSenckenberg, 171: 341-361; Frt

The Case for Sinking Homo erectus. 100 Years of Pithecanthropus is Enough!

Milford H. Wolpoff, Ann Arbor (U.S.A.) Alan G. Thome, Canberra (Australia) Jan Jelinek, Bmo (Czech Republic) Zhang Yinyun, Beijing (peoplesRepublic of China)

Abstract Some half-century ago, Homo erectuscame to replace a variety of geographicallydistinguished genera, including "Pithecanthropus","Sinanthropus", "Meganthropus","Atlanthropus", in order to clarify the evolutionary processby burying a nomenclaturethat was obscuring it. With the discoveriesand theoretical advancesof recent decades,the continued use of Homo erectushas now becomea major impedimentto Wtderstandingthe Pleistoceneevolution of humans. Moreover,there is good reasonto regard the taxon itself as invalid.

We proposehere to mergeHomo erectuswithin the evolutionary speciesHomo sapiens.The origin of Homo erectus lies in a cladogenicevent at least 2.0 myr ago. We view the subsequentlineage as culturally and I.!hysicallyadapted to an in- creasinglybroad range of ecologies,ultimately leading to its spreadacross the old world prior to the beginning of the Middle Pleistocene.Homo erectusdiffers from Homo habilis in a numberof ways. The vast majority of these distinctions also charac- terize Homo sapiens.The few distinctions of Homo sapiensthat are not sharedwith Homo erectusappear to be responsesto, or reflections of, continuing evolutionary trends of increasing cultural complexity, increasing brain size, and the progressive substitutionofteclmology for biology.

Homo erectusis a polytypic species,divided into severaldistinct geographicvariants which each show at least some geneticcontinuity with the geographicvariants of the polytypic speciesHomo sapiensthat is reflected in sharedunique combi- nations of morphologicalfeatures. There is no distinct boundarybetween Homo erectusand Homo sapiens in time or space, and cladogenesisdoes not seemto mark the origin of Homo sapiens.Instead, the characteristicsof Homo erectusand Homo sapiensare found to be mixed in seeminglytransitional samplesfrom the later Middle Pleistoceneof every region where there are human remains. The regional ancestryof Homo sapienspopulations makes monophyly impossible for the speciesif the earlier populationsare in a different species.We interpretthis to mean that there is no speciationinvolved in the emergenceof Homo sapiens from Homo erectus.These reasonscombine to require that the lineage be regardedas a single evolutionary species.

Kurzfassung Vcr ungefahr einem halben Jahrhundertist der BegrifTHomo erectusan die Stelle einer ganzenReihe geographisch unterschiedenerGattungen, wie "Pithecanthropus","Sinanthropus", "Meganthropus", "Atlanthropus", getreten,urn eine No- menklatur zu ersetzen,welche die stattgefundenenEvolutionsprozesse bis dahin eher verbarg. Infolge der Entdeckungenund theoretischenFortschritte der letzten Jahrzehnteist der fortgesetzteGebrauch des BegrifTesHomo erectus nunmehr seIber zu einemHindernis geworden,wenn es darum geht, die Evolution des Menschenim Pleistozlin zu verstehen.Mehr noch,es gibt mittlerweile gute GrtInde,dieses Taxon tiberhauptals ungultig zu betrachten. Daher schlagenwir vcr, Homo erectusin der evolvierendenArt Homo sapiens aufgehenzu lassen. Der Ursprung desHomo erectusliegt in einem AbspaltungsprozeB,der mindestens2 Millionen JahrezurUckliegt. Wir betrachtendie darausher- vorgegangeneEntwicklungslinie als kulturell wie physischan eine zunehmendbreite Spanne van Umweltbedingungenange- paBt,was schonvcr Beginn des Mittel-Pleistol1ins zu ihrer Ausbreitunguber die Alte Welt filhrte. Homo erectusunterscheidet sich van Homo habilis in vieltaltiger Weise. Die Mehrzahl dieserUnterschiede charakterisiert auch den Homo sapiens.Die wenigenUnterschiede, die der Homo sapiensnicht mit clemHomo erectusteilt, erscheinenwie Reaktionenauf Entwicklungs- tendenzenin Richtung auf zunehmendekulturelle Vielfalt, zunehmendeGroBe des Gehirns und fortschreitendenErsatz yon Biologie durch Technologie. Homo erectusist eine polyt}1'ischeArt, aufgesplittertin mehreregeographische Varianten, yon denenjede zumindest einige genetischeKontinuitiit mit entsprechendengeographischen Varianten des ebenfalls polyt}1'ischenHomo sapiensauf- weist. Dies drtickt sich in einzigartigen gemeinsamenMerkrnalskombinationen aus. Es gibt keine deutliche Grenze, weder zeitlich noch rliumlich, zwischenHomo erectusund Homo sapiens.Der Ursprung des Homo sapiens scheint nicht auf einen AbspaltungsprozeBzurUckzugehen. Statt dessenstellt man in jeder Region mit Funden aus der Zeit des vermutlichenOber- gangsim splitenMittelpleistozlin fest, daBsich charakteristischeZuge desHomo erectusmit solchendes Homo sapiensvermi- scheu. Die regionale Abstammungdes Homo sapiensstilnde im Widerspruchzur Annahme seiner monophyletischenEntste- hung,wilrde man die fIilheren Populationeneiner anderenArt zuordnen.WiT interpretierendies in clemSinne, daBdie Entste- hung des Homo sapiensaus clemHomo erectusmit keiner Speziationverbunden war. Das fllhrt zu der Forderung,diese Ent- vlicklungslinie als eine einzige evolvierendeArt anzusehen.

Introduction More than 50 years after DUBOISdiscovered the menclaturethat was obscuring it (MAYR, 1950). Per- first true "missing link", and 40 years after SCHWALBE haps ironically, with the discoveries and theoretical (1899) provided the first comparative analysis of the advancesof the decadessince, the continued use of Trinil skullcap, workers suchas LEGRos CLARK(1940) Homo erectus may now have becomethe major im- argued that a variety of similar forms be subsumedin pedimentto understandingthe Pleistoceneevolution of "Pithecanthropus". Ten years later the variation was humans.There are good reasonsto considerthe taxon described at an even lower taxonomic level as Homo invalid. We propose here to merge Homo erectus erectus replaced "Pithecanthropus" in order to clarify within the evolutionaryspecies Homo sapiens. the evolutionary processby burying a taxonomic no-

Understanding Polytypism Perhaps the most important impetus for exam- Australians is to be found elsewhere, and not all of the ining the question of whether Homo erectus and Homo descendants of the Javans live in Australia.sapiens ~ .. should be regarded as different species, derives W .11 t t d th O I t . < from an appreClatilDn..., of the problems Involved In EIDENREICH ...,. I us ra e IS evo u lonary pat- . ~ '" . t t. th I t. f lytyp ' . W ' th tern as a trellis, WIth main vertical lines representing .. In erpre mg e evo u Ion 0 po IC species. I ... ~ ..the centers of evolution, and dIagonal connections ~ regard to the evolution of Pleistocene Homo, there are be ",veen th em re fl ec ti. ng th e pa tt ems 0 f gene t IC" mter- two hlstonc. sources for these problems. FIrst, there IS h Thi Id .t t . htfi d b 't, f . h h b . d b ...c ange. s wou seem quI e s ralg orwar, ut '. a senes 0 questions , t at ave een raise y .er cnticlsm aft hi s dea th the theory was nuSIn.. t erpre t e d ..m an I 1 - C; of WEIDENREICH s (1943, 1946) polycentric theory; I t ti. th t d d th rt. II ' 1"- .,.. us ra on a repro uce e ve Ica Ines 0 f d escent actually, questions of HOWELLS nusrendenng of the b t did t ta. th d. II ' C ..u no re In e lagona Ines 0 f gene exc h ange- C'; theory. d. Second, there are problems Involved m under- H ' (1959 ' 236) " c d I b " th Thi h I . f I .. th OWELLS. an e a ra eory. s was "\ stan mg t e evo ution 0 po ytyplC species at stem t"'; b t d t W d h t . a UJ u e 0 EIDENREICH an c arac enze d as th e f'- from the acceptance of species...eory defimtions that simply th .In w hi c: h " H uman evo I u t .Ion progresse d .In eac h tii do not allow past polytyplC species to have existed. f th Id ( t. II art) l: corner 0 e wor , essenla y ap ilom w hat was ~ Because pf his (at that time) unique experience happening in the other comers" (p, 234), ~ with the human evolutionary record for three conti- . t WElDEN REICH h d . t . fth tt HOWELLS asked of this theory, how one could ~ nen s, a a presclen vIew 0 e pa em ..' .. f h I t . H t d d tht h I explain the Independent evolution of four major races: 0 uman evo u Ion. e con en e a uman evo u- " .."WEIDENREICH has at least four different evol,wg tion was best understood as a network of Interconnected h .. I '. l: . uman vaneties, Ivmg lar apart, moving a h ea d b Y fi ItS populations that retaIned regIonal continuIty WIthin at d t rt d . th ' . I I '" . least some geographIc ..an areas, whIch he called centers s a s ' pro uClng elr own specla pecu Ian ties 0 f (h " I ...; " ) E h f th fi . I form ...Yet these four careers at last converged to pro- ence po ycenUJsm. ac 0 e our major evo u- d th ki d f h A d all . ...uce e same n 0 man eve ere. n nu- tionary centers he proposed retaIned differences at the ryw,. " . II I (1947) d Id be d . tl I t d t th raculously enough, breasted the tape at the same time racla eve , an cou Ifec y re a e 0 e 235 races of today. At the same time there was significant (p. ). gene flow between regions and the evolving species HOWELLSdiscounted and discredited the theory maintained its unity as a whole. For instance, in sum- because he could provide no answer as to how this marizing his understanding of Australasian evolution, could have happened. This is not surprising, in view of he wrote (1943: 249-250): "at least one line leads from the fact that the "four independent converging lines" Pithecanthropus and Homo s%ensis to the Australian were actually a consequence of HOWELLS'misinterpre- aborigines of today. This does not mean, of course, that tation of WEIDENREICH'Stheory. Yet, in spite of its I believe all the Australians of today can be traced back origin in a misrepresentation, the assumption that there to Pithecanthropus or that they are the sole descendants must be parallel independent lines in a regional of the Pithecanthropus -Homo s%ensis line." continuity view of human evolution has taken on a life of its own. For instance, it has become a centerpiece of In other words, while WEIDENREICHbelieved similar criticisms of the Multiregional Evolution con- that significant aspects of the skeletal morphology of cept (THORNEand WOLPOFF,1981; WOLPOFF,Wu, and Aboriginal Australians were to be found in their re- THORNE,1984) that builds on WEIDENREICH'spublica- gional antecedents, the Javan Homo erectus people, he tions -see KLEIN (1990), STRINGERand ANDREWS never regarded this as a unique ancestral-descendant (1988), BRACE (1991), ROUHANI (1989), FAGAN relation. He posited that some of the ancestry of the (1991), and LEWIN(1991b). This misrepresentation has its own intellectual history, independentof the evolu- consideration,the lack of evolutionary modeling for'polytypic tionary model it misquotes,and has becomeone of the species.This is surprising, in view of the sig- important sources of criticisms raised about explana- nificant number of living polytypic vertebrate species tions of Multiregional Evolution. mostof us are familiar with (MAYR,1969: 38-50), such Even given the correct interpretation of as bears,dogs, baboons, and other mammals including WEIDENREICH'Spolycentric trellis model, the question e. g. Homo sapiens. Other polytypic species include remains as to how the geographicdemes he envisioned birds of paradiseand pigeons,kingsnakes and pythons, for Homo erectuscould have remaineddistinct acrossa varanid and scincid lizards, Salamanders and tree speciesboundary, if Homo sapiensevolved out of this frogs. In contrast,polytypic fossil speciesare virtually ancestral species. WEIDENREICHhimself, as discussed unknown (VAN V ALEN, 1966), and the question we below, never regarded this as a serious problem since find interestingis why this is so. In particular, we are in his "Single Species Hypothesis" he regarded all concernedthat this might be an artifact of how species PleistoceneHomo (as known then) as Homosapiens. are defined in the fossil record, theoretically as well as However, WEIDENREICH'Shypothesis was not operationally [LEVINTONand SIMON, 1980; and see widely accepted, which brings us to a more general KIMBELand RAK (1992)J.

SpeciesDefinitions

It has been suggested that the virtual absence of extendedover time. The conceptprovides a clear link descriptions of polytypic species in the fossil record betweenthe biological speciesand chronospecies.In- could be an artifact of applying the criteria of mor- deed,BOCK (1986) regardsthe biological speciesas a phospecies (i.e. phylogenetic species) (WOLPOFF,Wu, paradigm that encompassesthe evolutionary species and THOR1~E,1984: 448; TuRNER and CHAMBERLAIN, concept,and we havelittle disagreementwith this since 1989). A number of authors have equated these, includ- the differencebetween a focus on gene flow vs. breed- ing most recently BOCK (1989: 54-55) who dismisses ing boundariesis one of emphasis.Both mechanisms both because "they fail to distinguish between the spe- playa significant role in partitioning and controlling cies and the phyletic lineage". Current usage of mor- geneticvariation within internally subdividedspecies. phospecies, or phylogenetic species, continue to define One weaknessof the biological speciesconcept polytypism out of existence on the species level. How- that is of particular concern to paleontologistsis the ever, we believe that the evolutionary species concept difficulty of translating the main criterion of the bio- of SIMPSON(1961), especially as developed by WILEY logical species-lack of actual or potential interbreed- (1978, 1981) provides a very productive alternative ap- ing -into usefulmorphological criteria. In contrast,the proach for examining the problems posed by evolutionaryspecies emphasizes a behavioral aspectof polytypism in the fossil record. WILEY (1978) defines a speciesthat is readily translated into morphological an evolutionary species as: "a single lineage [one or criteria: the evolutionary importance of behavioral more demes that share a common history of descent not mechanisms that promote species recognition shared by other demes] of ancestor-descendentpopula- (PAn"ERSON,1985): Concerningspecies, one aspectof If'" tions which maintains its identity from other such line- estimating whetherhvo forms have different identities s ages and which has its own evolutionary tendencies is to examinevarious mechanismsthat promote iden- If and historical fate." tity. The most important aspects for many species )- We find this a useful definition, and one which include recognition systemspermitting individuals to 1- corresponds closely to the species concept we have discriminate betweenother members of their species developed over the last decade to deal with the special and membersof other species(WILEY 1981: 25). problems engendered by the existence of evol\ing A corollary of this definition is that the identi- y polytypic species within Homo. fication of successivespecies along a single evolu- IS tionary lineage,i.e. phyletic speciation,is unnecessary Jf The evolutionary species concept by implication retains the essence ofMAYR'S (1963) biological species (WILEY,1978; BONDE,1981; BOCK,1986). The basis SOl for this assertionis of some interest in our considera- e- -reproductive isolation -while avoiding many of its deficiencies: lack of time-depth, absence of operational tions of the Homo erectus problem. We accepttwo of its're criteria, and perhaps most importantly emphasis on the reasonsproposed for setting this widely accepted reproductive ties (gene flow) as a major cohesive force preceptaside (WILEY, 1981: 39-41) as being particu- 'at (LEWOJl.'TIN,1974). The biological species concept does larly valid. Theseare: (1) the arbitrariness of phyletic .fe not pro\ide for an explanation of cohesion with species speciesrecognition (see SIMPSON,1961: 165), and (2) of the arbitrary mechanismsconsequentially thought to ,n- that is fundamental to the evolutionary species concept; promote phyletic speciation.Dismissal of the phyletic nd that is, the observation that reproductive boundaries are :a- a critical aspect of the main characteristic shared across speciationconcept, of course,does not imply rejection ,vs the populations of a species -its evolutionary pattern of phyletic evolution, or of gradualism as an explana- tion of it (contra GINGERICH,1985). Quite the opposite, ..\N (WILEY, 1981; WOLPOFF,Wu, and THORNE, 198~). las Evolutionary species can be seen as biological species as BOCK(1986, 1989) has pointed out, speciation(the 344

of phyletic lineages) results from the normal processes cord, that led him to proposea single specieshypothe- 'of phyletic evolution, under "the special boundary sis for all Pleistocenehumans (except australopithe- conditions of an external isolating barrier". RIGHTMIRE cines). (1990: 181-186) provides a very balanced discussion of this problem. Yet, he defends the phyletic speciation With many systematists,we accept the precept concept. He criticizes evolutionary species (p. 185) for that cladogenesisis the only way new species are encompassing too much variation ("indi-viduals as- formed,whether or not one of the daughterspecies can signed to one 'species' differ substantially more in size be separatedfrom the parental one, as discussedby and form than do representatives of living groups") WILEY(1981: 34- 35). In the fossil record, the absence while ironically it is WEIDENREICH'S(1943) observa- of evidence reflecting cladogenesisin the origin of tion of just the opposite, that the normal range of Homo sapiens from some or many populations of variation in living humans encompasses most of the Homo erectuswould support the hypothesis that only variation found in the hominid Pleistocene fossil re- one evolutionaryspecies is represented.

Polytypic Species: Defl out of Existence? As noted above, we have been concerned with Two consequencesof this extraordinary asser- the possibility that past polytypic species may have tion are: (1) that it should not be possibleto recognize been "defined out of existence" and suspect that their antecedentfonDS leading to existing human subgroups perceived absence from the evolutionary record might (i.e. today's races have no discernible ancestors),and (2) that either the majority of the living human sub- }. better reflect the ~pecies definition used than paleo- (: biological reality (see also TuRNER and CHAMBERLAIN, groups will soondisappear, or that they will all merge - 1989: 121). In response to this contention, it has been into one single panrnictic morpho ~: asserted that the absence of past polytypic species is not , It follows that past polytypic speciescannot be ""' an artifact of definitions, but instead that "a poor fossil uncovered.If they existed at all they were ephemeral, C record of polytypic species ...may... reveal something "t'C and one does not expect a fossil record of ephemeral important about the lack of stability of geographical organisms. When there "seems" to be evidence of a to populations across geological time" (KIMBEL, 1991: polytypic speciespersisting over time, it must be a con- 362). sequenceof misinterpretingthe fossils. We believe that The explanation offered for this supposition is in this mainly demonstratesthe validity of our initial the form of a conunent that is neither specific nor concern.If pastpolytypic speciesare not defined out of documented. According to (KIMBEL 1991: 362-363): existence,they are argued out of existence. "intraspecific geographic differentiation ...is largely What we understand of this argument is that ephemeral. The vast majority of local populations of since polytypic species cannot be recognized in the polytypic species either become extinct or merge with past, we are to assumethat they don't exist (or didn't parent populations as environments change and shift exist for any significant length of time.). As KIMBEL geographically over time and, as such, certainly could (1991: 363) puts it: "To treat morphologically diverse not be expected to leave behind much of a fossil record, clusters in the fossil record as subspeciesconfuses an let alone one with substantial time depth. Since signifi- arbitrary tool of taxonomywith evolutionaryunits." cant morphological diversification among early homi- nids could hardly have proceeded without the geo- But we contend instead that this stricture con- graphic isolation and subsequent expansion of differ- fuses evolutionary units by use of an arbitral)' taxon- entiated local populations, it is far more likely that the omy. We, in contrast,believe that it is the presentthat fossil record of this diversity reflects the successful should be a guide for interpreting the past and not visa establishment of new species." versa.

SpeciesRecognition -Its Role in Polytypic Species

A species mate recognition system may be criti- been less well developed. In internally subdivided cal in maintaining reproductive boundaries between social species with gene flow between demes, there-are species, especially at the time of speciation (PAT- multiple levels of cohesiveness. In these cases there is TERSON, 1985). Mate recognition is of particular some level of cohesiveness "ithin the demes them- importance among sibling species, or when an adaptive selves. This localized pattern "ithin species may reflect radiation brings closely related species into competition WRIGHT'S (1967) plateaus of stable adaptations, but (MAYR, 1963). By maintaining reproductive bounda- there are other factors which also act to maintain co- ries, it is \videly regarded as a mechanism that helps hesiveness in such demes. One of these which we be- maintain species cohesion. lieve is important in humans and other social species is An understanding of the role of the mate rec- the problem of differential interactions oriented by kin oenition in internally subdivided species, however, has recognition, an aspect of behavior that is a necessary ined requirement for any inclusive fitness considerations. most of the externallyvisible featuresshowing regional We proposethat certain mechanismsof mate recogni- continuityare found. tion, balancing learned behaviorand genetic predispo- sition to recognize certain resemblances,have evolved Despite the demic differences described above, to meetthis problem in polytypic specieswhere migra- no human geographic deme, today or in the past, shows tions and mate exchangesprovide a specialopportunity a set of unique evolutionary changes that would be for interactionswith unrelatedindividuals. expected to characterize a distinct evolutionary species. When taking the problem out of the human These demes are not ancestral-descendent populations, arena, where kinship and alliance are of significant each with their own distinct evolutionary tendencies importance in mate choice,phenetic similarities playa and historical fates. Unique evolutionary tendencies role for other species.The complex nature of this role and a historic fate characterize the species as a whole, derives from the fact that simple similarity would en- and not any subdivision of it. The unique evolutionary hance mating with sibs, or other close kin. How a bal- tendencies in all Pleistocene Homo populations, widely ance is reached is suggestedby a study of Japanese recognized for a number of decades, involve (among quail (BATESON,1982). These birds showa clear pref- other things) expansion of cranial capacity, reductions erence for first cousins, remaining in their proximity in the posterior dentition, and in skeletal robusticity. significantly more often than in proximity with birds These characterize the species as a whole and involve with other degreesof relationship (including sibs) and documented changes in every geographic region with a unrelatedbirds. In fact, the quails spentthe leasttime fossil record since the beginning of the Middle in the proximity of their sibs. Time spent in proximity Pleistocene. They are unique to no region. is directly related to mate choice in this species We believe that some of the unique features (BATEsoN,1978). , within eachlocal region, are important in kin identifi- The most strongly preferred matesare slightly cation. On the other hand, other skeletal features that different from individuals that are familiar from early show regional continuity clearly have no expression life (BATEsoN,1982). Recognition, in the quails, is that is readily observableduring life (incisor shovel- based on the plumage. In humans recognition of indi- ling, foramen ovale form, mylohyoid foramen form, viduals is mainly basedon features of the face and we nasal sill morphology). These probably better reflect do not believe it is a coincidencethat the face is where geneticcontinuities that are fully non-adaptive.

Geographic Speciesin Homo: Implications for Modem Variation Severalauthors have proposedthat Homo erec- clearly not applicableto the living biota since it fails to Ius is a composite taxon that, in reality, includes a admit the crucial distinction between differentiated c number of different geneticallydistinct lines, or species populations where the irreversible genetic event of 't (ANDREWS,1984; CLARKE,1990; KENNEDY,1991; speciationhas and has not intervened."Moreover, the L STRINGER,1984; TATTERSALL,1986; WOOD, 1984). conceptaddresses none of the issues important in the c These proposals certainly support our assessmentof biological speciesconcept, such as the question of how n persistentregional differencesin Homo, but we do not cohesivenessis established (for instance whether concur ,vith the specieslevel of taxonomic difference through? common adaptive plateau, homeostatic central to the proposals,as detailed below. In many mechanisms,or external boundariesto gene flow). In \- cases the basis for this interpretation lies in the general phylogenetic species do not conform to the It "phylogenetic species concept" (ELDREDGEand geneticnotion of a speciesas protectedgene pools. The a CRACRAFT,1980), which defines speciesas "a diag- concept,in all, placespaleontologists between the jaws nosable cluster of individuals within which there is a of a dilemma, To ignore it is to dismiss the most un- parental pattern of ancestryand descent.and beyond ambiguousand easily reproduciblemeans of defining which there is not" (p. 92). somethingcalled a pastspecies, while to employ it is to

Ph I ., d '" h d b th make comparisons of past '\lth living species into .v ogenetic species are IStInguIS eye .. f . I d ' d fi compansons between apples and oranges, and to dIS- presence( 0 hione ) or moredi umquet thi y . enve eatures (E nuss the answers that many studies of livIng species autapomorp es accor ng 0 s VIew LDREDGE.. . and CRACRAFT,1980: 107; TATTERSALL,1986, 1991; proVIdekfor questionsabout how the evolutIonaryproc- KE is 91) Thi d fi .. th ' esswor s, \- NNEDY,19 .s e ImtIon means at speciesare, :.t as BOCK (1989: 55) perhaps best puts it, Support for the interpretation of multiple geo- It "what(ever)a taxonomistrecognizes as such". graphic speciesin Middle and UpperPleistocene Homo )- mainly rests on the assumptionof validity for the phy- ;:- The phylogeneticview of specieshas no basis in logeneticspecies concept, since it relies on establishing is the biological speciesconcept. As T An'ERSALL admits the presenceof long lasting morphologicaldifferences .n (1991: 80): "the PhylogeneticSpecies Concept~ which betweendemes (especially in ANDREWS(1984), al- defines the speciesas the minimal diagnosableunit, is ry though see TuRJl.'ERand CHAMBERLAIN(1989) and

:dre BRAUER(1990) for contrary views on whetherthe dif- COON, 1962; THORNEand WOLPOFF,1981; WOLPOFF, ferences actually occur). However, if there actually Wu, and THORNE, 1984; HUBLIN, 1986; JELINEK, were multiple geographic species of Homo in the 1980a, 1981, 1982; Wu, 1987, 1988b; VAN VALEN, Middle and Upper Pleistocene, evolution within this 1986; POPE, 1988, 1991; ECKHARDT, 1987, 1989; adaptiveradiation would involve a heretoforeunknown ETLER, 1991). However, the fact that there is only one magnitude of homoplasy.The parallel trends in these polytypic species of living humans, the evidence link- species,involving brain size expansion,posterior den- ing some of the geographic variants of that species to tal reduction, postcranial gracilization, toral degen- the geographic variants of the Middle Pleistocene, and eration, among other changes,would provide an un- the precept that the present should be used to interpret precedentedopportunity to developinnovative explana- the p~st, combine to provide a convincing reason to set tions. contentions of speciation within Homo erectus aside. Moreover, if the geographic variants of Homo The view that Homo erectus is a single polytypic erectusare to be regardedas valid species,several of species continues to receive wide support from a num- the living races of Homo sapienswould have to be re- ber of scholars, many of whom do not agree on other gardedas valid speciesas well. This is becausein many aspects of the human evolutionary pattern (cf. BRAUER, casesthe sameunique combinationof skeletalfeatures 1990; BILSBOROUGHand WOOD, 1986; DELSON, 1990; that differentiate the geographic variants of Homo ECKHARDT, 1987; HOWELLS, 1980; TuRNER and erectus have been found to differentiate geographic CHAMBERLAIN, 1989; WOLPOFF, Wu, and THORNE, variants of living people (WEIDENREICH,1943, 1946; 1984).

, The "Transition" Problem

Where is the boundary betweenHomo erectus nuchal toms were like the Zhoukoudian remains. He and Homo sapiens?There are four areas of the world has more recently come to agree with most others in where the fossil records are sufficient to examine this classifyingthe specimenas Homo sapiens(1990). question. In Indonesia, the Ngandongsample has de- The entirety of this sample might, in fact, be fied taxonomic assessment.It is not that workers lack allocated to early Homo sapiens (WOLPOFF,I 980a). opinions, but rather that the opinions do not appear The three earliest of the crania are of special interest, reconciliable. In Africa, Europe, and Asia, samples since they best reflect a mix of characteristics. Sale variously allocated to Homo erectusand Homo sapiens seemsto be the most ancient of the crania (HUBLIN, appearto overlap in time during the later Middle Pleis- 1986; CLARKE,1989). It is a very small vault that lacks tocene (approximately 200-400 kyr). In all three re- supraorbitalsand face. Sale is at the low end of the gions, the overlaps may simply be the result of poor Homo erectus range, approximating 880 cc. in endo- dating. If the dating is accurate, however,do we have cranial volume and has other erectus features such as evidenceof one speciesreplacing another with popula- thick cranial vault bones.However, it lacks a superior tions that are contemporaneousfor a period of time, or nuchal line and the nuchal musculatureis clearly de- do we have evidence of the inability to draw a distinctboundaryveloped pathologically. The cranium is rounded, almost between the two taxa becausethere is noboundary?globular, as seenfrom above,and it is quite gracile in Can the seeming chronological and ana- the Homo erectuscontext. Seemingly modern features tomical overlap of different morphs be taken as evi- of the occiputare probablyattributable to this condition dencethat a transition is taking place at this time, per- (HUBLIN,1985), and the specimen is as difficult to hapsmore noticeablebecause the rate of changemay be classify as it is to date. The Thomas Quarry frontal, fasterthan that earlier or later? which may not be fully adult, is small. The supraorbi- tals are moderatelythick and projecting, and by itself In the African fossil record this period is very this region resemblesHomo erectus. The remaining well represented.The latest crania show clear conti- portion of the vault, however,suggests a shorter and nuit), with recent and modern African populations more roundedcranium, and in this respectit resembles (JEL~EK,1980b). Earlier, Lake Ndutu, Kabwe,Elands- Sale. The other of the older crania in this sample is fontein, Bodo, Sale, and ThomasQuarry crania seemto Bodo, a very large specimenwith an extraordinaryfa- date to this general span (HUBLIN,1986). RIGHTMIRE cial size (its nasal breadth is the largest known in (1990) regards the former four as Homo sapiensand Homo). The size and robustnessof the facial region, the later as Homo erectus,while admitting to somesa- including supraorbitals, provide an anatomical link piens-like features in Sale, and some erectus-likefea- with African Homo erectus,but it is clearly a much tures in Broken Hill and Ndutu. HOWELLS(1980) re- larger cranium than any vault traditionally allocatedto gards Kabwe and Elandsfontein as Homo erectus this taxon. The later crania are also characterizedby a specimens (Bodo was not well known in 1980). morphologicalmix of features,and it would be difficult CLARKE(1976) first regarded Ndutu as a new subspe- to establisha clear boundary,morphologically or tem- cies of Homo erectus, largely based on the unexpected porally, betweenAfrican Homo erectus and Homo sa- conclusion that the occipital contours and "thickened" piens basedon thesecrania. As CLARK(1989: 565) put

346 it, there is a "variable degreeof mixing of Homo erec- very conceivably could be attributed to sexual dimor- Ius and moderncharacteristics." phism although the temporal difference may playa role Crania from the later Middle Pleistocene of as well. It is most reasonable to conclude that this is a Europe include Steinheim, Swanscombe, Biache, sample with mixed morphology, partitioned into two Arago, and Petralona that are usually consideredto be distinct subsets because of the marked sexual dimor- early Homo sapiens (WOLPOFF,1980b and references phism and earlier dates for the males. therein -although not by STRINGER(1981) who inter- This period in China is rich with crania, in- prets the differences among them differently). In fact cluding the Zhoukoudian H3 (CHIU et aI., 1973; Wu there is a tendencyto considerall of the Europeanfinds and LIN, 1983) and Hexian (Wu, 1983; Wu and Dong, as Homo sapiens (c.f. HOWELL,1960), although some 1985) crania which are often regarded as Homo erectus find this problematic for the more robust specimens. (Wu and DONG, 1985; Wu and LIN,1983). Moreover, Bilzingsleben is sometimesregarded as Homo erectus there are specimens from Jinniushan (Ls, 1989), Dali (VLCEK,1978), and Vertesszoloshas beendescribed as (Wu, 1981; Wu and Wu, 1985), and Chouhu (ZHANG, a Homo erectus/sapiensspecimen by THOMA(1966). In 1984, 1986) that have been allocated to Homo sapiens. fact, it is likely that the main source of variation be- Of these, the Zhoukoudian specimen is clearly the most tween these specimensis not phylogenetic at all, but erectus-like, and is the oldest. When it was described, rather combines time and sexual dimorphism. This CHIU et aI. (1973) pointed to a number of progressive variability is exacerbatedby the likelihood that the trends the specimen suggests, when compared with the males (petralona, Bilzingsleben, and Vertesszolos)ap- earlier locality 1 materials, including brain size ex- pearto be earlier than the females,but its presencein a pansion and supraorbital reduction (also see Wu and single population is confirmed at Atapuerca. LIN, 1983). Progressive trends in the dental remains Bilzingsleb~ is the only specimenconsistently from Zhoukoudian were more recently established by regardedas Homo erectus.Yet, when the fragments of ZHANG (1991). He was able to show that a pattern of this specimenare comparedwith the much more com- reduction characterized the later to earlier teeth from plete cranium from Petralona, numerous similarities the site that was very similar to other cases of Middle emerge.The frontals sharea high nasion position with Pleistocene hominid dental reduction. a flat, projecting glabellar region above it. The central Hexian would appear to be later, and in many supraorbital region projects markedly anterior to the respects is unlike the Zhoukoudian remains and instead baseof the low-angled frontal squama,and the anterior resembles the Indonesian humans (WOLPOFF,1985). margins of the supraorbitalspreserved in Bilzingsleben This suggests that there may have been a gradient in suggest that both specimensalso share a low upper morphology running from the north to the south of East s facial angle. There is also significant similarity in the , Asia, and that some of Hexian' s features may better occiputs,where comparisonsare possible,although the reflect this gradient than its taxonomic affiliation. The occipital plane of PetraIon a is somewhatlonger, mainly more southerly specimens have proven extremely dif- s due to the presenceof a very large extra-suturalbone at r'. ficult to classify, as is discussed below. Chouhu lambda. The angle betweenthe nuchal plane and the (ZHANG, 1984; ETLER, 1990), the most fragmentary of vertical face at the back of the occiput is virtually ;t the specimens, is quite unlike former hvo in occipital identical, as is the form of the low, vertically thick and morphology but in our view could easily represent a horizontally extensive nuchal torus, lacking any step as the local Homo erectus condition evolves into a S supratoralsulcus or resorbedarea aboveit. Both crania more modem morphology. It is much like the European It are very thick, especially posteriorly. The third male, 0 female described above. Jinniushan and Dali both mix representedby the occipital from Vertesszolos,shares erectus and sapiens characteristics. Jinniushan, which some of these similarities although the erodedtorus is " may be considerable later (POPE, 1992), combines a more medially restricted while the nuchal plane is large and thin cranium ,vith a gracilized posterior (LO If much longer than the occipital plane (WOLPOFF,1977) ., 1989), and an anterior frontal and face that is reminis- ;, -a feature particularly reminiscentof Homo erectus. j cent of the L2 Zhoukoudian female. As a whole the S sample shows a mix of features that characterize Homo lS None of the females have been regarded as erectus and Homo sapiens locally (JELiNEK, 1982). 1- Homo erectus,and these specimensdiffer especiallyin Each specimen has a different combination of these n:1, the occipital regions, preserved in Steinheim, features. Swanscombe, and Biache. Yet, the (low) parietal In Java, the only hominids that can be related in k heightJ(high) occipital breadth relation of the three is any way to the transition question are represented by h very archaic. The size of the crania, especiallyStein- the Ngandong cranial population. Instead of two mor- 0 heim that may be only slightly larger than 900 cc phologies in this case, there has simply been a long a (WOLPOFF,1980b), and the thicknessand projectionof term disagreement about how to classify these people. It the Steinheim supraorbital tori, combine to revealfea- Some include them in Homo erectus or its equivalent \- tures that are more archaic than often perceived.The (cf. LE GROSCLARK, 1955) "Pithecanthropus" (JACOB, 1- greatestdifferences betweenthe more and less archaic 1967, 1981; SANTA LucA, 1980; HOWELLS, 1980; lIt specimensin this sample are occipital differencesthat RIGHTMIRE,1990, 1991), while others include them in Homo sapiens (THORNE,1971; WOLPOFF,1980a), no clear boundary. The important point is that there which can be seenyet again as indicating the difficulty appearsto be a transition over similar time spans in in placing individuals in a transition period. Those,\rho four different places. There does not appear to have have hitherto considered the Ngandong people to be beena single origin of an anatomically distinct species Homo erectus (cf. RIGHTMIRE,1991) may now haveto in one region. In our view the taxonomic affiliations of reconsidertheir position in the light of new datesthat the spe'cimensdiscussed above are problems of classifi- indicate this site may be as recent as 100,000 years cation and not of the perception of the evolutionary (BARTSTRA,1988). If maintenanceof a Homo erectus process.In some casesfeatures attributable to the hvo statuscontinues to be preferred for Ngandong,on ana- speciesare mixed throughoutthe sample,while in oth- tomical grounds, it would have to be seenas evidence ers the sampleas a whole has proved difficult to clas- that there is an overlap of Homo erectus and Homo sify. For this reasonwe regard the solution to the taxo- sapiens of at least 200,000 years duration. If true, this nomic problem as arbitrary, and a matter of taste, would also make the taxonomy of certain Pleistocene rather than as a matter of systematics.What is more Australians very problematic, as many of them show interesting,from our perspective,is the fact that there the most detailed anatomicalsimilarities with the late are any problems of classification at this boundary at Pleistocene Indonesians (THORNE,1984; WOLPOFF,all. Should there have been cladogenesisat about this 1991). time, we would expect taxonomic clarity rather than In sum, for those regions where appropriate taxonomic confusion. There is a real contrast between fossils can be found, there are difficulties in clearly es- the pattern of hominid variation during this period and tablishing where the Homo erectus -Homo sapiens the patternof variation at the end of the Pliocene,when boundarylies. We interpret this to indicate that there is Homo erectusfirst appeared.

, Problem of Defining Modern Homo sapiens-What can be defined? Another aspectof the lack of a clear boundary problem of trying to define instead of diagnose, as betweenHomo erectus and Homo sapiens is the ab- some would lead us to believe (cf. KIMBELand RAK, senceof an operational definition for Homo sapiens.It 1992). If diagnostic features are a consequenceof is always possible that this absencereflects the validity species,their absencehas powerful implications for of KIMBELand RAK's (1992) remark that "there cannot speciesrecognition. KIMBELand RAK (1992) note: "the be any definition of a speciestaxon becausespecies are diagnosisof a speciestaxon is a list of 'symptoms' of individuals, not classes".Yet they do admit that a spe- genealogicalcohesion; diagnostic charactersare those cies has diagnostic charactersand in defenseof their that are held to distinguishone genealogicallycohesive own favored speciesconcept.. the phylogeneticone, they systemof parental ancestry and descent from other assert: "The phylogenetic speciesconcept ...does not similar and/or closelyrelated systems." require that charactersof organismalclusters in a spe- cies be autapomorphic [but see KENNEDY,1991], only The inability to diagnosethis particular species that the organismalclusters themselves be diagnosable implies, then, the inability to show it is actually a dis- as representinga distinct reproductivelycohesive line- tinct genealogicallycohesive system. age. ...speciesmay be diagnosableby any unique setof In our opinion, the reasonfor this state of affairs characterstates." is a consequenceof both the pattern of evolution and But what if the speciesis not diagnosableby any the 'hidden agenda' in the definitions that have been set of characterstates? It is difficult to avoid the impli- proposed.The unsuccessfulattempts to uniquely define cations of the fact that there is great difficulty in diag- Homo sapiens began with the assumption that the nosing Homo sapiens skeletally in a manner that can EuropeanNeandertals were not Homo sapiens, and be consistentlyapplied acrossthe geographicrange of developed criteria to identify Homo sapiens that the living species. showedthis to be the case.These criteria do not work becausethey began with the \vrong assumption. The Two definitions for Homo sapienshave beenof- underlying assumptionfor any definition of Homo sa- fered in recent years (HOWELL,1978; DAY and piens must be about who is in it, and not \vho is not STRINGER,1982). Both of these equate Homo sapiens (WOLPOFFand THORNE,1991). We must begin with with modern Homo sapiens. The HOWELLdefinition the preceptthat all recentand living people are Homo was never operationalized,and in any event involves sapiens(THORNE and WOLPOFF,1991). A definition far too many criteria to be applied to most fossil re- that begins here includes many, if not all, Neandertals mains. The DAY and STRINGER(1982) definition "ith in our species,as well as numerous individuals from or ,vithout the slight modifications provided by other regions presentlyexcluded. But such a definition STRINGERand ANDREWS(1988), was found to exclude \vould have its own problems, not the least of which is a large number of recentand living Aboriginal Austra- the exclusionof someMiddle or Late Pleistoceneindiv- lians from the human species (WOLPOFF,1986: iduals from Homo sapienswhile other contemporaries BROW~. 1990), and the authors accept that it is not are included (ZHANG,1986). Does this mean that there valid (personalcommunication). This is not simply a were t\VOcontemporary species of Homo, coexisting at e different times in different places, or is it an artifact of quenceof the single recentorigin theory). Instead what n the definition? we find is that different features show distributions that e vary across both time and space, which irrevocably We believe the latter is the case.The underlying confusesclade and grade variation; that is, variation .s If problem is that there does not seem to be a single across demes and variation within demes over time. unique recent origin for all living populations, and KIMBEL and RAK (1992) quite correctly say that therefore it is understandablethat we cannot find a "diagnosticcharacters do not make the species",but we 0 single set of features that uniquely identifies them believe that the absenceof diagnostic charactersgoes a 1- (such a unique set of charactersis an expectedconse- long way towards unmaking it. 5- Does a Definition of Homo erectusinclude Homo sapiens?

-e LE GROSCLARK (1955) argued persuasivelythat. large body size; c the diff~ring ge~g~aphicvariants"of p~~-~omosapie~ls .large cranial vaults (virtually all of the specimens It populatIons ( Pithecanthropus, Smanthro-pus, lie above the maximum Homo habilis brain size IS "Meganthropus","Atlanthropus", and so on) should be thus far reported of750 cc). In lumped in a single taxon Homo erectus.He provided a , .' , .'. . d fi ,. fi h . h. h fi II . h." t .anteriorly projecting, vertically thick and continu- n e lrutlon or t e specIesw IC , 0 owing IS presen .. " h t h I t. d t d.st.n ous supraorbital tori, separated from the frontal d to past approac 0 uman evo u Ion, serve 0 I I - b I I . . h u ~ u . Th. d ' t.squama y supratora su CI. n gutS llomo erectus J.romlloma sapiens. IS ISInC-. ' tion became the focus of other definitions of Homo. flattenIng of the frontal squama broken by a me- erectus,For instance, HOWELLS(1980: I) introduces topic eminence; his review with th~ observation"we should begin con- .presence of a sagittal keel; sideration of Homo erectus by consideringHomo sapi- .reduced facial prognathism; ens". The idea that Homo erectusshould be defined in b d I ~ IS . d . f th ( . roa erung an owenng 0 e J.ace speci fi IC all Y relation to Homo sapiens has become almost ubiqUItous ' '. (see HUBLIN,1986; HOWELL,1978; RIGHTMIRE,1988), expansion of the ml~ace: zyg?maxlllal?' regIon, of B h.. th .t d. t. fi th t tak b th nasalbreadth, and orbital Size,WIth reduction of the ut t ISIS e OppOSIe Irec Ion rom a en y e I ~ d d. or evolutionary...owerprocess. It IS, after all, Homo sapiensthat J.acean man Ibl e)'' . IC evolved from Homo erectus. The more useful and in- .lateral nasalaperture eversion; J[ formative question is what separatesHomo erectus .angulation of the nasalbones in the sagittal plane, iC from its australopithecineancestors. To our knowledge and ;C only three workers have addressedthis issue,WOOD. expansionof their breadth relative to the breadth of cr (1984), RIGHT~fIRE(1990) and WOLPOFF(1991). the nasal aperture (while the breadth of the nasal RIGHT~fIRE(1990: 189) provided a short list of apertureitself expands); CS is- nine Homo erectus cranial features "that seemto be .low temporalsquama with a posterior border that is derived ...relative to earlier Homo". This list included unarched; almos~all ~ffeatures .that WO?D (1984: 102) found to .maximum parietal breadth at the mastoid (or asteri- rs be denved In the specIes,relative to a last commonan- onic) notch. Id cestorof .,. all Homo species.RIGHTMIRE pointed out that .t. hIC' kemng .' 0 if t h e crania ' I wa II rnvo. I vrng. b ot h COrtl- . ~n "there IS vanatlon, and not all characters are expressed d d. 1"" ' nc in each of the individuals assignedto Homo erectus". ces an IP oe; he In fact, we find that three of the nine featureshe dis- .kyphosis (flexion) of the cranial base; nd cussed do not characterize enough Homo erectus .marked angulation between the tympanic and the tat specimens to be accurate in a diagnosis, Omitting petrous; rk these, the characters remaining are italicized in the list. presence of an occipitomastoid crest and digastric he below (note, however, that while referring to the same sulcus. a- character states,they are often defined somewhatdif- d.: h I h . h . lot fieren tl)y .nuchal .a Istlnctline nuc and a markingtorus, overa strongangrng angulation t e supenor be- ithno WOLPOFF'S(1991) description of Homo erectus tweenoccipital and nuchal planes; on uniquenessesis also in relation to Homo habilis, its. transversely wide occiput with expanded nuchal us contemporaf)' and possible ancestor. This is summa- muscle attachmentarea; 1m rized .and ex-panded?elow. ~haracters that vary in .dominance of sagittal length of the nuchal plane on magnItude or expressionare Included, as long as the overthe occipital plane length. is variation does not include expressionsthat are common '. IV- in Homo habilis. Many of the comparisons involve. coincidenceof InIon and oplstocraruon; Ies features that are sexually dimorphic in both species, .relative size reduction in the later erupting premo- ::re and the comparisonstake sex into account. In all cases lars and molars, in the context of general reductions at the adult morphologyis described: in the posteriorteeth;

~. J:>U

.lateral maxillary incisor expansion relative to the find a featureas a rare variant in a large numberof central incisors; crania. Nonetheless,it is evident that the great ma- jority of features that distinguish Homo erectus .reduction in the magnitude of sexual dimorphism, from Homo habi/is, are preservedin a significant as expressedmetrically and morphologically, espe- numberof Homo sapiensspecimens: cially sexualdimorphism in the canines; .large body size; .reduction in relative forearm length; .large cranial vaults .appearance of transversethickening of the femoral .vertically thick continuous supraorbital tori,which neck; when presentare separatedfrom the frontal squama .weakened femoral pilaster and by supratoralsulci; .low pilastric index. .flattening of the frontal squamabroken by ametopic Taking into accountthe uneven preservationof eminenceand presenceof a sagittalkeel; cranial, mandibular, and body parts between regions, .low facial prognathism; thesefeatures in combinationdescribe normal variation in all of the taxon. With Homo sapiens excluded from .broad shortfaces (with broad nosesand large orbits, the analysis, these are the autapomorphiesthat can be and continuing reduction of the lower face and demonstratedfor Homo erectus. mandible); But Homo sapiensshould now be considered,as .lateral nasalaperture eversion; the questionasked here is to what extentHomo sapiens .paracoronal angulation of the nasal bones, and can be characterized,as a whole or in part, by these maintenanceof nasal bone breadth relative to the features. This question is not as simple as it might seem,as we envisagethree different forms that shared breadthof the nasalaperture; unique features linking Homo erectus and Homo .low temporalsquama with a posteriorborder that is sapienstake. First, there are features that characterize unarched; earlier populations of the species, but become rare, attenuated or extinct later (for instance in living .maximum parietal breadthat the mastoid(or asteri- populations).Second, there are featureswhich in Homo onic) notch; sapiensare further developedor more prominently ex- .kyphosis (flexion) of the cranial base; petrotym- pressedthan in Homo erectus.Third, there are features whose expressionchanges, in some casesconsiderably. panic angulation; Variation in this pattern can occur once, or a numberof .presence of an occipitomastoid crest and digastric times (cranial thickness, supraorbital development,and sulcus; some of the facial changesexemplify this). Thesethree possibilities delimit the question of how Homo sapiens .a distinct nuchal torus, overhanging the superior relates to Homo erectus.The fact that they follow dif- nuchalline; ferent patterns of variation reflects the evolutionary .relative size reduction in the later erupting premo- process,but the fact that they expressa similarity in lars and molars,in the context of generalreductions derived charactersreflects a phylogeneticrelationship. in the posteriorteeth; Given these provisos, in the listing below, we .lateral maxillary incisors remain substantiallyex- indicate those apomorphic features of Homo erectus pandedrelative to the centralincisors; that also characterize some Homo sapiens. To weigh this comparisonin the direction of maximizing differ- .reduced sexualdimorphism, asexpressed metrically ence, we restrict our Homo sapienssample to the last and morphologically,including that of canines; 30,000 years, a period during which it is universally .short forearms; accepted that the only humans are unquestionably modern Homo sapiens. Moreover, we want to make it .long legs; clear that presencein Homo sapiensmeans commonly presentand does not mean that we were only able to .thick femoral necks.

Do Homo erectusAutapomorphies Preclude a Homo sapiensAncestry? ANDREWS(1984) identified severalautapomor- rived traits not shared with the other group, then the phies in Homo erectus,and attempted to use them to hypothesisthat they are conspecific can be rejected" show that Homo erectus, by any definition, was not a (KENNEDY,1991: 376). ANDREWS'study is widely probable ancestor of Homo sapiens. The underlying cited as one of the more detailed of the recentcladistic assumptioncan be stated: "If it can be demo,nstrated analysisof the three speciesof Homo: habi/is, erectus, that either H. erectus or H. sapiens ...possessedde- and sapiens.The analysis, however,was basedon se- lected data from the literature (unfortunately not em- Even more troublesomeis the fact that the par- phasizing facial characters). From it he recognizeda ticular autapomorphies ANDREWSclaims for Homo short list of Homo erectusautapomorphies (with which ereclusare virtually all invalid. In particular the keel- we do not particularly concur): frontal and parietal ing, thickened vault bone characterstate (especially if keels,thickened cranial vault bone, angulartorus, inion the details of thickening are left unspecified as in the well separatedfrom endinion, mastoid fissure separat- ANDREWSpresentation), angular torus, and separation ing the anterior face of the process from the petrosal of inion from endinion characterize numerous late crest, and a recess present between the entoglenoid Pleistocene,Holocene, and living humans. This brings processand the tympanic plate. focus to anotherdifficulty we have with the ANDREWS analysis (as well as those by STRINGER(1984) and Using this list he developed two far-reaching WOOD(1984)). Most of theseso-called autapomorphies proposals: (I) most Homo ereclus specimensthat are they use are particularly common in the Australian derived for these character states are from east Asia, variant of Homo sapiens, and regularly occur else- and the African specimensare only linked to the Asian where. Are we to conclude that this makes Australians ones through plesiomorphic characters; (2) the exis- a different species,just as the AustralasianHomo erec- tence of the autapomorphiesindicates that Homo erec- Ius variant is said to be a different species?Of course Ius could not be the direct ancestor of Homo sapiens. not, but it is for just this reasonthat the precept of us- He concluded that the African and east Asian Homo ing the presentas a guide to understanding the past ereclusspecimens were in different speciesand that the forces one to reject the contention of geographic spe- direct ancestorof Homo sapienswas Homo habilis. ciation in Homo ereclus. We have great difficulty \vith the contentionthat In our analysis of the features distinguishing autapomorphiescan provide the basis for positing that Homo ereclus from Homo sapiens,we were able to one species cannot be the ancestor of another. We identify only a few Homo ereclus characters that we would prefer to believe that these specieswere capable canconclude are not also sharedby Homo sapiens: of evolution and that features can change over time. Furthermore, HUBLIN(1989) makes quite the opposite .Anteriorly projecting supraorbitaltori; casewith regard to Homo erectus autapomorphies.He .thickening of the cranial wall, involving both corti- arguesthat most of these uniquely derived featurescan cesand diploe; be related to the unusual skeletal mass that especially .strong angulation between occipital and nuchal characterizesthe later Homo erectusremains. As these planes; are aspectsof a single morphological complex,a rever- .dominance of sagittal length of the nuchal plane sal would not be unexpectedand an ancestryfor Homo over the occipital plane length; sapienscould not be excluded. For reasonsdiscussed .coincidence of inion and opistocranion; above,we also do not acceptthe phylogenetic species .weakened femoral pilasterand a low pilastric index. conceptas valid. Subsequentlywe too rejectthe associ- ated notion that a single or a few autapomorphicfea- Of these, some actually are present in recent tures can define distinct species,because this would Homo sapiens,albeit rarely, and many are common in mean the breakup of polytypic species, past and pre- earlier (or so-called "archaic") Homo sapiens. More- sent,and in particular would require the interpretation over, all of thesedifferences can be linked to the evolu- that living humans consist of several distinct species. tionary trends that span both species, of increasing Instead \\'e contend that for a phylogenetic analysisto cultural complexity, increasing brain size (and there- provide support for a hypothesisof ancestry,it must be fore cranial vault size), and the morphological conse- through sister group analysis showing closenessof quencesof the progressive substitution of technology relationship, and not from the assumption that the for biology. We contend that this short list does not evolutionary process does not operate in creating provide a justifiable or valid foundation for suggestions significant diversity within species. that Homo erectusis not ancestralto Homo sapiens.

Homo erectusas a Stage of Human Evolution SCHWALBE(1904), and much later ARAMBOVRG as the ancestorof Homo sapiens,the question remains (1958),outlined a model of human evolutionemphasiz- whether the appearanceof Homo sapiens involved a ing a progressionthrough universal stagesof organiza- gradual (although perhaps accelerated)transition, or tion. In this evolutionary model Homo erectus is seen cladogenesis.The nature of the transition, and the ab- as the ancestorof Homo sapiens, clearly not a point of sence of clear boundaries provides convincing evi- contention among most paleoanthropologists,except dence, in our view, that no species split occurred in for those who accept both the "Eve Theory" for the theseMiddle Pleistocenehominids. recent African appearanceof modern humans and a Evolutionary trends within the Homo erectus definition of Homo erectus which restricts this species samplesupport this interpretation. Whether the sample to eastand southeastAsia. If Homo erectus,in its en- is considered in its entirety (WOLPOFF, 1984; tirety or only one geographic portion of it, is regarded RIGHTMIRE,1990), or regional sub-samplesare de- scribed (for instance, see ZHANG'S1991 analysis of the of speciesin the Hominidae is expectedto range be- Zhoukoudian teeth or RIGHTMIRE'S1990 discussionof tween 3 and 11, with a mean expectation of under 6. brain size change in Indonesia), numerouscharacteris- Without Homo erectusthe minimum possible number tics of Homo erectus show significant evolutionary of hominid speciescovering the approximately 7 myr change in the direction of Homo sapiens throughout of the Hominidae's existencewould appearto be 5, in the Early and Middle Pleistocene.These evolutionary our view: Australopithecusafarensis, A. boisei, A. afri- trends are not all in the same direction. Thus, as brain canus, Homo habilis, H. sapiens. Virtually all workers size and many of the cranial dimensions increase,pos- believe there are more speciesthan this, and we con- terior tooth size and the dimensions of the mandibular clude that comparedwith expectationsbased on other corpusdecrease. primates, the problem with hominid taxonomy is clearly not that there are too few species! Therefore, no single underlying cause,such as a change in body size, could account for these trends. Fourth, we can ask whether subsuming Homo Homo sapienswould appearto be their consequence. erectus within Homo sapiens results in an unrealisti- cally large amount of phenetic diversity within the Given the absence of cladogenesis between resulting species.That this is an important questionis them, the question remains whether the specimens suggestedby the fact that at least one worker believes presentlyassigned to Homo erectus can validly be con- that, even as presentlydefined, Homo sapiensencom- sideredboth a stage in human evolution and part of the passestoO" much variation: "Most human fossils of the species Homo sapiens. There are four ways of ap- last half-million years are conventionally referred to proaching this issue. the speciesHomo sapiens,which is therebystretched to First, if the demesof the polytypic speciesHomo limits which would strain the averagemammalian ge- erectus are by in large related to the demes of Homo nus" (TAlTERSALL,1991: 80). But, is this actually sapiens,as THOItlA(1973) points out, the two taxa must true? be merged for the principle of monophyly to apply, even though there might be consistentdifferences that WEIDENREICHreported quite the opposite con- characterizeall of the Homo sapiensdemes. Otherwise, clusion as a result of systematic comparison of fossil Homo sapienswould not consist of a group of geogra- and living human variation with Canis farni/iaris. Of phic demes, their last common ancestor,and all the courseone might argue that the variation in dogs is an descendantsof that ancestor. invalid comparisonbecause of their domesticationand breeding into distinct types. Yet, domestication is a Of course,this argument is weakenedsomewhat processoften usedto describethe later stagesof human by the fact that genetic continuity reflected in some evolution, and the recognized dog breeds have been featuresis not the continuity of populations. createdby diversifying varieties and establishing high frequenciesof different allele combinations,and not by Second, species need not be regardedas static, different genes.They therefore remain as a valid model and contrary to the assessmentsof some authors a of the amount of genetic variation encompassedwithin phylogenetic approachdoes not rest on the assumption a polytypic species,even though their phenotypic vari- that punctuated equilibrium is the only mode of evolu- ation is packagedin an unusualway. tionary change (WILEY, 1981). As discussedabove, the evolutionary species concept allows for evolutionary Moreover, the magnitude of variation within change. In fact, the degree of variability is irrelevant to living humansis often underestimated.The impetus of the issue since the species question is one of phyloge- WEIDENREICH'S"single specieshypothesis" was, after netics and not of phenetics. Without cladogenesisit all, his experiencefrom comparative anatomy,that the seemsquite unjustified to arbitrarily divide this lineage individual details of humanfossils could be largely in- into species,whether or not there is significant evolu- corporatedwithin the range of variation of living peo- tionary change within it. And in fact, within a broader ples. We concur,as it is our experiencethat the range primate context there is not that much difference be- of variation in recentHomo sapiens,whether in a sin- tween Homo erectus and Homo sapiens. The former gle fossil sample (for instancethe Klasies River Mouth can largely be subsumedwithin the range of the later. Cavemandibles 16424and 13400)or in a single living population of the species, often encompasseswhat Third, we can examine the questionof whether others have regarded as multiple taxa (WOLPOFF, subsumingHomo erectus within Homo sapiensresults 1992). If this range is taken as "normal" for horninids in an unrealistically small amount of phylogenetic of the Pleistocene,TAlTERSALL's concern is clearly diversity in the hominids. STANLEY(1979) argues that unwarranted. rates of speciation differ through the span of a taxon, and are highest early on (a pattern that well-fits the Finally, we can ask whether the amount ofvari- Hominidae). This would suggestthat the longer a taxon ation within a speciesis a valid concern. T AlTERSALL exists, the lower the average rate of speciation within himself usesa speciesdefinition that relies on a distinct it. Summarizing a large body of data about speciation beginning (cladogenesis)and a distinct end (extinction rates, FOLEY(1991: 416-417) predicts that the number or a secondcladogenesis) to delimit the boundaries,as we do. Unlike us, he further acceptsthe criterion of a unacceptableand unnecessarystipulation. The fact is single unique link, establishinga unique ancestral-des- that for speciesto be validly related in a phylogenetic r cendant pattern of relationship among the members. scheme,their definition must allow for the pattern of r Nowhere within the phylogenetic definition he uses descentto be establishedand for this the sizes of the does the question of amount of variation appear,and entities are not important. This is the very sort of 1\ we agree with this. The evolutionaryspecies is also not phenetic concern that the change in paradigms from s delimited by the amount of variation it encompasses, species-groups-as-similar-entitiesto species-groups-as- for if it were, this would place an arbitrary limit on the related-entitiesis supposedto haveavoided. r amount of evolution that could occur in a species,an s Precedents for Sin] Homo erectus Our contention that Homo erectus cannot be assumptions;he acceptedthe most important element validly distinguished as a different speciesfrom Homo of the biological speciesconcept, that a speciesdefines sapiens has been embraced by a number of workers the limits to gene flow, and acceptedthe phylogenetic before us, spanning the last half-century. The first to view that taxa are monophyletic. He concluded that addressthis issue, from the perspectiveof the modem given its geographicdifferentiation and ancestralposi- con~eption of what constitutes Homo erectus, was tion, Homo erectuscould not exist as a valid separate e WEIDENREICH.In discussing the nomenclatureof the taxon but mustbe regardedas an earlier portion of the 0 Zhoukoudianremains (1943: 246), he admitted that his polytypic speciesHomo sapiens. 0 use of "Sinanthropuspekinensis" wasa convenience"... without any "generic" or "specific" meaning or, in On the basis of his allometric studies, HEMMER other words, as a ';latinization" of PekingMan". (1967, 1969) concluded that the ground-plan ("An- lage") of the cranium "is the same from the earliest His view of the Zhoukoudian taxonomy was definite membersof the erectusgroup (e.g., "Pithecan- quite explicit, and one with which we concur: "... it thropus IV" from Sangiran, Java)to the modem races would not be corr~ct to call our fossil "Homo peki- ,il of H. sapiens Since these features vary among the nensis" or "Homo erectuspekinensis"; it would be best )f recentraces no less significantly than betweendifferent to call it "Homo sapienserectus pekinensis". Othenvise In fossil groups,or betweenfossil and recent populations, it would appear as a proper "species",different from Id it is impossibleto draw a line anywherefor speciesde- "Homo sapiens", which remains doubtful, to say the a limitation unless one intends also to split up recent least." WEIDENREICH'sprimary reasons for this ap- 1\ man into severalspecies. Therefore it seemsnecessary proachwere threefold. First, as mentionedabove much :1\ to include all of these fossil and recent groups in the of the range of variation in Homo erectuscan be sub- single speciesH. sapiens" (1969: 179). ;h sumed within the normal range of variation of recent JY and living Homo sapiens. Second,the fossil races of lei Homo erectus correspondto the living racesof Homo The problem of consistencyhas plagued all of in thosewho have tried to equatefossil and modem geo- sapiens.Third, the total amount of variation in Homo -i. graphic variation in the same taxonomic scheme.The sapiens,including fossil and living, is not especially great, in his comparativeview. For instance,he found idea of separatespecies for living populationsis unac- the differencesamong dog varietiesto be at the level of ceptable,and clearly incorrect, [although some have In regardedthe notion as valid, even in this century (HILL, of subfamily variation in the higher primates -he held that the variation among dog varieties madea reason- 1940)].The preceptof using the presentto interpret the er pastis clearly the acceptableapproach for HEMMER,as able comparisonfor the human/gorilla differences-far he well as for ourselves. n- exceedingthe variation within Homo sapiens. '0- THOMA(1962, 1973) presentedan evolutionary Although primarily concerned \vith early homi- "e " model for Pleistocenehominids that was very similar to nid evolution, ROBINSON(1967: 98) addressed the n- COON'S(1962), in that he envisioned(1973: 532) "at Homo erectus question as part of his general revision lth least three collateral phyla [that] passedthrough three of hominid taxonomy. He argued that in the broad view ng evolutionary phases, within the framework of a spe- of human evolution, "most of the obvious physical lat cies Becausethe recentend points of [each] phylon change had already occurred" at the time of the appear- FF, is not separatedfrom the others at a specific level, the ance of Homo erectus. All subsequent human popula- ids taxonomic limits of Homo sapiens must be necessarily tions were mainly characterized by a single evolution- rly extended in the past back to the point of ramification" ary trend, in his view, "the realization of the cultural in order for it to be monophyletic. potential". He believed Homo erectus and Homo sapi- ens should therefore be subsumed in the single species ri- In this later precept, he differed from COON, Homo sapiens, because it has priority. It is ironic that if LL who chose to retain the two species, erectus and sapi- this vie\\-point now becomes known as the "single spe- lCt ens, and then stumbled badly in trying to explain how cies hypothesis", as WOOD (quoted in LEWIN 1991b) on different populations evolved from the one to the other. and KIMBEL (1991) propose, ROBINSONwould have to as THOMA'S analysis was based on a combination of two be considered one of its earlier adherents.

king JELiNEK(1978, 1980a, 1981) consideredHomo (JELiNEK,1981: 88). HUBLIN(1986) took the position erectusas a subspeciesof Homo sapiens,because there that cladistically, Homo erectus does not exist as a are no clear boundariesbetween them -morphological, taxon separatefrom Homo sapiens. But he was not chronological, behavioral, or environmental. He con- fully comfortablewith his conclusion,and subsequently cludes from this that there is no genetic barrier between tied his interpretation of the species to the punctua- them. Focusing on the transitional specimensspanning tional approach. HUBLINwas willing to accept Homo the erectus-sapiensboundary, he concluded that the erectus as a pre-sapiensgrade if it could be demon- African, European, north Asian, and Indonesiansam- strated that it exhibited evolutionary stasis. On the ples can eachbe describedas normal variations within other hand, if significant evolution could be shown the limits of a single species.Moreover, he raised the "Homo erectus should be included in Homo sapiens" issue of criteria for species definitions, questioning (p. 184). In his 1986paper he did not clearly define his whether global morphology, regional morphology, position, but since we are convinced that there is in- chronology, or cultural traditions provide the more deed a record of significant evolutionary changes in valid means of separatingthe species.In his view, the Homo erectus,we take HUBLIN'Sposition to be in sup- anatomicallinks betweenMiddle and Late Pleistocene port of our own. populations from different regions make it impossible to regard some, such as Ngandong, as Homo erectus R.E.F.LEAKEY(1989: 55) recognizesthe pattern while other contemporariesare Homo sapiens. "Have of Pleistocenehuman evolution we have described: "I we any solid scientific grounds on which to consider do not favour the idea that the modem form of our Middle PleistoceneEuropean finds, with earlier mor- specieshad a single geographical origin. The fossil phological cranial changes,as Homo sapiensand the evidence from widely separated parts of the world extra-Europeanfinds evolving in the samedirection but indicatesto me that Homo sapiens in the modem form in somewhatdiffer~nt degreeand time sequenceof ad- arosefrom populations of the more archaic form wher- aptationinto different conditions as Homo erectus?The ever it was established;and that similarly, thesearchaic whole mode and the processof the hominid evolution- forms arose from establishedpopulations of so-called ary processshows that there are not, and that in the Homo erectus.There are specific examples that cannot past [there] could not have beendifferences at the spe- be brushedaside." cies level, but only at the subspecieslevel, whetherthe cerebralisationprocess -as only one part of the mosaic He proposedto use the beginning of encephali- of evolutionary changes-started earlier or later. The zation, surelyone of the unique hominid trends in both logical consequenceof such a situation is to lead us to morphologicaland metric change,to define the begin- consider the different African, European, and Asian ning of a single evolving lineage, and argued (p. 57) finds of H. erectus type as Homo sapiens erectus" that: "the enlargedand more complex brain arose only (JELiNEK,1978: 427-428). once, in the lineage embracingboth Homo erectus and He argued further that Homo sapiens is a Homo sapiens." He added: "I am increasingly of the polytypic species and its evolution is mosaic. "The vie," that all of the material currently referred to as anatomical links betweenMiddle and Late Pleistocene Homo erectusshould in fact be placed within the spe- populations are important for an-other reason-these cies sapiens [which would] project Homo sapiens as a links are strongly regional. If the differential diagnosis speciesthat can be traced from the present, back to a betweenHomo erectus and Homo sapiens cannot be little overtwo million years," other than by convention,and ...this conventionmust be different for different geographicalregions, then the This would employ evolutionary speciescriteria value of such a difference should be critically consid- for justifying the process of merging Homo erectus ered. ...It is time to replacethe paleontologicalspecies with Homo sapiens,since it establishesspecies identity with a biological one. ...Paleontological taxonomy through the persistenceof a long term, clearly impor- cannot be in contradiction with ...biological facts" tant, evolutionarytrend.

The Origin of Homo erectus The phylogeneticrelationships and taxonomyof crania such as ER 1590 and 1805 are very slightly the hominid populations immediately ancestral to earlier (FEIBEL et al., 1989) but still after the earliest Homo erectus are in a state of flux (STRINGER,1986; cranium ER 2598. The putative ancestor of Homo erec- TOBIAS,1989, 1991; WooD, 1987). Despite this, we tus, Homo habilis (sensu lato), does not only predate believe that at the moment, the follo\\ing tentative the Homo erectus occurrences at Lake Turkana and conclusions can be drawn. Homo erectus appears in suT\ives this emergence by at least several hundred east Africa, certainly at Lake Turkana, as a distinct de- thousand years in a seemingly unaltered form in Bed II, finable cranial morphology (WALKERand LEAKEY, Olduvai Gorge (cf. OH 13, in comparison \\ith ER 1978) by at least 1.9 million years ago (raising the 1813). The most reasonable interpretation of these facts possibilit). that it may have evolved somewhereelse in is that Homo erectus arises as a cladogenic event. Africa). The earliest complete definitive cranium, ER Characters that diagnose Homo erectus com- 3733,just postdatesthis figure, while more problematic monly appear together in virtually all Homo erectus specimens, subject to the vagrancies of preselvation. tures into a pattern thereafter commonly found in all Many of thesefeatures can also be found, albeit rarely' humans of the Lower Pleistocene,and later, strongly and in isolation, in others attributed to Homo. A com- suggeststhat there had been the type of adaptive and bination of body size variation (midget to basketball morphological reorganization that characterizes new player height) and the marked sexual dimorphism of species.This supportsthe implications drawn from the these earlier hominids confuses the meaning of this existenceof Homo habilis specimensthat postdatethe earlier distribution, but what is most important, we be- earliestHomo erectusremains -Homo erectusarose in lieve, is the isolated nature of the occurrences.Thus, a distinct cladogenicevent and therebyhas a definable n we find Homo erectus features in the lower border of beginning. the nasal aperture and inferior maxilla of STS 52, the cranial base of STS 19/58,the cranial rear of STW 53, O~r interpretationof hominid systematicsat the the midface and temporo-mastoidregion of SK 80/847, time of the earliestHomo erectus in East Africa, prior the endocranial size of ER 1470, and the frontal and to the beginning of the Pleistocene,is that there are a occipitomastoid morphology of ER 1813. That several minimum of three hominid species,as WALKERand Plio-Pleistocenespecimens each have a few different LEAKEY(1978) suggestedmore than a decadeago. We Homo erectusfeatures does not make themHomo erec- hope this will reassurethose who are concerned that n Ius, but also does not negate the validity of those fea- our hidden agendais to "bring the single specieshy- tures in diagnosing Homo erectus, when found in pothesisback" (cf. WOOD,quoted in LEWIN(1991b); combination. The coalescenceof these and other fea- KIMBEL,1991).

Summary: Why] the Taxa? To review. for the reasons summarized below That is, there is no evidenceof an appearanceof we proposehere to merge the Homo erectussample in a new combination of features separating earlier and the expanded evolutionary speciesHomo sapiens. Its later populations in one area (which contrasts with the origin is in a cladogenic event that is at least 2.0 myr. eventsat the origin of Homo erectus).The characteris- We view the subsequentlineage as culturally adaptedto tics of Homo erectusand Homo sapiens are mixed in an increasingly broad range of ecologies, ultimately transitional samplesthat are found in the later Middle leading to its spread across the world some half or Pleistoceneof every region where there are hominid more million years later. Homo erectus,the name gen- remains. We interpret these data to mean that there is It erally acceptedfor our lineage in the Early Pleistocene, no speciationinvolved in the emergenceof Homo sapi- differs from contemporaryHomo habilis in a numberof ens from Homo erectus. The absenceof a cladogenic ways. The vast majority of thesedistinctions also char- event creating a distinct boundary at the "origin" of acterize Homo sapiens; that is, the diagnosis of Homo Homo sapiens, together with the related patterns of Id erectus relative to Homo habilis largely characterizes pol.)1)'pismin both "species",provides an explanation le15c- Homo sapiensas well. Every one of the few characters for the inability to develop a valid morphological that are not shared appearto change in responseto the definition of Homo sapiens.These reasonscombine to evolutionary trends of increasing cultural complexity, require that the lineage be regardedas a single evolu- increasing brain size, and the progressivesubstitution tionary species. of technology for biology. Homo erectus,thus, is not a We view this evolutionary speciesas spanning static species.It shows a number of evolutionarytrends the entire Pleistocene.The species is geographically in the direction of Homo sapiens. ia'IS pol.)1)'piconce humanpopulations begin to migrate out of Africa, 1.4-1.2myr. This pol.)1)'pismis characterized ly Moreover, Homo erectus is a polytypic species, in morphologicalcombinations which include (but are ,r- divided into several distinct geographicvariants which not restricted to) a number of features we believe are show continuity with the geographic variants of the used in kin recognition. The main evolutionary ten- polytypic speciesHomo sapiens through the sharing of dencies in the speciesare linked to the evolving cul- unique combinations of morphological features.There tural systemand its role as both a causeand an effect in is no distinct boundary between Homo erectus and what we regard as the human biocultural evolutionary lly Homo sapiens in time or space. process. ~5t 'c- Behavioral Evolutiol in a Single Species Ite While the fusion of Homo erectus and Homo ture of human behavior -the unique nature of biocul- Id sapiensas proposedhere is a purely biological concept tural feedbackas the mechanismdriving our learning, ::.dII,:R:t5 that removesmuch of the confusionand misconception adaptation,and social organization. The idea that a set inherent in the presenttaxonomic dichotomy, this ac- of cultural and morphologicaltraits can be interrelated tion also has a clear messagefor perceptionsof our stems from DARWIN(1871) and forms the basis of behavioral evolution. We believe the notion of Homo much anthropologicalthinking about the human evo- n- sapiens as a single evolving species throughout the lutionary process,beginning with WASHBURN(1960). us Pleistocenemakes clear the central distinguishing fea- The cultural evidenceassociated with Early Pleistocene

Merge1d f

Homo sapiens, as defined here, represents a major dertal (MERCIERet al., 1991) makes it clear that change in complexity. The dramatic expansionof the changesin behaviorthrough time are by no meanswell brain caseis a clear reflection of major changesin our correlated \vith concurrent biological developments,if 'capacityto learn and develop vastly expandedcultural the juxtaposition of industries and hominids in the Le- systems. vant had not already done so. More broadly, prehis- torians devising frameworks with artifactual or adap- For many years, interpretations of the biological tive changes linked to alleged speciation events can aspects of human evolution have been driven by the now focus on an interpretation of the archaeological divisions made by prehistorians (BRoSEand WOLPOFF, record that is essentially independentof phylogenetic 1971). In Europe, for example, the Late Pleistocene changes.The present global and regional variants of skeletal morphology has been broadly divided on the modernhumanity must continue to be a warning of the basis of Mousterian and Upper Paleolithic stone indus- dangers of drawing simple links between the com- tries. However, the late date for the St. CesaireNean- plexities of biological and behavioralsequences.

Subdivisions fo r Homo sapiens One final immediate question that stems from a origins. Clearly, we do not envision this process as dramatically enlarged Homo sapiens is: how do we re- involving a speciationevent. fer to all the various time and space segments of this evolving complex of morphologies and cultures? As The division of early, middle, and late as here our merging of Homo erectus into Homo sapiens sug- proposedfollows the samedivisions used for the Pleis- gests, we feel strongly that the difficulties in examining tocenePeriod. Tying our definitions of the sapiensdi- the lineage have been exacerbated by excessive tax- visions to the geochronologicaldivisions of the Pleisto- ceneconfers the advantageof having faunal, geologic, onomizing (e.g. HOWELL, 1978; GRO-VES, 1989; CLARKE, 1990). We do not wish to see temporal sub- and paleoenvironmentaldata directly associatedwith species (cf. STRINGER,HOWELL, and MELENTIS, 1979) our schemefor referring to the segmentsof Homo sapi- take the place of temporal species, as this would recre- ens. Therefore,we proposethe use of the geologic di- ate the very problems we hope to solve. visions for Homo sapiens. Further subdivisionsare, of course,consistent for both time and space. Thus, the The fossil record of Homo sapiens can be most Zhoukoudiansample of Locality 1, spanning 400-600 validly discussed in a manner that is totally outside of a kyr, would be an example of Chinese Homo sapiens taxonomic context, since geographic subdivision at the from the Middle to late Middle Pleistocene.The Euro- subspecies level in the past would require a subspecies pean Neandertalswould be an example of European interpretation of distinctions between the races of to- Homo sapiensfrom the earlier part of the Late Pleisto- day, and the division of the lineage into temporal sub- cene. species would recreate all of the difficulties solved with the merger of the species. We would prefer to see the Finally, we recognizethat not all specimensare comparisons that are the requisite of any evolutionary well dated, and that for some the provenancesand/or study be made on the basis of the familiar, simple, and dates may never be accurate. One could potentially generalized divisions of early, middle, and late Homo, criticize our schemewith the argument that without sapiens. We would follow this with a fourth division of information abOutage, it is unclear what Homo sapiens recent, or living. Such a course would allow the con- group to put a specimenin, while taxonomic categories tinuing assessment of the establishment of our species basedon anatomy can be defined regardlessof a speci- within Africa as early sapiens. The period would cover men's age. But, in fact, the taxonomic categoriesbased the cladogenic event that separated sapiens from its on anatomy within PleistoceneHomo have also not parent and sister species sometime before the beginning proven to be particularly unambiguousto define and of the Pleistocene, up until the beginning of the Middle the confusion of grade and clade characters makes Pleistocene, some 780,000 years ago. This arrangement these categoriesinherently inaccurate; we would say, would also make clear that Homo sapiens is the only on the whole more inaccurate than the geochronologi- hominid ever to live outside of Africa. cal schemewe proposehere. Moreover, without infor- mation about age,the importance of a specimenin any Middle Homo sapiens would cover the period of evolutionary scheme is minimized in any event, no anatomical and cultural establishment of humans out- matter how evident its taxonomyappears to be. In all, side of Africa up until approximately 130,000 years we contend that systematics based on morphology ago, when it is clear that there are the beginnings of alone within the speciesHomo sapiens provides no major changes developing in some areas. This period is advantageover the non-taxonomic (or perhaps,better, the subject of current debate about "modern human" ataxonomic)approach we suggest.

Conclusions A major objection to the merger of Homo erec- speciesspanning the last two million years.We believe, tus into Homo sapiens is the difficulty that some will however,that \"e have reachedthe point in thinking have in conceiving of and accepting a single evolving aboutthe later phasesof our evolution when the con- 357

cept of Homo erectus -as a stage,a named ancestor, that only a single genus, Homo, should be used to and as a time and morphological segmentof our devel- describeour past two million years of evolution. We oping lineage -has clearly begunto obscurea clearun- think that similarly, the time has come to recognize derstanding of the dynamic processesthat have taken that only a single polytypic evolutionaryspecies should place. The complexity of the physical and cultural be used for this period. For us it is not so much a events of the last two million years are not made "sinking" of Homo erectus,as a more basic recognition clearer by a proliferation of taxa -the stamp collector that Homo sapienshas priority. This taxonomic ad- approach that is central to the phylogenetic species justmentwill help make clear what many workers have come to realize, that we have been \vise for much concept. For the last thirty-odd years we have recognized longerthan has beenthought.

Acknowledgments We thank the curators of the fossil evidencethat we haverelied on. We have found the detailed study of the original specimensto be of immeasurablevalue in consideringthe problemsraised in this paper.We are deeplyindebted to JensFran- zen for inviting us to participate in this 4th InternationalSenckenberg Conference, marking the 100th anniversaryof Dubois' Trinil discovery. It is Mark Flynn of the University of Missouri who first suggestedthat many of the seeminglynon- adaptive characteristicswe have found to be regionally distinct through the Middle and Upper Pleistocenemay well be recognition markers for peripheral populations,and therebysubject to stabilizationthrough a balanceof gene flow and local selection.We are very grateful for this suggestionas it has cometo playa significantrole in our thinking. We thank David Frayer, Con Chil- dress, Sang-HeeLee, Tracey Crummett, B.A.Wood and Jeff Miller for their many helpful, substanuveimprovements to the text. . Bibliography ANDREWS,P.J. (1984): An alternative interpretation of the characters used to define Homo erectus.- In ANDREws, P. and J.L.FRANZEN(eds.): The Early Evolution of Man, with Special Emphasis on Southeast Asia and Africa.- Cour. Forsch.- Institut Senckenberg, 69: 167-175. - ARAMBoURG,C. (1958): Les stades evo1utifs de l'humanite.- The Leech (Raymond A. Dart Commemorative Number), 28: 106- f Ill. BARTSTRA,G.J. (1988): The age of Solo man and Java man.- In PREscoTT,J.R. (ed.): Early Man in the Southern Hemisphere. Archaeometry: Australasian Studies 1988 (Supplement): 12-22; Adelaide (University of Adelaide Press). BATESON,P. (1978): Sexual imprinting and optimal outbreeding.- Nature, 273: 659-660. BATESON,P. (1982): Preferences for cousins in Japanese quail.- Nature, 295: 236-237. Bn.SBOROUGH,A., and B.A. WOOD(1986): The nature, origin, and fate of Homo erectus.- In WOOD,B.A. and P.ANDREwS(eds.): Major Topics in Primate and Human Evolution: 295-316; Cambridge (Cambridge University Press). BOCK, W.J. (1986): Species concepts, speciation, and macroevolution.- In IWATSUKI,K., RAVEN, P.H., and W.J.BOCK (eds.): e Modem Aspects of Species: 31-57; Tokyo (University of Tokyo Press). Ir BONDE, N. (1981): Problems of species concepts in paleontology.- In MARTINELLI,J. (ed.): Concept and Method in Paleon- y tology: 19-34;Barce10na (University of Barcelona Press). It BRACE, C.L. (1991): Cultural innovation and the mechanism for the emergence of modem morphology (abstract).- Amer. J. 'S :s Phys. Anthropol., Suppl., 12: 52. BROSE,D.S., and M.H. Wo1pofT (1971): Early upper Paleolithic man and late middle Paleolithic tools.- American Anthropolo- 1- d gist,73: 1156-1194. BROWN,P. (1990): Osteological defmitions of "anatomically modem" Homo sapiens: a test using modem and tenninal Pleisto- )t cene Homo sapiens.- In FREEDMAN,L. (ed.): Is Our Future Limited by Our Past?: 51-74.- Proc. Third Conf. Australas. d Soc. Hum. BioI.; Nedlands (Centre for Human Biology, University of Western Australia). :s CHIU CHUNGLANG,Gu YSMIN, ZHANGYINYUN, and CHANGSHENSHUl (1973): Newly discovered Sinanthropus remains and stone

1- artifacts at Choukoutien.- Vertebrata Palasiatica, 11: 109-131. r- CLARK, J.D. (1989): The origins and spread of modem humans: a broad perspective on the African evidence.- InMELLARs, P., and C.B.STRINGER(eds.): The Human Revolution: Behavioura1 and Biological Perspectives on the Origins of Modem IY 10 Humans: 565-588; Edinburgh (Edinburgh University Press). CLARK, W.E. LE GROS(1940): The relationship between Pithecanthropus and Sinanthropus.- Nature, 145: 70-71.

~Y CLARK, W.E. LE GROS(1955): The Fossil Evidence for Human Evolution.- Chicago (University of Chicago Press). 10 CLARKE,R.J. (1976): New cranium of Homo erectus from Lake Ndutu, ianzania.- Nature, 262: 485-487. r, CLARKE,R.J. (1990): The Ndutu cranium and the origin ofHomosapiens.- J. Hum. Evo1., 19 (6-7): 699-736. COON,C.S. (1962): The Origin of Races.- New York (Knopf). DARWIN, C. (1871): The Descent of Man, and Selection in Relation to Sex.- London (John Murray). . DAY, M.H., and C.B.STRINGER(1982): A reconsideration of the Omo Kibish remains and the erectus-sapiens transition.- In L 'Homo erectus et la Place de l'Homme de Tautavel panni les Hominides Fossiles, 2: 814-846; Nice (Louis-Jean Sci- Ig entific and Literary Publications). 358

&LSON, E. (1990): Speciesand speciesconcepts in paleoanthropology.-In HECHT,M.K. (ed.): Evolutionary Biology at tile Crossroads:141-145; Flushing (QueensCollege Press). DELSON,E. (1991): How many speciesof Homo in Europe?-In:NATO Scientific Affairs Division Advanced ResearchWork- shop:Les Premiers PeuplementsHumains de l'Europe;Paris (NATO). ECKHARDT,R.B. (1987): Evolution eastofEden.- Nature,326: 749. ECKHARDT,R.B. (1989): Matching molecularand morphologicalevolution.- Human Evolution, 4 (4): 317-319. ELDREDGE,N., and J.CRACRAFT(1980): PhylogeneticPatterns and tile EvolutionaryProcess.- New York (Columbia University Press). ETLER,D.A. (1990): A case study of tile "erectus"-"sapiens"transition in Asia: Hominid remains from Hexian and Chaoxian Counties,Anhui Province,China.- KroeberAnthropol. Soc.Pap., 71-72: 1-19. FAGAN,B. (1990): The Journeyfrom Eden:The Peoplingof Our World.- London(Thames and Hudson). FEffiEL,C.S., BROWN,F.H., and I. McDouGALL(1989): Stratigraphiccontext of fossil hominids from tile Omo Group deposits: Northern Turkana Basin, Kenya and Ethiopia.-Amer. J. Phys.Anthropol., 78 (4): 595-622, FOLEY,R.A. (1991): How manyspecies of hominid shouldthere be?- J. Hum. Evol., 30 (5): 413-427. GINGERICH,P.D. (1985): Speciesin tile fossil record: concepts,trends, and transitions.-Paleobiology, 11: 27-41. GROVES,C.P. (1989): A Theory of Human and PrimateEvolution.- New York (Oxford University Press). HEMMER,H. (1967): Allometrie-Untersuchungenzur Evo!.utiondes menschlichenSch1idels und seinerRassentypen.- Stuttgart (G.Fischer). IfEMJl,ffiR,H. (1969): A n~w view of tile evolution ofman.- CurrentAnthropology, 10(2-3): 179-180. HILL, W.C. Osman(1940): Anthropologicalnomenclature.- Nature, 145: 260-261. HOWELL,F.C. (1978): Hominidae.-1n MAGLIO,V.J., and H.B.S.CooKE(eds.): Evolution of African Mammals: 154-248;Cam- bridge (Harvard University Press). HOWELLS,W. W. (1959): Mankind in tile Making.- GardenCity (Doubleday). HOWELLS,W.W. (1980): Homo erectus-who, when andwhere: a survey.-Vb. Phys. Anthropol.,23: 1-23. HUBLIN,J-J. (1985): Human fossils from tile north African Middle Pleistoceneand tile origin of Homo sapiens.-1n DELSON,E. (ed.): Ancestors:The Hard Evidence:283-288; NewYork (Alan R. Liss). HUBLIN,J-J. (1986): Somecomments on tile diagnosticfeatures of Homo erectus.-1nNOVOTNY, V. V., and A.MIZERovA(eds.): Fossil Man. New Facts, New Ideas. Papers in Honor orlan Jelinek's Life Anniversary.-Anthropos (Brno), 23: 175- 187. HUBLIN,J-J. (1987): Qui rut l'Ancetre de I 'Homo sapiens?-Pour la Science,113: 26-35. HUBLIN,J-J. (1989): Les characteresderives d'Homo erectus:Relation avec I'augmentationde la masse squelettique.- In GIACOBINI,G. (ed.): Hominidae:Proceedings of tile 2nd InternationalCongress 6fHuman Paleontology:199-204; TUrin (JacaBooks). JACOB,T. (1967): SomeProblems Pertaining to the RacialHistory of tile IndonesianRegion.- Utrecht (DrukkerijNeerlandia). JACOB,T. (1981): Solo man and Peking man.-1n SIGMON,B.A., and J.S.CYBULSKI(eds.): Homo erectus.Papers in Honor of DavidsonBlack: 87-104 (combinedbibliography at end of volume);Toronto (University of Toronto Press). JELINEK,J. (1978): Homo erectusor Homo sapiens?-Recent Advances in Primatology,3: 419-429. JELINEK,J. (1980a): EuropeanHomo erectusand tile origin of Homo sapiens.-1nKONIGSSON, L.K. (ed.): Current Argument on Early Man: 137-144;Oxford (Pergammon). JELiNEK,J. (1980b): Variability and geography.Contribution to our knowledgeof Europeanand north African Middle Pleisto- cenehominids.- Anthropologie (Brno), 18: 109;.114. JELINEK,J. (1981): Was Homo erectusalready Homo sapiens?Les Processusde L 'Hominisation.-C.N.R.S International Collo- quium, No. 599: 91-95. JELiNEK,J. (1982): The eastand soutileastAsian way of regionalevolution.- Anthropos (Brno), 21: 195-212. JONES,R. (1989): East of Wallace's line: issuesand problemsin tile colonizationof tile Australian continent.- In MELLARS,P., and C.B.STRlNGER(eds.): The HumanRevolution: Behaviouraland Biological Perspectiveson tile Origins of Modem Humans:743-782; Edinburgh(Edinburgh UniversityPress). KENNEDY,G.E. (1991): On tile autapomorphictraits of Homo erectus.-J.Hum. Evol., 20 (5): 375-412. KIMBEL,W.H. (1991): Species,species concepts, and hominid evolution.-J. Hum. Evol., 20 (4): 355-371. KIMBEL,W.H., and Y.RAK (1992): The importanceof speciestaxa in paleoanthropologyand an argumentfor tile phylogenetic conceptof the speciescategory.-1n Kl\.1BEL, W.H., and L.B.MARTIN(eds.): SpeciesConcepts and Primate Evolution.- s New York (Plenum). . s KLEIN,R.G. (1990): The HumanCareer. Human Biological and CulturalOrigins.- Chicago(University of ChicagoPress). LEAKEY,R.E. (1989): Recent fossil finds from EastAfrica.-1n DURANT,J.R. (ed.): Human Origins: 53-62; Oxford (Clarendon Press). LEVINTON, J.S., and C.M.SIMON (1980): A critique of the pqnctuated equilibria model and implications for the detection of speciation in the fossil record.- Systematic Zoology, 29: 130-142. LEWIN,R. (199Ia): Neanderthals puzzle the anthropologists.- New Scientist (20 April): 27. LEWIN, R. (1991b): La Naissance de 1 'Anthropologie Moleculaire.- La Recherche, 22: 1242-1251. LEWONTIN,R.C. (1974): The Genetic Basis of Evolutionary Change.- New York (Columbia University Press). 10 ZUN'E (1989): Mosaic evolution of Jinniushan archaic Homo sapiens -ill KRANTz, G., and Ho CHUANKUN(eds.): Human Variation in the Pacific Region.- Proceedings of the Circum-Pacific Prehistory Conference, Seattle. MAYR, E. (1951): Taxonomic categories in fossil hominids.- Cold Spring Harbor Symposia on Quantitative Biology, 15: 108- n 118. MAYR, E. (1963): Animal Species and Evolution.- Cambridge (Belknap Press). MAYR, E. (1968): The role of systematics in biology.- Science, 159: 595-599. MAYR, E. (1969): Principles of Systematic Zoology.- New York (McGraw-Hili ). MERCIER, N., VALLADAS, H., JORON,J-L.,REYss, J.-L., LEvEQUE, F., and B. VANDERlvffiERSCH(1991): Thermoluminescence dating of the late Neanderthal remains from Saint-Cesaire.- Nature, 351: 737-739. PATrERSON,H.E.H. (1985): The recognition concept of species.- ill VRBA, E. (ed.): Srecies and Speciation: 21-29; Pretoria (Transvaal Museum Monograph, 4 ). POPE,G.G. (1988): Recent advances in Far Eastern Paleoanthropology.- ill SIEGEL,B..1., BEALS, A.R., and S.A.TYLER (eds.): Annual Review of Anthropology.- Annual Reviews (Palo Alto), 17: 43-77. POPE,G.G. (1991): Evolution of the zygomaxillary region in the genus Homo and its relevance to the origin of modern hu- mans.-J. Hum. Evol., 21 (3): 189-213. POPE,G.G. (1992): The craniofacial evidence for the emergence of modern humans in China.- Yb. Phys. Anthropol., 35: 243- 298. RIGmMIRE, G.P. (1988): Homo erectus and later Middle Pleistocene humans.- ill SIEGEL,B.J., BEALS, A.R., and S.A.TYLER (eds.): Annual Review of Anthropology.- Annual Reviews (Palo Alto), 17: 239-259. RIGmMIRE, G.P. (1990): The Evolution of Homo erectus. Comparative Anatomical Studies of an Extinct Human Species.- Cambridge (Cambridge University Press). RIGmMIRE, G.P. (1991): The dispersal of Homo erectus from Africa and the emergence of more modern humans.- J. Anthro- pol. Res., 47 (2): 177-191. ROBINSON,J. T. (1967): Variation and taxonomy of the early hominids.- Evolutionary Biology, 1: 69-100. ROUHANI,S. (1989): Molecular genetics and the pattern of human evolution: plausible and implausible models.- ill MELLARs, P., and C.B.S1RINGER(eds.): The Human Revolution: Behavioural and Biological Perspectives on the Origins of Mod- ,n ern Humans: 47-61;Edinburgh (Edinburgh University Press). m SANTA LocA, A.P. (1980): The Ngandong Fossil Hominids.- Yale Univ. Publ. Anthropol., 78; New Haven (Yale University Press). SARICH,V.M. (1971): Human variation in an evolutionary perspective.- ill: DoLHINOW,P., andB.SARIcH(eds.): Background for Jf Man: 182-191; Boston (Little Browa). SCHWALBE,G. (1899): Studien fiber Pithecanthropus erectus Dubois.- Z. Morphol. Anthropol., 1: 16-240. SCHWALBE,G. (1904): Die Vorgeschichte des Menschen.- Braunschweig (Friedrich Viewug). m SIMPSON,G.G. (1961): Principles of Animal Taxonomy.- New York (Columbia University Press). STANLEY,S.M. (1979): Macroevolution: Pattern and Process.- San Francisco (Freeman). S1RINGER,C.B. (1981): The dating of European Middle Pleistocene hominids and the existence of Homo erectus in Europe.- Anthropos (Brno), 19: 3- 14. S1RINGER,C.B. (1984): The definition of Homo erectus and the existence of the species in Africa and Europe.- ill ANDREWS, P., and J.L.FRANZEN(eds.): The Early Evolution of Man, with Special Emphasis on Southeast Asia and Africa.- Cour. Forsch.-illstitut Senckenberg, 69: 131-143. STRINGER,C.B. (1986): The credibility of Homo habilis.- ill WOOD,B.A., and P.ANDREwS(eds.): Major Topics in Primate and Human Evolution: 266-294; Cambridge (Cambridge University Press). STRINGER,C.B. (1987): A numerical cladistic analysis for the genus Homo.- J. Hum. Evol., 16 (1): 135-146. S1RINGER,C.B., and P.ANDREwS (1988): Genetic and fossil evidence for the origin of modern humans.- Science, 239: 1263- 1268. ic S1RINGER,C.B., and R.GRON (1991): Time for the last Neanderthals.- Nature, 351: 701-702. l.- STRINGER,C.B., HOWELL, F.C., and J.MELENTIS(1979): The significance of the fossil hominid skull from Petralona, Greece.- J. Archaeol. Sci., 6: 235-253. TATrERSALL,I. (1986): Species recognition in human paleontology.- J. Hum. Evol., 15 (3): 165-175. In TATrERSALL,I. (1991): What was the human revolution?- J. Hum. Evol., 20 (I): 77-83. THOMA,A. (1962): Le deploiement evolutifde I'Homo sapiens -Anthropologica Hungarica, 5: I-III.

m

Wu RUKANG,and LIN SHENGLONG(1983): Peking man.- Scientific American, 248 (6): 86-94. Wu RUKANG,and Wu XINZffi (1982): Comparison of Tautavel man with Homo erectus and early Homo sapiens in China.- In L 'Homo erectus et la place de I'Homme de Tautavel parmi les Homindes Fossiles, 2: 605-616; Nice (Louis-Jean Sci- entific and Literary Publications). In.:S, Wu XINzm (1981): A well-preserved cranium ofan archaic type of early Homo sapiens from Dali, China.- Scientia Sinica, 24 (4): 530-41. Wu XINZffi (1988): Comparative study of early Homo sapiens from China and Europe.- Acta Anthropol. Sin., 7 (4): 287-293. he Wu XINZffi, and Wu MAOLIN (1985): Early Homo sapiens in China.- In Wu RUKANG,and J.W.OLSEN(eds.): Palaeoanthro- pology and Paleolithic Archaeology in the People's Republic of China: 91-106; New York (Academic Press). ZHANG YINYUN (1984): Human fossil newly discovered at Chaoxian, Anhui.- Acta Anthropol. Sin., 4: 305-310. ZHANG YINYUN (1986): The problems caused by dating of the Homo erectus and early Homo sapiens in China. Treatises in 'd- Commemoration of the 30th Anniversary of the Discovery of Maba cranium: 127-132; Beijing (Chinese Academy of Sciences). ~r- ZHANG YINYUN (1991): An examination of temporal variation in the hominid dental sample from Zhoukoudian Locality 1.- Acta Anthropol. Sin., 10 (5): 85-95.

The authors' addresses:Prof. Dr. Milford H. Wolpoff, PaleoanthropologyLaboratory, Department of Anthropol- ogy, University of Michigan, Ann Arbor, MI 48109-1382,USA; Prof. Dr. Alan G. Thome, Departmentof Prehis- tory, ResearchSchool of Pacific Studies,Australian National University, Canberra2601, Australia; Prof. Dr. Jan Jelinek, Anthropos Institute, Moravian Museum,Bmo, CzechRepublic; Dr. Zhang Yin}lln, Institute for Verte- brate Paleontologyand Paleoanthropology,Beijing 100044,People's Republic of China. , n-

9-

id 5;

!J1 Ie

:le ~

~i-

ill A

th

iD.