Evolutionary Anthropology 195

ARTICLES

The Changing Face of

BERNARD WOOD AND MARK COLLARD

Since its introduction in the 18th century, the genus Homo has undergone a contemporary Europeans. Interest- number of reinterpretations, all of which have served to make it a more inclusive ingly, these additions to the genus taxon. In this paper, we trace this trend towards greater inclusiveness, and explain Homo were made four before how it has affected the way Homo is defined. We then demonstrate that the current the first fossils were allocated to H. criteria for identifying species of Homo are difficult, if not impossible, to operate using sapiens.4 paleoanthropological evidence. We discuss alternative, verifiable, criteria, and show The inclusion of the that when these new criteria are applied to Homo, two species, and skeleton within Homo expanded the Homo rudolfensis, fail to meet them. We contend that the lowest boundary of Homo ranges of both the cranial and postcra- should be redrawn so that the earliest species in the genus is , or nial morphology of the genus. The early African . The appearance of this species at around 1.9 Myr morphology of the specimen, to- appears to mark a distinct shift in hominin adaptive strategy involving morphological gether with evidence gleaned from dis- and behavioral innovations. coveries made prior to 1859 and there- after in Western Eurasia, show that Neanderthal crania differ from those The genus Homo was established by CHANGING INTERPRETATIONS of living and fossil H. sapiens in sev- Carolus Linnaeus1 as part of the bino- eral respects.5,6 Typically, they have The process of relaxing the criteria discrete and rounded supraorbital mial system he introduced in the tenth for allocating species to Homo has ridges, faces that project anteriorly in edition of his Systema Naturae. As proceeded in a series of steps. Some of the midline, laterally projecting and conceived by Linnaeus, the genus these coincided with the recognition rounded parietal bones, a rounded, Homo subsumed two species; the name of new fossil species. Others resulted posteriorly projecting occipital bone, Homo sapiens was attached to the from the discovery of previously un- a derived nasal morphology, large inci- more diurnal of the two. Within H. known features of a species or from sor teeth, and postcanine teeth with sapiens, Linnaeus recognized six reinterpretations of existing evidence. large root canals. Their brains were as groups. Four of these are geographical large as, if not larger, than those of variants drawn from the four conti- modern . Postcranial peculiari- nents, Africa, America, Asia and Eu- Homo neanderthalensis ties include limb bones with stout rope, that were known to Linnaeus. The first step in the process of broad- shafts and relatively large joint sur- The other two groups, namely the wild ening the scope of Homo was William faces, especially well-marked areas for and the monstrous men, are of socio- King’s2 decision to recognize a new the attachment of a muscle that helps historical rather than biological inter- Homo species for the Neanderthal to control movement at the shoulder, est. The same is true of the second specimen. He considered naming a and an elongated pubic ramus of the species, also called Homo sylvestris, new genus for the Feldhofer skeleton, . Despite the latter trait, there is Homo troglodytes which is part myth but eventually decided that it was suf- no indication that the and part orangutan. ficiently similar to H. sapiens to war- were anything other than upright, ob- rant its inclusion in Homo as H. nean- ligate, bipeds. derthalensis King, 1864. In the same Deliberate burials or rock falls have , George Busk3 reported to the resulted in an unusually good sample British Association for the Advance- of immature Neanderthals. These Bernard Wood is at the Department of ment of Science on what we now show that the peculiar cranial and Anthropology, George Washington Univer- know to be a Neanderthal cranium postcranial morphology of the Nean- sity, 2110 G Street NW, Washington, DC from Gibraltar. Although Busk ac- derthals occurs in the skeletal remains 20052, USA, and Origins Pro- gram, National Museum of Natural History, knowledged the strength of the resem- of very young children as well as Smithsonian Institution, Washington, DC blance between the Gibraltar cranium adults.7,8 This is evidence that these 20560, USA. E-mail: [email protected]. and the one from the Neanderthal features are under genetic control. Mark Collard is at the Department of An- cave, he judged the former to belong Thus they cannot be dismissed as be- thropology, University College London, to H. sapiens, albeit a member of the ing behaviorally induced and, in conse- Gower Street, London WC1E 6BT, UK. E-mail: [email protected]. species that was more similar to living quence, taxonomically and phyloge- Tasmanians and Australians than to netically irrelevant. 196 Evolutionary Anthropology ARTICLES

Box 1. Principles and Practice of Nomenclature

‘‘. . . biological nomenclature has to be an exact tool that will convey a precise meaning for persons in all generations.’’ J. Chester Bradley in the Preface to the 1st edition of the International Code of Zoological Nomenclature129 The need to have conventions for example, Article 31, subsection (a) (ii) and E, provide useful recommenda- the application and use of names for (i.e. 31, (a) (ii)) sets out the procedure tions about matters such as the choice species groups has been recognized to be followed for choosing the ending of title and whether illustrations are since Linnaeus1 introduced the bino- of a new species name depending on necessary, but these are recommenda- mial system in 1758. The first formal whether the person to whom the name tions and need not be followed. Thus, code to regulate zoological nomencla- refers is male or female or, when there for example, researchers are not re- ture was devised by Hugh Strickland, is more than one person a mixture of quired to give the etymology of a and both Charles Darwin and Richard males and females. This may seem species name nor to specify a type Owen were on the Committee that arcane, but the provisions of the Code specimen. However, the original de- presented it to the British Association do bring uniformity to a system that scription is the only opportunity to for the Advancement of Science in could easily descend into chaos. It specify a type specimen. Subse- 1842. Entitled a ‘‘Series of Proposi- must be emphasized that the Code quently, the only way to identify a tions for Rendering the Nomenclature has both ‘‘rules’’ and ‘‘recommenda- replacement for the type is to select of Zoology Uniform and Permanent,’’ it tions.’’ If the rules are not followed, one of the original specimens as the was widely adopted in France, Italy, then a nomenclatural act may be over- ‘‘lectotype.’’129,133 and the United States. The code was turned, but the failure to follow recom- The Code recommends, but does further refined by discussions, at and mendations cannot invalidate a pro- not stipulate, that the binomial should between, successive International posal. be followed by the name of the de- Congresses of Zoology in 1889, 1895, When a new species is established, scriber who has priority and the date and 1898. The direct descendant of the correct sequence must be fol- of publication: for example, Sinanthro- Strickland’s ‘‘Propositions’’ was the lowed; in the examples given below, pus pekinensis Black, 1927.134 If sub- ‘‘Re`gles internationales de la Nomen- the relevant articles and subsections sequent research suggests that the clature zoologique,’’ published in 1905. are given in parentheses. The pro- species should be subsumed into a This code remained in force until 1961, posed name must be available and it genus with priority, then the original when an international commission pub- must be published. The name must be author’s name is put in parenthesis, lished the first edition of the Interna- a word of more than one letter or a and the name of the reviser is added. tional Code of Zoological Nomencla- compound word (11, (h), (i)), that is Thus, Sinanthropus pekinensis Black, ture, hereafter referred to as the Code. capable of being rendered into Latin 1927 effectively became Pithecanthro- In 1973, responsibility for the Commis- (11, (b), (iii)). It must not have been pus pekinensis (Black, 1927) von sion and the Code was transferred used for another taxon, and Article 13, 16 from the organizers of the Interna- (a), (I) stipulates that its publication Koenigswald and Weidenreich, 1939 tional Zoological Congresses to the must include ‘‘a description or defini- when the latter authors reduced the International Union of Biological Sci- tion that states in words characters differences between the two genera to ences in 1973, which sponsored the that are purported to differentiate the no more than the differences between third and most recent edition of the taxon.’’ The Code does not specify ‘‘two different races of present man- 19,20 Code in 1985.130 how comprehensive these descrip- kind’’ (p. 928). First Mayr, and then 15 The purpose of the Code130 is to tions should be, and this imprecision later Le Gros Clark, ceased to distin- ‘‘promote stability and universality in leads to potential difficulties: one re- guish between Pithecanthropus erec- the scientific names of and to searcher may deem a description to tus from Indonesia and Pithecanthro- ensure that the name of each taxon is be satisfactory, while another may pus pekinensis from Zhoukoudian, and unique and distinct’’ (p. 3). Its basic judge it to be inadequate. Disagree- subsumed Pithecanthropus into Homo. principle is priority, in which normally ment about whether or not Alexeev131 Thus, because of the priority it enjoys, ‘‘the valid name of a taxon is the oldest did this adequately for Homo rudolfen- the species Anthropopithecus erectus available name applied to it’’ (pp. 262– sis lies at the heart of a recent ex- Dubois, 1892, which two years later 263). The Code is effective but cum- change about hominin nomencla- became Pithecanthropus erectus bersome to use. It is made up of 88 ture.132,133 A ‘‘publication’’ is defined (Dubois, 1892) Dubois, 1894135 pro- Articles that are given Arabic numbers very precisely in the Code. It must be vides the species name Homo erectus (1–88) and five Appendices (A–F). public, intended for the scientific re- (Dubois, 1892) Mayr, 1944 for the Each article has subsections that are cord, and freely available, either gratis new, amalgamated taxon. Although identified by lower-case letters, and or for purchase, but not by private this is the full, formal, name of the these sometimes are broken down subscription or in a limited addition. All species, authors normally use the into sub-subsections using either Ara- copies of the publication must be iden- shortened version, Homo erectus (Du- bic or lower-case italic numbers. For tical. Two appendices to the Code, D bois, 1892). ARTICLES Evolutionary Anthropology 197

Homo heidelbergensis thropus erectus Dubois, 1892 should Dubois, 1894. The next taxonomic revi- be transferred to the genus Homo, and sion was signaled when von Koenig- The second step in the process of was completed by Le Gros Clark15 swald and Weidenreich16 concluded relaxing the criteria for including fos- exactly a hundred years after H. nean- sil evidence within Homo was the addi- that Sinanthropus and Pithecanthro- derthalensis was incorporated into tion of Schoeten- pus differed no more than ‘‘two differ- Homo. Until the was ratio- sack, 1908. The type specimen of H. ent races of present mankind’’ (p. 928). nalized, the main subsets of the hy- heidelbergensis is a found in This conclusion was formalized four podigm of what is now referred to as 17 1907 during excavations to extract years later when Weidenreich sank H. erectus were attributed to five gen- sand from a quarry at Mauer, near Sinanthropus into Pithecanthropus. era. Two of these, Pithecanthropus and Heidelberg, Germany.9 The next evi- Just over a decade later in the first Meganthropus von Koenigswald, 1950, dence within Europe came in 1933 edition of The Fossil Evidence for Hu- were known from Java, and one, Sinan- 18 from a gravel pit at Steinheim, Ger- man Evolution, Le Gros Clark placed thropus Black, 1927, from China. Two many, but in the meantime evidence both Sinanthropus Black, 1927 and had been found in Africa (Kabwe, Meganthropus von Koenigswald, 1950 1921 and 1925). These remains and into Pithecanthropus (Dubois, 1892). others like them were initially labeled . . . the hypodigm of H. In the meantime, however, Mayr had as ‘‘archaic’’ H. sapiens, but now they proposed that Pithecanthropus,19 are being increasingly often referred erectus has had a Meganthropus,20 and Telanthropus20 be to as H. heidelbergensis.10–13 complicated transferred to Homo. Lastly, in the The brain cases of H. heidelbergensis nomenclatural history. second edition of his text, Le Gros are often, but not always, smaller than Clark15 proposed that Atlanthropus as those of modern humans (for ex- The type specimen was well as Pithecanthropus be transferred ample, Steinheim), but they are al- originally referred to as to Homo. ways more robustly built, with large Compared with H. sapiens, H. nean- ridges above the orbits and a thick- Anthropopithecus derthalensis, and H. heidelbergensis, ened occipital region. The Mauer man- erectus Dubois, 1892, but fossils attributed to H. erectus have a dible has no chin, and the corpus is smaller neurocranium, a lower vault, substantially larger than those of mod- was transferred two a broader base relative to the vault, ern Europeans. Postcranially, the years later to and more complex premolar roots. shapes of the limb bones are much like They also have a substantial, essen- those of H. sapiens except that the Pithecanthropus erectus tially continuous torus above the or- shafts of the long bones are generally Dubois, 1894. The next bits, behind which is a sulcus. There is thicker, with higher robusticity indi- taxonomic revision was usually a sagittal torus and an angular ces. They have some, but not all, of the torus that runs toward the mastoid features of H. erectus crania, but lack signaled when von process. The occipital region is sharply the derived features of Neanderthal Koenigswald and angulated, with a well-marked supra- crania. toral sulcus, and the inner and outer If there is to be a single species Weidenreich concluded tables of the vault are thickened. De- name to cover the archaic material spite the relatively large number of 12,13 that Sinanthropus and from Europe, Africa and Asia, then crania that had been recovered from the name with priority is H. heidelber- Pithecanthropus differed Java, China, and elsewhere, relatively gensis Schoetensack, 1908. However, no more than ‘‘two little was known about the postcranial if there was evidence that the two different races of present morphology of what was to become H. other regions sampled equally good erectus. Discoveries from East African species, then the name that would mankind...’’ sites provided crucial evidence in the have priority for the African species form of a pelvis and femur from Oldu- would be H. rhodesiensis Woodward, vai Gorge (OH 28), two fragmentary 1921. Further, if the Ngandong mate- partial skeletons from East Turkana rial is not to be included in H. erectus African genera, Telanthropus Broom (KNM-ER 803 and 1808), and the un- (see below), the appropriate name for and Robinson, 1949 from southern usually well-preserved skeleton from a distinctive Asian ‘‘archaic’’ Homo Africa and Atlanthropus Arambourg, West Turkana (KNM-WT 15000). The species would be H. soloensis Oppe- 1954 from North Africa made up the cortical bone of the postcranial skel- noorth, 1932. balance. eton is generally thick. The long bones This has meant that the hypodigm are robust; the shaft of the femur is of H. erectus has had a complicated relatively flattened from front to back Homo erectus nomenclatural history. The type speci- while that of the tibia is relatively The third step in the process of men was originally referred to as An- flattened from side to side in compari- expanding the range of morphology thropopithecus erectus Dubois, 1892,14 son with those of other Homo species. within Homo began in 1944, when but was transferred two years later to These are referred to as platymeria Mayr19 recommended that Pithecan- Pithecanthropus erectus (Dubois, 1892) and platycnemia, respectively. How- 198 Evolutionary Anthropology ARTICLES ever, all the postcranial elements are among them KNM-ER 1590, 1802, subsumes a wide range of cranial mor- consistent with a habitually upright 1805, 1813, and 3732, that were infor- phology, with endocranial volumes posture and . mally referred to as ‘‘early Homo.’’ The ranging from just less than 500 cm3 to Some workers have suggested that morphology of KNM-ER 1470 showed about 800 cm3. The also the ‘‘early African’’ component of H. the unique combination of an ‘‘ad- vary in size, with those from the larger erectus is significantly more primitive vanced’’ neurocranium with a ‘‘primi- individuals having robust bodies and than is H. erectus sensu stricto and tive’’ australopith-like face. The pres- premolar teeth with complex crowns have proposed that this subset of the ence of these two morphologies in the and roots. hypodigm be assigned to a separate same cranium posed a difficulty for The discovery of OH 62 was particu- species, H. ergaster Groves and Maza´k, researchers. Which was the homo- larly significant in connection with the 1975.10,11,21–24 However, not all re- plasy, the large brain or the large, postcranial skeleton of H. habilis. Al- searchers accept the need to subdivide broad face? Alone among the early though the preservation of this speci- the H. erectus hypodigm in this commentators, Walker33 cautioned men is poor, enough of the skull is way.25–30 that KNM-ER 1470 may represent a preserved to allow virtual certainty large-brained . Most that the partial skeleton does not be- Homo habilis researchers chose the face as the site long to boisei. Thus, un- of homoplasy34,35 and argued that the less it is the first evidence from Bed I The fourth step in the process of of a novel taxon, OH 62 must belong to relaxing the criteria for allocating spe- H. habilis, the only other hominin cies to Homo came in 1964, when species known from that time range at Leakey, Tobias, and Napier, set out the The morphology of Olduvai Gorge. Although several iso- case for recognizing a new species for KNM-ER 1470 showed the lated postcranial specimens from Bed the ‘‘gracile’’ hominid remains recov- I had been attributed to H. habilis,31 it ered at Olduvai Gorge from 1960 on- unique combination of was subsequently pointed out that it is ward. Even more contentiously, they an ‘‘advanced’’ at least equally likely that this postcra- proposed that the new species should nial evidence belongs to P. boisei.47 If be accommodated within the genus neurocranium with a the logic we have set out is followed, Homo as H. habilis Leakey, Tobias, the discovery of OH 62 provided the 31 ‘‘primitive’’ australopith- and Napier, 1964. The addition of H. first unequivocal postcranial evidence habilis to the genus Homo meant that like face. The presence of H. habilis. It was all the more signifi- 18 Le Gros Clark’s diagnosis of Homo of these two cant, therefore, that OH 62 apparently needed amendment. This involved had limb proportions that were at 31 Leakey and coworkers relaxing some morphologies in the least as ape-like as those of individuals criteria such as brain size, so that the same cranium posed a attributed to Australopithecus afaren- relatively small-brained (600 to 700 sis.37,48 If the associated skeleton 3 difficulty for researchers. cm ) crania from Olduvai could be KNM-ER 3735 from also included. It was claimed that other Which was the proves to belong to H. habilis,36,49,50 criteria, such as dexterity, an erect homoplasy, the large this will provide further evidence of posture, and a bipedal gait, did not that taxon’s postcranial morphology. need to be changed because Leakey, brain or the large, broad 31 Tobias, and Napier’s interpretation face? of the functional capabilities of the H. Changing Functional habilis remains from Olduvai was such Interpretations of H. habilis that the type specimen and the para- types complied with these functional The final contribution to the increas- large neurocranium allied the new cra- criteria.31 Ultimately, new discoveries ing inclusiveness of Homo came as the nium with Homo. As a consequence, and the reinterpretation of existing result of reassessing the functional from 1973 onward, the genus Homo evidence has led others to offer rather implications of the original postcra- subsumed a substantially wider range different functional assessments of the nial remains attributed to H. habilis 36 same material. of facial and basicranial morphology from Olduvai Gorge. The type and than it did prior to the discovery of paratypes of H. habilis included fossil KNM-ER 1470. evidence from both the forelimb (OH Early Homo In due course, important additional 7) and the hindlimb (OH 8, 10, and The fifth step in the process of chang- specimens from Koobi Fora (for ex- 35).31 Some investigators have argued ing and relaxing the criteria for allocat- ample, KNM-ER 1590, 1802, 1813, that OH 8 and 35 are from the same ing species to the genus Homo began and 3732)36 and Olduvai Gorge (OH individual,51 but an analysis of the in 1972 with the discovery of KNM-ER 62)37 were added to the early Homo shapes of the reciprocal joint surfaces 1470. Although this specimen was not hypodigm, as was fossil evidence from suggests otherwise.52 The initial assess- formally assigned to H. habilis (ini- Members G and H of the Shungura ment of the functional implications of tially it was attributed to Homo sp. Formation,38–40 Member 5 at Sterkfon- the evidence from the leg and foot indet.32), it was the first in a sequence tein,41,42 and Member 1 at Swart- stressed the ways in which the Olduvai of fossil discoveries at Koobi Fora, krans.43–46 This additional material material resembled H. sapiens,53 but ARTICLES Evolutionary Anthropology 199

Box 2. Species Groups Within Homo

The following taxa have been pro- King, 1864.2 , western Species Homo ergaster Groves and posed as component species of the Eurasia. Type: , adult Maza´k, 1975.136 Plio-Pleistocene, Af- genus Homo; temporal and geographic calotte and partial skeleton. Found at rica and possibly Eurasia. Type: distribution and the type specimen are Neanderthal, Elberfield, Germany, 1856. KNM-ER 992, adult mandible. Found given for each taxon. Species Homo erectus (Dubois, in Okote Member, Koobi Fora Forma- Genus Homo Linnaeus, 17581 [in- 1892)14 Mayr, 1944.19 Pleistocene, Af- tion, Area 3 at Koobi Fora, East Tur- cludes, for example, Pithecanthropus rica and Eurasia. Type: Trinil 2, adult kana, NFD, Northern Kenya, 1971. Dubois, 1894; Protanthropus Haeckel, calotte. Found at Trinil, Ngawi, 1891. Species Homo rudolfensis (Alex- 1895; Sinanthropus Black, 1927; Species Homo heidelbergensis eev, 1986)131 sensu Wood, 1992.23 Cyphanthropus Pycraft, 1928; Megan- Schoetensack, 1908.9 Pleistocene, Af- Pliocene, East Africa. Type: KNM-ER thropus Weidenreich, 1945; Atlanthro- rica and Eurasia. Type: Mauer 1, adult 1470, adult cranium. Found in the pus Arambourg, 1954; Telanthropus mandible. Found at Mauer, Heidel- Upper Burgi Member, Koobi Fora For- Broom and Robinson, 1949]. Plioce- berg, Germany, 1907. mation, Area 131 at Koobi Fora, East ne–present, world-wide. Species Homo habilis Leakey, To- Turkana, NFD, Northern Kenya, 1972. Species Homo sapiens Linnaeus, bias and Napier, 1964.31 Pliocene, Species Bermu- 1758.1 Pleistocene–present, world- Africa. Type: OH 7, partial calotte and dez de Castro et al., 1997.137 Pleis- wide. There is no designated type hand bones. Found at site FLKNN 1 in tocene, Europe. Type: ATD6-5—man- specimen for H. sapiens. Bed I at Olduvai Gorge, Serengeti, dible and associated teeth. Found at Species Homo neanderthalensis Tanzania, 1960. Gran Dolina, Atapuerca, Spain, 1994.

authors who have considered these ape-like carpus and phalanges, to a in order to accommodate H. habilis. specimens in more detail were more thumb that some have interpreted as The second criterion is the possession cautious. For example, they stressed being compatible with pulp-to-pulp of language. Tobias,67 in particular, that the knee was imperfectly adapted opposition.60–63 has championed the link between to bipedalism54 and that the foot may Homo and the ability to communicate not have been from an individual ca- through spoken language. He based pable of modern human-like striding HAVE THE LIMITS OF HOMO BEEN this primarily on evidence from endo- bipedalism.55 However, in their re- cranial casts and wrote that ‘‘in the vised diagnosis of the genus Homo, STRETCHED TOO FAR? endocranial casts of H. habilis, for the Leakey and coworkers31 were clear Depending on the type of taxonomic first time in the early hominid fossil that inclusion in Homo implied ‘‘ha- interpretation adopted,64 the genus record, there are prominences corre- bitual erect posture and bipedal gait’’ Homo contains as many as seven fossil sponding to both a well-developed and, with respect to the upper limb, species: H. neanderthalensis, H. heidel- speech area of Broca and a secondary evidence of a ‘‘power grip’’ and a bergensis, H. erectus, H. habilis, H. speech area of Wernicke. These are ‘‘simple and usually well developed ergaster, H. rudolfensis, and H. anteces- two of the most important neural bases precision grip’’ (p. 7). sor. These species run the gamut of for language ability in the human With regard to the OH 8 foot, reas- morphology from the australopith- brain’’ (p. 836). H. habilis was, he sessments of its functional morphol- like postcranial skeletal morphology claimed, ‘‘the first language-depen- ogy have stressed its potential for and proportions of H. habilis to the dent ’’ (p. 840). The third crite- climbing and pointed out that it re- very different morphology and limb rion is the ability to manufacture stone tains several of the morphological fea- proportions of modern humans. Not- tools.31,67 The connection between tures seen in living nonhuman pri- withstanding the existence of compre- stone-tool manufacture and Homo is a mates.51,56–58 Researchers have hensive morphological diagnoses of longstanding one68 that Kenneth Oak- suggested that although OH 8 has the Homo,15,31,65 in practice species have ley made explicit in the content and articular mechanisms that convert the been and are being assigned to that the title of his book Man the Tool- foot into a rigid lever during the sup- genus on the basis of four criteria. The Maker.69 Although for five years Zinjan- port phase of walking,57 it apparently first criterion concerns absolute brain thropus70 was credited with being ‘‘the lacks some functional elements such size. The notion of a ‘‘cerebral rubi- oldest yet discovered maker of stone as the lateral deviation of the heel and con’’ for membership in Homo is tools’’ (p. 493), the mantle was swiftly the propulsive great toe that are pre- closely linked with Keith,66 who lo- passed on to H. habilis.31 Thereafter, sent in H. sapiens.59 Similarly, consid- cated it at 750 cm3, midway between the putative link between stone tools erations of the Olduvai OH 7 hand the highest gorilla and the lowest ‘‘ab- and Homo has been maintained by have suggested that earlier functional origine’’ endocranial volumes (pp. 205– workers such as Hill and coworkers71 interpretations may need to be revised 206). In Leakey, Tobias, and Napier’s31 and Kimbel and coworkers.72 The in the light of evidence that it displays diagnosis of Homo, the rubicon was fourth criterion for including species a mosaic of features ranging from an lowered to its present level of 600 cm3 in Homo is the possession of a modern 200 Evolutionary Anthropology ARTICLES

Figure 1. An illustration of the conflict between defining a genus using cladistic and evolution- ary systematic criteria. With which group of species should species D be classified? If it is grouped with species A, B, and C, the resulting genus would be monophyletic but not adaptively coherent, whereas if it is grouped with species E, F, and G, the resulting genus would be adaptively coherent but not monophyletic.

human-like precision grip based on a of any of the early hominins.62 Thus with the evolutionary systematic well-developed, opposable pollex.31 there is a need to rethink the criteria method of classification, states that a It is now evident that none of these we use to assign species to Homo. genus is a species or group of species criteria is satisfactory. The ‘‘cerebral of common ancestry that occupies a rubicon’’ is problematic because cra- different ecological situation or ‘‘adap- nial capacity is of questionable biologi- tive zone’’ than that occupied by the cal significance unless it is related to While it is attractive to species of another genus.95 A group of estimates of body mass.73 Likewise, link language with the species of common ancestry under there is compelling evidence that lan- appearance of the this definition can be either monophy- guage function cannot be reliably in- letic, comprising a common ancestor ferred from the gross appearance of genus Homo, there is and all its descendants, or paraphy- endocasts.74–77 Moreover, the language- little sound evidence to letic, comprising a subset of a mono- related parts of the brain are not as phyletic group. In the second defini- well-localized as earlier studies had support such a scenario. tion, which is associated with the implied. While it is attractive to link The connection between cladistic method of classification, a language with the appearance of the genus is a group of species that are genus Homo, there is little sound evi- Homo and stone-tool more closely related to one another dence to support such a scenario. The manufacture is also than they are to any species assigned connection between Homo and stone- to another genus.90 In other words, tool manufacture is also difficult to difficult to substan- under this definition a genus can only substantiate, for there is now over- tiate . . . be monophyletic; it cannot be paraphy- whelming evidence that for much of letic. the Plio-Pleistocene in East Africa These definitions differ critically hominin species were both synchronic over the classification of ancestral spe- and sympatric. The earliest stone tools, A REVISED DEFINITION OF THE cies that have different adaptive strate- which come from about 2.6–2.3 Myr GENUS HOMO gies than their descendants, as well as deposits in East Africa, were almost over terminal taxa that form a mono- certainly contemporaneous with both Systematists are currently debating phyletic group with one taxon but early Homo and Paranthropus.78–81 Fur- the definition of the genus category as share an adaptive strategy with an- thermore, functional morphological part of a wider discussion about the other taxon. For example, in Fig. 1 analyses of the hands of the early implications of the cladistic method of species A, B, and C are monophyletic hominins have either suggested that a phylogenetic reconstruction for the and adaptively similar, but their sister modern human-like grip is not re- Linnaean system of classification.89–94 taxon, species D, shares an adaptive stricted to Homo63,82–88 or indicated At present, there are two main, com- strategy with a monophyletic group that we cannot yet be certain about peting definitions of the genus cat- comprising species E, F, and G. With the potential range of precision grips egory. The first, which is associated which taxon should species D be clas- ARTICLES Evolutionary Anthropology 201 sified? Should it be grouped with spe- one taxon but which shares an adap- PHYLOGENETIC RELATIONSHIPS cies A, B, and C, in which case the tive strategy with another can be over- OF FOSSIL HOMO SPECIES resulting genus would be monophy- come by recognizing it as a monotypic Hominin phylogenetic relationships letic but not adaptively coherent? Or genus. We suggested, therefore, that a have been reconstructed using cladis- should it be grouped with species E, F, genus should be defined as a species, tic analysis for more than two de- and G, in which case the resulting or monophylum, whose members oc- 64 cades, but no published analysis has genus would be adaptively coherent cupy a single adaptive zone. included all the fossil hominins or but not monophyletic? The latter op- Cladistics is the most effective examined all the possible hypotheses tion would be selected under the evolu- method of identifying monophyletic of relationships among the species. tionary systematic definition, whereas groups of species, but how can we best determine hominin adaptive strate- Nevertheless, there is a clear consen- the former option would be favored gies? For a species to persist long sus that H. ergaster, H. heidelbergensis, under the cladistic definition. enough to be sampled in the hominin and H. neanderthalensis are more The evolutionary systematic defini- fossil record, individuals must be able closely related to H. sapiens than they tion of the genus category is rejected to maintain themselves in homeosta- are to australopith genera. The rela- by cladists because they do not con- sis despite fluctuations in ambient tem- tionships of H. ergaster have been ex- sider paraphyletic taxa to be real evolu- perature, humidity, and the availabil- amined in three studies,23,36,96 all of 90 tionary units. However, defining gen- ity of water. They also have to procure which suggested that H. ergaster shares era solely on the basis of monophyly is and process sufficient food to meet a common ancestor with H. sapiens to equally problematic, for this definition their minimum requirements for en- the exclusion of Australopithecus or does not specify how many species ergy and essential nutrients. Last, they Paranthropus. Notably, in a bootstrap should be included in a genus. Be- must be able to mate in order to reanalysis using PAUP (1,000 replica- cause all species are related to one produce offspring. The ways in which tion, heuristic, TBR)97 of Strait, Grine, another, a genus defined on the basis and Moniz’s96 character-state data ma- of monophyly could comprise, depend- trix a (H. ergaster, H. sapiens) clade ing on personal preference, between was recovered in 98% of the repli- three species and all species that have Cladistics is the most cates.64 ever existed. Of course it would be effective method of The only exhaustive examination of possible for systematists to agree, for identifying monophyletic the relationships of H. heidelbergensis example, that a genus should com- among the has suggested prise no more than two species and groups of species, but that H. heidelbergensis is the sister their common ancestor, a tribe no how can we best taxon of modern H. sapiens to the more than two genera, a family no exclusion of all the other taxa in the more than two tribes, and so on. Pro- determine hominin sample.98 A bootstrap reanalysis of the vided sister taxa were granted equiva- adaptive strategies? character-state data matrix using PAUP lent rank, such a system would yield (options as above) found a sister group objectively defined taxa. However, it relationship between H. heidelbergen- would also demand an impracticably sis and H. sapiens in 100% of the 64 large number of categories. Moreover, the members of a hominin species replicates. In line with this, when MacClade99 was used to alter the topol- because such a phylogenetic classifica- meet these fundamental requirements ogy of the most parsimonious tree so tion, by definition, imparts information comprise that species’ adaptive strat- that H. heidelbergensis was the sister only about descent, it ignores the equally egy. Thus, if H. rudolfensis, H. habilis, taxon of the (A. africanus, H. rudolfen- important modification component of H. ergaster, H. erectus, H. heidelbergen- sis, Paranthropus) clade, the tree length evolution.94 sis and H. neanderthalensis have been increased markedly from 297 to 335, 64 allocated to the correct genus, two We have suggested elsewhere that and the consistency index (CI) de- conditions must be met. First, cladis- the evolutionary systematic definition clined from 0.66 to 0.58.64 tic analyses should indicate that the of the genus category should be modi- No cladistic analysis has included species are more closely related to H. fied so that paraphyletic taxa are inad- H. neanderthalensis and examined all missible. Because the cladistic method- sapiens than they are to the australopi- the possible hypotheses regarding rela- ology cannot distinguish between ths (the shorthand we will use for the tionships among the in-group taxa. ancestor-descendant and sister-group species within Australopithecus, Paran- Nevertheless, the number of almost relationships, there is no way of recog- thropus, and ). Second, certainly derived cranial and postcra- nizing ancestors empirically. In prac- functional analyses should indicate nial similarities between H. neander- tice, therefore, the problem of how to that the strategies used by the fossil thalensis and H. sapiens is such that it classify an ancestral species having an Homo species to maintain homeosta- is highly unlikely that H. neandertha- adaptive strategy that differs from that sis, acquire food, and produce off- lensis is more closely related to the of its descendants simply does not spring are more similar to the strate- species assigned to australopith gen- arise. On the other hand, the problem gies used by the H. sapiens than they era than it is to H. sapiens. The close of how to classify a terminal species are to the strategies employed by the relationship between the Neander- that forms a monophyletic group with australopiths. thals and modern humans is reflected 202 Evolutionary Anthropology ARTICLES in the recent debate as to whether the a common ancestor with H. sapiens to the species assigned to Australopithe- Neanderthals should be considered a the exclusion of Australopithecus or cus or Paranthropus (too little is cur- separate species or included within H. Paranthropus. Moreover, even in the rently known about Ardipithecus to sapiens as a subspecies.10,11,100 cladograms in which H. habilis and H. include it in these analyses). However, It is also clear from the analyses that rudolfensis are grouped with the other H. habilis and especially H. rudolfensis H. erectus shares a common ancestor Homo species, the links are weak. For are as likely to form clades with Austra- with H. sapiens to the exclusion of the example, Wood36 found that a clado- lopithecus or Paranthropus as they are australopith genera, although the rela- gram in which H. rudolfensis is the with H. sapiens.64 tionship is possibly less reliable than sister taxon of Paranthropus was only those linking H. ergaster, H. heidelber- one step longer than the cladogram in gensis, and H. neanderthalensis to H. which Homo is monophyletic. Like- ADAPTIVE STRATEGIES OF FOSSIL sapiens. The relationships of H. erec- wise, a reanalysis of Strait, Grine, and HOMO SPECIES Moniz’s96 data using MacClade indi- tus within Hominini have been ana- Many aspects of a primate’s pheno- 23,98 cated that their favored cladogram is lyzed in two studies, both of which type help it maintain homeostasis, ac- only two steps shorter than one in suggest that H. erectus is more closely quire food, and produce offspring. which H. rudolfensis is the sister taxon related to H. sapiens than it is to However, not all of these aspects can of Paranthropus and only three steps Australopithecus or Paranthropus. be reliably reconstructed from the fos- shorter than a cladogram in which H. However, a bootstrap reanalysis (1,000 sil record. Arguably, the most impor- replication, heuristic, TBR) of the sec- tant of those that can be determined ond study’s data matrix using PAUP using palaeontological evidence are indicated that the clade linking H. . . . neither H. habilis nor body size and shape, locomotor behav- erectus and H. sapiens to the exclusion ior, relative brain size, pattern of devel- of Australopithecus and Paranthropus H. rudolfensis can be opment, and the relative size of the 64 was not supported at the 50% level. assumed with any masticatory apparatus. Using these as When MacClade was used to constrain criteria, what evidence is there that H. the cladogram so that H. erectus was degree of reliability to be erectus, H. ergaster, H. habilis, H. heidel- the sister taxon of a (Paranthropus, A. more closely related to bergensis, H. neanderthalensis, and H. africanus, H. rudolfensis) clade, tree H. sapiens than they are rudolfensis share a functionally coher- length increased by just six steps to ent adaptive strategy with H. sapiens 303 and the CI declined by just 0.02 to to species allocated to rather than with the australopiths? We 64 0.64. other genera. have reviewed this evidence else- In contrast, neither H. habilis nor H. where64 and concluded that on the rudolfensis can be assumed with any basis of the information that is avail- degree of reliability to be more closely able about early hominin body size related to H. sapiens than they are to habilis is the sister taxon of Paranthro- and shape, locomotion, development, species allocated to other genera. Six pus.64 and the relative size of their mastica- studies have investigated the relation- Figure 2 depicts what we consider, tory apparatus, the hominins fall into ships of H. habilis and H. rudolfensis on current evidence, to be the best two broad groups (Box 3). within Hominini.23,36,96,98,101,102 In three The first group has a relatively low estimate of the phylogenetic relation- of the favored cladograms (Fig. 1A in body mass; a body shape that, in terms ships between the species assigned to Strait, Grine, and Moniz,96 Fig. 7.2 in of thermoregulation, is interpreted as Homo or to the other hominin genera. Wood,36 and Fig. 3 in Chamberlain being better suited to a relatively closed It should be noted, however, that there and Wood101) H. habilis and H. rudo- environment; and a postcranial skel- are other, more pessimistic assess- lfensis are more closely related to the eton that suggests a combination of ments of what can be determined other Homo species than either is to terrestrial bipedalism with proficient Australopithecus or Paranthropus. In about hominin relationships using cla- climbing. The first group also has teeth 103 the fourth cladogram (Fig. 4 in Cham- distic methods. Indeed the results of and jaws that were apparently adapted berlain and Wood101), H. habilis is the a recent study suggest that the type of to a considerably more mechanically sister taxon of a (A. africanus, Paran- morphology preserved in the hominin demanding diet than that of H. sapi- thropus, H. rudolfensis, H. erectus, H. fossil record may not be a reliable ens, and a developmental schedule sapiens) clade, and H. rudolfensis is source of information about phylog- more closely resembling that of apes 104,105 more closely related to Paranthropus eny. It is clear from Figure 2 that than modern humans. This broad than it is to H. sapiens. In the remain- the current interpretation of the genus grouping may subsume more than one ing cladogram (Fig. 1a in Lieberman, Homo does not satisfy the first condi- adaptive strategy, but the details are Wood, and Pilbeam102), H. habilis is tion for a genus, for the fossil species beyond the scope of this review. the sister taxon of H. ergaster, but that assigned to it do not unequivocally The second group has a larger body clade is more closely related to A. form a monophyletic group with H. mass, a more modern-human-like, africanus than it is to H. rudolfensis. sapiens. It is probable that H. erectus, open-habitat-adapted physique, and a The most parsimonious cladograms H. ergaster, H. heidelbergensis, and H. postcranial skeleton consistent with a do not, therefore, consistently indicate neanderthalensis are more closely re- form of locomotion similar to that of that H. habilis and H. rudolfensis share lated to H. sapiens than any of them to modern humans (terrestrial bipedal- ARTICLES Evolutionary Anthropology 203

Figure 2. Consensus cladogram showing the relationship between hominin species. Because A. afarensis is consistently placed as the basal taxon, there are grounds for removing it from Australopithecus.96

Box 3. Summary of Results of Functional Analyses Carried Out by Wood and Collard64

(H) indicates a modern human-like pattern, (A) an australopith-like pattern, and (I) an Intermediate pattern.

Body Size Body Shape Locomotion Jaws and Teeth Development Brain Size H. rudolfensis ?? ? A A A H. habilis AA A A A A H. ergaster HH H H H A H. erectus H? H H ? I H. heidelbergensis H? H H ? A H. neanderthalensis HH H H H H

ism with, in adults, a limited ability that the link between relative brain We presently know too little about for climbing). The teeth and jaws of size and adaptive zone is more com- Ardipithecus ramidus and Australopi- the second group were apparently plex than is conventionally assumed.106 thecus anamensis to be confident about adapted to a diet which, when in- With varying degrees of certainty, even an approximate assessment of gested, had similar mechanical proper- species within Australopithecus, in- their adaptive strategies. When these ties to that of H. sapiens, and its devel- cluding the recently announced Austra- observations are combined with the opmental pattern was more modern lopithecus garhi,107 and Paranthropus, uncertainty about the phylogenetic re- human-like. This group may also sub- together with H. habilis and H. rudo- lationships of H. habilis and H. rudol- sume more than one adaptive zone. lfensis, can all be assigned to the first fensis, it is clear that the species cur- It is noteworthy that relative brain group, whereas H. erectus, H. ergaster, rently assigned to Homo do not form a size does not group the fossil homi- H. heidelbergensis, and H. neandertha- monophylum whose members occupy nins in the same way. This suggests lensis can be assigned to the second. a single adaptive zone. In other words, 204 Evolutionary Anthropology ARTICLES

to generate more robust and reliable cladistic hypotheses. It is for these reasons that we have suggested else- where that, pro tem, it is most appro- priate to refer to the two species dis- placed from Homo as Australopithecus habilis and Australopithecus rudolfen- sis.64 In the case of A. habilis, this means reverting to a taxonomy that at one time or another was supported or proposed by Tobias (see Washburn112), Howell,113 Robinson,114 Pilbeam,115 and Walker,33 among others. We antici- pate that the results of future research may well justify erecting one or more new genera for these taxa in order to 132 Figure 3. Oblique views of KNM-ER 1470, the lectotype of Homo rudolfensis, and KNM-ER better reflect hypotheses about their 1813, the best-preserved cranium of Homo habilis sensu stricto.31 It has been proposed that both these taxa should be transferred from Homo to Australopithecus. These species would phylogenetic relationships and adap- thus become, respectively, Australopithecus rudolfensis (Alexeev, 1986)130 Wood and Collard, tive regimes. It may also be necessary 199964 and Australopithecus habilis (Leakey, Tobias, and Napier, 1964)31 Wood and Collard, to reassess the taxonomic homogene- 1999.64 ity of A. habilis.45,46 with the hypodigms of H. habilis and important it is to be circumspect about EVOLUTIONARY IMPLICATIONS H. rudolfensis assigned to it, the genus linking hominin species in an ancestor- One of the main objectives of paleon- Homo fails to measure up to the crite- descendant sequence on the basis of tology is the identification of adaptive ria for a genus. assumed synapomorphies.108 radiations. These are episodes of spe- As if these were not sufficient cies generation and diversification, and TAXONOMIC IMPLICATIONS grounds for taxonomic caution, there are usually associated with a shift in also is the fact that cranial and dental ‘‘grade.’’116 Some refer to the process When the twin criteria of mono- remains, those parts of the hominin as the emergence of a new adaptive phyly and adaptive coherence are ap- fossil record normally employed in zone,95,117 the species diversification plied to the fossil species presently phylogenetic analyses, may be reliable being based on the appearance of one included within Homo, it is apparent for reconstructing the phylogenetic re- or more ‘‘key innovations.’’118–120 Re- that on both counts H. neanderthalen- lationships of species and genera.91,103- searchers have argued that there is a sis, H. erectus, H. heidelbergensis, and 105,107,111 Thus, we cannot be as confi- concept in neontology, the ‘‘fitness- H. ergaster in large measure satisfy dent about the composition of hominin generating’’ or ‘‘G-function,’’121 which both criteria, whereas H. habilis and clades as we would like to be. Further, provides a model for how grades may H. rudolfensis signally fail to meet because the identification of clades is have emerged in the fossil record.122 64 them. This suggests that H. habilis one of the two components of the The fitness-generating function is and H. rudolfensis should be removed genus definition we have put forward, an equation used to calculate the fit- from Homo. decisions about higher-order taxa ness of a phenotype.123 The equation If H. habilis and H. rudolfensis are should be conservative until we can takes into account ‘‘all the fitness trade- removed from Homo, does a new ge- demonstrate that we have the means offs in terms of the costs and the nus or new genera have to be estab- lished, or is it possible to transfer the two species to an existing genus or genera? Although H. habilis is adap- tively like the australopiths, several aspects of its cranial morphology dis- tinguish it from Australopithe- cus.23,36,50,67,107,108 Likewise, although some aspects of cranial morphology align H. rudolfensis with Paranthro- pus,102 other evidence, such as enamel structure,36,109,110 suggests that, par- ticularly until associated postcranial remains of H. rudolfensis are forthcom- ing, it may be premature to incorpo- rate that species into Paranthropus. Moreover, the mosaic nature of the A. Figure 4. Lateral views of KNM-ER 3733, a representative of Homo ergaster, and KNM-ER 1813, a garhi cranial remains reemphasize how representative of Australopithecus habilis. ARTICLES Evolutionary Anthropology 205 benefits an organism receives for do- If we accept that the lower bound- way we suggest, may still be too inclu- ing its business a certain way in a ary of Homo should be redrawn to sive. Although it may be difficult to particular time and place....’’122 (p. include H. ergaster but exclude H. habi- detect in the hominin fossil record, the 204). Thus, a new ‘‘adaptive type’’ is a lis and H. rudolfensis, what is the basis emergence of a complex spoken lan- species grouping with a function that of the hypothesis that this would re- guage and the shift to the manufacture has a less severe fitness trade-off than sult in a genus that more closely corre- of precision tool-kits128 may have co- that of the group it supersedes. Rosen- sponds to one adaptive zone? The most evolved as joint key innovations that zweig, Brown, and Vincent123 provide likely candidate for a key innovation is ushered in, at least in behavioral terms, a compelling example from among the diet, for this would be consistent with a new and distinct adaptive zone. Did reptiles. In the Americas, and increas- the reduction in tooth and mandible an increasingly sophisticated hominin ingly in the Old World, true vipers are size. However, it does not seem to culture blunt the process of species being replaced by members of the pit correspond in all particulars to the generation and diversification and thus viper group, comprising rattlesnakes, type of dietary shift that is envisaged block what in other circumstances copperheads, and coppermouths. It is would have been a classic adaptive hypothesized that the reason for the radiation? success of the latter group relates to a We suggest that if the key innovation in its visual system. ACKNOWLEDGMENTS True vipers have to trade off visual genus Homo is to have Bernard Wood is supported by The acuity against spectral band width; any meaning, its lower Henry Luce Foundation, and Mark they cannot focus sharply on both boundary should mark a Collard by The Wellcome Trust. We infrared and visible light. Members of are grateful to Arthur Cain for translat- the pit viper group have overcome this distinct shift in hominin ing the 10th edition of Linnaeus’ Sys- problem by developing loreal pits for adaptive strategy tema Naturae and to A. Chamberlain the reception of light in the infrared for making his data available. We also part of the spectrum, reserving their involving morphological thank Fred Grine and Alan Bilsbor- ‘‘real’’ eyes for the reception of visible and behavioral ough for helpful comments on an ear- light. By avoiding the compromise be- lier version of this paper. We are also tween wavelength and sharpness of innovations. In other grateful to the NERC and The Lever- image, pit vipers have effected a fit- words, it should be a new hulme Trust for past support for re- ness trade-off that is superior to that adaptive type. The search that has been incorporated into of the true vipers. In other words, this study. Rosenzweig, Brown, and Vincent123 additional interpretative claim that we are observing the emer- challenge is that this gence of a new grade of vipers. REFERENCES We suggest that if the genus Homo is adaptive shift must be 1 Linnaeus C. 1758. Systema naturae. Stock- to have any meaning, its lower bound- capable of identification holm: Laurentii Salvii. ary should mark a distinct shift in 2 King W. 1864. The reputed fossil man of the Neanderthal. Q J Sci 1:88–97. hominin adaptive strategy involving in the hominin fossil 3 Busk G. 1865. On a very ancient cranium from morphological and behavioral innova- record.... Gibraltar. Rep Br Assoc Adv Sci Bath, 1864:91– tions. In other words, it should be a 92. new adaptive type. The additional in- 4 Lartet L. 1868. Une se´pultive des Troglodytes du Pe´rigord (craˆnes des Eyzies). Bull Soc Anthro- terpretative challenge is that this adap- pol Paris 3:335–349. tive shift must be capable of identifica- in the ‘‘expensive-tissue hypothesis.’’124 5 Trinkaus E. 1983. The Shanidar neanderthals. tion in the hominin fossil record and Although there is skeletal evidence New York: Academic Press. therefore must involve those parts of that the relative gut size of H. ergaster 6 Stringer C, Gamble C. 1993. In search of the neanderthals. London: Thames and Hudson. the skeleton and other hard tissues was reduced as compared to those of 7 Tillier AM. 1989. The evolution of modern that are relatively abundant in the the australopiths, H. ergaster does not humans: evidence from young Mousterian indi- hominin fossil record. It is highly un- show evidence of any substantial shift viduals. In: Mellars P, Stringer C, editors. The 64 human revolution. Edinburgh: Edinburgh Univer- likely that interpretations based on the in absolute or relative brain size. sity Press. p 286–297. fossil record are going to falsely iden- These changes do not appear until 8 Rak Y, Kimbel WH, Hovers E. 1996. On Nean- tify adaptive shifts (a false-positive much later in hominin evolution. We derthal autapomorphies discernible in Neander- 125 tal infants: a response to Creed Miles et al. J Hum result or Type II error). Indeed, it is have suggested elsewhere, as have Evol 30:155–158. 126,127 much more likely that many adaptive others, that extra-oral prepara- 9 Schoetensack O. 1908. Der Unterkiefer des shifts will escape detection by those tion of food by cooking may be the key Homo heidelbergensis aus den Sanden von Mauer who have to rely on the fossil record innovation. Others have pointed to the bei Heidelberg. Leipzig 1–67. 10 Tattersall I. 1986. Species recognition in hu- for evidence of their existence. For likelihood that a switch to the con- man paleontology. J Hum Evol 15:165–175. example, there may be bony evidence sumption of underground storage or- 11 Tattersall I. 1992. Species concepts and spe- of the viper’s loreal pit, but it would be gans such as tubers may have oc- cies identification in . J Hum difficult to unravel the details of Roz- curred around this time.126,127 Evol 22:341–349. 12 Rightmire GP. 1996. The human cranium from ensweig, Brown, and Vincent’s ex- It is also possible, indeed probable, Bodo, Ethiopia: Evidence for speciation in the ample from the fossil record. that Homo, even when redefined in the Middle Pleistocene. J Hum Evol 31:21–39. 206 Evolutionary Anthropology ARTICLES

13 Rightmire GP. 1998. Human evolution in the 37 Johanson DC, Masao FT, Eck GG, White TD, ogy of the hind foot utilizing a multivariate Middle Pleistocene: the role of Homo heidelbergen- Walter RC, Kimbel WH, Asfaw B, Manega P, analysis. J Hum Evol 31:269–291. sis. Evol Anthropol 6:218–227. Ndessokia P, Suwa G. 1987. New partial skeleton 59 Lewis OJ. 1972. The evolution of the hallucial 14 Dubois E. 1892. Palaeontologische onder- of Homo habilis from Olduvai Gorge, Tanzania. tarsometatarsal joint in the Anthropoidea. Am J zoekingen op Java. Versl Mijnw, Batavia 3:10–14. Nature 327:205–209. Phys Anthropol 37:13–34. 15 Le Gros Clark WE. 1964. The fossil evidence 38 Howell FC, Coppens Y. 1976. An overview of 60 Susman RL, Creel N. 1979. Functional and for human evolution: an introduction to the study the from the Omo succession, Ethio- morphological affinities of the subadult hand of palaeoanthropology, 2nd ed. Chicago: Univer- pia. In: Coppens Y, Howell FC, Isaac G, Leakey (OH 7) from Olduvai Gorge. Am J Phys Anthropol sity of Chicago Press. REF, editors. Earliest man and environments in 51:311–332. the Lake Rudolf Basin. Chicago: University of 16 von Koenigswald GHR, Weidenreich F. 1939. 61 Susman RL, Stern JT, Jungers WL. 1984. Chicago Press. p 522–532. The relationship between Pithecanthropus and Arboreality and bipedality in the Hadar hominids. Sinanthropus. Nature 144:926–929. 39 Boaz NT, Howell FC. 1977. A gracile hominid Folia Primatol 43:113–156. cranium from Upper Member G of the Shungura 17 Weidenreich F. 1943. The skull of Sinanthro- 62 Marzke MW. 1997. Precision grips, hand mor- Formation, Ethiopia. Am J Phys Anthropol 46:93– pus pekinensis: A comparative study of a hominid phology, and tools. Am J Phys Anthropol 102:91– 108. skull. Palaeontol Sinica series D 10:1–484. 110. 40 Coppens Y. 1980. The differences between 18 Le Gros Clark WE. 1955. The fossil evidence 63 Susman RL. 1998. Hand function and tool Australopithecus and Homo; preliminary conclu- for human evolution, 1st ed. Chicago: University behavior in early hominids. J Hum Evol 35:23– of Chicago Press. sions from the Omo Research Expeditions stud- ies. In: Ko¨nigsson LK, editor. Current argument 46. 19 Mayr E. 1944. On the concepts and terminol- on early man. Oxford: Pergamon Press. p 207– 64 Wood B, Collard M. 1999. The human genus. ogy of vertical subspecies and species. National 225. Science. 284:65–71. Research Council Committee on Common Prob- 65 Robinson JT. 1968. The origin and adaptive lems of Genetics, Paleontology and Systematics 41 Hughes AR, Tobias PV. 1977. A fossil skull radiation of the . In: Kurth G, Bulletin No. 2, p 11–16. probably of the genus Homo from Sterkfontein, Transvaal. Nature 265:310–312. editor. Evolution and hominisation, 2nd ed. Stut- 20 Mayr E. 1950. Taxonomic categories in fossil tgart: Fischer Verlag. p 150–175. hominids. Cold Spring Harbor Symp Quant Biol 42 Clarke RJ. 1985. Australopithecus and early 66 Keith A. 1948. A new theory of human evolu- 15:109–118. Homo in Southern Africa. In: Delson E, editor. Ancestors: the hard evidence. New York: Alan R. tion. London: C.A. Watts. 21 Andrews P. 1984. An alternative interpretation Liss. p 171–177. of the characters used to define Homo erectus. 67 Tobias PV. 1991. Olduvai Gorge: Vol. 4. The Cour Forsch Inst Senckenberg. 69:167–175. 43 Clarke RJ, Howell FC. 1972. Affinities of the skulls, endocasts and teeth of Homo habilis. Swartkrans 847 Hominid cranium. Am J Phys Cambridge: Cambridge University Press. 22 Wood BA. 1984. The origin of Homo erectus. Anthropol 37:319–336. Cour Forsch Inst Senckenberg 69:99–111. 68 Wood B, Collard M. 1999. Is Homo defined by 44 Grine FE, Strait DS. 1994. New hominid fos- culture? Proc Br Acad 99:11–23. 23 Wood BA. 1992. Early hominid species. J Hum sils from Member 1 ‘‘Hanging Remnant,’’ Swart- Evol 22:351–365. 69 Oakley KP. 1949. Man the toolmaker. London: krans, Formation, South Africa. J Hum Evol British Museum (Natural History). 24 Wood BA. 1994. Taxonomy and evolutionary 26:57–75. relationships of Homo erectus. Cour Forsch Inst 70 Leakey LSB. 1959. A new fossil skull from 45 Grine FE, Demes B, Jungers WL, Cole TM. Olduvai. Nature 184:491–493. Senckenberg 171:159–165. 1993. Taxonomic affinity of the early Homo cra- 71 Hill A, Ward S, Deino A, Curtis G, Drake R. 25 Turner A, Chamberlain AT. 1989. Speciation, nium from Swartkrans, South Africa. Am J Phys 1992. Earliest Homo. Nature 355:719–722. morphological change and the status of African Anthropol 92:411–426. Homo erectus. J Hum Evol 18:115–130. 46 Grine FE, Jungers WL, Schultz J. 1996. Phe- 72 Kimbel WH, Walter RC, Johanson DC, Reed 26 Rightmire GP. 1990. The evolution of Homo netic affinities among early Homo crania from KE, Aronson JL, Assefa Z, Marean CW, Eck GG, erectus. Cambridge: Cambridge University Press. East and South Africa. J Hum Evol 30:189–225. Bobe R, Hovers E, Rak Y, Vondra C, Chen Y, Evensen NM, Smith PE. 1996. Late Pliocene 27 Rightmire GP. 1993. Variation among early 47 Wood BA. 1974. Olduvai Bed I post-cranial Homo and tools from the Hadar Forma- Homo crania from Olduvai Gorge and the Koobi fossils: a reassessment. J Hum Evol 3:373–378. tion (Kada Hadar Member), Ethiopia. J Hum Fora region. Am J Phys Anthropol 90:1–33. 48 Hartwig-Scherer S, Martin RD. 1991. Was Evol 31:549–561. 28 Brauer G. 1994. How different are Asian and ‘‘’’ more human than her ‘‘child’’? Observa- 73 Martin RD. 1983. Human brain evolution in African Homo erectus? In: Franzen JL, editor. 100 tions on early hominid post-cranial skeletons. J an ecological context. New York: American Mu- years of Pithecanthropus: the Homo erectus prob- Hum Evol 21:439–449. seum of Natural History. lem. Cour Forschungs-Institut Senckenberg. 171: 49 Leakey REF, Walker A, Ward CV, Grausz HM. 74 Galaburda AM, Pandya DN. 1982. Role of 301–318. 1989. A partial skeleton of a gracile hominid from architectonics and connections in the study of 29 Brauer G, Mbua E. 1992. Homo erectus fea- the Upper Burgi Member of the Koobi Fora primate brain evolution. In: Armstrong E, Falk D, tures used in cladistics and their variability in Formation, East , Kenya. In: Giaco- editors. Primate brain evolution. New York: Ple- Asian and African hominids. J Hum Evol 22:79– bini G, editor. Proceedings of the 2nd Interna- num Press. p 203–216. 108. tional Congress on Human Paleontology, pp 167– 30 Kramer A. 1993. Human taxonomic diversity 173. Milan: Jaca Books. 75 Gannon PJ, Holloway RL, Broadfield DC, Braun AR. 1998. Asymmetry of chimpanzee pla- in the Pleistocene: Does Homo erectus represent 50 Wood BA. 1992. Origin and evolution of the num temporale: humanlike pattern of Wernicke’s multiple hominid species? Am J Phys Anthropol genus Homo. Nature 355:783–790. 91:161–171. brain language area homolog. Science 279:220– 51 Susman RL, Stern JT. 1982. Functional mor- 222. 31 Leakey LSB, Tobias PV, Napier JR. 1964. A phology of Homo habilis. Science 217:931–934. new species of the genus Homo from Olduvai 76 Gilissen E, Amunts K, Schlaug G, Zilles K. 52 Wood BA, Aiello L, Wood C, Key C. 1998. A Gorge. Nature 202:7–9. 1998. Left-right asymmetries in the temporopari- technique for establishing the identity of ‘‘iso- 32 Leakey REF. 1973. Evidence for an advanced etal intrasylvian cortex of common chimpanzees. lated’’ fossil hominin limb bones. J Anat 193:61– Am J Phys Anthropol 26(suppl):86. Plio-Pleistocene hominid from East Rudolf, 72. Kenya. Nature 242:447–450. 77 Gannon PJ, Kheck NM. 1999. Primate brain 53 Napier JR. 1964. The evolution of bipedal 33 Walker A. 1976. Remains attributable to Aus- ‘language’ area evolution: anatomy of Heschl’s walking in the hominids. Arch Biol (Lie`ge) gyrus and planum temporale in hominids, hylo- tralopithecus in the East Rudolf succession. In: 75(suppl):673–708. Coppens Y, Howell FC, Isaac GL, Leakey REF, batids and macaques and of planum parietale in editors. Earliest man and environments in the 54 Davis PR. 1964. Hominid fossils from Bed I, Pan troglodytes. Am J Phys Anthropol 28(suppl): Lake Rudolf Basin. Chicago: University of Chi- Olduvai Gorge, Tanganyika: a tibia and fibula. 132–133. Nature 201:967. cago Press. p 484–489. 78 Kibunjia M, Roche H, Brown FH, Leakey 34 Leakey REF. 1973. Evidence for an advanced 55 Day MN, Napier JR. 1964. Hominid fossils REF. 1992. Pliocene and Pleistocene archaeologi- Plio-Pleistocene hominid from East Rudolf, from Bed I, Olduvai Gorge, Tanganyika: fossil cal sites west of Lake Turkana, Kenya. J Hum Kenya. Nature 242:447–450. foot bones. Nature 201:968–970. Evol 23:431–438. 35 Bilsborough A, Wood BA. 1988. Cranial mor- 56 Lewis OJ. 1983. The joints of the evolving foot. 79 Kibunjia M. 1994. Pliocene archaeological phometry of early hominids I. facial region. Am J part III: the fossil evidence. J Anat 131:275–298. occurrences in the Lake . J Hum Phys Anthropol 76:61–86. 57 Lewis OJ. 1989. Functional morphology of the Evol 27:159–171. 36 Wood BA. 1991. Koobi Fora research project, evolving hand and foot. Oxford: Clarendon Press. 80 Wood BA, Wood CW, Konigsberg LW. 1994. vol. 4: hominid cranial remains. Oxford: Claren- 58 Kidd RS, Higgins PO, Oxnard CE. 1996. The Paranthropus boisei—an example of evolutionary don Press. OH8 foot: a reappraisal of the functional morphol- stasis? Am J Phys Anthropol 95:117–136. ARTICLES Evolutionary Anthropology 207

81 Semaw S, Renne P, Harris JWK, Feibel CS, 102 Lieberman DE, Wood BA, Pilbeam DE. 1996. 121 Brown JS, Vincent TL. 1987. A theory for the Bernor RL, Fesseha N, Mowbray K. 1997. 2.5 Homoplasy and early Homo: an analysis of the evolutionary game. Theoretical Popul Biol 31:140– million-year-old stone tools from Gona, Ethiopia. evolutionary relationships of H. habilis and H. 166. Nature 385:333–336. rudolfensis. J Hum Evol 30:97–120. 122 Rosenzweig ML, McCord RD. 1991. Incum- 82 Napier JR. 1962. The evolution of the hand. 103 Corruccini RS. 1994. In: Corruccini RS, Cio- bent replacement: evidence for long-term evolu- Sci Am 204:2–9. chon RL, editors. Integrative paths to the past: tionary progress. Paleobiology 17:23–27. 83 Napier JR. 1962. Fossil hand bones from paleoanthropological advances in honor of F. 123 Rosenzweig ML, Brown JS, Vincent TL. 1987. Olduvai Gorge. Nature 196:409–411. Clark Howell. Englewood Cliffs: Prentice Hall. p Red Queen and ESS: The coevolution of evolution- 84 Susman RL. 1988. Hand of Paranthropus ro- 167–183. ary rates. Evol Ecol 1:59–94. bustus from Member I of Swartkrans: fossil evi- 104 Collard M, Wood BA. 1998. Cladistics and 124 Aiello LC, Wheeler PE. 1995. The expensive- dence for tool behaviour. Science 240:781–784. the estimation of hominid phylogeny. Am J Phys tissue hypothesis: the brain and the digestive 85 Susman RL. 1991. Who made the Oldowan Anthropol 26(suppl):122. system in human and primate evolution. Curr stone tools? Fossil evidence for tool behaviour in 105 Collard M, Wood BA. submitted. Reliability Anthropol 36:199–221. Plio-Pleistocene hominids. J Anthropol Res 47: of human phylogenetic hypotheses. Proc Natl 125 Collard M, Wood B. 1999. Grades among the 129–151. Acad Sci. African early hominids. In: Bromage T, Schrenk 86 Susman RL. 1994. Fossil evidence for early 106 Deacon TW. 1990. Problems of ontogeny and F, editors. African biogeography, climate change hominid tool use. Science 265:1570–1573. phylogeny in brain size evolution. Int J Primatol and early hominid evolution. New York: Oxford 87 Ricklan DE. 1990. The precision grip of Austra- 11:237–282. University Press. p 316–327. lopithecus africanus: anatomical and behavioral 107 Asfaw B, White T, Lovejoy O, Latimer B, 126 O’Connell JF, Hawkes K, Blurton Jones NG. correlates. In: Sperber GH, editor. From apes to Simpson S, Suwa G. 1999. : 1999. Grandmothering and the evolution of Homo angels: essays in anthropology in honor of Phillip a new species of early hominid from Ethiopia. erectus. J Hum Evol 36:461–485. V. Tobias. New York: Wiley Liss. p 171–183. Science 284:629–635. 88 Hamrick MW, Churchill SE, Schmitt D, 108 Wood BA. 1996. Origin and evolution of the 127 Wrangham RW, Jones JH, Laden G, Pilbeam Hylander WL. 1998. EMG of the human flexor genus Homo. In: Meikle WE, Howell FC, Jablon- D, Conklin-Brittain N. in press. The raw and the pollicis longus muscle: implications for the evolu- ski NG, editors. Contemporary issues in human stolen: cooking and the ecology of human origins. tion of hominid tool use. J Hum Evol 34:123–136. evolution, memoir 21. San Francisco: California Curr Anthropol. 89 Clayton WD. 1983. The genus concept in Academy of Sciences. p 105–114. 128 Brooks AS. 1996. Behavior and human evolu- practice. Kew Bull 38:149–153. 109 Beynon AD, Wood BA. 1986. Variations in tion. In: Meikle WE, Howell FC, Jablonski NG, 90 Stevens PF. 1984. The genus concept in prac- enamel thickness and structure in East African editors. Contemporary issues in human evolu- tice—but for what practice? Kew Bull 40:457– hominids. Am J Phys Anthropol 70:177–193. tion, Memoir 21. San Francisco: California Acad- 465. 110 Beynon AD, Wood BA. 1987. Patterns and emy of Sciences. p 135–166. rates of enamel growth in the teeth of early 91 Harrison T. 1993. Cladistic concepts and the 129 Bradley J. 1961. International Code of Zoo- hominids. Nature 326:493–496. species problem in hominoid evolution. In: Kim- logical Nomenclature adopted by XV Interna- bel WH, Martin LB, editors. Species, species 111 Hartman SE. 1988. A cladistic analysis of tional Congress of Zoology, London, July 1958. concepts and primate evolution. New York: Ple- hominoid molars. J Hum Evol 17:489–502. London: International Trust for Zoological No- num Press. p 345–371. 112 Washburn SL. 1963. Behavior and human menclature. 92 de Queiroz K, Gauthier J. 1994. Toward a evolution. In: Washburn S, editor. Classification phylogenetic system of biological nomenclature. and human evolution, Chicago: Aldine de Gruyter. 130 Ride WDL. 1985. International code of zoo- Trends Ecol Evol 9:27–31. logical nomenclature, 3rd ed. Berkeley: University 113 Howell FC. 1965. Comment on ‘‘New discov- of California Press. 93 Valentine JW, May CL. 1996. Hierarchies in eries in Tanganyika: Their bearing on hominid biology and paleontology. Paleobiology 22:23–33. evolution,’’ by PV Tobias. Curr Anthropol 6:399– 131 Alexeev VP. 1986. The origin of the human 94 Knox EB. 1998. The use of hierarchies as 401. race. Moscow: Progress Publishers. organizational models in systematics. Biol J Lin- 114 Robinson JT. 1965. Homo ‘‘habilis’’ and the 132 Kennedy GE. 1999. Is ‘‘Homo rudolfensis’’ a naeus Soc 63:1–49. australopithecines. Nature 205:121–124. valid species? J Hum Evol 36:119–121. 95 Mayr E. 1963. species and their evolu- 115 Pilbeam D. 1972. The ascent of man. New tion. Cambridge, MA: Belknap. York: Macmillan. 133 Wood BA. 1999. ‘‘Homo rudolfensis’’ Alexeev, 1986—fact or phantom? J Hum Evol 36:115–118. 96 Strait DS, Grine FE, Moniz MA. 1997. A 116 Huxley JS. 1958. Evolutionary processes and reappraisal of early hominid phylogeny. J Hum taxonomy with special reference to grades. 134 Black D. 1927. On a lower molar hominid Evol 32:17–82. Uppsala Universitets a˚rsskrift. 6:21–38. tooth from the Chou Kou Tien deposit. Palaeonto- 97 Swofford DL. 1991. Phylogenetic analysis us- 117 Simpson GG. 1961. Principles of animal logia Sinica (N.S.D.)7:1–28. ing parsimony, version 3.0s. Champaign: Illinois taxonomy. New York: Columbia University Press. 135 Dubois E. 1894. Pithecanthropus erectus, eine Natural History Survey. 118 Miller AH. 1949. Some ecologic and morpho- menschena¨hnliche Uebergangsform aus Java, 98 Chamberlain AT. 1987. A taxonomic review logic considerations in the evolution of higher Batavia Landes-Druckere. and phylogenetic analysis of Homo habilis. Ph.D. taxonomic categories. In: Mayr E, Schu¨ z E, edi- thesis, the University of Liverpool, England. tors. Ornithologie als Biologische Wissenschaft. 136 Groves CP, Maza´k V. 1975. An approach to 99 Maddison WP, Maddison DR. 1992. Mac- Berlin: Carl Winter. p 84–88. the taxonomy of the Hominidae: gracile Villafran- chian hominids of Africa. Cas Miner Geol 20:225– Clade: analysis of phylogeny and character evolu- 119 Jensen JS. 1990. Plausibility and testability: 247. tion, version 3. Sunderland, MA: Sinauer Assoc. assessing the consequences of evolutionary inno- 100 Krings M, Stone A, Schmitz RW, Krainitzk vation. In: Nitecki MH, editor. Evolutionary inno- 137 Bermudez de Castro JM, Arsuaga JL, Carbon- H, Stoneking M, Pa¨a¨bo S. 1997. Neandertal DNA vations. Chicago: Chicago University Press. p ell E, Rosas A, Martinez I, Mosqueria M. 1997. A sequences and the origin of modern humans. Cell 171–190. hominid from the Lower Pleistocene of Atapu- 90:19–30. 120 Hunter JP. 1998. Key innovations and the erca, Spain: possible ancestor to Neandertals and 101 Chamberlain AT, Wood BA. 1987. Early homi- ecology of macroevolution. Trends Ecol Evol modern humans. Science 276:1392–1395. nid phylogeny. J Hum Evol 16:119–133. 13:31–36. ௠ 1999 Wiley-Liss, Inc.