The Omo-Turkana Basin Fossil Hominins and Their Contribution to Our Understanding of Human Evolution in Africa BERNARD WOOD AND MEAVE LEAKEY The Omo-Turkana Basin, including the hominin fossil sites around Lake Turkana and the sites along the lower reaches of the Omo River, has made and continues to make an important contribution to improving our murky understanding of human evolution. This review highlights the various ways the Omo-Turkana Basin fossil re- cord has contributed to, and continues to challenge, interpretations of human evolu- tion. Despite many diagrams that look suspiciously like comprehensive hypotheses about human evolutionary history, any sensible paleoanthropologist knows that the early hominin fossil record is too meager to do anything other than offer very provi- sional statements about hominin taxonomy and phylogeny. If history tells us any- thing, it is that we still have much to learn about the hominin clade. Thus, we sum- marize the current state of knowledge of the hominin species represented at the Omo-Turkana Basin sites. We then focus on three specific topics for which the fossil evidence is especially relevant: the origin and nature of Paranthropus ; the origin and nature of early Homo ; and the ongoing debate about whether the pattern of human evolution is more consistent with speciation by cladogenesis, with greater taxo- nomic diversity or with speciation by anagenetic transformation, resulting in less tax- onomic diversity and a more linear interpretation of human evolutionary history. We use a relatively speciose taxon- omy for African hominins that places more stress on grades than clades (Fig. 1; Box 1). 1 The need to provide columns that represent the time between the first and last appearan- ces of the taxa in the hominin fossil record complicates the diagram, but there is a relatively simple underlying pattern. The first unambiguous homi- nins occur ca . 4 Ma. Before that, the fossil record is poor, despite one 4.4 Ma impressively complete associated skeleton (ARA-VP-6/500) from Aramis in the Middle Awash. It is not pres- ently possible to be sure that any of the possible early hominin taxa really do belong to the hominin clade. All we know of the earliest true hominins comes from sites in East Africa. Moreover, recently published dating evidence suggests that we may not see evidence of hominins in southern Africa until as late as ca . 2.5 2 2 Ma. Between 2.5 and 2.0 Ma, the hominin fossil record seems to resolve into two trends, one toward even larger chew- ing teeth with thick enamel, and one in which postcanine tooth size dimin- ishes, brain size increases, and obli- gate bipedalism becomes established. The taxa emphasized in bold in Fig- ure 1 are the ones we believe are rep- resented in the hominin fossil record from the Omo-Turkana Basin. We summarize the species within each grade in the historical order the taxa were recognized, not in their tempo- ral order. In doing so, we emphasize the important role the hominin fossil record from the Omo-Turkana Basin (Appendix 1) has played in developing our current understanding of hominin evolution (Box 2). ARCHAIC HOMININS Australopithecus afarensis Johanson, 1978 The type specimen, LH 4, was recovered in 1974 from Laetoli in Tanzania. The majority of the Au. afarensis hypodigm comes from the Ethiopian site of Hadar, but other ARTICLE Bernard Wood is University Professor of Human Origins at The George Washington University and adjunct senior scientist at the National Museum of Natural History, Smithsonian Institution. In 1968, he took part in the first field season at what was then called East Rudolf. He is the author of the 1991 monograph on the hominin cranial remains from Koobi Fora. He edits the Wiley-Blackwell Encyclopedia of Human Evolution. E-mail: bernardawood@gmail.com Meave Leakey is a research professor in the Department of Anthropology, Stony Brook Uni- versity, New York. She has carried out field work in the Turkana Basin over the past four deca- des and currently co-leads the Koobi Fora Research Project with her daughter Louise. She is an Explorer-in-Residence of the National Geographic Society, which has funded field work in East Africa for many decades and currently, in large measure, funds the KFRP field research at Turkana. Meave’s research interests include the evolution of cercopithecids and hominids, as well as the study of past faunal assemblages as proxies for reconstructing paleohabitats and past environmental changes. Together with Richard and Louise Leakey, and in collaboration with Stony Brook University, Meave is developing the Turkana Basin Institute with two research facilities to the east and west of Lake Turkana. E-mail: meaveleakey@gmail.com Key words: Australopithecus ; Homo ; Paranthropus ; taxonomy; East Africa V V C 2011 Wiley Periodicals, Inc. DOI 10.1002/evan.20335 Published online in Wiley Online Library (wileyonlinelibrary.com). Evolutionary Anthropology 20:264–292 (2011) sites in Ethiopia, such as Dikika, Maka, and Woranso-Mille, have also contributed important fossil evidence. The temporal range of Au. afarensis is ca 3.8 2 3 Ma ( ca 4 2 3 Ma, if the Belohdelie frontal is confirmed as belonging to Au. afarensis ). The Au. afarensis hypodigm includes a well- preserved skull, partial and frag- mented crania, many lower jaws, an associated skeleton (A.L.-288) that preserves just less than half of the skeleton of an adult female, and suffi- cient limb bones to be able to esti- mate stature and body mass. Com- parative evidence suggests that the hind limbs of Au. afarensis are sub- stantially shorter than those of a modern human of similar stature. The morphology of the pelvis and the relatively short lower limb suggests that although Au. afarensis was capa- ble of bipedal walking, it was not adapted for long-range bipedalism. This indirect evidence of the loco- motion of Au. afarensis is comple- mented by the discovery at Laetoli of several trails of fossil footprints. These provide direct graphic evidence that at least one contemporary hominin, pre- sumably Au. afarensis , but possibly Kenyanthropus platyops , was capable of bipedal locomotion. However, the Laetoli prints are less like those of modern humans than the 1.5 Ma foot- prints from Koobi Fora, which are pre- sumed to be of pre-modern Homo 2 Despite this compelling direct evi- dence of bipedalism, the upper limb of Au. afarensis , especially the hand 3 and shoulder girdle, 4 retains morphology that most likely reflects a significant element of arboreality. Although some argue there is rela- tively little sexual dimorphism in this taxon, most researchers interpret the Au. afarensis hypodigm as consistent with high levels of canine and body mass sexual dimorphism. Most body mass estimates of Au. afarensis range from ca 30 2 45 kg; its estimated endocranial volume ranges from 400 2 550 cm 3 . The taxon Australopi- thecus bahrelghazali Brunet et al 1996 may be a regional variant of Au. afarensis or it may conceivably sample K. platyops Figure 1. Distribution of hominin taxa through time as discussed in the text. Taxa present in the Omo Turkana Basin appear in boxes. Box 1. Grades Explained The grade concept sorts taxa into broad functional groups. Although the boundaries of grades are sometimes frustratingly ‘‘fuzzy,’’ grades are the best way to sort hominin species into more inclusive units until we can be sure that we are generating reliable hypotheses about the relationships among hominin taxa. 1 We use the genus names Australopithecus , Para- nthropus , and Homo even though Australopithecus is very likely not a monophyletic group. Both of us also doubt that Paranthropus and Homo are monophyletic groups as they are presently interpreted. The grades used in this review are ‘‘Archaic hominins,’’ ‘‘Hyper-megadont archaic hominins,’’ ‘‘Transitional hominins,’’ ‘‘Pre-modern Homo ,’’ and ‘‘Anatomi- cally modern Homo .’’ ARTICLE Omo-Turkana Basin Hominins 265 Evidence from the Omo-Turkana Basin The first discoveries of specimens currently attributed to Au. afarensis from the Turkana Basin were from two localities, Brown Sands and White Sands, in the Usno Formation in the Lower Omo Valley. 5 To date, the only other possible evidence of Au. afarensis from the Turkana Basin is a fragmented cranium (KNM-ER 2602) dated to a little under 3.4 Ma from Koobi Fora. 6 A fragment of mandibular corpus (KNM-ER 20432) and several isolated teeth estimated to be ca 3.95 Ma from the Koobi Fora Formation at Kubi Algi in the Allia Bay region were assigned to Australopithecus cf. Au. afarensis when they were first described. 7 However, once Au. anamensis was recognized as a distinct taxon, the Kubi Algi specimens have been con- sistently assigned to the latter taxon, although Kimbel and colleagues 21 described ways in which their mor- phology was intermediate between the two taxa. The hominins recovered from Loth- agam 8,9 were at one time considered to belong to Au. afarensis , 10 then more conservatively to Hominidae indet., 11 but temporally they fit better with ei- ther Ardipithecus ramidus 12 or Ardipi- thecus kadabba Two isolated teeth from the uppermost Upper Nawata Member, along with the KNM-LT 329 mandibular fragment from the lower- most Apak Member, are probably close in age to ca . 5 Ma and thus repre- sent the earliest known hominins (or hominids) from the Omo-Turkana Ba- sin. Hominin specimens have also been recovered from South Turkwel, a ca . 3.5 Ma site south of the Turkwel River on the western shores of the lake, as well as from the Kaiyumung Member of the Nachukui Formation at Lothagam. The South Turkwel specimens, which include a mandibu- lar fragment, some associated hand bones, a proximal pedal phalanx, and tooth fragments, were attributed to Australopithecus sp. indet. 13 They may belong to either Au. afarensis or Ken- yanthropus platyops . Fragmentary fos- sils from Fejej originally assigned to Au. afarensis 14,15 are more likely to belong to Au. anamensis Australopithecus anamensis Leakey, Feibel, McDougall, and Walker, 1995 The first evidence of the taxon we now refer to as Australopithecus ana- mensis , the distal end of a left hu- merus, was recovered from Kanapoi in the Omo-Turkana Basin in 1965 by a team from Harvard University led by Brian Patterson. He suggested that the specimen, which is now known as KNM-KP 271, but which they referred to as Kanapoi Homi- noid 1, ‘‘may prove to be Australopi- thecus .’’ 16:66 Nearly thirty years elapsed before the next hominins were recovered from Kanapoi in 1994. It was one of these latter dis- coveries, KNM-KP 29281, that became the holotype of Au. anamen- sis 17 The other sites to contribute to the hypodigm of Au. anamensis are Allia Bay 18 and Fejej 14,15 in the Omo- Turkana Basin and Aramis and Assa Issie 19 in the Middle Awash study area in Ethiopia. The temporal range of Au. anamensis is ca . 4.2 2 3.9 Ma. Evidence from the Omo-Turkana Basin The type specimen and most of the Au. anamensis hypodigm comes from sites within the Omo-Turkana Basin, and most of that comes from Kanapoi. With the exception of one specimen (KNM-KP 29287), all of the Kanapoi component of the Au. anamensis hypo- digm derives from a ca . 100-ka time interval represented by fluvial sedi- ments in the Kanapoi Formation, which are sandwiched between two tuffaceous beds securely dated at 4.17 and 4.07 Ma. Fossils from the Loko- chot Member of the Koobi Fora For- mation at Allia Bay are slightly younger than those at Kanapoi. The majority of the Allia Bay specimens were recovered from the excavation of a bone bed at a site known as 261-1 just below the Moiti Tuff, dated at 3.95 Ma. Three specimens recovered ca . 1 km to the east of the 261-1 locality in a small drainage that cuts through the Moiti tuff and a radius recovered just north of Sibiloi are estimated to be of similar age. The Omo-Turkana Basin contribu- tion to the Au. anamensis hypodigm is dominated by cranial remains. The teeth of Au. anamensis are more primi- tive than those of Au. afarensis with respect to their parallel-sided tooth rows, mandibular canine morphology, asymmetry of the premolar crowns, and relatively simple crown morphol- ogy of the deciduous first mandibular molars, but are reminiscent of Par- anthropus in other features, including the low cross-sectional profiles and bulging sides of the molar crowns. The distal humerus attributed to Au. ana- mensis (KNM-KP 271) is australopith- like and a tibia (KNM-KP 29283) has features associated with obligate bipedalism. The KNM-KP 31724 capi- tate differs from the Au. afarensis capi- tate in that the articular facet for meta- carpal II faces laterally, as it does in apes, suggesting little capacity for rotation at the second carpometacar- pal joint. 20 Thus, the postcranial skele- ton of Au. anamensis from the Omo- Turkana Basin points to that taxon being bipedal, but with some primitive forelimb features. 20 White and coworkers 19 supported the proposal made by others 21 that although the hypodigms of Au. ana- mensis and Au. afarensis are taxo- nomically distinct, there is evidence, such as relative canine size, upper canine morphology, enamel thick- ness, and femoral shaft morphology, of the type of morphological continu- ity one would expect to result from anagenetic transformation within an evolving lineage. Kimbel and col- Although the hypodigms of Au. anamensis and Au. afarensis are taxonomically distinct, there is evidence . . . of the type of morphological continuity one would expect to result from anagenetic transformation within an evolving lineage . . . 266 Wood and Leakey ARTICLE Box 2. Timeline of Hominin Research in the Omo-Turkana Basin Figure B1. ARTICLE Omo-Turkana Basin Hominins 267 Figure B2. 268 Wood and Leakey ARTICLE leagues 21:134 provide compelling evi- dence that Au. anamensis and Au. afarensis are ‘‘parts of an anageneti- cally evolving lineage, or evolution- ary species.’’ Recent discoveries from the Woronso-Mille study area in Ethiopia 22 support this hypothesis. Kenyanthropus platyops Leakey and colleagues, 2001 In 1982, the Koobi Fora Research Project (KFRP) initiated exploration of richly fossiliferous sites in the Nachukui Formation exposed in the drainages of the Lomekwi and Toper- nawi rivers to the west of Lake Tur- kana. 23 During field work between 1982 and 1985, two hominins (KNM- WT 8556 and 16006) were recovered and initially assigned to Au. afaren- sis , 24 even though it was evident that they differed in several ways from the mandibles of this taxon. While they have not been formally attributed to Kenyanthropus platyops , it is probable that they belong to that taxon rather than to Au. afarensis 25 Field expeditions in 1998 and 1999 led to the recovery of 34 isolated teeth, a small maxillary fragment, and a well- preserved temporal bone from the Kataboi and lower Lomekwi members of the Nachukui Formation dating to 3.53 2 3.3 Ma. In the last few days of the 1999 field season, a cranium was discovered in situ in a mudstone 8 m below the Tulu Bor Tuff and 12 m above the Lokochot Tuff. The cra- nium’s location in the section points to an estimated age of 3.5 Ma. Although the specimen is unfortunately dis- torted by multiple matrix-filled cracks and exhibits lateral skewing, the gen- eral morphology of the face appears unusually flat and long. One upper molar was sufficiently well-preserved to allow metrical dental comparisons. The distortion of the cranium compli- cated its interpretation, but its unusual morphology was judged sufficient to justify the establishment of a new hominin genus and species, Kenyan- thropus platyops , 26 with the cranium KNM-WT 40000 as the holotype and the maxilla KNM-WT 38350 as the par- atype. It is likely that most of the remaining 34 specimens referred to earlier, including three mandible frag- ments and a maxilla fragment that date to between 3.53 and 3.3 Ma, belong to K. platyops . However, because none of the anatomical regions preserved in the type and paratype were duplicated in the rest of the sample, the authors re- served judgment about the taxonomy of these remains and the two mandi- bles, which had only recently been referred to Au. afarensis 24 The main reasons that Leakey and colleagues 26 did not assign the Lomekwi material to Au. afarensis are its reduced subnasal progna- thism, anteriorly situated zygomatic root, flat and vertically orientated malar region, relatively small but thick-enameled molars, and an M 2 that falls below the known ranges of other contemporary hominin species. Some morphology of the new genus, including the lack of prognathism, is Paranthropus -like, yet K. platyops lacks the postcanine hyper-megadon- tia that characterizes Paranthropus The authors noted that the face of the new material resembles that of Homo rudolfensis Kenyanthropus platyops apparently displays a hith- erto unique combination of facial and dental morphology, but White 27 claimed the new taxon is unneces- sary because, in his view, the type specimen of K. platyops is just a dis- torted Au. afarensis cranium. Spoor, Leakey, and Leakey 28 provide a cogent response to White’s criticism that the specimen is so distorted that no taxonomically useful morphology remains. In any event, the ‘‘dis- tortion’’ argument does not account for the actual and estimated small size of the postcanine teeth. Evidence from the Omo-Turkana Basin All the presently known fossil evi- dence of K. platyops comes from the Omo-Turkana Basin, but it is possi- ble that presently unpublished fossils from Chad may belong to K. pla- tyops . If this is the case, K. platyops would be a junior subjective syno- nym of Au. bahrelghazali HYPER-MEGADONT ARCHAIC HOMININS Paranthropus boisei (Leakey, 1959) Robinson, 1960 The first evidence of an East Afri- can hyper-megadont hominin was OH 3, which consists of two decidu- ous teeth recovered in 1955 from lo- cality BK in Bed I at Olduvai Gorge. The significance of the discovery was not appreciated until the recovery of the OH 5 cranium from locality FLK, also in Bed I at Olduvai Gorge, in 1959. The cranium was initially allocated to a novel genus and spe- cies, Zinjanthropus boisei , but in 1964 Leakey, Tobias, and Napier proposed that Zinjanthropus be reduced to a subgenus within Aus- tralopithecus , as Australopithecus ( Zinjanthropus ) boisei 29 Subse- quently, Tobias recommended that the subgenus Zinjanthropus be sunk into Australopithecus 30 In spite of the implications of monophyly for the southern African and East Afri- can hyper-megadont taxa, in this review we use Paranthropus boisei , a name combination first used by Rob- inson in 1960. 31 The next-recovered evidence of a hyper-megadont archaic hominin was a well-preserved mandible, Pen- inj I, recovered in 1964, which was referred to as ‘‘probably a late mem- ber of the Olduvai species, Au. boi- sei .’’ 32:27 A proximal femoral frag- ment, OH 20, was found at Olduvai Gorge in 1959, but it was not recog- nized as a hominin until 1968 (hence its relatively high ‘‘OH’’ number). It was initially referred to as Australopi- thecus cf. Au. boisei , 33 but its mor- phology is also consistent with that of Homo habilis . Apart from evidence from sites in the Omo-Turkana Ba- sin, the next East African site to yield the remains of P. boisei was Cheso- wanja, where KNM-CH 1, a right hemiface and anterior cranial base, was recovered from the Chemoigut Formation in 1970. The morphology of the face and the absolute size and proportions of the dentition were judged to be characteristic of P. boi- sei 34 It was suggested that the facial and cranial morphology of KNM-CH 1 (the latter was interpreted as being consistent with a large endocranial volume) were evidence that it sampled ‘‘a population of evolved ro- bust australopithecines, most likely descended from Au. boisei .’’ 34:513 But others 35 were not persuaded that the cranial capacity of KNM-CH 1 was large, nor were they convinced that ARTICLE Omo-Turkana Basin Hominins 269 the cranial morphology of KNM-CH 1 departed in any significant way from that of P. boisei The Cheso- wanja cranium is now more usually interpreted as a small-bodied, prob- ably female representative of P. boi- sei 36,37 Two additional hominin specimens from Chesowanja, KNM- CH 302 and KNM-CH 304, have also been assigned to the species. 38 Fur- ther evidence for P. boisei in East Africa came in 1993 when KGA 10- 525, the first well-preserved skull of P. boisei , was recovered at Konso (initially called Konso Gardula) in Ethiopia. 39 Evidence from the Omo-Turkana Basin Several large isolated postcanine tooth crowns were recovered from the Shungura Formation in 1967 and 1968, but were not announced until 1969. 40 Isolated teeth resembling those of OH 5 and Peninj 1 (for example, Omo L628-1) and mandi- bles similar to Peninj 1 (for example, L74a-21 and L7a-125) were assigned to P. boisei 41,42 A fragmentary adult cranium (Omo 323-1976-896) recov- ered from Member G of the Shun- gura Formation in 1976 almost cer- tainly belongs to this species as well. 43,44 Most of the P. boisei hypodigm comes from Koobi Fora. 37 Two hemi-mandibles with robust bodies (KNM-ER 403 and 404), plus an abraded and edentulous palate (KNM-ER 405), were recovered in 1968 by the first field expedition to the east side of what was then called Lake Rudolf. Cranial evidence came the following year with the recovery of KNM-ER 406, a well- preserved but edentulous adult cranium, and KNM-ER 407, the pos- terior part of an adult calvaria. A year later, researchers recovered KNM-ER 732, a partial adult cra- nium preserving the majority of the vault, the right side of the face, and part of the right side of the cranial base, and KNM-ER 733, a frag- mented adult skull that preserves sufficient taxonomically valent mor- phology to allow it to be assigned to P. boisei Additional sub- ca 2 Ma cranial remains belonging to P. boi- sei have been found at Koobi Fora, among them KNM-ER 13750 and KNM-ER 23000, as well as on the opposite side of the lake at West Turkana (KNM-WT 17400). 45–48 De Lumley and Marchal 50 claim that a distal humerus recovered from Box 3. Taxonomic State of Play Paraustralopithecus aethiopicus Arambourg and Coppens, 1968 In 1968, Camille Arambourg and Yves Coppens 51 came to the conclu- sion that a hominin mandible from the Shungura Formation (Omo 18- 1967-18 or sometimes just Omo 18-18) departed sufficiently from the mor- phology of Peninj I and the mandibles of Paranthropus robustus to merit the introduction of a new genus and species. Subsequently, Howell 36 assigned the mandible to Australopithecus africanus , thus making Paraus- tralopithecus aethiopicus a junior subjective synonym of Au. africanus . In 1980, one of the original authors, Yves Coppens, implied that Par. aethio- picus should be sunk into Australopithecus boisei 42 No one currently rec- ognizes Paraustralopithecus as a separate genus. However, researchers who consider the pre-2.3 Ma hyper-megadont hominins from Omo-Shun- gura (for example, L. 338y-6) and West Turkana (for example, the KNM- WT 16005 mandible and the KNM-WT 17000 cranium) to belong to a spe- cies other than Paranthropus boisei have used either Australopithecus aethiopicus or Paranthropus aethiopicus for such a taxon; we use the lat- ter. The fossil evidence for P. aethiopicus is similar to that for P. boisei except that the face of the former taxon is more prognathic, the cranial base is less flexed, the incisors are larger, and the postcanine teeth are not so large or morphologically specialized. But since there is only one rela- tively complete P. aethiopicus cranium, the warnings of Smith 133 about making taxonomic inferences based on small samples are especially relevant. Pithecanthropus rudolfensis Alexeev, 1986 In 1986, Valery Alexeev established a new hominin species, Pithecanthro- pus rudolfensis , because, based on the published evidence, he concluded that the KNM-ER 1470 cranium could not be accommodated within Homo habilis . Groves 82 transferred the new taxon to Homo . Subsequently, it was more comprehensively differentiated from Homo habilis . In addition to the larger endocranial volume of H. rudolfensis (based on the KNM-ER 1470 cranium and the KNM-ER 3732 calotte), other differences between H. rudolfensis and H. habilis include the location of the maximum width of the face (the former taxon is widest across the midface; the latter is widest across the upper face). Other researchers prefer not to recognize H. rudol- fensis as a separate taxon. Instead, they include its hypodigm within a more inclusive interpretation of H. habilis Homo ergaster Groves and Maza ́ k, 1975 In the early 1970s, Colin Groves and Vratislav Maza ́ k reviewed the pub- lished evidence for the fossil hominins from East Rudolf and concluded that a new Homo species was needed to accommodate fossils that, in their view, could not be accommodated within Homo habilis or Homo erectus Researchers have used the taxon name Homo ergaster for East African hominin remains (for example, KNM-ER 992, KNM-WT 15000) that they consider to be generally more primitive than Asian Homo erectus and that lack the more extreme expressions of some of its derived features, such as a thick inner and outer table and sagittal keeling. Most researchers acknowledge that these differences exist but are inclined to view them as regional variants or size-related variations within H. erectus , the species that has priority. 270 Wood and Leakey ARTICLE strata from the Koobi Fora Forma- tion exposed at Fejej belongs to P. boisei . More details about the P. boi- sei specimens from the Omo-Tur- kana Basin can be found in a recent review. 49 The earliest evidence of P. boisei is ca . 2.3 Ma. The most recent evidence is dated to 1.4 Ma 39 from Konso in Ethiopia. Unfortunately, because there are no major East African hominin sites in the period between ca . 1.4 and 1.0 Ma, we do not know for how long P. boisei persisted during that interval. There is no evi- dence of it after 1.0 Ma. Paranthropus aethiopicus (Arambourg and Coppens, 1968) Wood and Chamberlain, 1987 A hominin mandible, Omo 18- 1967-18 (or Omo 18-18), was found in the Shungura Formation in 1967. The authors of the paper announc- ing the find concluded that it departed sufficiently from the mor- phology of both the Peninj I mandi- ble and the mandibles of Paranthro- pus robustus to merit the introduc- tion of a new genus and species, Paraustralopithecus aethiopicus (Box 3). 51 In this review, we recognize this taxon as a distinct species but include it within Paranthropus as P aethiopicus Two other hyper-megadont homi- nin fossils, the mandible KNM-WT 16005 and the cranium KNM-WT 17000, which date to between 2.6 and 2.3 Ma were recovered from West Turkana in 1985. 52 Three years later, Gen Suwa 53 reviewed the hyper-megadont dental evidence from the Shungura Formation. He showed that the crowns of the mandibular premolars before ca . 2.3 Ma had smaller talonids and were in other ways less derived than those younger than 2.3 Ma. A frag- mented immature cranium from Member E, L338y-6, recovered from the Shungura Formation in 1969 and 1970-72, was initially assigned to P. boisei , but its age of ca . 2.4 Ma and the morphology it shares with KNM-WT 17000 are consistent with it belonging to P. aethiopicus 54,55–57 Evidence from the Omo-Turkana Basin Most of the relatively modest P. aethiopicus hypodigm comes from the Omo-Turkana Basin. The excep- tions are two maxillary fragments, one from Malema 58 and a probable P. aethiopicus maxilla, EP 1500/01, recovered from Laetoli in 2001. 59 Presently, the oldest known evidence of P. aethiopicus are the ca 2.6 Ma 46,49 fossils from Shungura For- mation and the ca 2.5 Ma fossils from West Turkana, 52 but the max- illa from the Ndolanya Beds at Lae- toli may be as old as 2.7 Ma. 59 The last documented appearance of P. aethiopicus is ca . 2.3 Ma. TRANSITIONAL HOMININS Homo habilis Leakey, Tobias and Napier, 1964 In 1964, Louis Leakey, Phillip Tobias, and John Napier announced the new species Homo habilis and designated OH 7 from FLK NN as the type specimen, with OH 4, 6, 8, and 13 as paratypes. 60 In due course, additional specimens from Olduvai were added to the hypodigm of H. habilis A relatively complete but crushed and deformed cranium OH 24 61 and the fragmentary associated skeleton known as OH 62 62 were the most significant of these. The discov- ery of OH 24 was important for two reasons. First, it resembled OH 13 and confirmed that the ‘‘advanced’’ morphology of the former was not unusual. Second, it was not found in Bed II (OH 13’s location), but near the base of Bed I. This placed it among the oldest of the specimens from Olduvai Gorge allocated to H. habilis , meaning that it was no lon- ger possible to argue that there was a temporal cline in H. habilis mor- phology, with the more ‘‘primitive’’ specimens at the base of Bed I and the morphologically ‘‘more advanced’’ fossils in Bed II. The implications of the OH 62 associated skeleton were rather different. Its limb proportions were evidently more primitive than those of any then-known Homo species. Indeed, some claimed that they were more primitive than those of Au. afaren- sis , 63 but other researchers have since suggested that, given the poor quality of the evidence on the lengths of the fragmentary long bones of OH 62, there is probably no significant difference between the limb proportions of Au. afarensis and H. habilis 64,65 However, if OH 62 belongs to H. habilis , as it almost certainly does unless paleoanthropol- ogists are prepared to recognize a new species for this one specimen, then most researchers subscribe to the view that the locomotor adapta- tions of H. habilis as interpreted from the hand and foot bones were not like those of pre-modern Homo (for example Homo erectus , Homo heidelbergensis ) and were more like those of archaic hominins. Another hominin found at Olduvai, as well as fossils from other African sites, have been judged to have affin- ities with H. habilis . A hominin max- illa (OH 65) discovered in Upper Bed I at Olduvai Gorge in 1995 has been assigned to H. habilis 66 The frag- mentary cranium Stw 53 and some isolated teeth from Member 5 at Sterkfontein are also said to resem- ble H. habilis 67 The same suggestion has been made with respect to SK 847 from Member 1 at Swartkrans, 68 a maxilla from the Kadar Hadar Member of the Busidima Formation at Hadar, 69,70 a temporal bone from the Chemeron Formation, 71 and some teeth from Drimolen. 72 Pro- posals that H. habilis remains have been recovered from sites beyond Africa 73 have not received wide ac- ceptance. The small endocranial vol- umes and relatively primitive mor- phology of the hominins recovered from Dmanisi led some observers to refer to them as ‘‘habilis-like,’’ 74 but the researchers who are most famil- iar with the material suggest that these specimens share sufficient derived features with Homo erectus sensu lato 75,76 for the Dmanisi homi- nins to be regarded as primitive rep- resentatives of that taxon. The mor- phology of the postcranial remains from Dmanisi is generally more like that of later Homo than that of H. habilis . As for the temporal span of H. habilis , the oldest of the speci- mens allocated or likened to that ARTICLE Omo-Turkana Basin Hominins 271 taxon is the A.L. 666-1 maxilla from Hadar ( c . 2.4 Ma) 69,70 the youngest is a maxilla from Koobi Fora, KNM-ER 42703, dated at 1.44 Ma. Evidence from the Omo-Turkana Basin Apart from the evidence from the type site, Olduvai Gorge, the most significant contribution to the collec- tion of fossils generally attributed to H. habilis has come from Koobi Fora. The evidence from Koobi Fora includes well-preserved crania (for example, KNM-ER 1805, 1813), a fragmentary associated skeleton (KNM-ER 3891), mandibles, and iso- lated teeth. 37 Nearly all the published material attributed to Homo habilis was found either in the early 1970s or shortly thereafter. Initially, these specimens were not allocated to a species, but were given the informal designation ‘‘early Homo .’’ Some of the hominin fossils recovered from Members G and H of the Shungura Formation (for example, a frag- mented cranium, L894-1, 77 two man- dibles, and isolated teeth 78,79 ) have also been assigned to H. habilis , as has an isolated tooth from the Nachukui Formation, West Tur- kana. 80 Three isolated mandibular teeth recovered from strata from the Koobi Fora Formation exposed at Fejej have been attributed to Homo aff. Homo habilis 50 Homo rudolfensis (Alexeev, 1986) Groves, 1989 Although Tobias 81 made a compel- ling case that all of the non-mega- dont archaic hominins from Bed I and Lower Bed II at Olduvai Gorge could be subsumed into H. habilis , it became clear during the first half of the 1970s that the early Homo cra- nial fossils (for example, KNM-ER 1470, 1590, and 3732 versus KNM- ER 1805 and 1813) and mandibular fossils (for example, KNM-ER 1802 versus KNM-ER 1501 and 1502) from Koobi Fora were less easy to interpret as representatives of a sin- gle taxon (Box 3). In a 1986 book about the fossil evidence of human evolution, Valery Alexeev named a new species, Pithecanthropus rudol- fensis , suggesting that this was justi- fied because of the morphological differences between the cranium KNM-ER 1470 from Koobi Fora and the fossils from Olduvai Gorge allo- cated to H. habilis Groves 82 sug- gested that P. rudolfensis should be sunk into the genus Homo because the genus Pithecanthropus had been made a junior synonym of Homo in 1940, so that its formal name is now Homo rudolfensis . If H. habilis does subsume more variation than is con- sistent with it being a single species, and if KNM-ER 1470 is judged to belong to a different species than the type specimen of H. habilis , OH 7, then H. rudolfensis is available as the name of a second early Homo taxon. Evidence from the Omo-Turkana Basin One of us (BW) has suggested that in addition to KNM-ER 1470, the hypodigm of H. rudolfensis consists of other cranial specimens, among them KNM-ER 1590 and 3732, and mandibles such as KNM-ER 1802, which come from a relatively small geographic area on the east side of Lake Turkana. 37 Some also consider the HCRP UR 501 mandible from Uraha in Malawi 83 to be part of the hypodigm of H. rudolfensis How- ever, in the absence of associated cranial and mandibular specimens, these suggestions cannot be verified. New material recently discovered at Koobi Fora and currently under study should provide new evidence on this issue. PRE-MODERN HOMO Homo ergaster Groves and Maz ak, 1975 Colin Groves and Vratislav Maza ́ k established this hominin species to accommodate fossil hominins recov- ered from Koobi Fora that, in their judgment, did not belong in any existing taxon (Box 3). 84 Before that 1975 publication, it had been tacitly assumed that these remains belonged to H. erectus . However, Wood 85 used the taxon name Homo ergaster for East African hominin remains that are generally more primitive than Asian Homo erectus and lack the more extreme expressions of some of the derived features (such as thick inner and outer table and sagittal keeling) seen in Asian Homo erectus Other researchers either acknowl- edge these differences exist or con- sider the regional differences to have been overstated. Either way, most are inclined to view them as regional variations within a single species, H. erectus . The holotype of H. ergaster is the mandible KNM-ER 992 found at Koobi Fora in 1971. Groves and Maza ́ k nominated KNM-ER 730, 731, 734, 803, 806-9, 820, 1480, 1805 as paratypes, but we, and most observers, refer KNM-ER 1805 to H. habilis Evidence from the Omo-Turkana Basin If the hominins from Dmanisi are assigned to H. ergaster , 75,76 then the Omo-Turkana Basin has not furnished the majority of the evidence for H. ergaster ; if they are not, then most of the evidence for H. ergaster comes from the Omo-Turkana Basin. Among the fossils from Koobi Fora presently assigned to H. ergaster are cranial (for example, KNM-ER 3733, 3883) and mandibular evidence (for example, KNM-ER 730, 820, 992), as well as a fragmentary, diseased associated skel- eton (KNM-ER 1808) and a second, less complete, associated skeleton (KNM-ER 803). KNM-WT 15000, an associated skeleton of a presumed ado- lescent male from Nariokotome at West Turkana, 86 is the most complete specimen belonging to H. ergaster from the Omo-Turkana Basin. As well as being remarkably free from plastic de- formation, at the time of its recovery KNM-WT 15000 was the best pre- served early hominin skeleton. Its sec- ond permanent molars are just in occlusion and the M 3 germs have probably only just grown complete crowns. Based on tooth micro- structure, 87,88 the most precise and accurate developmental age estimate suggests an age of ca . 8 years (range, 7.5–8.9 years); the modern human equivalent would be ca . 12 years old. Even though nearly all researchers agree that this skeleton is correctly 272 Wood and Leakey ARTICLE attributed to Homo , the indications are that its life history was not like that of modern humans. 89 The projected stature estimate for an equivalent adult is 185 cm, 90 but Ohman and co- worker 91 and Graves and colleagues. 92 argue for smaller stature estimates. The projected body mass for an equiv- alent adult is 68 kg. 90 In 2001, a 1.55 Ma calvaria (KNM- ER 42700) was recovered from the Koobi Fora Formation at Ileret. 93 Its dimensions and estimated endocra- nial volume are smaller than the smallest calvaria from Dmanisi. It is small enough to overlap with some of the earlier crania included in H. habilis , yet it also displays morpho- logical features typical of Asian H. erectus . The initial assessment of the KNM-ER 42700 calvaria 93 suggested that the small size is consistent with a high degree of sexual dimorphism (as compared, for example, with OH 9) within a single African and Asian Homo erectus taxon. Fragmentary but taxonomically distinctive evidence in the form of an occipital (KNM-ER 2598) from the Upper Burgi member located 4 m below the KBS Tuff provides the cur- rent ca 1.87 Ma first appearance date for H. ergaster . The last appear- ance date is c . 1.5 Ma. Homo rhodesiensis Woodward, 1921 or Homo helmei Dreyer, 1935 At least two hominin cranial speci- mens from the Omo-Turkana Basin evidently belong to Homo , but they lack the distinctive cranial and man- dibular morphology of H. ergaster or Homo sapiens . Two taxa have been proposed for comparable fossil evi- dence from other African sites. The first, Homo rhodesiensis , was intro- duced by Sir Arthur Smith Wood- ward for the cranium (Kabwe 1 or E 686) and limb bones recovered from the Broken Hill lead mine at Kabwe in Zambia. 94 Woodward reasoned that a new species of Homo was needed because the Broken Hill fos- sils did not match the three Homo taxa known at that time, H. sapiens , H. erectus and H. neanderthalensis ). For some, H. rhodesiensis is a junior subjective synonym of Homo heidel- bergensis However, for those who regard H. heidelbergensis as an exclu- sively European pre-modern Homo taxon that already shows evidence of distinctive H. neanderthalensis mor- phology, there is a need for a sepa- rate African pre-modern Homo taxon. A second taxon, Homo helmei , was established to accommodate a partial cranium discovered in 1932 by T. F. Dreyer in Florisbad, South Africa. 95 This taxon is used by those who regard the morphology of the Florisbad 1 cranium as intermediate between H rhodesiensis and H. sapi- ens . Its vertical face, brow ridge size, and frontal squama inclination dis- tinguish it from the former, yet it retains features such as a large brow ridge, receding frontal, and broad lower vault that distinguish it from anatomically modern humans. Other specimens suggested to belong to the same ta