New hominin genus from eastern Africa shows diverse - Biolinguagem

New hominin genus from eastern Africa shows diverse - Biolinguagem

articles New hominin genus from eastern Africa shows diverse middle Pliocene lineages Meave G. Leakey*, Fred Spoor², Frank H. Brown³, Patrick N. Gath...

616KB Sizes 0 Downloads 0 Views

Recommend Documents

Iridoids from Scrophularia Genus
01.01.2015 - side and harpagide are of particular interest to botanists because of their potential ... This work has bee

Central and Eastern Africa : IRIN update from 97.11.17 to 15.01.98
Gerishon Kirima said insanitary condition's in Nairobi's slums had contributed to the outbreak. Humanitarian sources tol

SEVEN NEW BARIUM MINERALS FROM EASTERN FRESNO
Seven new barium-bearing silicate minerals have been found in eastern Fresno County, california. The minerals occur in s

A new araneid genus from the Galapagos Islands (Araneae: Araneidae)
For some years I have been puzzled about the generic placement of a male specimen from the Galapagos Islands. Additional

A. NEW GENUS OF 'WiNTER MOTHS
Nov 17, 1980 - EASTERN CALIFORNIA AND WESTERN NEVADA imam A. PQWELL .... in. female Winter moths in the Northern Hemisph

A New Species of the Scorpionfish Genus Maxillicosta from the
Mar 7, 2005 - A New Species of the Scorpionfish Genus Maxillicosta from the Southeast. Coast of .... CSIRO H 4358-01, 52

Early hominin diversity and the emergence of the genus - Isita-org.com
Ian Tattersall, and the two reviewers of this paper. References. Alemseged Z. .... Guy F., Pilbeam D., Likius A., Mackay

OTEPIC International Permaculture School in Eastern Africa
OTEPIC International Permaculture School in Eastern Africa. Summary: In Kitale, Western Kenya, OTEPIC ("Organic Technolo

Adetogramma (Polypodiaceae), a new monotypic fern genus
Bachman S, Moat J, de la Torre J, Scott B (2011) Supporting Red List threat ... Bon eld JK, Smith KF, Staden R (1995) A

A new genus and two new species of land snail from geographically
Oct 31, 2016 - The Camaenidae is a family of land snails that is particularly diverse in the rainforests of eastern Quee

articles

New hominin genus from eastern Africa shows diverse middle Pliocene lineages Meave G. Leakey*, Fred Spoor², Frank H. Brown³, Patrick N. Gathogo³, Christopher Kiarie*, Louise N. Leakey* & Ian McDougall§ * Division of Palaeontology, National Museums of Kenya, P.O. Box 40658, Nairobi, Kenya ² Department of Anatomy & Developmental Biology, University College London, WC1E 6JJ, UK ³ Department of Geology & Geophysics, University of Utah, Salt Lake City, Utah 84112, USA § Research School of Earth Sciences, The Australian National University, Canberra ACT 0200, Australia

............................................................................................................................................................................................................................................................................

Most interpretations of early hominin phylogeny recognize a single early to middle Pliocene ancestral lineage, best represented by Australopithecus afarensis, which gave rise to a radiation of taxa in the late Pliocene. Here we report on new fossils discovered west of Lake Turkana, Kenya, which differ markedly from those of contemporary A. afarensis, indicating that hominin taxonomic diversity extended back, well into the middle Pliocene. A 3.5 Myr-old cranium, showing a unique combination of derived facial and primitive neurocranial features, is assigned to a new genus of hominin. These ®ndings point to an early diet-driven adaptive radiation, provide new insight on the association of hominin craniodental features, and have implications for our understanding of Plio±Pleistocene hominin phylogeny. The eastern African hominin record between 4 and 3 Myr is represented exclusively by a single species, A. afarensis, and its possible ancestor, Australopithecus anamensis, which are commonly thought to belong to the lineage ancestral to all later hominins1,2. This apparent lack of diversity in the middle Pliocene contrasts markedly with the increasingly bushy phylogeny evident in the later hominin fossil record. To study further the time interval between 4 and 3 Myr, ®eldwork in 1998 and 1999 focused on sites of this age at Lomekwi in the Nachukui Formation, west of Lake Turkana. New hominin discoveries from Lomekwi, as well as two mandibles and isolated molars recovered previously3 (Table 1), indicate that multiple species existed between 3.5 and 3.0 Myr. The new ®nds include a well-preserved temporal bone, two partial maxillae, isolated teeth, and most importantly a largely complete, although distorted, cranium. We assign the latter specimen to a new hominin genus on the basis of its unique combination of primitive and derived features.

Description of Kenyanthropus platyops

Order Primates LINNAEUS 1758 Suborder Anthropoidea MIVART 1864 Superfamily Hominoidea GRAY 1825 Kenyanthropus gen. nov. Etymology. In recognition of Kenya's contribution to the understanding of human evolution through the many specimens recovered from its fossil sites. Generic diagnosis. A hominin genus characterized by the following morphology: transverse facial contour ¯at at a level just below the nasal bones; tall malar region; zygomaticoalveolar crest low and curved; anterior surface of the maxillary zygomatic process positioned over premolars and more vertically orientated than the nasal aperture and nasoalveolar clivus; nasoalveolar clivus long and both transversely and sagittally ¯at, without marked juga; moderate subnasal prognathism; incisor alveoli parallel with, and only just anterior to, the bicanine line; nasal cavity entrance stepped; palate roof thin and ¯exed inferiorly anterior to the incisive foramen; upper incisor (I1 and I2) roots near equal in size; upper premolars (P3, P4) mostly three-rooted; upper ®rst and second molars (M1 and M2) small with thick enamel; tympanic element mediolaterally long and lacking a petrous crest; external acoustic porus small. Kenyanthropus can be distinguished from Ardipithecus ramidus by its buccolingually narrow M2, thick molar enamel, and a temporal NATURE | VOL 410 | 22 MARCH 2001 | www.nature.com

bone with a more cylindrical articular eminence and deeper mandibular fossa. It differs from A. anamensis, A. afarensis, A. africanus and A. garhi in the derived morphology of the lower face, particularly the moderate subnasal prognathism, sagittally and transversely ¯at nasoalveolar clivus, anteriorly positioned maxillary zygomatic process, similarly sized I1 and I2 roots, and small M1 and M2 crowns. From A. afarensis it also differs by a transversely ¯at midface, a small, external acoustic porus, and the absence of an occipital/marginal venous sinus system, and from A. africanus by a tall malar region, a low and curved zygomaticoalveolar crest, a narrow nasal aperture, the absence of anterior facial pillars, a tubular, long and crestless tympanic element, and a small, external acoustic porus. Kenyanthropus lacks the suite of derived dental and cranial features found in Paranthropus aethiopicus, P. boisei and P. robustus (Table 2), and the derived cranial features of species indisputably assigned to Homo (For example, H. erectus s.l. and H. sapiens, but not H. rudolfensis and H. habilis)4. Type species Kenyanthropus platyops sp. nov. Etymology. From the Greek platus, meaning ¯at, and opsis, meaning face; thus referring to the characteristically ¯at face of this species. Speci®c diagnosis. Same as for genus. Types. The holotype is KNM-WT 40000 (Fig. 1a±d), a largely complete cranium found by J. Erus in August 1999. The paratype is KNM-WT 38350 (Fig. 1e), a partial left maxilla found by B. Onyango in August 1998. The repository is the National Museums of Kenya, Nairobi. Localities. Lomekwi localities are situated in the Lomekwi and Topernawi river drainages in Turkana district, northern Kenya (Fig. 2). The type locality LO-6N is at 038 54.039 north latitude, 0358 44.409 east longitude. Horizon. The type specimen is from the Kataboi Member, 8 m below the Tulu Bor Tuff and 12 m above the Lokochot Tuff, giving an estimated age of 3.5 Myr. The paratype is from the lower Lomekwi Member, 17 m above the Tulu Bor Tuff, with an estimated age of 3.3 Myr.

Cranial description and comparisons

The overall size of the KNM-WT 40000 cranium falls within the range of A. afarensis and A. africanus. It is preserved in two main parts, the neurocranium with the superior and lateral orbital margins, but lacking most of the cranial base; and the face, lacking

© 2001 Macmillan Magazines Ltd

433

articles the premolar and anterior tooth crowns and the right incisor roots. Most of the vault is heavily distorted, both through post-mortem diploic expansion and compression from an inferoposterior direction (Fig. 1a, b). The better preserved facial part shows some lateral skewing of the nasal area, anterior displacement of the right canine,

and some expansion of the alveolar and zygomatic processes (Fig. 1c±d), but allows for reliable assessment of its morphology. Only the right M2 crown is suf®ciently preserved to allow reliable metric dental comparisons. It is particularly small, falling below the known ranges of other early hominin species (Fig. 3a). Likewise, the

Table 1 Hominin specimens from the lower Lomekwi and Kataboi Members KNM-WT

Description

Year

Mandible fragment: symphysis, right body with RP3 ±RM1, isolated partial RM2, RM3, LP3 LM1/2 RM3 Left mandible fragment with M2 fragment and M3 Partial RM3 crown LM1/2 crown LM1/2 RM1/2 crown fragment RM1/2 Partial RM1/2 crown LM1/2 crown Partial LM2/3 LM1/2 crown Right maxilla fragment with I2 and P3 roots and partial C; mandible fragment with partial P4 and M1 roots RM1/2 crown Partial RM1/2 LdM2 crown RM1/2 crown Left maxilla fragment with P3 and P4 roots and partial M1 Partial RM1/2 Partial RM1/2 crown Partial RM1/2 crown

1982

N. Mutiwa

Discoverer

LO-5

Locality

1982 1985 1985

N. Mutiwa M. Kyeva N. Mutiwa

LO-4 LO-5 LO-4E

RP3, 9.8, 12.4; RP4, 11.3, 12.6; RM1, 13.7, 12.9; RM2, NA, NA; RM3, (17.5), (14.1); LP3, 9.8, 12.5 NA, (11.5) 13.3, 14.6 M2, NA, NA; M3, 15.3, 13.1

Measurements (mm)

1999 1999 1999 1999 1999 1999 1999 1999 1999 1999

M. Eregae M. Eregae M. Eregae M. Eregae R. Moru N. Mutiwa J. Erus G. Ekalale J. Erus J. Erus

LO-4E LO-4E LO-4W LO-4E LO-4E LO-4E LO-4W LO-4E LO-4E LO-4W

NA, 14.8 13.1, 12.1 12.1, 11.5 NA 11.5, 12.3 NA 12.8, 12.7 NA 12.8, (11.3) NA

1998 1998 1998 1998 1998

M. Eregae M. Mutiwa R. Moru W. Mangao B. Onyango

LO-9 LO-5 LO-5 LO-5 LO-5

12.8, 12.2 NA 11.7, 9.6 13.5, 12.6 LM1: (10.5), (12.0)

1998 1998 1998

LO-5 LO-9 LO-9

NA, 11.5 NA 12.8, NA

1998 1998

W. Mangao M. Eregae M. Eregae & J. Kaatho G. Ekalale G. Ekalale

LO-5 LO-5

12.8, 11.8 RI2, 7.5, 7.5, 9.1; LM3, 15.3, 13.2

1998 1998

M. Eregae R. Moru

LO-5 LO-5

1998 1998 1998 1998 1998 1998 1998 1998 1999 1998

R. Moru R. Moru R. Moru R. Moru R. Moru R. Moru R. Moru R. Moru J. Erus P. Gathogo

LO-5 LO-5 LO-5 LO-5 LO-5 LO-5 LO-5 LO-5 LO-6N LO-5

RM1, 12.7, 11.6; RM2, 13.9, 12.2 I1, NA, (8.0), (11.5); LI2, 7.6, .5.9, 8.3; LP3, (9.3), (12.0) RM1/2, 12.9, 14.3 NA 16.0, 14.5 NA NA NA NA NA RM2, 11.4, 12.4 NA

...................................................................................................................................................................................................................................................................................................................................................................

8556

8557 16003 16006 38332 38333 38334 38335 38337 38338 38339 38341 38342 38343 38344 38346 38347 38349 38350 38352 38355 38356 38357 38358 38359 38361 38362 39949 39950 39951 39952 39953 39954 39955 40000 40001

RM1/2 Associated RI2, LM2 fragment, LM3, RM3 fragment, four crown fragments Associated RM1, RM2 Associated (partial) germs of I1, LI2, RC, LRP3, LRP4 Associated partial LM1/2, RM1/2 Partial LP4 RM3 RM1/2 fragment LM1/2 LM1/2 fragment Two tooth fragments LC fragment Cranium Right temporal bone

................................................................................................................................................................................................................................................................................................................................................................... Dental measurements taken as in ref. 34. Mesiodistal crown diameter followed by buccolingual or labiolingual diameter, and for incisors and canines, labial crown height. Values in parentheses are estimates. NA, Not available. L or R in the `Description' column indicates the left or right side. C, upper canine; d, deciduous.

Table 2 Derived cranial features of Paranthropus, and their character state in K. platyops and H. rudolfensis Paranthropus aethiopicus

Paranthropus boisei

Paranthropus robustus

Kenyanthropus platyops

Homo rudolfensis

Large Hyperthick Thick Present Gutter Strong Medial Smooth Straight, high Present

Large Hyperthick Thick Present Gutter Moderate Medial Smooth Straight, high Present

Moderate Hyperthick Thick Present Gutter Moderate Medial Smooth Straight, high Present

Small Thick Thin Present Flat Weak Lateral Stepped Curved, low Present

Moderate Thick Thin Present Flat Weak Lateral Stepped Curved, low Present

Concave, dished Wide Aligned High, present Marked Medial

Concave, dished Wide Aligned High, present Marked Medial

Concave, dished Wide Aligned High, present Marked Medial

Flat Tall More vertical Low, absent Moderate Posteromedial

Flat Tall More vertical Low, absent Moderate Posteromedial

Present Lateral Shallow Present Unknown

Present Lateral Deep Present Present

Present Lateral Deep Absent present

Absent Medial Moderate Absent Absent

Absent Medial Moderate Absent Absent

...................................................................................................................................................................................................................................................................................................................................................................

Upper molar size Enamel thickness Palatal thickness Incisor alveoli close to bicanine line* Nasoalveolar clivus Midline subnasal prognathism Upper I2 root to lateral nasal aperture Nasal cavity entrance Zygomaticoalveolar crest Anteriorly positioned zygomatic process of maxilla* Midface transverse contour Malar region Malar orientation to lateral nasal margin Facial hafting, frontal trigone Postorbital constriction Initial supraorbital course of temporal lines Tympanic vertically deep and plate-like Position external acoustic porus Mandibular fossa depth Foramen magnum heart shaped Occipitomarginal sinus

................................................................................................................................................................................................................................................................................................................................................................... Hypodigm of H. rudolfensis as in ref. 35. See refs 1, 8, 11, 36±40 for detailed discussions of the features. * Character states shared by Paranthropus and K. platyops.

434

© 2001 Macmillan Magazines Ltd

NATURE | VOL 410 | 22 MARCH 2001 | www.nature.com

articles estimated M1 crown size of KNM-WT 38350 (Table 1) corresponds to minima for A. anamensis, A. afarensis and H. habilis, and is below the ranges for other African early hominins5±7. Molar enamel thickness in both specimens is comparable to that in A. anamensis and A. afarensis. CT scans show that both P3 and P4 of KNM-WT 40000 have a lingual root and two well-separated buccal roots. This morphology, thought to be the ancestral hominoid condition8, is commonly found in Paranthropus, but is variable among species of Australopithecus. The P3 of KNM-WT 38350 has three wellseparated roots (Fig. 1e). Its P4 seems to be two-rooted, but the deeply grooved buccal root may split more apically. Relative to M2 crown size, the canine roots of KNM-WT 40000 are smaller in crosssection at the alveolar margin than in Ardipithecus ramidus and A. anamensis, similar in size to A. afarensis, A. africanus and H. habilis, and larger than in P. boisei. Exposed surfaces and CT scans demonstrate that the I1 and I2 roots in KNM-WT 40000 are straight and similar in size. At the level of the alveolar margin the cross-sectional area of the I2 root is about 90% of that of the I1 root, whereas this is typically 50±70% in other known hominid taxa. The incisor alveoli of KNM-WT 40000 are aligned coronally, just anterior to the bicanine line, and the overlying nasoalveolar clivus is ¯at both sagittally and transversely. There is no canine jugum visible on the preserved left side, re¯ecting the modest size of the canine root. At 32 mm (chord distance nasospinale to prosthion) the clivus is among the longest of all early hominins. Subnasal prognathism is

moderate, expressed by a more vertically orientated clivus than in nearly all specimens of Australopithecus and Paranthropus (Fig. 3b). The nasal aperture lies in the same coronal plane as the nasoalveolar clivus and there are no anterior facial pillars (Fig. 1a±c). The nasal aperture is small and narrow, in contrast to the large, wide aperture in A. africanus and P. robustus. The midface of KNM-WT 40000 is dominated by the tall malar region (Fig. 3c) with a low and curved zygomaticoalveolar crest. At a level just below the nasal bones the transverse facial contour is ¯at (Fig. 1b). In both KNM-WT 40000 and KNM-WT 38350 the anterior surface of the zygomatic process of the maxilla is positioned between P3 and P4 (Fig. 1a, d, e), as is commonly seen in Paranthropus, but more anteriorly than in most Australopithecus specimens9 or in H. habilis. The supraorbital region is Australopithecus-like, lacking both a frontal trigon as seen in Paranthropus, and a supratoral sulcus as seen in H. habilis (but not H. rudolfensis). Relative postorbital constriction (frontofacial index) of KNM-WT 40000 is similar to that in Australopithecus, H. rudolfensis and H. habilis, and less than in P. boisei (estimated frontofacial index9 = 70). Its temporal lines converging on the frontal squama have a posteromedial course throughout (Fig. 1b). Around bregma the midline morphology is not well preserved, but the posterior half of the parietals show double, slightly raised temporal lines about 6 mm apart. These contribute posteriorly to indistinct compound temporal/nuchal lines. The original shape of the severely distorted mastoids cannot be reconstructed, but other

Figure 1 Holotype KNM-WT 40000 a, left lateral view (markers indicate the plane separating the distorted neurocranium and the well-preserved face). b, Superior view.

c, Anterior view. d, Occlusal view of palate. Paratype KNM-WT 38350. e, Lateral view. KNM-WT 40001. f, Lateral view. g, Inferior view. Scale bars: a±c, 3 cm; d±g, 1 cm.

NATURE | VOL 410 | 22 MARCH 2001 | www.nature.com

© 2001 Macmillan Magazines Ltd

435

articles parts of the left temporal are well preserved. The tubular tympanic lacks a petrous crest and forms a narrow external acoustic meatus with a small aperture. This combination constitutes the primitive hominin morphology, also seen in Ar. ramidus and A. anamensis (Fig. 3d). The mandibular fossa resembles that of specimens of A. afarensis and A. africanus. It is moderately deep, and the articular eminence, missing its lateral margin, is cylindrical with a moderately convex sagittal pro®le. The preserved posterior half of the foramen magnum suggests that it was probably oval in shape, rather than the heart shape seen in P. boisei and probably P. aethiopicus. Regarding the endocranial aspect, the reasonably well preserved occipital surface lacks any indication of the occipital/marginal venous sinus system characteristic of A. afarensis, P. boisei and P. robustus. Bilateral sulci suggest that the transverse/sigmoid sinus system was well developed. Endocranial capacity is dif®cult to estimate because of the distorted vault. However, comparing hominin glabella±opisthion arc lengths8 with that of KNM-WT 40000 (259 mm; an estimate in¯ated by diploic expansion) suggests a value in the range of Australopithecus or Paranthropus. The sex of KNM-WT 40000 is dif®cult to infer. Interpretation of the canine root size proves inconclusive without a suitable comparative context. The small M2 crown size could suggest that the specimen is female. However, the close proximity and slightly raised aspect of the temporal lines on the posterior half of the parietals is not seen in known female hominin crania, including the Paranthropus specimens KNM-ER 732, KNM-ER 407 and DNH7, and suggests that KNM-WT 40000 could be male. With incisor alveoli close to the bicanine line and anteriorly positioned zygomatic processes, the face of KNM-WT 40000 resembles the ¯at, orthognathic-looking faces of both Paranthropus and H. rudolfensis cranium KNM-ER 1470. However, KNM-WT 40000 lacks most of the derived features that characterize Paranthropus (Table 2), and its facial architecture differs from the latter in much the same way as has been described for KNM-ER 1470 (refs 8, 10). Facial ¯atness in Paranthropus results from the

v

v

The right maxilla fragment KNM-WT 38343A preserves three wellseparated P3 roots, and its damaged canine seems low-crowned when compared with A. afarensis canines of similar size and degree of wear. The right temporal bone KNM-WT 40001 lacks the squama and petrous apex, but is otherwise well preserved (Fig. 1f, g). It shows a combination of characters not seen in any other hominin specimen. The projecting mastoid process is rounded, with an anteriorly positioned tip. It has a well-developed digastric fossa in the form of a deep, narrow groove that runs posterolaterally from the stylomastoid foramen, fully demarcating the mastoid process from the adjacent nuchal plane. The tympanic element is long, inferosuperiorly shallow and lacks a petrous crest. The external acoustic porus is the smallest of any known hominin temporal bone (Fig. 3d). The articular eminence is as broad mediolaterally (38 mm) as in P. aethiopicus and P. boisei, and similar to the largest found in A. afarensis. Compared with KNM-WT 40000 the eminence is relatively ¯at sagittally, and the mandibular fossa is shallow. The partial mandibles KNM-WT 8556 and KNM-WT 16006 have been assigned to A. afarensis3. However, KNM-WT 8556 shows a more derived morphology than this species by having a ¯at, more horizontal post-incisive plane, a more superiorly positioned genioglossal pit, a molarized lower fourth premolar (P4) and a large M3

v

v

v

v

v

.b. e ...

v

v

Additional material

.. L ... ...o k...a l..a l ..e i ... ... ... ... . . . . . . wi .. . . . ... LO-6N e.k ... 3°55' ... ... m Lo . . . . . . ... . ... . ... ... .. ... . .. ... ... . ... .... ... ... ... ... ... ... ... . . .. ... ... . LO-6S .. . . ... .. ... ... .. . . . ... ... .. ... ... ..... ... ... .... .... . . LO-4E ... ... .. LO-4W ... ... .. LO-9 .. ... ... ... .. ... . . . .. ... . LO-5 . ... ... .. . ... ... ... ... ... .. ... LO-10 . . . . . . . ... .. ... ... Ka n ga t u k u s eo ...... 3°50' ... ... ... ... ... . ... .. ... ... .... ... . n .. a w ... i ... ... .. . ... Lomekwi and Lake Lokalalei Members ...K a t Lonyumun and ... ...a b o i Turkana Kataboi Members ... ... ... ... Miocene volcanic rocks Kataboi

Na

36° E

forward position of the anteroinferiorly sloping malar region, whose main facial surface approximates the plane of the nasal aperture, but whose orientation contrasts with the more horizontally inclined nasoalveolar gutter11. In KNM-WT 40000 and KNMER 1470, it is the ¯at and orthognathic nasoalveolar clivus that aligns with the plane of the nasal aperture, whereas the anteriorly set, tall malar region is more vertically orientated. KNM-WT 40000 lacks the derived short nasal bones and everted lateral nasal margin of KNM-ER 1470, and is less orthognathic in the midfacial region than this specimen; however, on balance this is the hominin face that KNM-WT 40000 most closely resembles.

v

v

ti

v

li

v

v

v

v

v

v

o Ri v e

v

v

v

v

v

v

r

v

v

v

v

v

v

v

v

v

.Ileret

v

Om

5° N

v

v

v

v

v

v

v

v

v

v

v

v

v

v

v

v

v

v

v

v

v

v

v

v

v

v

v

v

v

v

v

v

v

v

v

v

v

v

v

v

3° N

r

v

v

v

pe

v

v

v

Lodwar

To

v

v

v

v

a

..

v

v

v

v

an

.

v v

urk e T

Kalakol

v

.Koobi Fora

Lak

4° N

v

v

v

v

v

v

v

v

v

v

v

0

50

v

v

v

v

v

v

36° E

v

v

v

v

v

v

0

Figure 2 Map showing localities of fossil collection in upper Lomekwi and simpli®ed geology. The boundary between the Kataboi and Lomekwi Members is the base of the Tulu Bor Tuff, indicated as a dashed line through LO-4E and LO4-W. Faults are shown as 436

v

km

v

km

v

v

5

3°45' 35°50'

thick lines; minor faults are omitted. LO-4E and LO4-W are of different shades to distinguish them from each other.

© 2001 Macmillan Magazines Ltd

NATURE | VOL 410 | 22 MARCH 2001 | www.nature.com

articles (ref. 3). Indeed, relative to its Australopithecus-sized M1 (refs 5, 6, 12, 13), the P4 and M3 crowns of KNM-WT 8556 are enlarged to an extent only seen in P. boisei (Fig. 3e, f). All unworn molars in the Lomekwi sample are characterized by low occlusal relief and numerous secondary ®ssures. Most of the lower molars, including the KNM-WT 16006 M3, have a well-developed protostylid, a feature that is usually absent in A. afarensis, but common in A. africanus14. The two I2s are lower crowned than in A. afarensis, A. africanus14 and P. robustus14. Inability to distinguish between ®rst and second molars makes meaningful intertaxon comparisons of these elements dif®cult.

considerably earlier. Furthermore, the presence in K. platyops of an anteriorly positioned zygomatic process in combination with a small M1 and M2 indicates that such characters are more independent than is suggested by developmental and functional models that link such facial morphology in Paranthropus with postcanine megadontia11,15. At present it is unclear whether the Lomekwi hominin fossils sample multiple species. Apart from the paratype maxilla KNM-WT 38350 with its small molar size and anteriorly positioned zygomatic process, the other specimens cannot be positively associated morphologically with the K. platyops holotype. These are therefore not included in the paratype series, and are left unassigned until further evidence emerges. Differences between the tympanic and mandibular fossa morphologies of the KNM-WT 40000 and KNM-WT 40001 temporal bones can perhaps be accommodated within a single species, but their shared primitive characters do not necessarily imply conspeci®city. Af®liation of the KNM-WT 8556 mandible with the K. platyops types is not contradicted by its molarized P4, which is consistent with an anteriorly positioned zygomatic process. However, its M1 is larger than would be inferred from the smaller upper molars of the types, and with a 177 mm2 crown area it is also larger than any in the combined sample of ten isolated M1s and M2s (139±172 mm2). One isolated M1/2 (KNM-WT 38362) is signi®cantly larger than the molars of the K. platyops types, whereas another (KNM-WT 38337) is similar in size to the holotype's M2. The marked differences of the KNM-WT 40000 cranium from established hominin taxa, both with respect to individual features

Taxonomic discussion

The hominin specimens recovered from the Kataboi and lower Lomekwi Members show a suite of features that distinguishes them from established hominin taxa, including the only contemporaneous eastern African species, A. afarensis. Compared with the latter, the morphology of K. platyops is more derived facially, and more primitive in its small external acoustic porus and the absence of an occipital/marginal sinus system. These ®nds not only provide evidence for a taxonomically more diverse middle Pliocene hominin record, but also show that a more orthognathic facial morphology emerged signi®cantly earlier in hominin evolutionary history than previously documented. This early faciodental diversity concerns morphologies that functionally are most closely associated with mastication. It suggests a diet-driven adaptive radiation among hominins in this time interval, which perhaps had its origins

55 50

P. boisei (4)

H. rudolfensis (1)

H. habilis (2)

P. boisei (4)

H. habilis (2)

P. robustus (2)

P. aethiopicus (1) P. robustus (6)

H. rudolfensis (1)

A. africanus (5) A. afarensis (6)

A. africanus (2)

H. habilis (4)

P. boisei (5)

H. rudolfensis (1)

P. aethiopicus (1)

A. afarensis (6)

A. garhi (1)

35 P. robustus (10)

20 A. africanus (9)

40

WT 40000

25

WT 40000

45

A. afarensis (2)

30

60

Ar. ramidus (1)

Relative molar height (%)

35

A. anamensis (1)

P. boisei (4)

40

H. rud.-hab. (8)

A. garhi (1)

P. robustus (24)

A. afarensis (7)

WT 40000

45

e

180

f

100

150 P4 area / M1 area (%)

100

60

80

70

130

110

Figure 3 Mean and range of characters of speci®ed hominins. a, Square root of M2 crown area (buccolingual ´ mesiolingual diameters). b, Angle of subnasal prognathism (nasospinale±prosthion to postcanine alveolar plane). c, Malar height8 relative to orbitoalveolar height (orbitale to alveolar margin aligned with malar surface). d, Area of the NATURE | VOL 410 | 22 MARCH 2001 | www.nature.com

WT 8556

H. habilis (3)

H. rudolfensis (1)

P. boisei (4)

90 P. robustus (6)

A. afarensis (9)

A. africanus (3)

Ar. ramidus (1)

A. anamensis (2)

WT 8556

H. habilis (5)

P. boisei (6)

H. rudolfensis (1)

P. robustus (2)

P. aethiopicus (1)

A. africanus (5)

50 A. afarensis (4)

A. anamensis (1)

Ar. ramidus (2)

60

WT 40000

20

M3 area / M1 area (%)

90 140

WT 40001

Area of external acoustic porus (mm2)

d

Ar. ramidus (2)

11

A. africanus (23)

13

70 65

50

Subnasal prognathism (degrees)

15

A. anamensis (5)

Square root M2 area (mm2)

17

c

55

19

A. anamensis (2)

b

a

external acoustic porus (p ´ long axis ´ short axis). e, Crown area of P4 relative to that of M1 ´100. f, Crown area of M3 relative to that of M1 ´100. All measurements are taken from originals, directly or as given in refs 8, 9, 14, 18, 37, 41±44, except for some South African crania taken from casts. Numbers in parentheses indicate sample size.

© 2001 Macmillan Magazines Ltd

437

articles and their unique combination, fully justify its status as a separate species. It is worth noting that comparisons with Australopithecus bahrelghazali cannot be made directly, because this species was named on the basis of the limited evidence provided by an anterior mandible fragment16. Speci®c distinction of A. bahrelghazali from A. afarensis has yet to be con®rmed17, and Lomekwi specimens differ from A. bahrelghazali in symphyseal morphology and incisor crown height. The generic attribution of KNM-WT 40000 is a more complex issue, in the absence of consensus over the de®nition of the genus category4. The specimen lacks almost all of the derived features of Paranthropus (Table 2), and there are no grounds for assigning it to this genus unless it can be shown to represent a stem species. However, the fact that the facial morphology of KNM-WT 40000 is derived in a markedly different way renders this implausible. As KNM-WT 40000 does not show the derived features associated with Homo4 (excluding H. rudolfensis and H. habilis) or the strongly primitive morphology of Ardipithecus18, the only other available

Thickness (m) Member 80

LO-4E LO-4W .. . . . . .......... . ................... . .. ... . .. .... .... ......... .. ....... ... . . . . . . . ................ ................ .. . . . ....... . .. .. . . .. .............. ............ . ............ . . .. ... ......... ..... ............. . . . .. .......... ....... 38337 . . ... .. ........ 38342 ............ . 38338 .. ... .... .... ....... ....... . . 38343 ............ ... ...... . . .. .......... .. . . 38334 .......... ......... . . . . . . . . . . .. 16006; 38332, -35 .... ... . . . .. . .. .. ............... . . . . .. . .. . .. . .. ..

.......... . . .. .. ..... .... .... ........... ............... . . ..

70

60

50

40

genus is Australopithecus. We agree with the taxonomically conservative, grade-sensitive approach to hominin classi®cation that for the moment accepts Australopithecus as a paraphyletic genus in which are clustered stem species sharing a suite of key primitive features, such as a small brain, strong subnasal prognathism, and relatively large postcanine teeth. However, with its derived face and small molar size, KNM-WT 40000 stands apart from species assigned to Australopithecus on this basis. All it has in common with such species is its small brain size and a few other primitive characters in the nasal, supraorbital and temporal regions. Therefore, there is no ®rm basis for linking KNM-WT 40000 speci®cally with Australopithecus, and the inclusion of such a derived but early form could well render this genus polyphyletic. In a classi®cation in which Australopithecus also includes the `robust' taxa and perhaps even species traditionally known as `early Homo'4, this genus subsumes several widely divergent craniofacial morphologies. It could thus be argued that the inclusion of KNM-WT 40000 in Australopithecus would merely add yet another hominin species

L o m e k w i

30

20

LO-5 (LO-9) ............. . .. . . .. .. . .. .. . . ... ...... .... .... . ....... .. ........ . .............. . . 38344, -55, -56 .................. .. . ..... ............. .. .. . .................. 38352,-57, -58; 40001 ........ . 38346, -49, -59 .... .... .. 8556, 8557,16003; 38347, ... .. ... 38350, -61, -62; 39949, -50, -51, -52, -53, -54, -55 ... ..... .. ... .... .. . . . . . . . .. ................ .... . .... β

....... .

LO-6N

α

......... ..

Tulu Bor Tuff (3.40 ± 0.03 Myr)

10

0

K a . . . .. 40000 . . t a b o .. . i ..... ............. ... . .. . .

– 10

.. . . . . . 38333

38339 Burrowed bed

38341

.. . ............... ..... ... . ... ........ .. LO-6S (pumice site) ........... .. .. .... ..... . .... .... .. .......... ......... . ..

Topernawi

..... ..... ................ ......... .. . ... . .... .... ...

–20

–30

–40

Lokochot Tuff (3.57 ± 0.05 Myr)

Moiti Tuff (3.94 ± 0.03 Myr) L o n y u m u n

Topernawi Tuff (3.96 ± 0.03 Myr)

Claystone Mudstone Siltstone

'Nabwal Tuff'

Silty sandstone

. .

Sandstone

........... Conglomerate Tuff ( ^ =altered)

Figure 4 Stratigraphic sections and placement of hominin specimens at sites in upper part of the Lomekwi drainage, west of Lake Turkana, northern Kenya. Specimen numbers are given without the pre®x KNM-WT, and those in bold are discussed in the text. Placement of specimens is relative to the nearest marker bed in the section. Italicized numbers show the relative placement of specimens at LO-9 on section LO-5. The burrowed bed, a useful local marker, is used as stratigraphic datum (0 m). Representative 40 Ar± 39Ar analytical data on the Moiti Tuff and on the Topernawi Tuff are given as 438

Supplementary Information. The date for the Tulu Bor Tuff is taken as the age of the Sidi Hakoma Tuff at Hadar25, which is consistent with the age of the Toroto Tuff (3.32 6 0.03 Myr)45 that overlies the Tulu Bor at Koobi Fora. The age of the Lokochot Tuff is assigned from its placement at the Gilbert/Gauss Chron boundary24,46. The tuff formerly thought to be the Moiti Tuff at Lomekwi22 has been informally called the `Nabwal tuff'47. Ages on the Tulu Bor and the Lokochot Tuffs are consistent with orbitally tuned ages of correlative ash layers in Ocean Drilling Program Core 722A in the Arabian Sea48.

© 2001 Macmillan Magazines Ltd

NATURE | VOL 410 | 22 MARCH 2001 | www.nature.com

articles with a derived face. This amounts to de®ning Australopithecus by a single criterion, those hominin species not attributable to Ardipithecus or Homo, which in our view constitutes an undesirable approach to classi®cation. Thus, given that KNM-WT 40000 cannot be grouped sensibly with any of the established hominin genera, and that it shows a unique pattern of facial and dental morphology that probably re¯ects a distinct dietary adaptive zone, we assign this specimen to the new genus Kenyanthropus. Despite being separated by about 1.5 Myr, KNM-WT 40000 is very similar in its facial architecture to KNM-ER 1470, the lectotype of H. rudolfensis. The main differences amount to the more primitive nasal and neurocranial morphology of KNM-WT 40000. This raises the possibility that there is a close phylogenetic relationship between the two taxa, and affects our interpretation of H. rudolfensis. The transfer of this species to Australopithecus has been recommended4,19, but Kenyanthropus may be a more appropriate genus. The identi®cation of K. platyops has a number of additional implications. As a species contemporary with A. afarensis that is more primitive in some of its morphology, K. platyops weakens the case for A. afarensis being the sister taxon of all later hominins, and thus its proposed transfer to Praeanthropus1,20. Furthermore, the morphology of K. platyops raises questions about the polarity of characters used in analyses of hominin phylogeny. An example is the species' small molar size, which, although probably a derived feature, might also imply that the larger postcanine dentition of A. afarensis or A. anamensis does not represent the primitive hominin condition. Finally, the occurrence of at least one additional hominin species in the middle Pliocene of eastern Africa means that the af®liation of fragmentary specimens can now be reassessed. For example, the attribution of the 3.3 Myr old KNM-ER 2602 cranial fragment to A. afarensis21 has been questioned8, and evaluating its af®nities with K. platyops is now timely.

Geological context and dating

KNM-WT 40000 was collected near the contact of the Nachukui Formation with Miocene volcanic rocks in the northern tributary of Lomekwi (Nabetili). It is situated 12 m above the Lokochot Tuff, and 8 m below the b-Tulu Bor Tuff (Fig. 4). Along Nabetili, the Lokochot Tuff is pinkish-grey and contains much clay and volcanic detritus. It is overlain by a volcanic pebble conglomerate, followed by a pale brown quartz-rich ®ne sandstone that includes a burrowed ®ne-sandstone marker bed 10±15 cm thick. The Lokochot Tuff is replaced by a thick volcanic clast conglomerate in the central part of Lomekwi. The contact between the ®ne sandstone and the overlying dark mudstone can be traced from Nabetili to the hominin locality. Locally the mudstone contains volcanic pebbles at the base, and it has thin pebble conglomerate lenses in the upper part at the hominin locality, and also contains CaCO3 concretions. The hominin specimen and other vertebrate fossils derive from this mudstone. Overlying the dark mudstone at the hominin site is a brown mudstone (8 m) that directly underlies the b-Tulu Bor Tuff. New 40Ar± 39Ar determinations on alkali feldspars from pumice clasts in the Moiti Tuff and the Topernawi Tuff, stratigraphically beneath the Lokochot Tuff, were instrumental in re-investigating the lower portion of this section. The new results yield a mean age for the Topernawi Tuff of 3.96 6 0.03 Myr; this is marginally older than the pooled age for the Moiti Tuff of 3.94 6 0.03 Myr. Previous investigations22,23 placed the Topernawi Tuff above the Moiti Tuff, mainly on the basis of the K/Ar ages on alkali feldspar from pumice clasts in the Topernawi Tuff (3.78, 3.71, 3.76 and 3.97 Myr, all 6 0.04 Myr)23. The older determination (3.97 Myr) was thought to result from contamination by detrital feldspar. South of Topernawi, however, the Topernawi Tuff has now been shown to underlie the Moiti Tuff, and to be in turn underlain by a tephra informally termed the `Nabwal tuff ', previously thought to be a Moiti Tuff correlative. The correct sequence is shown in Fig. 4, and NATURE | VOL 410 | 22 MARCH 2001 | www.nature.com

the new 40Ar± 39Ar age data on the Moiti Tuff and Topernawi Tuff are provided as Supplementary Information. Linear interpolation between the Lokochot Tuff (3.57 Myr old)24 and Tulu Bor Tuff (3.40 Myr)25 yields an age of 3.5 Myr for KNMWT 40000, and 3.53 Myr for the burrowed bed. KNM-WT 38341 from immediately below the burrowed bed has an age near 3.53 Myr. KNM-WT 38333 and 38339, from between the burrowed bed and the a-Tulu Bor Tuff lie between 3.4 and 3.5 Myr. Other specimens from LO-4, LO-5, and LO-6 lie 16±24 m above the bTulu Bor Tuff, with ages near 3.3 Myr. Assuming linear sedimentation between the Tulu Bor Tuff and the Lokalalei Tuff (2.5 Myr)23, specimens from LO-9 are around 3.2 Myr. The probable error on these age estimates is less than 0.10 Myr. Palaeogeographically, the mudstone that contained KNM-WT 40000 at LO-6N was deposited along the northern margin of a shallow lake that extended to Kataboi and beyond26,27. Laterally discontinuous volcanic pebble conglomerates within the mudstone record small streams draining from hills to the west. Carbonate concretions at the hominin level are probably pedogenic, and indicate regional conditions with net evaporative loss. Other specimens between the burrowed bed and the Tulu Bor Tuff were also preserved in lake-margin environments, as is the case for KNM-WT 38341 that was collected below the burrowed bed. At LO-5, and in the upper part of LO-4E, strata were laid down by ephemeral streams draining the basin margin, principally the ancestral Topernawi, which deposited gravels in broad, shallow channels, and ®ner grained materials in inter¯uves. Specimens preserved in ¯oodplain deposits of the ancestral Omo River that occupied the axial portion of the basin include those at LO-9, those less than 6 m above the Tulu Bor Tuff at LO-4E, and KNM-WT 38338. Thus, there is evidence for hominins occupying ¯oodplains of major rivers, alluvial fans, and lake-margin environments 3.0±3.5 Myr ago. There is reasonable evidence that water sources were available to these hominins in channels of the ephemeral streams, and also possibly as seeps or springs farther out into the basin.

Palaeoecology and fauna

Faunal assemblages from Lomekwi sites LO 4, LO 5, LO 6 and LO 9 indicate palaeoenvironments that were relatively well watered and well vegetated. The relative proportions of the bovids in the early collections from these sites indicate a mosaic of habitats, but with predominantly woodland and forest-edge species dominating22. Comparisons of the Lomekwi faunal assemblages with those from the few known hominin sites of similar age, Laetoli in Tanzania, Hadar in Ethiopia and Bahr el Ghazal in Chad, are of interest in view of the different hominin taxa represented. Hadar and Bahr el Ghazal, like Lomekwi, represent lakeshore or river ¯oodplain palaeoenvironments28,29, whereas Laetoli was not located near a water source; no aquatic taxa nor terrestrial mammals indicative of swamp or grassy wetlands were recovered30. The faunal assemblages of all four sites indicate a mosaic of habitats that seems to have included open grasslands and more wooded or forested environments22,28,29,31,32; the assemblages differ primarily in the indication of the nature of the dominant vegetation cover. Although the mammalian faunal assemblage from Lomekwi is more similar to that from Hadar than to that from Laetoli, some mammalian species represented are different. At Lomekwi, Theropithecus brumpti is common and is the dominant cercopithecid, as it is elsewhere in the Turkana Basin at this time. This species is generally considered to indicate more forested or closed woodland habitats. In the Hadar Formation, Theropithecus darti is the common Theropithecus species and is associated with lower occurrences of the water-dependent reduncines and higher occurrences of alcelaphines and/or Aepyceros, which indicates drier woodlands and grasslands33. Differences in the representation of other common species at the two sites that are less obviously linked to habitat include Kolpochoerus limnetes, Tragelaphus nakuae and Aepyceros

© 2001 Macmillan Magazines Ltd

439

articles shungurensis at Lomekwi, as opposed to K. afarensis, T. kyaloae and an undescribed species of Aepyceros at Hadar (K. Reed, personal communication). The general indication is that the palaeoenvironment at Lomekwi may have been somewhat more vegetated and perhaps wetter than that persisting through much of the Hadar Formation. At both sites more detailed analyses will be essential to further develop an understanding of how subtle temporal changes in the faunal assemblages relate to hominin occurrences. Note added in proof: If the hominin status of the recently published Lukeino craniodental specimens49 is con®rmed, this would support the suggestion that small molar size is the primitive rather than the derived hominin condition. M Received 31 January; accepted 16 February 2001. 1. Strait, D. S., Grine, F. E. & Moniz, M. A. A reappraisal of early hominid phylogeny. J. Hum. Evol. 32, 17±82 (1997). 2. Ward, C., Leakey, M. & Walker, A. The new hominid species Australopithecus anamensis. Evol. Anthrop. 7, 197±205 (1999). 3. Brown, B., Brown, F. & Walker, A. New hominids from the Lake Turkana Basin, Kenya. J. Hum. Evol. (in the press). 4. Wood, B. A. & Collard, M. C. The human genus. Science 284, 65±71 (1999). 5. Leakey, M. G., Feibel, C. S., McDougall, I. & Walker, A. C. New four-million-year-old hominid species from Kanapoi and Allia Bay, Kenya. Nature 376, 565±571 (1995). 6. Leakey, M. G., Feibel, C. S., McDougall, I., Ward, C. & Walker, A. New specimens and con®rmation of an early age for Australopithecus anamensis. Nature 393, 62±66 (1998). 7. Kimbel, W. H., Johanson, D. C. & Rak, Y. Systematic assessment of a maxilla of Homo from Hadar, Ethiopia. Am. J. Phys. Anthrop. 103, 235±262 (1997). 8. Wood, B. A. Koobi Fora Research Project Vol. 4 (Clarendon Press, Oxford, 1991). 9. Lockwood, C. A. & Tobias, P. V. A large male hominin cranium from Sterkfontein, South Africa, and the status of Australopithecus africanus. J. Hum. Evol. 36, 637±685 (1999). 10. Bilsborough, A. & Wood, B. A. Cranial morphometry of early hominids: facial region. Am. J. Phys. Anthrop. 76, 61±86 (1988). 11. Rak, Y. The Australopithecine Face (Academic, New York, 1983). 12. Lockwood, C. A., Kimbel, W. H. & Johanson, D. C. Temporal trends and metric variation in the mandibles and dentition of Australopithecus afarensis. J. Hum. Evol. 39, 23±55 (2000). 13. Moggi-Cecchi, J., Tobias, P. V. & Beynon, A. D. The mixed denitition and associated skull fragments of a juvenile fossil hominid from Sterkfontein, South Africa. Am. J. Phys. Anthrop. 106, 425±465 (1998). 14. Robinson, J. T. The dentition of the Australopithecinae. Transv. Mus. Mem. 9, 1±179 (1956). 15. McCollum, M. A. The robust australopithecine face: a morphogenetic perspective. Science 284, 301± 305 (1999). 16. Brunet, M. et al. Australopithecus bahrelghazali, une nouvelle espeÁce d'Hominide ancien de la reÂgion de Koro Toro (Tchad). C.R. Acad. Sci. Ser. IIa 322, 907±913 (1996). 17. White, T. D., Suwa, G., Simpson, S. & Asfaw, B. Jaws and teeth of A. afarensis from Maka, Middle Awash, Ethiopia. Am. J. Phys. Anthrop. 111, 45±68 (2000). 18. White, T. D., Suwa, G. & Asfaw, B. Australopithecus ramidus, a new species of early hominid from Aramis, Ethiopia. Nature 371, 306±312 (1994). 19. Wood, B. & Collard, M. The changing face of genus Homo. Evol. Anthrop. 8, 195±207 (2000). 20. Harrison, T. in Species, Species Concepts, and Primate Evolution (eds Kimbel, W. H. & Martin, L. B.) 345±371 (Plenum, New York, 1993). 21. Kimbel, W. H. Identi®cation of a partial cranium of Australopithecus afarensis from the Koobi Fora Formation. J. Hum. Evol. 17, 647±656 (1988). 22. Harris, J. M., Brown, F. & Leakey, M. G. Stratigraphy and paleontology of Pliocene and Pleistocene localities west of Lake Turkana, Kenya. Cont. Sci. Nat. Hist. Mus. Los Angeles 399, 1±128 (1988). 23. Feibel, C. S., Brown, F. H. & McDougall, I. Stratigraphic context of fossil hominids from the Omo Group deposits, northern Turkana Basin, Kenya and Ethiopia. Am. J. Phys. Anthrop. 78, 595±622 (1989). 24. McDougall, I., Brown, F. H., Cerling, T. E. & Hillhouse, J. W. A reappraisal of the geomagnetic polarity time scale to 4 Ma using data from the Turkana Basin, East Africa. Geophys. Res. Lett. 19, 2349±2352 (1992). 25. Walter, R. C. & Aronson, J. L. Age and source of the Sidi Hakoma Tuff, Hadar Formation, Ethiopia. J. Hum. Evol. 25, 229±240 (1993). 26. Brown, F. H. & Feibel, C. S. in Koobi Fora Research Project, Vol. 3, Stratigraphy, artiodactyls and paleoenvironments (ed. Harris, J. M.) 1±30 (Clarendon, Oxford, 1991). 27. Feibel, C. S., Harris, J. M. & Brown, F. H. in Koobi Fora Research Project, Vol. 3, Stratigraphy, artiodactyls and paleoenvironments (ed. Harris, J. M.) 321±346 (Clarendon, Oxford, 1991). 28. Johanson, D. C., Taieb, M. & Coppens, Y. Pliocene hominids from the Hadar Formation, Ethiopia

440

(1973±1977): stratigraphic, chronologic and paleoenvironmental contexts, with notes on hominid morphology and systematics. Am. J. Phys. Anthrop. 57, 373±402 (1982). 29. Brunet, M. et al. The ®rst australopithecine 2,500 kilometres west of the Rift Valley (Chad). Nature 378, 273±240 (1995). 30. Leakey, M. D. & Harris, J. M. Laetoli, a Pliocene Site in Northern Tanzania (Clarendon, Oxford, 1987). 31. Harris, J. M. in Laetoli, a Pliocene Site in Northern Tanzania (eds Leakey, M. D. & Harris, J. M.) 524± 531 (Clarendon, Oxford, 1987). 32. Kimbel, W. H. et al. Late Pliocene Homo and Oldowan tools from the Hadar Formation (Kadar Hadar Member), Ethiopia. J. Hum. Evol. 31, 549±561 (1996). 33. Eck, G. G. in Theropithecus, the Rise and Fall of a Primate Genus (ed. Jablonski, N. G.) 15±83 (Cambridge Univ. Press, 1993). 34. White, T. D. New fossil hominids from Laetolil, Tanzania. Am. J. Phys. Anthrop. 46, 197±230 (1977). 35. Wood, B. Origin and evolution of the genus Homo. Nature 355, 783±790 (1992). 36. Suwa, G. et al. The ®rst skull of Australopithecus boisei. Nature 389, 489±492 (1997). 37. Keyser, A. W. The Drimolen skull: the most complete australopithecine cranium and mandible to date. S. Afr. J. Sci. 96, 189±193 (2000). 38. Kimbel, W. H., White, T. D. & Johanson, D. C. Cranial morphology of Australopithecus afarensis: a comparative study based on a composite reconstruction of the adult skull. Am. J. Phys. Anthrop. 64, 337±388 (1984). 39. Walker, A., Leakey, R. E., Harris, J. H. & Brown, F. H. 2.5-Myr Australopithecus boisei from west of Lake Turkana, Kenya. Nature 322, 517±522 (1986). 40. Kimbel, W. H., White, T. D. & Johanson, D. C. in Evolutionary History of the ``Robust'' Australopithecines (ed. Grine, F. E.) 259±268 (Aldine de Gruyter, New York, 1988). 41. Asfaw, B. et al. Australopithecus garhi: A new species of early hominid from Ethiopia. Science 284, 629± 635 (1999). 42. Grine, F. E. & Strait, D. S. New hominid fossils from Member 1 ``Hanging Remnant'' Swartkrans Formation, South Africa. J. Hum. Evol. 26, 57±75 (1994). 43. Johanson, D. C., White, T. D. & Coppens, Y. Dental remains from the Hadar Formation, Ethiopia: 1974±1977 collections. Am. J. Phys. Anthrop. 57, 545±603 (1982). 44. Tobias, P. V. in Olduvai Gorge Vol. 4 (Cambridge Univ. Press, 1991). 45. McDougall, I. K-Ar and 40Ar/39Ar dating of the hominid-bearing Pliocene±Pleistocene sequence at Koobi Fora, Lake Turkana, northern Kenya. Geol. Soc. Am. Bull. 96, 159±175 (1985). 46. Brown, F. H., Shuey, R. T. & Croes, M. K. Magnetostratigraphy of the Shungura and Usno Formations, southwestern Ethiopia: new data and comprehensive reanalysis. Geophys. J. R. Astron. Soc. 54, 519± 538 (1978). 47. Haileab, B. Geochemistry, Geochronology and Tephrostratigraphy of Tephra from the Turkana Basin, Southern Ethiopia and Northern Kenya. Thesis, Univ. Utah (1995). 48. deMenocal, P. B. & Brown, F. H. in Hominin Evolution and Climatic Change in Europe Vol. 1 (eds AgustõÂ, J., Rook, L. & Andrews, P.) 23±54 (Cambridge Univ. Press, 1999). 49. Senut, B. et al. First hominid from the Miocene (Lukeino Formation, Kenya). C.R. Acad. Sci. Paris 332, 137±144 (2001).

Supplementary information is available on Nature's World-Wide Web site (http://www.nature.com) or as paper copy from the London editorial of®ce of Nature.

Acknowledgements We thank the Government of Kenya for permission to carry out this research and the National Museums of Kenya for logistical support. The National Geographic Society funded the ®eld work and some laboratory studies. Neutron irradiations were facilitated by the Australian Institute of Nuclear Science and Engineering and the Australian Nuclear Science and Technology Organisation. We also thank the Ethiopian Ministry of Information and Culture, the National Museum of Ethiopia, B. Asfaw, Y. Bayene, C. Howell, D. Johanson, W. Kimbel, G. Suwa and T. White for permission to make comparisons with the early Ethiopian hominins and numerous people including N. Adamali, B. Asfaw, C. Dean, C. Feibel, A. Grif®ths, W. Kimbel, R. Kruszynski, K. Kupczik, R. Leakey, D. Lieberman, J. Moore, K. Patel, D. Plummer, K. Reed, B. Sokhi, M. Tighe, T. White, and B. Wood for their help. Caltex (Kenya) provided fuel for the ®eld expeditions, and R. Leakey allowed us the use of his aeroplane. The ®eld expedition members included U. Bwana, S. Crispin, G. Ekalale, M. Eragae, J. Erus, J. Ferraro, J. Kaatho, N. Kaling, P. Kapoko, R. Lorinyok, J. Lorot, S. Hagemann, B. Malika, W. Mangao, S. Muge, P. Mulinge , D. Mutinda, K. Muthyoka, N. Mutiwa, W. Mutiwa, B. Onyango, E. Weston and J. Wynn. A. Ibui, F. Kyalo, F. Kirera, N. Malit, E. Mbua, M. Muungu, J. Ndunda, S. Ngui and A. Mwai provided curatorial assistance. The Leakey Foundation awarded a grant to F.B. Correspondence and requests for materials should be addressed to M.G.L. (e-mail: [email protected]).

© 2001 Macmillan Magazines Ltd

NATURE | VOL 410 | 22 MARCH 2001 | www.nature.com