Late Miocene Teeth from Middle Awash, Ethiopia, and Early Hominid Dental Evolution

See allHide authors and affiliations

Science  05 Mar 2004:
Vol. 303, Issue 5663, pp. 1503-1505
DOI: 10.1126/science.1092978


Late Miocene fossil hominid teeth recovered from Ethiopia's Middle Awash are assigned to Ardipithecus kadabba. Their primitive morphology and wear pattern demonstrate that A. kadabba is distinct from Ardipithecus ramidus. These fossils suggest that the last common ancestor of apes and humans had a functionally honing canine–third premolar complex. Comparison with teeth of Sahelanthropus and Orrorin, the two other named late Miocene hominid genera, implies that these putative taxa are very similar to A. kadabba. It is therefore premature to posit extensive late Miocene hominid diversity on the basis of currently available samples.

The phylogenetic status of the earliest hominid genera Sahelanthropus, Orrorin, and Ardipithecus (16) and the definition of the family Hominidae (710) are in debate. By what derived characters should the hominid (1, 11) clade be recognized? Bipedality might be an arbiter of hominid status, but “bipedality” involves a large and complex set of anatomical traits and is not a dichotomous character. Femora attributed to Orrorin tugenensis at ∼5.8 million years ago (Ma) constitute the earliest postcranial evidence for early hominid bipedality (2, 12). However, the O. tugenensis femora are different from those of later hominids such as Australopithecus afarensis (13). Indeed, some question whether Au. afarensis was habitually bipedal (1417), despite evidence indicating that some form of bipedality appears to have arisen before 5 Ma (1, 2).

Early hominids have also been identified through dental characters. For example, a functional canine–third premolar (C/P3) honing complex is present in fossil and modern apes but absent in all hominids. Among late Miocene hominids, derived features of the lower canines of Sahelanthropus tchadensis and A. ramidus kadabba have contributed to the recognition of these two taxa as early hominids. However, only one partial upper canine of S. tchadensis and a complete crown of O. tugenensis are known. Here, we describe teeth from the late Miocene of the Middle Awash, Ethiopia.

In 2001, hominid fossils from the Middle Awash that were dated 5.2 to 5.8 Ma (18) were assigned to A. r. kadabba (1). Continued work in November 2002 produced additional remains. An upper canine and a lower third premolar are among six new teeth (19) recovered from Asa Koma Locality 3 (ASK-VP-3), with an estimated age of 5.6 to 5.8 Ma (18). The new fossils illuminate hominid dental evolution near the divergence of apes and humans, and they require revision of the taxonomic status of A. r. kadabba.

The following is a description of A. kadabba sp. nov., elevated here from the subspecies A. r. kadabba (1): Order Primates Linnaeus, 1758; Suborder Anthropoidea Mivart, 1864; Genus Ardipithecus White, Suwa, and Asfaw, 1995; Species Ardipithecus kadabba sp. nov.

Holotype, localities, and horizons. Same as described for A. r. kadabba (1).

Referred material. The holotype and paratype of A. r. kadabba listed in (1), and ASK-VP-3/400 (right upper canine); ASK-VP-3/401 (right upper first molar); ASK-VP-3/402 (left upper first molar); ASK-VP-3/403 (left lower P3); ASK-VP-3/404 (lingual half of left upper P4); ASK-VP-3/405 (right upper P4 fragment).

Diagnosis. A. kadabba differs from fossil and extant apes by features enumerated in (1) and by the presence of a clearly defined anterior fovea of the lower P3, demarcated by a foldlike buccal segment of the mesial marginal ridge. It differs from extant apes and O. tugenensis by its more circular upper canine crown outline in occlusal view. In addition to the features distinguishing A. r. ramidus from A. r. kadabba (1), A. kadabba also differs from A. ramidus by the morphology of the upper canine (more basal termination of the mesial and distal apical crests) and morphology of the lower P3 (more asymmetrical crown outline and relatively smaller anterior fovea).

Specimen ASK-VP-3/400 is a right upper canine lacking root and some crown base (Fig. 1; the preserved buccolingual and crown height dimensions of the canine are >10.1 mm and >15.5 mm, respectively). The mesial crest runs from mesial crown shoulder to crown tip, losing sharpness toward the apex. This mesial crest bears a long vertical facet from occlusal contact with the lower canine, extending from the apex and approximating the mesial crown shoulder. The crown tip shows no apical wear except minimal surface polishing. A similar configuration is often seen in chimpanzees at an early wear stage. In later hominids such as Au. afarensis, this mesial occlusal facet is not as steep and may be associated with relatively extensive apical wear. The mesial lingual groove is deep toward the base and shallow toward the apex. The rest of the lingual crown face is relatively flat and lacks the apparently stronger hollowing seen in the lone upper canine of O. tugenensis (BAR 1425′00) and in chimpanzee homologs (where a lingual cingulum usually contributes to this effect). The mesial shoulder of ASK-VP-3/400 is relatively more basally placed than in A. ramidus. This leads to a primitive, tall, pointed crown outline in lateral view.

Fig. 1.

(A) Lingual, labial, mesial, and occlusal views of ASK-VP-3/400 (top two rows), and buccal, distal, lingual, and occlusal views of ASK-VP-3/403 (bottom two rows). Scale bars, 1.0 mm (each segment). (B) Lateral views of a female common chimpanzee and A. kadabba upper and lower canines and premolars (upper canine ASK-VP-3/400, lower canine STD-VP-2/61, upper premolar ASK-VP-3/160 reversed, lower premolar ASK-VP-3/403 reversed). (C) Bivariate plot of upper canine mesiodistal length and mesial crest length (in millimeters; see text for discussion). The Pan troglodytes sample includes five males and 11 females; the P. paniscus sample includes six males and 10 females. The Au. afarensis/anamensis sample consists of L.H.-3, L.H.-6, A.L.333x-3, and KNM-KP 35839; the A. ramidus sample consists of ARA-VP-6/1 and ARA-VP-1/300. Mesiodistal length was taken either across the mesial and distal shoulders, or, in modern apes, at the crown bases adjacent to the shoulders when the latter yielded slightly larger dimensions. (D) Trajectory of evolutionary change in the early hominid canine, lingual view. Bottom canine is A. kadabba (ASK-VP-3/400 reversed) from Asa Koma; middle canine is A. ramidus (ARA-VP-6/1) from Aramis; top two canines are Au. afarensis from Hadar (A.L.400-1b reversed) and Laetoli (L.H.-3). Arrows indicate temporal placement of the fossils. [Illustrations by Luba Gudz]

Relative to basal diameters, the Asa Koma upper canine crown is taller than in O. tugenensis, but the latter's canine has an even more basal mesial shoulder than that of ASK-VP-3/400. Au. afarensis, Au. anamensis, and A. ramidus upper canines have relatively lower crowns than those of female Pan troglodytes. The new A. kadabba upper canine lies at the margin of the female P. troglodytes distribution (Fig. 1). The O. tugenensis upper canine is relatively more elongate mesiodistally, as are the homologs of female chimpanzees.

ASK-VP-3/403 is a left lower P3 lacking the mesiobuccal crown base and distolingual enamel. The roots are missing except for a portion immediately below the lingual cusp. The crown is asymmetric in occlusal outline, more so than in any known hominid P3. A small facet occupies the mesiobuccal crown face close to the protoconid apex, caused by contact with the lingual surface of an interlocking upper canine. The transverse crest descends from the tip of the protoconid to the metaconid, which is hardly expressed as a distinct entity. It forms a high divide between the larger posterior fovea and the smaller, more restricted anterior fovea. In its occlusal crown outline, the premolar is narrow lingually and broad buccally. The distobuccal face has a vertical enamel ridge delineated by a small groove anterior to it. The anterior fovea is constricted and positioned at the mesiobuccal angle of the occlusal surface, apparently defined only by the buccal segment of the mesial marginal ridge. The mesial protoconid crest is longer than the distal. The protoconid and metaconid are basally defined by a strong constriction of the mesial crown contour. Comparison with A. ramidus shows that ASK-VP-3/403 is more asymmetric and therefore more similar to Miocene and extant apes. This corresponds with the Asa Koma upper canine and the previously described morphology of upper third premolars and lower canines of A. kadabba (1).

Functional honing is the phenomenon by which all extant and Miocene apes sharpen the distal edge of the upper canine against the mesiobucally extensive buccal crown surface of the lower P3. The distolingual surface of the upper canine crown is often extensively scalloped by honing wear. Even in the case of female ape canines that are not well sharpened, a distal wear strip extends and faces distolingually because of its occlusal relationship with the lower P3. Of all the African and Eurasian Miocene apes, Ouranopithecus has been recognized as having a C/P3 complex that approximates the hominid condition (20). We interpret Ouranopithecus as nonhominid because its upper canine is sharpened by wear and because its masticatory specializations are divergent from African hominids predating Australopithecus. In contrast, the Pliocene hominids Au. afarensis, Au. anamensis, and A. ramidus usually show vertically elongate wear facets on the distal crest of the lower canine and on both mesial and distal crests of the upper canine. The latter facet is positioned on the distal crest itself and does not carve into the upper canine lingual crown face. In some cases, occlusal overlap does create a slight lingual bevel on the upper canine and a narrow facet along the mesial crest of the occluding P3 (Au. afarensis specimens such as A.L.128-23 and L.H.-5, for example). No known A. ramidus (5, 6), Au. afarensis (21, 22), or Au. anamensis (23, 24) lower P3 exhibits any sign of a mesiobuccally oriented facet on its buccal crown face. The primitive morphology of the Asa Koma C/P3 complex closely approaches the extant and fossil ape conditions, and thereby helps identify A. kadabba as a species distinct from A. ramidus. The single Orrorin upper canine is apparently not worn enough to assess a possible honing condition (2). The few known C/P3 elements of S. tchadensis are described as showing no sign of functional honing (3), despite its apparent antiquity (25).

Late Miocene hominid fossils from Chad, Ethiopia, and Kenya have been ascribed to three different genera and interpreted as evidence of phylogenetic diversity (26) despite close anatomical resemblances in comparable parts. The small constellation of derived characters of the A. kadabba and S. tchadensis dentitions (1, 3) and the O. tugenensis femora (12) is shared exclusively with later hominids. Collectively, these fossils provide important outgroup comparisons to geologically younger A. ramidus and Au. anamensis, confirming these later taxa as cladistically hominid on the basis of their increasingly derived C/P3 complex. The distinctly primitive characters in the A. kadabba dentition—such as the projecting, interlocking upper and lower canines, and the asymmetric lower P3 with buccal wear facet—imply that its last common ancestor with chimpanzees and bonobos retained a functioning C/P3 complex. But wear on the upper and lower canines of Sahelanthropus (3) and the lower canine of A. kadabba from Alayla (1) suggest a lack of consistently expressed functional honing in these earliest hominids. Now that dental remains of the earliest hominids and their descendants are available, a trajectory of canine morphological change from the late Miocene hominids to Pliocene A. ramidus, Au. anamensis, and Au. afarensis can be envisioned (Fig. 1).

Metric and morphological variation within available small samples of late Miocene teeth attributed to A. kadabba, O. tugenensis, and S. tchadensis is no greater in degree than that seen within extant ape genera. Despite claims of molar enamel thickness differences among these late Miocene fossils (2), we question the interpretation that these taxa represent three separate genera or even lineages. Given the limited data currently available, it is possible that all of these remains represent specific or subspecific variation within a single genus.

References and Notes

View Abstract

Navigate This Article