Mandibular Remains Support Taxonomic Validity of Australopithecus sediba

See allHide authors and affiliations

Science  12 Apr 2013:
Vol. 340, Issue 6129, 1232997
DOI: 10.1126/science.1232997


Since the announcement of the species Australopithecus sediba, questions have been raised over whether the Malapa fossils represent a valid taxon or whether inadequate allowance was made for intraspecific variation, in particular with reference to the temporally and geographically proximate species Au. africanus. The morphology of mandibular remains of Au. sediba, including newly recovered material discussed here, shows that it is not merely a late-surviving morph of Au. africanus. Rather—as is seen elsewhere in the cranium, dentition, and postcranial skeleton—these mandibular remains share similarities with other australopiths but can be differentiated from the hypodigm of Au. africanus in both size and shape as well as in their ontogenetic growth trajectory.

Mandibular remains are well represented in the hominin fossil record and have been included among the holotypes of numerous early hominin taxa, including Australopithecus sediba from Malapa (1, 2). The suggestion (35) that insufficient consideration was given to intraspecific variation when Au. sediba was diagnosed as a discrete taxon can be further investigated by using recently recovered mandibular remains. Most notable among these are two conjoined fragments, UW88-128 and UW88-129, that can be refit to the previously published mandibular specimen UW88-54 (1) to form a near complete right hemi-mandible with complete dentition of an adult individual, Malapa Hominim 2 (MH2) (Fig. 1). These new portions allow us to examine the premolar and M1 regions of the mandibular corpus of an adult probable female of Au. sediba and provide a first look at the mandibular incisors and premolars of this taxon. In addition, we can now directly compare the mandibular canines of probable male and female individuals of Au. sediba.

Fig. 1 Right hemi-mandible of adult probable female individual MH2 from Malapa.

Conjoined fragments UW88-128 and UW88-129 refit to UW88-54 in (A) occlusal view, (B) medial view, and (C) lateral view. The coronoid process is truncated because of damage. Scale bar, 5 cm.

As is seen in the cranium and postcranial skeleton of Au. sediba (1, 614), the reassembled mandible of MH2 exhibits a mosaic of morphological characters that align it with specimens attributed to both Australopithecus and early Homo. For instance, the mandible of MH2 shares with Au. africanus such features as a moderately developed lateral prominence, moderately developed lateral superior and marginal tori, weakly delineated anterior and posterior marginal tubercles, a well-developed ectocondyloid buttress, and a well-developed lateral eminence of the ramus. However, it differs from Au. africanus and resembles specimens of early Homo by evincing relief between the alveolar prominence and the subalveolar fossa; possessing a small, steeply inclined postincisive planum; and revealing weakly developed endocoronoid and endocondyloid buttresses, which result in a triangular planum that is ill-defined along its inferior extent (15). However, gross morphology in hominin mandibles tends to be variable, as is seen in a direct comparison of MH1 and MH2; thus, we turn to size and shape analyses of mandibular corpora to assess the taxonomic validity of Au. sediba.

Linear dimensions of mandibular corpora have been shown to be taxonomically informative within fossil representatives of the genus Homo (2). In absolute measures, the mandibular corpus of MH2 is slightly taller and narrower than that of MH1 (Fig. 2 and table S1). There is considerable overlap in mandibular corpus height between Au. afarensis, Au. africanus, Au. sediba, H. habilis, H. erectus, and even H. sapiens. The corpus of MH1 is narrower than Au. africanus at the level of the P4; it overlaps with the lower end of the range of Au. africanus at the level of the M1, although only a single specimen of Au. africanus, the otherwise large-toothed Stw 498, is narrower. MH1 overlaps with H. habilis and H. erectus in width at the level of the P4 and M1, although again plotting near the lower end of the range of both taxa. The newly reassembled mandible of MH2 is narrower still than MH1 at the level of the P4 and M1, plotting below the range for Au. africanus, H. habilis, and H. erectus; it appears among the narrowest in the African hominin fossil record. Calculating corpus cross-sectional area by using the formula for an ellipse, MH1 and MH2 have similar values at the level of both the P4 and M1. The cross-sectional areas of the mandibular corpora of Au. sediba are small at the level of the P4 and M1, plotting at or below the lowest end of the range of values for Au. africanus and within the ranges of H. habilis and H. erectus. In linear dimensions, mandibular corpus width—and to a lesser extent cross-sectional area—thus appear useful for discriminating Au. sediba from Au. africanus and for aligning the former taxon with specimens attributed to early Homo. This is particularly the case with the adult specimen MH2.

Fig. 2 Mandibular corpus measures of Au. sediba compared with other early hominin taxa.

(A) Corpus height at P4. (B) Corpus width at P4. (C) Corpus area at P4 computed by using the formula for an ellipse [π(height/2) * (width/2)]. (D) Corpus height at M1. (E) Corpus width at M1. (F) Corpus area at M1 computed by using the formula for an ellipse [π(height/2) * (width/2)]. Measures for Au. afarensis, Au. boisei, H. habilis, H. rudolfensis, and H. erectus specimens are from (2325); measures of Au. africanus, Au. sediba, Au. robustus, and H. sapiens specimens were taken by D.J.d.R. on original materials.

To better understand the nature of differentiation between early hominin taxa in both size and shape of mandibular corpora, we replicated the approach of Lague and colleagues (2) using specimens attributed to species of Australopithecus and early Homo in Africa and Georgia (15). Seven of the eight measures originally used (2) could be replicated in the reassembled mandible of MH2 and form the basis of our analyses. Randomization of “distinctness values” (RDV) (2, 16) reveals that within-group variance is less than between-group variance; thus, it appears that the mandibles of Australopithecus and Homo that we examined do indeed reflect a taxonomic signal. We therefore investigated the size and shape of the mandibular corpus of Au. sediba in relation to other hominins. We performed a multivariate analysis of variance (MANOVA) (15), the overall results of which can be visualized as a plot of between-group differences, scaled to within group variances (Fig. 3A). The major result illustrated in this plot is that H. sapiens are clearly distinct from early hominins, especially the robust australopiths, whereas Au. sediba clusters with the nonrobust australopiths and early Homo. This major axis indicates that human mandibles are small overall, have relatively narrow corpora, and relatively mesiodistally (MD) elongated (though small overall) canine alveoli (table S2).

Fig. 3 Visualization of the first two canonical axes from the MANOVA on form.

(A) Including humans. (B) Excluding humans. Details of canonical plot interpretation are available in the text, and computation is available in supplementary text S1. Measures for Au. afarensis, Au. boisei, H. habilis, H. rudolfensis, and H. erectus specimens are from (2325); measures of Au. africanus, Au. sediba, Au. robustus and H. sapiens specimens were taken by D.J.d.R. on original materials. Plot aspect ratio indicates the relative magnitude (eigenvalues) of the respective canonical axes.

Among the early hominins, Au. sediba most closely matches this pattern. A planned contrast within the MANOVA results explicitly comparing Au. sediba and Au. africanus revealed a significant difference between them (F8,23 = 2.55, P = 0.037). The difference between these two taxa primarily involved the smaller overall size of the mandible, the relatively deeper mandibular corpus, and the relatively MD elongated premolar row in Au. sediba when compared with that of Au. africanus (table S3). In other words, the mandibles of Au. sediba are not only scaled differently than those of Au. africanus, they are also shaped differently. Removal of humans from the ordination resulted in a primary axis that principally defined differences between the robust australopiths and the remaining early hominins (Fig. 3B). However, the secondary axis—approximately one-third the magnitude of the first—separated Au. sediba clearly from all other early hominins (planned contrast, P = 0.034). This effect was due to Au. sediba having a relatively MD elongated canine alveolus (table S3). Excluding size as a variable results in ordinations similar to those in Fig. 3, demonstrating that size is not unduly influencing the analysis (fig. S1).

The mandibular molars of Au. sediba were recorded as being small, similar to specimens attributed to early Homo (1), and we now have data on the remaining mandibular dentition from MH2 (figs. S2 and S3). Although the incisors of MH2 are too worn to reliably reconstruct their mesiodistal (MD) MD lengths, the buccolingual (BL) breadth of the I1 plots at the lowest end of the range of Au. africanus and H. erectus but within the range of Au. robustus and H. sapiens. The BL breadth of the I2 plots at the lowest end of the range of Au. africanus, but within the range of Au. robustus, Au. boisei, H. erectus, and H. sapiens. The only specimen of Au. africanus that overlaps with MH2 in the BL breadth of both incisors is Stw 151, a specimen described as more derived toward Homo than are the remainder of the Sterkfontein Member 4 sample (17). As in MH1 (1), the MD length of the canine of MH2 is small, plotting below the range for Au. africanus and H. habilis, and within the range of H. erectus and H. sapiens. In BL breadth, the canine of MH1 plots near the lowest end of the range for Au. africanus and close to the mean for H. habilis and H. erectus, whereas the canine of MH2 falls below the range of these early Homo taxa as well as Au. africanus. Indeed, the canine of MH2, and to a lesser extent of MH1, is among the smallest in the hominin fossil record. Likewise, the premolars of MH2 are small and generally plot at or below the lower end of the ranges seen in Au. africanus and H. habilis, appearing especially MD shortened; where the P4 of Au. sediba overlaps with Au. africanus, only a single specimen attributed to Au. africanus, Stw 112, appears smaller than MH2. Conversely, the premolars of Au. sediba plot within the ranges of H. erectus and H. sapiens dentitions.

The shape of the premolars is also diagnostic; occlusal outline analysis of both the P3 and the P4 by using elliptical function Fourier analysis (EFFA) positions them as outliers relative to both Au. africanus and Au. robustus, demonstrating that these teeth differ in both size and shape from other South African australopiths (fig. S4) (15). Although the complete dentition of MH2 appears small in size similar to that of Homo, it shares with Australopithecus a pattern of increasing tooth size along the molar row (fig. S5).

Using mandibular molars, it was originally estimated that the dentition of MH1 was between 8.1 and 10.7% larger than that of MH2 (1). We can now extend this to include canines: The canine of MH1 is ~9.6% larger than the canine of MH2 in MD length and ~20.0% larger in BL breadth. Converting this to a square root of the basal crown area, the canine of MH1 is ~15% larger than that of MH2. This can be compared with canine dimorphism levels derived for other hominins by using the “mean” method of estimation (15, 18). With a value of 1.15, Au. sediba is intermediate between taxa such as Au. afarensis (1.21) and Au. boisei (1.21) on the one hand and taxa such as Au. africanus (1.11), Au. robustus (1.11), H. erectus (1.13), and H. sapiens (1.07) on the other. Although the “mean” method tends to overestimate dimorphism in fossil species for which sex is generally unknown (18), with probable male and female individuals Au. sediba provides a known sample. Sample sizes are small for the hominin taxa, although these data nonetheless serve to highlight the dimorphism evident in the otherwise small canines of Au. sediba. Although the canines of Au. sediba are small and lack the diagnostic lingual relief seen in Au. africanus, a pattern shared with early Homo, they reveal a degree of canine dimorphism that approaches levels seen in several other australopith taxa and that is also close to that of H. erectus.

Last, we tested whether Au. sediba shares a common pattern of development with Au. africanus using Euclidean distance matrix analysis (EDMA) (15). We examine the ontogeny of mandibular shape in the fossil species Au. sediba, Au. africanus, and H. erectus and in the extant species H. sapiens and Pan troglodytes. Juvenile samples are roughly the same developmental age as MH1 (pre-M3 eruption), and therefore by necessity this comparison captures only one stage of development. EDMA was performed on mean forms of extant samples and on individual fossil juvenile and adult specimens as representatives of their respective species (table S4) (1922). In addition, simulated hypothetical juvenile Au. sediba forms were created by using species-specific growth patterns derived from Au. africanus and H. erectus in order to explore the effect of these respective growth patterns when applied to Au. sediba (15). The overall EDMA results can be visualized in a principal coordinates plot (PCOORD) (Fig. 4) and show that Au. sediba displays a pattern of growth that differs from Au. africanus, P. troglodytes, and H. sapiens [further discussion is available in (15)]. Rather, the growth pattern of Au. sediba is most similar to H. erectus. Au. sediba also differs in magnitude of growth from Au. africanus and H. erectus as well as both of the extant taxa (Fig. 4). Because these data are scaled to facilitate cross-species comparisons, the magnitude of change should be interpreted as the magnitude of shape change during growth rather than absolute size. Additionally, the simulated hypothetical juvenile forms and their associated growth trajectories further support a distinction between mandibular growth in Au. sediba versus both fossil hominin species [further discussion is available in (15)], with neither a H. erectus–like nor an Au. africanus–like pattern or magnitude of growth resulting in simulated final juvenile forms similar to the actual MH1 juvenile, which further suggests that ontogenetic change in the mandible of Au. sediba is unique.

Fig. 4 Ontogenetic trajectories of mandibular shape variability for Au. sediba (MH1,MH2), Au. africanus (A,a), H. erectus (E,e), H. sapiens (H,h), P. troglodytes (C,c), and two simulated trajectories for Au. sediba (S-A,S-E).

Upper- and lowercase labels indicate adults and juveniles, respectively, and are connected by a line that indicates a “growth trajectory” in shape space. Black dotted lines denote simulated growth trajectories connecting MH2 with “hypothetical” juvenile forms. Details of interpretation are available in the text. Details of the samples and analysis can be found in (15). Linear measurements for all specimens are derived from laser surface scans taken by L.S. and K.B.C. on original material.

The small size of the newly recovered mandibular corpus of Au. sediba—combined with its reduced dentition, in particular of the canines and premolars—separate it from its otherwise close morphological affinity, Au. africanus. The distinction between Au. sediba and Au. africanus extends to more than just a size difference; the two taxa differ in shape as well, with Au. sediba possessing a relatively smaller and deeper corpus, and a relatively elongated premolar row when compared with those of Au. africanus. Likewise, the premolars of Au. sediba differ in occlusal outline shape from those of Au. africanus and Au. robustus, plotting as distinct outliers relative to the South African australopiths. The divergent growth trajectory of Au. sediba compared with those of both extant and fossil taxa further highlights the singular nature of the mandible of Au. sediba. We interpret these substantial size and shape differences as upholding the taxonomic distinction initially proposed for Au. sediba (1), in particular as compared with specimens of Au. africanus. And where the Au. sediba mandibles differ from those of Au. africanus, they appear most similar to representatives of early Homo in both size and shape.

Supplementary Materials

Methods and Results

Supplementary Text

Figs. S1 to S5

Tables S1 to S5

References (2633)

References and Notes

  1. Methods and results are available as supplementary materials on Science Online.
  2. Acknowledgments: We thank the South African Heritage Resource agency for the permits to work on the Malapa site and the Nash family for granting access to the Malapa site and continued support of research on their reserve. The South African Department of Science and Technology, and the African Origins Platform (AOP), the South African National Research Foundation, the Institute for Human Evolution (IHE), the Palaeontological Scientific Trust (PAST), the Andrew W. Mellon Foundation, the L.S.B. Leakey Foundation Baldwin Fellowship, the United States Diplomatic Mission to South Africa, the French Embassy in South Africa, the National Geographic Society, the A. H. Schultz Foundation, the Oppenheimer and Ackerman families, Sir Richard Branson, and the Program to Enhance Scholarly and Creative Activities and the International Research Travel Award Grant of Texas A&M University all provided funding for this research. D.J.d.R. holds a Ray A. Rothrock ’77 Fellowship at Texas A&M University, which provided additional funding. The University of the Witwatersrand’s Schools of Geosciences and Anatomical Sciences and the Bernard Price Institute for Palaeontology provided support and facilities. The authors declare that they have no real or apparent conflicts of interest. We thank E. Mbua, P. Kiura, V. Iminjili, and the National Museums of Kenya; P. Msemwa and the National Museum and House of Culture of Tanzania; S. Potze of the Ditsong Museum; B. Billings of the School of Anatomical Sciences of the University of the Witwatersrand; W. Seconna of the Iziko South African Museum; and L. M. Jellema of the Cleveland Museum of Natural History for access to comparative fossil and extant primate and human materials in their care. Numerous individuals have been involved in the ongoing preparation and excavation of these fossils, including C. Dube, B. Eloff, C. Kemp, M. Kgasi, M. Languza, J. Malaza, G. Mokoma, P. Mukanela, T. Nemvhundi, M. Ngcamphalala, S. Jirah, S. Tshabalala, and C. Yates. Other individuals who have given substantial support to this project include B. de Klerk, C. Steininger, B. Kuhn, L. Pollarolo, B. Zipfel, J. Kretzen, D. Conforti, J. McCaffery, C. Dlamini, H. Visser, R. McCrae-Samuel, B. Nkosi, B. Louw, L. Backwell, F. Thackeray, M. Peltier, J. Cheverud, and D. Roach. The Au. sediba specimens are archived at the Evolutionary Studies Institute at the University of the Witwatersrand. All data used in this study are available upon request, including access to the original specimens, by bona fide scientists.
View Abstract

Navigate This Article