Australopithecus sediba at 1.977 Ma and Implications for the Origins of the Genus Homo

See allHide authors and affiliations

Science  09 Sep 2011:
Vol. 333, Issue 6048, pp. 1421-1423
DOI: 10.1126/science.1203697


Newly exposed cave sediments at the Malapa site include a flowstone layer capping the sedimentary unit containing the Australopithecus sediba fossils. Uranium-lead dating of the flowstone, combined with paleomagnetic and stratigraphic analysis of the flowstone and underlying sediments, provides a tightly constrained date of 1.977 ± 0.002 million years ago (Ma) for these fossils. This refined dating suggests that Au. sediba from Malapa predates the earliest uncontested evidence for Homo in Africa.

The dolomite caves near Johannesburg, South Africa, have received material from the surface for at least the past ~3 million years (1) and contain rich fossil assemblages, including early hominin fossils. At the site of Malapa, remarkably well-preserved remains of numerous individual hominins are attributed to a newly discovered species, Australopithecus sediba, interpreted as being potentially ancestral to Homo (2). The age of the fossils was estimated as 1.78 to 1.95 Ma on the basis of faunal correlation, U-Pb dating, and paleomagnetic data (3). This estimated age interval postdated published dates for the oldest potential representatives of Homo in the African fossil record, leading some to suggest that Au. sediba was too young to be considered an ancestor of Homo (4, 5).

The earliest species that is commonly accepted as a member of the genus Homo is H. erectus sensu lato, as the generic status of H. habilis and H. rudolfensis has been disputed (6). The oldest relatively complete cranium widely accepted as belonging to this species, KNM-ER 3733 from the Koobi Fora Formation of Kenya, dates to ~1.78 million years ago (Ma), although a series of more fragmented cranial remains and a single os coxa from Koobi Fora that are referred to H. erectus date to ~1.88 to 1.90 Ma (7). Other probable early H. erectus fossils are known from Swartkrans in South Africa at ~1.80 to 1.90 Ma (8) and Dmanisi in Georgia at ~1.78 to 1.85 Ma (9). Thus, ~1.90 Ma represents a reliable first appearance datum for this species (10, 11). Either H. habilis or H. rudolfensis is widely regarded as being the ancestor of H. erectus, suggesting that fossils attributable to one or both of these species could predate ~1.90 Ma. Indeed, a number of fragmented and/or isolated cranial and dental remains reported to be older than 1.90 Ma have been assigned to Homo. However, most of these purported early Homo fossils are of equivocal taxonomic assignment, or their age is uncertain (12) [supporting online material (SOM) text S1], and thus do not conclusively demonstrate the existence of Homo before 1.90 Ma.

A single maxillary specimen from Ethiopia, A.L. 666-1, is considered by some to be the strongest evidence for early Homo before 1.90 Ma (12, 13), although its isolation from other material renders taxonomic diagnosis difficult (SOM text S1). Although A.L. 666-1 shares features with Homo, it is unclear whether these features are unequivocally diagnostic of the genus, in particular given that this ~2.33-million-year-old fossil represents the only compelling potential representative of Homo before ~1.90 Ma (12) (SOM text S1). The oldest fossils that can be more confidently assigned to H. habilis and/or H. rudolfensis, in this case relatively intact or complete cranial remains, date to approximately the same age as H. erectus at ~1.88 to 1.90 Ma at Koobi Fora. Thus, excepting A.L. 666-1, the appearance of these three species of early Homo is otherwise effectively contemporaneous.

Substantial challenges in dating hominin fossils in South Africa are the lack of volcanic strata and the complex stratigraphy of the cave sediments. However, uranium-lead (U-Pb) dating of calcium carbonate cave rocks (speleothems or flowstones) (8) and paleomagnetic analysis of the speleothems and sediments surrounding the fossils (14) are increasingly providing accurate and precise ages for the South African sites. The last major geomagnetic field reversal was 0.78 Ma, but a number of short events and excursions have also occurred, some lasting just 3000 to 20,000 years (15). Recent studies of the South African paleocaves indicate that such short geomagnetic field excursions can be accurately recorded in these deposits (3, 14). Here we report on newly exposed stratigraphy at the Malapa site, including flowstone units (Figs. 1 and 2 and fig. S4) that, combined with new U-Pb and paleomagnetic data, allow us to narrow the deposition of the Au. sediba–bearing deposits to one of these short geomagnetic field events, the Pre-Olduvai event at ~1.977 Ma.

Fig. 1

Geological map of Malapa pits 1 and 2, showing the distribution of the principal rock facies as on 18 November 2010, including freshly exposed flowstone 2, and the sample locations for U-Pb and palaeomagnetic analysis.

Fig. 2

A NE-SW cross-section sketch map through the Malapa site showing the distribution of sedimentary facies, position of hominin fossils, and U-Pb and palaeomagnetic sample locations, together with U-Pb ages and normal and reversed polarities as reported in this paper and in (3). The sample locations of the dating and paleomagnetic samples have been projected onto the section within their correct stratigraphic position.

At Malapa, the Au. sediba fossils (2) are encased in poorly sorted, massive peloidal sandstone (classified as facies D) underlain by a thick flowstone layer (flowstone 1) (3). Flowstone 1 is dated to 2.026 ± 0.021 Ma and records a normal magnetic polarity near its base, correlated to the Huckleberry Ridge event, followed by the reversed polarity of the Matuyama chron (3) (Fig. 1). Normal and intermediate polarity from overlying flowstone drapes and clastic deposits containing Au. sediba fossils were originally interpreted as representing the Olduvai event between 1.78 and 1.95 Ma (3).

Recent excavation of the site away from the original pit exposure (pit 1) revealed fossil-bearing deposits in a second, shallow pit (pit 2), as well as dolomite roof blocks over 1 m in diameter and more flowstone layers (Figs. 1 and 2). Overlying the hominin-bearing layer of facies D is facies E, which is traceable across the E-wall of pit 1. In the southeast corner of pit 1, facies E directly overlies flowstone 1. Here the layering in facies E is disturbed and partly truncated by a small (100 × 40 cm) dolomite block (block 1), which preserves an internal layering that dips at 42° toward 57° (Fig. 1). A second, larger (8 × 4 m) block (block 2) consisting of dolomite and cave sediment rests on top of the smaller block and occupies the center of the currently exposed deposits between pit 1 and pit 2 (Fig. 1). Cave sediments in block 2 cover the top surface of the block and consist of peloidal grainstone with abundant calcite fenestrae [facies C (3)] that were deposited in an upper cave chamber before it collapsed into the chamber that trapped the Au. sediba fossils. Resting on top of block 2 is a third block (block 3) consisting of consolidated cave sediment (facies C), with internal layering that dips in the opposite direction from block 2 (Fig. 2). Block 3 is covered by a layer of flowstone (flowstone 2), up to 80 cm wide and around 7 cm thick, that forms a vertical curtain along the broken north side of the block, indicating that the flowstone was deposited in situ after the block had fallen, making flowstone 2 younger than the hominin-bearing layers of facies D and E in pit 1. The flowstone is covered by a massive peloidal grainstone, facies F, which shows horizontal layering and graded bedding. Facies F drapes blocks 2 and 3 in pit 2 and probably represents the same stratigraphic horizon as the layer of facies F at the top of the sequence in pit 1 (Figs. 1 and 2). The stratigraphy in pits 1 and 2 indicates that the fossil-bearing units of facies D and E in pit 1 were deposited between the formation of flowstones 1 and 2. Therefore, a U-Pb date for flowstone 2 will provide a direct upper age limit for the Au. sediba fossils.

Two exposures of flowstone 2 (subsamples M6 and M7; Fig. 1, SOM text S4, fig. S4, and table S3) were dated using the U-Pb dating method (16, 17). Dates were obtained using multicollector inductively coupled plasma mass spectrometry MC-ICPMS techniques from thin, U-rich (on average 0.4 parts per million) layers within the flowstone unit identified through laser-ablation ICPMS traverses. The resulting age is 2.048 ± 0.140 Ma. This age provides a maximum age range of 2.19 to 1.91 Ma for the capping flowstone. Flowstone 1 at the base of the sequence, dated to 2.026 ± 0.021 Ma (2.05 to 2.01) (3), is within error of the dates for flowstone 2.

The U-Pb dates pin the palaeomagnetic sequence for the stratigraphic section flanked by flowstones 1 and 2 to between 2.05 and 1.91 Ma. Flowstone 2 records a reversed polarity that, combined with the U-Pb ages (>1.91 Ma), indicates that this layer must have formed before the beginning of the Olduvai normal-polarity subchron at ~1.95 Ma. Additional subsamples were taken from the fossil-bearing units between flowstones 1 and 2 that confirm the normal and intermediate polarities for facies D (3) (SOM text S3 and table S2). Considering that flowstone 2 formed before 1.95 Ma, normal polarity for the fossil-bearing sediments can only be explained if they were deposited during the short-lived Pre-Olduvai excursion at ~1.977 Ma (15, 18), the only geomagnetic field reversal between 1.95 Ma and the Huckleberry Ridge event at 2.05 to 2.03 Ma (SOM text S3) The duration of the Pre-Olduvai event has been estimated at 3000 years (18), thus providing an age of 1.977 ± 0.002 Ma for the Au. sediba fossils.

Thus, the fossils of Au. sediba are older than the oldest uncontested representatives of the genus Homo. As such, Au. sediba cannot be precluded a priori as a potential candidate ancestor of Homo based on the age of the fossils from Malapa. Taken as a whole, the species Au. sediba, H. habilis, and H. rudolfensis all present morphological features that imply adaptive evolution toward the Homo grade as expressed in H. erectus. These involve such changes as increased brain size and organization, dentognathic reduction, derived thermoregulatory capabilities (a projecting nose), increased body size, biomechanical reorganization of the pelvis for locomotion, relative lower limb elongation, enhanced bipedal characteristics of the foot (a longitudinal arch), and the potential for tool use and manufacture (2, 1922). None of these three species exhibits a comprehensive constellation of these features; thus when all three are viewed as a whole, they appear to form an adaptive mosaic. This conceivably implies an adaptive radiation of hominins at ~2.0 Ma. Under such a scenario, one would expect substantial homoplasy to be evident; consequently, we will face great difficulty in sorting out the phylogenetic relationships of several closely related species. In particular, discerning the specific ancestor of H. erectus, and placing this ancestor in the correct genus (i.e., Australopithecus versus Homo), will be a most challenging endeavor, especially when only fragmentary fossil remains are preserved.

Supporting Online Material

SOM Text S1 to S4

Figs. S1 to S8

Tables S1 to S3

References (2373)

References and Notes

  1. White (10) agreed that the first appearance date of H. erectus could be reliably established, although he accepted 1.75 Ma as the probable date.
  2. We have not seen the excavation report cited in (36), thus it could potentially confirm the provenience of the specimen.
  3. Acknowledgments: We thank the South African Heritage Resources and the Nash family. Funding was received from the South African Department of Science and Technology, South African National Research Foundation, Institute for Human Evolution, University of the Witwatersrand, University of the Witwatersrand’s Vice Chancellor’s Discretionary Fund, National Geographic Society, Palaeontological Scientific Trust, Andrew W. Mellon Foundation, Ford Foundation, U.S. Diplomatic Mission to South Africa, French Embassy of South Africa, Oppenheimer and Ackerman families, and Sir Richard Branson. Thanks also to the University of the Witwatersrand’s Schools of Geosciences and Anatomical Sciences, Bernard Price Institute for Palaeontology; Gauteng Government, Gauteng Department of Agriculture, Conservation and Environment, Cradle of Humankind Management Authority; E. Mbua, P. Kiura, and V. Iminjili at the National Museums of Kenya for access to comparative specimens; and the University of Zurich 2010 Field School. For the fossil preparation, we thank C. Dube, C. Kemp, M. Kgasi, M. Languza, J. Malaza, G. Mokoma, P. Mukanela, T. Nemvhundi, M. Ngcamphalala, S. Jirah, S. Tshabalala, and C. Yates. Others who have given significant support are B. de Klerk, W. Lawrence, C. Steininger, B. Kuhn, L. Pollarolo, J. Kretzen, D. Conforti, C. Dlamini, H. Visser, B. Nkosi, B. Louw, L. Backwell, F. Thackeray, M. Peltier, and M. Klinkmüller. Further acknowledgements are given to Texas A&M University (The Ray A. Rothrock ’77 Fellowship, the Program to Enhance Scholarly and Creative Activities, and the International Research Travel Assistance Grant of Texas A&M University), Swiss National Research Foundation (PBBEP2-126195), Australian Research Council (DP0877603), University of Melbourne (McKenzie Fellowship), Liverpool University Geomagnetism Laboratory staff, AfricaArray, James Cook University, and the BHPBilliton Geosciences staff development program.
View Abstract

Stay Connected to Science

Navigate This Article