Spatial sampling bias influences our understanding of early hominin evolution in eastern Africa

Abstract

The eastern branch of the Eastern African Rift System (EARS) is the source of a large proportion of the early hominin fossil record, but it covers a tiny fraction (ca. 1%) of the continent. Here we investigate how this mismatch between where fossils are preserved and where hominins probably lived may influence our ability to understand early hominin evolution, using extant mammals as analogues. We show that the eastern branch of the EARS is not an environmentally representative sample of the full species range for nearly all extant rift-dwelling mammals. Likewise, when we investigate published morphometric datasets for extant cercopithecine primates, evidence from the eastern branch alone fails to capture major portions of continental-scale cercopithecine cranial morphospace. We suggest that extant rift-dwelling species should be used as analogues to place confidence intervals on hominin habitat reconstructions. Furthermore, given the north–south orientation of the eastern branch of the EARS, morphoclines that are not aligned along this major north–south axis are likely to be poorly sampled by sites in the eastern branch. There is a pressing need for research on the geography of early hominin morphoclines to estimate how morphologically representative the hominin fossil sample from the eastern branch may be. The Eastern African Rift System (EARS) is a key location for the hominin fossil record, but the fact that it samples a narrow section of the continent has long been known. The authors tackle this known (but largely unaddressed) bias by sampling the distribution and morphospace of extant mammals in the rift, showing that the eastern branch of the EARS fails to capture the full range of diversity and morphology. This approach could be helpful to place confidence intervals on extinct habitat reconstructions, controlling for spatial bias.