Tracing the paths of modern humans from Africa.

In the 1980s, genetic and fossil evidence began to call attention to Africa’s preeminence in the origins of modern human populations (1), but this evidence could be interpreted in two fundamentally different ways (2). Was Africa’s role greater than other continents because it always harbored a larger human population (size) or because modern humans arose in Africa first and subsequently expanded their range across the world (time)? In the 2000s, improvements in DNA sequencing technology and genetic sampling of more present day human groups made it possible to accurately characterize the genetic diversity of groups from different regions of the world, and it became clear that within-group genetic diversity decreased predictably with increased geographic distance from subSaharan Africa (3, 4). Subsequently, similar, albeit weaker, relationships were found between within-group variation in aspects of skeletal morphology (cranial, dental, and pelvic measurements) and distance from subSaharan Africa (5–8). These results effectively settled the size vs. time debate, because it was hard to imagine how the observed geographic distribution of within-group diversity could have arisen without a recent range expansion of modern humans from Africa. However, how exactly did the spread out of Africa happen? In PNAS, Reyes-Centeno et al. (9) are, to my knowledge, the first to use both morphological data (cranial measurements) and genetic data (single nucleotide polymorphisms) to explicitly evaluate competing models for the expansion of modern humans from Africa. They conclude that after modern humans left Africa, they first went south, only later heading north in a second expansion wave. The most straightforward option for the peopling of Europe is an expansion wave that passed through the Levant (north along the eastern Mediterranean coast), but there are other possibilities to consider for the peopling of Asia and Australia. Modern humans may have initially migrated along the coasts of the Arabian Peninsula and southern Asia, allowing them to quickly reach Australia, and only later ventured north into Europe and northern Asia. This “southern route” model of two waves of expansion was motivated, in part, by the need to accommodate 60,000-y-old dates for the occupation of Australia (10, 11), but subsequent assessments of the Australian archaeological record (12) have shifted the timing in the minds of most scholars to 45,000 y ago. Intuition suggests that it may have been easier for modern humans to migrate latitudinally along broadly similar coasts (Fig. 1), but in contrast to the northern route for which there is substantial archaeological evidence (13, 14), so far there is little evidence for the southern route. Often, morphological and genetic datasets are studied with distinct analytical approaches by different research teams, and, after the fact, the insights provided by each are synthesized. Reyes-Centeno et al. take an uncommon and more sophisticated approach; one that should become more widespread in studies of morphology as more and more genomic data become readily available. Using analyses grounded in evolutionary theory (quantitative and population genetics), they are able to handle seemingly dissimilar morphological and genetic evidence within a common quantitative framework. If assumptions are made about the inheritance of cranial characteristics, cranial shape variation can be used to estimate the genetic dissimilarity of pairs of human groups. Together, these dissimilarity estimates combine to form a matrix of pairwise distances. An analogous distance matrix can be generated from patterns of DNA sequence variation. Once these matrices are calculated, they can be related through evolutionary theory to different models for the expansion of modern humans out of Africa, as long as certain assumptions hold, such as that human groups diverged morphologically and genetically by genetic drift (chance changes in gene frequencies) alone. Following the logic of this approach, if modern humans expanded in a single wave, neighboring human groups would always be expected to be quite similar to each other in their morphological and genetic distances, because they arose at about the same time from the same wave. In contrast, if there were two expansion waves, neighboring human groups derived from different waves would be expected to be quite different from each other relative to neighboring groups derived from the same wave, because groups derived from different waves would trace their shared ancestry all of the way back to before modern humans left Africa. Some readers may balk at the suggestion that the morphological distances between human groups were generated by genetic drift rather than natural selection, but multiple studies have demonstrated that patterns of human cranial variation are consistent with this assumption (15, 16). Potentially more problematic is the assumption that present day patterns of morphological and genetic variation primarily reflect ancient expansions from Africa rather than more recent migrations or gene flow. What are the broader implications of the southern route model? Most intriguingly, it may complicate scenarios for interbreeding between expanding modern humans and Neanderthals (17). If Australasia was colonized by an earlier expansion wave than the one that peopled Eurasia, then why do all non-Africans, including Australian and Fig. 1. Arabian Peninsula coast (Oman). Photo courtesy of Wikimedia Commons/Ji-Elle.