The biogeography of hominid evolution

The present consensus of scientific opinion is that man's closest living relatives are the African apes, chimpanzees and gorillas, which combine with humans in the subfamily Homininae (Pilbeam, 1996). This conclusion is based largely on molecular evidence (Ruvolo, 1997), but there is also a body of opinion that is based mainly on morphological evidence which indicates that the closest relative to humans are the orangutans (Schwartz, 1984; Grehan, 2006). Folinsbee & Brooks (2007) address this issue using PACT analysis of area cladograms. There is much dispute over the relative merits of morphological and molecular evidence in resolving this issue (Patterson, 1987), and here I will look to palaeontological and ecological sources to consider this issue. The early record of fossil apes is exclusively African, and the first record outside Africa was about 15.5 million years ago (Ma). After this time, apes are less abundant in the fossil record, and during the period preceding the emergence of the line leading to humans they are almost nonexistent in Africa. What does this mean? Absence of fossils does not necessarily mean absence of ancestral hominids in Africa, but absence of fossils does make it difficult to examine possible morphological or migratory trends. One way around this problem is to examine other mammalian groups that have better fossil records and which are associated with hominins. Folinsbee & Brooks use hyaenids and proboscideans as comparative samples, and their area cladograms show that these two groups had early radiations in Africa during the early Miocene (before 15 Ma), dispersal out of Africa at around 15 Ma, followed by re-entry into Africa in the later part of the middle Miocene (c. 10–8 Ma). They show that the hominid lineage (great apes and humans) followed the same migratory pathways during the middle Miocene. There is some fossil evidence supporting part of this model for hominid migrations. Emigration of apes from Africa at the end of the early Miocene is well established and it is currently being proposed that at least one lineage of fossil apes re-entered Africa in the middle Miocene (Andrews & Kelley, in press). This is the genus Kenyapithecus, a fossil ape from 14 Ma deposits in Kenya but now known from earlier fossils (c. 15.5 Ma) in Turkey. In other words, movement of apes between the continents was happening both ways at this stage in the middle Miocene, made possible by the establishment of land bridges (Rögl, 1999). Note, however, that Kenyapithecus and late Miocene fossil apes lack adaptations linking them with Hominidae (Pilbeam, 1996), while on the contrary some of the European apes such as Dryopithecus do have such characters. This has led to the suggestion that hominids may have arisen in Eurasia (Begun, 1994) and that the ancestor of the African apes and humans (subfamily Homininae) re-entered Africa in the late Miocene, in accord with the PACT analysis of Folinsbee & Brooks. This proposal rests on the twin assumptions that the paucity of fossils in Africa during the period 12–8 Ma is real and not an artefact of sampling, and that if fossils are found in the future they will be seen to lack hominine derived characters. Folinsbee & Brooks argue that the hominid fossil record is close to complete, so that their area cladograms provide realistic results for the dispersal of the hominids; however, this is a phylogenetic argument, and I would argue from an ecological point of view that hominids (and other primates) are greatly under-represented in the African fossil record throughout the Miocene. Many of the fossil apes known from Africa were ecologically similar to living monkeys, not to apes, and in many environments today there can be ten or more sympatric species of monkey. Even in the richest sites in the early Miocene, fossil apes are much less diverse than this, and overall there are far fewer species known from single time horizons. I would estimate that the African fossil record known today samples less than 30% of the species that existed throughout the Miocene. Considering now the ecological source of evidence, it is self-evident that the right environmental conditions have to exist for movement of animal species across and between continents. In the middle Miocene, when the first ‘return to Africa’ occurred (namely the case of Kenyapithecus mentioned above), much of eastern Africa and southern Europe was still covered in tropical and subtropical woodlands, and dispersal of fossil apes that were still arboreal (although with some terrestrial adaptations) was possible provided that land bridges were present between Africa and Eurasia. Later in the Miocene, however, coincident with the uplift of the Himalayas and the development of the Asian monsoon climate, northern Africa dried out (De Menocal & Rind, 1993) and forest areas disappeared. Although there may have been wooded corridors across this region, for example supporting suids adapted to woodland conditions, it is unlikely that they would have supported hominins which were dependent on trees for shelter and fruit for eating. As a result, this limited dispersal across northern Africa to animals with four essential ecological attributes: they had to be mobile, ground-living, large and able to subsist on widely dispersed food types. It is no coincidence that in their analysis, Folinsbee & Brooks investigate hyaenas and proboscideans. Hyaenas today are mobile terrestrial species, and they are widespread across Africa and Eurasia because they are carnivores able to draw on a food source that is globally ubiquitous, namely other animals. Similarly, proboscideans are also mobile terrestrial species, and as the largest of land mammals, both in the Miocene and today, they are able to survive in environments from desert to tropical forest. To demonstrate this relationship between ecological attributes and dispersal distances of individual species, Fig. 1 shows the average species range sizes of African mammals for four size categories and demonstrates that species of large terrestrial carnivores have the greatest mean population ranges amongst living species. What does this mean for human evolution? The average species range size of African mammals for four size categories. The range size data are linear ranges based on distributions over an equal area grid of 158 km square, so 1 unit = 158 km (Andrews & O'Brien, 2000). The human species today has a geographic range unique for vertebrates, encompassing the four corners of the world. This first manifested itself during the Pliocene, when human ancestors (hominins) expanded over much of sub-Saharan Africa, a range unusual not just for primates but for mammals in general. Towards the end of the Pliocene, after 2 Ma, we have the first fossil evidence of expansion out of Africa, and by the beginning of the Pleistocene most present classifications have a single species (Homo erectus) with a range from southern Africa to Indonesia. This is a distribution that is only matched today by large terrestrial carnivores, and it strongly suggests that the early Pleistocene expansion and dispersal into Asia could only have been possible for a large species with a carnivorous diet. During the late Miocene, however, when Folinsbee & Brooks suggest ancestral hominins re-entered Africa along with hyaenids and proboscidea, the fossil apes were small, partly arboreal frugivores/herbivores and lacked the dispersal mechanisms enjoyed by the other two groups which enabled them to achieve inter-continental geographic ranges. The ecological conditions supporting earlier migrations had now changed, and while the phylogenetic conclusion is interesting, it is not supported ecologically. Folinsbee & Brooks have added to the issue of the origin of the human lineage, but final resolution must await the discovery of further fossil evidence. Editors Dov Sax and Robert Whittaker