Genetic relationships between human groups were first studied by comparing populations - an approach having problems of resolution and dating, but results were largely consistent with an African ancestry for anatomically modern humans (AMH). Following specific lineages rather than populations has since revealed a detailed geography of migrations, revolutionising our knowledge of the peopling of the world, giving stronger proof of the recent near replacement of all human species by AMH.
     Africa is the most likely geographical origin for a modern human dispersal. The basic questions are how many founding exits of AMH can be seen in the fossil or archaeological record, which of these are evidenced genetically, which routes were taken, and when, how, and why.
     There is growing consensus for a single southern dispersal of AMH via the mouth of the Red Sea, around the coasts of the Indian Ocean - initially to Bali, but ultimately to Melanesia and Australia, and to the Americas.
     Although evidence for a single successful ex-African AMH lineage is clear, the scenario is surprising and counterintuitive. In the absence of severe drift, a single exit group would be expected to involve multiple founder lineages spreading to different Eurasian locations.
     A single successful African exit for AMH has several implications. The simplest but most important is that the number of possible subsequent routes decreased. Whichever route was taken initially, the model has to explain how both Europe and Asia could have been colonised from the same single exit group.
     There are genetic reasons for identifying the southern route across the mouth of the Red Sea as the most likely. Dated archaeological evidence for the exit is lacking, because sea-level rise has drowned most coastal remains. Much of the relevant datable evidence lies either in on the Indian subcontinent, or on the eastern side of the Indian Ocean. The first archaeological evidence of occupation of the island of New Guinea has been radiocarbon dated to 49,000 BP.

Edited from Philospohical Transaction of The Royal Society (6 February 2012)