There is a new paper about a recently discovered skull from Dmanisi, Georgia, an early Homo erectus. Compared with a set of other erectus skulls from the same area and from about the same time, it looks as if this population had a fair amount of variation – so some are arguing that various skulls in Africa that were assigned to several different species may well all be members of a single fairly variable species. Possibly evidence for lumpers, in their eternal struggle with splitters.
Milford Wolpoff, in reaction to this work, said “Everyone knows today you could find your mate from a different continent and it is normal for people to marry outside their local group, outside their religion, outside their culture,” Wolpoff told Agence France Presse. “What this really helps show is that this has been the human pattern for most of our history, at least outside of Africa,” he added. “We don’t have races. We don’t have different subspecies. But it is normal for humans to vary, and they have varied in the past.”
That’s all wrong. For most of human existence mating with people from substantially different groups was rare – usually very rare. For one thing, there were no letters for your baby to send, let alone aeroplanes to take. Long-distance travel was very rare. Most people never even met anyone from another race or subspecies.
If two populations exchange exchange genes much, they can’t become very different via drift. To be exact, you don’t expect to see much divergence in neutral gene frequencies as long as at least one individual is exchanged per generation. You heard me right: ONE. Assume that Old Stone Age people in sub-Saharan Africa were effectively a single population, while Neanderthals were another. Judging from the measured genetic distance between those two groups, gene flow was very low – averaging quite a bit less than one individual per generation for hundreds of thousands of years.
In terms of populations that exist today, Bushmen seem to have been isolated – really isolated, less than 1 mating across the line per generation – for something like 150,000 years. Until fairly recently. In the Holocene, the last 10,000 years or so, there’s been a whole lot of shaking going on. Many Bushmen groups have a fair amount of Bantu admixture, and surprisingly, they also seem to have a few percent of ancestry picked up from an expansion of pastoral peoples that (in part) traces all the way to the Middle East [Joe Pickrell’s work]. But that’s all recent.
Populations can also come to differ significantly because of natural selection, rather than drift. That can happen much more rapidly. Gene flow between populations interferes with this process, but it takes a lot more gene flow to stop selection than it does to stop drift.
None of this means that individuals from long-isolated populations couldn’t mate successfully, given the chance. It takes a long time for that kind of infertility to develop – more than a million years, on average. Lions and tigers, which have been separated for about two million years, can still produce fertile female offspring, although male ligers. are sterile. This makes another point: fertility isn’t all or nothing. There can be degrees of interfertility. A given kind of hybrid mating might have an average fertility that was 90%, or 50%, or 10% of normal. Haldane’s rule says that, in mammals, male hybrids are more likely to take it in the shorts. Or, as genetic distance between parents increases, fertility problems show up first in male offspring.
If you knew how long ago Neanderthals and Denisovans split off from the ancestors of anatomically modern humans (something like half a million years), it was possible to see that AMH could probably mate successfully with those archaic humans. In the course of expanding out of Africa, AMH had to run into Neanderthals, and so some degree of admixture was likely. And since even a wee bit of admixture is enough to transmit alleles with a selective advantage, that admixture had some functional consequences [upcoming paper by Sankararaman]. None of this means that Neanderthals weren’t significantly different from modern humans.
It doesn’t mean that subspecies didn’t exist, or don’t exist. And the existence of some interfertility doesn’t mean that there were no fertility problems in AMH-Neanderthal crosses. The amount of Neanderthal ancestry is considerably lower than average on the X-chromosome [about a third of the normal fraction] . This suggest a certain amount of hybrid incompatibility, since incompatibility loci are known to concentrate on chromosome X [recent work from Sriram Sankararaman at the Reich lab] .
The odds are fair that a modern human could successfully mate with a Floresian hobbit, assuming that it’s really an island-dwarfed version of Homo erectus. Even though it was three feet tall and had a chimp-sized brain. Somehow that doesn’t convince me that races and subspecies don’t exist. Nor does it convince me that mating with people from very different populations was common. Mating across the lines happened before the Holocene, and sometimes, when it transmitted adaptive alleles, it was very significant.
But it was rare.