"Out of Africa" jumps the shark?

The December 2004 Discover magazine ran an article called "The hidden history of men" covering Spencer Wells research into ancient migrations from Y chromosome evidence. In the February 2005 issue, there is a letter to the editor about the article. The letter asks a very reasonable question about the colonization of the Americas: if genetic evidence excludes a date earlier than 20,000 years ago for the arrival of the first humans, then what are we to make of even older archaeological sites-as ancient as 50,000 years ago?

The magazine runs a reply by Spencer Wells that is a model of creative thinking. He divides current opinion on the initial habitation of the Americas into two camps:one camp favoring a settlement from Siberia within the past 20,000 years, and another suggesting an earlier migration as early as 40,000 years ago. He then argues that any sites as early as 50,000 years ago probably are not evidence of human habitation, but instead are evidence of natural processes-in other words, they can't be real archaeological sites.

But in the event that the sites are documented as being real evidence of human activity, Wells has another option. He suggests that the sites may be evidence of the activity of an earlier human species, not related to living Native Americans. Presumably this ancient species would be a descendant of Homo erectus, as has been suggested for the dwarf hominids from Flores:

H. erectus was primarily a tropical or subtropical species, and there is no evidence suggesting that it ever managed to live in the arctic. Perhaps one day we will find the remains of a subpopulation that adapted to life in northern latitudes in the same way that Neandertals did in Europe, leaving them poised to make the leap into the Americas. . . . The genetic evidence we have is clear: all non-Africans came out of Africa within the past 60,000 years, and the ancestors of today's Native Americans entered the Americas within the past 20,000 years. If there were other hominids living here when modern humans arrived, they must have died out.

Wells oversimplifies the current diversity of opinion on the origins of Native Americans, but this kind of simplification is understandable in a short reply to a letter. He eliminates discussion of the Clovis-first school of thought, but today this hypothesis is in decline. He also omits discussion of whether migrants came only from Siberia or whether they also came from other parts of the world: in particular other parts of Asia or Europe. In my view these hypotheses are not especially likely, and they rely either on older genetic evidence that is now equivocal, or archaeological or anatomical resemblances that may reflect similarities other than close relationship. But even so, it is fair to say that there is more than a two-fold division of opinion on the issues.

Of greater concern is his simplification of the genetic evidence for migration time. He is correct that the Y chromosome supports a relatively recent entry into the New World-with the youngest haplotype shared by both Asians and Native Americans appearing to give a time bracketed between around 15,0000 and 12,000 thousand years ago. This is in fact the strongest genetic evidence about the date of movement of people into the New World, with other genes much more equivocal about the date. Researchers of mtDNA have suggested a much earlier date of entry, or an earlier population expansion, while examination of other nuclear genetic evidence has been less informative. So it is fair to say that the genetic evidence does support an initial habitation after 20,000 years ago, but the strength of the evidence leaves something to be desired-certainly if it were compared to archaeological evidence for an earlier entry. The problem with a hypothesis of an earlier habitation is clearly that the archaeological evidence for it is weak.

So where does Wells get his idea about the survival of Homo erectus and its habitation of the New World? Clearly if he is unwilling to abandon the assumption that human genes document a movement out of Africa within the past 60,000 years, he has little choice. Certainly humans could not have reached the New World so quickly after their origin in Africa. But Wells it is wrong to be so confident about this date. He appears to have a sort of tunnel vision favoring evidence from the Y chromosome above all others. What if other genes give an earlier date for movement out of Africa? Could the Y chromosome just have the wrong chronology?

Or, more critically, what if the Y chromosome is giving us evidence of ancient patterns of natural selection? After all, the distribution of variation of the Y chromosome is clearly a nonequilibrium distribution, in contrast to other nuclear genes. Is there any reason to think that humans cannot have reached the New World as early as 50,000 years ago, and only later have acquired the selected Y chromosome variant? The fallacy is the assumption of identity between the genetic distributions and population distributions. Unless we can put this assumption behind us, we will have no end of wild hypotheses to continue to support it.