Center and edge

Fifty years ago in science:

Various ideas and data more or less closely related to the present hypothesis are apparently widespread in the literature (e.g., Matthew, 1915; Lutz, 1916; Timoféeff-Ressovsky, 1940, among others), and it would be interesting but difficult to trace the historical continuity of thoughts along this line. For instance, there is the venerable notion that variation in the genus Homo has proceeded prevailingly outward from more central continental areas into peripheral Europe, riding on successive population waves. There seem to be no serious objections to this point of view. It may well be that speciation in early man depended on central-peripheral processes as outlined in the present hypothesis. In fact, I predict that good evidence for this will be available in a very short time (Brown 1957:265).

Well, maybe not a short time, but one or another variant of this idea have certainly dominated both before and after that paper.

A problem is that there are at least two different evolutionary patterns that can lead to the outcome of variable center and less variable periphery populations. One is the repeated proliferation of colonizing species at the center and their partial or complete replacement of populations at the periphery. This is the "centrifugal speciation" model. The other is radial gene flow directed from the center to the periphery.

The difference between these two alternatives is really a matter of degree. Both scenarios depend on a common source-sink population dynamic. The peripheral populations absorb more migrants from the center than they produce; the center populations produce more new genetic variants. Possibly the center population includes distinct morphs with different colonizing potential, and the peripheral populations ultimately receive only the best colonizers. The difference between the "centrifugal speciation" model and simple radial gene flow is a matter of clumpiness -- do genes tend to move together with each other in discrete migrations (with possible population replacements) or individually with continuous genetic exchanges?

Another way to put this distinction is to consider the behavior of clines of variation over time. With recurrent colonization from a central population, clines of variation tend to move in synchrony with each other as multiple genetic systems (possibly coadapted gene complexes) disperse with new colonists into the peripheral populations. At the extreme of speciation, all genes move together, establishing new clines entirely (although some genes may introgress from the original peripheral populations). In contrast, under long-term radial gene flow, geographic clines may evolve nearly independently from each other for different genes (depending on linkage and interactions among them).

There are two scenarios in which the two patterns can be differentiated. If one of the peripheral populations expands in size and itself becomes a colonizer (possibly impinging or displacing a central population), the overall pattern of variation may be noticeably different from the simple centrifugal pattern. Or, as Thorne and Wolpoff (1981) suggest, recurrent gene flow from the central populations may be opposed by selection in the peripheries, leading to long-term stable clines.

Which pattern characterized human evolution? Almost certainly, it was different at different times and in different places -- sometimes major dispersals, other times long-term gene flow.


Brown WL, Jr. 1957. Centrifugal speciation. Q Rev Biol 32:247-277.

Thorne AG, Wolpoff MH. 1981. Regional continuity in Australasian Pleistocene hominid evolution. Am J Phys Anthropol 55:337-349. doi:10.1002/ajpa.1330550308