Genetic Structure of Small Populations

Abstract

This review covers a body of recent literature on formal approaches to the analysis of the population genetics of small populations. My approach is critical rather than comprehensive and several topics are not considered at all. Because of length limitations, this is a very incomplete review, and it thus reflects my personal perception of new and significant work. Simulation seems not to have made any substantial contribution yet to anthropological population genetics, for example, and I cite only the excellent review by MacCluer (39). Similarly there are a number of good informal papers on the structure of mating practices in regions (e.g. 6, 41), but these are now difficult to incorporate into formal procedures. The organization of this review is by methodology, since that is the topic around which most debate is centered and since few regional or any other sort of generalities have emerged yet. My conclusion will be that studies of the genetic structure of small populations have made particular and incidental contributions to formal genetics, to regional history and prehistory, to epidemiology, and to several other fields to which they are peripheral, but that they have not advanced our understanding of human evolution in a global sense. The sample sizes available have been too small to allow reliable inferences about natural selection; the extensive occurrence of what is presumably local random genetic drift has little or no consequence for evolution over long time periods over large areas; and the presumed selective agents in the various environments of these peoples differ greatly so that few of the generalizations which have been put forward hold for many groups. As it became clear that natural selection ofreasonable magnitude would not be detectable in sample sizes such as could be obtained from primitive populations, and as the idea gained currency in biology that much polymorphism could be the transient outcome of the occurrence of selectively neutral mutation, the thrust of many of these studies became the study of local random genetic drift. This leads to comparison of the amount and pattern of variability in local gene frequencies with the level of variability predicted by the amount and pattern of gene exchange