Genic Variation in Natural Populations

By characterizing the struggle to measure variation as a struggle between two opposing views, the classical and balance theories of population genetics, I have in part obscured the underlying problem of making a functional whole out of theory and observation. Suppose that the experiments of Wallace and of Mukai had proved beyond any reasonable doubt that new mutations were, on the average, heterotic, or that we found the evidence from the great variety of selection experiments absolutely compelling. Then the balance hypothesis would be admitted and we would know that there was a tremendous amount of genetic heterozygosity at most loci. Yet even then we would be no closer to an empirically and dynamically sufficient genetic description of populations. Knowing that populations in general must be highly heterozygous will not tell us how fast or how far natural selection can go in changing a particular character, nor will it allow us to make any inferences about the past history of populations and geographical races, or about the genetic processes in speciation. The answers to these latter questions, which comprise the task that evolutionary geneticists have set themselves, require the estimation of the quantities that appear in population genetic theory; they require the characterization of gene frequencies. The whole of population genetic theory remains an abstract exercise unless the frequencies of alternative