Estimation, variance and optimal sampling of gene diversity II. Diploid locus.

Nei's analysis of diversity at a diploid locus is extended to a population subdivided into a large number of subpopulations. The diversities and the heterozygotes frequency are defined with respect to the total population and unbiasedly estimated in a two-stage random cluster sampling. The fixation indices F IS, F IT andF ST are derived, then inter- and intra-population variances of the estimated parameters are studied. We show that there is a unique sample size per population which yields the best accuracy in estimatingF ST and F IS, respectively, at a given locus. These results are illustrated with an analysis of DNA diversity in a forest tree and compared to those obtained under the Hardy-Weinberg assumption.