A GENERAL pattern that emerges from electrophoretic analyses of protein variation in natural populations is that alleles with intermediate frequencies are rare. This results, graphically, in a U-shaped frequency distribution when both allele frequencies from a two-allele locus are plotted. Supporters of the neutralist theory of population variation believe that most protein polymorphisms found in nature are the result of the random drift of neutral mutations and argue that a U-shaped distribution is expected for the case 4Ns 1. From this we conclude that the analysis of allele frequency distributions (or, equivalently, relative contributions to heterozygosity) does not provide a sensitive means of resolving the neutralist and selectionist controversy.