We investigated the effects of genotypic frequencies on egg-to-adult viabilities in pairwise combinations of four strains of Drosophila melanogaster. The experiments involved mixture of a total of 42,000 eggs in varying proportions under controlled densities and observation of surviving adults. Viabilities were found to depend on frequencies in several genotypic combinations. In the most extreme case, the absolute viability of cn;bw females increased monotonically from 54% when common to 70% when rare. The results illustrate several statistical and methodological problems that might explain why some experiments have failed to detect frequency-dependent viabilities. These problems include heterogeneity between replications, sex differences in susceptibility to competition, and strong dependence of the experimental outcome on the choice of competitor genotypes.