1. The relative viability of brown homozygote was measured in various levels of larval population density which paralleled with the number of female parents in a culture bottle. The results were shown in Table 1 and Fig. 1, and confirmed the fact that the superiority in viability of brown homozygote increased with larval population density.2. Two series of competition experiments were conducted: series S, in which ten or less parental flies were sampled generation after generation, and series L, in which generally larger number of parents were sampled. By assuming that the zygotic selection coefficients were determined only by the relative viability of three genotypes, and by using Fig. 1, random change in frequency of brown genes were calculated from generation to generation.3. In both series the distributions of the amounts of random change relative to expected standard deviation of change in gene-frequency could be regarded to have mean values of 0. This result showed that the above assumption was practically correct.4. Average effective population size was considered to be smaller than average apparent size (N) on account of, at least, three factors, i, e., (i) the departure of parental sex-ratio from one-to-one, (ii) the variability of the numbers of progeny of each parent and (iii) the parental inbreeding. Thus, the formulaN=αβN.was presented, where α (0<α≤1) and β (0<β≤1) were the coefficients measuring influences of the factors (i) and the other factors, respectively.5. The values of α and β, and those of their poducts were calculated as follows:Series S: α=0.84, β=0.35-0.62, αβ=0.29-0.52.Series L: α=0.92, β=0.22-0.30, αβ=0.20-0.27.The values of β, especially in series L, seemed to be too small in comparison with the results of several antecedents. It could be interpreted, at least in part, as a result of circumstances that the number of effective parents was only a small fraction of total parents sampled; and such circumstances were considered to be remarkable in crowding condition as in series L.