Survival (viability) of newly eclosed adults ofD. simulans st and D. melanogaster Or R-C, which had excess quantities of dead yeast available throughout life, was measured daily for 7 days in an experiment where adult density (6 levels), species frequency (6 levels) and a2p tissue content (2 levels) were varied factorially. A separate experiment comparing viability in different types of experimental unit also was done. Similar experiments were done for D. simulansst competing against D. melanogastery w. The entire data were subjected to a least-squares (unbalanced, missing-plot) analysis of variance.D. melanogaster Or R-C had a higher average viability than D. melanogastery w or D. simulans st, which were equal. The competing strain of D. melanogaster influenced D. simulans st viability -viability being higher when y w was the competitor. Viability decreased over the seven day period but at different rates for the three strains. Increasing density reduced viability for all three strains, but species frequency effects, although significant, were generally not consistent. Females had higher viability than males in both D. mdanogaster strains, but the reverse was true in D. simulans st. 32p .lowered viability and experimental unit type altered viability. Numerous inter. actions were significant.Adult density was shown to have a delayed effect on viability -the delay before the appearance of the effect (an increase in death rate) being decreased as density rose. The term 'variably delayed density dependent' has been adopted to describe the fitness component, adult viability. Some high density populations showed a readjustment (a decrease) in their death rate as a reaction to the effect of reduced density caused by high early mortality.