Background: Progress in understanding senescence requires information about age-dependent changes in individual organisms. However, for experimental systems such as Drosophila melanogaster we usually do not know whether population trajectories are an accurate guide to patterns of individual aging, or are artifacts of population heterogeneity.

Objective: In experimental cohorts of D. melanogaster, population trajectories of age-specific fecundity typically plateau in old age. Here we ask whether fecundity plateaus can be explained by hidden heterogeneity in reproductive life spans (RLS).

Methods: Using published and original data from five experimental populations, we examined fecundity trajectories in subpopulations stratified by RLS, and in total populations with age expressed relative to RLS for each individual fly. We also executed computer simulations of reproductive senescence in which individuals vary in RLS, and show linear decline in fecundity with increasing age.

Results: In subgroups of flies with similar RLS, the senescent decline of fecundity is generally linear. Population trajectories in which age is expressed relative to individual RLS also exhibit linear or slightly accelerated decline. Simulations demonstrate that observed levels of variation in RLS are sufficient to produce plateaus in the mixture trajectories that are very much like those observed in experiments, even though fecundity declines linearly in individuals.

Conclusions: Late-life fecundity plateaus in D. melanogaster are artifacts of population heterogeneity in RLS. This conclusion applies to both inbred and outbred populations, and both "wild" and lab-adapted stocks. For reproductive senescence in this model system, population trajectories are not an accurate guide to individual senescence.