The genetic variance of mortality rates in Drosophila melanogaster increases with age early in life, but declines at intermediate ages. A simulation study was done in order to evaluate two competing explanations for this age-specific pattern: (1) demographic heterogeneity, and (2) binomial sampling effects. The pattern can be explained by demographic heterogeneity among and within genotypes. In contrast, binomial sampling variance is not sufficient to explain this age-specific pattern of genetic variance in mortality rates. A previous publication that rejected the 'heterogeneity explanation' in favor of binomial sampling is shown to be mistaken.