Inbreeding depression varies considerably among populations, but only some aspects of this variation have been thoroughly studied. Because inbreeding depression requires genetic variation, factors that influence the amount of standing variation can affect the magnitude of inbreeding depression. Environmental heterogeneity has long been considered an important contributor to the maintenance of genetic variation, but its effects on inbreeding depression have been largely ignored by empiricists. Here we compare inbreeding depression, measured in two environments, for 20 experimental populations of Drosophila melanogaster that have been maintained under four different selection regimes, including two types of environmentally homogeneous selection and two types environmentally heterogeneous selection. In line with theory, we find considerably higher inbreeding depression in populations from heterogeneous selection regimes. We also use our data set to test whether inbreeding depression is correlated with either stress or the phenotypic coefficient of variation (CV), as suggested by some recent studies. Though both of these factors are significant predictors of inbreeding depression in our study, there is an effect of assay environment on inbreeding depression that cannot be explained by either stress or CV.