We manipulated experimental populations of the housefly (Musca domestica L.) under three inbreeding schemes (fast, slow, and punctuated) to partition out the influences of different means and variances in the rate of inbreeding, per generation, while controlling for the final level of inbreeding as a constant. One treatment used constant fast inbreeding (11% per generation; N_e = 4 for 4 generations), for a comparison to one that was consistently slow (3% per generation; N_e = 16 for 14 generations). The third followed a model for serial founder-flush events. Each founder-flush episode involved a one-generation pulse of fast inbreeding (N_e = 4) followed by two generations of very low (or no) inbreeding, yielding high intergenerational variation (i.e., for an average inbreeding rate of 4% per generation). Allozyme assays showed that we achieved the intended final inbreeding coefficient of about 37%. All inbreeding schemes decreased fitness levels in terms of egg-to-adult viability, development time, and male mating success relative to the outbred control. The consistently fast inbreeding protocol had more pronounced reductions in fitness, relative to the other two inbreeding schemes. In comparison to the fast and punctuated regimes, the consistently slow protocol preserved evolutionary potential (as assayed by the genetic divergence of subpopulations exposed to different environments) in egg-to-adult viability, and (albeit anecdotally) reduced the extinction probabilities, especially in a novel environment. The punctuated treatment did not optimize the potential for purge as predicted, but instead reduced fitness, evolutionary potential, and environmental responsiveness (as measured by genotype-by-environment interactions). This founder-flush treatment also had the highest extinction probabilities. Longer periods of population flush might be necessary to purge effectively in a punctuated scheme. We conclude that the rate of inbreeding, independent from the final level, can have important effects on population fitness, environmental responsiveness, and evolutionary potential.