This experiment was designed to study the relationship between rate of inbreeding and observed inbreeding depression of larval viability, adult fecundity and cold shock mortality in Drosophila melanogaster. Rates of inbreeding used were full-sib mating and closed lines of N=4 and N=20. Eight generations of mating in the N=20 lines, three generations in the N=4 lines and one generation of full-sib mating were synchronised to simultaneously produce individuals with an expected level of inbreeding coefficient (F) of approximately 0.25. Inbreeding depression for the three traits was significant at F=0.25. N=20 lines showed significantly less inbreeding depression than full-sib mated lines for larval viability at approximately the same level of F. A similar trend was observed for fecundity. No effect of rate of inbreeding depression was found for cold shock mortality, but this trait was measured with less precision than the other two. Natural selection acting on loci influencing larval viability and fecundity during the process of inbreeding could explain these results. Selection is expected to be more effective with slow rates of inbreeding because there are more generations and greater opportunity for selection to act before F=0.25 is reached. Selection intensities seem to have been different in the three traits measured. Selection was most intense for larval viability, less intense for fecundity and, perhaps, negligible at loci influencing cold shock mortality.