The rate of adaptation to a competitive laboratory environment and the associated inbreeding depression in measures of reproductive fitness have been observed in populations of Drosophila melanogaster with mean effective breeding size of the order of 50 individuals. Two large wild-derived populations and a long-established laboratory cage population were used as base stocks, from which subpopulations were extracted and slowly inbred under crowded conditions over a period of 210 generations. Comparisons have been made of the competitive ability and reproductive fitness of these subpopulations, the panmictic populations produced from them by hybridization and random mating and the wild- or cage-base populations from which they were derived. After an average of approximately 180 generations in the laboratory, the wild-derived panmictic populations exceeded the resampled natural populations by 75% in fitness under competitive conditions. The cage-derived panmictic population, after a total of 17 years in the laboratory, showed a 90% superiority in competitive ability over the corresponding wild population. In the inbred lines derived from the wild-base stocks, the average rate of adaptation was estimated to be 0.33 +/- 0.06% per generation. However, the gain in competitive ability was more than offset by inbreeding depression at an initial rate of approximately 2% per generation. The effects of both adaptation and inbreeding on reproductive ability in a noncompetitive environment were found to be minor by comparison. The maintenance of captive populations under noncompetitive conditions can therefore be expected to minimize adaptive changes due to natural selection in the changed environment.