Abstract. -When Drosophila simulans and Drosophilafunebris were cultured together in population cages over many generations, there was a prolonged period ofapparently stable coexistence followed by a rapid exclusion of D. funebris. As both species maintained large population sizes in monocultures it follows that the extinctions of D. funebris in the mixed-species cultures must have been caused by D. simulans. The time to extinction of D. funebris ranged from 26 weeks in one cage, to between 40 and 48 weeks in the other five.To test the idea that an evolutionary increase in competitive ability ofD. simulans had OCCUlTed during the course of its interaction with D. funebris, a single-generation experiment was set up. In this experiment the interspecific competitive ability of a population of D. simulans that had been in competition with D. funebris for 44 weeks was compared to that of a stock population that had had no previous contact with D. funebris.In this experiment both stock and precompeted populations of D. simulans increased the eggto-adult development time ofDifunebris. However, precompeted D. simulans caused a significantly greater increase in the development time of D. funebris than did stock D. simulans. Thus D. simulans had evolved an increase in competitive ability as a result ofits interaction with D.funebris.Development time is important because in the population cages the resource bottles-in which the larvae reside-were replaced every three weeks. An increase in development time ofD. funebris in the multigeneration experiment similar to that observed in the single-generation experiment would lead to a rapid decrease in adult population size, resulting in the extinction of this species, as was observed to happen.Received November 6, 1989. Accepted June 5, 1990.The question of how a species responds lize interactions), two main conclusions can to selection pressures resulting from com-be drawn from the laboratory studies carpetitive interactions with other species is ried out so far. First, a common outcome both important and complex. It is impor-in such studies is for no evidence of evotant because the kind of evolutionary re-lutionary changes to be found (e.g., Park and sponsecanaffectthestabilityofcoexistence, Lloyd, 1955;Lerner and Ho, 1961; Dawthereby having a potentially major role in son, 1972 for Tribolium; Sokal et al., 1970 determining the diversity of guilds; and for Musca; Sulzbach, 1980; Pruzan-Hotchcomplex because rarely do we know the bi-kiss et al., 1980 for Drosophila). This is deological or genetic basis of changes in com-spite an almost certain bias in the literature petitive ability, which makes any predic-in favor of publication of positive results. tions or interpretations about the character, Second, in those studies in which there were or suite of characters, very difficult.at least some positive results, these were Before one can begin to answer questions often inconsistent between replicates (e.g., on the relative importance of the various Futuyma, 1970;Barker, 1973; Sulzb...