Many species require captive breeding to save them from extinction, with reintroduction into the wild being the eventual aim of most programs. Adaptation to captive environments typically results in reduced fitness under wild conditions. Consequently, unintentional adaptation during captive breeding programs may seriously compromise the success of reintroduction programs. However, there is little experimental evidence on the rate or extent of adaptation for captive populations maintained under benign captive conditions for extended periods of time. To investigate the dynamics of genetic adaptation to captivity, large captive populations of Drosophila melanogaster were assessed for relative fitness under captive conditions for up to 87 generations in captivity. Captive fitness increased to 3.33 times the initial fitness over 87 generations. The pattern of adaptation was curvilinear, with an exponential curve providing the best fit. Fitness reached 25% of its maximum within 6 generations, 50% within 15 generations, 75% within 31 generations and 95% within 67 generations. The model predicted that the asymptotic level of fitness reached would be 3.38 times the initial fitness. Thus, very large genetic adaptations to captivity may occur under relatively benign captive conditions. Captive populations destined for reintroduction need to be managed to minimise genetic adaptation to captivity.