Stocking with offspring of local wild fish, so-called supportive breeding, is often advocated as an alternative to stocking domesticated fish. However, it is important to ensure that supportive breeding does not result in inbreeding and loss of genetic variability. We analysed eight microsatellite loci in samples of wild and hatchery-reared brown trout (Salmo trutta) from three populations subject to supportive breeding. For calibrating statistical procedures, we included two test samples of reared offspring for which the precise number of parent fish was known and a sample from a further wild reference population. Three different statistical procedures were used to detect population bottlenecks and loss of variability: (i) a randomization test for comparing allelic diversity between samples; (ii) estimates of effective number of breeders from gametic-phase disequilibrium; and (iii) a test for assessing population bottlenecks based on detecting deviations from mutation-drift equilibrium. All three procedures were useful but they also exhibited different strengths and limitations, with the test for population bottlenecks probably being the single most useful procedure for routine monitoring. In two populations subject to supportive breeding, there were strong indications of reduced effective population sizes, and significant genetic differentiation was observed between different samples from the same population.
