1. We suggest that two important hypotheses underlie the translocation of animals for conservation management, namely that performance is higher for: (i) animals acclimated to a novel site before release (soft releases), compared with those not acclimated (hard releases); and (ii) wild-caught animals compared with captive-bred animals. Despite the increased use of translocations, these hypotheses remain largely untested. 2. Before reintroductions of the dormouse Muscardinus avellanarius, experimental translocations were conducted. These examined the performance of released dormice in relation to season, release methods and animal origin (captive-bred or wild-caught). 3. Early releases (May, June) lost body mass at a significantly higher rate than late releases (August, September). This resulted from seasonal food shortage and could be reversed by supplementary feeding. All released dormice appeared not to locate some important food resources. 4. Soft releases continued to nest at their release point and utilized supplementary food. Hard releases often dispersed and ignored food provided. These differences have profound consequences for reintroductions and hypotheses to explain them are proposed. 5. Wild-caught and captive-bred dormice responded differently to pre-release pens. After release, captive-bred animals travelled significantly less far, in total and from nests, each night. This would have reduced their chances of locating natural food sources and suggests that they are slower to establish in novel habitat. 6. The experiments support the hypotheses underlying translocations and imply that there may also be a common mechanism underlying the behaviour of translocated animals.