Extensive individual variation in spatial behaviour is a common feature among species that exhibit migratory life cycles. Nowhere is this more evident than in salmonid fishes; individual fish may complete their entire life cycle in freshwater streams, others may migrate variable distances at sea and yet others limit their migrations to larger rivers or lakes before returning to freshwater streams to spawn. This review presents evidence that individual variation in migratory behaviour and physiology in salmonid fishes is controlled by developmental thresholds and that part of the variation in proximal traits activating the development of alternative migratory tactics is genetically based. We summarize evidence that alternative migratory tactics co-exist within populations and that all individuals may potentially adopt any of the alternative phenotypes. Even though intra-specific genetic divergence of migratory tactics is uncommon, it may occur if female competition for oviposition sites results in spawning segregation of alternative phenotypes. Because of their polygenic nature, alternative migratory tactics are considered as threshold traits. Threshold traits have two characteristics: an underlying 'liability' trait that varies in a continuous fashion, and a threshold value which is responsible for the discreetness observed in phenotypic distribution. We review evidence demonstrating that body size is an adequate proxy for the liability trait controlling the decision to migrate, but that the same phenotypic outcome (anadromy or residency) may be reached by different developmental pathways. The evidence suggesting a significant heritable component in the development of alternative migratory tactics is subsequently reviewed, leading us to conclude that alternative migratory tactics have considerable potential to respond to selection and evolve. We review what is known about the proximal physiological mechanisms mediating the translation of the continuous value of the liability trait into a discontinuous migratory tactic. We conclude by identifying several avenues for future research, including testing the frequency-dependent selection hypothesis, establishing the relative importance of adaptive phenotypic plasticity in explaining some geographic gradients in migratory behaviour and identifying the physiological and genetic basis of the switching mechanisms responsible for alternative migratory tactics.