Conventional narratives explain fish migrations in term of requirements (food, mates, habitats, etc.), with adequate temperatures being optional. Here, using the example of a (commercially extinct) stock of Black Sea mackerel (Scomber scombrus), we suggest that seasonal migrations are driven by seasonal temperature cycles. Therein, temperature acts as a constraint determining where the fish can be at any given time, and not a one of several factors which they would consider when choosing between alternative migration routes. Generalizing, we suggest that temperature should generally be an explicit part of hypotheses about the migratory behaviours of marine fishes. For illustration of what may occur when this is not the case, it is suggested that the non-consideration of temperature in a model of North Atlantic mackerel migration may have led, among the researchers concerned, to a sense of complacency with respect to the climate change-induced changes in the phenology of this fish in the North Atlantic, whose distribution and migration are misleadingly seen as "stochastic".