Process in the evolution of bird migration and pattern in avian ecogeography

Current ideas about the evolution of bird migration equate its origin with the first appearance of fully migratory populations, and attribute its evolution to a selective advantage generated by increased breeding success, gained through temporary emigration from resident populations to breed in under‐exploited seasonal areas. I propose an alternative hypothesis in which migration first appears as a temporary directional shift away from the breeding site outside the reproductive period, in response to seasonal variation in the direction and/or severity of environmental gradients. Fully migratory populations then appear through either extinction of sedentary phenotypes, or colonisation of vacant seasonal areas by migrants. Where colonisation occurs, resident ancestral populations can be driven to extinction by competition from migrants which invade their range outside the breeding season, resulting in fully migratory species. An analogous process drives the evolution of migration between high latitudes and the tropics, since extension of breeding range into higher latitudes may drive low latitude populations to extinction, resulting in an overall shift of breeding range. This process can explain reverse latitudinal gradients in avian diversity in the temperate zone, since the breeding ranges of migratory species concentrate in latitudes where they enjoy the highest breeding success. Near absence of forest‐dwelling species among Palaearctic‐African migrants is attributable to the lack of forest in northern Africa for much of the Tertiary, which has precluded selection both for southward extension of migration by west Palaearctic forest species, and northward breeding colonisation by African forest species.