Vagrancy fails to predict colonization of oceanic islands

Aim: Theories of island biogeography typically predict a close link between patterns of immigration and the dynamics of colonization and extinction on islands, but relatively little is known of the immigration process owing to the difficulty of observing immigration directly. We assess the relationship between propagule immigration and colonization success using records of out-of-range 'vagrant' animals compiled by amateur and professional biologists. Location: A global sample of 66 oceanic islands and archipelagos. Methods: We compiled a database of landbird records from each archipelago including occasional visitors (vagrants) and colonizers (defined as non-endemic species with breeding populations on islands). Using a database of species traits for all the world's passerine and near-passerine birds (migratory status and range size), we compared the characteristics of vagrants and colonizers and assessed how island characteristics influenced vagrant and colonizer communities. Results: We show that variation in the propensity for vagrancy to oceanic islands world-wide is a surprisingly poor predictor of colonization success across species on a global scale. Colonization success is positively related to global range size, and is higher for migrants than residents (particularly on higher-latitude islands). Species richness of vagrant communities increases with the area and local isolation of island groups, and decreases with isolation from continents and latitude. Colonizer species richness increases with island area and decreases with isolation and latitude, and is little influenced by vagrant species richness. Main conclusions: Contrary to expectations, we find that the capacity for trans-oceanic dispersal may not be a key determinant of oceanic island colonization amongst landbirds.