Species pool structure determines the level of generalism of island parasitoid faunas

Aim To examine whether island parasitoid faunas are biased towards generalists when compared with the mainland and their species pool, and to evaluate the effects of climate, island characteristics and regional factors on the relative proportions of idiobionts (i.e. generalists) and koinobionts (i.e. specialists) of two parasitic wasp families, Braconidae and Ichneumonidae. Location Seventy‐three archipelagos distributed world‐wide. Methods We used data on the distribution and biology obtained from a digital catalogue and several literature sources. We related level of generalism, measured as the ratio between the number of idiobiont and koinobiont species, to climatic, physiographic and regional factors using generalized linear models. We compared models by means of Akaike weighting, and evaluated the spatial structure of their residuals. We used partial regressions to determine whether the final models account for all latitudinal structure in the level of generalism. Results Islands host comparatively more idiobionts than continental areas. Although there is a latitudinal gradient in the level of generalism of island faunas correlating with both environmental factors and island characteristics, the most important determinant of island community structure is their source pool. This effect is stronger for ichneumonids, where generalism is higher in the Indomalayan region, arguably due to the higher diversity of endophytic hosts in its large rain forests. Main conclusions The level of generalism of island parasitoid faunas is largely constrained by regional factors, namely by the structure of the species pool, which emphasizes the importance of including regional processes in our understanding of the functioning of ecological communities. The fact that generalist species are more predominant in islands with a large cover of rain forests pinpoints the importance of the indirect effects of ecological requirements on community structure, highlighting the complex nature of geographical gradients of diversity.