• Birds;
  • climate;
  • diversity;
  • island biogeography;
  • isolation;
  • richness–energy;
  • species–area;
  • species–energy;
  • species richness


Aim  The goal of our study was to build a global model of island biogeography explaining bird species richness that combines MacArthur and Wilson's area–isolation theory with the species–energy theory.

Location  Global.

Methods  We assembled a global data set of 346 marine islands representing all types of climate, topography and degree of isolation on our planet, ranging in size from 10 ha to 800,000 km2. We built a multiple regression model with the number of non-marine breeding bird species as the dependent variable.

Results  We found that about 85–90% of the global variance in insular bird species richness can be explained by simple, contemporary abiotic factors. On a global scale, the three major predictors — area, average annual temperature and the distance separating the islands from the nearest continent — all have constraining (i.e. triangular rather than linear) relationships with insular bird species richness. We found that the slope of the species–area curve depends on both average annual temperature and total annual precipitation, but not on isolation. Insular isolation depends not only on the distance of an island from the continent, but also on the presence or absence of other neighbouring islands. Range in elevation — a surrogate for diversity of habitats — showed a positive correlation with bird diversity in warmer regions of the world, while its effect was negative in colder regions. We also propose a global statistical model to quantify the isolation-reducing effect of neighbouring islands.

Main conclusions  The variation in avian richness among islands worldwide can be statistically explained by contemporary environmental variables. The equilibrium theory of island biogeography of MacArthur and Wilson and the species–energy theory are both only partly correct. Global variation in richness depends about equally upon area, climate (temperature and precipitation) and isolation. The slope of the species richness–area curve depends upon climate, but not on isolation, in contrast to MacArthur and Wilson's theory.