Strength of the 'island rule' in birds is positively associated with absence of avian predators

Abstract

The similar characteristics shared by island environments have been shown to lead to common patterns of adaptations in island species, commonly referred to as the 'insularity syndrome'. A well-known example is the 'island rule', where large species become smaller on islands and small species become larger, leading to well-known cases of dwarfism and gigantism. This pattern was recently verified on a global scale, but the mechanisms underlying it have been poorly investigated. Here, we focused on the role of released pressure from predation and competition experienced by island birds. Using 120 pairs of endemic island species and their mainland sister relatives, we first verified that the island rule was detected in our dataset, and then evaluated the effects of the numbers of raptors and interspecific competitors on the evolution of the insular species' body mass. We found a strong effect of predation on body mass evolution, with a stronger island rule for species occurring on islands with no raptors, while the pattern disappears in their presence. However, we did not find an effect of competition on this pattern. Our study shows the importance of considering ecological interactions for understanding patterns of body size evolution, and the exceptions to those patterns.

Keywords: body mass; competitors; ecological release; insularity; raptors.

Figure 1.

(a) Presence/absence of raptors changes the relationship between the body mass ratio (lnRR) and the body mass of the mainland species while (b) presence/absence of competitors belonging to the same genus does not, and (c) presence/absence competitors belonging to the same family on islands has a marginal effect on this relationship. Points and lines in blue indicate absence and in violet presence of predators or competitors. The intercepts correspond to the binary models performed each with one variable (predation, competitors of the same genus or family). See models in electronic supplementary material, table S1 in appendix S4.

Figure 2.

Fixed estimates of PGLS models using non-binary (left) and binary (right) variables for accounting for predation and competition. See electronic supplementary material, table S2 in appendix S4.