To eat and not be eaten: modelling resources and safety in multi-species animal groups.

Using mixed-species bird flocks as an example, we model the payoffs for two types of species from participating in multi-species animal groups. Salliers feed on mobile prey, are good sentinels and do not affect prey capture rates of gleaners; gleaners feed on prey on substrates and can enhance the prey capture rate of salliers by flushing prey, but are poor sentinels. These functional types are known from various animal taxa that form multi-species associations. We model costs and benefits of joining groups for a wide range of group compositions under varying abundances of two types of prey-prey on substrates and mobile prey. Our model predicts that gleaners and salliers show a conflict of interest in multi-species groups, because gleaners benefit from increasing numbers of salliers in the group, whereas salliers benefit from increasing gleaner numbers. The model also predicts that the limits to size and variability in composition of multi-species groups are driven by the relative abundance of different types of prey, independent of predation pressure. Our model emphasises resources as a primary driver of temporal and spatial group dynamics, rather than reproductive activity or predation per se, which have hitherto been thought to explain patterns of multi-species group formation and cohesion. The qualitative predictions of the model are supported by empirical patterns from both terrestrial and marine multi-species groups, suggesting that similar mechanisms might underlie group dynamics in a range of taxa. The model also makes novel predictions about group dynamics that can be tested using variation across space and time.