Interference competition, the spatial distribution of food and free-living foragers

Studies of interference competition among foraging animals generally assume that variation in the spatial distribution of food can be neglected. This assumption may be problematic as resource defence experiments suggest that such variation is of the essence in some interference mechanisms. Interpretation of the results of field experiments on this topic, however, is hard because most studies used univariate statistics to analyse multivariate data. Because in free-living foragers interference and patch selection are connected behaviours, treatment effects on these responses are best studied simultaneously, through multivariate analyses. We performed a field experiment in which we provided wild ruddy turnstones, Arenaria interpres, with experimental plots that varied in the distance between a fixed number of so-called food pits, and, using multivariate statistics, we studied effects on the combination of the turnstones' behaviour and abundance. We found that when food pits were more spaced out, turnstones were present in higher numbers, while interacting less with each other. Nevertheless, turnstones spent about the same amount of time digging for food, our measure of intake rate, at each interpit distance. These findings imply that to reliably predict the combination of the number, intake rate and aggression of turnstones, the spatial distribution of food has to be known. We would not have reached this conclusion if we had used univariate statistics. In addition, we argue that multivariate statistics helps to clarify the way field experiments on the spatial distribution of food are to be interpreted.