Abstract

To test between the contradictory predictions of resource defence theory and a modified hawk-dove model, I examined the effect of patch size and competitor number on aggression among foraging house sparrows, Passer domesticus . The birds were observed at feeding tables of seven different sizes, each table doubling in surface area from 0.014 m 2 to 0.922 m 2 , allowing for a broader quantitative description of changes in aggression with patch size than previously examined. Resource defence theory predicts that small patches will be more defendable than large patches and that aggression will peak at intermediate competitor numbers at defendable patches. In contrast, a modified hawk-dove model predicts an increase in aggressiveness with competitor number; the model makes no predictions about patch size. Contrary to resource defence theory, birds did not defend an entire patch, even when it was as small as 0.014 m 2 . As a result, aggression among the birds decreased more gradually with increasing patch size than expected by an all-or-none threshold model of resource defence. Moreover, birds fought more frequently and intensely at high competitor numbers, a result that is more consistent with the predictions of a modified hawk-dove model than the dome-shaped relationship predicted by models of resource defence. Interestingly, females were more aggressive and foraged faster than males. Because females were able to compete more effectively when individuals were aggressive, the proportion of females increased as patch size decreased. My results have implications for the truncated phenotype ideal free distributions.