Item

Abstract

1. Ideal free models for unequal competitors predict a truncated competitor phenotype distribution among patches when relative payoffs of phenotypes vary across patches. Partial truncation is expected under field conditions. This prediction was tested in a field experiment with an overwintering population of mute swans (Cygnus olor). 2. Two food patches were generated in which adult and subadult swans were expected to have different relative success. In one patch ('clumped'), pieces of bread were thrown over a small area on the water surface. In the other patch, pieces of bread were scattered over a larger area. 3. When only one patch was offered at a time (no choice situation), adult swans were more successful than subadult swans in the 'clumped' patch, but were similarly successful in the 'scattered' patch. Relative payoffs of adult and subadult swans differed significantly between patches. 4. When the two patches were offered simultaneously, black-headed gulls (Larns ridibundus) competed with the swans to a considerable extent in some replicates. The gulls appeared to be the poorest competitors. They snatched more bread in the scattered than in the clumped patch. Both classes of swans avoided the scattered patch but not the clumped patch with increasing competition from gulls. Gulls preferred the scattered patch and swans increasingly preferred the clumped patch under gull competition, creating a partially truncated distribution. 5. Without the four replicates in which the gulls had consumed more than 25% of the bread, the percentage of adult swans that chose the clumped patch was significantly higher than the percentage of subadult swans that chose that patch. This is the first experimental verification of a partially truncated phenotype distribution. Subadult swans significantly preferred the scattered patch whereas adults tended to prefer the clumped patch. This distribution was predicted from the 'no choice' experiment, where competition by gulls had been similarly weak.