Natural Selection of Colony Size in a Passerine Bird

1. The breeding success of fieldfares Turdus pilaris was studied in 56 colonies and 16 solitary pairs in a mountain birch area of N Sweden, and in 40 colonies and 70 solitary pairs in S Sweden. The analysis is based on a statistical method that treats colonies, not nests, as independent sample units. 2. The probability of adult survival was the major difference between the areas. Adult breeding birds ran a much higher risk of predation in S than in N Sweden; the risk increased with colony size. 3. Predation on eggs and nestlings decreased with increasing colony size in both areas. 4. The total risk of offspring loss owing to predation (of either parents or offspring) decreased with increasing colony size in the north, but not in the south. 5. Among chicks in surviving nests, the risk of death from starvation increased with colony size in both areas. 6. Mobbing frequency, examined by presenting a nest predator 10 m from the nearest nest, was higher in colonies than at solitary nests. 7. The results show that nest predation selects for larger colonies, where mobbing of predators is more effective, whereas chick starvation selects for smaller colony size. Protection against nest predation, rather than increased foraging efficiency, therefore seems to be the main advantage of colonial nesting in fieldfares. 8. In S Sweden, but not in the north, predation on adults (apparently mainly by tawny owls Strix aluco, absent in the northern study area) selected for smaller colony size. In accordance, mean colony size was less than half as large as in the northern area. 9. Although much local variation remains unexplained, colony size in the fieldfare is in part adapted to predation risks, which vary among areas.