- 1The hypothesis of site-dependent population regulation predicts that birds utilize available nesting sites in a pre-emptive (ideal despotic) manner, leading to density dependence in heterogeneous habitats as poorer sites are used at higher population densities. At small population sizes adaptive site choice protects populations against fluctuations (the buffer effect).
- 2Common guillemots Uria aalge (Pontoppidan) breed at high density on sea-cliffs. The population breeding on the Isle of May, Scotland increased by 60% between 1981 and 2000. A good nest-site is a prerequisite for successful breeding and there is much competition for the best sites. Throughout this period, site use correlated with two measures of site quality, and photographs taken in 1936 show that this pattern has been extremely stable.
- 3The data indicate declining quality of sites that remain available as the population has increased. Site-dependent regulation was evident in that average breeding success declined over the years, but no declining trend was detected in the best (and most preferred) sites.
- 4An individual guillemot generally uses the same nest-site from year to year, but a minority move, usually less than 2 m, between breeding seasons. These movements can be involuntary or voluntary. Involuntarily moving birds that had occupied very good sites before moving often spent several years as non-breeders (floaters) close to their previous site before breeding again, and then occupied poorer sites. Voluntarily moving birds significantly improved their site quality by moving.
- 5Birds responded both to the physical site characteristics and to their own experience (breeding failure) when abandoning a site. Their behaviour thus combined aspects of the ‘win-stay, lose-switch’ strategy with direct assessment of site quality. Our data set is inconclusive with respect to the public information hypothesis, i.e. whether birds use conspecific cues of breeding success when settling in a new site.