1. Natal dispersal distance (NDD; the distance between natal site and recruitment site) was studied in relation to proximal factors including sex, cohort, recruitment age, nesting density, nest site quality; and also within-colony movements of prospecting prebreeders among 133 recruits of two cohorts (1979–80) in a large, dense colony of western gulls Larus occidentalis (Audubon) on South-east Farallon Island, California. NDD was also studied in relation to ultimate factors, assessed from total number of chicks fledged to 1994, survival of breeding adults, and fitness-related life history traits, including brood size and hatching date. The Farallones colony is an unusual case in that it includes 50% of the species population and 90% of all western gulls breeding within a radius of 1000 km. Hence, potential recruits could choose between locations within the Farallones colony, but had limited choice to recruit at other colonies.
2. Among prebreeders, distance from natal site to prospecting site increased with age among gulls aged 2 to 4 years, and stabilized in 4- to 8-year-olds. Most 2-year-olds prospected on their natal sites.
3. Compared to other bird species, natal philopatry was strongly developed in both sexes, but was significantly greater in males (median distance between natal and recruitment site among males, 17 m; median distance among females, 50 m).
4. More philopatric males nested in areas of higher density and with a tendency for more nest cover (P = 0·080, when controlling for the effect of density). A non-linear relation between NDD and recruitment age among males resulted from shorter NDD in 4-, 5- and 6-year-olds compared to 3-year-olds and males older than 6 years. There was no relation between NDD and recruitment age in females.
5. Males who had hatched early in the season (individuals who were, on average, more dominant, with higher prebreeder survival and recruitment probability; Spear & Nur 1994) had shorter NDD than males who hatched later.
6. Fitness costs associated with natal philopatry were detected in both sexes. More philopatric males survived significantly less well than did less philopatric males. Average number of chicks fledged among experienced females (6–11 years breeding experience) was significantly lower among more philopatric compared to less philo patric individuals, as was cumulative number of chicks fledged for all females (as of 1994).
7. Thus, whilst philopatry was well developed in these gulls, the trait appeared maladaptive. This paradoxical result may be related to an extended period of poor food supply (1989–94) in the Gulf of the Farallones. We hypothesize that fitness costs associated with philopatry reflect different life history strategies where high philopatry may be part of a high effort strategy characterized by higher reproductive effort and lower survival. These results are consistent with the view that relative advantages and disadvantages between life history strategies could lead to selective equilibrium, depending on environmental conditions.