We used stable-isotope analysis (SIA) to evaluate trophic relationships in an Antarctic seabird community. We determined natural abundances of stable-nitrogen (delta(15)N) and stable-carbon (delta(13)C) isotopes from blood samples (n = 283) from adults and chicks of 4 Antarctic fulmarine petrel species (Fulmarus glacialoides, Thalassoica antarctica, Daption capense and Pagodroma nivea) during 2 consecutive breeding seasons, 1994/1995 and 1995/1996, and from representative prey items. Our objectives were to use the isotope approach to infer trophic status and diet composition within and between species, addressing interspecific and temporal variability within this seabird community, and to investigate potential age-related differences in assumed trophic position within species. Prey delta(13)C values ranged from -26.8 parts per thousand in amphipods to -23.9 parts per thousand in adult Antarctic silverfish. Seabird delta(13)C values ranged from -25.3 parts per thousand in Antarctic petrel chicks to -23.8 parts per thousand. in cape petrel adults. Prey delta(15)N values ranged from 4.0 parts per thousand in euphausiids to 10.7 parts per thousand in adult Antarctic silverfish. Seabird delta(15)N values ranged from 8.4 parts per thousand in Antarctic petrel adults to 12.0 parts per thousand in snow petrel chicks. There was considerable interspecific overlap in assumed trophic positions amongst the 4 petrel species, and we conclude all species consumed fish and krill. Despite this apparent overlap, the range in delta(15)N values for petrels corresponded to the equivalent of 1 full trophic level, and estimated trophic level varied with both species and age. A simple trophic level model, constructed based on the delta(15)N data, predicted trophic levels ranging from 2.3 in krill to 4.7 in snow petrel chicks. Snow petrels and Antarctic fulmars tended to have higher delta(15)N values than Antarctic and cape petrels, suggesting a higher proportion of fish in their diets. Petrel chicks consistently had higher delta(15)N values than adults, which suggests trophic segregation between adults and chicks. We discuss advantages of selectively provisioning chicks with higher trophic level prey. Extensive overlap and a relatively narrow range of delta(15)N values are consistent with a food web comprised of few trophic steps.