Summary
The thesis contains 13 research papers, of which 10 published and 3 in various stages of submission/acceptance. Papers in chapters 2-7 deal with results from ship-based research on top predator communities or individual bird species in the pelagic environment. Chapters 8-14 are on land-based studies of the fulmarine petrels. All chapters have separate summaries, so here I can focus on a few main conclusions. Also, repeating what has been achieved in the past is only useful in terms of what it means for the future. Counting seabirds and marine mammals at sea is highly impopular with research funding agencies, as it is hardly considered to be a ‘scienctific’ activity. I most seriously disagree, and hope to have contributed to a little change in opinion in the first section of this thesis. Predator counts, conducted by proper methods and in combination with other disciplines of research, have high research value. Spatial and temporal patterns in density-distributions of top predators are a crucial tool to unravel the ecology and differentiation of the individual species, and the role of predator-communities in the ecosystem. Furthermore, top predators are also the signalling posts drawing attention to situations where our perception of ecosystem functioning may be incomplete or even wrong. Few biological components of marine ecosystems can be quantified as easily and in as much detail as top predators. By inserting observed top predator values in existing ecosystem models, it is possible to check such models and to possibly indicate the directions in which the models could be incomplete. This is what I have attempted to do in chapter 5. First main conclusion is that existing food web models for the Southern Ocean ecosystem underestimate complexity of trophic interactions. At the top predator end of the food web, there is reason to doubt the key role of krill in channeling primary production to top predators. Secondly, results strongly indicate that our