Abstract: Feeding ecology and isotopic niche of the wandering albatross Diomedea exulans were investigated in the poorly studied population on the Kerguelen Islands and compared to that on the Crozet Islands. Fish (48% by mass) and cephalopods (46%) were similarly important in chick food at Kerguelen, while cephalopods (87%) dominated the diet at Crozet. Fish prey included mainly deep-sea species, with the Patagonian toothfish Dissostichus eleginoides being the main item. Cephalopod beaks were identified, most of which were from adult oceanic squids. Albatrosses preyed upon the same taxa at both localities, but in different proportions. Histioteuthis atlantica (30% by number), Galiteuthis glacialis (13%), and Kondakovia longimana (10%) were the main squid prey at Kerguelen, while K. longimana (35%) and H. eltaninae (23%) dominated at Crozet. Chick feather ?15N values were higher in wandering albatrosses than in other oceanic seabirds of the 2 communities, indicating that the wandering albatross is an apex consumer together with the sperm whale and sleeper shark that have similar ?15N values. Satellite-tracked wandering albatrosses foraged in local subantarctic waters and farther north, with some Crozet birds overlapping with those from the Kerguelen population in western Kerguelen waters. Anthropogenic items (e.g. plastic fragments, hooks) were found in half the food samples. All fishery-related items were from the local toothfish fishery. The high number of hooks from Crozet indicated the presence of a fairly large number of illegal longliners in the area during the Austral winter 1998. A review of the feeding habits of Diomedea spp. highlights the need for more dietary investigations to achieve effective conservation and management of this endangered group of charismatic seabirds.

Abstract: In the marine environment, track and dive parameter data (obtained using Argos or GPS tags and time?depth recorders) are commonly used to provide proxies for foraging behaviour of marine predators. However, their accuracy is rarely assessed. Recently, the addition of head-mounted accelerometers has allowed for detection of prey capture attempts (PCAs) at sea, allowing for more accurate estimations of foraging behaviour. Despite increased numbers of such devices being deployed, their use is still marginal compared with other tools which measure track and dive parameters. The objectives of our study were (1) to identify the most relevant combination of tracking and diving metrics in predicting the frequency of PCAs in female southern elephant seals Mirounga leonina from the Kerguelen Islands, and (2) to apply it to a broader range of individuals for which only tracking and diving data were available. The results of our models were consistent with the optimal foraging theory as well as the optimal diving theory. The model with the best predictive performance was the one that combined both tracking and diving information. However, most of the variability in the number of PCAs could be solely explained by changes in the diving behaviour of seals. Finally, we used the best predictive model on 20 individuals, which had not been fitted with accelerometers, to determine their main foraging zones. The behavioural indicators established in this study constitute a useful ecological tool for population monitoring and conservation purposes.