Diet composition of starry smooth-hound Mustelus asterias and methodological considerations for assessing the trophic level of predatory fish.

The stomach contents of 640 starry smooth-hound Mustelus asterias from the north-east Atlantic were examined. The diet was dominated by crustaceans (98.8% percentage of index of relative importance, %IRI), with the two main prey species being hermit crab Pagurus bernhardus (34% IRI) and flying crab Liocarcinus holsatus (15% IRI). Ontogenetic dietary preferences showed that smaller individuals [20-69 cm total length (LT ) n = 283] had a significantly lower diversity of prey than larger individuals (70-124 cm LT , n = 348); however, 18 prey species were found exclusively in smaller individuals and eight prey taxa were found exclusively in larger individuals. Larger commercially important brachyurans such as edible crab Cancer pagurus and velvet swimming crab Necora puber were more prevalent in the diet of larger individuals. Specimens from the North Sea ecoregion had a lower diversity of prey types for a given sample size than fish from the Celtic Seas ecoregion. Whilst cumulative prey curves did not reach an asymptote, this was primarily due to the high taxonomic resolution utilized and 95% of the diet was described by just seven crustacean taxa. The trophic level (TL) was calculated as 4.34 when species-level prey categories were used. This fine-scale taxonomic resolution resulted in a TL estimate close to a whole level above that estimated using wider taxonomic groupings. This large bias has important methodological implications for TL studies based on categorized prey data, particularly those of predatory fish.