Prey selection by sparrowhawks,Accipiter nisus: relative predation risk for breeding passerine birds in relation to their size, ecology and behaviour

Predation is a fundamental ecological process, but there are few studies of predation risk for adult, breeding birds. In this study, we quantified Relative Predation Risk (RPR) for 46 species of passerine birds preyed on by sparrowhawks during the breeding season in south-western Sweden. The sparrowhawk is a major predator of smaller birds in Eurasian forests and woodland. Near nesting hawk pairs, prey abundance was estimated by line transects. From prey remains at hawk nests, we estimated the frequency of prey species in the diet of hawks. For each hawk pair, RPR values for different prey species were calculated. The mean RPR for each prey species, based on hawk pairs as sample units, indicated degree of overrepresentation (positive RPR) or underrepresentation (negative RPR) of the prey in the diet. Prey species with a high positive RPR nested mainly near human habitations (villages or towns), at forest edges and in farmland. However, there were marked seasonal changes in RPR. Forest birds, such as the goldcrest and wren, had the highest RPR in April. Later, the hawks seemed to hunt mainly in the habitats mentioned above, where prey were more abundant or easier to catch, resulting in increasing RPR for the house sparrow, yellowhammer, greenfinch and other species. RPR values were analysed in relation to prey body size (mass) and relative density, as well as prey foraging height and nest height (categorized from the literature). In addition, the mean perch height and mean exposure of prey species were quantified in the field. Most predation was due to the smaller male hawks, providing food for their mates and young, RPR increased with prey body size up to a mass of about 40 g, then declined with increasing body size (larger prey being more difficult to catch), RPR decreased with increasing relative density of prey species. Nest height was not correlated with RPR, but foraging height seems to be an important factor: RPR decreased with increasing foraging height. A similar result was found for mean perch height, when larger prey (over 40 g) and singing individuals were excluded from the field data. RPR increased with mean exposure of prey species when larger prey and singing birds were excluded; a combination of perch height and exposure improved the correlation with RPR. Singing birds were generally perched high in the vegetation and may not suffer high predation risk, contrary to common belief. The results of this study are discussed in relation to temporal and geographical variation in RPR and antipredator adaptations in passerine birds.