A key factor analysis of a sparrowhawk population.

During 1972-86, the Sparrowhawk breeding population in Eskdale, southern Scotland, remained fairly stable, with nest numbers fluctuating between 29 and 39 in different years. The annual production of young varied rather more, between 44 and 86 in different years. In a key factor analysis, the 'overwinter loss' operating in the period between the fledging of young and subsequent recruitment to the breeding population, emerged as the 'key factor', explaining 77% of the variance in total annual loss, and largely accounting for the pattern of change in breeding numbers. Overwinter loss in each new cohort was also the only form of loss which was density dependent. In a simulation model this loss was found to be capable, on its own, of stabilising the nesting population at the mean level observed. Losses during the breeding season were inversely correlated with losses over the ensuing winter. Thus, good breeding seasons, when many young were produced, were followed by high losses over winter, while poor breeding seasons were followed by reduced losses over winter. No density dependence was detected at any stage of breeding, or in the annual mortality of established breeders.