Density-dependent mortality in wintering shorebirds is critical to shorebird population dynamics, but is difficult to study and demonstrate. Most winter mortality of small shorebirds probably results from predation by raptors. Over 13 winters, I studied Eurasian Sparrowhawk Accipiter nisus predation on wintering Dunlins Calidris alpina at a rocky shore in south-east Scotland, a site apparently of marginal suitability for wintering Dunlins. Most Dunlins on the study site were juveniles (first winter birds), and winter juvenile mortality rate due to Sparrowhawk predation was strongly and positively related to Dunlin density (maximum winter count). There was no indication of any marked effect of weather on Dunlin winter mortality rate, except that mortality was slightly greater in winters with lower temperatures. Dunlin mortality was unrelated to the mortality through Sparrowhawk predation of a preferred prey, the Redshank Tringa totanus, or to sighting rates of hawks, suggesting that increased hunting of Redshanks by hawks or a numerical response by hawks were not involved in density-dependent mortality of Dunlins.