Pink-footed geese Anser brachyrhynchus caught in western Denmark during March–April 1990–1992 were X-rayed to detect shotgun pellets. Among first-year and older geese, pellets were detected in 25% (N = 69) and 36% (N = 286), respectively. A simple theoretical model is proposed, relating frequencies of pellet carriers in different age classes, adult survival and the annual rate at which pellets are inflicted upon geese. This model resulted in estimates of pellets being inflicted upon 1,000 first-year and 800 older geese annually. However, annual bag is ca 1,000 first-year and 2,000 older geese, and thus, expectedly, pellets should be inflicted upon approximately two older geese for each first-year. From this it is concluded that the model leads to unrealistic estimates because the observed frequencies of first-year and older pellet carriers are inconsistent; if 25% of the geese carry pellets after their first hunting season, more than the observed 36% of the older geese should be carriers. One explanation of this could be statistical uncertainty of the observed frequency of first-year carriers. A frequency of 15% pellet carriers (lower 95% c.l.) after the first hunting season could account for the inconsistency. As an alternative explanation, lower survival of pellet carriers is proposed. A more general model including differential mortality predicts that survival rates of 87% and 78% for non-carriers and carriers, respectively, could explain the observed carrier frequencies. Use of lower 95% c.l. leads to estimates of pellets being inflicted upon a minimum of ca 0.5 goose per bagged one. If differential survival exists, this ratio will be somewhat higher.