Abstract
Although much of life-history theory assumes increased mortality at certain stages (e.g., migration), survival rates are rarely estimated on a seasonal basis within the annual cycle of migratory species. We estimated variations in seasonal survival rates in a long-lived, hunted species in the presence of short-term (between consecutive seasons) and long-term (between years in the same season) dependence in sighting probabilities. We also tested the two contrasting hypotheses that hunting mortality is either compensatory or additive to natural mortality. This study was conducted on adult female Greater Snow Geese (Chen caerulescens atlantica) from 1990 to 1998, and is based on 3890 neck-banded birds and 13 657 resightings on the northern breeding grounds in summer, and southern staging areas in spring and autumn. Birds were 10-20% more likely to be seen in autumn and spring if they were seen on the previous occasion (summer and autumn, respectively). Birds were 30-40% more likely to be seen in autumn and spring if they were last seen in the same season in the previous year. Differences in behavior according to family status (pres-ence or absence of young) and heterogeneity in site fidelity may explain the dependence in sighting probabilities; failure to account for this may lead to biased survival estimates. Monthly survival rates from spring to summer (3-mo period) and summer to autumn (2.5-mo period) were equal (0.989 0.003, mean 1 SE) and showed little variation over the years, even though the two lengthy migratory flights (3000 km each) and breeding occurred during these periods. In contrast, monthly winter survival (from autumn to spring, 6.5 mo) was lower than during the other seasons and varied significantly over the years (range 0.936 0.021 to 0.993 0.008). Mean survival of adult females (corrected for neck band loss) was 0.96 from spring to autumn, 0.86 during winter, and 0.83 0.05 for the whole year. Natural mortality (i.e., excluding hunting) was equal among seasons and did not vary over the years, which suggests that mortality risk is not increased during migration or reproduction. There was a significant inverse relationship between winter survival and the kill rate (b 1.21 0.56), which suggests that hunting mortality was additive to natural mortality. This is probably a general feature of long-lived species because their low and relatively constant natural mortality rate does not allow them to compensate for an additional source of mortality such as hunting. Contrary to life-history theory, we did not find evidence that migration or reproduction entailed a survival cost in this long-distance migrant bird.