Fledging size and survival in snow geese: timing is everything (or is it?)

In many birds, body size at ̄ edging is assumed to predict accurately the probability of subsequent sur vival, and size at ̄ edging is often used as a proxy variable in analyses attempting to assess the pattern of natural selection on body size. However, in some species, size at ̄ edging can vary signi ® cantly as a function of variation in the environmental component of growth. Such developmental plasticity has been demonstrated in several species of Arctic-breeding geese. In many cases, slower growth and reduced size at ̄ edging has been suggested as the most parsimonious explanation for reduced post̄ edging survival in goslings reared under poor environmental conditions. However, simply quantifying a relationship between mean size at ̄ edging and mean survival rate (Francis et al., 1992) may obscure the pattern of selection on the interaction of the genetic and environmental components of growth. The hypothesis that selection operates on the environmental component of body size at ̄ edging, rather than the genetic component of size per se, was tested using data from the long-term study of Lesser Snow Geese (Anser c. caerulescens) breeding at La Pe rouse B ay, Manitoba, Canada. Using data from female goslings measured at ̄ edging, post̄ edging sur vival rates were estimated using combined live encounter and dead recovery data (B urnham, 1993). To control for the covariation between growth and environmental factors, survival rates were constrained to be functions of individual covariation of size at ̄ edging, and various measures of the timing of hatch; in all Arctic-breeding geese studied to date, late hatching goslings grow signi® cantly more slowly than do early hatching goslings. The slower g rowth of latehatching goslings has been demonstrated to re ̄ ect systematic changes in the environmental component of growth, and thus controlling for hatch date controls for a signi® cant proportion of variation in the environmental component of growth. The relationship between size at ̄ edging, hatch date and survival was found to be signi® cantly non-linear; among early hatching goslings, there was little indication of signi® cant diþ erences in survival rate among large and small goslings. However, with increasingly later hatch Correspondence: E. G. Cooch, Department of Natural Resources, Cornell University, Ithaca, New