Allometric scaling and Bayesian priors for annual survival of birds and mammals.

Allometric theory predicts that instantaneous mortality rates scale with body mass with a negative quarter power. Such a relationship would mean that the survival rate of one species is partly predictable from the survival rate of other species. We develop allometric regression models for annual adult survival of birds and mammals, using data collected from the literature. These models conform to the predictions of the allometric theory; the value of negative one-quarter for the scaling parameter is within the 95% credible interval, which is [-0.31, -0.10] for birds and [-0.35, -0.15] for mammals. The predictions are very well supported when evaluated using an independent set of data. The regression models can be used to provide objective and informative Bayesian priors for annual adult survival rates of birds and mammals or to act as a point of comparison in new studies.