DEATH AND DISAPPEARANCE - ESTIMATING MORTALITY RISKS ASSOCIATED WITH PHILOPATRY AND DISPERSAL

Models dealing with the relationship of inbreeding to dispersal, the e volution of cooperative breeding, and the adaptive significance of rep roductive suppression all involve a parameter describing the probabili ty that emigration ends in death. Surprisingly, estimates of this para meter (or equivalently, of the probability of dispersing successfully) are virtually absent from the literature. There are formidable diffic ulties in estimating dispersal risk by direct observation of emigrants , but we contend that many investigators have already collected demogr aphic information that can be used to estimate it indirectly. The root problem is that the proportion of the population that emigrates from its natal group or home range, e, is not generally observable. Neverth eless, for each value of e there is only one possible survival rate of emigrants (s(e)) and of philopatric animals (s(p)) that can explain t he overall survival rate. For study sites that are neither sources nor sinks, we describe a graphical framework within which census data and behavioral observations can be used to place bounds on e, s(e), and s (p). We use data from a Serengeti population of the dwarf mongoose (He logale parvula) to illustrate how different types of data and assumpti ons allow us to construct better estimates of these parameters. For dw arf mongooses, the approach reveals unexpected differences in dispersa l risk among sex and age classes.