Abstract
A K-sample tag-recapture model for an open population is derived which allows for animals with different histories of capture to have both different probabilities of survival and of future capture. Previous stochastic models are special cases of this one. Maximum likelihood estimators of the population sizes at each sampling time are derived for models allowing differing degrees of unequal survival and catchability. Methods of testing between these models are derived by conditioning on sufficient statistics.

This publication has 0 references indexed in Scilit: