A mark-recapture survival analysis.

Long-term marking studies are often used to investigate survival rates in animal populations. In many studies, a few individuals are caught regularly, whereas others may only rarely be caught. Existing mark-recapture and tag return models often make unrealistic assumptions and are not robust when these fail. A procedure is developed that uses both mark-recapture and tag return data to estimate the survival curve. The effect of differential trappability is assessed. A method that is deliberately subjective is proposed for discarding seriously biased estimates, so that estimation is not greatly influenced by the various departures from equal trappability that may occur. -Author