Analysis of survival probabilities of female common toads, Bufo bufo

Abstract

Recent advances in the statistical theory of mark-recapture analysis have markedly enhanced our ability to estimate survival probabilities in natural populations (Burnham et al., 1987; Pollock et al., 1990; Lebreton et al., 1992). Here, we report an analysis of a data set drawn from a published study of female common toads, Bufo bufo (Kuhn, 1994a). In this otherwise statistically well analysed paper, capture-mark-recapture histories of thousands of toads were not analysed statistically but inferences were drawn nonetheless. This resulted in a serious underestimation of survival probabilities. We wish to correct these survival estimates and advocate the use of robust statistics for drawing inferences. We will also show that such an analysis can reveal unexpected patterns when using drift fences around amphibian breeding sites.