Résumé
In monomorphic species, determination of sex from behavior is prone to errors. Ignoring this fact renders problematic the estimation of sex-specific parameters. We develop capture-recapture survival models that account for uncertainty in the assessment of sex. There is a legitimate concern that some parameters may not be identifiable in these models. We examine parameter redundancy for four basic models with survival and encounter probabilities constant or time-dependent. We further develop a more refined and more appropriate model for an Audouin's gull data set where four distinct behavioral clues have been used. We examine how useful it is to incorporate the least reliable of the clues and the genetic determination of sex available for only a handful of individuals. We finally discuss the implications of our findings for the design of field studies as well as further directions of research.