Résumé
•We propose a Bayesian hierarchical model for comparative demography studies.•The model accounts for phylogenetic dependence and correct for imperfect detection.•We illustrate the method using the survival–body mass relationship of seabird species.•Our results show that ignoring phylogeny may lead to spurious conclusions.
Species in comparative demography studies often have a common phylogenetic or evolutionary ancestry and hence, they cannot fully be treated as independent samples in the statistical analysis. Although the serious implication of ignoring phylogeny has long been recognized, no attempt has been made so far to account for the lack of statistical independence due to phylogeny in multi-species mark–recapture comparative demography studies. In this paper, we propose a Bayesian hierarchical model that explicitly accounts for phylogenetic dependence among species, and to correct for imperfect detection, which is a common phenomenon in free-ranging species. We illustrate the method using individual mark–recapture data collected from 16 seabird species of the order Procellariiformes. Data on body mass and phylogeny of these species are compiled from literature. We investigate the relationship between adult survival and body mass with and without accounting for phylogeny. If we ignore phylogeny, we obtain a positive survival–body mass relationship. However, this relationship is no longer statistically significant once phylogenetic dependence is taken into account, implying that survival may actually depend on an unmeasured variable that is correlated with body mass due to a shared dependence on phylogeny. The proposed model allows the integration of multi-species mark–recapture data and phylogenetic information, and it is therefore a valuable tool in ecological and evolutionary biology.