EARLIER RECRUITMENT OR EARLIER DEATH? ON THE ASSUMPTION OF EQUAL SURVIVAL IN RECRUITMENT STUDIES

In species with deferred breeding, the population can be viewed as composed of phenotypes with different underlying age of first breeding defined at birth. The distribution of underlying age of first breeding within cohorts and the realized distribution of age of recruitment in individuals that survived from birth to first breeding (assessed in the breeding segment of the population) may differ. Realized age of recruitment is governed by the product of two demographic components: [local survival probability from birth to age i − 1] × [transition probability from the “prebreeder” state to the “breeder” state between age i − 1 and i]. The usual approach to gain insight into selective pressures shaping age of recruitment is to address covariation of realized age of first breeding with population size or social and environmental factors. Ideally, one should also conduct comparisons among groups of individuals encountering different conditions at a given age as prebreeders. However, in many species with deferred breeding, individuals are not encountered between birth and first breeding. Consequently, approaches to estimating recruitment probability in the absence of data from prebreeders have been developed. Some of these measures of recruitment ignore variation in the size of the pool of prebreeders among groups (as opposed to transition probability). Unless survival before full recruitment is known to be identical in the groups, inferences about the causes of variation in realized age of first breeding among groups are difficult: such differences may result from differences in survival probability before all the individuals make the transition between states.

We assessed the consequences of differences in prebreeding survival among groups on realized age-specific recruitment probability using numerical simulations. Depending on the specific scenario, realized age-specific breeding proportions were delayed or advanced compared to those expected according to the true underlying age-specific transition probabilities. We consider an example of this problem using data from a population of a seabird species, the kittiwake (Rissa tridactyla), where data from prebreeders and breeders were available. Consistent with results from our numerical simulations, we show that transition probabilities directly estimated (the underlying local recruitment probabilities) and those derived from analysis of data from the breeding segment of the population are different.