RESUMEN
The trade-off between current and future reproduction plays an important role in demographic analyses. This can be revealed by the relationship between the number of years without reproduction and reproductive investment within a reproductive year. However, estimating both the duration between two successive breeding season and reproductive effort is often limited by variable recapture or resighting effort. Moreover, a supplementary difficulty is raised when nonbreeder individuals are not present sampling breeding grounds, and are therefore unobservable. We used capture-recapture (CR) models to investigate intermittent breeding and reproductive effort to test a putative physiological trade-off in a long-lived species with intermittent breeding, the leatherback sea turtle. We used CR data collected on breeding females on Awa:la-Ya:lima:po beach (French Guiana, South America) from 1995 to 2002. By adding specific constraints in multistate (MS) CR models incorporating several nonobservable states, we modelled the breeding cycle in leatherbacks and then estimated the reproductive effort according to the number of years elapsed since the last nesting season. Using this MS CR framework, the mean survival rate was estimated to 0.91 and the average resighting probability to 0.58 (ranged from 0.30 to 0.99). The breeding cycle was found to be limited to 3 years. These results therefore suggested that animals whose observed breeding intervals are greater than 3 years were most likely animals that escaped detection during their previous nesting season(s). CR data collected in 2001 and 2002 allowed us to compare the individual reproductive effort between females that skipped one breeding season and females that skipped two breeding seasons. These inferences led us to conclude that a trade-off between current and future reproduction exists in leatherbacks nesting in French Guiana, likely linked to the resource provisioning required to invest in reproduction.