Predicting reproductive success in captive Puerto Rican parrots (Amazona vittata)

Abstract

The productivity of captive breeding programs can limit efforts to restore endangered species. The ability to accurately predict which captive pairs will breed successfully improves breeding productivity by allowing managers to focus their efforts on pairs that are likely to succeed, and to re-assort pairs with a low probability of success. This could facilitate population restoration efforts that depend on captive programs. I tested whether pair duration and affiliative behaviors predict reproductive success in a captive population of the endangered Puerto Rican parrot (Amazona vittata). I systematically observed 18 pairs in the months prior to breeding, and tracked their reproductive success through the following breeding season. I then modeled various aspects of reproductive success as a function of pair duration, allopreening, vocal duetting and allofeeding, and used an informational approach to select the best models. The number of eggs laid could not be predicted with the measured variables. The best model for the number of chicks hatched included the independent variables pair duration, allopreening frequency and allofeeding frequency. The most important response variable, number of chicks fledged, was best predicted by pair duration and allopreening frequency. I conclude that both pair duration and allopreening behavior are useful for predicting reproductive success in pairs of captive Puerto Rican parrots. I recommend that managers monitor potential breeding partners for allopreening behaviors and manage populations to maximize average pair duration.