DOES REPRODUCTIVE SITE CHOICE IN A NEOTROPICAL FROG MIRROR VARIABLE RISKS FACING OFFSPRING?

Abstract

For organisms whose juvenile stages develop in discrete habitats, the strategy adults use to select patches directly impacts reproductive success. In most environments, the magnitude of juvenile mortality risks varies spatially and temporally among patches, often at several scales. Little is known, however, about how closely adults track such variation in site quality and how the predictability of risks facing juveniles influences the strategy adults adopt. I addressed these questions in a three-year study of Edalorhina perezi, an iteroparous frog that reproduces in forest pools in the southwestern Amazon Basin. Using observations and experiments, I quantified the spatial and temporal variation in direct and indirect mortality risks facing tadpoles at 18–34 pools. Data on larval risks were compared with observations of adult pool choice, including nest distribution and movement trends. Direct risks, pool drying and insect predation, accounted for 95% of estimated tadpole mortality. The indirect effect of tadpole density on mortality was small, slightly increasing tadpole drying risk. When pool risks were combined, persistent pools were predictably lower in risk early and late in the rainy season, while ephemeral pools were slightly advantageous midseason. At the population level, site choice shifted seasonally from persistent to ephemeral pools. Nonetheless, three lines of evidence suggest that adults were sensitive to local variation in larval risks. First, movement patterns were consistent with yearly changes in predation risk. Second, adults favored newly formed pools, which had lower predator densities and were likely to be more productive. Third, adults favored pools without recent reproductive activity. Whether fixed or flexible, the selective advantage for observed site choice was not evident during the last 50 d of the rainy season, when a majority of adults used ephemeral pools at a time when persistent pools appeared favorable for tadpoles. This study suggests that E. perezi responds to variation in juvenile risks at several scales when selecting reproductive sites. Yet adults appeared to be strikingly imperfect at tracking risk. Errors could arise due to cues that differ in their reliability or if adults shift from risk-sensitive to risk-prone strategies. Moreover, the behavior of conspecifics may constrain optimal site choice. Corresponding Editor: D. N. Reznick.