Abstract
Call surveys are an effective technique for detecting the presence and activity of breeding male frogs. Such surveys have been used to quantify breeding activity at a site under the assumption that male chorusing activity appropriately reflects breeding consequences, such as the number of oviposition events. However, only a few studies have actually examined the relationship between chorusing activity and breeding consequences in the field. In this study, I examined the relationship between chorusing activity (the number of male calls recorded during a five-minute period every night) and the number of oviposition events (number of oviposited egg masses during the night) of the Otton frog, Babina subaspera, with regard to the time lag between calls and oviposition. I constructed nine generalized linear mixed models (GLMMs) to explain the number of oviposition events by chorusing activity on the same night and on nights 1 to 7 days before the oviposition events. The Akaike information criterion (AiC) of the GLMM was lowest when the number of calls from nights 2 days before the oviposition events was used, indicating that breeding consequences in Otton frogs reflect the chorusing activity of 2 days prior. This study shows that frog call surveys can be reliable tools with which to represent breeding activity at a site as long as the time lag between chorusing activity and breeding consequences is considered.
S-Estimation and a Robust Conditional Akaike Information Criterion for Linear Mixed Models