Adjusted phonotactic reactions to sound amplitude and pulse number mediate territoriality in the harlequin poison frog.

The information content of signals remains one of the central questions in animal communication. Auditory signals might contain information that allows receivers to estimate the distance as well as the size or the motivational state of senders. Proper differential reactions by receivers could be especially important for territorial species and lead to behavioural adjustments towards intruders according to the threat level: the perceived risk of losing territory. Therefore, territorial individuals should decode signal parameters that indicate the sender's size, distance, and motivation / treat level, reacting accordingly. To test this hypothesis, we studied the phonotactic reaction in territorial males of the poison frog Oophaga histrionica, after analyzing the variability of spectral and temporal parameters of natural calls. Next, we constructed synthetic calls by manipulating sound amplitude (sound pressure level: SPL), gross (inter-call intervals) and fine-temporal (number of pulses) structure of synthetic signals. In this way, we simulated near and far intruders with potentially variable sizes, motivational states or threat levels. Then, we conducted playback experiments using these synthetic calls to determine 1) how perceived proximity of vocal competitors (determined by SPL) affect the behavior of receivers, and 2) how variation in the temporal structure of calls (inter-call interval and call pulse number) impact the behavior of receivers. We also asked whether signals convey body-size related information by analyzing the relationship between call parameters and body size of the males receiving the experimental stimuli. Consistent with our hypothesis, males recognized the variations in the SPL of calls. Males attacking the simulated opponent increased the number of pulses per call, while those retreating, kept pulse number unchanged (a graded aggressive signal). On the other hand, call traits resulted poor predictors of body size, and neither SPL nor temporal call traits predict whether a resident will attack an opponent. Our data demonstrate that males discriminate and use SPL to estimate the sender's distance and use the call's fine temporal traits and to adjust the competitive/aggressive reaction using bimodal signals. They further suggest that body size assessment is not always an important factor in vocally- mediated agonistic interactions.