Evolutionary correlates of bioacoustic diversity in the Amazonian poison-dart frog Epipedobates (Allobates) femoralis

The proposed project aims at testing a general model an the potential causes for acoustic differentation between populations of a pan-Amazonian frog species. The cotnmunication System of the poison-dart trog Epipedobates femoralis (Dendrobatidae) is an excellent model for studying the relative contributions of acoustic environment, geographic distances and barriers between populations, distances to postpleistocenie dispersion centers, as well as morphological and genetical traits to geographic variation in male anuran`s vocal and spacing behavior. Advertisement calls of this diamal frog species can easily be recorded and analyzed, and are alread~ known to van geographically. Stereotyped behavioral tresholds and responses to playbacks of synthetic Balls permit quantitative sampling of both emitter`s and receiver`s behavior. The wide distribution range of the species includes sites strongly differing in acoustic enviromnent, and allows the choice of study sites corresponding to the predictions of several competing hypotheses an the origin of high biodiversity in Amazonia. The proposed project concentrates an sampling geographical variation in call characteristics of male E. femoralis and male`s reactivity to synthetic playbacks, and an ascertaimnent of acoustic environment for eight studv sites. Study sites are distributed thrortghout the Amazonian basin including the Andean foothills. Data an concomitant differences in morphology and genetic structure of selected populations will be collected. We intend to genotype 20-30 individuals per study site with RFLP (Restriction Fragment Length Polymorphism) analysis of a 308 bp fragment of the mtDNA cytochrome b locus. Further genetic information will be obtained by sequencing mitochondrial 12S and 16S rRNA genes of 1-2 specimens per study site. We plan to analyze distances between study sites and their relation to geographic barriers and dispersion centers. Data will enter a correlational matrix and will be tested with Mantel tests. Results should elucidate the significance of the tested factors in the evolution of this species` geographic differentiation and should contribute to the controversy an the origin of high differentation rates in Amazonian biota.

With the project FWF 15345 we considerably improved our understanding of a male-male communication system in an anuran species and the knowledge of evolutionary adaptations to acoustic environments in general. As a main result of the project we found a geographic variation in the receiver but not the sender component of the acoustic communication system of the poison-arrow frog Allobates femoralis due to masking interference in frequency by a second frog species, Epipedobates trivittatus. This constitutes the first correlational evidence of the commonly invoked adaptation for anuran communication in acoustic assemblages. Detected call differences of A. femoralis in eight populations in the Amazon basin and its correlation with genetic divergence and geographic distance indicate a differentiation of mating traits that follows mainly a pattern of isolation by distance. Additionally, geographic variation in the advertisement call is not or just weakly influenced by the isolating effect of large rivers for this ground-dwelling frog. We learned that the prominent upward sweep and the fixed temporal pattern of the advertisement call are irrelevant for species recognition. In playback experiments males reacted also to pure tones with natural temporal pattern within the species-specific frequency range as well as to an inverse sweep signal. Conspecific calls were perceived with the basilar papilla of the inner ear. Under field condition the probability of response to conspecific playback- signals was sensitive to the number of broadcasted notes per call and furthermore matched the mean spectral and temporal properties of the population. Finally, for the first time for anuran amphibians we could artificially elicit aggressive behaviour by using an electromechanical model mimicking male A. femoralis. Fighting is evoked only in presence of a dynamic bimodal signal, where both temporal and spatial integration of visual (pulsating vocal sac) and acoustic (advertisement call) stimuli is necessary for physical attacks. We will continue our rewarding research by experimentally studying the evolutionary aspects of acoustic communication for the reproductive success of male A. femoralis in the recently granted FWF project 18811 (2006 - 2009).