Acoustic adaptation hypothesis in macro- and micro environments: An analysis of frog calls

Abstract

The acoustic adaptation hypothesis states that acoustic signals are structured so as to maximize their performance under the constraints of the environmental acoustics that characterize their native habitats. However, less attention has been paid to the channels in which acoustic communication takes place and the specific constraints that different channels put on acoustic signals. For frog calls, the channel characteristics depend on two factors: the macro environment—such as the habitat; and the micro environment—such as the sites of the sender and receiver. The data in our analysis came from recordings of 81 species of frogs in French Guyana. These species span a variety of macro environments, such as forest, coastal marsh, and savanna, and a variety of micro environments, such as on vegetation, on ground, and under litter. Two methods—spectrogram correlation and hidden markov models—are used to derive a similarity matrix between these calls. The results indicate that calls show a higher similarity within the same macro environment than between different macro environments. On the other hand, measurements of spectral and temporal features suggest that the micro environment also plays a role, while its effect is more salient for frequency components than for time components.