Can the acoustic adaptation hypothesis predict the structure of Australian birdsong?

The acoustic adaptation hypothesis (AAH) predicts that long-distance signals will be structured so as to maximize their transmission fidelity. Previous studies testing the hypothesis on birdsong have provided equivocal support. The best support comes from large-scale comparative studies and those studies where habitat is characterized as "open" versus "densely vegetated." In the first case, sufficient statistical power is present to detect even small effects on song structure, whereas in the later case the "effect size" of the habitat may be sufficiently large. Most studies have focused on Holarctic or Neotropical species, which may ultimately share a common evolutionary history. In this study, Australian birds were chosen for a phylogenetically independent test of key predictions of the AAH. Specifically, birds in open habitats were predicted to sing songs with higher frequencies, greater bandwidth, have a greater probability of having overtones, and be emitted at a quicker rate than birds inhabiting densely forested habitats. Acoustic measurements were made on commercially available recordings of 121 species of Australian birds from 41 different families. Analyses controlled for variation explained by body mass (using ANCOVA), and phylogeny (using genus pairs analyses). We found only modest support for the AAH. Our finding that birds in open habitats produced higher frequency vocalizations and greater bandwidth vocalizations is also consistent with hypotheses about signal structure facilitating auditory distance assessment. Forest birds may therefore rely on cues other than frequency-dependent attenuation for ranging.