LAUSR

ABSTRACT Geographic variation in acoustic signals may arise in vocal learning species due to processes of cultural evolution and population dynamics, but few studies have evaluated geographic vocal variation for large-bodied non-oscine species such as parrots. We evaluated similarity in a contact call of the Military Macaw (Ara militaris) among three populations along the coast of Jalisco, Mexico. We compared contact calls among sites using two approaches i) analysis of specific acoustic parameter measures using GLMM on Principal Components; and ii) evaluation of spectrogram similarity using cross-correlation with a Mantel test to evaluate site and distance effects. Acoustic parameter analysis found that incorporating the site where recordings were obtained significantly explained variation in call features, for both the complete and reduced, balanced dataset. The spectrogram cross-correlations similarity analysis indicated an association with site in spectral similarity of calls, and that call similarity decreased with distance. Our results demonstrated the accumulation of small, fine-scale changes in Military Macaw calls with distance, suggesting that large-bodied non-oscines such as macaws may be able to maintain connectivity among sites by dispersal, facilitating call diffusion, while limited movements among some populations may account for the differentiation among sites in call features.