Latitudinal variation in whistle structure of the estuarine dolphin Sotalia guianensis

Abstract

[Analyses of intraspecific variation in behaviour can provide insights into the process of behavioural diversification. In this study we characterized geographic variation in whistle structure of the estuarine dolphin, Sotalia guianensis, at 10 sites of a ∼4000 km transect along the coast of Brazil. Whistles were recorded, digitized (N = 527), and six acoustic parameters measured from spectrograms. ANOVA revealed geographic variation in all six parameters. Post-hoc analyses indicated that pairs of sites with similar whistle parameters tended to be adjacent to each other, and discriminant functions analysis (DFA) revealed that the greatest differences in whistle structure tended to occur between the most distant sites. Whistle parameters did not, however, express simple clinal variation, and DFA was a poor predictor of site (29.0% accuracy). Mantel tests revealed significant correlations between geographic distance and variation in three of six whistle parameters. Linear regression analyses for our data set, pooled with data from Azevedo & Van Sluys (2005, J. Acoust. Soc. Am. 117:1456), revealed significantly lower starting and minimum frequencies for southern populations relative to northern populations. A geographic discontinuity in these two features is observed, varying on either side of the easternmost tip of South America. Possible ecological and evolutionary explanations for these patterns are discussed., Analyses of intraspecific variation in behaviour can provide insights into the process of behavioural diversification. In this study we characterized geographic variation in whistle structure of the estuarine dolphin, Sotalia guianensis , at 10 sites of a ∼4000 km transect along the coast of Brazil. Whistles were recorded, digitized ( N = 527), and six acoustic parameters measured from spectrograms. ANOVA revealed geographic variation in all six parameters. Post-hoc analyses indicated that pairs of sites with similar whistle parameters tended to be adjacent to each other, and discriminant functions analysis (DFA) revealed that the greatest differences in whistle structure tended to occur between the most distant sites. Whistle parameters did not, however, express simple clinal variation, and DFA was a poor predictor of site (29.0% accuracy). Mantel tests revealed significant correlations between geographic distance and variation in three of six whistle parameters. Linear regression analyses for our data set, pooled with data from Azevedo & Van Sluys (2005, J. Acoust. Soc. Am. 117:1456), revealed significantly lower starting and minimum frequencies for southern populations relative to northern populations. A geographic discontinuity in these two features is observed, varying on either side of the easternmost tip of South America. Possible ecological and evolutionary explanations for these patterns are discussed.]