A comparative study of sexual selection on song repertoire size in the avian subfamily Icterinae (Passeriformes:Emberizidae).

Abstract

Sexual selection has been proposed as an explanation of the evolution of song repertoire size in birds. I tested hypotheses of sexual selection on song repertoires by conducting a comparative study of song and syllable repertoires in the subfamily Icterinae; icterines show extreme interspecific variation in mating system, so different species can be predicted to differ in potential for sexual selection. If comparative studies are done with some knowledge of the phylogeny of the group in question, it can be determined whether observed states of traits such as song repertoire represent independent evolutionary events. I reviewed the phylogeny of the Icterinae, based on skeletal, myological, and behavioral traits taken primarily from the literature, as well as vocalizations and plumage characters. I was unable to resolve the phylogenetic relationships of most icterine species, but I identified some groups within the Icterinae that are probably monophyletic. To test hypotheses of sexual selection I determined whether repertoire size was associated with two indices of potential for sexual selection: mating system and sexual size dimorphism. Size dimorphism was determined by a morphometric study based on measurements bird skins. Sexual size dimorphism is highly correlated with both female body size and mating system; predictions about repertoire size variation did not depend on which index of sexual selection I used. I combined a literature review of repertoire size in icterines with my field recordings of several icterine species. Song repertoire was extremely variable between icterine species. There was no overall association of song repertoire size with polygyny or size dimorphism, or with presence or absence of territoriality; within different icterine clades, different patterns of association were observed. Song repertoire showed more geographic variation in two highly polygynous species (Quiscalus mexicanus and Cacicus cela) than in two monogamous species (Quiscalus quiscula and Icterus galbula). Models predicting that sexual selection results in evolution of traits in arbitrary directions thus have some support; models predicting directional increase of repertoire size through sexual selection do not.