As a consequence of vocal learning, the songs of male songbirds often vary geographically, We examined three proximate hypotheses that predict variation in the incidence of local song sharing among species: the annual cycle, breeding season duration,and epiphenomenon hypotheses. To test these hypotheses, we compiled data from the literature on local song sharing, annual cycle, breeding location isophane (an index of breeding season duration), repertoire size, and phylogenetic relationships of 65 populations in the family Fringillidae. Mapping characters on the phylogenetic tree revealed that local song sharing, migratory status, and bleeding location isophane were randomly distributed on the phylogeny. Breeding isophane and migratory status were themselves correlated and had significant effects on the probability of sharing local songs. As predicted by the annual cycle and breeding season duration hypotheses, males in sedentary populations and populations breeding at low isophanes were more likelyto share local songs. In subsets of the data, the annual cycle hypothesis received more support than did the breeding season duration hypothesis. Neither hypothesis explained more than 17% of the variation in the incidence of song sharing, and we suggest that detailed information on song function, male dispersal, and constraints on the timing of song learning will be needed to better predict the occurrence of local song sharing.
展开▼