Geographic Variation of Song Form within and among Chestnut-Sided Warbler Populations

Abstract

Chestnut-sided Warbler (Dendroica pensylvanica) song repertoires include both accented-ending (AE) and unaccented-ending (UE) songs, and the two song categories are used in different behavioral contexts. To determine if the song categories also differ in patterns of spatial distribution, I analyzed local geographic variation in song form within a Massa- chusetts population and also compared songs from this local population with songs from other geographic regions. I found that: (1) UE repertoires of individuals were more similar to those of nearby individuals in the local population than to those of more distant birds; (2) most UE song types and song elements tended to cluster on groups of adjacent territories; and (3) UE songs from geographically distant sites tended not to match types from the main study site. In contrast, my findings indicate that: (1) AE-repertoire similarities were not significantly related to local interindividual distances; (2) AE song types were dispersed at random over the local area; and (3) nearly all AE song types from distant sites were identical to the AE types found at the main study site. The contrast between variable, geographically clustered UE songs, and stereotyped, geographically dispersed AE songs suggests that the two song categories serve different communication functions. Received 21 December 1994, accepted 28 June 1995.