No abstract

Previous theory and research have suggested that bird species with song repertoires in general, and song sparrows (Melospiza melodia) in particular, cannot readily discriminate between the songs of neighbors and strangers. In a recent study (Stoddard et al. 1991) we showed that song sparrows can in fact discriminate neighbors from strangers on the basis of song. In this study we sought to demonstrate that song sparrows can make the finer discrimination between individual neighbors and that they can do so on the basis of a single song type. We compared the response of territorial males to song playback of neighbors and strangers at three locations : the neighbor's regular boundary, the opposite boundary, and the center of the territory. The birds showed strong neighbor-stranger discrimination at the regular boundary but not at the opposite boundary, nor in the center of the territory. The differences in song discrimination between different boundary locations indicate that song sparrows associate particular songs with particular territories, effectively discriminating between individual neighbors on the basis of song. Song repertoires themselves do not interfere with neighbor recognition to the extent originally postulated. As speakers are moved inside the territory from the border, however, the degree of discrimination diminishes. We believe that differences in speaker placement may have contributed to the variability in neighbor-stranger discrimination observed in previous studies of the song sparrow and perhaps other repertoire species as well. This interpretation is consistent with data from another song sparrow population showing that half the territory takeovers are by immediate neighbors.

Song playback to song sparrows (Melospiza melodia) in a resident population in Washington state showed that the tendency of birds to respond with the same song type (match) depended on the identity of the singer. Matching rates were high to 'self song (60%) and 'stranger' song (50%) and low to 'neighbor' song (20%, not significantly above chance level). The higher matching rate to stranger song was particularly interesting, since the neighbor test songs were generally more similar to the subjects' songs than were the stranger test songs (the self songs, of course, were the most similar). The importance of the neighbour–stranger contrast, in addition to song similarity, in eliciting song matching confirms similar conclusions from earlier studies on the great tit (Parus major) and western meadowlark (Sturna neglecta).

No abstract