Song length variation serves multiple functions in the white-crowned sparrow

A wide variety of hypotheses has been proposed to explain the structural diversity in bird song repertoires. Song diversity is frequently described in terms of song "types" or within-type "variants." Male Puget Sound whitecrowned sparrows, Zonotrichia leucophrys pugetensis, produce variants of their single adult song type by altering the number of repetitions of syllables in the terminal trill. We tested whether variation in trill length correlated with distance to the receiver and with signaling context as predicted by the eavesdropping avoidance hypothesis and the strategic signaling hypothesis. In accordance with the eavesdropping avoidance hypothesis, males sang variants with shorter trills, and sang quieter and less frequently when near their mate during the incubation phase than during spontaneous singing while unpaired. Males also sang variants with short trills, but at a high rate and variable amplitude when within 10 m of an opponent during close male-male territorial interactions. In agreement with the strategic signaling hypothesis, males decreased trill length immediately before chasing an opponent, but did not change length consistently prior to flight. We conclude that the occurrence of short quiet songs sung near the mate agrees with predictions of the eavesdropping avoidance hypothesis, while short songs sung near other males are best explained by the strategic signaling hypothesis. Trill length variation may be a conventional signal of aggressive intentions in male-male contests stabilized by receiverimposed retaliation costs.