Song type matching as threat: a test using interactive playback

Neighbouring song sparrows, Melospiza melodia, in our Seattle population share song types and sometimes respond to neighbour song by type matching (replying with the same song type being sung) or repertoire matching (replying with a shared song other than the type being sung). Based on our modification of the threat hypothesis of Krebs et al. (1981, Animal Behaviour, 29, 918–923), according to which, type matching is a threat, we develop two predictions concerning type matching. (1) A bird will be more likely to escalate when type matched by his neighbour than when repertoire matched. (2) When type matched, birds who escalate will continue to sing the matched song type, while those who de-escalate will switch off the matched type or stop singing. To test these predictions we conducted an interactive playback experiment that simulated a bird in an adjoining territory issuing a song reply to a singing subject. We exposed subjects to three song playback conditions, each on a different day: a type match, a repertoire match and an unshared stranger song (to provide an estimate of each subjects maximum response for comparison with their responses to type matching). Subjects responded consistently and most aggressively to stranger song. Subjects responded more aggressively to type-matching playback than to repertoire-matching playback, supporting our first prediction. In type-matching trials, subjects did not always respond aggressively, and those who stayed on the same song type throughout the trial responded more aggressively than those who switched to a different song, supporting the second prediction.