Communication of Intentions in Agonistic Contexts By the Pigeon Guillemot, Cepphus Columba

Abstract

This observational study of communication behavior of the pigeon guillemot (Cepphus columba) in agonistic contexts studied two questions: (1) What factors influence the outcome of interactions between male territory owners and intruders? and (2) Is information about the signaller's intention contained in signals? Pigeon guillemots nest in rock crevices which they defend against conspecifics. Intruders tend to visit single birds on territories as opposed to pairs or unoccupied territories. Male territory owners give a long (mean = 85 sec) series of whistled notes, the Hunch-whistle display, and enter the nest crevice when a female visits. Males also Hunch-whistle spontaneously and females approach, enter the nest, and appear to inspect the male's nest-site. Mutual Hunch-whistles usually involve male owners and male intruders, and tend to occur soon after a bird occupies a territory. The outcome of interactions between male territory owners and non-mate intruders was independent of the intruder's sex. Intruders retreated without escalation occurring in 51 % of cases and the owner attacked in 2 % of the cases. Fights involving both birds were rare and tended to occur early in the breeding season when both males had spent time on the territory in the previous week. No injuries were witnessed in over 600 aggressive interactions. Territory owners won 21 of 23 fights with male intruders. Size and age did not appear to contribute to fighting success. A pay-off asymmetry in favor of owners best explains the dominance of owners over intruders. Analysis of behavior sequences using a three-dimensional contingency table revealed that displays were significantly associated with different responses by recipients and with different subsequent behaviors by the signaller. The response also had an effect on the displayer's subsequent behavior. Flight by recipients decreased the probability of subsequent attack by signallers but other, less predictable forms of interaction between response and subequent behavior also occurred. Territory intruders flew after 43 % of the occurrences of the Neck-stretched display; no owner display preceded attack more than 14% of the time. Variation in the length and frequency of Hunch-whistle notes reflect changing probabilities that the signaller will sit, move, or attack. The owner/intruder dichotomy best explains which bird attacks/flies in an interaction. Variability in the bird's vocalization provides information about when flight or attack will occur. Differences in costs associated with attack versus flight may explain why flight is potentially more predictable than attack.