Alarm signals emitted by animals may not be all that they seem. But a good example has been identified in the whistling sound of a crested pigeon's wings when it takes flight in response to a predator.
In keeping with the principle that there's safety in numbers, animals gain various anti-predatory benefits from forming groups. One such benefit is that an individual in a group can be alerted to danger by a group-mate, rather than by detecting the predator itself. As they recount in Proceedings of the Royal Society, Hingee and Magrath1 have compiled convincing evidence of an unusual example of this behaviour.
Intra-group alarm signalling obviously requires communication about predators to occur between group members. But the reliability of these alarm calls is often unclear2, in that they could be given deceptively to reduce competition for food3. Although the evidence for deception is equivocal, false alarms may be common if there is little cost to emitting an alarm call when no predator is in fact present. In many species, group members that detect a predator do not give an obvious alarm vocalization, but simply flee.
In the limited previous experimentation2 carried out on this subject, group-mates have apparently struggled to differentiate between individuals leaving the group because they've seen a predator and those leaving for other reasons. Hingee and Magrath1, however, show that the rattle-like whistling sound generated by the flapping of a fleeing crested pigeon (Ocyphaps lophotes) can be reliably associated with flight in response to a predator, and that this information is used by group-mates to trigger their own anti-predatory behaviour.
The reliability of this signal stems from its nature: the whistling sound is generated by the movement of air across the wing (an audio file accompanies the article online). Because birds must take flight to produce the signal, the high energetic cost of flapping flight will discourage false alarms. Flock-mates can differentiate between the noises made by different types of flight, and in the authors' experiments they took flight in response to a playback of sound emitted by birds in alarmed flight but not to the sounds that birds make in other, non-predator-driven, departures from the flock. Test flocks took flight immediately in 11 of 15 trials with an alarm whistle and in none of 15 cases with a non-alarm whistle. Such non-alarm whistles tend to be of lower tempo and amplitude than alarm whistles.
Hingee and Magrath1 also performed studies with alarm whistles whose amplitude had been experimentally reduced before being played to test flocks. In these experiments, flocks generally did not flee, but showed higher levels of vigilance immediately after the playback than did flocks subjected to control playbacks. However, even non-alarm whistles may be useful to flock-mates, because such flight sounds still induced an increase in vigilance. Indeed, there was a trend for vigilance to increase with increasing tempo of such flight sounds (that is, the more similar they became to alarm whistles). So the birds may not simply get information on whether or not an attack is imminent, but may also be able to extract useful information about the severity of risk.
Hingee and Magrath provide morphological evidence that the shape of the eighth primary feather in this species is unusual, and they link this feature to the characteristic noise of the bird in flight. Experimental manipulation to confirm this link would now be valuable, as it could have a bearing on the likely generality of this mode of anti-predatory communication and on its evolution within the crested pigeon.
If the unusual feather structure of this species is essential for signal production, then the whistle probably evolved because it modifies the behaviour of the signal receivers in such a way as to benefit the whistler. This might come about because the original detector of a predator benefits (through dilution of risk or sensory confusion of the predator) if other flock-mates flee along with it. There may also be particular benefits from warning a mate or kin. Or perhaps the intended receiver of the signal is the predator, if the signal reliably indicates that that whistler is already taking rapid avoidance flight and so the predator would be better focusing its attention on other flock-mates. Thus, other flock-mates might be eavesdroppers that gain useful information from a signal that evolved for other reasons. Finally, it could even be that the noise production originally evolved because it benefited the signaller in a context unrelated to predation; it might be relevant that the courtship behaviour of the crested pigeon involves opening and closing the wings.
As for the generality of this type of anti-predatory communication, it is certainly true that birds taking flight (Fig. 1) or mammals taking to their heels can both be noisy. The limited previous empirical study2 does not provide strong evidence of a general ability of group-mates to distinguish between predator-induced and other departures from a group. However, Hingee and Magrath's research suggests that we should return to this question, and perhaps concentrate on the noise of fleeing as a signal that is difficult to fake, and on species that gather in structurally complex environments (such as scrubland) where noise might be easier to detect than visual evidence of a fleeing group-mate. No matter what, we have not heard the last of the whistling pigeon.
Hingee, M. & Magrath, R. D. Proc. R. Soc. Lond. B doi:10.1098/rspb.2009.1110 (2009). | Article |
Lima, S. L. Anim. Behav. 50, 1097–1108 (1995).
Munn, C. A. Nature 319, 143–145 (1986).