What cues can a bird use to indicate the threat of an approaching predator? The most direct, of course, is to keep the head up and keep a look out. However, this clearly limits the time available for bending down and pecking. Similarly, keeping an eye out for movements of other members of the flock could be informative – if you see a neighbour taking off without the usual warm-up, or if you see everyone else flying away, you begin to wonder why – but can also be costly in terms of time spent feeding. Alarm calls from other members of the flock can be an effective means of warning, and need not limit the time spent feeding. However, these can also be very untrustworthy ‘cry wolf’ signals, benefitting the caller to the detriment of the listener, who might be persuaded to unnecessarily fly away from the food source. An alternative that bridges these extremes is listening to the sounds of take-off of other flock members. Might aspects of these sounds be used as indicators of threat? What are these aspects? And can fearful flights be distinguished from normal take-offs?

These questions were recently approached by Mae Hingee and Robert Magrath (Australian National University, Canberra), studying the crested pigeon. Also known as the ‘whistle-winged’ pigeon, these birds produce a characteristic sound when they flap, probably due to vibrations of the 4th primary, which is only half the width of neighbouring flight feathers. The sounds of flights were recorded for pigeons in both normal, relaxed take-offs, and ascents (presumably panicked) induced by throwing a gliding model of a predatory hawk over the flock. Alarmed take-offs were both louder (at a similar distance) and had higher ‘element cycle rates’ indicating higher wingbeat frequencies. This is relatively unsurprising: worried birds flap harder and faster. But do members of a flock take any notice of this acoustic information?

To determine this, Mae and Rob performed playback experiments on flocks of unsuspecting pigeons using the sounds of alarmed and non-alarmed take-off flight, and observed how nearby members of the flock responded. When faced with the sounds of an unfamiliar pigeon in alarmed flight at full volume, the majority of flocks fled. When exposed to non-alarm flight sounds, very few birds responded. When non-alarm flight sounds were amplified to equate to alarmed sounds, there was still little response; when alarm sounds were quietened to equate to non-alarmed noises, some flocks fled, while some birds stayed. However, those staying showed a significant increase in ‘vigilance’ (they spent more time with their heads up, looking around), indicating that the wingbeat frequency, quite apart from the sound volume, worried the flock.

If a signal as simple as wingbeat frequency is the key to indicating alarm, why have these pigeons evolved special flight feathers and a ‘wing whistle’ sound? Details of the flight sound that are presumably related to the wing whistle mechanism (for instance the fundamental frequencies) were statistically poor at differentiating alarmed flights from un-alarmed flights. So why evolve the special wing sounds, and not stick to the atonal ‘whooshing’ sounds that are an inevitable consequence of flapping? One possibility is simply volume – whistles can be considerably louder than whooshing – to catch a neighbour's attention; another, presumably, is some degree of species specificity. There are obvious disadvantages to taking off in fright whenever a bird of a slightly smaller species, and consequently higher, panic-inducing flap frequency, flies nearby.

R. D.
Flights of fear: a mechanical wing whistle sounds the alarm in a flocking bird.
Proc. R. Soc. B