COLLECTIVE DETECTION OF PREDATORY ATTACK BY SOCIAL FORAGERS - FRAUGHT WITH AMBIGUITY

A cornerstone in the study of anti-predatory vigilance in socially feeding vertebrates is the idea of collective detection, that all members of the group are unambiguously alerted to an attack as long as it is detected by at least one group member. Collective detection in mixed flocks of emberizid sparrows, however, is an ambiguous and uncertain process. In these flocks, detectors of attack impart little information directly to their flockmates. Non-detectors infer the possibility of an approaching threat based upon the occurrence of departures from the flock, and appear unable to distinguish non-threat-induced departures from those induced by an approaching threat. Under such circumstances, theory suggests that multiple detections of attack (or multiple departures) may be required to incite alarm in the flock. Experiments demonstrated such an effect: multiple threat-induced departures over a short time interval had a markedly greater effect on the escape behaviour of non-detectors than did single threat-induced departures. The effect of departures was moderated by their environmental context. Non-detectors responded more readily to departures that (1) originated from the outer portion of the flock, or (2) occurred when feeding far from protective cover. These sparrows also showed a range of behavioural responses to departures, from no response, to alertness, to immediate, full-blown escape. This ambiguous, departure-based form of collective detection implies that individuals must rely considerably on personal vigilance to detect predatory attack. Such a reliance may leave animals more vigilant than suggested by models of vigilance based upon the conventional version of collective detection. (C) 1995 The Association for the Study of Animal Behaviour