The escape behavior of the cockroachPeriplaneta americana

Summary1.The escape response of the cockroachPeriplaneta americana to the predatory strike of the toadBufo marinus was studied by cinematography and stop-frame analysis. Various manipulations permitted us to appraise the role of the turning behavior mediated by the cereal wind receptors (Camhi and Tom, 1978) in the escape from toads.2.Cockroaches with their cerci covered showed fewer successful escapes from toad strikes than did either normal cockroaches or those with their sternites covered as a control (Table 1).3.The cockroach's escape behavior was similar to its behavioral response to wind puffs (Camhi and Tom, 1978). Similarities included an initial pivot away from the toad (Figs. 1 and 2) and similar initial movements of the metathoracic legs.4.The displacement of air produced by a toad's strike was recordable at the position of the cockroach throughout essentially the entire strike. If wind were the only cue by which the cockroach responded to the toad, the cockroach must have received a supra-threshold wind stimulus at an average time of 75 ms before the toad's tongue appeared. This calculation is the sum of the average time during the strike when the cockroach began its escape movements (17 ms before the toad's tongue appeared [Fig. 4]), and the mean latency of the escape response to wind puffs (58 ms [Roeder, 1963; Camhi and Tom, 1978]). Therefore, the mean wind speed at 75 ms before the toad's tongue appeared should represent an adequate stimulus to evoke the escape behavior. The mean wind speed at this time was 22 mm/s.5.Controlled wind puffs directed at the cerci of restrained cockroaches evoked escape movements of the legs. The mean value of the peak wind speed of a just supra-threshold puff was only 1–5 mm/s.6.Visual, auditory, vibrational and olfactory cues from the toad, in the absence of wind, did not evoke escape behavior.7.These results suggest that the extreme sensitivity of the cockroach's wind-detecting system permits the use of wind as a major channel of information for detecting predators under semi-natural conditions.8.These findings are discussed with reference to (a) the problems this detection system might encounter in a natural environment (b) implications for the neural control of escape behavior in the cockroach, and (c) comparative aspects of escape behavior in different animals.