The foraging behaviour and sensory basis of prey detection in the long-eared bat, Myotis evotis, was examined. Bats were adept at gleaning moths from both the ground and a bark-covered vertical trellis. Behavioural experiments in which sensory cues were selectively isolated or removed showed that preygenerated (i.e. moth-fluttering) sounds were both sufficient and necessary for prey detection by gleaning M. evotis. However, prey sounds did not have to be fluttering sounds as bats also detected wingless moths. Echolocation and visual cues were non-essential for prey detection. Ultrasonic recordings showed that echolocation was inconsistently used by gleaning M. evotis; calls were detected in only 58% of hovers and 65% of attacks. When echolocation was used, bats ceased calling on average 179 ms before attacking, and feeding buzzes (the rapid increase in pulse repetition rate associated with prey capture) were never recorded. This study demonstrates how the sensory modality of prey detection, in combination with a particular foraging strategy, can have a marked influence on prey availability. It appears that for M. evotis, the availability of moths as prey is increased by a prey detection system biased towards the detection of prey-generated sounds. © 1992 The Association for the Study of Animal Behaviour.