Echolocating bats, Eptesicus fuscus, were trained in two distinct behavioral tasks to investigate the images they perceive of a sonar point target. In the first task, bats were trained in a two-alternative forced-choice procedure to detect electronically simulated target echoes at a range of approximately 57 cm. Half of the trials in the detection task contained echoes from a stationary target (simulated by a fixed echo delay) and half contained echoes from a jittering target (simulated by an echo delay alternating between two time values over successive sonar emissions). In the second task, bats were trained in a two-alternative forced-choice procedure to discriminate between electronically simulated stationary and jittering targets, centered about a range of 57 cm. Both target detection and target jitter discrimination performance were assessed as a function of jitter magnitude, with jitter values ranging from 0-60 microseconds (corresponding to a change in distance of 0 to 10.3 mm). In both detection and discrimination tasks, the bat's performance changed cyclically with the magnitude of echo jitter. Specifically, when the phase of the playback echoes was unchanged, performance levels were poorest at 0 and 30 microseconds, and when the phase of the echoes alternated by 180 deg from one to the next, performance levels were poorest at 15 and 40-50 microseconds. The results suggest that Eptesicus is sensitive to the phase reversal of echoes and thus have implications for assessing receiver models of echolocation.