SUMMARYThe echolocation sounds of Japanese CF-FM bats (Rhinolophus ferrumequinum nippon) were measured while the bats pursued a moth (Goniocraspidum pryeri) in a flight chamber. Using a 31-channel microphone array system, we investigated how CF-FM bats adjust pulse direction and beam width according to prey position. During the search and approach phases, the horizontal and vertical beam widths were ±22±5 and ±13±5deg, respectively. When bats entered the terminal phase approximately 1m from a moth, distinctive evasive flight by G. pryeri was sometimes observed. Simultaneously, the bats broadened the beam widths of some emissions in both the horizontal (44% of emitted echolocation pulses) and vertical planes (71%). The expanded beam widths were ±36±7deg (horizontal) and ±30±9deg (vertical). When moths began evasive flight, the tracking accuracy decreased compared with that during the approach phase. However, in 97% of emissions during the terminal phase, the beam width was wider than the misalignment (the angular difference between the pulse and target directions). These findings indicate that bats actively adjust their beam width to retain the moving target within a spatial echolocation window during the final capture stages.

Echolocation sounds of Rhinolophus ferrumequinum nippon as they approached a fluttering moth (Goniocraspidum pryeri) were investigated using an on-board telemetry microphone (Telemike). In 40% of the successful moth-capture flights, the moth exhibited distinctive evasive flight behavior, but the bat pursued the moth by following its flight path. When the distance to the moth was approximately 3-4 m, the bats increased the duration of the pulses to 65-95 ms, which is 2-3 times longer than those during landing flight (30-40 ms). The mean of 5.8 long pulses were emitted before the final buzz phase of moth capture, without strengthening the sound pressure level. The mean duration of long pulses (79.9 ± 7.9 ms) corresponded to three times the fluttering period of G. pryeri (26.5 × 3 = 79.5 ms). These findings indicate that the bats adjust the pulse duration to increase the number of temporal repetitions of fluttering information rather than to produce more intense sonar sounds to receive fine insect echoes. The bats exhibited Doppler-shift compensation for echoes returning from large static objects ahead, but not for echoes from target moths, even though the bats were focused on capturing the moths. Furthermore, the echoes of the Telemike recordings from target moths showed spectral glints of approximately 1-1.5 kHz caused by the fluttering of the moths but not amplitude glints because of the highly acoustical attenuation of ultrasound in the air, suggesting that spectral information may be more robust than amplitude information in echoes during moth capturing flight.

Analysis of the bat's reactions to relevant target echoes enables us to directly assess biosonar performance. Here, we recorded the sonar broadcast and its echoes the bat received during flight by using an on-board telemetry microphone (Telemike) mounted on the bat's back. Telemike recordings confirmed that flying bats adjust the amplitude and frequency of their sonar broadcasts to compensate for increases in echo amplitude and for Doppler-shifts. For insect capturing, the bat exhibited Doppler-shift compensation for echoes from the static target ahead, but not for echoes from the target moth even though the flying bat attended to the moth for capture. Positive and negative Doppler shifts (acoustic glints) caused by insect fluttering were observed in the constant-frequency component of observed echoes, which synchronized with wingbeat cycle of the moth. Combined frequency and amplitude compensation for the static target may be advantageous for detection of acoustic glints of target prey. We also constructed multiple-microphone arrays for tagging wild aerial-feeding insectivorous bats. Not only the location of the bat, but also direction and directivity of the bat's broadcast can be measured. This will allow us to investigate 3-D search algorithm of multiple targets by the bat. [supported by JSPS and ONR]