The echolocation of flying insects by bats

Abstract

• 1. Bats of the genus Myotis (M. lucifugus, M. subulatus leibii and M. keenii septentrionalis) have been studied while pursuing and capturing small insects under laboratory conditions. It is apparently important to provide fairly large numbers of such insects in order to elicit insect catching behaviour indoors.

• 2. Insect catches are individually directed pursuit manoeuvres; each insect is detected, located, and intercepted in flight within about half a second.

• 3. Certain individual bats caught mosquitos (Culex quinquefaciatus) and fruit flies (Drosophila robusta and D. melanogaster) at remarkably high rates which could be measured conservatively by the gain in weight of the bat. Sometimes a bat would average as many as 10 mosquitos or 14 fruit flies per minute during a period of several minutes. In four cases motion pictures showed two separate Drosophila catches within half a second.

• 4. The orientation sounds of the hunting bat are adjusted in a manner that seems appropriate for the echolocation of single insects one at a time. There is a search phase before the occurrence of any apparent reaction to the insect. In this phase the frequency drops from about 100 to 50 kilocycles during each pulse of sound, and the pulses are emitted by M. lucifugus at intervals of 50 to 100 milliseconds.

• 5. When an insect is detected the search phase gives way to an approach phase characterized by a progressive shortening of the pulse-to-pulse interval and, if necessary, a sharp turn towards the insect. In this phase the pulse duration may shorten somewhat, but the frequencies remain approximately the same as in the search phase or drop slightly.

• 6. When the bat is within a few centimetres of the insect there is a terminal phase in which the pulse duration and interval between pulses shorten to about 0·5 millisecond and 5 or 6 milliseconds respectively. Contrary to a conclusion reached earlier on the basis of much less adequate data (Griffin, 1953), the frequency drops in the terminal phase, sometimes to 25 or 30 kilocycles. This is the buzz, which also occurs in many cases when the bat is dodging wires or landing.

• 7. The distance from the insect at which detection occurs can be judged by the shift from search to approach patterns. This distance of detection is commonly about 50 cm. for Drosophila, and it occasionally may be as much as a metre with fruit flies or mosquitos.

• 8. Two M. lucifugus which had become adept at catching Drosophila in the laboratory were exposed to broad band thermal noise either at low frequencies (0·1–15 kilocycles) or high (20–100 kilocycles). The low frequency noise had an approximately uniform spectrum level of about 50 decibels per cycle band width (re 0·0002 dyne/cm²) from 0·1 to 8 kilocycles. It was thus very loud compared to the flight sounds of Drosophila which have a fundamental frequency of a few hundred cycles/second and a maximum sound pressure level of 20–25 decibels at the distances of detection by these bats. The high frequency noise was of low and varying intensity, but it discouraged or prevented insect catching. The low frequency noise, on the other hand, had no effect on insect catching; the bats gained weight in this noise (and in the dark) just as rapidly as in the quiet. Although bats sometimes detect insect prey by passive listening to sounds emanating from the insects themselves, these experiments appear to us to establish conclusively that small and relatively silent insects are often detected by echolocation.