When a bat is cruising around looking for its next meal it sends out its ultrasonic pulses and listens carefully to the tell-tale echoes of a moth fluttering by. Despite the sophistication of a bat's echolocation system,moths have a few defences of their own. The dogbane tiger moth responds to a bat's attacking clicks with clicks of its own when the bat gets too close,either interfering with the bat's echolocation or warning the bat that it faces a bitter mouthful. However the moth has to choose very carefully when to click at a bat; too early draws unnecessary attention to itself, and leaving it too late is very risky. James Fullard and colleagues at the University of Toronto and Cornell University investigated which aspects of a bat's calls the moths use to decide whether to defend themselves or not (p. 2481).

A bat can vary many of the characteristics of its echolocating calls, such as the frequency or intensity of the calls. It also alters the duration of the individual pulses or the time between the start of each pulse, known as pulse period, which in turn affect the duty cycle, which is the percentage of the total time a calling bat is making a sound. A moth could use any of these characteristics to identify an attacking bat, but the question is, which one?

First the team recorded from the moths' auditory neurons to test their response to different frequency sounds, and find their auditory threshold. They found while moths' auditory neurons are sensitive between 30 and 50 kHz,they didn't show any difference in their clicking behaviour to different frequencies, showing that moths are essentially tone deaf.

Having ruled out frequency, they turned to a technique which allowed them to test if moths could tell the difference between two sets of bat calls. They attached the moths to a pin with wax and suspended them above a speaker in a dark room, before repeatedly playing a sequence of bat pulses to them and recording their clicking response. Once the moths had habituated –stopped responding to the bat signals – they changed one or more aspects of the signal to see whether the moths could tell the difference and start clicking again.

They found that once the moths had habituated to a particular signal, they had to change the duty cycle by 60% or more before the moths started responding again. Next they tested to see how the moths responded to varying duty cycles, and found that they were most sensitive to pulse periods of 20 ms. They were less sensitive to longer and shorter pulse periods, showing that the response was tuned.

Finally to show that pulse period was the primary parameter that the moths were responding to, they habituated the moths to `searching' bat signals and then tested them with `attacking' bat signals. Both signals had the same duty cycle, but the pulse period changed as the bats switched from searching to attacking. They found that the moths responded very strongly to the switch between a searching and attacking bat, showing that pulse period is the defining feature they use to identify when they are in danger. However they also found that their response was influenced by the intensity of the signal. If an attacking bat is pulsing to another insect 30 m away, then the moth is not going to respond because the intensity is too low, and it doesn't want to draw attention to itself. But if the intensity is right, and the bat is bearing down on the moth and bombarding it with high intensity, short period signals, it will click to avoid being snatched and eaten.

Fullard, J. H., Ratcliffe, J. M. and Christie, C. G.(
). Acoustic feature recognition in the dogbane tiger moth, Cycnia tenera.
J. Exp. Biol.