• Ratcliffe, John M.
  • Fullard, James H.
  • Arthur, Benjamin
  • Hoy, Ronald


Moths and butterflies flying in search of mates risk detection by numerous aerial predators; under the cover of night, the greatest threat will often be from insectivorous bats. During such encounters, the toxic dogbane tiger moth, Cycnia tenera uses the received intensity, duration and emission pattern of the bat's echolocation calls to determine when, and how many, defensive ultrasonic clicks to produce in return. These clicks, which constitute an acoustic startle response, act as warning signals against bats in flight. Using an integrated test of stimulus generalization and dishabituation, here we show that C. tenera is able to discriminate between the echolocation calls characteristic of a bat that has only just detected it versus those of a bat actively in pursuit of it. We also show that C. tenera habituates more profoundly to the former stimulus train (‘early attack’) than to the latter (‘late attack’), even though it was initially equally responsive to both stimuli. Matched sensory and behavioural data indicate that reduced responsiveness reflects habituation and is not merely attributable to sensory adaptation or motor fatigue. In search of mates in the face of bats, C. tenera's ability to discriminate between attacking bats representing different levels of risk, and to habituate less so to those most dangerous, should function as an adaptive cost–benefit trade-off mechanism in nature.


Experiments were conducted near Chaffey’s Lock, Ontario, Canada at the Queen’s University Biological Station (QUBS). Cycnia tenera were reared from eggs collected from wildcaught females and raised to pupae on dogbane (Apocynum androsaemifolium) and Indian hemp (A. cannabinum). Pupae were over-wintered at the University of Toronto in constant temperature rooms at 48C (12 L : 12 D regime) for several months and then transferred to constant temperature rooms at 258C (16 L : 8 D). Adults emerged two to three weeks later at QUBS and were allowed to mature for 12–24 h before being used as subjects. Both early and late attack groups (described below) comprised 10 moths: 5 males and 5 females, for a total of 20 moths.

Moths were tethered from their dorsal thorax using wax and rigid wire in a dark chamber lined with sound-absorbing foam. Moths were allowed to acclimate for 20 min and remained relatively motionless throughout trials. Acoustic stimuli and tymbal clicks were detected using a condenser microphone (CM16, Avisoft, Berlin, Germany) equidistant from moth and speaker, and recorded using an acquisition board (sampling rate ¼ 250 kHz, UltraSoundGate 416-200, Avisoft) connected to a PC running Avisoft RECORDER. Digital recordings (.wav files) were subsequently analysed using BATSOUND PRO v. 3.2 (Pettersson Elektronik AB, Uppsala, Sweden). We noted the number of click modulation cycles (MCs; the 14–20 clicks produced over the course (approx. 20 ms) of a single tymbal’s collapse and recovery) and each MC’s onset time as referenced to the beginning of each trial (i.e. the time at which the first pulse reached the moth’s ipsilateral ear).