- University of Toronto
Summary
I analyzed the auditory characteristics of a variety of tympanate moths from areas representing low and high levels of bat (= echolocation signal) diversity. Moths sampled in two African (high diversity) sites reveal significantly higher sensitivities than those analyzed at an Ontario site (low diversity). These sensitivities are particularly pronounced at both low-frequency (5–25 kHz) and high-frequency (80–110 kHz) ranges. I suggest these auditory differences are due to the corresponding increased sensory requirements of tropical moths having to cope with a greater density and diversity of bats (predation pressure) present there.
Methodology
The auditory characteristics of tympanate moths from three study areas representing high and low bat diversities were analyzed in this study. Moths were captured from ultraviolet lights at (i) Queen's University Biology Station, Ontario, Canada (hereafter referred to as QUBS), 44"34' N, 79" 15' W (1 978- 1979); (ii) Lamto Research Station, CBte d'Ivoire, Africa (LAMTO), 6"13' N, 5"02' W (1978); and (iii) Sengwa Wildlife Research Area, Zimbabwe, Africa (SENG), 18"lO'S, 28"13' E (1979). Common moths which gathered at the lights were selected and voucher specimens of each species tested were collected for later identification (vouchers now reside with: the British Museum (Natural History) (Ivorian); author (Zimbabwean and Canadian)).
Neurological auditory threshold curves were computed for each moth tested using techniques described in Fenton and Fullard (1979). Briefly, afferent tympana1 nerve activity was monitored with stainless steel, extracellular, hook electrodes and thresholds to differing continuous pure tones, of 5 kHz increment frequencies from 5 to 110 kHz were determined. Thresholds were taken as that stimulus intensity (in decibels re 20 pPa) causing the first perceptible change in most sensitive auditory cell activity. Values were recorded to randomly chosen frequencies and were checked at the end of a stimulus run. Any values which were remeasured and exhibited a 5 dB or more discrepancy resulted in rejection of the entire run.
Audiograms were then measured to give an approximation of their areas beneath the arbitrary value of 90dB (0.63 Pa) (Fig. 1) (cf. Masterton et al. 1968). Threshold intensity values were converted to pascals to avoid logarithmic biasing and audiogram areas were measured as threshold values x kilohertz in 5-kHz steps. Total audiograrn area (total sensitivity) was subdivided to give sensitivities in four frequency bandwidths: (i) low (5-25 kHz), (ii) low middle (30-50kHz), (iii) high middle (55-75 kHz), and (iv) high (80-110 Hz). Values for all ranges were measured for normality (D'Agostino's test (Zar 1974)) and subsequently compared using a modified Student's t-test for samples with unequal variances (Sokal and Rohlf 1969). In addition to these sensitivity values, Qlo dB values were computed. A unit's QlodB is defined as the ratio of the unit's best frequency I bandwidth at 10 dB above threshold and describes that unit's degree of tuning, high Qlo d~ values representing narrowly tuned fibers and low Qlo dB values representing broadly tuned fibers.