Abstract
Deterrents that use acoustics to guide fish away from dangerous areas depend on the elicitation of avoidance in the target species. Acoustic deterrents select the optimum frequency based on an assumption that highest avoidance is likely to occur at the greatest sensitivity. However, such an assumption may be unfounded. Using goldfish (Carassius auratus) as a suitable experimental model, this study tested this as a null hypothesis. Under laboratory conditions, the deterrence thresholds of individual goldfish exposed to 120 ms tones at six frequencies (250-2000 Hz) and four Sound Pressure Levels (SPL 115-145 dB) were quantified. The deterrence threshold defined as the SPL at which 25% of the tested population startled was calculated and compared to the hearing threshold obtained using Auditory Evoked Potential and particle acceleration threshold data. The optimum frequency to elicit a startle response was 250 Hz; different from the published hearing and particle acceleration sensitivities based on audiograms. The difference between the deterrence threshold and published hearing threshold data varied from 47.1 dB at 250 Hz to 76 dB at 600 Hz. This study demonstrates that information obtained from audiograms may poorly predict the most suitable frequencies at which avoidance behaviours are elicited in fish.