Auditory-evoked potentials (AEPs) have become a widely utilized measure of hearing sensitivity. Most investigators use pharmacological paralysis to reduce myogenic noise and immobilize the animal for stable electrical recordings, but additional anesthesia is generally not used because the most commonly available fish anesthetic, the cholinergic antagonist tricaine methanosulfate (MS222), is known to disrupt hair cell and primary afferent physiology. Anesthetic agents that do not interfere with auditory function would be a useful adjunct to paralytic immobilization and would reduce any possible distress incurred by prolonged immobilization. In this report we tested the opiate anesthetic fentanyl and compared hearing thresholds in immobilized versus immobilized and anesthetized animals. Short-term effects of mild MS222 anesthesia were also measured via evoked potential audiometry. Animals were tested before and after fentanyl injection (100, 500 and 2500 microg g(-1) fish body-weight) using standard evoked potential audiometry. Tone pips, 0.2-3 kHz, from an aerial loudspeaker served as stimuli. Fentanyl altered evoked potential waveforms slightly but did not alter estimated threshold sensitivity. These results suggest fentanyl be considered as a possible addition to AEP techniques in goldfish (Carassius auratus) and poikilothermic vertebrates generally.