The ototoxic action of salicylate was investigated in the guinea pig by perfusion of both salicylate and bromosalicylate through scala tympani. The results qualitatively confirmed experiments using intravenous administration in cats (Stypulkowski, 1990), showing dose-dependent elevations in compound action potential (CAP) thresholds, increases in cochlear microphonics (CM) and level-dependent reductions in 2f1-f2 acoustic distortion products. The endocochlear potential was not significantly affected and iontophoretic injection of salicylate into scala media had no measurable effect on CAP thresholds, consistent with an action on the basolateral walls of the hair cells. Perfusion with indomethacin produced effects similar to those of the salicylates, but at non-physiological doses. Together with the great effectiveness of 5-bromosalicylate, this suggests that salicylate does not act by inhibiting prostaglandin synthesis. The results are qualitatively consistent with the proposition that salicylates act on the basolateral walls of the outer hair cells. However, the magnitude of the CM increases, particularly at high drug concentrations, and the fact that salicylate reduced, but did not eliminate the effects of olivocochlear efferent stimulation on CM amplitude indicate that a simple explanation for salicylate effects based solely on a conductance increase in the outer hair cell membranes may be inadequate.