Auditory brainstem responses (ABR) and distortion product otoacoustic emissions (DPOAE) have been applied to the evaluation of peripheral auditory function. To date, no comparison of their relative sensitivities to aminoglycoside toxicity has been reported. The purpose of this study was to compare click evoked ABR testing and 2f1-f2 DPOAEs as detectors of cochlear damage induced by gentamicin treatment in guinea-pigs. ABR thresholds to click stimuli were recorded. DPOAE amplitude input/output functions were recorded using three different primary tone level conditions. In one condition, L1 was changed relative to a fixed L2. In the next condition, L2 was changed relative to a fixed L1. In the third condition, L1 and L2 were both changed while maintaining a consistent L1-L2 difference. Baseline L1-L2 differences were determined by adjusting L2 to produce the maximum DPOAE amplitude in each ear of each subject. Guinea-pigs were treated for a four week period with intramuscular injections of gentamicin. ABRs and DPOAEs were monitored and compared to baseline recordings or untreated control groups. DPOAE testing detected cochlear damage earlier than ABR testing. Changes in DPOAE input/output functions were noted after two weeks of treatment while changes in ABR threshold were not identified until after three weeks of treatment. The sensitivity of the DPOAE testing was related to the stimulus conditions utilized. L1-L2 differences were determined in the control group and for the initial test session by fixing L1 at 75 dB SPL and lowering L2 until the maximum DPOAE amplitude was obtained for each ear. All subsequent testing was begun using the optimal L1-L2 difference for each ear. The stimulus condition in which only L1 was changed and L2 was fixed resulted in the least sensitive DPOAE indicator of cochlear damage. The condition in which both L1 and L2 were changed, but the L1-L2 difference remained constant, resulted in the most sensitive indicator of damage. The onset and degree of cochlear damage secondary to gentamicin treatment was subject dependent. This study demonstrates that 2f1-f2 DPOAE testing is preferable to click evoked ABR testing for early detection of gentamicin toxicity of the cochlea. It also indicates that DPOAE stimulus parameters must be considered when developing test protocols. Specifically, recording the DPOAE amplitude input/output function while maintaining an effective L1-L2 difference is preferable to changing either L1 or L2 individually.