Two-tone suppression has been observed at different levels along the auditory pathway. It has been shown at the level of the auditory nerve and in the mechanical response of the basilar membrane (BM). Geisler et al. (1990) postulated that suppression originates from the saturation of outer hair cell receptor currents, which in turn results in saturation of a positive feedback, and thus causes a reduction in gain. As a consequence, responses on the basilar membrane are less compressive in the presence of a fixed-level suppressor. Such a linearization should result in a decreasing amount of suppression with increasing level of a suppressed tone. The present study measures how suppression of the response to a tone depends on the level of that tone when the level of a suppressor is kept constant. A forward-masking paradigm was used in which a short 4-kHz probe presented at a fixed level was masked by a longer duration 4-kHz masker. In the suppression condition, the same probe was masked by the 4-kHz masker presented together with a fixed-level 4.8-kHz suppressing tone. In both cases, the level of the masker necessary to mask the probe was estimated using a 3IFC tracking procedure. The task was repeated for different levels of the probe, which allowed for estimating suppression at different levels of the suppressed tone (masker). The results are consistent with a linearization of the BM response in the presence of a fixed-level suppressing stimulus. Different levels of a suppressor led to different slopes of the masking-growth functions. Implications of this result will be discussed. [Work supported by Grant No. DC00683 from NIDCD].