1. This study examines the effect on auditory desensitization of electrically stimulating the crossed olivocochlear bundle (COCB) at the floor of the fourth ventricle. Auditory desensitization was induced by a loud high-frequency pure tone exposure and measured as temporary threshold shifts (TTS) in the sensitivity of the compound action potential recorded from the cochlea. COCB stimulation simultaneous with the loud sound exposure reduced the TTS. This reduction was contingent on the COCB stimulus being presented as a continuous burst for the entire duration (1 min) of the exposure. 2. The reduction in TTS could be abolished by prior administration of strychnine. The action of strychnine on these TTS effects of continuous COCB stimulation paralleled its action on the classical COCB effects elicited by pulsed short COCB trains. If the action of strychnine on the classical COCB effects was allowed to reverse, then continuous COCB stimulation reduced TTS as effectively as before. 3. The most effective COCB stimulus was found to be one that was presented at a high rate of stimulation simultaneous with the exposure. The COCB effect on TTS was also found to be a tonic one; smaller but significant reductions in TTS could still be obtained with the exposure presented 5 min after COCB stimulation though not when the delay was 10 min. The tonic reductions in TTS appeared to occur without any persisting changes at the cochlea. Normal cochlear responses remeasured in the delay between the stimulus and exposure were not altered. 4. It was hypothesized that the persisting effect responsible for TTS reductions did not occur at the cochlea but at some central site facilitated by antidromic action potentials along the COCB fibers. Subsequent exposure to loud sounds would activate the central site primed by the prior COCB stimulus. This hypothesis was tested by stimulating the COCB alone as before, but then lesioning the fibers before presenting the exposure. Persistent cochlear effects of the COCB stimulus should have still resulted in a reduction in TTS. However, if the persistent effect was at a more central location, lesioning the fibers would allow afferent input to act at the facilitated central location but would not allow subsequent expression of COCB effects at the cochlea. In this case, no reductions in TTS could be expected--precisely the results that were obtained in these experiments. Thus the COCB system appeared to have a "memory" component facilitated by prior stimulation and activated by a subsequent exposure.