Previous studies have shown that: (1) raphe neurons respond to acoustic and vestibular stimuli, some with a latency of 10-15ms; (2) alterations of the raphe nuclei alter the acoustic startle reflex; (3) the dorsal raphe nucleus (DRN) is the major source of serotonergic neurons; and (4) approximately 57% of the DRN neurons are nonserotonergic. In the present study, cholera toxin subunit-B (CTB) was injected into cat cochleas, and the brain tissue was examined after a survival period of 5-7 days. Aside from neurons which were known to project to the inner ear, i.e., olivocochlear and vestibular efferent neurons, a surprising new finding was made that somata of a subpopulation of DRN neurons were intensely labeled with CTB. These CTB-labeled neurons were densely distributed in a dorsomedian part of the DRN with some in a surrounding area outside the DRN. The present results suggest that a novel raphe-labyrinthine projection may exist. A future study of anterograde labeling with injections of a tracer in the DRN will be needed to establish the existence of a raphe-labyrinthine projection more thoroughly. A raphe-labyrinthine descending input, together with an ascending input from the inner ear to the DRN through intervening neurons, such as the juxtaacousticofloccular raphe neurons (JAFRNs) described by Ye and Kim, may mediate a brain stem reflex whereby a salient multisensory (including auditory and vestibular) stimulus may alter the sensitivity of the inner ear. As a mammal responds to a biologically important auditoryvestibular multisensory event, the raphe projections to the inner ear and other auditory and vestibular structures may enhance the mammal, sability to localize and recognize the sound and respond properly. The raphe-labyrinthine projection may also modulate the inner ear, s sensitivity as a function of the sleep–wake arousal state of an organism on a slower time course.