There is no previous scientific evidence to support the anecdotal and self-report literature proclaiming the adverse affect of caffeine on the auditory disorder tinnitus. The present study investigated the relation between these two factors. Three models of caffeine action were postulated, implicating cyclic 3'5'-AMP, adenosine, glucose, adenine and hypoxia. It was predicted that the stimulatory action of caffeine would increase or vary the sensation of tinnitus resulting from excitation anywhere within the auditory system. Sixteen subjects with unilateral versus bilateral or head tinnitus were randomly administered a placebo, 100mg or 300mg of caffeine using a double-blind procedure. Pitch matches to a tone, thresholds of sensitivity, loudness matches to a tone at the matched tinnitus frequency and to broadband noise, and the subjective ratings of tinnitus frequency and loudness were recorded prior to, and 30 or 60 minutes after oral ingestion of caffeine. Caffeine did not affect the pitch of tinnitus or tone threshold but noise thresholds were elevated after 100mg but not 300mg of caffeine and loudness matches to noise increased after both 100mg and 300mg of caffeine. Changes in loudness matches to a tone and subjective ratings of loudness were dose- and time-dependent. The results were discussed in relation to the proposed models and mechanisms of action. Results suggest that for certain tinnitus sufferers caffeine has a stimulatory effect and that the cessation of caffeine-containing beverages would eliminate gross increases or variations in tinnitus. For others, however, the results suggest a depressant action, raising the possibility of using caffeine as a therapeutic, management technique.