Ninety-two middle-aged and elderly patients with disabling tinnitus participated in a double-blind randomized clinical trial comparing nortriptyline (a tricyclic antidepressant) to placebo. The study was stratified for presence (n = 38) or absence (n = 54) of current major depression (by DSM-III criteria). Both active drug and placebo were given for 6 weeks following a dose adjustment phase; the median nightly dose of nortriptyline was 100 mg. The two primary outcome variables were global satisfaction questions: "Has the medication helped you in any way?" and "Has your tinnitus improved?" Sixty-seven percent of nortriptyline patients stated the drug had helped them, versus 40 percent of placebo patients (chi-square = 7.14, p = 0.008). However, tinnitus severity was not significantly affected by nortriptyline (active: 43%; placebo: 30%; chi-square = 1.567, p = N/S). Benefit was more likely to be reported by depressed patients, by patients with insomnia, by women, and by patients without cervical musculoskeletal disease. Nortriptyline is useful in some patients with disabling tinnitus, but has not been shown to directly affect tinnitus sensation. Placebo effects were strongly significant and must be considered important in tinnitus therapy. It is difficult to specify the most appropriate outcome measures for tinnitus therapeutic trials.