Patients with moderately severe depressive illness were matched for age, sex, severity of symptoms and number of previous episodes of affective illness. Eleven pairs of patients were randomly assigned to four weeks' treatment with either viloxazine (100 mg three times a day) or imipramine (50 mg three times a day). Both treatments were associated with significant reductions in the scores on depression rating scales. With a sequential testing plan based on repeated significance tests and formed into a skewed restricted plan, no difference in the two treatments was shown. Significantly fewer side effects were reported by patients receiving viloxazine.