We compared the efficacy of unilateral nondominant (n = 15) and bilateral (n = 31) electroconvulsive therapy (ECT) in melancholia in 46 consecutive, nonrandomly assigned medication-free patients with endogenous depression. Seizure duration was recorded and, if a seizure lasted <25 s, the stimulus was immediately readministered. There were more missed or brief seizures that required restimulation with unilateral ECT than with bilateral ECT. After five treatments, blind assessments on the Hamilton Rating Scale for Depression showed a 57% improvement in the bilateral group as compared with a 19% improvement in the unilateral group. The unilateral group received more total treatments (mean = 9.9) than did the bilateral group (mean = 7.7). At the conclusion of all ECT treatments, 72% of patients receiving bilateral ECT were substantially improved and 14% partially improved, compared with 32% of patients receiving unilateral ECT with substantial improvement and 30% partial improvement. This difference between the two groups, however, had disappeared by the time of discharge. Although unilateral ECT may induce fewer side effects, we find bilateral treatments to be more efficacious, at least on a short-term basis.