Recent reports have challenged the assumption that unilateral and bilateral electroconvulsive therapy (ECT) are equally effective in all cases. In a prospective study, hospitalized manic patients were randomly assigned to treatment with either lithium or ECT. The first six patients received nondominant unilateral treatment but failed to respond. However, they rapidly remitted when switched to bilateral ECT. Accordingly, the next 11 manic patients to receive ECT were assigned to bilateral treatment. They were compared with the first six patients on weekly ratings of mania and other psychopathology. Both ECT groups improved but therapeutic responses were delayed until after switchover to bilateral treatment in the patients who began with unilateral ECT. We conclude that patients in the manic state should receive bilateral rather than unilateral ECT.