Antipsychotic dosing for acute mania has not been well studied. Combined treatment with lithium and an antipsychotic is the most common treatment, but additional antimanic efficacy of a lithium-antipsychotic combination beyond that of an antipsychotic alone has not been well demonstrated. Furthermore, the possibility that lithium could affect antipsychotic dose requirement is believed to have never been studied. In this study, 63 acutely psychotic bipolar manic inpatients were randomly assigned to receive double-blind treatment with 1 of 2 haloperidol doses, 25 mg/day or 5 mg/day, for 21 days. In addition to haloperidol, subjects were randomly assigned to receive concomitant treatment with placebo, standard lithium, or lorazepam 4 mg/day. The high haloperidol dose produced greater improvement and more side effects than did the low dose. Lithium added to the low dose produced a markedly greater clinical response than did the low dose alone. Lorazepam did not improve the outcome for the patients receiving low-dose haloperidol. The clinical response produced by high-dose haloperidol was not enhanced by adding either lithium or lorazepam. All treatment effects emerged by the fourth day of treatment and persisted. Used alone, a haloperidol dose of 5 mg/day is too low for most manic patients, but concomitant lithium produces a dose-dependent enhancement of haloperidol response. Lorazepam 4 mg/day was insufficient to produce an advantage when added to low-dose haloperidol.