Lithium carbonate, haloperidol, and chlorpromazine hydrochloride were compared in a double-blind controlled study with severely ill hospitalized manics. Lithium carbonate and haloperidol produced a highly significant improvement of manic symptoms without sedation. Although producing considerable sedation, chlorpromazine did little to alter the underlying mania qualitatively. Qualitative differences between lithium carbonate and haloperidol indicate that, while haloperidol has a more dramatically rapid impact on behavior-motor activity, lithium carbonate acted more evenly on the entire manic picture, with total normalization realized during active treatment. The majority of lithium carbonate-treated patients met discharge criteria at study termination, but not the patients receiving either neuroleptic drug. The rating scales are not sensitive enough to monitor manic psychopathology; this accounts for the lack of statistically significant differences among drug groups at treatment termination, despite the widely disparate discharge rates.