In 1968, the Veterans Administration and the National Institute of Mental Health organized a multi-hospital collaborative study on the effectiveness of lithium carbonate in acute mania. Two hundred and fifty five newly-admitted patients with this disorder were randomly assigned to lithium or chlorpromazine treatment for a three-week period. Patients were classified as highly active or mildly active on the basis of hyperactivity shown at admission. Results have been reported in a previous paper (7). Briefly, it was found that chlorpromazine was significantly superior to lithium in treating the highly active patient: chlorpromazine acted more quickly than lithium, produced significantly fewer treatment failures, and had a lower incidence of severe side effects. The difference between treatments was less clear-cut with mildly active patients, although lithium appeared to be the better treatment. Both drugs effectively reduced manic symptomatology, but lithium produced fewer severe side effects and left the patient less sluggish and fatigued. The poor results with lithium among highly active patients were attributed to the drug's relatively slow onset of action. As a group, highly active patients on lithium showed relatively little change in manic behaviour during the first week; it was ten days before some patients showed any improvement. Almost 40 per cent of the highly active patients had to be taken off lithium before the end of three weeks because of unmanageable behaviour or toxicity. In contrast, chlorpromazine, with its quick onset of action, generally controlled manic symptoms