A large subgroup of lithium-resistant manic patients are rapid cyclers and as many as 82% of them exhibit poor responses to lithium. Thus, a substantial percentage of poor responses to lithium is accounted for on the basis of rapid cycling. Although controlled trials have demonstrated the efficacy of carbamazepine for the treatment of rapid cycling bipolar disorder, the response to carbamazepine frequently deteriorates. Furthermore, its ability to auto-induce and hetero-induce drug metabolism complicates its routine use. These findings suggest that substantial numbers of rapid cyclers do not respond to either carbamazepine or lithium and that additional mood stabilizers are needed. Our recent findings on 101 rapid cycling bipolar patients continue to support the impression that valproate has marked antimanic efficacy and poor to moderate antidepressant properties. Most patients with mixed states exhibited good antimixed state responses but then became depressed. Predictors of a good antimanic response included decreasing or stable episode frequencies and non psychotic mania. Predictors of a good antidepressant response were non psychotic mania worsening over the years of the illness and absence of borderline personality disorder comorbidity. These open prospective trials, as well as other positive reports of valproate's efficacy in bipolar rapid cycling, await replication with ongoing, controlled maintenance trials.