Individuals with the rapid cycling form of bipolar disorder represent 13% to 20% of the bipolar population. Although lithium remains the treatment of choice for classic bipolar disorder, failure rates as high as 72% to 82% have been reported for lithium among those who have the rapid cycling variant. Treatment alternatives, including the use of divalproex sodium and carbamazepine, have shown promise for this often treatment-refractory group of patients. Predictors of positive outcome for the acute and prophylactic management of mania with divalproex sodium have emerged; they include nonpsychotic mania, the occurrence of decreasing or stable episode frequencies, mild mania, and mixed states. Predictors for positive acute and prophylactic antidepressant responses to divalproex sodium include nonpsychotic mania, increasingly severe mania, and the absence of borderline personality. Mixed results have been reported for studies using carbamazepine therapy for the treatment of rapid cycling bipolar disorder. Some investigators have reported success with carbamazepine in conjunction with other medications, while others have not. A psychopharmacologic algorithm for the treatment of rapid cycling bipolar disorder is proposed. There is a growing opinion among psychiatrists that patients who rapidly cycle should be treated with an anticonvulsant prior to lithium. However, until homogeneous cohorts of rapid cyclers undergo at least random assignment to different open treatments, these recommendations remain controversial.