A double-blind crossover trial was used to evaluate carbamazepine as the sole maintenance treatment of chronic, nonmanic schizophrenic outpatients whose conditions had been stabilized with the use of neuroleptics prior to study. Criteria of treatment effectiveness included the number of patients relapsing and time to relapse over a 95-day neuroleptic-free period during which either carbamazepine or placebo was administered. Relapse was determined by the concordance of psychiatric ratings and independent clinical judgements indicating significant worsening. Results for 27 patients (13 receiving carbamazepine and 14 receiving placebo) involved in the first phase of this treatment comparison were nondifferentiating. Corroborating descriptive findings in the second phase were available for 14 of these patients. There was no evidence supporting the existence of a treatment-relevant subgroup defined by episodic dyscontrol phenomena.