To evaluate the clinical potential of sulpiride for the treatment of schizophrenic patients, a double-blind study was performed comparing fixed doses of sulpiride (800 mg daily) and chlorpromazine (400 mg daily). Twenty-five schizophrenic (RDC) patients participated in each treatment group. Antipsychotic effects were evaluated by CPRS and NOSIE ratings before and after 1, 2, 4 and 8 weeks of treatment. Interrater reliabilities for CPRS items and subscales were satisfactory. Treatment with sulpiride or chlorpromazine resulted in a significant reduction of psychotic morbidity as estimated by CPRS and global ratings. CPRS scores reflecting autism were significantly reduced in all ratings of sulpiride-treated patients, but only after four weeks in the chlorpromazine group. Total NOSIE scores indicated improvement in both treatment groups. A significant difference in favour of sulpiride was obtained for the NOSIE subscale "retardation". Extrapyramidal side effects occurred at a similar frequency in both treatment groups. Autonomic side effects occurred to a greater extent in chlorpromazine-treated patients. Lactation was reported only in four sulpiride-treated patients. Liver transaminase enzymes in serum were markedly elevated only in chlorpromazine-treated patients. The results indicate that sulpiride has a marked antipsychotic effect which is at least not inferior to that of chlorpromazine. A better effect on autistic components of behaviour was demonstrated for sulpiride. The results indicate a higher risk of lactation but a lower risk of anticholinergic side effects and liver toxicity for treatment with sulpiride than with chlorpromazine.