Sixteen out of 25 hebephrenic and paranoid schizophrenic patients completed a double-blind cross-over study with sulpiride and haloperidol. The patient sample was relatively chronic: Median age was 35 years (range 26-53 years), median duration of illness 10 years (4-35 years), and median duration of neuroleptic treatment 5 years (1-28 years). Each patient was treated with sulpiride/haloperidol in random order for 12 weeks with a drug-free period before each treatment phase. Mean total BPRS (Brief Psychiatric Rating Scale) score was reduced from 25 to 15 (P less than 0.05) during sulpiride (800-2800 mg/day, median 1600 mg/day), and from 28 to 15 (P less than 0.01) during haloperidol (6-18 mg/day, median 12). There were no significant differences between the groups with respect to total BPRS score, single items or symptom clusters. However, in a few, very "chronic", disturbed, and long-term treated patients, haloperidol appeared more beneficial than sulpiride. Autonomic side effects and parkinsonism tended to occur more frequently during haloperidol than during sulpiride, but no significant differences were found. It is concluded that sulpiride, a specific dopamine-2 receptor blocker, has antipsychotic effect, not significantly different from haloperidol, but may produce slightly less side effects.