Schizophrenic patients in long-term neuroleptic monotherapy with clozapine (n = 100) and perphenazine, flupenthixol or zuclopentixol (controls, n = 100) were evaluated for extrapyramidal side effects (EPS) (blind) as well as other side effects and mental condition (non-blind). In both groups the patients had received neuroleptic treatment for a total of 14 years (median) and the present antipsychotic (clozapine or control drug) for 5 years. Thus the clozapine-treated patients had previously received traditional neuroleptics for 9 years (median). The study was both retrospective (0.3-19 years for clozapine, 0.3-24 years for control drug, by means of chart information) and prospective (1 year, with video-controlled evaluation of EPS). There was a significantly lower prevalence of tardive dyskinesia (TD) in clozapine treated patients than control patients, although prior to this treatment there were more TD patients in the clozapine group (P < 0.05). This lower level of TD in the clozapine group was related to a lower induction of new cases (P < 0.001) and a tendency towards greater disappearance of TD in the clozapine than in the control group (P = 0.07). Clozapine treated patients without TD had started clozapine and ceased traditional neuroleptics at an earlier age than those with TD. Parkinsonian signs were seen in 33% of the clozapine patients versus 61% of the control patients, mainly as hypokinesia; tremor in 3% versus 11% and rigidity in 0 versus 19%. Psychic akathisia was found in 14% versus 40% and motor akathisia in 7% versus 29% of the patients, all differences significantly in favor of clozapine. Clozapine treated patients also had less neuroleptic-induced emotional indifference and depression, but more autonomic side effects than controls.