Fifty-three patients with acute exacerbations of Research Diagnostic Criteria schizophrenic, schizoaffective (mainly schizophrenic), and other nonaffective psychoses completed 24 or 28 days of treatment with randomized, fixed, double-blind doses of 10, 20, or 30 mg of oral fluphenazine hydrochloride daily. In the sample as a whole, improvement was not predicted by dose but was negatively related to duration of illness and of lifetime hospitalization, and to the presence of akathisia during the study (which was unrelated to chronicity). But among patients showing 40% or greater improvement in positive symptoms, percent improvement was predicted by dose and dose per kilogram of body weight; this was not the case for negative symptoms. Severity of acute extrapyramidal symptoms (excluding acute dystonia, dyskinesia, and akathisia) was significantly correlated with dosage per kilogram. Doses greater than 0.2 mg/kg per day were associated with greater clinical improvement but also with a high incidence of extrapyramidal symptoms; doses over 0.3 mg/kg per day were associated with more severe extrapyramidal symptoms. These preliminary results suggest that there is a linear relationship between fluphenazine dosage and acute outcome, and that this relationship is observed in patients whose conditions improve to a criterion level. It is suggested that the nonresponder group may include many patients in whom dose is not relevant because they are unable (for a variety of reasons) to respond to the study treatment conditions; excluding them from analysis may allow a significant dose-response relationship to be observed. Akathisia deserves further study as a possible predictor of nonresponse.