All patients with a discharged diagnosis of schizo-affective disorder (53 cases) admitted to the University of Iowa Department of Psychiatry during 1971-73, were selected for the present study. Schizophrenia (chronic psychotic disorders) and affective disorders (episodic remitting disorders with affective symptoms) were diagnosed among parents and sibs on the basis of chart material. The frequency of schizophrenia was 0.9%; the frequency of affective disorders, 11.8%. These results were compared to data for the relatives of cases of schizophrenia and affective disorders. There were no significant differences for the rates of schizophrenia among the relatives. The rate of affective disorder among the relatives of the schizo-affectives was significantly higher than for the relatives of schizophrenics (3.2%), and not significantly different from the relatives of affective disorder cases (8.3%). An analysis of outcome, precipitants, and age of admission also support the view that schizo-affective disorder should not be routinely classified as schizophrenia for research purposes.