The authors undertook a chart review to examine the use of the diagnosis of schizophrenia, schizo-affective type, in clinical practice. Of 27 patients given this diagnosis over a 3-year period, 13 were found to have evidence of bipolar course in their illnesses. For both the bipolar and unipolar groups, the most striking finding in first-degree relatives was the prominence of affective conditions. The bipolar group had a statistically significant earlier age for first psychiatric treatment and previous number of hospitalizations. Symptoms noted on admission were mostly affective, and the schizophrenic symptoms reported were noted by authors to be considered overinclusive or unreliable by many clinicians. Both groups received treatment with antipsychotic and antidepressant medication. Six of 13 bipolar patients and no unipolar patients were treated with lithium carbonate. Five bipolar patients met research criteria of Feighner et al. (1972) for primary affective illness and another met the criteria for schizophreniform illness. One unipolar patient met criteria for probable primary affective illness and another met the criteria for probable schizophreniform illness. The authors concluded that the diagnosis of schizo-affective illness, as used in day to day clinical practice, does not identify a group of schizophrenic patients nor a homogeneous patient group, and when both affective and schizophrenic features appear in a patient with a bipolar illness, the diagnosis of manic-depressive illness, not schizophrenia, should be given first consideration.