The authors determined morbidity risks for psychiatric illnesses in the families of probands with schizophrenia-spectrum personality disorders. Subjects were recruited from the community through newspaper advertisements. Subjects were identified as having schizophrenia-spectrum personality disorders (N = 30) if they met at least three, four, or three DSM-III-R criteria for schizoid (N = 14), schizotypal (N = 20), and/or paranoid (N = 15) personality disorder, respectively. The comparison subjects had no psychiatric diagnoses (N = 8) or had other personality disorders (N = 12); none of the subjects in either group had any DSM-III-R axis I diagnosis. Trained interviewers collected family history information about the relatives of the two groups; the interviewers were blind to the probands' diagnoses. The risks for schizophrenia, other functional psychoses, and schizophrenia-spectrum personality disorders were significantly higher in the relatives of subjects with schizophrenia-spectrum personality disorders than in the families of the comparison subjects. The high rate of schizophrenia in the families of probands with schizophrenia-spectrum personality disorders is consistent with the previous findings of higher than normal rates of these personality disorders in the biological relatives of schizophrenic patients. The significance of the high rate of unspecified functional psychoses is unclear. Use of the family study method, by which valid differential diagnosis of psychoses is possible, is indicated. The results from the current study do not rule out the possibility that the schizophrenia-spectrum personality disorders are related to psychoses in general rather than specifically to schizophrenia.