The distinction between personality disorder diagnoses is often unclear because of the frequent cooccurrence of one or more of the diagnoses. To date, studies using sample sizes large enough to evaluate the rates of cooccurrence in the less prevalent personality disorders have not been conducted. The Structured Interview for Diagnosis of Personality-Revised, a semistructured instrument designed to yield reliable personality disorder diagnoses, was used to evaluate 1116 subjects for the presence of DSM-III-R personality disorder diagnoses. Cooccurrence rates and odds ratios were calculated for each pair of diagnoses. There was a high degree of cooccurrence between the cluster A personality disorders, between the cluster C personality disorders, and between narcissistic, borderline, and histrionic disorders. In addition, there was also a high frequency of avoidant personality disorders within the cluster A personality disorder diagnoses. Possible explanations for the high degree of cooccurrence and the potential implications for the DSM-IV personality disorder diagnoses are discussed.