• The validity of the concept of the borderline personality disorder was investigated by comparing assigned diagnoses of 117 consecutive inpatients using Gunderson's Diagnostic Interview for Borderlines (DIB), the Minnesota Multiphasic Personality Inventory (MMPI), the diagnostic criteria of the Diagnostic and Statistical Manual, ed 3 (DSM-III), and the checklist described by Spitzer et al. Twenty-one patients (13 women, eight men) were identified by the DIB as borderline; ten patients (eight women, two men) met DSM-III criteria for borderline personality disorder. A characteristic 8-4-2 mean profile on the MMPI emerged for the DIB-diagnosed borderline patients. The relationship between the DIB and (1) DSM-III diagnosis of borderline, (2) Spitzer unstable type borderline, and (3) MMPI characteristics are consistent, but of a low order of magnitude, thus lending at least partial support to the validity of the borderline construct. Nevertheless, the parsimonious use of the borderline diagnoses by DSM-III when compared with the DIB raises serious questions regarding the usefulness of the borderline concept.