Despite some recent favorable findings, there has not been strong empirical support for the validity of DSM-III-R--and now DSM-IV--personality disorder (PD) clusters. In this study, Axis II symptom ratings on 320 personality disordered patients were used to obtain dimensional scores for the 11 DSM-III-R PDs. The dimensional scores for the DSM PDs were subjected to a principal component analysis with orthoginal rotation. Three factors emerged having eigenvalues > or = than 1. The pattern of factor loadings for the individual PDs were not consistent with the DSM Cluster. Rather, the factor loadings were quite consistent with Millon's theory of personality. These results are discussed in light of the clinical benefits provided by employing empirically determined and theoretically anchored model for organizing the Axis II disorders.