This study examines whether the differences in the cluster-analytically generated classes-nonvital versus vital depression-are dimensional (quantitative) rather than categorical (qualitative). To this end, we used various pattern-recognition methods based on principal component analysis (PCA), e.g., display methods (PC plotting), eigenanalysis, and SIMCA (statistical isolinear multiple components analyses). We found several arguments supporting the dimensional hypothesis that the nonvital and vital classes constitute relevant stages (continuous categories) in the continuum of illness-severity. Nevertheless, we found some arguments supporting the categorical hypothesis that the cluster-analytically generated classes are qualitatively different with reference to the similarity of the vital symptoms. Our findings suggest that a nosological or categorical classification is possible from the moment that one component (i.e., the vital component) is quantitatively prominent to the extent that it has become qualitative. As the overall severity of illness increases, vital symptoms emerge which, grouped together, shape a new symptom profile (i.e., vital depression). Thus, our results favor the hypothesis that there are simultaneous quantitative (dimensional: overall severity of illness) and qualitative (categorical: vital symptoms) differences between the nonvital and vital depression groups. © 1990.