Abstract
Data collected by a single observer on 147 schizophrenic patients were subjected to clustering analysis. The results produced the hypothesis that schizophrenic illnesses directly after childbirth are a separate disease entity. This hypothesis was not disproved by experimental testing. Several disease entities may be included in the term schizophrenia. If this is so, the methods used in generating and testing the hypothesis that puerperal schizophrenia is a separate disease may provide a systematic method of classifying the various illnesses.