In his 1933 article, Bruno Schulz reported a follow-up and reanalysis of the schizophrenic probands and their relatives first studied by Rüdin 20 years earlier that sought to clarify whether schizophrenia was a valid "unit-character" for Mendelian genetic analysis. He proposed a range of subgroupings of probands, particularly traditional subtyping, presence or absence of identifiable causal influences, and outcome. He then compared those subgroupings in several ways, most commonly by the risk for schizophrenia in their siblings and by the level of resemblance among proband-sibling affected pairs. Of his many findings, those of greatest interest included (1) probands with possible and probable physical causes, particularly those with head trauma, had substantially lower risk of illness in siblings, (2) probands with a hebephrenic subtype had a striking elevation of risk for schizophrenia in siblings, (3) probands with psychological causes had higher rates of good outcome, (4) proband-sibling pairs resembled one another for the classical schizophrenic subtypes, and (5) an absence of any cases of schizophrenia in siblings of a small group of schizophrenic probands with birth complications, convulsions, and skull deformities. Schulz used this sample in a fundamentally different way than Rüdin. Rather than seeking for Mendelian transmission patterns, Schulz used family data to evaluate hypothesis about clinical/etiological heterogeneity, thereby presaging many subsequent family studies of psychiatric disorders. While Schulz did not claim to have proved the etiologic heterogeneity of schizophrenia, he raised important questions, still unanswered, about whether schizophrenia is a legitimate "unit-character" appropriate for genetic analysis.