The main reason for the inconsistent findings in schizophrenia research is the lack of diagnostic conformity. This has not changed markedly following the introduction of modern operational diagnostic systems. Taking schizophrenia as a disease entity or assuming schizophrenia spectrum psychoses to represent a continuum of diseases without any clear dividing lines, the results of family and twin studies point to a multifactorial etiology based on a polygenic mode of transmission. Further, then it has to be assumed a familial continuum from schizophrenia to affective psychosis and other spectrum disorders. However, in family and twin studies based on Leonhard's classification, there is clearcut evidence that schizophrenic spectrum psychoses have to be divided into clinical and etiological subgroups with a completely different genetic background. For example, systematic catatonia is, for the most part, a sporadic disease, whereas periodic catatonia aggregates in families in a manner consistent with a major gene effect. Further, the results indicate that schizophrenic spectrum psychoses consist of three main valid categories: cycloid psychoses, unsystematic schizophrenias and systematic schizophrenias. In the case of cycloid psychosis and systematic schizophrenias, genetic loading seem to be very low, while "environmental" factors, for example, birth complications, may play an important etiological role. Unsystematic schizophrenias, however, are predominantly inherited and "environmental" factors are not very prominent.