The preliminary analyses of our data from the twins of schizophrenics ascertained in 16 years' consecutive admissions to a psychiatric clinic, together with our re-analysis of the earlier major twin studies, suggest that we are dealing with replications of the same experiment. Differences among the studies are more apparent than real. One important source of heterogeneity is the unequal representation in a sample of the two sexes. Higher female concordances in two particular studies (Rosanoff and Slater) could be an artefact of sampling; or it could be associated with more environmental variability for males or with some aspect of the process of identification. Another source, perhaps the most significant, relates to the severity of schizophrenia in the proband. Chronic schizophrenics residing in or admitted to long-stay hospitals show the highest concordances. The issue of sampling from the resident population versus consecutively is not as crucial as the type of hospital. Kallmann's twins were mostly from consecutive admissions, and an analysis of Slater's series according to whether the index case was "resident" or "consecutive" showed no difference in concordance (Gottesman and Shields, 1966). Table 6 presents a summary of the 11 major twin studies of schizophrenia analysed in terms of the important dimensions; a [SEE TABLE VI IN SOURCE PDF.] simultaneous consideration of the effects of these dimensions on concordance rates has led us to the conclusion above. All the concordance rates cited are without age corrections and in terms of pairs, not index cases.
A third kind of heterogeneity is that between studies in the Scandinavian peninsula and adjoining Finland and those conducted thus far in the rest of the world. In the former (Essen-Möller, Tienari, and Kringlen), MZ probands tended to be matched with co-twins who presented eccentricities of character which we might label as schizophrenic-like, borderline, or schizoid. The possibility arises that there may be real reasons, genetic and/or environmental, why "concordance" in Scandinavian twin studies does not show up at the level of a clear cut psychosis. There is, however, no reason to expect concordance rates to be the same throughout the world, even in studies using the same design. Rates can vary according to the frequency of genes and their penetrance (Huizinga and Heiden, 1957). The continuing evolution of populations could have led to a genotype more adapted by natural selection to harbouring the genetic component of schizophrenia (Gottesman, 1965). There are many climates and cultures where twin studies have yet to be carried out.
We shall correct the de-emphasis on environmental factors in schizophrenia in future papers. Unfortunately, such factors as have been implicated in the literature have not proved to be specific for schizophrenia. From the fact that the identical twin of a schizophrenic is at least 42 times as likely to be schizophrenic as a person from the general population, and a fraternal twin of the same sex 9 times as likely, it would appear that genetic factors are largely responsible for the specific nature of most of the schizophrenias. Note that postulating a genetic specific aetiology for schizophrenia means only that the gene or genes are necessary, not that they are sufficient, for the disorder to occur (Meehl, 1962).
If we may be permitted a forecast, it would be that the mysteries of the schizophrenias will be solved in our lifetimes. The solution will come from the syntheses of research findings in such seemingly disparate areas as twins reared apart (Shields, 1962; Juel-Nielsen, 1965), fostered and non-fostered children of schizophrenics and normals, population genetics, biochemistry, and molecular and developmental biology.