Murray et al (1985) have proposed a method for using biological markers and information about family history to reduce the heterogeneity in a disease such as schizophrenia. They propose that families which are heavily loaded with illness are most likely to be segregating for a major locus and therefore should be used for studies of genetic marker or other biological traits that are thought to be related to a genetic etiology. They propose that patients without a family history (sporadics) of an illness should be investigated for hypothesised environmental components since they are the cases where environmental factors are most likely to play a large role. They give an example from their own data on Ventricular Brain Ratios (VBR) in a sample of schizophrenic twins (Reveley et al, 1984) where schizophrenic twins without any family history (FH-) of a major psychosis had significantly larger ventricles than did schizophrenic twins with a family history (FH+). They conclude that while there is a genetic determinant to ventricular size within the normal range, large ventricles reflect an environmental aetiology that is more important in sporadic cases than in those with a familial pattern to the illness. Subsequently, Reveley & Chitkara (1985) found that singleton schizophrenic patients who were FH-had a significantly larger VBR than did controls while FH+ patients had a mean midway between FH-patients and controls. While this is an attractive hypothesis for the structural brain changes seen in some schizophrenic patients, results of other studies are not consistent with these findings (Nasrallah et al, 1983; Schulsinger et al, 1984; Owens et al, 1985; DeLisi et al, 1986). For example, in our own data (DeLisi et al, 1986), ‘familial’ schizophrenics had larger ventricles than did controls and risk factors thought to be environmental (head injuries and birth complications) were found to be present in that sample.