A theoretical model should explain all of the observed facts. Both the model of schizophrenia which proposes many genes of small effect (Reference McClellan, Susser and KingMcClellan et al, 2007) and that which proposes few gene mutations of large effect (Reference Craddock, O'Donovan and OwenCraddock et al, 2007) explain many of the observed facts. In particular, the model of many genes of small effect explains the observed spectrum of mental illness, from bipolar disorder, through schizoaffective disorder to schizophrenia.

We have both observed schizophrenia occurring in particular families. One of us (M.A.) has studied a number of South-Asian families with multiple members with schizophrenia. In some of these families, patients of a second generation developed the illness at a much younger age than their parents and their illness was more severe. Thus far we have assumed that these observations were related to the concentration of many genes of small effect within these families.

One of us (M.B.-P.) has also observed the same effect, known as anticipation, in a group of families in Slovenia. Thirty-six parent–offspring pairs with schizophrenia were studied. First hospital admission was used as a proxy for disease onset. In the off-spring group, mean age at onset was identified as 23.5 years whereas this was 39.6 years in the parent group. There was a higher mean total number of days of hospitalisation in the first 5 years of treatment in the offspring group (223 v. 161), and a higher mean number of hospitalisations over the same period in the offspring (7.27 v. 7.51) (both results statistically significant). These two measures were used as a proxy for increased intensity of illness. The offspring had a higher level of education but demonstrated fewer working years and had fewer children (Reference BlincBlinc, 2002).

What arises is the question of whether the ‘many genes of small effect’ or the ‘few genes of large effect’ model is best suited to explaining this observation of anticipation of schizophrenia.