Congenital heart disease occurs in approximately 1 per cent of liveborn children, but in a much higher percentage of those aborted spontaneously or stillborn. To detect as many as possible with CHD, including those with mild lesions, very intensive studies are needed. Studies that are not so intensive, especially those done before modern diagnostic techniques were in general use, considerably underestimated the incidence of CHD in liveborn children. It appears that the incidence of CHD and of the various individual lesions does not differ in different countries or at different times, providing the ascertainment of CHD is complete and accurate. The commonest form of CHD is the ventricular septal defect, which occurs in 30 to 40 per cent of all children with CHD. The risks of recurrence in siblings and of transmission to future generations depends on the exact mode of inheritance involved. Approximately 5 to 8 per cent of CHD is due to gross chromosomal abnormalities, and the recurrence risk is that of the chromosomal derangement itself. Because many children with these chromosomal lesions die in infancy or have reduced fertility, the risk to future generations is relatively low. About 3 per cent of CHD is due to classical Mendelian gene effects, with correspondingly high recurrence risks in first-degree relatives. Most CHD has lower risks of recurrence and transmission than those predicted by Mendelian single-gene action. The popular explanation for their inheritance has been the interaction of polygenic effects and the environment, but recent studies of the recurrence and transmission risks of various forms of CHD do not fit this model well. The alternative model is a single gene defect modulated by random events. The recurrence risks for future siblings are 2 to 6 per cent, and for offspring are 1 to 10 per cent, but in a few families the recurrence and transmission risks may be much higher.