A preliminary analysis of twins or triplets with heart defects, ascertained in five centres, confirms earlier suggestions that monozygotic (MZ) twins are over represented among twins with heart defects, even after excluding persistent ductus arteriosus and conjoined twins. An MZ twin individual has a risk of cardiovascular malformation approximately twice that of DZ twins and singletons. It is suggested that the twinning process itself affects one of the pair. Disturbance of laterality ('mirror imaging') is probably a more important mechanism than twin-twin transfusion. Inappropriate use of the twin method in the past has caused the importance of genetic factors in the etiology of congenital heart defects to be underestimated. Nevertheless, twins do provide a useful illustration of the likely importance of epigenetic factors in heart development.