It has been reported that the frequency of congenital heart defects (CHD) in children of an affected parent is now three to five times what it was 20 years ago. One conceivable cause of this is that patients with CHD of a severity that would have precluded parenthood before the advent of treatment are now able to (and consequently do) have children once the defect has been corrected. The mean severity of CHD among parents who have it would have thereby increased and this might, in turn, have increased the frequency of CHD in their children. But the latter would occur only if severely affected parents confer a greater risk on their children than parents who are mildly affected. To determine whether this is the case, we investigated 83 probands, each of whom had had an atrial septal defect secundum whose dimensions had been recorded at the time of operation. Of their 166 children, 17 had CHD. The defects of probands with affected children were not found to be significantly different in size from the defects in probands whose children were all normal. Analysis of the hemodynamic data available yielded a similar result. Our findings decisively exclude surgical correction as a significant explanation for the increased recurrence of atrial septal defect.