In the past 10 yr the reported incidence of ventricular septal defect (VSD) has increased more than two-fold in two American birth-defects surveillance systems--the Metropolitan Atlanta Congenital Defects Program and the nationwide Birth Defects Monitoring Program. To test the hypothesis that this increase was caused by improved clinical diagnosis, resulting in the diagnosis of minor VSDs, we conducted a follow-up investigation of infants with VSD in the Atlanta area. We compared the rate of spontaneous closure by 1 yr of age of VSDs diagnosed during 1970--72 with the closure rate during 1975--76. If the recently increased incidence was caused entirely by better diagnosis of minor defects that are likely to close spontaneously, we would have expected a much higher rate of spontaneous closure in the later period. In fact, the rate of spontaneous closure by 1 yr was almost identical for both periods (23.8% and 23.9%). The considerable recent increase in the reported incidence of VSD cannot be simply due to better diagnosis.