The exponential correction of the QT interval (eQTc) was studied using 12,543 resting electrocardiograms obtained in children (first and seventh graders) who were screened for heart disease in Kagoshima City. The values of eQTc to minimize the effect of the RR interval were estimated using following exponential model, eQTc = (QT interval)/(RR interval)--k. The parameter k = 0.31 was found to be the most useful exponent. The distribution of eQTc values showed a hiatus or break at probability points near p < 0.001 or p < 0.0005. Tentative criteria for an abnormally long eQTc were selected using these points and the following convenient data: 0.430 and 0.435 for the male and female first graders, and 0.440 and 0.445 for the male and female seventh graders, increased by 0.005, respectively. Abnormal values must ultimately be selected using data on patients with the long QT syndrome; however, our findings suggest that the exponential correction of the QT interval may be useful in large-scale screening programs for detecting a QT prolongation in young school children.