We evaluated the correlationship between the predicted defect size at preoperative examination and the actual defect size at operation room, by examining 69 cases of ventricular septal defect operated at the deparment of Thoracic and Cardiovascular surgery, Kyungpook University Hospital from January 1988 to December 1990. We excluded cases associated with other cardiac anomalies. Of the 69 cases, 39 are male and 30 female, forming 1.3:1 sex ratio in males favor. Their age range from 6 months to 16 year, and 4.3 on the average Their body weights are from 6 to 45kg and 15 on the average. According to Soto`s classification, perimembranous type costitutes 42 cases[61%], doubly committed subarterial type 23 cases[33%], and muscular type 4 cases[6%]. The average diameter of defect size is 8.0$pm$3.5mm measured in 2D-echocardiogram, 5.6$pm$3.4mm in angiogram, and 7.4$pm$4.4mm in operative field. There is statistically significant correlation between the size from 2D-echocardiogram and actual defect size[p=0.001], and no significant difference between the two. Especially in the cases without anurysmal formation, they are nearly the same. Cardiothoracic ratio, pulmonary to systemic flow ratio, pressure ratio and resistance ratio also have statistically significant correlation. Main pumonary artery to descending aorta diameter ratio is correlated with the actual defect size. There is statistically significant correlation between the size from angiogram and actual defect size with some difference.