In patients with atrial septal defect (ASD), color flow Doppler echocardiography provides visualization of the transseptal jet, the maximal dimension of which can be assumed to correspond to the maximal dimension of the true orifice. To test whether color flow Doppler echocardiography can provide an alternative method for measurement of ASD size, we studied 63 consecutive patients with echocardiographic evidence of ASD. In 48 patients the maximal dimension of the jet was measured in the parasternal, apical, or subcostal four-chamber view or in the parasternal short-axis view. In the remaining 15 patients transesophageal echocardiography was performed because of transthoracic views were inadequate. The transesophageal studies also measured, from two-dimensional images, the maximal transverse discontinuity in the atrial septum. All patients underwent surgical repair, during which the surgeon directly measured the maximal dimension of ASD. Linear regression equations were performed to compare transthoracic and transesophageal dimensions to those measured at operation. Correlation coefficients were as follows for transthoracic versus surgical measurements: r = 0.745, standard error = 4.35, p less than 0.001. Transesophageal measurements derived from both two-dimensional images and echocardiographic jet width showed similar excellent correlation with surgical measurements (n = 0.91, standard error = 4.33, p less than 0.001; and r = 0.919, standard error = 4.42, p less than 0.001, respectively). We conclude that ASD size derived from color flow Doppler echocardiography shows a good correlation with the anatomic maximal dimension observed at operation. Both transesophageal color flow Doppler echocardiography of jet width and direct surgical measurement of the defect provide an accurate estimation of ASD size.