Noninvasive monitoring of patency of aorta-coronary bypass grafts can be achieved with reasonable accuracy using the continuous wave directional Doppler and a pencil probe. The character of the graphic record and the auditory signal generated by the flow through these new vessels perfusing the myocardium are distinctive because of both the pattern of flow and also the fixed and relatively immobile position of these grafts in the anterior mediastinum. A total of 226 aorta-coronary bypass grafts were monitored in the first postoperative week: 82 to the right coronary artery (RCA), 90 to the left anterior descending (LAD), and 56 to the circumflex and lateral ventricular branches. Interpretation of patency in these latter vessels was consistently unreliable and was abandoned early in the study. This unreliability was probably due to their short superficial course and their juxtaposition to the aorta and the pulmonary artery. In contrast, flow through grafts to the RCA and the LAD could be established in approximately 90%. Of 82 grafts to the RCA, flow could not be detected in four (5%) and was doubtful in two (2%). In 90 grafts to the LAD, no flow was observed in four (4%) and was doubtful in five (6%). Comparing noninvasive Doppler results with angiography in 34 grafts, no false negatives were found, but there were two or 10% false positives. Flow characteristics in a functioning internal mammary implant were also found to be quite different from those in an intact opposite mammary artery. This form of monitoring requires some practice and experience but is quick, cheap and noninvasive. It offers an acceptable degree of accuracy in monitoring aorta-coronary bypass grafts to the RCA and the LAD when the conduits are in the anterior mediastinum.