Coronary angiography provides important anatomic information about coronary artery stenoses. However, it is unclear whether specific assumptions about hemodynamic impact can be made from the angiogram. We therefore studied the relationship between angiographic severity of coronary stenoses and the mean diastolic pressure gradient measured directly at coronary bypass operation. We studied 29 stenoses selected because they were the only lesions in a given vessel, they were well seen in two angiographic views, and there was no change in clinical status of the patient between angiographic and surgical procedures. Fourteen stenoses were in the left anterior descending artery, nine were in the right coronary artery, and six were in the left circumflex artery. Correlation between mean diastolic gradient and percent stenosis was good (r = 0.78, p = 0.001) and especially so for lesions in the left anterior descending artery (r = 0.84, p = 0.001). Lesions over 90% had a wide range of gradients, while lesions less than 90% tended to have more predictable gradients. Collaterals invariably identified vessels with stenoses causing major pressure gradients. History of myocardial infarction was associated with major gradients in supplying vessels. Length of stenosis was not an important influence on gradient over a stenosis. We conclude that in some defined instances, generalizations about hemodynamic (and presumed functional) impact of stenoses can be made from the angiographic assessment, but precise assumptions are not possible.