The effect of exercise on the development of intercoronary collaterals and on left ventricular function is controversial. Twenty male patients (mean age 48 years, range 36-54 years) who had suffered an acute myocardial infarction were randomly allocated to an exercise group (10 patients) and a control group (10 patients). Both groups underwent coronary angiography, left ventricular function studies and myocardial perfusion studies with labeled microspheres, before and after the 7-month experimental period. Both groups had similar extent of disease as measured angiographically and both had mild progression of disease. Neither group showed changes in extent of callateralization, myocardial perfusion or left ventricular function. The exercise group had a significant increase in anginal threshold and a significant (p less than 0.01) decrease in heart rate at a given work load. Exercise, therefore, does not appear to affect progression of disease, myocardial perfusion, extent of collateralization, or left ventricular function in patients with coronary artery disease.