The importance of the coronary collateral circulation as a cause of false negative exercise tests was studied in 37 patients who had a 90 percent or greater isolated stenosis of the luminal diameter in one major coronary artery. Sixteen patients had large collateral vessels and 21 patients had either minimal or no collateral circulation. Myocardial scintigraphy was performed in 22 of the 37 patients. The final treadmill time was similar in both groups (521 +/- 192 versus 554 +/- 144 seconds [mean +/- standard error of the mean]). The presence and depth of S-T segment depression was not influenced by the degree of collateralization. The sensitivity of the exercise electrocardiogram was greater for patients with disease in the left anterior descending than in the right or left circumflex coronary arteries (95 versus 60 percent, p < 0.03). Among the 22 patients with thallium-201 scintigrams, myocardial perfusion defects were more common in patients without collateral circulation (100 versus 40 percent, p < 0.01) and involved more myocardial segments (p < 0.005). The overall sensitivity of exercise electrocardiography for the detection of single vessel stenosis 90 percent or greater was similar to that of myocardial scintigraphy (81 versus 73 percent). In conclusion, large intercoronary collateral vessels are not a cause of false negative exercise electrocardiograms in patients with single vessel disease but are a common cause of false negative scintigrams. Large collateral vessels may limit the quantity and location of myocardial ischemia, but exercise electrocardiography and thallium-201 scintigraphy may detect different aspects of ischemia.