A 12 lead electrocardiogram was recorded during treadmill exercise in 57 patients with variant angina in whom coronary angiography was performed. Thirty six patients performed exercise tests with and without calcium antagonists, and 21 performed them only with calcium antagonists. In 55 patients calcium antagonists had prevented spontaneous attacks of variant angina for more than two days before the test. The other two patients were given a single dose of diltiazem (90 mg) two hours before the test. Exercise testing without calcium antagonists induced ST segment elevation with chest pain in nine patients, ST segment depression in 10 (nine with chest pain), and no important shift of the ST segment in 17. Five patients had severe coronary stenosis (greater than or equal to 75%) and all of them showed positive response. Thirty one patients had no important coronary stenosis and 14 of them showed positive response. The sensitivity of the exercise test in detecting a coronary stenosis greater than or equal to 75% was 100% without calcium antagonists but the specificity was low (55%). When the exercise test was done in patients taking calcium antagonists, only two (specificity 96%) of 48 patients without severe coronary stenosis showed positive response (elevation of ST segment in one and depression in another) whereas all nine patients with severe coronary stenosis had a positive response (depression of ST segment in six and elevation in three (sensitivity 100%). It is concluded that exercise testing with calcium antagonists may be a useful method for detecting severe coronary stenosis in patients with variant angina.