Two exercise tests were performed with an intervening rest period of 45 minutes in a group of 13 subjects with previously identified exercise-induced ventricular arrhythmias and no resting arrhythmias. Both normal subjects and patients with heart disease were included in the group. The level of stresss was equal in both tests as judged by similar rate-pressure products at peak exercise. There was a significant decrease (P less than 0.05) in the number of VPCs induced by exercise during and after the second test. When the number of VPCs on test I and test II in the same patients were compared, a regression line fitted the data well (r = 0.92). Analysis of the recovery periods revealed significant (P less than 0.01) decreases in systolic blood pressure at one and three minutes post exercise, comparing the second to the first test. The underlying mechanism may be decreased myocardial oxygen demand during the second test as the lowered rate-pressure products during recovery (P less than 0.01) reflect. The results of this study indicate that tests of effectiveness of an antiarrhythmic drug should not be based solely on a decrease in the amount of severity of ventricular irritability between two successive exercise tests, one immediately before and the other following administration of the drug.