The effects of aspirin (4.0 g/day) given orally to eight patients with variant angina were observed. An exercise stress test performed in the morning was positive in two of seven patients during placebo administration, whereas a test performed in the afternoon at the same exercise work load resulted in negative findings. During aspirin administration, the afternoon exercise test repeatedly provoked anginal attacks associated with electrocardiographic changes (S-T segment elevation in five and S-T depression in two). Rate-pressure product at the end of the exercise test during aspirin administration was significantly lower than that during placebo administration (p less than 0.01). During aspirin administration, the frequency of angina increased markedly, and the attacks occurred not only during the night or early morning but also in the daytime in six of the eight patients. Our observations suggest that aspirin, in this large dose, reduces the capacity for exercise and provokes exercise-induced coronary arterial spasm in patients with variant angina.