The cardiorespiratory responses to exercise and forced hyperventilation were measured in 17 unselected patients with syndrome X (angina, positive exercise test, normal coronary arteriogram, no other cardiovascular disease) and compared with those in 15 healthy subjects. Forced hyperventilation produced hypocapnia and metabolic alkalosis but no chest pain or electrocardiographic change. Patients with syndrome X showed reduced maximum oxygen consumption with an increased respiratory exchange ratio at peak exercise, confirming that exercise was limited by skeletal muscle perfusion--and thus that the increase in cardiac output with exercise is limited in syndrome X as in heart failure. Arterial carbon dioxide tension (PCO2) homoeostasis during exercise was normal but the ventilatory cost of carbon dioxide excretion was increased in syndrome X (as in heart failure). End tidal PCO2 measurements correlated only poorly with arterial PCO2 in individual patients with syndrome X, providing a possible explanation for previous reports, based on end tidal PCO2 of inappropriate hyperventilation. Patients with syndrome X did not show inappropriate hyperventilation but they did show hyperventilation that was appropriate to maintain normal arterial PCO2 in the face of reduced cardiac reserve.