Pharmacological stress testing may be used in the diagnosis of coronary artery disease when there are contra-indications to the use of conventional exercise protocols. The responses to such testing using arbutamine and to conventional treadmill exercise were compared in eight patients. Respiratory gas analysis and cardiovascular observations were performed during both tests. For an equivalent increment in heart rate, both protocols increased systolic blood pressure and serum lactate. Minute ventilation and oxygen consumption also rose during both protocols, but much more so with exercise. The end-tidal partial pressure of CO(2) [35.1 (S.D. 3. 1) to 30.8 (6.6) mmHg] and the dead space/tidal volume ratio (V(D)/V(T)) [0.37 (0.09) to 0.33 (0.08)] fell significantly during arbutamine infusion, but the respiratory exchange ratio did not change during either protocol. Oxygen pulse, a marker of stroke volume, did not change significantly after arbutamine, but rose markedly after exercise [arbutamine, 3.9 (1.1) to 3.37 (0.7) ml. min(-1).beat(-1); exercise, 4.7 (1.4) to 16.1 (4.6) ml.min(-1). beat(-1) (P<0.0001 compared with baseline); difference between peak responses: P<0.0001]. We conclude that arbutamine simulates some of the physiological responses to exercise, although a number of these responses are less marked than during conventional exercise, in particular cardiac output (oxygen pulse). An increase in ventilation is produced, possibly due to direct stimulation of arterial chemoreceptors. These data suggest that the main action of arbutamine is to increase central drive rather than to establish peripheral demand.