Magnetic resonance imaging is a new technique for the investigation of cardiac anatomy and function, but few attempts have been made to assess its value in the evaluation of coronary artery disease because of the problems of performing exercise in the magnet. In this review, we describe our experience using dobutamine as a substitute for exercise. We have assessed regional left ventricular wall motion by cine magnetic resonance imaging and global function from aortic flow and acceleration measurements using cine magnetic resonance velocity mapping. We studied regional wall motion in 23 patients presenting for coronary angiography because of chest pain. Of 21 with reversible dobutamine-induced thallium perfusion defects, 20 (95%) had dobutamine-induced wall motion abnormalities corresponding closely in site and size. We studied aortic flow in 25 similar patients and we compared the results with five normal volunteers with a low likelihood of coronary artery disease. A failure to increase peak aortic flow acceleration (peak, rate of change of flow) by 15 L/s2 was the most significant parameter associated with coronary artery disease. A reduction in peak flow acceleration at peak stress occurred in patients with moderate or severe myocardial ischaemia as determined by thallium myocardial perfusion imaging. Dobutamine is a valuable pharmacological agent which can induce regional and global abnormalities of left ventricular function in patients with coronary artery disease during magnetic resonance imaging.