Left ventricular function was studied in 17 patients with ischaemic heart disease and compared with 4 patients with normal left ventricular function. The patients in the homogeneous group of ischaemic heart disease were further subdivided into those 'without angina' (n=5) and those 'with angina' (n=12), depending upon the presence of angina during supine leg exercise at the time of definitive study. At rest there was no significant difference in the heart rate, cardiac output, stroke volume, and left ventricular end-diastolic pressure (LVEDP) in the three groups. During exercise the cardiac output and stroke volume were significantly depressed and LVEDP was significantly raised in the ischaemic heart disease group as a whole but within this group failed to show any significant difference in patients with and without angina. The left ventricular end-diastolic volume (LVEDV) and end-systolic volume (LVESV) measurements showed clear separation of these three groups only on exercise. On exercise, there was decrease in LVEDV and LVESV (P less than 0.05; P less than 0.02) in the group with normal left ventricular function, no change in the group with ischaemic heart disease without angina, and striking increase in LVEDV and LVESV in the group with ischaemic heart disease and angina (P less than 0.01 and P less than 0.02, respectively). This angiographic method of assessing left ventricular function shows clear separation of the three groups and also highlights the significance of angina. Ejection fraction (EF), a commonly measured parameter of left ventricular function, failed to reflect consistent changes on exercise as compared to values at rest which emphasizes the limitations of the measurement of ejection fraction at rest.