T wave and ST segment abnormalities in 20 asymptomatic men aged 18 to 55 were investigated because they were identical with myocardial ischaemic changes, and the professional livelihood of the subjects was jeopardised. Coronary arteriograms showed unobstructed arteries in all except one in whom a 50 per cent lesion of the left anterior descending artery was present. Left ventricular angiograms showed a normal contraction pattern, Ejection fractions were normal in 12 and increased in 8. Three characteristic electrocardiographic patterns were observed: flat or inverted T waves in leads II, III, aVF, and V4 to 6 designated type 1; deep T inversion particularly evident in leads V2 to 5 designated type 2, and minor ST segment depression in the inferior and lateral leads without T changes designated type 3. Characteristically, type 1 changes were temporarily suppressed by either beta-blockade or an overnight rest and were more abnormal in the standing position. Type 2 and 3 changes were relatively uninfluenced by these manoeuvres. Maximal treadmill exercise tests were positive in 6 and borderline or negative in 14. When repeated after oxprenolol all tests were negative. Echocardiograms showed asymmetric septal hypertrophy in 3 subjects (ratio of greater than 1.5 between ventricular septum and posterior left ventricular wall). After normalisation by an overnight rest, type 1 T wave abnormalities were reproduced by intravenous adrenaline infusion (0.024 to 0.18 microgram/kg/min) but not by noradrenaline or by adrenaline after prior administration of oxprenolol. When the T waves had remained deeply inverted before infusion despite rest (type 2) adrenaline infusion normalised them and again noradrenaline was without effect. This effect was also prevented by oxprenolol. Type 3 changes were uninfluenced by catecholamine infusion. Plasma catecholamine estimations suggest that catecholamine hypersecretion and hypersensitivity may both be relevant, particularly the latter. The apparent bimodal response of the ventricular myocardium to adrenaline infusion is not surprising since in vitro experiments suggest that reversal of T wave polarity in either direction may result from summation of changes in action potential duration in different parts of the heart. Such changes may be unimodal, that is both areas involved show lengthening or shortening of action potential duration, but by occurring at different rates may lead to a bimodal change in the differences in duration which generate the T wave.