The effect of halothane on left ventricular (LV) function in healthy, nonmedicated patients before operation was studied. Using local anesthesia, with the aid of fluoroscopy, catheters were placed in the left ventricle, thoracic aorta and pulmonary artery. Thermodilution cardiac output derivatives and left ventricular pressure indices were measured awake and during anesthesia with halothane, 0.9 and 1.8 per cent (end-tidal). Ventilation was controlled to maintain Pa(CO 2) and pH(a) at the awake values. Dose-dependent decreases in cardiac index (3.98 → 3.48 → 2.86 l/min.m 2), stroke volume index (50.1 → 44.6 → 38.5 ml/m 2), LV dP/dt(max) (1,440 → 1,200 → 1,000 mm Hg/sec), dP/dt(max)/IP (20.7 → 17.0 → 15.0 sec -1), cardiac work (11.76 → 8.15 → 5.52 x 10 2 mm Hg/ml/min.kg), and mean aortic pressure (MAP) (94 → 80 → 69 mm Hg) were accompanied by increased left ventricular end-diastolic pressure (LVEDP) (11 → 12 → 14 mm Hg), without change in heart rate or systemic vascular resistance (SVR). Preload (LVEDP) and afterload (MAP and SVR) changes could not account for the demonstrated depression in ventricular function. Thus, in man, as in the dog, halothane has a direct depressant effect on LV function.