To validate the accuracy of human left ventricular (LV) volume measured by the conductance catheter method, conductance volume was compared with LV volume measured by biplane angiography in 19 patients with ischaemic heart disease. Angiographic LV volumes were calculated frame by frame and matched with instantaneous conductance volumes. Calibration was determined by both gain constant, 1/alpha, and parallel conductance, alpha V c. The gain constant, 1/alpha, was the ratio of stroke volume measured by the thermodilution method to that measured by the conductance catheter. The parallel conductance, alpha Vc, was estimated by the saline injection method. There was a good correlation between the corrected conductance volume (Vcc) and the angiographic volume (V angio) such that Vcc = 0.94 V angio + 5.3 ml, r = 0.94, P < 0.001. There were no differences in the correlation coefficients between normal and depressed hearts. Compared with the angiographic data, LV end-diastolic volume, end-systolic volume and ejection fraction were determined accurately by the conductance catheter. We conclude that the conductance catheter method, corrected for gain constant and parallel conductance, can accurately and continuously estimate the LV volume throughout the cardiac cycle in the diseased human heart.