Multiple gated blood pool (MUGA) and contrast ventriculographic studies were performed within 24 h in 80 patients, 20 with 120 normokinetic wall segments and 60 with wall motion abnormalities in 239 of 360 wall segments. Three methods of evaluation of the radionuclide ventriculograms were compared with the results of the biplane contrast ventriculography which served as a standard: (1) qualitative analysis of the cine mode, (2) analysis of parametric scans (amplitude and phase images) and the phase histogram obtained by Fourier analysis, (3) quantitative determination of regional ejection fraction. Normal values were obtained from 20 patients with normal wall motion in the contrast angiogram. The overall sensitivity for the detection of wall motion abnormalities of high degree was 96% for method 1, 95% for method 2, and 90% for method 3, for those of low degree 72% for method 1, 63% for method 2, and 75% for method 3. Combining methods 2 and 3 the sensitivity was increased to 99% for high grade and to 81% for low grade wall motion abnormalities. The two methods showed a complementary effect because of different sensitivities in dependence of the localization of the wall motion abnormality. Although a high efficiency for the evaluation of left ventricular function was provided by the qualitative analysis of the cine mode the combination of Fourier analysis and the determination of regional ejection fraction should be preferred. It increases the sensitivity for the detection of wall motion abnormalities. Moreover, it offers quantitative data which improve the reproducibility and decrease the observer variability.