Left ventricular stroke volumes derived by two-dimensional echocardiography (2D echo) were compared with thermodilution and cineangiography measurements in closed-chest dogs before andone hour after proximal LAD occlusion. Stroke volume was calculated from end-diastolic and end-systolic volumes reconstructed by two models: 1) Simpson's rule employing left ventricular length and five short-axis cross-sectional areas; 2) a simplified volume formula (V = 5/6 area . length), utilizing a single short-axis area at either the mitral valve or midpapillary muscle level. The comprehensive Simpson reconstruction yielded a good correlation of 2D echo stroke volume against thermodilution (r = 0.89) over a range of normal (N = 14) and ischemic (N = 8) states. The simplified formula provided a satisfactory correlation (r = .90, N = 22) when using the midpapillary cross-section, which encompassed the induced ischemic dys-synergy. In contrast, when using the mitral valve level cross-section above the site of ventricular asymmetry, there was no significant statistical correlation. Comparison of cineangiography with 2D echo volume reconstruction based on the simplified formula with the midpapillary muscle level section yielded good correlations for stroke volume (r = 0.87) and ejection fraction (r = 0.97). Intraobserver and interobserver variability of duplicate echo stroke volume measurements was 8% and 10%, respectively. We conclude that 2D echocardiography in dogs permits quantitation of left ventricular stroke volume in normal and ischemic states.