The short-axis area-length method for the estimation of left ventricular mass (LVM) and volume was validated in 24 formaldehyde-fixed canine hearts, using the truncated ellipsoid model. Estimates of the short-axis muscle area were highly correlated with directly measured values (R = 0.92; P less than 0.001). Similarly, LVM calculated using the area-length method showed good correlation with the actual weight of the left ventricle (LVW) (R = 0.85; P less than 0.001). When the regression equation was used, retrospectively, to correct the LVM estimates, the correlation between the actual LVW and the calculated LVM was markedly improved (R = 0.96; P less than 0.001). In awake dogs instrumented for measuring transverse and long axis ventricular dimensions using sonomicrometry, short-axis two-dimensional echocardiograms were used to convert the sonar external transverse dimensions to true diameters. The combined use of the corrected diameters and the regression equation to calculate LVM in the intact dog resulted in values that were highly correlated with the actual LVW (R = 0.95; P less than 0.001). Application of this technique for monitoring LVM in awake instrumented dogs demonstrated the method to be reproducible from day to day and to be sensitive enough to detect serial changes in mass such as during the development of left ventricular hypertrophy.