To determine the reproducibility of two-dimensional exercise echocardiography, duplicate studies were performed on the same patients a median of 14 days apart. Because measurements are operator-dependent, interobserver variability was calculated for two experienced readers who interpreted the findings independently in a blinded manner. A high degree of interobserver agreement was found in evaluation of both ejection fraction measurements and wall motion abnormalities. Readings for ejection fraction immediately after exercise taken on different days could be estimated within 4% of the values measured in the first test; similarly measured wall motion score index was within 6% of that in the first test. Ejection fractions and wall motion scores were highly correlated between tests 1 and 2. The correlation coefficients between tests 1 and 2 were 0.92 for both the pre- and postexercise ejection fractions and 0.98 for both the pre- and postexercise wall motion scores. Quantitative two-dimensional echocardiography immediately after exercise is highly reproducible, providing a valuable tool for assessing serial changes in left ventricular function. © 1989.