Objective - To evaluate the effects of positioning and number of repeated measurements on intra- and interobserver variability of echocardiographic measurements in dogs. Design - Prospective study. Animals - 4 healthy dogs. Procedure - Each observer performed 24 examinations, separately assessing each dog 6 nonconsecutive times (3 times with the dog in lateral recumbency and 3 with the dog in a standing position). Variables evaluated included M-mode measurements of left ventricular end-diastolic and left ventricular end-systolic diameters, left ventricular free-wall thickness in diastole and systole, interventricular septal thickness in diastole and systole, left ventricular shortening fraction, and 2-dimensional measurements of the left atrial diameter-to-aortic diameter ratio. Results - All coefficients of variation (range, 3.4% to 26.6%) were similar between operators and positions and were <15% for 27 of 32 values. For both operators, repeatability of the measurements was better for left ventricular end-systolic diameter, left ventricular free-wall thickness in diastole, left ventricular free-wall thickness in systole, and the left atrial diameter-to-aortic diameter in the standing position, and similar for both positions for shortening fraction and left ventricular end-diastolic diameter. No effect of cardiac cycle was observed. Conclusions and Clinical Relevance - Within-day variability of conventional echocardiography performed with the dog in the standing position was at least as good as that obtained with the dog in lateral recumbency for most measured variables. Single measurements of each variable may be sufficient for trained observers examining dogs that do not have an arrhythmia. The standing position should be used, particularly for stressed or dyspneic dogs.