Sonographic measurements often give considerable diagnostic significance. However, large discrepancy between the real anatomic sizes and sonographic measurements could cause an incorrect diagnosis. Authors carried out a series of experiment using a phantom and two scanners and the results are as follows : 1. There is a little difference in errors between two scanners, and the measurement does not almost always correspond to the real distance. 2. The rates of errors of the measurements are almost fixed regardless of the positions of the reflectors. 3. There is no relationship between the transducer MHz and error rate. 4. As the error is derived from the scanner electronics it isnecessary to calibrate the scanner periodically, and if not calibrated one should consider the error rate of the scanner inthe anatomical measurement .