Inaccuracies in total intmuterine volumes calculated using the prolate el~ lipse equation have been reported. No previous study has examined all the sources of error. In this study, a comprehensive approach was undertaken. Measurements were obtained from scans of the pregnant uterus in the prone position using an automated water-path scanner (Octoson) and in the supine position using standard static B-mode scanners. Several conclusions could be drawn : 1) From the Octoson prone scans, uterine volumes obtained using the prolate ellipse formula were markedly different from the true uterine volumes obtained by the summation of stepped areas. This showed that the prolate ellipse formula was inaccurate. 2) From the static supine scans, many observer inconsistencies were found in uterine volumes obtained from the prolate ellipse formula. This made the prolate ellipse formula unreliable. 3) Previously published graphs calculated from the prolate ellipse equation, comparing fetal age with total intrauterine volume, were found to vary accuracy, presumably as a result of 1 and 2. A more accurate approach is proposed. Using the outer uterine wall as the boundary, the stepped areato-volume values of transverse scans taken at 3-cm intervals were found to closely approximate true volumes, with an average error ofonly 3.5 per cent. Since these measurements encompass the intrauterine contents and the myometrium , it is suggested that the term •'total uterine volume'" be used instead of "total intrauterine volume."