Objective: To compare the accuracy of routine ultrasonic and clinical birth weight estimation. Methods: The study sample included 1717 women with singleton pregnancies, admitted in early labor with an ultrasonic estimated fetal weight (EFW) performed during the preceding week. Clinical EFW was obtained before rupture of the membranes by the attending senior resident, who was unaware of the ultrasonic EFW. Accuracy was determined by the percentage error, the absolute percentage error, and the proportion of estimates within 10% of the actual birth weight (birth weight ± 10%). Statistical analysis was done by the paired t test, the comparison of correlated variances, the Wilcoxon sign test, and the χ2 test. Actual birth weight in the study sample averaged 3334 ± 607 g (± standard deviation, [SD]) and ranged between 690 and 5320 g. Results: The means of all error terms of the clinical EFW were significantly smaller than those of the ultrasonic EFW. However, the rates of estimates within 10% of birth weight were not significantly different (72 and 69%, respectively). In birth weights less than 2500 g, both methods overestimated the birth weight, but the mean errors of the ultrasonic EFW were significantly smaller than those of the clinical EFW. The ultrasonic EFW had significantly higher rates of birth weight ± 10% than the clinical EFW (63 compared to 49%, respectively). In the 2500-4000 g birth weight, only the clinical EFW had no systematic error, whereas the ultrasonic EFW underestimated the birth weight. The mean errors of the clinical EFW were significantly smaller and the rate of birth weight ± 10% significantly higher than those of the ultrasonic EFW. In the birth weight greater than 4000 g, both methods underestimated the birth weight, and the mean errors and the rate of estimates within 10% of birth weight were similar for both methods. Conclusion: Clinical estimation of birth weight in early labor is as accurate as routine ultrasonic estimation obtained in the preceding week. In the lower range of birth weight (less than 2500 g), ultrasonic estimation is more accurate; in the 2500-4000 g range, clinical estimation is more accurate. In the higher range of birth weight (greater than 4000 g), both methods have similar accuracy.