Ultrasound estimation of fetal age is a routine, safe, and valuable diagnostic procedure. The overall accuracy of ultrasound measurement of femur length and biparietal diameter was investigated using a dowel phantom, chicken bones, and a bottle. Objects were imaged in a water bath using commercially available static, linear, and sector scanners. Transducer type and frequency, depth, object dimension, and observer variation were evaluated. Our results show that there was no appreciable difference between transducer types; the overall accuracy of the measurements improved with increasing transducer frequency; the lateral measurements consistently overestimated the object length; the absolute error in lateral measurements was between 0-12 mm; the absolute error in axial measurements was between 0-4 mm; the interobserver variability was +/- 1.3 mm, and intraobserver variability was +/- 1.0 mm. We conclude that transducer type does not affect measurement accuracy, measurement errors are greater in the lateral dimension as opposed to the axial dimension, proper quality assurance and performance evaluation procedures are necessary to ensure optimal results, and femur length measurements will continue to be useful in evaluation of gestational age.