This study was designed to evaluate the accuracy of regression models for menstrual age prediction using single versus multiple fetal ultrasonographic measurements (biparietal diameter, head circumference, abdominal circumference, and femur length). The models tested had been previously developed from a study of middle-class white patients in a private hospital. The current study population consisted of 300 indigent black and Hispanic patients seen in a county hospital. This study demonstrated prospectively that the use of multiple parameters in estimating fetal age offers a significant advantage over any single parameter used alone and that the regression equations developed from a middle-class white population appear to be applicable to fetuses from a population with different socioeconomic and racial characteristics.