OBJECTIVE: Our Purpose was to determine the utility of screening for humerus and femur length shortening in prenatal detection of Down's syndrome (trisomy 21). STUDY DESIGN: On the basis of the biparietal diameter, predicted humerus and femur lengths were derived from a sample control group of 350 karyotypically normal fetuses. The regression formulas were prospectively applied to 45 consecutive fetuses with Down's syndrome and 942 fetuses with normal karyotypes. Ratios of measured humerus length/predicted humerus length and measured femur length/predicted femur length were calculated for each fetus. RESULTS: Eleven of 45 (24.4%) fetuses with Down's syndrome had short humerus (measured humerus length/predicted humerus length ratio of less-than-or-equal-to 0.89), and 11 (24.4%) had short femur (measured femur length/predicted femur length ratio of less-than-or-equal-to 0.91. In comparison, among 942 karyotypically normal fetuses only 43 (4.5%) showed short humerus and 44 (4.7%) showed short femur (p < 0.05). Fetuses with both short humerus and short femur carried an 11-fold greater risk of Down's syndrome (risk ratio = 11.1; 95% confidence interval = 5.0 to 25.0). CONCLUSION: Prenatal ultrasonographic detection of short humerus to short femur, or both significantly increase the risk of Down's syndrome; this information may be useful in screening programs.