Using dual-energy X-ray absorptiometry (DXA) as a standard method for determining body composition in children, we evaluated the accuracy of skinfold-thickness measurements (with the Slaughter et al equations, which are based on triceps and calf skinfold-thickness measurements), bioelectrical resistance (BR; with the Kushner el al equations and age-specific hydration constants), and other clinical measurements (individual skinfold thicknesses and body mass index) for the assessment of body fat in children. We studied a heterogenous group of 49 boys and 49 girls, aged 6.6 +/- 1.4 y and weighing 24.1 +/- 5.9 kg. Fat mass estimated by DXA was significantly lower than fat mass measured by skinfold thickness, even though fat mass measurements by these two techniques were strongly related to each other. Fat mass estimated by DXA was also significantly lower than fat mass measured by BR, and the model R2 and SEE were not as strong as for the skinfold-thickness technique. Fat mass estimated by DXA also correlated with other clinical indexes such as triceps skinfold thickness, body mass index, body weight, and subscapular skinfold thickness. In forward-regression analysis, subscapular skinfold thickness, body weight, triceps skinfold thickness, sex, and height2/resistance estimated the value for fat mass measured by DXA with a model R2 of 0.91 and an SEE of 0.94 kg fat mass. These studies suggest that existing techniques for assessing body fat in children may be inaccurate. We provide new anthropometric equations based on the use of DXA as a criterion that provide accurate and precise measures of body fat and fat-free mass in white children aged 4-9 y. This approach provides estimates of body fat standardized to a known laboratory standard of chemical analysis of carcasses.