The influence of physical activity on body mass components has been studied using a sample of Moscow children. 195 girls and 259 boys of Russian ethnicity from 12 to 17 years old were investigated cross-sectionally in 2005 in different Moscow schools. According to the level of physical activity they were divided into three groups: 1-those who did not take part in regular physical exercise (44 boys and 50 girls); 2-those who took part in special sports programs in general education schools (82 boys and 82 girls); 3-students of special sports schools with a high sports ranking (133 boys and 63 girls). The program included anthropometric measurements, evaluation of sexual maturation indices, somatotypes, and "functional" traits (diastolic and systolic blood pressure, pulse rate, hand grip, etc). For the study of body composition, bioelectrical impedance analysis (BIA) was used. The estimates of body mass components were also calculated using the anthropometric measurements. For the fat component, the estimates obtained by BIA and the anthropometric methods were highly correlated: r=0.85-0.88. Age changes of BIA measurements and body components were analysed. With multiple regression analysis it was shown that BIA measurements are dependent on a great number of morphological and functional traits, with the most informative sets of traits being selected. The degree of physical activity has a strong effect on body components: the contents of fat-free mass (FFM) and total body water (TBW) significantly increase, and the fat mass (FM) in girls decreases.