BACKGROUND: Human body is an irregular geometrical one, so it is very difficult to measure its volume.OBJECTIVE: To establish a multiple regressive equation by taking body height and body mass as the independent variables and body volume as the dependent variable, and select optimally the body volume calculative model of male college students.DESIGN: A single-sample univariate analysis.PARTICIPANTS: Twenty-four male students aged 18-22 years were selected from Zhejiang Gongshang University between January 2003 and October 2004.METHODS: The body height, body mass and body volume of the male students were measured. The indexes of body height and body mass were measured with the nationally-recognized fitness test instrument, and the indexes of body volume were measured with a self-made iron container with a diameter of 0.95 m and height of 1.20 m. There was a mark scale for height in the container, water was infused to a fixed height, then the student slowly immersed himself into the water completely and the height difference was recorded. Body volume (m3)=(0.95÷2)2×3.141 59×height difference. The measured data were statistically calculated. A regression equation in two unknowns was established by taking body height and body mass as the independent variables and body volume as the dependent variable with the systematic software for statistical treatment of physical education scientific research data, and the optimal selection of the regression equation was completed.MAIN OUTCOME MEASURES: The measured data of body height,body mass and body volume and the calculative results of the regression equations were observed.RESULTS: ① A regression equation in two unknowns for calculating body volume was established: (^)y=0.006 16 +0.000 022 ×body height +0.000 756×body mass. ② A regression equation in one unknown for bodyvolume and its optimal selection: The linear equation was (^)y=0.000 8×body mass+0.009 2, the logarithm equation was (^)y=0.050 8 Ln(body mass)-0.152 4, the power equation was (^)y=0.001 8×body mass 0.840 9, the exponent equation was (^)y=0.025 8×e0.012 6x, and the multiple correlation coefficient R2=0.992 1-0.997 3, all were close to 1, indicating that the body volume predicted by models was highly correlated with the actual one [r＞r0.001(24-2) , P ＜ 0.001]. The predicted values of the 4 models had no differ-ence from the actual one. ③ Linear equation had the best simplicity in measurement and calculation in the 5 regression equations. ④ The indexes of physical quality, which were highly correlated with body volume,were body shape, physical function and physical quality.CONCLUSION: The index of body volume is one of the important indexes, which cannot be neglected in the study of physical quality. Analyzing from the simplicity of measurement and calculation and the linear goodness-of-fit of equation, the regression equation in one unknown is the best.