抄録
The purpose of the present study was to examine the distribution of subcutaneous fat and to derive several equations to predict body density (BD) using an amplitude modulation type of ultrasonoscope (A-mode ultrasonoscope, FUKUDA FT-100). Subjects were 188 male physical education major college students ranging in age from 18 to 24 years. Fifty subjects who were randomly selected out of the 188 subjects were measured for BD by the underwater weighing technique and were used to derive the equation for estimating the BD. Four points (scapular, triceps, suprailiac, and thigh) of subcutaneous fat which had been commonly selected, height, and weight were measured. The four measurements of fat for the 188 subjects indicated rather small means and small standard deviations respectively. Further-more, histograms of those measurements tended to show a significant skewness for low values and deviated from the normal probability curve (p<0.01). Regarding the means, they were almost all the same except for suprailiac measurements. Suprailiac measurements showed more large values and were distributed rather more widely than the other measurements. Derivations of the multiple regression equations from anthropometric measurements were made using the Wherry-Doolittle test selection method (Clarke & Clarke, 1972). Four measure-ments (triceps, suprailiac, height, and weight) were selected by the Wherry-Doolittle method. The multiple regression equation based on these four measurements was as follows : BD = 1.03894 - 0.00541 1(X1) - 0.000891 (X2) + 0.000506(X3) - 0.000353(X4) · · · · · · · · · Formula(1) (R= 0.660, SEE= 0.00603) where BD : body density (g/cm3), X1 : triceps(mm). X2 : suprailiac(mm), X3 : height(cm), X4: weight(kg), R : multiple correlation coefficient, and SEE : standard error of estimation. On the other hand, a more simple equation based on only fat measurements was as follows : BD = I .10907 - 0.004085(X1 ) - 0.001295(X2) · · · · · · · · · Formula(2) (R=0.631, SEE=0.00624) Correlation coefficients of these formulas were not sufficiently high and SEES of the formulas were not so small. It was shown that derivation of equation to predict BD with high accuracy was not easy when physical education major students were selected as subjects. However, considering the homogeneous nature of the subjects examined in the present study, the present formula could be expected to be applicable when used with groups with similar characteristics.
