188 A BIOMETRIC STUDY OF BASAL METABOLISM IN MAN. 



Now, if we return to the differences in these three tables and con- 

 sider together the three criteria of excellence of prediction each of 

 which has some advantages but neither of which is perfect as a basis 

 for a generalization concerning the value of the two methods under 

 consideration, we note the following points : 



1. The results in the first difference column show that prediction 

 from the two direct measurements stature and body-weight gives more 

 accurate results than the method of calculation from body-surface 

 area by the Du Bois height-weight chart heretofore employed. 



2. The second difference column suggests that when the more 

 accurate method of prediction by means of linear regression equations 

 suggested in this volume is substituted for the old method slightly more 



TABLE 72. Comparison of square root of mean-square deviation (in calories) from the actual caloric-output, 

 of heat-production, calculated on the one hand from multiple regression equations involving body-weight and 

 stature and on the other from (a) the mean heat-production per unit of body-weight and of surface by the 

 Du Bois height-weight chart and from (b) the regression of total heat on body-weight and on surface area 

 by the Du Bois height-weight chart. 



* The differences in these columns are obtained from the first column of this table and the entries of the pre- 

 ceding tables as follows: column II from III of table 62; column III from III of table 68; column IV from I of 

 table 62; column V from I of table 68. 



accurate predictions may be made from body-surface area than from 

 multiple regression equations involving height and weight. 



3. The third difference column shows that practically without 

 exception (25 out of 27 tests) better prediction can be made from 

 multiple regression equations than by considering heat-production in 

 the individual as given by (body- weight X mean heat-production per 

 kilogram in the control series). 



4. Even when the superior method of predicting from the regression 

 of heat-production on body-weight introduced in this paper is employed 

 instead of the older method, the multiple regression equation in which 

 prediction is based on both stature and body-weight gives far better 

 results (as shown by the preponderance of negative signs in the final 

 difference column) than prediction from weight alone. 



