202 METABOLISM AND GROWTH FROM BIRTH TO PUBERTY. 



actually measured and as predicted from the curve in figure 27 (total 

 calories referred to weight) and likewise the heat predicted from the 

 curve in figure 39 (total calories referred to surface) . For each of these 

 series of values, the differences between the predicted and actual and 

 the percentage differences are given. 



Considering first the prediction of metabolism from the curve of 

 total calories referred to weight, we find the plus and minus signs are 

 fairly equal in number, as would be expected from the way the curve 

 was laid on. The magnitude of the differences ranges from to 152 

 calories, but since the children vary in weight from 2.7 to 30.9 kg., 

 percentage differences alone have value for comparison purposes. 

 In 11 instances the error of prediction is 20 per cent or more; but 10 

 out of these 11 girls weigh 10.6 kg. or less, thereby confirming the ob- 

 servation made with boys that the error is smaller with the heavier 

 children. 



TABLE 35. Comparison of the actual basal metabolism of girls with the metabolism predicted: 

 (a) from body-weight, 1 and (b) from body-surface. 2 (Average values. 3 ) 



i See figure 27, p. 142, and table 36, p. 206 



3 Averages obtained from data in table 34, pp. 199 to 201. 



2 See figure 39 p. 164. 



The summaries showing the average differences between predicted 

 and actual and the percentage differences for girls are given in table 35. 

 The average error of prediction for all the girls on this basis is there 

 given as 9.7 per cent. Since we note that the errors are greater with 

 the earlier weights, the girls may be separated into those weighing 

 above or below 10 kg. We find that those below 10 kg. have a devia- 

 tion of 11.8 per cent and above 10 kg. 7.5 per cent, thus furnishing 

 a mathematical demonstration of our conclusion from the inspection 

 of the figures in column h of table 34 that the accuracy of the prediction 

 increases with the increase in weight. If a division is made at 15 kg., 

 as was done for the boys, the percentage deviation for the girls with 

 this or a higher body-weight is 6.1 per cent. 



A special reason for separating the girls at 10 kg. is based upon the 

 fact that we found that there was no sexual differentiation in the 

 metabolism of boys and girls up to the weight of 10 kg., so on this 



