390 THE SCIENTIFIC MONTHLY 



metabolism exhibits the same orderliness that biometric wor 

 has shown to prevail in the variation of other biological meas 

 urements. When data are somewhat more numerous, it wi 

 be profitable to carry the analysis further by fitting theoretic* 

 frequency curves to these frequency distributions. 



Measured statistically, the variability of these subjects ma 

 be expressed by standard deviations of 204.66 calories in me 

 and of 155.18 calories per twenty-four hours in women, or b 

 coefficients of variation, i. e., of standard deviations expresse 

 as percentages of the means, of 12.54 per cent, in men and c 

 11.50 per cent, in women. 



The statement that variation in the total daily heat produ&amp;lt; 

 tion of adults is measured by a coefficient of 11.5 to 12.5 pe 

 cent, will mean very little to the non-statistical reader until h 

 can compare these with coefficients for characters with whic 

 he is more familiar. In our series stature shows a coefficier 

 of variation of 4.39 in men and 3.20 in women, body weight 

 coefficient of variation of 16.06 in men and 20.35 in womei 

 pulse rate at complete rest a coefficient of variation of 10.99 i 

 men and 12.01 in women. Thus basal metabolism shows a var 

 ability far greater than stature but less than body weight an 

 of roughly the same order of magnitude as pulse rate. 



Basal metabolism is, therefore, rather highly variable. Th 

 reader will have noted, however, that the foregoing polygor 

 and constants are based upon the total daily heat productions c 

 adults in presumably good health but of various body weight 

 statures and ages. It has already been suggested that bas, 

 metabolism is related to these physical characters. 



We must now inquire whether the observed variability i 

 heat production is due in part to differences in bodily dimei 

 sions. This influence has been considered so great that son 

 physiologists have asserted that heat production per squai 

 meter of body surface is a constant. 



That heat production expressed in calories per square met* 

 of body surface is not a constant in any exact sense is shov 

 by Fig. 2, in which the ordinates represent the frequencies 

 total heat production per square meter of body surface as esi 

 mated by the Du Bois height-weight chart. 



These diagrams show clearly that heat production p 

 square meter, like total daily heat production, is a variab 

 function. In both cases the frequencies decrease as the magi 

 tudes of the constants diverge more widely, in both the pi 

 and the minus direction, from the average for the whole serif 



The fact that a large variability in daily heat product! 





