I.— PHYSIOLOGY 191 



extent of the variation in the volume of oxygen. For A.L.L. the oxygen 

 intake per two hours was 30-1 litres at a respiratory quotient of 0-834 

 and 20-6 litres at a quotient of 0-932 (a variation of 9-5 litres) and for 

 A.H.M., 29-4 litres at a quotient of 0-762 and 18-9 litres at a quotient 

 of 0-947 (^ variation of 10-5 litres). Such variations are surely quite 

 outside any possible error due to analysis. The fact that all control 

 experiments in which alcohol was burnt in the calorimeter show a high 

 degree of accuracy in the measurement of the oxygen intake when com- 

 pared with the heat output confirms the view that the error indicated in 

 figs. I and 2 cannot be accounted for by experimental errors in the 

 measurement of the oxygen. 



Similar results have been obtained with a number of individuals who 

 had been observed in the smaller calorimeters at Boston, as well as with 

 a number of pathological cases described by Du Bois and his colleagues, 

 and with Murlin and Lusk's observations on dogs. We have also shown 

 that they cannot be accounted for by a time lag in the taking up of heat 

 by the calorimeter. 



Constancy of Carbon Dioxide Output. — -In figs, i and 2 the values of the 

 oxygen intake and the carbon dioxide output corresponding to the heat 

 differences have been plotted ; the carbon dioxide remains remarkably 

 constant throughout the range of respiratory quotients ; the oxygen 

 intake, on the other hand, diminishes as the quotient increases. Benedict 

 and Carpenter comment on the constancy of the carbon dioxide in their 

 experiments. Those authors who have made a number of determinations 

 of the basal metabolism on the same individual have in general obtained 

 a constant COj and a fall in the oxygen with rise of quotient ; but when 

 the subject has been taking a carbohydrate diet there is a small but definite 

 rise in CO2 with rise in quotient and a fall in the oxygen as before. 



An Alternative Method of Indirect Calorimetry. 



As it became clear that the Zuntz-Schumburg figures could no longer 

 be used for the purpose of calculating the heat generated by the body, 

 the necessity arose of seeing whether any other relationship existed 

 between the o.xygen intake, the carbon dioxide output and the heat 

 production from which it would be possible to calculate the heat if the 

 oxygen intake and the carbon dioxide output were known — in other 

 words : Is indirect calorimetry a possibility } To answer this question 

 the plan adopted was to see whether there was any direct relation between 

 the carbon dioxide output and the heat produced, and between the oxygen 

 and the heat. In fig. 3, 337 observations by Du Bois and his colleagues 

 on the basal metabolism have been plotted, and the alcohol control 

 experiments as well. In these metabolism experiments the body tempera- 

 ture was not above 37-5° C. and protein figures are used — i.e. no 

 deduction has been made for the protein metabolism. 



The position of the points in the diagrams shows that there is a linear 

 relation between the CO2 and calories (correlation coefficient 0-950 



