I.— PHYSIOLOGY 189 



by control experiments in whicli alcohol was burnt in the calorimeter 

 under experimental conditions and the measured heat, oxygen con- 

 sumption, and carbon dioxide production were compared with the quan- 

 tities predicted from the known weight of alcohol burnt, and the results 

 obtained indicate a high degree of accuracy. So satisfactory were these 

 control experiments that it was felt to be necessary to give only one typical 

 experiment and a summary of the results of all experiments of the same 

 type which were made between October 1903 and April 1904 with the 

 calorimeter of 5,000 litres capacity at the Wesleyan University, Middle- 

 town, U.S.A. In two of their subjects, A.L.L. and A.H.M., the oxygen 

 intake and the carbon dioxide output, and the heat generated, were simul- 

 taneously determined with the subject at rest and in the post-absorptive 

 state at two hourly intervals over periods of eight hours for a number of 

 days. There are 24 sets of determinations for A.L.L. and 21 sets for 

 A.H.M. The ' non-protein ' values for oxygen intake and carbon dioxide 

 output and the calories of heat generated were calculated on the usual 

 assumption that the nitrogen excreted during any period corresponds to 

 an amount of protein which has been completely oxidised during that 

 period ; if now the oxygen, carbon dioxide and heat (in calories) resulting 

 from this oxidation are deducted from the total figures, the remaining 

 amounts must be solely due to the metabolism of carbohydrate and fat. 

 A recent study which is not yet complete throws some doubt on this 

 method of calculation, particularly for the higher quotients, and no 

 allowance for protein has been made in the later results described in this 

 Address. I have not found that the conclusions are invalidated by this 

 omission. 



Using these data for the subjects A.L.L. and A.H.M. , the theoretical 

 heat from the oxygen intake and the respiratory quotient was calculated 

 by the Zuntz-Schumburg method, and this heat (indirect heat) was com- 

 pared with the heat actually measured by the calorimeter (direct heat). 

 The difference between the indirect and direct heat was expressed as a 

 percentage of the direct heat and related to the respiratory quotient by 

 a graphic method. The result is shown in figs, i and 2. The horizontal 

 line at zero in the top part of the diagrams represents the condition where 

 the indirect and direct heats are identical ; the ordinates represent the 

 percentage differences between the indirect and direct heats. Satis- 

 factory agreement between the heat calculated and the direct heat will 

 depend on the percentage differences, indicated by the black dots, being 

 evenly distributed above and below the horizontal zero line throughout 

 the range of respiratory quotients. This is obviously not the case. The 

 positions of the points suggest a systematic error ; for the lower quotients 

 the indirect heat is too high and for the higher quotients too low. The 

 direct and indirect heats only agree at a respiratory quotient of about 

 0-785 and not elsewhere. At this particular point the respiratory 

 quotient represents a combustion of carbohydrate and fat in the ratio 

 of I to I -36. It is unlikely that there can be a very large error in the first 

 two factors on which the Zuntz-Schumburg values are based (viz. the 



