I90 



SECTIONAL ADDRESSES 



heats of combustion of carbohydrate and fat and the chemical composition 

 of these substances), it is therefore in the third factor that we must look 

 for the error ; and this must mean that it is not possible to calculate the 

 proportion of carbohydrate to fat actually burnt from other respiratory 

 quotients than o • 785. 



When these results were first reported at the Guy's meeting of the 

 Physiological Society in 1932, Prof. Benedict suggested that they might 

 be due to errors in measuring the oxygen. If, for instance, the volume 

 of oxygen was too low, the calculated value for the heat would be too 



Fig. I. 



Fig. 2. 



Comparison of indirect and direct calorimetry. o = Oxygen ; + = COj ; 

 two-hourly determinations during first day of fasting. Subject at rest. In 

 fig. I two observations coincided and so only 20 are shown. 



low ; at the same time the respiratory quotient would be too high, so 

 that the position of the point would be low down on the right of the 

 diagram and, similarly, if the oxygen were too high the point would be 

 high up and to the left ; in fact, the distribution would be the same as in 

 the figures. The oxygen would be particularly liable to error because it 

 was measured both by noting the loss in weight of an oxygen cylinder and 

 also by determining the change in oxygen content of the calorimeter at 

 the beginning and end of the experiment. Any error in the analysis of 

 the gas content of the calorimeter would be magnified in the final result 

 if the volume of the calorimeter was large. On the other hand, the carbon 

 dioxide, which has only a small influence on the indirect heat, was pre- 

 sumably accurate, because as the calorimeter was a closed circuit it could 

 not escape. The validity of this explanation can be tested by noting the 



