INTRODUCTION. 13 



tion. As has been pointed out in the earlier report, 1 the relationship between 

 the oxygen consumption and calories and the carbon-dioxide production and 

 calories was a relatively constant one for diabetics, that is, approximately 3.3 

 calories for every gram of oxygen consumed or carbon dioxide produced. On 

 the completion of the later experiments it was ascertained that these ratios 

 were abnormally low, both for oxygen and for carbon dioxide, while the 

 respiratory quotients as finally determined were in full accordance with the 

 earlier work. A further examination of the results, together with a series of 

 control experiments, showed that the apparatus measured the heat radiated 

 from the body most accurately, and the error was finally found to be in the 

 determination of the water-vapor. It was just at this time that fundamental 

 alterations in the method of determining the water-vapor were being intro- 

 duced, and unquestionably the errors in heat measurement can be directly 

 ascribed to this. Under these conditions, therefore, we deem it inexpedient to 

 discuss in this report the heat as measured, but as all of the control experiments 

 point toward the highest degree of accuracy in the measurement of the carbon 

 dioxide and oxygen, we can confidently base our discussion upon these two 

 factors. Since in an earlier publication the relationships between oxygen con- 

 sumption, carbon-dioxide production, and heat production were so thoroughly 

 studied, and since these well-known relationships play an important role in 

 so-called "indirect calorimetry," the data for the gaseous exchange in this new 

 series of experiments have an increased significance. 



'Benedict and Joslin, loc. cit., p. 219. 



