92 A RESPIRATION CALORIMETER. 



check experiments the discrepancies are not so great as during work 

 experiments with men, but it is clear that slight disturbances in either 

 the heat-radiating surface or in the heat- absorbing surface will make 

 considerable differences in the average temperature of the total volume 

 of air. 



With alcohol check experiments this factor is practically a constant 

 one, and while the electrical resistance thermometers may not represent 

 the average temperature in the system, at the same time, owing to the 

 constancy of the thermal gradient, they probably record accurately any 

 differences in temperature, and it is these alone which affect the volume 

 of air. 



In experiments with man the temperature factor could be eliminated 

 were it possible to have the heat-radiating surface constant throughout 

 the whole period. With variations in position, muscular activity, 

 changes in clothing, bedding, etc., however, it is very difficult, if not 

 indeed absolutely impossible, to measure the differences of the average 

 temperature of the air at the beginning and end of each period. This 

 is most noticeable in the case of work experiments, where in the morn- 

 ing at 7 a. m. the subject is lying quietly in bed asleep, and at the end 

 of the period, 9 a. m., he is riding a bicycle ergometer at a high rate 

 of speed. Again, at n o'clock at night the subject is sitting dressed, 

 possibly reading or writing, and at i a. m. he is sound asleep, covered 

 with bed-clothing. It is during these two periods, when the widest 

 variation in bodily activity naturally takes place, that we find the 

 greatest discrepancies in the measurement of the volume of oxygen. 

 These discrepancies are indicated generally by abnormal values for the 

 "respiratory quotient." (Seep. 184.) In the particular experiment the 

 results of which are presented in this report (p. 177) these discrepancies 

 do not appear. 



CONCLUSION REGARDING THE ACCURACY OF THE OXYGEN COMPUTATION. 



This method of calculation is found to be far more practicable and 

 on the whole more satisfactory at the present state of our experimenta- 

 tion than a method depending upon analyses of air at the end of each 

 period ; for the difficulties in securing proper temperature measure- 

 ments of the large volume of air affect alike the calculation of the total 

 residual oxygen whether the analysis of a sample is made by the most 

 approved methods or whether the computation method is employed. It is 

 therefore believed that the errors involved in the system of calculations 

 as above outlined are certainly no greater than those that necessarily 

 occur in using direct analyses of oxygen under present conditions. 



