104 COMPARISONS OF RESPIRATORY EXCHANGE. 



eter which would give a graphic record of the movements of the spi- 

 rometer bell during the last 5 minutes of the period. An assistant, by 

 pressing a key, connected this recording device in circuit with a signal 

 magnet when the kymograph drum was started; the key was then 

 pressed each minute until the end of the period. If the subject was 

 absolutely quiet and there was no marked change in the temperature 

 or the barometric pressure, the decrease in volume would be relatively 

 constant. If, however, the pointer showed an upward movement 

 of the spirometer bell, the experimental period was continued until the 

 record showed that the subject was perfectly quiet. In this way the 

 expansion of the air due to any movement of the subject was controlled 

 and the volume of the apparatus as shown by the spirometer at the 

 end of the period was the actual volume. 



The measurement of the average temperature of the air may also 

 be affected by the fact that the subject, when he first enters the appa- 

 ratus, warms the clothing with which he is covered and then warms the 

 air next to the clothing; accordingly, during the preliminary period and 

 possibly during the first period of the experiment, there may be a 

 gradual warming of the air in the chamber which is not indicated by 

 the thermometers. The record of the temperature at the beginning 

 of the experimental period would therefore be lower than the actual 

 average temperature of the apparatus, since the thermal gradient 

 between the man and the air next to the wall had not been established. 

 The resulting measurement of the oxygen content of the air in the 

 chamber would therefore be too large, the calculated oxygen consump- 

 tion too small, and the respiratory quotient too high. The chamber 

 itself is actually a large air thermometer, the changes of which are 

 shown by the movements of the diaphragm of the tension equalizer 

 or the bell of the spirometer. 



The difficulties met with in securing the average temperature and 

 the fact that the errors in the determination of the residual carbon 

 dioxide and water- vapor all affect the calculation of the oxygen content 

 of the air make the determinations of the oxygen consumption extremely 

 difficult and the experimental period should be of such length that these 

 errors would play a very small percentage role. Many of the experi- 

 ments in this comparison are relatively short and it would undoubtedly 

 have been desirable to have had them longer. It was impracticable, 

 however, to keep the subject inside the chamber for a longer period and 

 still have him sufficiently quiet for comparison purposes. 



SOUBCES OF ERROR IN EXPERIMENTS WITH THE BENEDICT RESPIRATION APPARATUS. 



As has already been shown in the description of the Benedict respira- 

 tion apparatus, 1 the principle is exactly the same as that of the bed 

 calorimeter, except that it is applied on a much smaller scale. There is 



'See p. 21. 



