APPARATUS AND METHODS USED IN THIS RESEARCH. 



49 



level of the alcohol in the burette was accurately read and the experi- 

 ment proper began. Thereafter it was only necessary to read this 

 burette accurately with a lens at the end of each experimental period of 

 30 to 40 minutes. 



L 



Fig. 4. Method of introducing alcohol in the alcohol check tests of respiration apparatus. 



The ventilating air-current was passed through the chamber at the 

 rate of approximately 35 liters per minute, and with a flame of this 

 size and regularity in the introduction of alcohol the amount of carbon 

 dioxide residual in the chamber at the end of each experimental period 

 was usually very constant. Since, however, we were dealing with 

 quantities of carbon dioxide amounting to 0.01 gram, it was necessary 

 in this exceedingly exact work to determine the residual amount of 

 carbon dioxide. This was obtained by drawing a sample of the air at 

 the end of each period, and determining the amount of carbon dioxide 

 by means of a modified 1 Pettersson-Palmquist gas-analysis apparatus, 

 which permits measurements of 0.5 or less per cent of carbon dioxide 

 to the third significant figure. Its manipulation is very simple and 

 has been rapidly acquired by a number of workers in the laboratory. 

 It should be stated that these residual analyses were not required in 

 the observations on infants at the hospital, but were necessary only to 

 secure the greatest degree of refinement in establishing the accuracy of 



Anderson, Journ. Am. Chem. Soc, 1913, 35, p. 162. 



