36 A RESPIRATION CALORIMETER. 



from the reservoir R. 2 . When the last traces of gas have left the capil- 

 lary on the pipette the stopcock C and the pinchcock are closed. 



The pipette is then shaken in the hand for five minutes, at the end 

 of which time, as has been shown by repeated tests, the absorption of 

 oxygen is complete. 



After filling the upper part of the rubber tube R with water the 

 pipette is connected with the capillary tube T. On opening the stop- 

 cock C in such a manner as to connect the capillary tube T with the 

 burette Bj previously filled with water, and by lowering the reservoir R 1( 

 the unabsorbed gas can be drawn over into burette B r R x is lowered 

 sufficiently to cause the reagent in the pipette to rise in the capillary, 

 completely fill the rubber tube R and pass along the capillary T, to a 

 graduation G, when the stopcock C is closed. After waiting about 

 three minutes for the water on the walls of burette B! to settle, the 

 reading of the level of water in E l is made by holding the reservoir Rj 

 immediately back of the burette in such a manner that the levels in 

 both tubes are the same. 



Both burettes B a and B. 2 are so graduated that they can be read accu- 

 rately to o.o i cc. The small burette B, is graduated from o to 20 cc. 

 The large burette B 2 is graduated only above 90 cc. , but from 90 to 

 100 cc. it is graduated in 0.05 cc. 



In computing the actual volumes of gas used it is necessary to take 

 into consideration the volume of the space in the connection between 

 the stopcock C and the two burettes, as well as that of the capillary 

 extension-tube T. Before being used for gas analyses, both burettes 

 were calibrated very accurately by filling with mercury. It was thus 

 found that to the reading obtained on burette B 2 there must be added 

 a constant 0.51 cc. for the volume of gas in the connections between 

 the burette proper and the end of the tube T, and to the volume as read 

 on burette B t there must be added a constant 0.34 cc. for the volume 

 of gas in the connections between the burette proper and the gradua- 

 tion mark G on the capillary tube T. 



In the analyses it is assumed in the first place that the difference in 

 time between the beginning and end of an analysis is so short that no 

 difference in barometric pressure will occur that need be taken into con- 

 sideration. It is furthermore assumed that in general, in the analysis 

 of oxygen, the relations between the volumes of nitrogen and of the 

 oxygen as originally measured are such that no fluctuations in temper- 

 ature ordinarily experienced will affect materially the percentage of 

 nitrogen in the sample of oxygen being analyzed. Consequently the 

 volume of gas plus the correction 0.51 cc. as measured on burette B, 

 and the volume of nitrogen as corrected with the constant 0.34 cc. as 



