78 



COMPARISONS OF RESPIRATORY EXCHANGE. 



dead space. All of the stopcocks should be absolutely tight. They 

 may be tested by drawing the air into the burette and then, after 

 connecting with the stopcock which is to be tested, putting the air in 

 the burette under pressure. If there is a leak the volume will gradually 

 decrease. A leak may also be shown by putting the air in the burette 

 under diminished pressure and the liquid in the potash pipette or the 

 potassium pyrogallate pipette will gradually rise, owing to the suction, 

 if the stopcocks connecting these parts leak. In manipulating the 

 apparatus care should be taken as far as possible to have the parts 

 under pressure when the setting of the potash levels is begun, otherwise 

 if there is suction the potash will rise into the connections, thus requiring 

 cleaning with acid and the lubrication of the stopcocks. 



Testing the Apparatus. 



The apparatus is regularly tested in this laboratory by analyses of 

 outdoor air. The outdoor air remains uniform in composition and the 

 standard for carbon dioxide and oxygen has been taken as 0.03 per 

 cent for the former and 20.94 per cent 1 for the latter. The limits of 

 accuracy commonly allowed have been 0.03 to 0.04 per cent in paral- 

 lels, this being a plus or minus error of 0.02 per cent. If the figures 

 obtained are not within theiimits of accuracy, the analysis is continued 

 or a search is made for the cause. Generally, however, with the labora- 

 tory form of the apparatus, it is not difficult to obtain duplicates within 

 0.01 or 0.02 per cent for both carbon dioxide and oxygen. 



The burette should be calibrated by some standard method of 

 calibration. 



PORTABLE FORM. 



The portable form of the Haldane apparatus has likewise been used 

 in this laboratory for the analysis of outdoor air and expired air. The 

 results of a series of analyses of outdoor air made with this apparatus 

 by G. A. D. are given in table 11. Phosphorus was used for the 

 absorption of oxygen. 



Table 11. — Results of analyses of atmospheric air with the 

 portable Haldane gas-analysis apparatus. 



x The value for oxygen used in this investigation is 20.94. Haldane gives 20.93 and Benedict 

 20.95. It is immaterial which value is used when the limits of error allowed are those given 

 above. 



