216 



JOSEPHINE ALBERTA WILLIS. 



used was as follows: A 160 cubic centimeter bottle, with a neck 

 1.5 centimeter in diameter was fitted with a rubber stopper 

 carrying a 10 centimeter long capillary tube reaching within one 

 centimeter of the bottom of the bottle. The outer end of the 

 capillary tube was fitted with a piece of heavy rubber tubing 

 carrying a clamp. This tube was fitted over the capillary of the 

 Haldane apparatus when the air to be analyzed was taken into the 

 burette of the apparatus. 



In making a determination, an approximately known volume 

 of carbon dioxide (made as described above) w r as transferred 

 from a nitrometer tube to the container filled with atmospheric 

 air and open to prevent increased pressure. The container was 

 then closed and a known amount of air, about 8 cubic centimeters 

 was withdrawn from it into the Haldane apparatus previously 

 filled with nitrogen made by removing oxygen and carbon 

 dioxide from atmospheric air. The percentages of oxygen and 

 carbon dioxide were then determined from this sample. The 

 stopper was withdrawn, the animal introduced and the bottle 

 quickly closed. Probable error at this point is negligible because 

 of the slow- diffusion of carbon dioxide. Since the air in the 

 bottle was put under slight negative pressure in the first analysis, 

 it was thought that some outside air might enter. However, 

 immediate analysis showed no change great enough to affect the 

 results of the experiments. Not more than three to five readings 

 were made on each bottle so that negative pressure produced 

 would not influence the normal reaction of the animal. 



By the above method, .using the Haldane apparatus, determi- 

 nations are accurate to 0.02 per cent. Thus the results after 

 taking into consideration individual differences, should be 

 comparable to at least o.i per cent. 



TABLE IV. 



SHOWING PERCENTAGE CHANGES IN COz AND Oj DURING EXPOSURE TO THE GAS 



FOR DEFINITE PERIODS OF TIME. 



