OXYGEN SUPPLY. 47 



was covered with a bell having an exit and a valve at the top. When 

 this valve was opened, the water entered the can of sodium peroxide 

 and gas was generated. The gas thus formed was remarkably pure, 

 containing only moisture. It was, therefore, still necessary to have 

 a drier. One objection to this apparatus was the fact that during 

 generation intense heat was formed which interfered with accurate 

 weighing. 



This method of supplying oxygen did not prove so practical as the 

 use of cylinders, and when it was found that the oxygen from the Linde 

 Air Products Company of Buffalo, New York, contained very little 

 nitrogen and practically no weighable amount of carbon dioxide and 

 water, their product was substituted. Small cylinders were obtained, 

 containing about 150 liters of the gas, with approximately 3 per cent 

 of nitrogen. 1 A reduction valve was attached by means of which the 

 flow of oxygen into the apparatus could readily be regulated. While 

 the quality of the oxygen and the method of admission were both 

 satisfactory, provided the reduction valve was in perfect condition, it 

 was frequently found that the reduction valve did not work properly 

 or that it was leaking. A Bohr experimental gas-meter of 1-liter 

 capacity was therefore tested in the spring of 1911 and adopted; at the 

 present time there are at least five of these meters in use in the Nutri- 

 tion Laboratory. 



The Bohr meter as set up and used is shown in figure 15 (page 41). 

 Each scale division corresponds to 5 c.c., while the numerals correspond 

 to 0. 1 liter. The whole meter is immersed in an aquarium j ar filled with 

 water. This insures uniformity of temperature throughout the meter 

 and surrounding medium, and precludes measurable temperature change 

 in a 15-minute experiment. A moistener is placed in front of the meter 

 so as to provide for the complete saturation of the air passing through 

 it, thus preventing the evaporation of the water in the meter. This 

 moistener consists of a wide-mouth bottle, c, in which a three-holed 

 rubber stopper provided with tubes is inserted. One tube dips below 

 the level of the water and the other provides for the exit of the gas. 

 A third tube, which extends from below the surface of the meter to 

 above the water in the aquarium jar, serves as a safety valve in case 

 there is back pressure. The use of this is referred to later. The 

 bottle is weighted down with shot. The thermometer inserted through 

 the cover of the aquarium jar indicates the temperature of the water. 



The requirements for accuracy in the use of the meter are accurate 

 measurements of the barometric pressure and the temperature, com- 

 plete saturation of the air with water- vapor, and a knowledge of the 

 mechanical factor of the apparatus. The first three conditions can 



'Formerly the impurity was considered to be nitrogen, but it has recently been found that thia 

 impurity is nearly all argon and our calculations are made upon this basis. 



