228 E. W. Morley — Volumetric Composition of Water. 



hundred or a thousand times more active. Now, partly to 

 obtain a solution of the present problem, I have determined* 

 the amount of aqueous vapor which fails to be absorbed by a 

 phosphorus jDentoxide tube containing twenty-five cubic centi- 

 meters, when the rate of the current of gas is two liters an 

 hour ; the volume of the aqueous vapor left unabsorbed is not 

 more than a thirty-million th of that of the air. If then we 

 make the surfaces exposed and the time of exposure compar- 

 able for the two absorbents, we may expect that the copper 

 will permit not over a hundred or a thousand times this quan- 

 tity to escape ; by increasing the exposed surface, and the time 

 of exposure, the absorption may be made as complete as we 

 please, for at the temperature used, we have reason to believe 

 that the reaction between copper and oxygen is not reversible. 

 Considering also the action of hydrogen which accompanies the 

 oxygen and is heated with it, it seems to me impossible to sup- 

 pose that the gas of my experiments contained any such quan- 

 tity of oxygen as 1/200000 its volume. 



Purity of the hydrogen obtained. — Each time when a globe 

 was filled with hydrogen for the determination of density, the 

 amount of nitrogen in the hydrogen was determined. The 

 first trial was made when a current of two or three liters an 

 hour had been evolved for three days ; the gas then contained 

 •00035 volumes of nitrogen. The amount became gradually 

 less as the amount of hydrogen obtained from the apparatus 

 increased. When the apparatus was not in use, it was always 

 left closed by fusion against the entrance of nitrogen except 

 by diffusion through the water in which the oxygen delivery 

 tube ended. When about six hundred liters had been obtained, 

 it became difficult to detect nitrogen ; that is, its amount was 

 less than one part in one hundred thousand. Sometimes con- 

 cordant determinations of the amount of nitrogen showed only 

 one part in two hundred thousand. 



Apparatus for storing hydrogen for several experiments 

 on the same sample. — Figure 2 shows the part of the apparatus 

 serving for certain processes in the determination of the com- 

 position of water. In my methods, it was necessary to secure 

 two considerable volumes of hydrogen which should be abso- 

 lutely identical in composition, one to be weighed, and one to 

 be analyzed. It was also necessary to preserve the second for 

 several days, with no possibility of contamination by leakage 

 or diffusion. Moreover, the sample for analysis must not come 

 in contact with fatty lubricants, nor that for weighing in con- 

 tact with mercury, from which it would take up mercury 

 vapor affecting its weight. It would no doubt have been 

 possible to have found a fatty lubricant which it would have 



* This Journal vol. xxxiv. p. 199, 1887. 



