E. W. Mbrley — Volumetric Composition of Water. 227 



It was desirable that the hydrogen to be analyzed should not 

 pass over any organic lubricant unless the lubricant were 

 proved not to give up any carbon compound to the gas. 

 Rather than investigate lubricants, I used syrupy phosphoric 

 acid on the glass stop-cocks, and at first used no fusible metal 

 valves, but prevented leakage while the exhaustion of the large 

 volumes e and n, fig. 2, was in progress by maintaining a 

 vacuum between the two stop-cocks by means of a second mer- 

 curial pump ; but the manipulation was troublesome, and if 

 the exhaustion was interrupted, the whole labor had to be 

 recommenced. 



In some of my experiments, I used hydrogen purified by 

 the process suggested by Chirikoff,* namely, by utilizing the 

 power of palladium to absorb hydrogen, and to give it off 

 again at a higher temperature. In such experiments, a tube 

 containing palladium was connected as shown at p, but a tube 

 containing phosphorus pentoxide should have been shown be- 

 tween o and p, intended to arrest any aqueous vapor which 

 might possibly be given off by the syrupy phosphoric acid at 

 m. A second drying tube was placed at q, to answer the same 

 purpose for the gas which should pass directly to r and the 

 following parts of the apparatus. 



The apparatus so far described serves to prepare pure hydro- 

 gen. In the case of all impurities to be feared, except oxygen 

 and nitrogen, their absence is capable of proof by direct ex- 

 periment, as sensitive reagents are applicable. It was not 

 expected that hydrogen free from nitrogen could be prepared, 

 and apparatus designed to measure the residual nitrogen was 

 constructed, and will shortly be described. The only question 

 remaining is as to the possible presence of oxygen after the 

 hydrogen had passed the tube containing heated copper at f. 

 To determine how much of a given substance fails to be ab- 

 sorbed by the best absorbent of the substance can be deter- 

 mined commonly only by indirect means which are of limited 

 application. But in regard to the absorption of oxygen by hot 

 copper, I think we can answer the question, and can perhaps 

 also judge whether the presence of a great quantity of hydro- 

 gen will lessen the completeness of the absorption. If we pass 

 a current of air containing a little aqueous vapor through a 

 tube filled with phosphorus pentoxide, and also pass a current 

 of air over heated copper in another tube, we can in some sort 

 compare the activity of the two absorptions by observing how 

 far each absorbent is visibly affected. I think we may judge 

 that the copper is not enormously less active then the phos- 

 phorus pentoxide. Suppose the phosphorus pentoxide is a 



* Journal Russ. Phys.-Chem. Soc, 1882, chemical part, p. 47. 

 Am. Jour. Sci.— Third Series, Yol. XLI, No. 243.— March, 1891. 

 15 



