18 



SMITHSONIAN MISCELLANEOUS COLLECTIONS 



VOL. 54 



hydride was exhausted by means of a Geissler pump to remove any 

 nitrogen which might have been present. In the preliminary investiga- 

 tion cited above, the latter precaution was neglected, which may account 

 for the low results. 



Between the palladium tube and the combustion tube a U tube was 

 interposed, containing phospliorus pentoxide. This was to determine 

 the amount of moisture in the hydrogen. The combustion tube was 

 filled with granular copper oxide, prepared by reducing the commercial 

 oxide in hydrogen, heating the metal so obtained to bright redness in a 

 vacuum, and then reoxidizing with pure oxygen. 



Upon warming the palladium tube, which was first carefully weighed, 

 hydrogen was given off and allowed to pass into the combustion tube. 

 When the greater part of it had been burned, the tube was cut off by 

 means of a stopcock and allowed to cool. Meanwhile a stream of nitro- 

 gen was passed through the combustion tube, sweeping hydrogen before 

 it. This was followed by a current of oxygen, reoxidizing the reduced 

 copper; and the copper oxide was finally cooled in a stream of dry air. 

 The water produced by the combustion was collected in a weighed bulb 

 tube, followed by a weighed U tube containing phosphorus pentoxide. 



A second phosphorus pentoxide tube served to prevent the sucking back 

 of moisture from the external air. The loss in weight of the palladium 

 tube, corrected by the gain in weight of the first phosphorus pentoxide, 

 gave the weight of hydrogen taken. The gain in weight of the two col- 

 lecting tubes gave the weight of water formed. All weights in the fol- 

 lowing table of results are reduced to a vacuum : 



Mean, 15.9514, ±: .0011 



In sum, 6.55880 grammes of hydrogen gave 52.30383 of water, whence 

 = 15.9492. 



In March, 1889, Lord Rayleigh ^ published a few determinations of the 

 ratio obtained by still a new method. Pure hydrogen and pure oxygen 



' Proc. Roy. Soc, 45, 425. 



