20 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 54 



bulb filled with phosphorus pentoxide, and the gaseous contents were 

 collected in a eudiometer. On cooling and weighing the apparatus, the 

 loss of weight, less the weight of gases pumped out, gave .the amount of 

 water produced by the traces of residual hydrogen under consideration. 

 This weight, added to the loss of weight when the original water was 

 expelled, gives the weight of oxygen taken away from the copper oxide. 

 Having thus the weight of hydrogen and the weight of oxygen, the 

 ratio sought for follows. Six results are given, but as they are repeated, 

 with corrections, in Noyes' second paper, they need not be considered 

 now. 



Noyes' methods were almost immediately criticised by Johnson,^ who 

 suggested several sources of error. This chemist had already shown in 

 an earlier paper ^ that copper reduced in hydrogen persistently retains 

 traces of the latter, and also that when the reduction is effected below 

 700°, water is retained too. The possible presence of sulphur in the 

 copper oxide was furthermore mentioned. Errors from these sources 

 would tend to make the apparent atomic weight of oxygen (referred to 

 hydrogen as unity) too low. 



In his second paper ^ Noyes replies to the foregoing criticisms, and 

 shows that they carry no weight, at least so far as his work is concerned. 

 He also describes a number of experiments in which oxides other than 

 copper oxide were tried, but without distinct success, and he gives fuller 

 details as to manipulations and materials. His final results are in four 

 scries, as follows : 



First Series. — Hydrogen from Zinc and Hydrochloric Acid. 

 Wt. of H. Wt. of 0. Ratio H : 0. 



Mean, 15.8973, ± .0032 



This series appeared in the earlier paper, but with an error which is 

 here corrected. 



' Chem. News, 59, 272. 

 ''Journ. Chem. Soc, May, 1879. 

 *Amer. Chem. Joiirn., 12, 441. 1890. 



