OXYGEN. 11 



In the sum total of these nineteen experiments, 840.161 grammes of 

 oxygen form 945.439 grammes of water. This gives, in percentages, for 

 the composition of water — oxygen, 88.864; hydrogen, 11.136. Hence 

 the atomic weight of oxygen, calculated in mass, is 15.9608. In the 

 following column the values are deduced from the individual data given 

 under the headings F and G : 



15-994 

 16.014 

 16.024 

 15-992 

 15.916 

 15.916 



15-943 

 16.000 



15-892 

 15-995 

 15-984 

 15-958 

 15.902 



15-987 

 15.926 



15-992 

 15-904 

 15.900 

 16.015 



Mean, 15.9607, with a probable error of ± .0070. 



In calculating the above column several discrepancies were noted, 

 probably due to misprints in the original memoir. On comparing col- 

 umns B and C with F, or D and E with G, these anomalies chiefly ap- 

 pear. They were detected and carefully considered in the course of my 

 own calculations ; and, I believe, eliminated from the final result. 



The investigation of Erdmann and Marchand * followed closely after 

 that of Dumas. The method of procedure was essentially that of the 

 latter chemist, differing from it only in points of detail. The hydrogen 

 used was prepared from zinc and sulphuric acid, and the zinc, which 

 contained traces of carbon, was proved to be free from arsenic and sul- 

 phur. The copper oxide was ma'de partly from copper turnings and 

 partly by the ignition of the nitrate. The results obtained are given in 

 two series, in one of which the weighings were not actually made in 

 vacuo, but were, nevertheless, reduced to a vacuum standard. In the 

 second series the copper oxide and copper were weighed in vacuo. The 

 following table contains the corrected weights of water obtained and of 

 the oxygen in it, with the value found for the atomic weight of oxygen 

 in a third column. The weights are given in grammes. 



* Journ. fiir Prakt. Chem., 1842, bd. 26, s. 461. 



