5 THE ATOMIC WEIGHTS. 



The third method indicated at the beginning of this dis- 

 cussion has been recently employed in part by J. Thomsen* 

 of Copenhagen. Unfortunately this chemist has not pub- 

 lished the details of his work, but only the end results. 

 These serve to confirm the values for oxygen fixed by other 

 methods, but they cannot well be included in the systematic 

 discussion. Partly by the oxidation of hydrogen over 

 heated copper oxide, and partly by its direct union with 

 oxygen, Thomsen finds that at the latitude of Copenhagen, 

 and at sea level, one litre of dry hydrogen at 0° and 760 

 mm. pressure will form .8041 gramme of water. According 

 to Regnault, at this latitude, level, temperature, and press- 

 ure, a litre of hydrogen weighs .08954 gramme. From 

 these data, = 15.9605. It will be seen at once that Thom- 

 sen's work depends in great part upon that of Regnault, 

 and yet that it affords an admirable reinforcement of the 

 latter. 



It is now plain, in conclusion, that all the different lines 

 of research point to an atomic weight for oxygen a little 

 below 16.00. Five distinct investigations confirm each other 

 wonderfully. Upon combining the values obtained by the 

 two chief methods we get the following final results : 



From synthesis of water O = 15.9642, it .0060 



From gaseous densities O = 15.9627, rfc .0043 



In the general mean the atomic weight of oxj^gen be- 

 comes 15.9633, with a probable error of ± .0035.t 



*Ber. d. Deutsch. Chem. Gesellschaft, 1870, s. 928. 



f Le Conte's correction of Regnault's figures introduced here would make O = 

 15.9622, instead of 15.9633. Difference, .0011. 



