OXYGEN. 7 



a more satisfactory kind. Indeed, the}^ are among the 

 classics of physical science ; and probably approach as near 

 to absolute accuracy as is possible for experiment. 



For hydrogen three determinations of density gave the 

 following results : 



.06923 



.06932 



.06924 



Mean, .069263, rb .000019 



For oxygen four determinations were made, but in the 

 first one the gas was contaminated by traces of hydrogen, 

 and the value obtained, 1.10525; was, therefore, rejected by 

 Regnault as too low. The other three are as follows : 



1.10561 

 I. 10564 

 1.10565 



Mean, 1. 105633, ±: .000008 



Now, combining the hydrogen and oxygen series, we have 

 for the atomic weight of oxygen, 15.9628, dr .0044.* 



Upon combining the result of Regnault's work with that 

 from Dumas and Boussingault's we get the following value : 



From Dumas and Boussingault O = I5-9538, ± .031 



From Regnault O = 15.9628, dz .0044 



General mean O = 15.9627, it .0043 



This result, it will be seen, agrees remarkably well with 

 that obtained in the experiments upon the synthesis of 

 water. 



* Since these computations were made, Professor John Le Conte has called my 

 attention to the existence of slight numerical errors in Regnault's own reductions. 

 As corrected by Le Conte, Regnault's figures give 1.105612 for the density of 

 oxygen, and 0.069269 for that of hydrogen. Hence the atomic weight of O be- 

 comes 15.961 1, instead of 15.9628. The difference is slight, but still it ought 

 not to be ignored. All the computations in the body of this work, having been 

 finished before I received Professor Le Conte's figures, must stand, nevertheless, 

 as they are. For further details Le Conte refers to Phil. Mag., (4,) 27, p. 29, 

 1864; and also to the Smithsonian Report for 1878, p. 428. 



