on the Atomic Weights of Bodies. 187 



the specific gravity of the oxygen gas. Three trials made 

 in this way all agreed with each other, and gave the specific 

 gravity of oxygen gas 1*1 117. — (See Annals of Philosophy, 

 xvi. 163.) 



During the course of the same year 1820, an elaborate 

 set of experiments on the specific gravity of oxygen, azote, 

 hydrogen, and carbonic acid gas, was made by Berzelius 

 and Dulong, with all that attention to minute accuracy 

 which distinguishes these eminent chemists. {Ann. de 

 Chim. et de Phys., xv. 386.) The specific gravities of oxy- 

 gen and azotic gases which they obtained are as follow : 

 Oxygen gas 1*1026 

 Azotic gas 0*976. 

 The specific gravity of oxygen gas as determined by these 

 chemists coincides almost exactly with the previous deter- 

 minations of Kirwan and of Biot and Arago. 



But, notwithstanding the great care bestowed upon these 

 experiments, it is demonstrable that they are not perfectly 

 correct. For the specific gravity of air deduced from them 

 (supposing air a compound of 21 oxygen and 79 azotic gas) 

 would not be 1 as it ought to be, but 1*002586. And, if 

 air be composed of 20 oxygen and 80 azote, its specific 

 gravity deduced from the above numbers would be 1*00132. 

 Such are the most accurate experiments to determine the 

 specific gravity of oxygen gas which have yet been made. 

 None of them is perfectly correct. But as the errors in all 

 probability lie on different sides, we may conclude, that by 

 taking the mean of the whole, we shall obtain a number 

 exceedingly near the truth. The following little table 

 exhibits these results. 



Specific gravity of oxygen gas according to 



Kirwan, 1*103 



Berzelius and Dulong, . 1*1026 

 Biot and Arago, . . . 1*103 

 De Saussure, .... 1*1352 

 Thomson, 1*1117 



5*5555 



Mean, = 11111 

 Thus, we obtain for the mean of the whole, the very mini- 



