on the Atomic Weights of Bodies. 263 



And the atomic weights of these bodies would be 

 Carbonic oxide, . . . 3-5 



Oxalic acid, 4*5 



Carbonic acid, .... 55 

 Carburetted hydrogen, . 2 

 defiant gas, .... 1'75 



Alcohol, 2-875 



Ether, 4-625 



These numbers have nothing in them to startle us. The 

 atom of carbonic acid being doubled the salts at present called 

 bi-carbonates would, in fact, be carbonates, and those called 

 carbonates would be di-carbonat.es. Analogy favours this 

 new notion. The bi-carbonates are much nearer the neutral 

 state than the carbonates; and, therefore, more likely to 

 contain an atom of each constituent. While the carbonates 

 possess all the characters which usually characterize the 

 sub-salts. 



Berzelius's atomic number for carbon being almost iden- 

 tical with mine, it is needless to institute a comparison 

 between them. 



For the same reason it will be necessary to double my 

 atomic weight for silicon and aluminum ; making the former 

 2 and the latter 2-5. These numbers multiplied into the 

 specific heats in the table would give us for a product the 

 mean number 0-375. But, as the specific heats of these 

 bodies are only hypothetical, it is needless to insist farther 

 on the subject. 



The specific heat of oxygen in the second compartment 

 of the table is given according to the experiments of Dela- 

 roche and Berard, while those of chlorine and bromine are 

 deduced from the experiments of Messrs. Haycraft, Delanve 

 and Marcet, according to whom, the specific heats of gases 

 and vapours are inversely as the square roots of then- 

 specific gravities. It is obvious, that none of these specific 

 heats can be depended on ; but, from the numbers in the 

 third column, we see that my atomic numbers must be 



true ones. 



Berzelius has adopted for chlorine and bromine numbers 

 which give the following results : 



I Atoms of I Specific Products , 

 1 Berzelius. I heat. 



Chlorine .... 2-21326|0-0827 0-183 

 Bromim', . . . • Ul»lK»|o-04w| 0-231 



