220 Professors Redtenbacher and Liebig 



they combine with oxide of silver, or, what is the same thing, 

 with metallic silver. These organic acids contain several 

 atoms of carbon, combined with oxygen and hydrogen in cer- 

 tain proportions, which can be ascertained with great facility. 

 It is evident, that if the formula? of these acids are known 

 with certainty, we obtain the sum of the weights of the atoms 

 of carbon, by subtracting the sum of the atoms of oxygen 

 and hydrogen which they contain from the atomic weights of 

 their compounds, as accurately determined by means of their 

 silver salts. The sum of the atoms of carbon thus ascer- 

 tained, must, if the formerly received number for carbon be 

 correct, be a multiple of it by a whole number, or it must in- 

 dicate how far that number differs from the true one. There 

 is no substance, the atomic weight of which we believe to be 

 known with greater certainty, and more precisely determined, 

 than that of silver; the continued and important applications 

 of it by Gay-Lussac, in his assay of silver in the wet way ; 

 indeed we may say, every one of his experiments on that sub- 

 ject is a fresh proof of its correctness. 



With respect to the atomic weight of hydrogen, there is 

 strong reason for thinking it to be rather less than was for- 

 merly supposed. In the three last analyses of water by Ber- 

 zelius and Dulong, the following numbers were obtained : — 



1. 2. 3. 



Oxygen 88*942 88*809 88*954 



Hydrogen... 11*058 11*191 11*046 



100*000 100*000 100*000 



Supposing the atomic weight of hydrogen to be 6*2398, 

 then the limits of error lie between the numbers 6-3055 and 

 6*2085. The difference between these numbers is 0*0970, 

 which would either increase the atomic weight of carbon from 

 76*437 to 76*534, or reduce it to 76*340. If we take the 

 mean, which is 6*2398, for hydrogen, then the limits of error 

 in the determination of the atom of carbon fall in the second 

 decimal place. Such variations result, as we know, from errors 

 of observation. 



The method which we have selected, enables us to estimate 

 the atomic weight of carbon in the condition in which it is 

 contained in organic compounds; and it requires only three 

 weighings : — first, that of the vessel in which the salt is 

 burned; secondly, that of the silver salt; and, thirdly, that 

 of the residual silver. In these three weighings no apparatus 

 is changed, and they take place in one and the same porcelain 

 vessel, of which the weight is constant. The height of the 



