LECTURE VI.] HISTORY OF CHEMISTRY. 95 



II. When a positive oxide combines with a negative one 

 (a base with an acid, for example), the oxygen in 

 the latter is a multiple, by a whole number, of that 

 in the former, and this number usually is, at the 

 same time, the number of the oxygen atoms in the 

 negative oxide. 



These are the only two rules which Berzelius advances in 

 1 8 1 9 in his theory of chemical proportions. In the translation of 

 the second edition of his Text-book (the first German edition) 

 new rules are added, called forth by Mitscherlich's discovery of 

 isomorphism, and by the relations which Dulong and Petit had 

 found (1819) between the atomic weights and the specific heats 

 of solid elements. 



Since both of these investigations are of the greatest 

 importance with respect to the views of Berzelius upon the 

 question now under discussion, I shall here introduce the 

 results of these investigations, and shall then proceed with the 

 consideration of the atomic weight determinations. 



Dulong and Petit proved, by exact experiments, 24 that the 

 products obtained by multiplying the specific heats of bismuth, 

 lead, gold, platinum, tin, silver, zinc, tellurium, copper, nickel, 

 iron, cobalt, and sulphur, by the respective atomic weights of 

 these elements, are almost identical ; and from this fact they 

 drew the conclusion that the same regularity would hold with 

 respect to all the elements and would lead to the exact 

 determination of their atomic weights. 



In order to establish the law, Dulong and Petit had 

 assumed the atomic weights of most of the metals, with 

 reference to that of sulphur, to be only half as great as 

 Berzelius had stated them in 1819. They assumed 201 for 

 the atomic weight of sulphur (O= 100), as Berzelius had done, 

 and then made Fe = 339, whereas Berzelius had adopted 693. 

 According to their law, the atomic weight of silver was only 

 one-fourth of Berzelius' number. In the cases of tellurium and 

 of cobalt, they arrive at results which are still further at variance 



24 Ann. Chim. [2] 10, 395. 



