354 0?i the Cause of Chemical Proportions. [May, 



gold, mercury, and copper, liave two salifiable oxides, the protoxide 

 of which in all these metals has a striking analogy. These protoxides, 

 which I distinguish by the Latin termination osstiw, di^er in so 

 striking a degree from the protoxides of iron, manganese, and 

 cerium, that there must be some general difference in their compo- 

 sition. The peroxides of the first named metals have more charac- 

 teristic marks of salifiable bases tlian the protoxides ; the contrary 

 is the case with the oxides of iron, manganese, and cerium. The 

 peroxides of the first class contain twice as much oxygen as tlie 

 protoxides, while in the three other metals the peroxides contain 

 only l-J- times as much oxygen as the protoxides. It appears from 

 this that the peroxides of platinum, gold, copper, mercury, are 

 proportional in their composition to the protoxides of iron, manga- 

 nese, and cerium ; that is to say, that both contain 2 volumes of 

 oxygen. This circumstance, added to the fact that the protoxide of 

 gold cannot possibly contain more than 1 volume of oxygen, has 

 induced me to conclude that the protoxides of rhodium, platinum, 

 gold, mercury, and copper, are composed of equal volumes of 

 radicle and oxygen. In consequence of all this, if, according to 

 my experiments, the protoxide of platinum is composed of 100 

 metal and S"287 o^ oxvgen, the volume of platinum must weigh 

 l-'067. 



15. Auriim, gold (Au). — A hundred parts of gold by my experi- 

 ments combine in the peroxide with 12-077 "^ oxygen ; in the 

 protoxide, with i of that quanthy. If this protoxide be Au + O, 

 as is j^robable, then the volume of gold weighs 2483*8. As the 

 peroxide of gold is Au + 3 O, it follows that a deutoxide Au + 

 12 O must exist. I have endeavoured to prove that the purple of 

 Cassius contains this deutoxide, and that the formation of jhe 

 pov.'der is inexplicable without the existence of this intermediate 

 oxide. M. Proust conceives that the purple of Cassius is a combi- 

 nation of metallic gold with the oxide of tin. But such a combi- 

 nation is contrai'v to every known analogy respecting the circum- 

 stances under wluch tb.c metals combine with other oxides. But 

 the solubility of the purple of Cassius in ammonia, which Proust 

 discovered, and which I have veiificd, proves sufficiently that tlic 

 gold in it is in the state of an oxide. Now the purple of Cassius 

 cannot contain protoxide of gold, because that oxide forms green 

 or yellow combinations ; neither can it contain the peroxide, because 

 it is formed by the reduction of tliat oxide to a lower degree of 

 oxidation. It appears, then, clearly to follow, that it must contain 

 ft deutoxide, which, like the deutoxide of rhodium, is not salifiable, 

 though it be capable of combining with the oxide of tin ; and by 

 means of this last, likewise, with ammonia. It seems to be the 

 deutoxide of gold that gives the ])urple colour to animal and veget- 

 able matter treated with muriate of go'J. 



16. Palladium (Pa). — The analogy between platinum and palla- 

 dium would lead one to expect that this last forms two oxides as 

 well as, the first. But I have only been able to discover a single 



