Combinations of Arsenic with Cobalt. 333 



any doubt as to the homogeneity and purity of this combina- 

 tion. If, therefore, it be admitted that the above-mentioned 

 quantitative composition affords a perfect symbol of the nature 

 of the combination, then the atomic relation of arsenic to the 

 other metals would be as the numbers 



7-74: 1776, 



that is, if it be assumed that the very slight quantity of sul- 

 phur occupies the place of a portion of the arsenic, and that 

 copper, cobalt and iron are here isomorphous. That this 

 actually is the case, will very clearly appear from the experi- 

 ments subsequently detailed. 



Second Combination. This is crystallized in large laminae 

 of metallic lustre, which for the greater part are so rooted in 

 the mass, that it was not possible to determine from the ter- 

 mination of the laminae the crystallographic system to which 

 they belong. Frequently several such laminae are super- 

 posed one upon the other, and project with dull superficies 

 beyond the surface of the rest of the metal. They have then 

 an appearance quite similar to that of the above-described 

 combination. The analysis was performed differently from 

 the former, inasmuch as the cobalt and iron were separated 

 from each other by the succinate of ammonia. 

 The result was as follows : 



Sulphur 0*50 



Arsenic 35-20 



Cobalt 31-35 



Iron 23-15 



Copper 8-90 99-10. 



The peroxide of iron separated by succinate of ammonia 

 evinced, indeed, before the blow-pipe the action of cobalt, but 

 in so slight a degree that it seemed of no importance. The 

 oxide of cobalt, however, on being dissolved in muriatic acid, 

 and treated with ammonia in excess, left behind 0*033 gram, 

 of the peroxide of iron ; the quantity of which, however, was 

 of course increased by some oxide of cobalt precipitated 

 with it*. 



* More accurate results, as to the application of succinate of ammonia 

 for the separation of cobalt from iron, are, as far as my experience 

 goes, extremely difficult to obtain. If it be attempted more perfectly to 

 neutralize the fluid previously to the addition of the succinate of ammonia, 

 the peroxide of iron may very likely prove to contain somewhat too much 

 cobalt. The method that I have recommended for the separation of these 

 two metals, it is true, has also its difficulties; however, in most cases I have 

 had better success with it, and scarcely ever worse, than with succinate of 

 ammonia. At the same time this very great advantage should be con- 

 sidered, that the solution containing cobalt, filtered from the iron, can 

 be immediately thrown down with potash ; while, in the separation by 



