43 o OF COBALT, &c. 



3. From a comparison of nickel, cobalt, arid 

 manganefe, it appears, that when pure of arfenic, 

 thcfe metals will fcarce melt in the fire; when 

 combined with arfenic, with which they enter 

 very eagerly into combination, they are no 

 longer iubjecl to the attraction of the magnet : it 

 like wife appears, that they can fcaree be entire- 

 ly iVee of iron if attrachibility by the magnet be 

 the tell of the prcfcnce of iron; when carefully 

 purified, they become ductile, and precipitate lu- 

 ruratediblutiomof niver/thusdinering Jrom iron. 

 Thefe qualities,- jull mentioned are, therefore, 

 common to tliefe three metals; but in other par- 

 ticulars they plainly ditler. For manganefe dif- 

 fers fo far from the rctt in its fpeciilc gravity 

 which is 6,850, and in other qualities peculiar to 

 itfelf, that uny peribn \\liu makes experiments 

 upun it, can have no doubt of its being a pecu- 

 Jiur fubtlunce. As to whut 1ms been lately af- 

 iertcdby a celebrated chemill, that nickel and 

 cobalt ave one metal, only indifferent forms, that 

 has indeed Tome ihew of probability, if what I 

 have above related concerning thole metals be 

 taken into conlideration. Yet proofs of their 

 diverilty arc not wanting. Nickel, when impreg- 

 nated with cobalt, cannot, without great dif- 

 ficulty, be feparatcd from it, and viceverjli; 

 this circumflance alone might be fullicient to 

 produce the miflakc. But nickel, when feparat- 

 cd from cobalt, canuor, with any addition of ar- 



fenic, 



