f >$ On the Biirificution. ofNkkeh 



golutioH had not acquired any colour ; and on evaporating 

 the ammonia no oxide of" nickel was obtained. 



As far as I have examined, the solutions of oxides in car- 

 bonate of ammonia possess the same properties as those ii^ 

 ammonia ; for although carbonate of potash does not de- 

 compose the latter, yet the former are decomposed by potash. 

 This is easily explained. Potash has a stronger affinity for 

 carbonic acid than annnonia has : when, therefore, a carbor 

 nated ammoniaeal solution is decomposed by it, a part of the 

 potash combines with the carbonic acid of the carbonate of 

 ammonia, and the remaining part decomposes what by the 

 action of the other has become ammoniaeal solution. 



Considering that the supposition of a combination of the 

 two alkalies would be strengthened if the experiments which 

 gave rise to it could be reversed, i. e. if it were possible to 

 precipitate with ammonia substances dissolved by potash, 

 I dissolved in the latter silica, alumina, and several metallic 

 oxides. On adding ammonia to these solutions, I at first 

 thouoht I had succeeded in obtaining precipitates ; but upoi| 

 examining the ammonia employed, I found that it con- 

 tained a small quantity of carbonic acid ; and that when am? 

 nionia free from it was used, no precipitate was in any case 

 obtained. Although these experiments did not succeed in 

 supporting the above supposition, yet they are by no meaij^ 

 fatal to it. It is probable that potash has a stronger affinity 

 for the substances which it dissolves than it has for anir 

 jTionia; and in this case, as no combination could be ef- 

 fected, no precipitation would ensue. In the first experi,- 

 ments potash and annnonla seem to possess a stronger af$.- 

 nity for each other than ammonia has for the oxides solul^lp 

 in it. 



Judging by the effects produced by water and pptasb, thr 

 ■affirxity of ammonia for the metallic oxides appears to he a^ 

 follows : 



Oxide ,pf cobajt. 

 Oxide of copper. 

 Oxide of silver, 

 Oxide of nickel. 



I reduced a quantity of the oxide of nickel, obtained i)f 

 the above-described process, and obtained a button pJf nie^gl 

 which exhibited the followirg properties : 



Colour — Dull yellowish vviiitc. 



Fracture— r-I*oliated. Specific gravity 8*51_. 



jFragile, but capable of slight extcnsiou by hammering. 



Strongly magnetic. 



In 



