On the Vurification of NkM. 315 



By comparing the above statements it appears proper, for 

 the separation of the oxides in question, that the solution 

 containing them should have the annnoaia in excess, and 

 be largely diluted before the addition of the potash. The 

 excess of ammonia prevents, for a considerable length of 

 time, any precipitation of the oxide of cobalt, while it pro- 

 duces no delay in the precipitation of the oxide of nickel. 

 Mere dilution precipitates a part of the oxide of nickel ; and 

 at the same time that it renders the remainder more easy of 

 precipitation, it prevents for several days any deposition of 

 the oxide of cobalt. The following experiment will show 

 that this method may be relied upon : 



To a measure of solution of oxide of nickel in ammonia 

 I added solution of potash as long as precipitation took 

 place. The precipitate was washed, dried, and weighed. As 

 it appeared probable that all the oxide of nickel might not 

 be precipitated, I heated the solution after filtration till the 

 ammonia had evaporated, but I did not obtain any further 

 quantity of oxide. 



To a similar measure of the same solution of oxide of 

 nickel I added a measure of solution of oxide of cobalt in 

 ammonia, and precipitated by potash. The precipitate ap- 

 peared to be pure oxide of nickel ; and, after drying, its weight 

 did not differ 1-lOth of a grain from that of the oxide of 

 nickel obtained in the former experiment. This experiment 

 was repeated with nearly similar results. 



Soda produces the same effects as potash, and appears 

 to act rather more readily ; but whether in smaller quantity 

 than the latter I have not tried. Carbonate of potash pro- 

 duces no effect. The decomposition of the ammoniacal 

 solutions by potash appears to depend upon a combination 

 of the two alkalies ; the compound possessing, as is usual 

 in chemical combinations, properties either partially or to- 

 tally differing from those of its constituents. That this pre- 

 cipitation do"cs not depend upon dilution has, I think, been 

 already shown ; and the following experiments more clearly 

 determine that it must be ascribcil to some other cause : 



To one measure of solution of aunnonia 1 added three mca- 

 gurcs of water. Moist oxide of nickel put into this dilute 

 solution quickly coloured it. After several days digestion 

 the solution decomposed by potash furnished nearly tour 

 grains of oxide. 



One measure of solution of animonia was mixed with 

 three measures of dilute solution of potash. After as long di- 

 gestion with moist oxide of nickel as in the foriner ease, tJie 



soliitiou 



