Dr. Kane on the Compounds of Ammonia. 25 



Hence he deduced the formula NO5 . NH3 + 3 ugo, which should give 

 Mercury = 86.46 



100.00 



In any ordinary case, where the error, unavoidable in manipulation, and to 

 which the collection of mercury in the metallic form by proto-chloride of tin, is 

 peculiarly liable, should necessarily tend to diminish the quantities obtained, and 

 consequently reduce the experimental, below the theoretical numbers, his analysis 

 should be considered as completely establishing the formula ; but here, there are 

 other circumstances which require to be taken into account, and which will lead 

 us to an opposite conclusion. 



It is evident that in the preparation of Hannehman's soluble mercury, there 

 is a tendency to error from the intermixture of a greyish material, and where 

 the whole, or nearly the whole of the solution has been precipitated at once, this 

 intermixture is unavoidable ; hence it is only the first portions that can be 

 obtained of the fine black colour which characterizes the pure substance. Now it 

 has been fully proved, that according as the decomposition proceeds, the quantity 

 of mercury in 100 increases in proportion as the colour becomes less deep ; and 

 hence the error, in estimating the composition of this body, must be opposite in 

 direction to what generally occurs, and must tend to render the proportion of 

 quicksilver above the truth. Thus, the result of Mitscherlich's theory is almost 

 precisely that obtained in my analysis of specimen No. 3, which was not black, 

 but dark grey ; and Mitscherlich himself indicates the powder which he analyzed 

 as grey ; he says, " Nach dem Trocknen ist die Farbe des Pulvers grau und es 

 darf sich kein metallisches quecksilber mechanisch herausdrucken lassen ;" and 

 also, if the solution which I employed had been precipitated all at once, there 

 should have been a dark grey precipitate, and its composition should have been 

 the mean of the composition of the powders given by the four equal and succes- 

 sive additions of water of ammonia, which average would almost coincide with the 

 result which Mitscherlich obtained. So marked is the production of this whitish 

 matter, that Soubeiran collected and analyzed it, and concluded from his results 

 that it was an ammonia sub-protonitrate with the formula no^ . NH3 -\- 4 ugo. He 



VOL. XIX. E 



