432 Professor Kane on the Action of Ammonia 
Here the mercury and chlorine agree pretty nearly with my determination, but the 
quantity of ammonia is only about one-half. I think it probable, that Guibourt deter- 
mined the ammonia by means of potash, thinking the decomposition perfect, as it is 
generally described, and as I myself at first considered it to be. In order to show 
the probability of this being the source of Guibourt’s error, I shall describe the fol- 
lowing, one of the earliest experiments I made in the matter. 
100 grains of white precipitate were mixed with a very strong solution of potash, 
in a flask from which a bent tube passed to a jar containing dilute muriatic acid. The 
flask was heated until most of the water in it had distilled over. ‘The liquor in the 
jar was then evaporated to dryness, and gave 11,5 grs. of sal-ammoniac, consisting of 
Muriatic acid 7,84 
Ammonia 3,66 
Now, supposing the loss to be oxygen, which was the method generally pursued in 
analyzing this substance hitherto, we should have the result— 
Mercury 78,60 
Chlorine 13,85 
Ammonia 3,66 
Oxygen 8,89 
These results agree so closely as to point out the manner in which Guibourt’s 
analysis, so correct in the chlorine and quicksilver constituents, became erroneous by 
reducing the ammonia to nearly one-half the actual amount. 
100,00 
I shall return to the causes of error in Soubeiran’s analysis; at present it is not 
necessary to advert to them, as he evidently analyzed a body quite different from white 
precipitate. 
The simplest view to take of the existence of the chlorine in this substance, is to 
suppose it united with half the mercury as corrosive sublimate ; it is almost the only 
view possible. Then, in what state is the remainder of the mercury? We may sup- 
pose it peroxidized, and the oxide united with the ammonia giving the formula 
(2Ch + Hg) + (Hg +2NH’) and the following numerical arrangements :— 
2atoms mercury 405,60 or 77,00 
2 oxygen 16,00 3,04 
2 chlorine 70,84 13,45 
Q ammonia 34,30 6,51 
526,74 100,00 
This agrees closely with Mitcherlich’s, and also with some of my own analyses. It 
differs, nevertheless, from the mean of my results in the quantities of mercury and 
chlorine, and particularly in the quantity of oxygen. This hypothesis supposes the 
existence of 3,04 per cent. of oxygen, a body of which I could not determine the 
