RESEARCHES ON THE AMMONI ACAL COMPOUNDS. 105 
This result differs in giving the amount of mercury and 
chlorine at a higher estimate than that derived from the re- 
sults ohtained by other chemists, with the exception of Gui- 
bourt, whose analysis on these points agrees remarkably with 
the above, but varies in making the ammonia about one-half 
the quantity, the balance being estimated as oxygen. This 
result was probably due to the mode of analysis adopted by 
Guibourt, based on the supposition that potassa liberated all 
the ammonia from the precipitate, the loss being supposed to be 
oxygen. But Prof. Kane has shown that only one-half of 
the ammonia is thus liberated, and that an analysis conducted 
in this way necessarily leads to this erroneous result from the 
inability of potassa to produce complete decomposition. 
The above formula does not, however, explain the pheno- 
mena resulting from the action of heat upon white precipitate, 
for a substance composed as above, should yield calomel and 
ammonia, not calomel, ammonia and nitrogen, when exposed 
to heat. 
Dumas' researches has led to the probability that, in oxamide, 
benzamide, &c, there exists abody composed of NH 2 , capable of 
combining with other bodies, as,for example, the bases of oxides, 
in the same manner as chlorine and cyanogen do. Considering, 
then, that this body is combined with one-half of the metallic 
mercury in the precipitate, while the other half is combined 
with the chlorine, there will result the bichloride of mercury 
on the one hand, and the biamidet of mercury on the other, by 
their union constituting the white precipitate, the formula for 
which would be 
HgCl 2 4-Hg2NH 2 
The examination of this formula will afford a very simple 
explanation of the effects of heat on this compound ; for, if 
three equivalents are considered to have undergone decompo- 
sition, there will be six equivalents of chlorine and six of mer- 
cury, to form the chloride of mercury, and twelve equivalents 
