36 H. ROSE ON THE COMBINATIONS OF THE 
obtained by the last method was carefully freed from adherent 
chlorine. 
If the protochloride is treated with dry ammoniacal gas, a 
white powder is obtained with evolution of heat, which is per- 
fectly soluble in water. 1:862 gramme of the well-saturated com- 
bination, dissolved in water, to which some nitric acid and then 
nitrate of silver were added, gave 3°323 grammes chloride of silver. 
1:368 gramme of the combination from a different preparation 
gave, on similar treatment, 24195 grammes chloride of silver. 
If the composition of the protochloride of arsenic-ammonia be 
calculated in 100, according to these experiments, the following 
results are obtained :— 
I. Il. 
C@hiorinetiginss hs ein a 402 43°63 
Arsene (y-s5 geeeii pe hageh Sls 30°89 
Ammonia «66 es ¢  B4B2 25°48 
100°00 100°00 
The second quantity of the compound examined retained 
some adhering ammonia; for it had been long preserved in a 
vessel filled with dry ammoniacal gas. The solution of this 
combination reacted, moreover, more strongly alkaline than that 
of the other quantity of the compound. 
The composition found agrees with the formula As Cl +7 N H, 
or rather 2 As Cl? + 7 N H®; the composition 1 calculated ac- 
cording to this contains in 100, 
Whiorinie: 2 Vo. wee ee) ee 4400 
Arsenic 07" PCr ne ant eh cee ee 
AMMONIA sso ee ee se 
100:00. 
In its solution by water there might be formed, supposing all 
the chlorine of the protochloride of arsenic converted into chlo- 
ride of ammonium, an acid arsenite of the oxide of ammonium, 
2 As + NH4 (the neutral salt is most probably As + N Hi’). 
T have e attempted i in vain to prepare a chloride of arsenic cor- 
responding to arsenic acid, in order to be able to examine its 
combination with ammonia. Liebig and Wohler had previously 
failed in preparing it*. I have obtained the same results as 
* Annalen der Pharmacie, Bd. xi, p. 150. 
