612 poisons : THEIR effects and detection. [§ 777. 
State. It is now transferred to a glass tube, open at top and bottom, 
but the top widened into a funnel ; this tube is firmly held perpendicu¬ 
larly on a glass plate, and the prepared sulphide hammered into a 
compact cylinder by the aid of a glass rod, which just fits the tube. 
The cylinder is now dried over a flame until no more moisture is to 
be detected, and then transferred into a glass tube 4 or 5 inches long, 
and with one end drawn to a point (the weight of this tube should be 
first accurately taken). The tube is connected with the following 
series :—(1) A chloride of calcium tube ; (2) a small bottle containing 
nitrate of silver solution ; (3) a hydrogen-generating bottle containing 
zinc and sulphuric acid. The hydrogen goes through the argentic 
nitrate solution, leaving behind any sulphur and arsenic it may contain ; 
it is then dried by chloride of calcium, and streams in a pure dry state 
over the cylinder of prepared sulphide (no error with regard to impurities 
in the gas is likely to occur ; but in rigid inquiries it is advisable to 
heat a portion of the tube, previous to the insertion of the cylinder, for 
some time, in order to prove the absence of any external arsenical source) ; 
when it is certain that pure hydrogen, unmixed with air, is being 
evolved, the portion of the tube in which the cylinder rests is heated 
slowly to redness, and the metallic arsenic sublimes at a little distance 
from the source of heat. Loss is inevitable if the tube is too short, or 
the stream of hydrogen too powerful. 
The tube, after the operation, is divided ; the portion soiled by the 
soda thoroughly cleansed, and then both parts weighed ; the difference 
between the weight of the empty tube and the tube + arsenic gives the 
metallic arsenic. This is the process as recommended by Mohr ; it 
may, however, be pointed out that the glass tube itself loses weight when 
any portion of it is kept red-hot for some little time ; and therefore, 
unless the crust is required in the original tube, it is better to divide 
it, carefully weigh the arsenical portion, remove the crust, and then 
re-weigh. The method is not perfectly accurate. The mirror is 
not pure- metallic arsenic, and if the white alkaline residue be 
examined, arsenic will be detected in it, the reason being bhat the 
arsenical sulphide generally contains pentasulphide of arsenic as well 
as free sulphur. Now the pentasulphide does not give up metallic 
arsenic when treated as before detailed ; nor, indeed, does the trisulphide, 
if mixed with much sulphur, yield an arsenical crust. It is therefore 
of great moment to free the precipitate as much as possible from 
sulphur, before attempting the reduction. 
The development of a reducing gas from a special and somewhat 
complicated apparatus is not absolutely necessary. The whole process 
of reduction, from beginning to end, may take place in a single tube by 
any of the following processes :—(1) The sulphide is mixed with oxalate 
of soda (a salt which contains no water of crystallisation), and the dry 
