stuffed with raw cotton at the end d g, and has a bent plate of copper or tinned iron hung over it at f,—open the stop-cock, allow a little gas to escape so as to expel the air in d e, and then kindle the gas at e, which must be contracted to a capillary opening. Keep the flame low, and hold the surface of a white porcelain vessel across the middle of it for a few seconds. If no stain be produced on the porcelain, there is no arsenic in the fluid. If a stain be formed, regulate the escape of gas by the stop-cock so that the fluid may not rise above the middle of the lower vessel of the apparatus, and apply the heat of a spirit-lamp flame to the tube d e on the left hand of the plate f, the purpose of which is to prevent the heat being communicated beyond that point. By and by, if there be arsenic in the fluid, a brilliant metallic ring will appear beyond f,owing to decomposition of arseniuretted-hydrogen gas. As soon as the crust is thick enough to present its properties characteristically, withdraw the spirit-lamp; place the tube e h so that the flame at e shall be completely within the ball, i; let the tube incline very slightly in the direction from k to l; and allow a stream of cold water to trickle down upon the portion k l, which should be wrapped in a single layer of calico. Oxide of arsenic will gradually condense, partly in white powder or minute sparkling crystals in the ball and between i and k, and partly between k and l in the form of a solution, which collects at the bend l. The solution which may be increased in quantity by boiling a little distilled water upon the powder in the ball and bend i k, is then to be subjected in small portions to the three liquid reagents, ammoniacal nitrate of silver, ammoniacal sulphate of copper, and hydrosulphuric acid.
Some experience is required to apply this process successfully. But with due attention it furnishes conclusive evidence with great delicacy and precision. A solution containing only a millionth part of oxide of arsenic will part with it readily in the form of arseniuretted-hydrogen; and the slightest trace of that gas in the hydrogen is indicated by the method recommended above.—The process is compounded of Mr. Marsh's original discovery, the supplementary test of reduction in the exit-tube recommended by Berzelius,[1] and the formation and examination of the oxide proposed by myself.[2]—With certain precautions and modes of manipulating, it is applicable to the most complex organic fluids, as well as to simple solutions.
The discovery of Mr. Marsh had not been long made before the test in its original simple form was found liable to divers important fallacies. It appeared, for example, that antimony yields very nearly the same appearance of metallic crust and of white powder, according to the position of the porcelain in the flame; that some porcelains glazed with oxide of zinc are similarly stained by a flame of simple
