Of ferromanganese and other very manganiferous substances, in which the manganese need not be determined with greater exactness than to 0.1 per cent., only 0.01 gram. is taken for a test; but of common pig, wrought iron, steel, iron ore, slags, etc., there is taken 0.5 to 1 gramme according to the supposed content of manganese and the desired exactness of the estimation. For instance one gramme iron, which has passed through a metal sieve with holes half a millimeter in diameter, is placed in a beaker 125 mm. in height and 60 mm. in diameter, and has added to it twenty cubic centimeters of hydrochloric acid of 1.12 specific gravity, which, with a well-fitting glass cover, is boiled for half an hour, in order that the combined carbon may be driven off in the shape of gas. After at least the half of the hydrochloric acid has been boiled away, there are added at least five cubic centimeters nitric acid of 1.2 specific gravity, partly to bring the iron to peroxide, partly to destroy the organic matters formed from the carbon, which might possibly be remaining and might tend to remove the color of the chameleon solution. The boiling is now continued till near dryness, when five cubic centimeters hydrochloric acid are added, after which the solution is boiled as long as any reddish-yellow vapors of nitrous acid are observed. When these have disappeared a drop of the liquid taken up on a small glass rod is tested with an newly prepared solution of red prussiate of potash (2 grammes in 100 cubic centimeters water), to ascertain whether there is any protoxide of iron remaining. First, when no indication of blue or green is visible, the test shows a pure yellow, it is certain that there are no reducing substances in the solution.