The method in question is recommended as easy, expeditious, and accurate. It consists in precipitating all the manganese in the state of peroxide, dissolving it in a ferrous solution so as to bring back the manganese to the manganous slate, and determining volumetrically, by means of potassium permanganate, the quantity of ferrous salt which has been converted into ferric. The method of rapidly precipitating manganese peroxide is peculiar. If we act upon cast-iron or steel with nitric acid and potassium chlorate in certain proportions, and boil the mixture, the manganese is completely precipitated in the state of peroxide insoluble in nitric acid, but retaining a small quantity of ferric oxide. Suppose that we have a sample of steel or manganiferous cast-iron containing less than 7 per cent of manganese. Three grammes are treated in a small flask with 40 c. c. of nitric acid, of sp. gr. 1.20, added little by little. The liquid is stirred, and ultimately heated to complete solution. It is withdrawn from the fire, and 15 grammes potassium chlorate are added, and then 20 c. c. of nitric acid at sp. gr. 1.40. It is boiled for about fifteen minutes, until the escape of chlorine ceases; all the manganese is found thrown down as peroxide; hot water is added, the mixture is filtered, and the precipitate washed with boiling water.

To dissolve the manganese peroxide thus obtained we measure exactly 50 c. c. of an acid solution of ferrous sulphate, made up with 40 grammes ferrous sulphate to 750 c. c. water and 230 c. c. sulphuric acid (full strength). The 50 c. c. are poured into the flask in which the sample has been dissolved, and to which a little peroxide adheres, and it is then poured upon the precipitate and the filter in a Berlin-ware capsule. The manganese peroxide dissolves very readily, transforming its equivalent of ferrous sulphate into ferric sulphate. The liquid is then diluted to 100 or 150 c. c. for the next operation. We then take a solution of permanganate formed by the same proportions as are used in determining iron by the process of Margueritte (5.65 grammes of the crystalline salt per liter of water), and determine its standard exactly. By means of this liquid we determine volumetrically the quantity of ferrous sulphate remaining in the solution of manganese. We take then 50 c. c. of the original solution of ferrous sulphate diluted as above, and determine the total ferrous salt.

The difference between the two determinations corresponds to the ferrous salt which has been peroxidized by the manganese peroxide. The quantity of iron thus peroxidized multiplied by 0.491 gives the quantity of manganese contained in the portion operated upon. In the case of a steel or cast iron containing but little manganese it is convenient to dissolve the peroxide in 25 c. c. only of the ferrous solution. Small Gay-Lussac burettes may then be used in the titration of only 0.010 meter internal diameter, and graduated into one-twentieth c. c., which allows of great exactitude in the determination. For a spiegeleisen not more than 1 gramme of the sample should be taken, and for a ferro-manganese 0.3 gramme.