Having found carbon present in all varieties that we examined, we naturally attempted its quantitative determination. Our first experiments did not prove successful. Chromic acid was tried, but this gave unsatisfactory results for the reason that it was impossible to control the action and at the same time secure complete oxidation of the phosphorus. With concentrated solutions the action is liable to become violent unless the temperature is kept low.

We also arranged an apparatus similar to that used in making phosphorus pentoxide on the small scale. The combustion took place in a bell jar filled with pure air, and after being thoroughly washed the gases were passed through baryta water. The operation was imperfect owing to the formation of red phosphorus, and to incomplete oxidation.

Finally we succeeded in obtaining satisfactory results by using nitric acid of 1.2 sp gr. The phosphorus was oxidized in a retort of 500 c.c. capacity. The retort was inclined so that any liquid condensing in the neck would run back. A glass tube fitted to the neck of the retort by means of gypsum served to convey the evolved gases into a wash bottle containing pure water. The latter was connected with a combustion tube containing in one end a layer of metallic copper about six inches in length, this served to decompose the oxides of nitrogen. The remainder of the tube was filled with copper oxide, which served to oxidize any carbon compound, which might be formed by the oxidation of the phosphorus, to carbon dioxide. After leaving the combustion tube the gases passed, first through a wash bottle containing water, then into one containing clear barium hydroxide, which was protected from the action of the air. All joints which were not of ground glass were made by means of gypsum. The operation was conducted as follows:—After 200 to 300 cubic centimetres of nitric acid (sp gr 1.2) and the weighted quantity of phosphorus had been introduced into the retort, a slow current of air free from carbon dioxide was drawn through the apparatus. The tubulus of the retort was then closed by means of a glass stopper, the combustion tube containing the metallic copper and copper oxide heated to a red heat, and a solution of baryta water rapidly filtered into the last wash bottle. The retort was then heated gently, after a time a regular evolution of gas takes place, and a precipitate gradually forms in the baryta water. At the end of the operation, air free from carbon dioxide is again drawn through the apparatus to remove all of the oxidation products. The precipitate is allowed to settle, the clear liquid is rapidly decanted through a filter. The precipitate is then washed, and quickly brought upon the filter paper. The filtering is done by means of a pump and is very rapid. The precipitate is then dissolved in dilute hydrochloric acid and the solution heated to boiling and the barium precipitated by sulphuric acid in the usual way. From the weight of barium sulphate obtained, the quantity of carbon in the phosphorus is readily calculated.

In some instances the experiment was varied by using a large quantity of phosphorus and allowing the action to continue for two or three hours, then weighing the phosphorus which remained unacted upon. In two instances the carbon dioxide was weighed directly by replacing the wash bottle containing the baryta water by weighed potash bulbs.

The following are the results obtained

I. 6.2272 grams Phosphorus gave .0300 gr BaSO₄ = .0057 grm CO₂ = .0016 gr C = .026%C.

II. 7.9545 grm Phosphorus gave .0324 gr CO₂ = .0088 gr Carbon = .111% Carbon

III. 8.8041 grm Phosphorus gave .0134 gr CO₂ = .00365 gr Carbon = .042% Carbon

IV. 9.0650 grm Phosphorus gave .0540 BaSO₄ = .0101 grm CO₂ = .00278 gr C = .031% C.

V. 16.4633 grm Phosphorus gave .1303 gr BaSO₄ = .0246 gr CO₂ = .0067 C = .041% C.

VI. 11 grams Phosphorus gave .0929 grams BaSO₄ = .0175 gr CO₂ = .00478 C = .043% C.