Two series of experiments were made. In the first series disks of steel 2.5 inches in diameter and 0.01 inch thick were employed. They were all cut from the same strip of metal, but some were "cold-rolled," some "annealed," and some "hardened." The total carbon was found to be: "cold-rolled," 1.108 per cent.; hardened, 1.128 per cent.; and annealed, 0.924 and 0.860 per cent. Some of the disks were submitted to the action of an oxidizing solution consisting of a cold saturated solution of potassium bichromate with 5 per cent. by volume of pure concentrated sulphuric acid. In all cases a blackish magnetic residue was left undissolved. These residues, calculated upon 100 parts of the disks employed, had the following compositions: "Cold-rolled" carbon, 1.039 per cent.; iron, 5.871. Annealed, C, 0.83 per cent.; Fe, 4.74 per cent. Hardened, C, 0.178 per cent.; Fe, 0.70 per cent. So that by treatment with chromic acid in the cold nearly the whole of the carbon remains undissolved with the cold-rolled and annealed disks, but only about one-sixth of the total carbon is left undissolved in the case of the hardened disk. The authors then give a resume of previous work on the subject.
In the second part they have investigated the action of bichromate solutions of various strengths on thin sheet-steel, about 0.098 inch thick, which was cold-rolled and contained: Carbon, 1.144 per cent.; silica, 0.166 per cent.; manganese, 0.104 per cent. Four solutions were used. The first contained about 10 per cent. of bichromate and 9 per cent. of HSO by weight; the second was eight-tenths as strong, the third about half as strong, the fourth about one and a half times as strong. In all cases the amount of solution employed was considerably in excess of the amount required to dissolve the steel used. A residue was obtained as before. With solution 1, the residue contained, C, 1.021; sol. 2, C, 0.969; sol. 3, C 1.049 the atomic ratio of iron to carbon was Fe 2.694: C, 1; Fe, 2.65: C, 1; Fe), 2.867 C, 1): sol. 4. C, 0.266 per 100 of steel. The authors conclude that the carbon in cold rolled steel exists not simply diffused mechanically through the mass of steel but in the form of an iron carbide, FeC, a definite product, capable of resisting the action of an oxidizing solution (if the latter is not too strong), which exerts a rapid solvent action upon the iron through which the carbide is distributed.