quality of steel, because we always began by making pure soft iron. From the zero point of decarburisation the hardest qualities of steel could be made, differing by almost imperceptible gradations, and depending on the number of pounds of rich carburet of iron added to the pure iron for that purpose.

The material had been proved in all respects suitable for the manufacture of ordnance, and, as I have already said, Colonel Eardley Wilmot and I had arranged, under contract, to erect a Bessemer plant in the old gun foundry at Woolwich, which was amply large enough for that purpose. This project, had it been carried out, would have rendered wholly unnecessary the erection of a second arsenal at Elswick, built under the guarantee of the British Government at a cost of £85,000. It must also be borne in mind that by my process we had the advantage of being able to make, if desired, malleable iron guns in a single piece without a weld or joint, by founding, or by the combined processes of founding and forging, with or without hoops; so that if malleable iron, and not steel, had in reality been the best material for the construction of ordnance, such guns could have been produced at Woolwich Arsenal, either as complete gun-castings, or as ingots to be forged, at a cost not exceeding £6 or £7 per ton if made of British iron, and not exceeding £10 per ton if made of Swedish charcoal pig-iron; whereas the Lowmoor iron bars used to make the coiled guns cost over £20 per ton, and were the mere raw material to start with. Nor did the Bessemer pure malleable iron, if used for guns, admit of any of the charges that had been made to depreciate the value of steel for that purpose, namely, that it was very uncertain in quality, and could not be obtained of the precise degree of carburisation and toughness required.

Such a charge could not possibly be made in reference to pure iron, which was wholly decarburised, a condition which it was impossible to mistake during its manufacture, for the huge white flame issuing from the converter suddenly drops when all the carbon is burnt out, a result which occurs with unerring certainty. At all events, if Bessemer steel could not be depended upon at Woolwich, Swedish charcoal pig-iron, wholly decarburised, could have been made in masses of 10 to 20 ton, at a cost not exceeding £10 per ton, and of, at least, 5 tons per square inch greater tensile strength than Lowmoor bars, as was proved by Colonel Wilmot's experiments at Woolwich Arsenal; while the cost of the huge unwelded mass would have been less than half the cost per ton of the bar-iron used to make a welded coil with its many imperfect junctions.

I should like to say a few words here about the broad distinctive characters of the two materials, wrought or bar-iron, and cast homogeneous iron or steel. I need scarcely remind the reader that bar-iron making begins with the process of puddling, which produces a ball or mass of iron that, in every case, is mechanically mixed with fluid scoria, and sometimes with sand and dry oxide or iron scale. From this crude material, puddle bars are made, and these are cut into lengths of 2 ft. or 3 ft., and formed into a bundle or pile, which is brought up to a welding heat in a suitable furnace, and then rolled into a merchant bar. This process of rolling and piling is repeated more than twice, and a bar is in this way produced, which to the eye appears, and is supposed, to have all its separate parts welded or united so as to form an undivided and indivisible mass. But this is not so. I have never seen a bar of wrought iron produced by puddling that, in two or three minutes, by a very simple treatment, I could not separate more or less perfectly into its component bars, which are in reality never thoroughly united, although they adhere more or less soundly. In fact, so imperfect is this adhesion called "welding," that whenever bar-iron is worked under the hammer, it is necessary to forge it at such a degree of heat as will continue the welding process; for by working it much below this temperature, the imperfectly coherent mass begins at once to separate at all the junctions between the several bars of which it is composed, and tumbles to pieces.

I will describe an experiment clearly illustrating this fact. Two pieces of ordinary commercial bar-iron of 1 in. square were heated to a blood-red heat, and put under a small steam-hammer, where they received several blows on alternate sides; the result was a complete disintegration of the mass, as shown in Fig. 62, Plate XXIV. The lower example was similarly treated on alternate angles, instead of on the flat sides. It may be supposed that the far-famed Lowmoor and other Yorkshire irons are exempt from this defect, but this is not so, the simple fact being that "best-best" iron has been piled more times than common iron, and the result of working it at a temperature that will not continue the welding process, only divides it into more numerous filaments than a bar of common iron. I may mention the fact that, on one occasion,