phosphorus. The weight of the charge, the temperature and moisture of the air, must necessarily affect the result somewhat; therefore the foreman first notes the number of revolutions up to the point when the spectroscope tells him that all the carbon has been burnt, and compares it with the estimated number. The difference, if any, is then applied to the calculated number of extra revolutions required for removing the phosphorus, and it depends upon the correctness of these several determinations whether the metal is pure or not. Even the temperature of the charge influences them, for the colder it is the more difficult is it to judge of the disappearance of the carbon lines; so that, generally speaking, it is very easy to be mistaken as to the time when the blast should stop. A danger to which the charge is also exposed is that due to the action of the highly phosphoric slag on the added spiegeleisen and ferro-manganese. It has been confidently stated, but not proved, that the affinity of carbon for oxygen, and of phosphorus for iron, is so great that, unless the latter stage of the process is hurried, phosphorus will leave the slag and re-enter the metal. This seems to be the reason why some works put the admixtures into the ladle and not into the converter. Where this is done the question arises whether the charge can then be thoroughly mixed. If not, this would explain occasional irregularities in the finished product, and these being more apparent in large plates than in bars and wires, may account for the strong dislike entertained towards basic Bessemer plates.

The Acid Bessemer Process differs from the one just described in so far as the lining of the converter is acid (silicic acid = ganister) and that no lime is added. The slag is generated by the combustion of the silicon (to silicic acid) and the iron (to iron oxide). Not a trace of phosphorus is removed by this process, and therefore the pig iron used must contain less than that to be allowed in the finished product. It must also contain a large percentage (2-3 %) of silicon for the production of sufficient heat.

The addition of spiegeleisen and ferro-manganese is effected just before casting, either in the converter or in the ladle.

The spectroscope enables the operator to judge, with reasonable accuracy, as to the percentage of carbon remaining. He could, therefore, either interrupt the blowing before all the carbon is burnt, which would save some of the costly admixtures, or he could wait till all the carbon has disappeared, and then reintroduce it with the necessary manganese. The latter of these two methods is the more reliable. An incidental difference between these processes is that in the basic process, during the necessary after-blow for removing the phosphorus, every trace of carbon disappears, while in the acid process this is not the case, so that by adding just sufficient ferro-manganese to remove all redshortness one would obtain a far weaker but also a more ductile material by the basic than by the acid process; the former is therefore used almost exclusively for producing the steel for soft wire.

The Open-Hearth Process was made available for producing mild steel only after Dr. Siemens had invented the regenerative chamber, and although the various shapes of the furnaces take the names of their respective designers, his name will always be associated with steel made by this process.

Heat regenerators consist of several vaults filled with loosely packed firebricks. Air is admitted through one of these chambers, and gas through another, and both are ignited in the furnace. The waste products are led to the chimney through the two other chambers, and heat them. The current is then reversed, and the air and the gas pass over the white-hot bricks of the latter chambers, and the waste products heat the former. The process is then continually repeated. Not only does this save nearly all the heat, which would have escaped with the hot gases, but the temperature of the furnace can easily be raised so high that even the firebrick lining would melt.

Within certain limits the temperatures can be regulated by admitting more or less excess air, and by altering the frequency of reversing the current. The flame might also be altered from an oxidising to a reducing one, but not without raising the temperature beyond the endurance of any firebrick. The gas used for firing these furnaces is produced at the works either by partly burning and partly distilling coal, in which case it consists of hydrocarbons, carbonic oxide, and nitrogen; or water gas is used, which contains carbonic oxide, hydrogen, and a smaller percentage of nitrogen. Recently part of the heat in the waste products has been used to distil the coal.

The Acid Siemens-Martin Process.-Pig iron is placed in the furnace, and when melted, iron ore and about 25 % scrap iron are added, until all the carbon and silicon are consumed, and the ore reduced to iron. Samples are repeatedly taken and tested mechanically, to judge of the condition of the bath, and when ready, spiegeleisen and ferro-manganese are added, and the charge is run. As in the Bessemer process, some of these additions may be saved by stopping the process of reduction at an early stage, and in some works the carbon is successfully reintroduced by adding it as powdered charcoal, or anthracite placed in the ladle or into the trough leading to it from the furnace.

During the early period of refining the various layers of the bath are of very different composition, as can easily be ascertained by taking a sample from the top or the bottom of the charge; but a natural mixing takes place, due to chemical action and to an evolution of gases. However, at the final stage, when the bath is nearly but not quite uniform, this action practically ceases. It is revived by the addition of iron ores, or of pig, or in some works by stirring with wooden poles, which evolve large quantities of gas.

The acid Siemens process does not remove either phosphorus or sulphur, and the pig and scrap used should therefore contain only traces of these impurities.

The Basic Siemens Process, using Pure Pig, is almost identical with the above, except that, on account of the use of a basic slag, almost every trace of phosphorus disappears. It is also found that both the carbon and manganese are very energetically attacked by the flame, and after the spiegel has been added, it is difficult to hit the right moment for running the charge. This difficulty is aggravated by the fact that the natural lowest limit of mild steel from these furnaces is below 20 tons, whereas with the acid Siemens furnace it is about 24 tons. When trying to make steel of 27 tons, a small

error in the admixtures, or in the time of casting, will have two to three times as much effect in the one case as in the other. However, with care the very best material, and of the intended hardness, is obtained.

On account of the costliness of the basic lining this process is hurried as much as possible, and six or even seven charges are sometimes got out of a furnace in twenty-four hours. Of course this is only possible if the refining process is curtailed, and with this object in view not more than about 20 to 25% of pig iron is used, while the rest is scrap. A further gain has been attempted by carrying out part of the process in a Bessemer converter where all the carbon and silicon are removed. The final reduction takes place in the basic Siemens furnace, where the last trace of phosphorus is abstracted.

Basic Refined Steel.'-C. E. Stromeyer, Eng. Scot., 1897, vol. xli. p. 227, deals fully with this and the next material. E. Bertrand, I. and S. I.,' 1897, vol. i. p. 115.

The Basic Siemens Process, using Phosphoric Pig, is carried out at a few works in this country. It is a tedious one, lasting about fifteen hours per charge; this is made up of iron ores, lime, and about 75 to 80% of pig, too poor in phosphorus for the basic converter and too rich for the acid Siemens process. The phosphoric acid which is generated is not volatile, and does not rise to the surface in bubbles like carbonic acid, and therefore cannot assist in mixing the charge, so that the process would be indefinitely prolonged if special means were not adopted, such as the occasional addition of pig and of iron ores, to produce gases. This is necessary up to the last stage, which, as in all other cases, consists in adding spiegeleisen and ferro-manganese.

The basic slag, which floats on the molten steel, is very thick, generally about 12 to 15 ins., and the furnace gases have little chance of acting on the metal. Few works use this process, and there all attempts to manufacture plates on which the same reliance can be placed as on the acid steel have as yet failed. With care good results could no doubt be obtained, but it seems that the effects of even slight carelessness on the part of the operator lead to bad results, and no tests have yet been devised which will detect them. Herr Knaut (Stahl und Eisen,' 1900, vol. xx. p. 783) has put about 500 tons of this basic steel into boilers, and after ten years' experience condemns it.

The Puddling Process has been dying out fast. In it the pig iron is melted in the presence of iron ores and slags by means of flames, which can be changed from oxidising to reducing ones as occasion arises. The temperatures, being very much lower than in all the previous processes, appear to assist at removing a large percentage of phosphorus and sulphur. Much hard labour and considerable skill are required to obtain good results. The final operation consists in extracting the intermixed slag from the iron, which is done under a blooming hammer.

Crucible Steel, as its name implies, is produced in crucibles. These are filled with carefully weighed quantities of blister steel, pure iron, or scrap steel. It is said to be giving way to Siemens steel even for guns, and as it is a very costly process, and has never been

used for boiler plates, it is unnecessary to enter into details, except to mention that when carrying out experiments on steel alloys it is of importance that the right fire-proof material should be used. Thus, when wishing to produce practically pure iron, basic crucibles must be employed; if a good percentage of carbon has to be retained, and for many alloys (manganese, silicon, aluminium), the melting must be done in plumbago, and for some other purposes acid crucibles are best, while for certain compositions the melting has to be carried out in various fire-resisting materials, and the molten metal mixed before casting.

The following books will be found to contain a very full account of the manufacture and properties of mild steel :-Dr. J. Percy, 'Iron and Steel,' London, 1864; M. H. Howe, 'The Metallurgy of Steel,' New York, 1890; Professor Ledebur, Handbuch für Eisenhüttenkunde,' Leipzig, 1884; V. Deshayes, 'Classement et Emploi des Aciers,' Paris, 1880; J. S. Jeans, Steel,' &c., 1890; Prof. A. Martens, Mitt. Berlin,' 1890, vol. ii.

The Influences of Impurities on the mechanical properties of mild steel. It may be remarked that absolutely pure iron or steel has not yet been produced and experimented upon, and the following remarks, therefore, only apply to the effects of additions made to average qualities.

The

Recently the subject has shown itself to be more complicated than was at first assumed, but it is hoped that microscopic researches will elucidate the matter. Thus it is now known that carbon and iron combine in various proportions, forming compounds which are dissolved throughout the mass of steel. These compounds can be distinguished under the microscope and also by attacking the metal with various acids, when the gases given off or the residues enable one to estimate the amount of carbon in its different conditions. These compounds give varying qualities to steel. Phosphorus behaves in the same way, and instead of always producing coldshortness, it is now believed that only one of its ferric compounds does this, which when treated with certain acids evolves PH3. phosphide of iron, on the other hand, is not dissolved in the mass of metal, but exists in the form of minute crystals, which, being very hard, cause the edges of cutting tools to get blunt. This explains why some steels with much phosphorus are not coldshort. It has not yet been shown how the phosphides can be changed. Probably sulphur has the effect of dissolving the phosphide of iron, and therefore accentuating coldshortness. It is known that carbon can be affected by various impurities. Silicon, tin, and copper seem to drive it out of the iron. Manganese, chromium, and tungsten seem to assist at dissolving it, the two latter converting it into hardening carbon.

For action of acids, see 'I. and S. I.,' 1888, p. 369, 1896, vol. i. p. 239, 1897, vol. i. p. 229; Comp. Rend.,' 1897, vol. cxxx. p. 148.

Carbon increases the tenacity and tempering qualities (if it does not even create them); it reduces the ductility and weldability and melting temperature.

Phosphorus has ascribed to it the chief blame for coldshortness and general treacherousness, but it seems that this is true only if

much carbon or sulphur is present. It also increases the liability to get burnt. It reduces the melting temperature.

Sulphur accentuates the bad effect of phosphorus, produces redshortness and greasiness as regards welding.

Arsenic increases the tenacity, reduces elongation and tempering qualities, increases coldshortness as well as redshortness, but only slightly. In chemical analysis it has been mistaken for phosphorus. Like phosphorus and carbon, it reduces the melting temperature.

Silicon reduces the elongation and melting temperature, prevents blowholes, increases tenacity and red- and coldshortness, but only in presence of carbon. It does not accentuate the evil effects of phosphorus.

Tin.-Authorities are conflicting. It seems to drive carbon out of cast iron.

Aluminium prevents blowholes, reduces the melting temperature, and hardens cast iron.

Manganese intensifies the influence of carbon, except as regards tempering properties, and neutralises the red- and coldshortness of phosphorus, sulphur, &c.

Nickel reduces tenacity and increases ductility, particularly as regards impact. It neutralises the influence of carbon, and perhaps of phosphorus, and it reduces corrosion.

Chromium, like Tungsten, intensifies all the influences of carbon, particularly as regards the tempering qualities.

Copper prevents blowholes, increases tenacity, reduces elongation, and produces redshortness, but only when above 3 per cent. Its influence on welding is doubtful.

An attempt to summarise these remarks is made in the following table:

Carbon Silicon Arsenic Phosph. Sulph. Copper Mang. Nickel Chrom.

,, tempered

+

+?

0

NOTE.-0 means that the property is not changed, that it is increased, that it is diminished by the impurity: ? means that the authorities are very conflicting; ... means that no information could be obtained.

More detailed information, particularly as to the amount of changes which these various impurities may produce, can be gained by consulting the following papers; but unfortunately where the information is precise the authorities are often in conflict, while generally their experimental results are exceedingly vague. This is due to the difficulty experienced in obtaining pure materials, to the accentuating and neutralising effects of the various impurities on