METHODS OF STEEL PRODUCTION, The Besseraer-Husket Process and its Precursors. - It has been known since the beginning of the 18th century at least that steel could be prepared by fusing together in crucibles cast and wrought iron ; thus the operation was performed in 1722 by 116aumur employing the heat of an ordinary forge ; whilst in the production of wootz it must have been observed centuries ago that by continued heating the badly prepared cakes (consisting partly of fused steel and partially of unfuseti iron) frequently resulting from the first crucible operation could be fused into one mass of somewhat less hard steel than that produced at first in the more successful operations. The possibility of producing steel by fusing together a malleable and a carbonized iron is evidently a simple deduction from the processes whereby a steel is produced by the direct addition of carbon to malleable iron, e.g., those of Mushet and of Heath (§ 35). Accordingly a few years after the latter patented the use of "carbnret " of manganese, he proposed (1845) a method for making steel on a larger scale that crucible operations would permit, viz., by fusing in a cupol; pig iron, running this into the bed of a steel-making furnace into the upper part of which the malleable iron was intro duced in bars so as to be heated up by the waste heat am gradually pushed forward so as to dissolve, as it were, i; the molten pig with formation of steel. This method i described by Siemens as being one which would doubtlea hare led to complete success had the regenerative principl been known to Heath, so as to enable him to obtain thi requisite intensity of heat and absence of cutting draugh essential to the proper combination together by fusion o the wrought and carbonized iron without oxidation ; it i: substantially one of the forms of steel making by mean: of the open hearth or regenerative processes now in use and known collectively as Siemens or Siemens-Marti] processes (see § 39). Other patents, amounting substan tially to. the same combination of wrought and cast irm by fusion so as to form steel, have been subsequently take; out by Price and Nicholson (1855), Gentle Brown, am Attwood (1862) ; a particular combination of this clan patented by Mushet in 1855 (consisting of the addition ti molten Bessemer blown decarbonized iron of fused spiegel eisen) has proved of the highest practical value (§ 27) notwithstanding that the non-removal of phosphorus an( sulphur to any marked extent in the ordinary blowini process render it applicable to certain kinds of pig iroi only; the recently invented "basic" process, however, bid fair to overcome this difficulty (see § 37).

The earliest form of converter patented by Bessemer, October 17 1855, consisted of a rectangular furnace with firebars at the sid instead of at thebottorn, so that a number of crucibles could be heater therein, each furnished with a tapping hole at the bottom, and pipe dipping to the bottom of the fused metal inside, through wide] air was to be blown, or a mixture of air and steam, the forme causing the temperature to rise, the steam having a cooling effect Two months later another patent was taken out, the use of spherical or egg-shaped vessel of iron lined with firebrick and sup ported by axes being the main novelty. In May 1856 a fixe( vertical cylindrical vessel, with blast pipes at the base and a tappirn hole, was patented ; not till some time later, however (after the introduction of Mushet's improvement of adding spiegeleisen), was the present form of converter arrived at, consisting of a pear-shaped or bottle-shaped vessel with tuyeres at the base, and supported upon axes, one of which being hollow serves as the tube by means of which the blast is communicated to the tuyeres, whilst the vessel can be rotated into any desired position round the axis (fig. 60) ; the use of hydraulic power to work the converter into position, and to manipulate the "ladle" into which the finished metal is run so as to be poured from it into moulds, soon followed. Subsequently various improvements in details of arrangement and construction of the plant have been introduced, notably by Holley in certain American works in the first instance ; amongst these may more particularly be noticed the use of somewhat longer converters (fig. 61), with movable bottoms, so that, as the tuyeres wear out (which occurs much more rapidly than does the destruction of the lining), new previously constructed bottoms can be introduced without entailing large loss

Converters. - The improved form of converter in use at the present day is indicated by fig. 61. It consists of a vessel in shape resembling a claret bottle with the neck somewhat shortened and slightly bent over sideways; this is made of stout boilerplate, lined internally with "ganister," a kind of sandstone possessed of peculiar physical properties, more especially of the power of binding together to a compact non-crumbly excessively infusible mass when ground to powder, moistened, and finally exposed to a high temperature. The base of the bottle is removable, being kept in position by bolts and nuts ; this being removed, a core is fitted concentrically within the shell, and the moistened ganister rammed lightly in ; sometimes the shell is also made of two parts thus treated separately and afterwards bolted together. The bottom is double, constituting an air chamber or " blast box," the top portion of which is a perforated cast iron plate, into which are fitted tuyeres consisting of slightly conical elongated bricks or plugs perforated longitudinally with holes (between 0.15 and 0.3 inch diameter) and fixed in at the thicker end ; these are composed of a mixture of fireclay and ganister, the space between them being filled up with ganister, so that the inner bottom is substantially the same as the lining sides of the converter, only perforated by a number of holes. The total number of orifices through which the blast thus comes is considerable, in order to spread it into numerous streams, the actual number varying with the size of the converter ; for a 5-ton converter a dozen or more tuyere bricks, each perforated with about as many holes, are usually employed. The lower plate of the air chamber is removable for the purpose of examining the tuyeres from time to time without removing the entire bottom.

The different portions of the vessel being put together, and the joints well grouted with ganister slip, the whole is gently dried by lighting a small coke fire inside, and by and by increasing this, the blast being turned on gently until the whole vessel is thoroughly dry and hot, when it is ready to receive a charge of molten pig. The blast is communicated to the blast-box by a curved tube reaching thereto from the hollow trunnion box, inside which one of the trunnions of the converter lies ; the other trunnion has a cogged wheel or pinion attached, by means of which the vessel can be rotated, a toothed rack gearing into the pinion and being worked directly by a hydraulic press or ram ; the rack and press were formerly arranged horizontally, but now are usually placed vertically or slantwise to save space. By means of an automatic valve the blast is shut off when the converter is in an inclined position, so that the level of the mouth is not lower than the tuyere hole then lowest ; in this position the molten metal is run in, the elevation of the tuyere holes preventing it from running into them ; whilst erecting the converter by means of the rack and pinion before the tuyeres sink below the molten mass, the blast is turned on at a pressure of some 1.5 atmospheres or more (21 lb per square inch and upwards) ; the pressure due to the column of molten metal being less than this, the air is forced through it without it being possible for the fused matter to run down into the blast-box. A large hood over the mouth of the converter and connected with a chimney or flue prevents the flame and ejected matter from being scattered about and injuring the workmen.

The ganister preferred for lining is a peculiar silieious deposit found under a thin eoal-seam near Sheffield, of almost eonchoidal fracture, therein differing from ordinary sandstones, and containing a few tenths per cent., or sometimes a little more, of lime and about the same amount of alumina, with small quantities of iron oxide and alkalies, the rest being silica ; analogous substances, however, are found in various other localities, e.g., in the Yorkshire, Northumberland, South Wales, and Shropshire Coal Measures. A well-prepared lining carefully repaired every few days will last several months, and even upwards of a year ; the bottoms, however, wear away much more rapidly, the tuyeres either becoming melted or dissolved away by the oxidation of the iron in their vicinity, the oxide produced exerting a marked solvent action on the fireelay and ganister bricks ; hence the advantage of Holley's removable bottoms. A damaged bottom can be removed and a new one put in its place, the joint being made up by ramming in ganister into the crevice whilst screwing up the bolts, without stopping the action of the converter for any lengthened period ; whilst with the older vessels it was requisite to cool down somewhat, knock out the damaged tuyere bricks, replace them by others, and ram in ganister between them from the. inside, arid finally to heat up again by an interim) fire before use, - the whole process causing much delay and waste of fuel, especially as the renewal of some at least of the tuyere bricks is requisite every few blows. In some American works, e.g., Bethlehem, instead of a rammed ganister lining, one composed of blocks of a peculiar sandstone is employed, set in ganister as mortar for the joints.

Subsidiary Appliauees. - The precise mode of arrangement of Bessemer plant varies in different works, but usually two converters are arranged to be worked together, one casting ladle being used for the pair, worked by a crane. In most of the European works the two converters are on opposite sides of the casting ladle, so that a large portion of the circle covered by it cannot be used for running ingots into moulds, being occupied by the converters ; in most American works the two converters are placed side by side, so as to leave a larger space for the casting bed when the metal is not used direct from the blast furnace. A range of cupolas is fitted up at a convenient distance, some larger for melting the pig, somo smaller for the spiegeleisen ; the molten metal is either run out from these direct into one or other of the converters through a shoot or gutter of iron lined with fireclay and sand, or is tapped into an intermediate collecting ladle, and when the requisite quantity is collected run rapidly into the converter either by "tipping" the ladle over so that the metal runs out together with some little amount of slag floating on its surface, or by means of a tapping hole at the bottom. This latter arrangement allows the charging of the converter to be more quickly effected ; when the iron is tapped directly from the blast furnace into the converter by means of an intermediate ladle, the cupolas are of course unnecessary, save those for melting the spiegeleisen. The blast is generated by an ordinary blowing engine, but at a considerably greater pressure than that used for blast farnaces, 20, 25, and even 30 lb per square inch pressure being employed. The casting ladle into which the contents of the converters are emptied by tilting them up sufficiently when the operation is finished is a large iron bucket lined with clay with a hole at the bottom filled up with a perforated firebrick, into which fits a stopper consisting of a stout iron rod covered with a thick fireclay tube to protect the rod from the fused steel ; when the ladle is full of molten metal, and the hole is closed by the stopper, the ladle'is swung round by means of a crane over the ingot moulds (of cast iron) ; the stopper is then lifted, when the fused steel runs out, the seorke floating on the top of the metal being thus retained and pure steel only poured. Sounder ingots are obtained by running the metal into a cavity communicating by firebrick tubes with the bottoms of the moulds, so as to fill them from below, than by filling them from above directly. Detailed descriptions of the plant employed in various first-class American Bessemer works are given by Holley and Lenox Smith in a series of articles in Engineering, 1877 and following years.

The mode of carrying out the operation is briefly as follows :the charge of pig iron being run into the converter, this is swung back into the vertical position, the blast being automatically turned on in so doing ; when the blowing is at an end, the converter is turned into a nearly horizontal position, the blast being thereby shut off ; a weighed quantity of fused spiegeleisen is then run in, and the total contents of the converter forthwith poured into the casting ladle ; formerly the converter was erected for a few seconds and the blast blown through to mix the spiegeleisen and blown metal, but that is now found to be unnecessary. The ingots are filially reheated and passed through the rolling mills after forging so as to reduce them to rails, bars, plates, &c., as required, the machinery for this purpose being identical in character with that employed for malleable iron (§ 25).

The following analyses illustrate the general composition of " Bessemer steel " as made in different European countries, the first two specimens being prepared by Bessemer's original process (without addition of spiegeleisen, § 27), and the others by the BessemerDiushet combination process (from report of E. Brusewitz to the Swedish iron office, Jern-Kontarets Annaler, 1871, 199).

be that when phosphorized metal is in presence of iron oxide in a fused state, or of a melted mixture of iron oxide and ferrous silicate containing an amount of the latter not above a certain limit, the tendency of the phosphorus is to become oxidized and converted into phosphate, which separates in the cinder, leaving a purer iron ; whilst on the other hand if the cinder is mainly silicate, especially of the metasilicate or " acid " type (R20, SiO2), the tendency is rather the other way, part of the metallic iron becoming oxidized whilst the phosphate is reduced, thus communicating phosphorus to the remaining iron.1 Accordingly, in the modern ordinary puddling process, especially when machine puddling and regular mechanical agitation are substituted for hand labour, and when plenty of fettling of ferric oxide (not largely silicious) is employed, and in Bell's and Krupp's purification processes, phosphorus is largely removed from the pig ; whilst on the other hand in the older method of puddling on sand bottoms and in the ordinary Bessemer blowing process the first action of the oxygen of the blast is concentrated rather on the silicon than on the iron (at least so far as the ultimate chemical change is concerned), and in consequence a highly silicious cinder results, so that removal of phosphorus by oxidation becomes impracticable ; this result, moreover, is intensified by the nature of the lining material (ganister) used for the converters ; similarly, in the Heaton process phosphorus is sometimes removed to a considerable extent and sometimes not, according as the soda produced by the decomposition of the nitrate and the iron oxide formed modify the character of the slag produced by the oxidation of the silicon and formation of silicates, rendering it of an " acid " or " basic " silicious character. Again, according to Riley, when the whole of the iron is reduced in the blast furnace, so that the cinder contains none, or practically none, the pig contains all the phosphorus present ; but if the cinder contains unreduced iron to any extent, it also retains a proportionate amount of phosphorus, being then much more basic. The temperature also seems somewhat to influence the reaction of iron oxide on phosphorus and of iron on phosphate ; the higher the temperature the more pronounced apparently is the tendency of the metal to retain phosphorus, i.e., the less is the tendency of the phosphorus to oxidize and separate from the metal. Guided by these considerations, Snelus took out a patent in 18722 for the use of lime or limestone as a lining material for the converter, and found that when a " basic " lining of this description was substituted for ganister a removal of phosphorus to a greater or lesser extent was actually brought about in the converter just as it is in the puddling furnace ; for some few years, however, this method was not applied on a commercial scale, but subsequently the subject was again examined by Thomas and Gilchrist, who finally succeeded in reducing the principles of Snelus's patent to successful practical operation, findinc, that by the use of a " basic" lining to the converter, and especially the addition of a small amount of lime, or lime mixed with "blue billy" or some other form of iron oxide such as mill scale, to the charge together with the continuance of the blow for some short period after the decarbonization is complete, the elimination of phosphorus (even from highly phosphorized pig, containing 1.5 to 2.0 per cent. of phosphorus) could be very largely effected, some 80 to 90 per cent. at least of the total phosphorus present becoming oxidized and converted into phosphates, this action chiefly taking place during the "after blow"; provided that the cinder is sufficiently basic, the iron does not oxidize during this after blow as it does in the ordinary "acid lined" converter process. Owing to the success of these operations, the " basic " process has been more frequently spoken of as the " Thomas-Gilchrist process" than as being what it substantially is, the principle of Snelus's earlier patent in a somewhat but not very largely modified form. Warned probably by the disasters of previous inventors, the rival competitors for the honour (and profit) of the practical production of ingot metal of fair quality from phosphorized pig have coalesced and united their forces, instead of opposing one another in costly litigation.

Processes based upon the general principle of making the cinder of the Bessemer converter more or less " basic," but considerably different in details from the basic lining method of Snelus, Thomas, and Gilchrist, have been proposed at one time or another by various inventors ; thus, in the earlier patents of Bessemer himself, the admixture of steam with air in the blowing operation was included, whilst steam had been previously employed as an adjunct in the refinery, the action being the formation of iron oxide (with evolution of hydrogen). Somewhat analogous uses of steam have been subsequently proposed by others, in some cases the phosphorus being alleged to be evolved as phosphoretted hydrogen (?), c.g., in Bull's process, in which the ordinary air blowing of a Bessemer converter or the effect of an oxidizing blast in a Siemens hearth, &c., on cast iron is first used to remove silicon and carbon, and then a jet of iteam used to remove phosphorus. The direct incorporation of oxide of iron (blue billy, &c.) alone with the material (by blowing into the converter along with the blast) has also been proposed by Pettitt ( § 24), the object in all cases being to assist the formation of phosphate, and thereby remove phosphorus from the iron by making the cinder highly basic.

During the oxidation of phosphorus a large amount of heat is evolved, so that the temperature rises during the after blow just as it does during the oxidation of silicon. According to various experimenters the heat of combustion of phosphorus is a little short of 6000, so that, as that of silicon is near to 8000, 4 parts of phosphorus are thermally about equivalent to 3 of silicon. Hence if a pig containing 2 per cent. of silicon will furnish sufficient heat to keep the blown metal and slag fluid, the same result will be attained by means of a pig containing about 0.5 per cent. of silicon and 2'0 per cent. of phosphorus, i.e., if the radiation loss be the same, and also the diminution in heat evolution clue to separating the non-metal from the iron,' as well as the fusibility of the slag. In actual practice the "basic " method is found to take a little longer time than the ordinary "acid" blow, so that the radiation loss is a been proposed to acid the basic matter in a highly heated or even molten state, and to make the air used for blowing traverse a chamber filled with coke on to which petroleum is allowed to drop, so that combustible vapours are blown into the converter along with the air, thus raising the temperature considerably, and preventing the blocking of the converter-mouth by solidification of cinder, and the production of metal too little heated to remain fluid during the pouring of the ingots, and consequently solidifying in the ladle forming a "skull." Wilks finds that the action of this arrangement is very satisfactory and effective in preventing " cold blows" from occurring. The same result is also producible by means of coal dust or other freely divided combustible matter blown in along with the blast.

With silicious pig iron the lining is apt to be attacked by the silica formed during the first part of the blowing operations ; this is partly but not wholly avoided by the addition of heated lime to the charge. Harmet has proposed, and at Witkowitz attempts have been made, to blow phosphorized and siliciuretted pig successively in two converters, one acid-lined for removing silicon and carbon, and the other basic-lined to get rid of the phosphorus, the blown metal being transferred from the one to the other ; much cooling of the metal was thus produced, and the operation was prolonged some fifteen minutes, so that the method was speedily abandoned.

According to Tiinner, in order to produce 100 parts of pure ingots of blown metal 122.5 parts of pig are requisite with " basic " converters, and only 118/ with the ordinary " acid " ones (waste in remelting being included in each case ; when the metal is used direct from the blast furnace, only 112 parts are required with acid. converters) ; hence an extra less of nearly 4 per cent. of metal is experienced in the basic process. The following analyses, by Muller, of specimens taken during a prolonged investigation at Horde illustrate the non-removal of phosphorus and sulphur during the earlier part of the blow, and indeed their slight increase (in percentage amount) owing to the oxidation of manganese, silicon, and carbon ; and also the rapid removal of phosphorus during the after-blow The composition of the slag produced during the basic process differs greatly from that of the ordinary method ; the following analyses illustrate the difference : - The phosphoric anhydride in the basic process slag appears to exist as calcium phosphate, and not as iron phosphate ; for after roasting to peroxidize iron, no soluble phosphate is dissolved out on digestion with ammonium sulphide, nor is any formed by fusion with sodium chloride, whilst sulphurous acid solution readily dissolves out phosphate (Pattinson and Stead).

Bask Lining Material. - Some of the earliest of Bessemer's experiments (made at Dowlais) on his air-blowing process were made with a converter lined with Stonrbridge firebrick, and in this vessel a Bessemer metal was made which contained only minute quantities of carbon and silicon, and much less phosphorus than the batches subsequently prepared from the same phosphorized. Welsh pig in other converters lined with ganister, the non-success of which substances as commercial products (owing to their brittleness clue to the phosphorus present) ultimately led to the use of hematite and other phosphorns-free pig only for Bessemerizing. Without doubt the superiority of the earliest samples was due to the less silicious character of the brick lining. The lining materials used by Snails in his experiments on the subject were lime and limestone, especially magnesian limestone ; those first employed by Thomas and Gilchrist were crushed limestone and sodium silicate solution added so as to moisten the limestone, which was then rammed into the converter like ganister. Bricks made of a limestone containing some alurnmous silicate and fired at a very high temperature were then employed, but were found to be subject to disintegration by moisture when kept in stock ; magnesian limestone bricks containing a little silica and alumina similarly prepared answered much better. Riley has found that freshly burnt pretty pure magnesian limestone mixed with about 10 or 15 per cent. of crude petroleum can be rammed into converters or moulded by hydraulic pressure into bricks, the substance in either case becoming compact and hard upon firing, so as to present a very durable lining material. Instead of petroleum, crude shale oil, resin oil, or tar may be used to moisten the lime. Tho bricks thus prepared do not slake or disintegrate on keeping, or even if immersed in water, provided they have been fired at a high temperature ; in this ease they contract slightly, whereas if only comparatively slightly heated they do not materially diminish in volume on firing.' into a rapidly revolving wheel in a water tank. Only hard steel for miners' picks, cutting tools, razors, dies, &c., were thus prepared, containing 01 to 1.3 per cent. of carbon, about 50 lb being melted in each crucible at one operation ; the crucibles lasted longer than in ordinary cast steel making, usually for some half dozen heats.

Rostaing effected the granulation of the iron by running the melted pig on to a rapidly revolvincr° horizontal cast iron disk, so that the liquid was scattered around (by centrifugal action) in globules which solidified into a kind of iron shot ; these being sifted into tolerably uniform sizes admitted of the production of a more nearly homogeneous metal than would otherwise have been obtained. The principles of Uchatius and Rostaing's patents, however, were long previously known and even patented ; thus in 1731 J. Wood obtained a patent for reducing iron " into small grains (according to art) by pouring it into water upon a wheel or roller turned briskly round, ' and for decarbonizing the granulated metal by fusing with various fluxes, lie., including iron cinder, scales, and scoria: ; whilst Mushet has described. as being in use long ago Cyfartha, method of granulating iron pouring on to a revolving horizontal stone in a water trough molten pig iron, stating that the granulated metal was fused with oxide of iron in the form of bloomery cinders, and thereby rendered less easily fusible and capable of welding,in short, that it was decarbonized to a greater or lesser extent. The method of purification or refining of pig iron proposed by Bell (§ 24) is substantially the Uchatius process applied on a large scale and at a somewhat lower temperature, with mechanical agitation and interruption at a certain stage. By stopping the process when the action has only gone on sufficiently long to oxidize the majority of the silicon and phosphorus without n:aterially affecting the carbon, a refined cast iron is produced ; whilst, were somewhat more carbon removed by longer action, a kind of Uchatius steel would result.

Ellershausen's process consisted in running molten pig iron upon a bed of iron oxide, sprinkling ore on the surface, and running on another layer of pig iron, and so on until a mass of alternate layers of pig and oxide is formed ; • on heating this the pig gradually becomes decarbonized, the reaction being in fact a kind of modification of the " malleable cast iron " process described in § 22. Finally the mass is forged and rolled. The results obtained were not uniformly satisfactory, whilst the fuel consumption was found to be large ; accordingly the method never became anything much more than an experimental process.

-Under the names of " Siemens process " and " Siemens-Martin process" are usually included several different modifications. In what may be called the older or original process, which was essentially the combination of pig iron and malleable iron fused together in a regenerative furnace hearth, the same kind of difficulty was met with that nearly proved fatal to Bessemer's original process, viz., that it is difficult if not impracticable to make sure of obtaining any required degree of carbonization of the resulting steel, so that the finer classes of steels cannot readily be thus made ; for rails, however, the process has been largely adopted. Much the same way of overcoming the difficulty was adopted in the SiemensMartin process as was used by Musket, thus giving the second modification, viz., making the atmosphere slightly oxidizing, and continuing the heating until the metal is decarbonized, when the required amount of carbon is added in the form of spiegeleiseu or ferro-manganese, and the steel forthwith cast. This modification is consequently substantially the refining process formerly adopted as a preliminary stage to puddling (§ 23) carried out a great deal further (so as wholly to decarbonize the metal) at a much higher temperature, and differs from the Bessemer blowing process mainly in this that the oxygen requisite to burn off the carbon and oxidize silicon, &c,, is made to play over the surface of the fused mass instead of passing through it. The decarbonization is carried out in precisely the same way (so far as principle is concerned) as that by which the oxidation of lead and base metal is effected in the ordinary process of gold and silver cupellation in an oxidizing atmosphere at a high temperature. These modifications, of which the second is now much the more largely adopted, are generally referred to as the " Siemens-Martin " process, or "Martin process." A third modification is substantially the Uehatius process carried out in a regenerative hearth instead of a crucible ; this is known as the "Siemens process" or "ore process," and consists in melting haematite pig, or other pig iron free from sulphur and phosphorus, and then adding in small quantities at a time an equally pure ore until a sample taken oat from time to time does not harden on plunging into water whilst still red hot ; to the fused iron spiegeleisen, &c., is then added as before ; in this way a somewhat larger quantity of steel is obtained from a given amount of pig, the ore becoming partly reduced whilst oxidizing carbon and silicon, &c. ; but this advantage is counterbalanced by the greater wear and tear owing to the larger amount of cinder formed and its corrosive action on the brickwork, and by the necessity for using somewhat more fuel. A fourth modification consists in a sort of combination of the Martin method and the ore process, the pig and scrap, &c., being fused together and the decarbonization being then effected, not by oxidation by the gases alone, but by that together with ore added to the mass.

When Siemens's precipitation process (§ 31) is used, or when the ore is reduced to spongy metal in a rotating furnace, &c. (§ 30), the resulting iron is readily converted into steel by simply adding it instead of malleable scrap to the fused pig in the above process,spongy metal when thus employed being made up into a sort of ball by stirring it up with fused magnetite, and the whole added to the fused pig ; another variety of combination of " ore process" and ordinary Siemens-Martin process is thus obtained, the finishing up of the metal by adding spiegeleisen, &c., in known quantity when complete decarbonization has been effected being the same in all cases. An analogous modification is that of Blair, who first prepares spongy iron by reducing the ore in much the same fashion as in Chenot's process (§ 30), and then fuses it up in a regenerative furnace with more or less pig to give a fluid bath to begin with.

In the ordinary working of the process when ore is not used, the materials employed are pig iron (free from sulphur and phosphorus) and malleable scrap of various kinds, together with scrap and waste Bessemer steel, crop ends of rails, &c. The pig being melted and the malleable iron raised to nearly a white heat (either in a separate furnace or by the waste gases before passing to the regenerator), the latter is gradually added to the former until the whole is liquid ; the heating is then continued, the flame being made somewhat oxidizing so as to gradually decarbonize the metal, until a sample of the metal drawn and cooled in water is scarcely hardened thereby ; at this stage the metal is virtually molten decarbonized iron, all silicon, manganese, and carbon having been removed by oxidation. To give the requisite steel character, a definite quantity of spiegel-eisen is added (or of ferro-manganese when a larger relative percentage of manganese is desired, or in certain cases of silico-manganeisen when silicon is wanted to be present), and the whole cast into ingots. Pig iron alone can be used in the first instance, only then a longer time is requisite to effect decarbonization. The following analyses by A. -Willis illustrate the relative-rates at which the oxidation of manganese, silicon, and carbon is effected, the quantity of the last not suffering any material diminution until the others have almost disappeared : - When pure ores are used in the ore process, no appreciable alteration takes place in the percentage of sulphur in the pig and scrap, but if sulphates (e.g., barium sulphate) be present in the ore, the resulting steel contains more sulphur than the pig and scrap used to the extent of about 30 per cent. of the sulphur present in the ore (Willis). When ferro-manganese is used to finish the process and prepare a soft steel, the requisite quantity of ferro-manganese is heated up either on the bed between the hearth proper and the regenerators on which the materials are heated up by waste heat or otherwise without fusion ; when spiegeleisen is used, the requisite qnantity may either be added in the same way, or fused in a cupola and then added ; burning out of more or less carbon and manganese is always a possibility where a cupola is used, wherefore when possible the spiegeleisen is fused in the hearth itself or its adjuncts. In order to obtain the best castings at Terre Noire, the decarbonized metal is treated with siliconeisen, and then allowed to remain for some twenty minutes in as nearlyneutral an atmosphere as possible, so that cinder may completely be separated by gravitation and the reaction of the silicon on the carbon oxide may be complete (§ 44) • the ferro-manganese is then added, and the casting proceeded In the Siemens-Martin process (where ore is not used) the yield of steel finally obtained is somewhat below the weight of metal originally employed, owing to oxidation ; in the ore process, on the other hand, a larger weight of steel is ultimately obtained than that of the metal used, owing to the reduction of iron from the ore. The consumption of fuel per ton of ingot steel is, however, somewhat higher in the ore process, owing to the larger quantity of slag. According to Gautier a considerable loss of fuel results if the gas producers are not placed close to the steel furnaces, even to the extent of one-third in some cases ; probably this figure is a little overestimated (see § 10).

In order to avoid the necessity of decarbonizing completely the metal and then recarbonizing by addition of spiegeleisen or ferromanganese, which is practically entailed by the difficulty experienced in finding out the exact composition of the partially decarbonized metal at any given stage of the operation, it has been proposed by Ryder to sample the steel and cast the sample into a small ingot of definite size and shape, and then to determine the magnetic qualities developed in the ingot by the influence of a powerful electromagnet, using a particular apparatus devised for the purpose. In this way a fairly correct estimate of the amount of residual carbon is obtained, rendering it unnecessary to prolong the operation of decarbonization further when the test shows that an amount is present sufficient to give a steel of the required quality on addition of a known amount of ferro-manganese. In the inventor's hands the method has been found to work successfully, enabling the open hearth operations to be considerably shortened as to duration. An improved form of apparatus for the purpose has also been described by Wattenhofen.

gas and blast and exit gases passing to the regenerators ; as the furnace revolves the liquid metal always forms a pool at the lowest portion, but any solid matter is carried round, alternately rising above the pool and being plunged beneath it ; the effect of this is greatly to facilitate melting down and also considerably to shorten the time requisite

Comparing the working of a furnace of this kind with one of similar dimensions but fixed bed, Hackney found that the output of steel was about double in a given time, and the coal used per unit weight of steel was less than one half, viz., 0.40 to 0.43 instead of 0.90 (8 to 8i ewts. per ton instead of 18 cwts.). At St Chamond an improved Pernot 7-ton furnace gave during three months working i the following results per unit of ingot steel :- Coal used for smelting 0-318 lighting, repairing, Sze. 0.156 Total 0.474 The metal used per unit of ingot steel was 1.06, the output being about 21 tons per day of twenty-four hours, the conversion taking about seven hours per charge. In English works where the Siemens process (" ore process ") is used with fixed hearths the yield of steel is somewhat in excess of the metal used originally, hut the time of working is inferior to that just mentioned, the yield with 5-ton furnaces being only some 14 tons per twenty-four hours.

Holley states that the removable Pernot furnaces set up in America (especially at Springfield, where 20-ton hearths have been recently erected) are highly satisfactory, especially as regards the ease with which repairs can be made • the hearth can be run out on Saturday night, and is cool enough to repair on Sunday ; firing up being commenced on Sunday night, the furnace is ready for the usual charge on Monday forenoon. Krupp's dephosphorizing pro cess is adopted to purify the pig before finishing in the Pernot hearth, the metal being run from the melting cupolas into the Krupp washing furnace, and thence into the steel furnace by means of a ladle. Essentially the Pernot furnace is an ingenious combination of various previously well-known principles, the rotating Circular bed having been previously used not only for puddling iron, but also for roasting ores, and the withdrawing carriage having also been employed previously in the manufacture of armour plates, whilst the inclined axis had also been previously used.

The Berard Process. - This method is essentially a sort of combination of the Bessemer converter principle and of the open-hearth method. A. double furnace is employed, heated by gas, and provided with movable tubes dipping into the melted metal, or with a tuyere at the base in the case of the first hearth, in which the metal is blown, and the carbon, &c., oxidized by means of an air-blast ; in the fellow hearth the metal is partially recarbonized by the gases from coal similarly blown into it, the object being to facilitate the removal of sulphur and phosphorus. - Finally the purified metal is treated with spiegeleisen or ferromanganese in the usual way.

Ponsard Furnace or Forno-Convertisseur. - This apparatus is essentially a combination of the Pernot furnace with the Bessemer converter, consisting of a hearth movable about an obliquely vertical axis (figs. 63, 64). Instead of rotating round and round on this axis, the hearth D only moves through half a revolution ; when in one position (as indicated in fig. 63) the surface of the molten metal is above the level of a series of tuyeres 0 fed by a blast pipe LMN under these circumstances the apparatus performs the functions of a Bessemer converter, the blast passing through the molten metal ; when it is half turned round, the tuyeres are raised above the surface of the metal and the blast is shut off, so that it then becomes an ordinary Siemens open hearth. The air blast is introduced, as in the Bessemer converter, through a hollow axis of rotation; the hearth is fixed, as in the Pernot furnace, upon a carriage or bogie K, so that it can be withdrawn and the metal tapped out at the tapping hole P. A gas producer A is attached, the gases from which are burnt as they are formed without cooling by passing through a long pipe, &c., or heating by a regenerator ; the air used to burn them, however, is heated by a regenerative arrangement consisting of a pair of chambers H filled with brick stacked in a peculiar way ; through one chamber the waste flame passes by the flue EFG, heating it up ; through the other one used alternately the air passes reaching the furnace by the flue CC. F is a chamber in the waste gas flue for the deposition of solid suspended matters, dust, &c., from the blowing operation, and G is a small bed for heating scrap, &c., before its introduction into the main hearth D. Owing to the regenerator only heating up the blast, the waste gases escape at a much higher temperature than with a Siemens regenerator; accordingly they may be advantageously utilized to produce steam.

According to Perisse (from whose paper - Journ. I. and S. Inst., 1878, 459 - the cuts, figs. 63 and 64 are taken) the average duration of a blow and subsequent operations in a 5-ton furnace is five to six hours when the metal is charged cold, whilst six to eight operations may be made in twenty-four hours when it is previously melted in a spare hearth ; 250 kilos (about 5 cwts.) of Mons coal were consumed per hour, or about 6 tons per twenty-four hours for 20 tons and upwards of production (charged cold), or for 30 to 40 tons (charged with fluid metal). Phosphorus is not materially eliminated, because, as in the ordinary Bessemer and Siemens-Martin hearths, there is a silicious flux always present, owing to the nature of the lining ; this objection, however, is readily remediable by simply using a "basic" lining (f 37), when more or less complete clephosphorization results, just as is the ease with the Pernot hearth when a large amount of fused iron oxide is added, as in Krupp's dephosphorizing process.

By omitting the movable bed, and substituting the ordinary fixed bed of a reheating furnace, the Ponsard forno-conyertisseur becomes changed into the Ponsard reheating furnace ; the advantages of this form of arrangement over that of Siemens's rcheater are said to be considerable saving of cost in construction and capability of producing steam by the waste heat.

sesquioxide 13'86 Ferric oxide 15'38 „ Silica 4.13 „ Water 8•68 „ Silica 25'88 per cent.

Ferrous oxide 54.57 „ Manganese oxide 17'17 „ Alumina and calcium carbonate, &c. 2.38 „ little or no magnetic action.

During the early period of the development of the fused steel industry, the richer manganeisons (containing 15 per cent. and upward of manganese) used for crucible steels were themselves prepared in crucibles, the term " ferro-mangauese" being applied to these products, " spiegeleisen " indicating the less manganiferous pig containing some 6 or 7 per cent. of manganese prepared in the blast furnace ; subsequently, however, the blast furnace spiegel-eisens of commerce were prepared much richer in manganese than formerly, whilst fer•o-manganese of 30 and upwards per cent. of manganese came into use prepared by Henderson's method, viz., by reducing upon the open hearth of a Siemens furnace a mixture of ma"ganese carbonate (obtained from bleaching powder residues or "still-liquor") and ferric oxide in presence of excess of carbon, a neutral or slight reducing flame being employed. The furnace bottom is made of coke ground up and consolidated, so as virtually to form a large carbon shallow crucible or basin, - the finely divided mixture being put in and the temperature raised to a low red heat for some hours. A metallic sponge is obtained which subsequently runs down to a regulus when the temperature is raised to a full white. An important influence in the amount of manganese the resulting pig contains less manganese ; and the same remark is also largely true of sulphur.

According to several chemists and metallurgists as the spiegeleisen of 7 or 8 per cent. manganese usually contains both find that the carbon percentage increases pari passe is derived are not exhibited by ferromanganese containing large amounts of manganese.

By mixing finely divided iron (sponge, filings, turnings of cast or wrought iron or steel, &c.) with finely powdered ammoniacal solution, and compressing lumps, hard masses are formed (after standing a few hours) which can be heated to a red heat without breaking up. By melting these in a small cupola furnace with a crucible or hearth of alumina, magnesia, or lime, or of hard carbon and a tuyere or iron-titanium alloys, or ternary alloys can be readily produced (Chronique de l'Industrie, 1873, ii. 235).

The following analyses illustrate the composition of various kinds of spiegeleisen, ferromanganese, and allied products: -

According to Snclus (Josera. I. and S. Inst., 1874, 68) the best results are obtained when a certain quantity of manganese remains mired:iced and escapes in the slag, which should have about the following composition to give the best results :- Silica 35 per cent.

Alumina 10 „ Lime . 30 „ Much more richly manganiferous metals than this have been obtained in the blast furnace by W. G. Ward (Cartersville, Georgia) by simply increasing the amount of lime added, and employing a smaller burden ; in this way ferromanganese of upwards of 50 per cent. has been manufactured readily, about three-fifths of the manganese in the charge being reduced. Analogous metals have boen obtained by many other smelters; thus specimens of ferromanganese containing upwards of 80 per cent. of inanganesc and prepared in the blastfurnace were exhibited in the Paris exhibition. According to Akermann, to produce ferromanganese of 85 per cent. in the blast furnace is a matter of no great difficulty, but the consumption of coke is about four times greater than that required for common pig iron, whilst the daily output is only about one-fourth of that of the latter. No advantage attends the production of manganeisen of upwards of about 80 percent., but rather the contrary, richer metal being vary brittle. It is noteworthy that, notwithstanding the use of much more fuel in a furnace smelting rich manganeisen, the escaping gases do not contain more carbon oxide than (and often not as much as) those of an ordinary blast furnace, the extra oxygen in the carbon dioxide escaping coming from the higher oxides of manganese employed as manganese ores. A description of the older modes of producing spiegeleisen in Germany is given in a report by Wiborg to the Swedish iron office (Jern-Kontorets Annaler, 1870 ; also in abstract in Journal I. and S. Inst., 1872, 138).

When lime fluxes are employed, it is indispensable that they should not contain phosphorus ; thus the spiegeleisen prepared in New Jersey from franklinite was found to contain phosphorus to a decidedly prejudicial extent when oyster shells were employed as flux, but ceased to do so when good limestone was used instead. It is noteworthy, however, that when a highly manganiferous iron is smelted a considerably larger preparation of phosphorus finds its way into the cinder and less Into the pig than is the case when filings, and silica melted in a wind furnace affords a cast iron containing 15-16 per cent. of silicon and nearly 3 per cent. carbon ; lime or calcium silicate on the other hand removes silicon from silicious iron when the two are melted together. It is a matter of usual belief that silicon expels sulphur from pig iron ; at any rate the conditions most favourable to production of highly silicious pig are not so favourable to the presence of sulphur, and hence grey silieious pig is much less sulphurized than white pig made from - the same materials. According to Riley the greater the percentage of silicon in silieoneisen the smaller is the amount of carbon present, so that when 15 or 20 per cent. of silicon is present the carbon is reduced to a very small amount.

Silleon-Hanganvisen. - Sil spiegeleisen (or manganese silicide, as it is sometimes termed) is prepared by the ordinary blast furnace methods of making rich spiegeleisens by increasing the amount of silica present and using large amounts of fuel. For the purpose of preventing " blowholes " in cast soft steels containing but little carbon, an alloy containing about 8 per cent. of silicon, 14-15 of manganese, and about 1.3 per cent. of carbon, has been extensively used at Terre Noire. The steel thus produced gave the following numbers (Euverte, Bulletin Soe. des leg. Civ., 1877) : - Chrontium-sted, or " Chromeisen," as it may conveniently be termed, is closely allied to manganeisen ; the capabilities of this substance for the most part yet remain undeveloped, but the effect of the chromium is analogous in some respects to that of manganese, in others to that of carbon, communicating a fine close texture with hardness and brittleness when present in any quantity. Chromeisens containing 10 per cent. and upwards of chromium have been prepared. by crucible operations, and of somewhat less richness in the blast furnace from ores containing a notable amount of chrome ironstone ; Sergius Kern obtained a very bard chromeisen containing 74 per cent. of chromium and 25 of iron by heating chrome ironstone and charcoal powder in graphite crucibles. The Tasmanian Iron Company produced from Tasmanian ores a pig containing some 6 or 7 per cent. of chromium and 4 of carbon ; according to Riley a mixture of this pig with ordinary hadmatite pig (half and half) would not puddle well, being incapable of welding and forming a bloom on account of the thick cinder ; with less chromeisen (one-tenth) the puddling period was prolonged ; a little of the chromium remained unoxidized in the iron without materially affecting its qualities ; at first the ordinary grey pig melted whilst the chromium pig remained unfused. By melting chromium pig containing upwards of 0.2 per cent. of sulphur with ferro-manganese, the sulphur becomes almost entirely eliminated, the resulting product only containing 0.035 per cent.; on adding it (in a fused state) to blown Bessemer metal a product was obtained not at all resembling good. manganiferous Bessemer steel, as it crumbled under the hammer at a red beat. The experience of a Sheffield firm as to the tempering qualities of chromium steel has been very unsatisfactory ; and Boussingault states that he wos unable to give to iron any of the useful properties of steel by adding chromium unless carbon were also present ; on the other hand, it has been stated. that in America mining tools made of chromium steel have been found to be more durable than any others, the tempering being readily managed. Yet again, R. Brown finds that by adding potassium dichromate to blown Bessemer metal or other varieties of steel a certain amount of chromium is reduced and counteracts the evil effects of phosphorus, steel so made being capable of being bent and twisted cold even though containing as much as 1 per cent. of phosphorus ; thus samples of steel so prepared gave the following numerical results (hunt 1. and S. Inst., 1879, 355) : - Tungsten-steels have been often brought forward as possessing valuable properties, but do not appear to have been largely manufactured in England. On the Continent they have attracted more attention ; thus Biermann of Hanover has prepared ferro-tungsten containing from 20 to 50 per cent. of tungsten and a few parts per cent. of manganese with the object of using the composition in the same way as ferro-manganese, i e., intermixing it with larger amounts of decarbonized iron so as to obtain a tungsteniferous steel of known composition. Griiner states that the hardness of steel is increased by addition of tungsten almost indefinitely, but that when more than 2 or 3 per cent. is present the metal becomes brittle from extreme hardness ; for lathe tools steel up to 8 per cent. tungsten may be used provided. they be submitted only to a mild oil tempering, for with water they would fly to pieces ; thus a hook tool used in a steel tire shop of the West of France Railway Company contained Tungsten.... 7.98 per cent.

Carbon 1.40 „ Silicon 0.22 In the manufacture of such steels the tungsten is introduced either as a triple manganese alloy (prepared on a Siemens hearth or in the blast furnace like ferro-manganese) or as crude fritted metallic tungsten obtained by reducing wolfram with a mixture of coal and tar, &c.

Levallois has patented in France the use of ternary nickeltungsten-iron alloys containing They are said to be very hard, but can he worked like ordinary cast steel ; they are prepared by putting the tungsten and nickel into a small soft iron tube together with a little flux (composed of 9 parts boric anhydride and 8 each of calcined quartz and washed calcium carbonate), and placing the tube in the midst of the rest of the iron in a crucible, the surface being covered with some of the flux. When melted up, the whole is poured into moulds in the usual way.

Manganese Bronze. - By melting ferro-manganese and copper together, or preferably mixing the two metals separately fused, all traces of oxide disseminated through the copper are removed by the agency of the readily oxidizable manganese, and a close-grained tough alloy results, capable of very many useful applications. The percentage of iron in the alloys usually prepared, however, being but small, the description of these products more properly belongs to the metallurgy of copper and manganese.