Carnegie Steel Company v. Cambria Iron Company (185 U.S. 403)/Opinion of the Court

Steel is a product, or, perhaps, more accurately, a species of iron, refined of some of its grosser elements, intermediate in the amount of its carbon between wrought and cast-iron, and tempered to a hardness which enables it to take a cutting edge, a toughness sufficient to bear a heavy strain, an elasticity which adapts it for springs and other articles requiring resiliency, as well as a susceptibility to polish, which makes it useful for ornamental and artistic purposes.

Pig iron, which was the original basis for the manufacture of all iron and steel, is made by the reduction of iron ore in large blast furnaces, which are filled with layers of ore, charcoal or coke, and flux. By the agency of this the iron is melted out and falls to the bottom of the furnaces, is drawn out through openings for that purpose into canals, and finally into molds, where it solidifies into what are termed pigs. Prior to the invention of Sir Henry Bessemer, steel was manufactured from a pig-iron base by a tedious and expensive process of refining in furnaces adapted to that purpose. The process was so costly that steel was little used except for cutlery and comparatively small articles. and was practically unknown in the construction of bridges, rails, buildings, and other structures where large quantities of iron were required.

In 1856 Bessemer discovered a process of purifying iron without the use of fuel, by blowing air through a molten mass of pig iron placed in a refractory lined vessel called a converter, whereby the silicon, carbon, and other nonmetallic constituents were consumed, and the iron thus fitted for immediate conversion into steel by recarbonization. The present process of recarbonization was a supplementary invention of Mushet, who accomplished it by the introduction of ferro-manganese, or spiegel-eisen, while the iron in a molten state was issuing from the converter, in which it had been purified, and was thus converted into steel. The process of running molten metal from blast furnaces into pigs and remelting them in cupola furnaces for use in a converter was termed the indirect process, and was generally used prior to the Jones invention.

His process is thus described by Bessemer in his patent of 1869: 'The most important of these operations consist in melting the pig metal, transferring it in the molten state to the converting vessel, blowing air through it, and converting it into a malleable metal, mixing the metal so converted with a certain quantity of fluid manganesian pig iron, pouring the mixed metals into a casting ladle, and running it from thence through a suitable valve into ingots or other molds, and the removal therefrom of the ingots or other cast masses when solidified.' This invention of Bessemer, simple as it appears, may be said not only to have revolutionized the manufacture of steel, and to have introduced it into large constructions where it had never been seen before, but to have created for it uses to which ordinary iron had been but illy adapted.

While in the Bessemer specification of 1856 it is said 'the iron to be used for the purposes of my present invention may be conveyed by a gutter in a fluid state direct from the smelting furnace where it has been obtained from the ore,' without the expense and delay incident to the intermediate cupola process, practical experience, in this country at least, showed that the refining of iron without first casting it into pigs, selecting or mixing the pigs, and remelting them, was attended with such expense that the entire abandonment of the practice was seriously considered. The difficulty was in the material variations between different portions of the same cast, and even different parts of the same pig,-an irregularity which was increased when the metal was drawn from several furnaces. There was added to this frequent changes in the character and composition of the ore, coke, and limestone flux with which the furnace was charged. The consequence was that the nonuniform chemical composition of the metal from the molten blast furnaces yielded products of steel, such as rails and beams, which were not only irregular chemically, but of irregular and uncertain final condition,-some sound, others of imperfect strength and full of flaws.

These irregularities were in a measure obviated, not only by a careful selection of pigs beforehand, but by the necessity of employing upon receiving ladles or reservoirs into which the product of one or more cupola furnaces was drawn off into such reservoirs, which were made large enough to hold the product of two or three furnaces, and from which the molten metal was withdrawn into the converters. Had the amount required for the converters in each case been the exact product of one or more cupolas, no reservoirs would have been necessary, but as the demand was variable, a storage of molten metal was required to retain the product of one or more cupolas, until it was required for the converters. Of course, as the product of two or more furnaces was drawn off into these receiving ladles, there would be some intermixing of those products, although the receiving ladles do not appear to have been used for that purpose, the operators relying more particularly upon the careful selection of pigs beforehand, to obtain the requisite uniformity for conversion into steel. The ladles being open at the top, the molten metal could not long be retained in them, and in the best practice it was so arranged that the withdrawals from the reservoir were made every few minutes, and without regard to the amount left in the reservoir after each withdrawal. It will be borne in mind that the object in either case, whether by direct or indirect process, is to obtain, as far as possible, a uniform product of iron for the converter.

'These results,' said one of the witnesses (Kennedy), speaking of the process used before that of Jones, 'are not obtained by the practice of taking metal from two blast furnaces by running a train of ladles in front of them and tapping into each ladle half a charge and following it from a second furnace. By such practice, of course, there is some independent equalization of the composition of each ladle or of the ladles of each group, but it affords no further advantage, and in fact would not obviate the difficulties of direct metal working. It does not enable the converter manager to foretell the character of each charge from the character of the preceding charge, and would therefore entail the uncertainties of operation and the irregularity of the product which the Jones method avoids.'

It had long been an object of manufacturers that steel should be made directly from the molten metal, as it comes from the blast furnaces, without having to pass through the intermediate or cupola process, which involved the casting of the furnace metal into pigs. These, after becoming cold, were assorted, broken up, recharged and remelted in a cupola furnace, and then placed in a converter for conversion into steel. By this cupola process a product, practically uniform in character and suitable for further treatment in the converters, was secured, but at the expense (more than 60 cents per ton) of rehandling and remelting the iron as it come from the blast furnaces, in cupolas, and the contamination of the metal with sulphur evolved from the coke in the process of remelting. The obstacles connected with this method and the difficulties attendant upon the use of the direct process are thus comprehensively set forth by Mr. Julian Kennedy, one of the experts:

'Ever since the invention of the Bessemer process it has been well recognized that great economics could be attained by transferring the molten metal from the blast furnace to the converter without allowing it to solidify. Until within a few years, however, this direct process, as it has been called, has not been generally used. It is easy to see why this was the case. The fluctuation in the chemical composition of the metal from the blast furnace was too great to allow that degree of uniformity of product in the Bessemer steel produced from it which is absolutely necessary in the case of steel rails, for example, which must be as reliable as human skill can make them, and where no reasonable expense can be spared to make them perfectly safe and trustworthy. A very few broken rails in a track, with the damage to property and human life which this might cause, would far more than offset any possible saving in a year's work, due to the use of the direct process. For this reason the practice, until within comparatively recent years, has been to cast the metal in pigs, then to analyze it and reject any portion not closely approximating a rigid specification in its chemical composition, and to select, mix, and then melt the approved metal in cupola furnaces. By this means very great uniformity of chemical composition of the remelted metal can be obtained, and good and reliable steel made from it with regularity and certainty.' Speaking of a time when the direct process (before that of Jones) had been in use for several years he said:

'After studying the results which had been obtained at the Edgar Thomson works and elsewhere in the use of the direct process, I consulted with Mr. James Gayley, and we agreed that in the building of a new works it would not be profitable to use direct metal, but that, on the contrary, the disadvantages resulting from the irregularity in the product were so great that it would be better to go to the expense of building and using cupola furnaces. We did not then perceive any means adequate to overcome these disadvantages.'

The difficulties connected with the prior devices are also stated in an article by Mr. Holley, published in 1877, from which we extract the following paragraph:

'Third. The embarrassing feature of the direct process is the irregularity in the heat-that is to say, in the silicon of the charges-resulting in the large amount of scrap due to too little of this element, and in the increased number of second-quality rails due to too much of it; while in France, where 3 to 5 per cent of manganese is the heating ingredient, there may always be an excess of this latter element without injuring the quality of the steel, although the variation of heat is here, also, a serious difficulty. In other words, it has not yet been practicable to work the blast furnace with sufficient regularity to realize approximately the theoretical advantages of the direct process.'

'Fourth. The obvious remedy is to mix a number of blast-furnace charges, so as to reduce the irregularity to a minimum. Two systems of doing this are on the eve of trial: The one is simply mixing so few charges in a tank that the metal will be drawn out before fore it chills; the other is to store a larger number of charges in a heated tank,-that is to say, in an immense open-hearth furnace.'

'A few words of history may be of interest. Mr Bessemer's early intention was to use blast-furnace metal direct. The earlier Bessemer practice, especially that in Sweden, was with metal right from the blast furnace. But this practice did not make headway, except where there was from 3 to 5 per cent of manganese in the pig blown, for reasons just mentioned; so that while it soon became standard at Terrenoire and elsewhere in France, as well as in Sweden, and to some extent in Germany, yet in England it was not only unused, but pronounced impracticable so late as September, 1874.'

This difficulty,-and it seems to have been so serious as to render the direct process commercially impracticable,-Jones sought to remedy, and did remedy, by creating a covered reservoir of molten metal between the blast furnaces and the converters, in which should always be maintained a large quantity of metal, happily termed by the district judge a dominant pool, which sould be drawn off in small quantities at a time, and replenished by a like quantity of metal from the blast furnaces. In this way, while the metals taken from the several blast furnaces might differ in their heat and constituent elements, yet, being received and mixed with the molten metal in the dominant pool, they were, when discharged from the reservoir, approximately, though not perfectly, uniform, the original variations having been lost in their mixture with the dominant pool. 'It is therefore plain,' says the district judge in his opinion, 'that with a mixer thus operated, it is possible to have wide variations in the composition of the blast furnace metal charges added, and at the same time the successive withdrawals for the Bessemer converter show quite small and gradual changes of composition. The heat of the detained mass is affected by the incoming charges just from the blast furnace, but the heat of such addition, whether relatively high or low, must mingle with, be modified by, and average with, the heat of the larger and dominating mass.' It is not insisted that this method gave absolutely uniform results, 'nor,' says the witness Fry, 'did the inventor, as I understood him, comprehend such, but, on the contrary, he recognized the practical impossibility of rendering uniform a continuous supply of metal, and desired only to reduce the abrupt changes of the several portions added to the gradual changes of the portions withdrawn, and this is what he worked out from his invention in a thoroughly practical way.'

While the patent in suit is for a process, and not for a mechanism, the process will be the more easily understood by a reference to the apparatus above reproduced, which consists of a reservoir, or closed receptacle, commonly termed a 'mixer,' lined with fire brick of sufficient thickness to retain the heat of the molten iron, and of such size and strength as to be capable of receiving and retaining a large amount-'say, 100 tons'-of molten iron. This reservoir is mounted upon journals, and is adapted to be tipped so as to receive at one end molten metal from the blast furnaces, carried to it in cars, and by being tipped in the other direction, to discharge the same into similar cars, in which it is carried to the converter. The essence of the invention lies in the fact that the tip is so regulated by a stop that the reservoir can never be wholly emptied, but a 'considerable quantity' of metal always remains,-a dominant pool, into which successive additions are received.

That the invention is one of very considerable importance is attested by the fact that it was not only put into immediate use in the Edgar Thomson works at Braddock, then owned by the plaintiff, but has since been adopted by all the leading steel manufacturers in this country, and by many similar works in Europe, where the patent was sold for 10,000. Mr. Carnegie, one of the witnesses, says of it: 'There were both advantages and disadvantages [in the direct process used prior to Jones's invention], but the disadvantages were so great that we often debated whether to abandon the process or not. We found it impossible to get a uniform quality of rails as well as by the cupola method. . . . When we were still anxiously struggling with the problem, and undecided whether to continue or abandon it, Captain Jones. . . told us that he believed he had invented a plan which would solve the problem. . . . We thought so well of the idea-I was so convinced of its reasonableness-that I directed him to go ahead with his invention. . . . Captain Jones did so, and almost from that day our troubles ended. He had scored a tremendous success; another step forward was taken in the manufacture of steel, and we are using the invention to-day. . . . Without this invention I believe that we should have abandoned the mode of running direct from the blast furnace. Above all things, the manufacturer has to regard the uniformity of product, the equality of rails; and this uniformity cannot be obtained without Jones's invention, as far as I know.'

It is true that what is termed the direct process was used in connection with the Bessemer invention in some foreign countries, notably Sweden and France, with more or less success, due to the peculiar character of the ores used in those countries; but such attempts in this country had proven practically failures, and had been abandoned. In regard to this the witness Kennedy said:

'The Jones method has made the direct process, which was attended with great danger and difficulties before the date of his invention, a thoroughly practicable and successful one. Instead of it being a question of great doubt whether to run the metal direct to the converter or remelt it, as it was up to the time of Jones's invention, no one would now think of building a new works containing both furnaces and converters without arranging to mix the metal by the Jones method, which not only effects an immense saving in the cost of operating the works, but enables a uniformly good product to be made, and also a purer product than can be obtained from cupola metal, which absorbs and is contaminated by sulphur from the coke which constitutes the fuel of the cupola.'

Indeed, the value of the process is not wholly denied, though much depreciated, by the defendant, which relies rater upon the fact that it was well known in the art, and that so far as it is described in the Jones specification and drawings it was not infringed by it.

1. We now proceed to an examination of the question of anticipation, in support of which a number of English patents are produced, which will be briefly considered: First, the British patent to Tabberner of 1856, the object of which was, as stated by the patentee in his specification, 'to dispense with the necessity of employing one or more large furnaces, and to use in lieu thereof several small furnaces, the combined capacities whereof are equal to that of one or more large furnaces, and to cause these small furnaces to discharge their contents at short intervals of time into one large reservoir, from which the molten metal may be drawn for casting from. . . . The principal features in this invention consist in directing the blast to the body or belly of the furnace, as well as to the hearth thereof, for the purpose of fusing or smelting the entire mass of ore in the furnace simultaneously, or nearly so. . . . The mode hitherto practised in smelting furnaces has been to direct blasts into the hearth only thereof, thereby requiring several hours to smelt or fuse the contents of a large furnace.' The specification is somewhat blind, and it is difficult to see what definite or valuable result is obtained by the use of several small instead of one large furnace, except, perhaps, a quicker heating and less delay in its practical operations; but it is sufficient for the purposes of this case to say of it that it contains no suggestion of a mixing of different casts for the purpose of obtaining a more uniform product, and that the invention has no relation to a further treatment or refining. It does contemplate the use of a reservoir, but there is no suggestion of a reservation in such reservoir, of a quantity of molten metal. It is not denied that the use of a reservoir from which molten metal may be drawn long antedated the Jones patent. But the best that can be said of the Tabberner patent is that, if the reservoir had been of sufficient size and properly constructed so as to never be completely emptied, it might have been adapted to carry out the Jones process; but there is no evidence that it was ever so constructed, or that the production of a uniform discharge from the reservoir was contemplated. That it could not have been intended for the purpose of carrying out the Bessemer process, or any other process, for the use of blast-furnace metal in a converter, is evident from the fact that the patent was nearly simultaneous with the Bessemer patent, of the existence of which the patentee appears to have been entirely ignorant.

The English patent to Deighton of 1873, for 'improvements in the arrangement and mode of working an apparatus for the manufacture of Bessemer steel,' contains the closest approximation to the principle of the Jones invention. If this does not anticipate, none does. The primary object of the patent seems to have been to prevent the loss of time while the converters are being cooled and relined or repaired, and again prepared for work, by providing that the converting vessel shall be so arranged that it can be readily detached from its actuating mechanism and lifted bodily out of its bearings by a suitable crane or other lifting mechanism, and a spare converter substituted in its place.

There is, however, a further provision in the patent, as follows:

'Instead of manufacturing Bessemer iron or steel from pig iron, which has been melted in cupolas, my invention also consists in taking the molten metal directly from the blast furnace to the converter, in which case I prefer to arrange the Bessemer plant in a line at a right angle to a row of two or more blast furnaces, and place a vessel to receive the molten metal tapped from two or more blast furnaces to get a better average of metal which will be more suitable for making Bessemer steel or metal of uniform quality, the vessel or receiver being placed on a weighing machine so that any required weight may be drawn or tapped from it and charged into the converter.'

The specifications provide for manufacturing Bessemer steel directly from the smelting furnace by employing gates or channels for molten metal from each furnace, leading to a reservoir, which is placed low enough to give fall for the molten metal to flow from the blast furnace to this reservoir, which forms a receptacle for mixing the molten metal from two or more of the smelting furnaces. From the reservoir the mixed molten metal is tapped, and flows down the swivel trough into the converter. By placing the reservoir on a weighing machine, it can be readily ascertained when the exact quantity required ascertained when from it into the converter. The sixth claim of the patent is for 'the system or mode of arranging and working Bessemer converters with a receiver or receptacle for mixing the molten metal from two or more smelting furnaces to get a more uniform quality of metal, substantially as hereinbefore described and illustrated by the drawings.'

While Deighton seems to have conceived the idea that uniformity of product was necessary to the successful use of the direct process, and might be attained by mixing the dischage from several blast furnaces in an open reservoir, standing between the furnaces and the converter, the dominant idea of the Jones invention, that a constant quantity of molten iron should always be kept in such reservoir to serve as a basis for such mixture and an equalizer of the different discharges, does not seem to have occurred to him. As the discharge pipe was located at the bottom of the reservoir, it was certainly possible to empty it entirely, and the testimony in the case indicates that this was the natural method of operation. If this were so, then the reservoir accomplished nothing beyond the mixing of each batch of metal introduced into it from the different blast furnaces. There is nowhere in the specification a suggestion of supplying to and withdrawing from the reservoir small amounts at a time, a constant quantity of metal being retained in the reservoir for the purpose of equalizing the different products of the blast furnaces. While the Deighton reservoir, if a cover had been added to it, might perhaps have been utilized for that purpose, there is no evidence that such use ever occurred to the inventor. Indeed, the absence of a cover to the reservoir is evidence, even to a nonexpert, that it was not contemplated that a permanent quantity of molten iron should be retained in it, since a radiation of heat would thereby be produced and the contents skulled or crusted over with a layer of refuse iron or slag. The testimony is clear that the Jones process cannot be carried on in an open reservoir, and the absence of a cover is conclusive that it is not so used.

It is insisted, however, that defendants have demonstrated, by practical experimentation with a plant constructed according to the specification of the Deighton patent, that the results are practically the same as those obtained by the Jones process. This plant, however, was constructed after suit brought, long after the Deighton patent had been allowed to expire, and with no opportunity afforded the plaintiffs to inspect the plant or witness its operation. The tank was fitted with a cover, and a constant pool of molten metal retained in it; but this was not the Deighton process, but the Jones process adapted to the Deighton device. Were this evidence admissible at all, we are satisfied that it is met by the fact that if the Deighton patent had been adaptable to the Jones process, it is scarcely possible that its merits should have failed to seize upon the attention of manufacturers, who would have brought the patent into general use, instead of allowing it to lapse for the nonpayment of a comparatively small fee. As something in the nature of the Jones process was needed to enable steel to be manufactured directly from the product of blast furnaces, the utility of the Deighton patent for that purpose would at once have been recognized and its success assured. But evidently that patent was not the final step in the accomplishment of the mixing process. It contributed nothing to the art of manufacturing steel, and, although issued in 1873, was allowed to lapse in 1876, after an apparently unprofitable existence for three years, by reason of the nonpayment of the stamp duty necessary to keep it alive. It is sufficient to say of it that it fails to disclose, fully and precisely, the essential features of the process covered by the Jones patent. Walker, Patents, § 54; Seymour v. Osborne, 11 Wall. 516, 555, 20 L. ed. 33, 42; Illinois C. R. Co. v. Turrill, 94 U.S. 704, 24 L. ed. 241.

Although Deighton was an employee of the Moss Bay Company of Workington, England, if any attempt were made by this company to make use of his process, it evidently amounted to nothing, since one of the writers, Snelus, contributing to the Journal of the Iron and Steel Institute, 1876, says: 'One great drawback to the direct casting process was that you could not always get your metal at the exact time you wanted it. He believed that it would be found that the great advantage the Bessemer works in Americal had was the intermediate receiving ladle, which was designed by Mr. Holley, and which was universally used there, although it was never used in England. The Moss Bay Company attempted to modify the thing some time ago, and put up a heating furnace; but that, to his mind, was a step in the wrong direction. Anyhow, the thing had failed, and no one in England, so far as he knew, was using any intermediate receiver between the blast furnace and the converters.'

This defense presents the common instance of a patent which attracted no attention, and was commercially a failure, being set up as an anticipation of a subsequent patent which has proved a success, because there appears to be in the mechanism described a possibility of its having been, with some alterations, adaptable to the process thereafter discovered. As hereinafter observed, a process patent can only be anticipated by a similar process. It is not sufficient to show a piece of mechanism by which the process might have been performed.

In the American patents to Durfee, Nos. 118,597 and 122,312, both of 1871, the desirableness of manufacturing steel directly from the blast furnace is recognized, and in his second patent he says: 'That in the manufacture of steel by the pneumatic or Bessemer process a great saving of fuel and iron, of wear and tear of furnaces, and of labor, would be effected were it possible to make uniformly good products of the desired temper by converting the crude iron immediately it is tapped from the blast furnace in which it is made. This plan has been and may still be practised to a considerable extent, but it has been found that, by reason of the irregular working of blast furnaces and the consequent varying character and quality of the crude iron produced, it was always very difficult and in most cases impossible to secure such uniformity in the converted metal as was essential to success in the business. Hence, at several establishments where the plan of taking the fluid iron as it was tapped from the blast furnace, and pouring it at once into the converter, had been practised, it has been abandoned, the proprietors preferring to incur the expense of handling and remelting the crude iron after it had been cast into pigs, in order thus to secure the advantage of carefully selecting and mixing the materials for each charge to be converted.'

He proposed to accomplish this by using a reverbatory gas furnace, into which the crude iron from the blast furnace is poured, and in which it may be mixed with other irons, and so treated as to insure uniformity. Pig iron of different qualities, or any metals or metalloids or fluxes, can be added and mixed with the metal as may be necessary to bring it to the required character. The process is so manifestly different from that described by Jones that it demands no further attention. if it were put in practice at all, it seems to have proved a failure, as, although an English patent was taken out by Durfee, it was allowed to lapse by reason of the nonpayment of the stamp duty.

Two American patents to James P. Witherow, No. 315,587 and No. 327,425, both issued in 1885, are pressed upon our attention. In the second patent, the only one necessary to notice, he restates the advantages of the direct process and the difficulties theretofore encountered in its practical operation. 'In the manufacture of steel by the pneumatic process, the converters are charged with molten metal, the product of the blast furnace. This metal is usually cast in the form of pigs, then remelted in the cupola as needed, before being charged into the converter. . . . It is very desirable to take advantage of the molten condition of the metal as it comes from the blast furnace for its use in the converter, because thereby the remelting of the metal and the expense of the construction of a cupola may be avoided. The charge of the converter is from one to five tons, and the casting of a blast furnace runs usually from ten to fifty tons. The difficulty of using the molten metal from the furnace to the converter consists in keeping the large quantity of metal from the latter in a proper molten condition for use in the former.' He proposed to remedy this by a reservoir provided with a suitable cover and with tuyeres 'which blow down upon the surface of the metal for the purpose of maintaining its heat and fluidity.' As this reservoir was apparently adapted to hold a single cast, and therefore must be emptied before another cast was received into it, it was impossible that Witherow intended by its use to practise the Jones process. There is no suggestion anywhere in the patent of a desire to retain a quantity of metal in the reservoir to serve as a basis for mixing the various products of the blast furnace, which was the dominant idea of the Jones patent. To anticipate a process patent, it is necessary not only to show that the prior patent might have been used to carry out the process, but that such use was contemplated, or that the leading idea of the Jones patent of maintaining a dominant pool in the reservoir was such a use of the Witherow patents as would have occurred to an ordinary mechanic in operating his device. Whether the reservoir in the Witherow patent was partly or fully emptied seems to have been a matter of complete indifference to the inventor, and the idea of maintaining a constant quantity therein seems to have never been conceived by him. His design seems to have been merely to provide a reservoir for the storage of the large quantity of metal from the blast furnaces, and to maintain its heat until the comparatively small quantities required in the converters had been drawn off for use. As he states in his specification: 'The metal is usually tapped from a blast furnace once in every six hours, and the quantity thus cast is many times in excess of the charge of a converter,' which 'is from one to five tons,' while 'the cast of a blast furnace runs usually from ten to fifty tons.' While the metal is tapped from the blast furnace once in every six hours, 'the time between charges in the converter is usually twenty minutes and upwards, and the metal in the furnace must be kept in condition to be tapped from time to time into the converter as needed.' This appears to have been the whole object of the invention.

The same remark may be made of all these prior devices. While all contemplate the reservoir between the blast furnaces and the converters, such reservoir is used for storage and for such incidental steps toward uniformity as the necessary mixing of the different products of the blast furnace would lead to, while in none of them is there a provision for supplying and withdrawing from the mixer such quantities of metal at a time and the retention of a considerable quantity of metal in the reservoir as a necessary prerequisite to that uniformity of product which was recognized as the great desideratum, and was the constant effort of manufacturers to secure. Granting that some of these devices may have been made use of to carry out the Jones process, none of them in practical operation seems to have been effective to secure the desired result. A process patent, such as that of Jones, is not anticipated by mechanism which might with slight alterations have been adapted to carry out that process, unless, at least, such use of it would have occurred to one whose duty it was to make practical use of the mechanism described. In other words, a process patent can only be anticipated by a similar process. A mechanical patent is anticipated by a prior device of like construction and capable of performing the same function; but it is otherwise with a process patent. The mere possession of an instrument or piece of mechanism contains no suggestion whatever of all the possible processes to which it may be adapted. New Process Fermentation Co. v. Maus, 122 U.S. 413, 428, 30 L. ed. 1193, 1198, 7 Sup. Ct. Rep. 1304. If the mere fact that a prior device might be made effective for the carrying on of a particular process were sufficient to anticipate such process, the absurd result would follow that, if the process consisted merely of manipulation, it would be anticipated by the mere possession of a pair of hands.

True, if the process were the mere function of a machine, another machine capable of performing the same function might be an anticipation; but this is not because a process can be anticipated by a mechanism, but because, as we have held in several cases, the mere function of a machine is not patentable as a process at all. Corning v. Burden, 15 How. 252, 14 L. ed. 683; Risdon Iron & Locomotive Works v. Medart, 158 U.S. 68, 39 L. ed. 899, 15 Sup. Ct. Rep. 745.

To enable the Jones process to be successfully carried out it is necessary (1) that the intermediate reservoir or mixer should be of large size, 'say, 100 tons' capacity; (2) that it be covered to prevent the access of cold air from without; (3) that it be provided with a stop, so that it may not be tilted so far as to be emptied of its contends; (4) that a quantity of molten metal so large as to absorb all the variations of the product of the blast furnace received into it and thus to unify the metals discharged into the converters, be constantly retained in it. None of the prior patents or processes to which we are referred meets these requirements. Indeed, it is scarcely too much to say that none meets more than one of them. When we add to this that none of them was ever used, or was ever susceptible of being used, without material alteration, to carry out the Jones process, it is evident that the defense of anticipation by prior patents rests upon a slender foundation.

Certain discussions, reported in the Journal of the British Iron and Steel Institute, are relied upon as embodying a description of the Jones process. Running through all these discussions there is the same idea of the difficulties experienced in the practical carrying out of the direct process by reason of the want of uniformity in the different products of the blast furnaces, and the possibility of remedying this and thereby doing away with the expense of remelting the pig iron in cupolas by a mixture of such products in a reservoir intermediate the furnaces and the converters; but the dominant idea of the Jones patent, of maintaining a permanent and large quantity of molten metal in the mixer for that purpose, does not seem to have occurred to any of the writers upon the subject. Through all these papers there is an admission of practical failure in the efforts theretofore made to obviate the difficulty, and a half-expressed hope that American ingenuity might ultimately solve the problem. Some of the expressions, taken by themselves, seem to foreshadow the Jones idea; but there was nothing in any of these discussions that filled the requirement of the law (Rev. Stat. § 4886) of a description in a publication sufficient to anticipate the patent.

In some of the very works where attempts had been made to adopt a direct process they were abandoned as unprofitable, and the Jones invention subsequently adopted. The witness David Evans, manager of certain iron works in England and Wales, sums up his testimony in the following answer: 'Prior to the invention of Captain Jones several firms used the direct process, but the results were not very satisfactory, as explained before, through want of uniformity. The results obtained gave a large number of defectives. But since the adoption of the mixer at the various works I have been engaged, we have reduced the defective or second-class rails fully one half, and also saved the remelting.' Indeed, it is stated by several of these writers that the adoption of the Jones invention reduced the defective rails to something like half of what they were before.

Out attention is also challenged to certain unpatented practices, among which is one known as the Whitney foundry practice for the casting of car wheels, wherein the metal is tapped from three cupolas into an open reservoir of eight to ten tons' capacity, permitted to mix and even up in it, and the charges withdrawn to be cast into car wheels, the reservoir being maintained half full. The practice was to run the metal from the cupola furnaces into the reservoir ladle until it was nearly full, then to begin pouring out charges into the casting ladles, while still continuing to pour metal into the ladle from the furnaces, the ladle being kept approximately full during the working day, when it was emptied and refilled on the following day. Aside from the fact that this process has only to do with cupola metal, uniformity in which was largely secured by a careful selection of the pig iron charged into the cupola furnaces, and had no reference whatever to the direct process of charging converters with the product of blast furnaces, it appears that, while Whitney recognized the fact that the charges of iron from the cupolas, when run together into the ladle, would mix, it appears that with this running together of the different charges the mixing operation ended. The maintenance of a permanent pool, and the constant pouring in and out in ladlefuls,-the essence of the Jones invention,-had nothing to do with the process. Indeed, it may be doubted whether the mixing of the cupola metal was of any substantial value. Evidently it suggested to no one the Jones process. It is now too late to insist that it would have been suggested to any mechanic of ordinary skill and intelligence. But if the Whitney practice were primarily for the purpose of mixing, and were adequate for that purpose when applied to cupola metal carefully selected beforehand, it might be, and evidently would have been, wholly inefficient when used for the purpose of unifying the products of blast furnaces,-in other words, for the Jones process; and it might and did require invention to make such changes as were necessary to adapt it to such purpose. Doubtless there was such mixing as the carefully selected cupola metal required for the purpose of manufacturing car wheels, but the fact that the Whitney practice was used for cupola metal has but little tendency to prove that it was adaptable without change to metal tapped from blast furnaces, which varied so largely in chemical composition.

The following observations of the District Judge are illustrative of the distinction between the Whitney foundry practice and the Jones process:

'We must avoid being misled by mere terms and subjects of work. While Jones and Whitney both desired the melting of metals, yet they had widely different objects in view. Whitney's purpose was to cast molten metal into a finished product; Jones's merely to prepare molten metal for further treatment, to wit, decarburizing it into steel. The sine qua non of purpose in Whitney was product uniformity. Uniformity of quality in car wheels is required, so they will stand strain and uniform wear.'

'In the Bessemer direct process you cannot secure, initially or by treatment, uniformity of molten metal. So far as yet developed, the best you can do is to make the nonuniformity gradual and not abrupt. In Whitney, nonuniformity, whether gradual or abrupt, would be alike fatal. In Whitney, relatively absolute uniformity is an essential of product and a sequence of material used. In Jones, uniformity is a nonessential-in fact, a nonattainable-attribute of product, and is a necessary nonsequence of material used. In Whitney, we remelt in a cupola metal which has already undergone the refining process of the blast furnaces. In Jones, we take metal direct from the furnace and discard the cupola. It will thus be seen that apart from the wide difference between the primary work of a huge blast furnace, the base of all metallurgy, and the cupola of the founder, a mere subdivision of that art, we find in the Jones and Whitney processes a substantial difference of purpose, of process, and of subject-matter of work.'

It should be borne in mind throughout the whole of this discussion that Jones never claimed to have succeeded in making a perfectly uniform product; that his object was to procure a uniformity which was adequate for the complete carrying on of the Bessemer process, or, as his second claim states, 'for further treatment,' and really to obviate the necessity of remelting the pigs, which had heretofore been regarded as preliminary to the further treatment by the Bessemer process.

Substantially the same remarks may be made with regard to the Kirk publication, which had to do only with the mixing of cupola metal. This publication was first held by the Patent Office to be an anticipation of the Jones process, the application for which was rejected upon that ground. Upon further consideration, however, and with some slight amendments, the application of Jones appears to have been reconsidered, and was finally granted.

An attempt was made to show that the Jones invention was anticipated by a practice, common in steel works prior thereto, of tapping iron from cupola furnaces into a receiving ladle, which became known as the Bassemer cupola ladle, from which it was poured into the converters. Molton iron was tapped from several cupolas into this ladle, from which a charge was drawn and delivered to the converter vessel. Of course, if the ladle were of greater capacity than was necessary to charge a single converter, a residuum of metal would be left in it; but this seems to have been merely an incident of the operation of the ladle, which was used primarily for storage, and to have been of no substantial benefit in securing uniformity of product, which can only be obtained by making the receiver of larger size, and retaining a considerable quantity of metal in it after each discharge. The witness Kennedy says of this process:

'The irons were carefully selected from the different piles to make up the cupola charges. . . . I have often seen the ladle drained in pouring into the converter. . . . It did not hold two full charges. . . . I never knew of the ladle being used for mixing purposes. If such was the practice I would have known it. . . . The capacity of the ladle was so small, and the size of the pool of metal, when there was a pool, was of such varying size, that I do not see how any mixing could be accomplished. . . . Q. 18. When was this ladle drained, and when would there be some metal left in the handle?-A. There would be no regularity in the process. The rate at which the converters take the metal does not always correspond with the rate at which the cupolas are melting.'

It is true the Jones patent is a simple one, and in the light of present experience it seems strange that none of the expert steel makers, who approached so near the consummation of their desires, should have failed to take the final step which was needed to convert their experiments into an assured success. This, however, is but the common history of important inventions, the simplicity of which seems to the ordinary observer to preclude the possibility of their involving an exercise of the inventive faculty. The very fact that the attempt which had been made to secure a uniformity of product seems to have been abandoned after the Jones invention came into popular notice is strong evidence tending to show that this patent contains something which was of great value to the manufacturers of steel, and which entitled Jones to the reward due to a successful inventor.

2. The phraseology of the patent and the amendments introduced in the Patent Office are made the subject of much criticism, apparently for the purpose of showing either that Jones did not understand what he had invented, or that the specification did not contain 'such full, clear, concise, and exact terms as to enable any person skilled in the art. . . to make, construct, compound, and use the same.' Rev. Stat. § 4888. If these criticisms are not altogether clear, they are pressed upon our consideration with an earnestness which challenges a careful consideration of the history of this patent in the Patent Office.

In his first application the patentee stated that 'the primary object of the invention is to provide means for insuring uniformity in the product of a Bessemer steel works or a similar plant, in which the metal from more than one blast furnace is employed to charge the converters. The product of the different furnaces, or of the same furnace at different times, varies in quality,. . . so that. . . the manufactured steel lacks uniformity in grade. To avoid this I employ suitably constructed reservoirs or vessels, into which the molten metal from the blast furnaces is put, the vessels being of proper capacity to hold a considerable charge of metal from a single furnace, or from a number of furnaces, and being adapted to retain the metal in a molten state for sufficient time to enable the different charges to mix and become homogeneous. . . . Such apparatus possesses also an additional advantage in that it makes it possible to dispense with cupola furnaces for remelting the pigs preparatory to charging the converters. The metal may be tapped from the blast furnace into ladles or trucks, carried to and discharged into the mixing reservoir or vessel, and there retained in a molten state until sufficient metal has been accumulated to charge the converters.'

It is true that he subsequently states, as observed in the opinion of the court of appeals, that 'the main feature of my present invention is the method of storing successive charges of molten metal in a receptacle before using it in converters or otherwise;' and hence it is insisted that the main feature of the invention was storage and not mixing; but the subsequent words of the same sentence, 'drawing portions of the metal from the receptacle without at any time removing the whole thereof, and from time to time replenishing the receptacle with fresh charges, which mingle with the residual molten metal already therein, for the purpose of rendering the successive tappings of metal uniform in quality,' convey a wholly different impression, and show that the primary object was that of mixing different charges for the purpose of securing uniformity in the metal when discharged into the converters. This appears still plainer in the claim appended to this specification: 'The process hereinbefore described, which consists in storing charges of molten metal in a covered receptacle provided with a heat-retaining lining, removing portions only of the molten contents of the said receptacle, without entirely draining or emptying the same, and successively replenishing the receptacle with fresh additions of molten metal, whereby the character of the several charges of metal so treated is equalized; substantially as described.' The word 'storing' was evidently used in the sense of pouring the metal into the reservoir or mixer, as essential to the maintenance of a dominant pool therein. The application was evidently considered as not sufficiently differentiating this from former patents, and was rejected upon reference to the Witherow patents and to Kirk's Founding of Metals. Certain slight amendments were then made in the specification, the claim verbally changed, and an argument submitted to the effect that the purpose of the Witherow patent was 'to receive and store the molten metal for the purpose of preventing the detention, incident to the necessity of discharging the contents of the blast furnace when there is no converter ready to receive it;' whereas the distinctive idea of the Jones patent was 'to have a receptacle capable of holding metal in a molten condition into which metal, it may be from several blast furnaces, is run from time to time, and from which metal is drawn for treatment in the converters, or otherwise as required.' This was evidently considered as still too indefinite, and the application was thought to be fully met by the description in Kirk's Founding of Metals, and was rejected.

Thereupon the application was again amended, its present phraseology adopted, and the distinguishing feature of the invention more clearly set forth. Without further suggestion the application was allowed, and the patent issued.

It is true the process is described in the second claim as a 'method of mixing molten metal,' from which we are asked to infer that it was intended to include the products of cupola as well as of blast furnaces, whereas in the very first sentence of the specification it is stated that 'in practice it is found that metal tapped from different blast furnaces is apt to vary considerably in chemical composition. . . . Especially is this so in the process of refining crude iron from the smelting furnace and charged directly into the converter without remelting in a cupola, and, although such direct process possesses many economic advantages, it has on this account been little practised.' The first claim of the patent is expressly for an improvement in the art of refining iron directly from the smelting furnace. The second claim apparently extends to the art of mixing all molten metals, but the specification, taken in connection with the disclaimer, which describes a process designed to dispense with the use of cupolas, shows that it was intended to include metal tapped from blast furnaces, and was probably intended to be limited to that. Whether the claim would be void if construed to include cupola metal it is unnecessary to consider. It clearly includes metal from blast furnaces, and is not rendered void by the possibility of its including cupola metal. The claim of a patent must always be explained by and read in connection with the specification, and as this claim clearly includes metal taken from blast furnaces, the question whether it includes every molten metal is as much eliminated from our consideration in this case as if it were sought to show that the word 'metal' might include other metals than iron. Were infringement charged in the use of an apparatus for mixing cupola metal, the question would be squarely presented whether the claim had been illegally expanded beyond the specification.

Much ingenuity and many words have been expended in an endeavor to prove that the plaintiff and defendant, as well as the courts, differed widely in their construction of the patent and of what Jones was trying to accomplish. Upon the theory of the defendant the circuit court 'did not attempt to construe the patent in any proper sense, but bent all its energies to wrest and torture the plain English of the patent into a meaning diametrically opposed to that which it bears on its face,' and to make it appear that the great trouble at the time Mr. Jones conceived his invention arose, not from any lack of uniformity in the percentages of silicon and sulphur, but were solely the natural difficulties incident to abrupt variations in the percentage of silicon present; and that his statement that the trouble in the Bessemerizing operation, which was the thing Jones had in mind to obviate, was absolutely irreconciable with the specification of the patent, because the sole object stated by Jones was to secure products, whether of Bessemer steel or otherwise, which would be practically homogeneous and substantially uniform in their contained sulphur and silicon, results which can only be obtained by mixing the iron to a substantial uniformity. Defendant further states its view of the case as follows:

'If, as a matter of fact, Mr. Jones, at the time he applied for his patent, had in view not only the process described by him for securing uniformity in the admixture of silicon and sulphur, but also another process, similar to that now used by the plaintiff and defendant, and by means of which the operation of Bessemerizing iron was made,. . . without securing uniformity in the product, then. . . it is manifestly clear from his patent and from all the surrounding facts that he deliberately and carefully suppressed any disclosure of this invention in his specification.'

One of the arguments that this was the case was that if Jones 'believed the method of use by which abrupt variations in silicon could be avoided without securing uniformity in product to be a patentable invention, there was even more reason for carefully suppressing all suggestion of such a mode of use in the patent for the manipulation to obtain uniform products, because the two processes are obviously alternative and inconsistent with each other, incapable of being claimed in one application, and therefore disclosure might have worked a loss of a possible grant of the patent for the alternative mode of use.'

It is true that its construction of the patent was pressed upon the courts by the defendant with great earnestness and elaborateness of detail, and appears to have created an impression of its soundness upon the circuit court of appeals, but the circuit court did not seem to look upon it as the turning point of the case, nor do we regard it as at all decisive. It seems to assume that the second claim can only be met by evidence of absolute uniformity of product, whereas all that is claimed is a uniformity in the constituent parts of molten metal preparatory to further treatment; in other words, to make it fit for further treatment in the converters, without the necessity of remelting in the cupolar furnaces. Or, as stated by the district judge: 'It is therefore plain that with a mixer thus operated it is possible to have wide variations in the composition of the blastfurnace metal charges added, and at the same time the successive withdrawals for the Bessemer converter show quite small and gradual changes of composition. The heat of the detained mass is affected by the incoming charges just from the blast furnace, but the heat of such addition, whether relatively high or low, must mingle with, be modified by, and average with the heat of the larger and dominating mass.'

With regard to this portion of the opinion, counsel for defendant observes:

'The judge of circuit court, having lost sight of the statutory requirements as to a full, clear, and concise statement of the invention, and having persuaded himself that it was his judicial duty to find a way if possible to protect the Carnegie Company in his monopoly of what Mr. Gayley and his colleagues claim ought to have been the invention described in the patent, adopted the ingenious view that the patent was to be construed as though it disclosed and covered two inventions, one having for its object to obtain a product substantially uniform in its contained silicon and sulphur, and the other having for its object the improvement in the operation of Bessemerizing iron which is incident to an avoidance in the successive charges of abrupt variations in contained silicon.' We have not, however, been able to persuade ourselves that the two processes are so alternative and inconsistent with each other as to render them mutually destructive, or to justify counsel in charging the district judge with an abdication of his judicial duty of deciding the case according to what he believed to be the law and the facts. We dismiss the subject with the simple observation that much more seems to have been made of it than it deserves, and that a reference to the second claim shows its object was to secure uniformity of the molten metal in its constituent parts preparatory to its further treatment, by which further treatment we are to understand the Bessemerizing process of converting metal into steel, and that any step in that direction would necessarily lead to an avoidance of abrupt variations in silicon and sulphur, while such avoidance of abrupt variations would in their turn only tend toward a greater uniformity of product.

Some criticism was made upon the action of the court in permitting a disclaimer of certain clauses in the specification, printed above in italics, which was made after the argument and upon the petition of the plaintiff, 'that at the hearing of this cause it was taken by surprise by the argument of the defendant that the portions of the specification now disclaimed enlarged the scope of the invention of the said letters patent beyond what your petitioner believes to be the import of the claims thereof.' Upon the hearing defendant seems to have insisted that certain portions of the specifications were broader than the second claim. Those parts of the specification therefore were disclaimed. As we had occasion to observe in Sessions v. Romadka, 145 U.S. 29, 36 L. ed. 609, 12 Sup. Ct. Rep. 799, 'the power to disclaim is a beneficial one, and ought not to be denied except where it is resorted to for a fraudulent and deceptive purpose.' In that case the plaintiff was permitted to enter a disclaimer of all the claims but the one in suit, the patentee having included in the patent more devices than properly could be the subject of a single patent. In the case under consideration the disclaimer was not of a claim but of certain statements in the specification, which if retained might be construed to have the effect of illegally broadening the second claim. The first statement disclaimed was that the invention might be practised by merely receiving a number of small portions of metal taken from different ladles, the mixing being performed merely by the act of pouring into the charging ladle. The use of the word 'merely' ignored the steps embodied in the second claim, where the mixing is not performed by merely pouring together the several charges into a ladle, but by maintaining a permanent quantity of metal in the reservoir, to which charges were alternately added and from which they were withdrawn. The other clauses were intended to disclaim the casting of the metal into pigs. We think there is no force in the criticism that a disclaimer may not extend to a part of the specification, as well as to a distinct claim. Hurlbut v. Schillinger, 130 U.S. 456, 32 L. ed. 1011, 9 Sup. Ct. Rep. 584; Schillinger v. Gunther, 17 Blatchf. 66, Fed. Cas. No. 12,458; Schwartzwalder v. New York Filter Co. 13 C. C. A. 380, 26 U.S. App. 547, 66 Fed. 152. Had the purpose of the disclaimer been to reform or alter the description of the invention, or convert the claim from one thing into something else, it might have been objectionable, as patents can only be amended for mistakes of this kind by a reissue. But the disclaimer in this case appears to have been made to obviate an ambiguity in the specification, and with no idea of obtaining the benefit of a reissue. If the clauses had the effect of broadening the patent the disclaimer removes the objection. If they did not, the disclaimer could do no harm, and cannot be made the subject of criticism.

It is insisted, too, that there is no mention in the second claim of a dominant pool, and that the words 'removing portions only of the composite molten contents of the receptacle without entirely draining or emptying the same, and successively replenishing the receptacle with fresh ununiform additions,' are satisfied by leaving a quantity of iron, however small, in the reservoir, and that it really includes nothing that was not well known before. It is true that neither the size of the reservoir nor the amount of metal to be left therein, after each discharge is made into the converter, is specified; but it is stated in the specification that this reservoir may be of any convenient size, 'holding, say, 100 tons of metal (more or less),' with the bottom of the discharge spout some distance above the bottom of the vessel, 'say, 2 feet in a hundred-ton tank, and more or less, according to the capacity of the vessel, the purpose of which is that when the metal is poured out of the spout a considerable quantity may always be left remaining and unpoured, and that whenever the vessel is replenished there may already be contained in it a body of molten metal with which the fresh addition may mix.' Though the size of the reservoir and the considerable quantity left therein as a dominant pool might have been described more definitely (but perhaps at the risk of an infringement being avoided by one using a receiver of a different size containing a different quantity), we think it is impossible to read this patent without gathering from it the dominant idea of Jones not to describe a reservoir for storage, with or without incidental mixing, but to provide a receptacle the main, if not the sole, object of which is to preserve therein a large and constant quantity of molten iron as a basis for a gradual unification of the product of several blast furnaces, or of several casts from the same furnace, and herein distinguishing it from all prior inventions. The specification of the patent is not addressed to lawyers, or even to the public generally, but to the manufacturers of steel; and any description which is sufficient to apprise them in the language of the art of the definite feature of the invention, and to serve as a warning to others of what the patent claims as a monopoly, is sufficiently definite to sustain the patent. He may assume that what was already known in the art of manufacturing steel was known to them, and, as observed by Mr. Justice Bradley in Webster Loom Co. v. Higgins, 105 U.S. 580, 586, 26 L. ed. 1177, 1179, 'he may begin at the point where his invention begins, and describe what he has made that is new, and what it replaces of the old. That which is common and well known is as if it were written out in the patent and delineated in the drawings.' We think this second claim not only describes with sufficient clearness the purpose of the patent to secure uniformity of the molten metal in its constituent parts preparatory to further treatment, but, read with the specification, sufficiently describes the process by which this uniformity may be secured by always preserving in the reservoir a sufficient quantity of molten metal to secure such uniformity of product. It is undoubtedly true that the storage feature appeared more prominently in the specification which was first rejected upon the ground that it was not sufficiently differentiated from prior patents than in that which was finally accepted, but there is nothing to indicate that Jones did not understand from the first that the distinguishing feature of his invention was the preservation of a considerable quantity of iron in the reservoir.

3. The question of infringement only remains to be considered, and, in the view we have taken of the prior devices, presents no serious difficulty. The court of appeals was of opinion that 'the defendant's reservoir, or accumulating ladle, complained of, is the same in principle as one which has been in use at the Cambria works ever since Bessemer steel was first manufactured there, with only this difference, that at first it was used at cupola, now at furnace.' If such were the fact, of course defendant would not be open to the charge of infringement. Undoubtedly it has the right to make use of all prior devices, and particularly such as had been used at its own manufactory. In order to understand the device made use of by the defendant prior to the Jones invention, we reproduce herewith with two small but easily understood cuts, taken from its brief, showing the character of the ladle known as the Bessemer intermediate ladle, used by it and generally by all American mills manufacturing steel by the Bessemer process.

[NOTE: MATERIAL SET AT THIS POINT IS NOT DISPLAYABLE (GRAPHIC OR TABULAR MATERIAL)]

It appears elsewhere in the testimony that the intermediate reservoir or ladle was from fifteen to eighteen tons' capacity, and the converter from six to eight tons; that the molten metal was tapped from the cupolas into the reservoir, and withdrawn for the converter, and as the intermediate ladle held considerably more than the amount of metal necessary to charge a converter, there was some incidental mixing; but the main and perhaps the only purpose of the reservoir was for storage, and that if any quantity of metal were left in the reservoir it was by accident rather than by design. It will be noticed, too, that the reservoir was open at the top. It does not appear to have been made use of in carrying out what is known as the direct process, the difference being that the cupola practice furnished a metal for the Bessemer converter that was uniform in composition, or practically so, while the direct metal was largely variable in composition.

The testimony further shows that, after the installaton of the Jones mixer at the Edgar Thomson works, Mr. Morgan, the defendant's mechanical engineer, visited and inspected these works, and obtained information as to their practical operation, and was advised by the superintendent as to the location and proper size of the mixer and its contiguity to the converters. Mr. Morgan does not deny this conversation, although he qualifies it by saying that he thought the Jones apparatus had grave defects. Shortly after this visit, and in the latter part of 1895, defendant installed an apparatus of its own for the operation of the direct process, which is herewith produced upon a small scale and in comparison with the Jones process. It consisted of a covered refractory lined and turtle-shaped vessel of about

[NOTE: MATERIAL SET AT THIS POINT IS NOT DISPLAYABLE (GRAPHIC OR TABULAR MATERIAL)]

300 tons' capacity, arranged to tilt, and having a spout at either side for receiving and pouring out the metal. The metal was brought to the mixer and poured in at one end, and, through a spout on the other side, was poured into a ladle, which supplied the Bessemer converters. The metal was supplied both from blast furnaces and cupolas, the former furnishing about two thirds, the latter about one third, of the metal used; but the metal from the cupola system was delivered by a ladle to the converter direct and not through the reservoir. The metal from the blast furnace entered the reservoir in about fifteen-ton ladle lots, and was withdrawn in approximately twelve-ton lots. The chief engineer of the company states that, 'in accordance with the natural way of using the reservoir, it is ordinarily kept well filled up.' That in the practical operation of the mixer or reservoir a large quantity of iron was retained for mixing purposes is evident from the fact that a chalk mark was made on the side of the mixer, which was not allowed to run below the floor, as a guide to the men who rotated or tilted the mixer, since, if the mark went below the floor and out of sight, they could not tell how much iron was left in the mixer. Under these instructions not to allow the chalk mark to go below the floor there was retained in the mixer about 175 tons of molten metal, amply sufficient for the purposes stated in the Jones patent. Its principle of construction was similar to that of the Jones mixer, and its operation identical. Indeed, defendant's engineer himself says: 'With the exception of additions of cupola metal, I do not know that there is any material difference between our practice and that described in the second claim' of the Jones patent. We agree, in the opinion of the circuit court, that 'it is quite clear, in view of these facts, that infringement takes place. That initial mixing rather than storage is the purpose of the reservoir is shown by the fact that the cupola metal is not stored, but served direct in ladles to the converter plant. And that the homogeneous mixture, once obtained, is used as a dominant pool to produce a graduated, nonabrupt product, is shown by the chalk line minimum limit of 175 tons. With such a permanent dominant pool in constant use, we are clear that respondent's practice infringes the second claim of the Jones patent in both letter and spirit.'

If the contents of the mixer used by defendant were allowed habitually to become empty in carrying out its process, there would be no infringement; but all the evidence contradicts this. In the Jones practice this cannot be done, since the mixer cannot be tilted beyond a certain point. In the defendant's mixer it can be done, but is not, since the operator is not allowed to tilt it beyond a certain point gauged by a chalk mark. This seems to be the only foundation for the charge so frequently reiterated, and in varying language, that the methods in use before the Jones process deprived that process of all novelty, and if novelty existed it was by reason of the varying modes of executing such methods; the inference from this being that, as the Jones method was old, it could only be treated as new because of the conduct of individuals in applying the method and their intentions, and that this reduces itself to the proposition that the Jones patent rests upon the mere intention of minds of persons. If we understand this argument correctly, it is that the prior method contemplated storing only, and the mixing was but an incident, while the Jones patent contemplates mixing as its main object and storage only as an incident.

This proposition that the application of this patent depends upon the individual intent of the operator overlooks the essential nature of a process patent. The directions and specifications of such a patent are addressed to those engaged and skilled in the art. It professes to disclose a method of procedure, not the particular instrumentality that may be employed. It may be, as suggested, that one person may, and in ignorance of the patented method, make use of a reservoir merely as such, and without any desire to avail himself of the patented process; but such a fact would not deprive the discoverer of the process of the protection of his patent. Such a supposed case might present a question of fact for a court or jury, and if it were made to appear that the party charged with infringement had, as in this case, changed the instrumentalities used by him after a new method had been disclosed, and particularly if he had for the first time used a special device necessary to that process, a jury might well refuse to believe and find that the defendant was only following the old methods of procedure, and not seeking to avail himself of the plaintiff's invention.

But we think the difference in the two processes may be illustrated by a very simple example: Let us imagine a reservoir containing, say, three quarts, and filled with one quart each of three liquids of different constituent parts, and withdrawn for further treatment at the rate of one or two quarts at a time. Necessarily there would be some incidental mixing, but it would occur at once that the main object of the reservoir was a retention of a sufficient quantity of the mixture to supply the receptacle for further treatment, and if no necessity existed for a longer retention of the liquid in the reservoir, it could be very quickly emptied by two discharges into the receiving vessel. Now, let us substitute for this reservoir a cask of, say, sixty quarts, into which the liquids of different constituent parts are poured in at one end from a multitude of receptacles, and discharged at the other end after remaining a certain time in the cask, and that this cask could not be tilted so far but what a quantity of liquid would be left within it amounting, say, to half its capacity. Now, if there be no distinction between these two operations there would be little left to the Jones process, the very vitality of which consists in the size of the cask relative to the ladles and the mixing of the various liquids poured into it before they are withdrawn.

If, as insisted by the defendant and found by the court of appeals, the reservoir now used is the same in principle as the one which had been in use at the Cambria Iron Works ever since the Bessemer steel was first manufactured there, and the same were adequate for the purposes of the direct process, why was any change made? Therein we think the court of appeals made its most serious error. The defendant had an unquestioned right to manufacture steel as it had been accustomed to do; but instead of that it abandons the Bessemer uncovered ladle of twelve to eighteen tons, and adopts a covered refractory lined reservoir of 300 tons' capacity, and makes use of it, not as before, for the storage of cupola metal, but for the mixing of Blast-furnace metal according to the direct process. This, too, was done immediately after Mr. Morgan's visit to plaintiff's works.

It is true that with the growth of the production of furnaces from fifty tons a day in 1872 to four or five hundred tons in 1895 all apparatus would naturally be increased in size; but why was the open reservoir theretofore used for cupola metal provided with a cover and enlarged in its capacity from fifteen to three hundred tons-twentyfold-while the converter was little more than doubled in size? Why was it so operated that 175 tons were left in the mixer as a dominant pool, if no infringement were contemplated? In the face of these facts the question so earnestly pressed by the defendant, whether the 'method of mixing molten metal,' covered by the second claim, was one for securing a substantial homogeneous composition of metal, to the end of getting a practically uniform product, or was one simply for the purpose of preventing sudden variations in the compositions of successive small portions drawn from the reservoir, without attaining substantial uniformity, loses most of its significance. We do not know how the process can be better described than in the specification itself: 'To provide means for rendering the product of steel works uniform in chemical composition.' The variations in such composition are said to be 'particularly in silicon and sulphur,' and the process to be one of mixing, whereby the particles of metal 'are diffused or commingled thoroughly among each other, and the entire charge is practically homogeneous in composition, representing in each part 'an average of a variety of uniform constituent parts, all the charges of the converter from time to time will be substantially uniform.' This, denuded of all hypercriticism, is the object of the Jones invention, which seems to be the only one yet devised for carrying on what is known as the direct process. If it be true that this process cannot be carried on without infringing the Jones patent, he is certainly entitled to a monopoly of the invention. If it can be, then every method theretofore known for carrying on such process was open to the defendant. But we think the change from the Bessemer intermediate ladle to the Jones mixer was a radical one, and was made for a purpose. That purpose was clearly the adoption of the Jones process.

It is true that before the facts were fully ascertained a stipulation was signed to the effect that the 'amount of molten metal in said mixer (defendant's) varies from nothing to its full capacity, depending on the supply and demand, the supply being generally sufficient to keep the mixer more than half full of molten metal, which metal remains molten therein.' It appears, however, that upon the facts being more fully ascertained, notice was given that, in so far as the stipulation varied from the facts appearing in the testimony of defendant's expert, it would be repudiated, and particularly that portion wherein it was said 'that the amount of the molten metal in the mixer varies from nothing to its full capacity.' As it clearly appears from the mouths of defendant's own witnesses (notably Mr. Morgan) that, in the usual operation of the mixer, the ordinary amount of metal kept in the reservoir was more than one half its capacity, we think that plaintiff's case should not be prejudiced by this stipulation. Stipulations are ordinarily entered into for the purpose of saving time, trouble, or expense, and in this case it recites that 'as defendant's counsel is expected to sail for Europe in a few days, and may not be back for about four months, it is therefore stipulated by counsel for both parties, to save delay, as follows.' But while the stipulation is undoubtedly admissible in evidence it ought not to be used as a pitfall, and where the facts subsequently developed show, with respect to a particular matter, that it was inadvertently signed, we think that, upon giving notice in sufficient time to prevent prejudice to the opposite party, counsel may repudiate any fact inadvertently incorporated therein. This practice has been frequently upheld in this and other courts. The Hiram, 1 Wheat. 440, 4 L. ed. 131; Hurt v. Hollingsworth, 100 U.S. 100, 103, 25 L. ed. 569, 570; Malin v. Kinney, 1 Cai. 117; Barry v. ''New York Mut. L. Ins. Co.'' 53 N. Y. 236.

In short, we are clearly of opinion that the reservoir now in use is used for entirely different purposes from the intermediate Bessemer ladle formerly employed; that the process carried on with it is identical with the Jones invention; and that its primary, if not its sole, use, is for mixing purposes, with necessarily incidental storage, while the Bessemer intermediate ladle was solely used for storage, with little, if any, thought of the advantages to be gained by an incidental mixing.

Discarding now all that does not bear directly upon the validity of the Jones patent, and dropping all superfluity of words, let us determine exactly what Jones has contributed, if anything, to the art of making steel. He undoubtedly found reservoirs of small size in use, in which were poured from receiving ladles enough molten metal to fill them, and from which a sufficient amount was discharged to supply a converter, usually about half the size of the reservoir. But in all these cases the fact whether any particular amount of metal was left in the reservoir was treated as a matter of indifference or accident, although there must have been necessarily some incidental mixing; and probably the metal as it ran into the converters approximated more nearly to uniformity than when it ran into the reservoir. The former methods were adequate for cupola metal, uniformity in which had been largely secured by a careful selection and breaking up of the pigs, but it had not proved a success for blast-furnace metal, except that it had been used to a very limited extent in foreign countries where the peculiar character of the iron ore had rendered it possible to carry on a direct process, although apparently by methods quite other than those employed by Jones. The principal step employed by Jones was to magnify the capacity of the reservoir about twentyfold, provide it with a cover, and to arrange that it should not be tilted beyond a certain point, in order that a 'considerable quantity' of molten metal might be retained in it for a sufficient time to accomplish a pretty thorough mixing, but little change having been made in the meantime in the size of the receiving ladles and converters. As the reservoir was designed to hold a large quantity of metal for a considerable time it must have been covered to obviate the contents being crusted over or sculled.

As soon as this method had proven to be successful by employment at the Edgar Thomson works, and had become so well known as to attract the attention of other manufacturers of steel, if found a ready sale, was adopted by all the leading manufacturers in this country, and was sold for use abroad for about $50,000.

It should be borne in mind that this process was one not accidentally discovered, but was the result of a long search for the very purpose. The surprise is that the manufacturers of steel, having felt the want for so many years, should never have discovered from the multiplicity of patents and of processes introduced into this suit, and well known to the manufacturers of steel, that it was but a step from what they already knew to that which they had spent years in endeavoring to find out. It only remains now for the wisdom which comes after the fact to teach us that Jones discovered nothing, invented nothing, accomplished nothing.

We cannot better conclude this opinion than by the following extract from the opinion of Mr. Justice Bradley in Webster Loom Co. v. Higgins, 105 U.S. 580, 591, 26 L. ed. 1177, 1181: 'But it is plain from the evidence, and from the very fact that it was not sooner adopted and used, that it did not, for years, occur in this light to even the most skilful persons. It may have been under their very eyes; they may almost be said to have stumbled over it; but they certainly failed to see it, to estimate its value, and to bring it into notice. . . . Now that it has succeeded, it may seem very plain to anyone that he could have done it as well. This is often the case with inventions of the greatest merit. It may be laid down as a general rule, though perhaps not an invariable one, that if a new combination and arrangement of known elements produce a new and beneficial result, never attained before, it is evidence of invention.'

The decree of the Circuit Court of Appeals is therefore reversed, and the case remanded to the Circuit Court for the Western District of Pennsylvania for further proceedings consistent with this opinion.

Mr. Justice White, with whom concurs Mr. Chief Justice Fuller, Mr. Justice Harlan, and Mr. Justice Brewer, dissenting: