Full opinion text
HAZEL, District Judge. In this suit in equity plaintiffs, by supplementary bill, geek relief by injunction and damages for alleged infringement by defendant corporation of ten United States letters patent, viz.: Nos. 1,572,343, 1,577,430, 1,502,705, 1,516,069, 1,570,660, 1,600,782, 1,607,498, 1,607,445, 1,597,955, and 1,609,045. The original bill also alleges, as an additional cause of action, unfair trade by the defendant company in relation to the asserted acts of infringement, and, by the supplemental bill, that defendant continues and threatens to continue such acts, to the irreparable loss and injury of plaintiffs. The Colonial Trust Company, as alleged, and as the evidence shows, has a mortgagee’s title to the properties, including the various patents in suit owned by the plaintiff Witherow Steel Corporation, which was acquired from its predecessor, the Witherow Steel Company, and, by reservations in the mortgage agreement, the Witherow Steel Corporation is exclusive licensee of all the involved patents, and, accordingly, I deem that both plaintiffs were properly joined. All the enumerated patents (for convenience designated Witherow patents) relate to the manufacture of automobile axles in nearly finished form — that is, front axle blanks and rear axle shafts, produced in series or strings and then readily sheared into singles from which front axles and rear axles are later finished by forging or machining parts — and a few other metallic articles (not involved), by what is described herein as> die rolling with flash and in a single pass. The art of rolling and shaping white or-molten metal ingots after reduction in size, in a broad sense, for commercial use, is very old. Modem die rolling, credited to Henry Cort in 1783, has greatly expanded, and, for perhaps fifty years or more, has been adapted for so-called parallel rolling and roll flattening (different types of die rolling) for shaping tie rods, concrete bars, plates, flats, rounds, angles, rails, knives, forks, coupling pins, and certain other metallic articles, each having uniform cross-section or a limited change in shape only. These limitations, as to what was accomplishable in rolling or die-rolling operations, were well understood, and, by Witherow’s invention, a new art in a familiar field, or a special class of die rolling, is claimed to have been evolved or developed, achieving a new and distinctive result, which was followed by abandonment of old die-rolling methods of fabricating certain rolled, metallic articles of nonuniform type and dimensions, or so-called complex articles, including front axle blanks and rear axle shafts for use in automobiles. The defenses interposed in the main are prior public use, prior patents, and publication, and specifically that plaintiffs’ adaptations were performable on the old die-rolling machines — no new die-rolling machines being described in any of the patents — and accordingly, Witherow’s concept was not a new art or method of die rolling as contemplated by section 4886, R. S.; in short, that die rolling complex articles and front axle blanks and rear axle shafts was inherent or a natural sequence in the customary die-rolling practices, and, further that neither the use of a declared rudimentary leader bar of the claims of certain Witherow patents, nor generally the rolling of blanks with a fin or flash, presented any serious difficulties of operation which were not remediable or surmountable by any one skilled in the rolling mill art. To establish validity of' the various patents in controversy and infringement by defendant of the 75 collective claims, and in support of the various defenses, a large number of witnesses were sworn on both sides, as the record, consisting of nearly 4,000 typewritten pages, and voluminous briefs, readily disclose. Defendant’s confidence in its belief that all the involved patents are clearly invalid is seemingly apparent by the fact that admittedly it began to die roll axle blanks and shafts for use in automobiles at its plant, after an inspection by its president of Witherow’s operations in such productions under the patents, adapting, as plaintiff avers, substantially the same mode of operation, and selling its axle forms to the automobile industry. The questions submitted are in many aspects highly technical and complicated. We are not particularly interested in; a rolling mill proper as a machine, or with mechanical parts, but our distinctive concern is with an art or process by which certain articles are fabricated. At the outset, a brief - definition of a few of the terms freely used at the hearing as bearing upon the principles specified in the patents and their application may not be amiss. Parallel rolling is a known method of rolling a hot metal bar of uniform cross-sectional size and form between a pair of smooth or grooved rolls for reducing its thickness and altering its shape; and, in die rolling a round bar, uniform grooves are graved or cut into the periphery of the roll or rolls, one ordinarily disposed above the other in their operations. The ordinary hot leader or bar is formed or pressed to the required size from a huge billet, and «the leader in its plastic state is fed into the rolling pass between the two rolls for compression, reducing the thickness and lengthening and shaping to form the particular article or series of articles. There is also a well-known shaping and forming of the leader by a method of roll flattening and rounding of parts. Defendant’s expert, Prof. Trinks, tersely and accurately defined die rolling to be “any rolling in whieh there is a variation in the contour of at least one of the two grooves of the rolls.” In parallel rolling the leader bar engages the pass or groove formed in a way to avoid irregular fins or overfills, since the material, if incumbered with such excess metal, is admittedly, in ordinary rolling with grooved rolls, a detriment and re garded as either commercially useless or as involving expense to put the material or articles in proper condition. Fins or overfills comprise thin extra portions of metal exuded sidewise from between grooved rolls, and, in common practice, to avoid them (for they may lap over the material or spread in such manner as to break the rolls), the roll surfaces were suitably designed or adjusted, or the leader, in parallel rolling, for example, made narrower than the grooves to eliminate the evil. In ordinary rolling, the leader generally fits into the groove, and may, in parts, be either longer or shorter than the matrix portion on the periphery of the rolls. Complex articles are understood to be die rolled and comprise articles requiring reduction in size or thickening at intermediate parts, or where the cross section and center lengths are dissimilar, as, for example, in front and rear axles which, by their contour, are required to have thinner and narrower portions between, but not at, the ends. Roll-set implies spacing and adjustment of the rolls to enable the exuding metal fin, overfill, or flash, as Witherow calls it, to spread sidewise of the groove; while the rudimentary leader, so-called in the specification and claims, is said to be distinguishable by its shape from a leader bar used in ordinary rolling, in that its length is reshaped relative to the pass and roll grooving, and corresponding to desired intermediate portions of the blank, or center to center length requirements. In die rolling Witherow axle blanks and shafts, definite flash zones, intentionally and with an essential object, appear uniformly along the opposite sides of each blank in a string of blanks, unlike the irregular sized and ragged fin or overfill in common rolling, and unlike forging flash which eoneededly lacks purpose or function. Defendant, however, defines flash as an ordinary fin or overfill with which the art has long been familiar, and, in the main, urges that the Witherow uniform flash must, on proper construction of the claims including such element, be limited to its formation in flash gutters during close contact of the two rolls, one placed above or adjoining the other. By tolerance is implied a deviation in dimensional length requirements of the matrieed designs and the axle blanks and shafts, or an agreed safety zone allowance, as counsel for defendant puts it, for the length variations within which the blanks or forms are acceptable to the automobile manufacturers. Before adverting to the original disclosure and claims in suit, it will be' an aid to understanding the intendment to refer to the earlier manner in which automobile axles or shafts were produced. There were several known methods of rolling or fabrication — first, by forming and shaping a front axle blank by the drop forging process, wherein the original impression was separately formed by the use of two heavy metallic dies. Hot metal was put into depressions on the roll surfaces, and both metal and rolls were forcibly driven together for shaping a front axle blank; while rear axle shafts were either drop forged or made by a so-called Ajax process of rolling or machining from bars. The Ajax method required several manual passes of the leader bar forwards and backwards between the matrieed rolls to f orm a unit, which proved to be a very slow method in production, and eoneededly was unsatisfactory. By Witherow’s process, the blanks are rapidly formed by the die rolls from the rudimentary leader, the speed of production corresponding to the speed of the run of the mill, and are formed connectedly in strings of 40 to 50 to each leader, or a single blank in perhaps less than a second. It is not denied that this method of die rolling axle blanks was first practiced at the Witherow plant. It is elicited that the modem method of assembling motorcar parts, prior to patented conceptions, required manufacturers to constantly have on hand a large supply of axle forms (sometimes a three months’ supply) for ready fitting and adjustment to chassis. The demand for satisfactory production of axle forms by speedier methods than was possible by drop forging, or by the Ajaxing method, or by machining, was most urgent, since it was supposed that, if the blanks could be satisfactorily die rolled, they could be bought cheaper at the mill than produced by any known methods. In as few words as possible I have included in the above outline, omitting some details, the relevant conditions of the' art of rolling and die rolling complex metal shapes and forms before the Witherow invention. In 1920, while engaged, as the evidence shows, in ordinary rolling operations, the patentee conceived the idea of supplying the then pressing demand for speedier production of motorcar axle blanks. He essayed to shape and form them by the old principle of rolling parallel and by eliminating undesirable fins or overfills, but he signally failed in all his endeavors and tests. The blanks produced by him could not be successfully used. In each instance they were defective, unaligned, bent, and incessantly longer or shorter than the necessary requirements prescribed by the prospective customer. They were irregular in size and incumbered with excessive and useless metallic fins, notwithstanding repeated trials and adjustments of the rolls. At great expense he struggled onward, hoping finally to solve the problem of producing blanks to meet the specified demands as to tolerances and lengths to fit connecting parts in the automobile chassis. After various failures, he conceived that the irregular metal expressed at the sides of the forms was the serious obstacle to achievement, and he then resolved upon a wide departure from the ordinary die-rolling practices. After many tests and experiments of his new idea, under "varying conditions with different kinds and shapes of leader bars and matrieed rolls, he eventually succeeded in getting a shapely and satisfactory contour and configuration. With a rudimentary leader bar, reshaped to obtain center to center requirements, he was able to produce axle blanks connected together in series, which made it possible for him to supply the trade in large quantities. He had nearly accomplished his full object. It became unnecessary for automobile manufacturers dealing with him to constantly carry a large stock of axles for use in motorcar assembly, for by his process he at all times was ready to supply dimensional requirements of axle blanks and fill demands for delivery. He realized, during his experiments, the importance of dimensional accuracy or center to center lengths and spacings, which were not securable in the adaptation of old rolling practices. By providing intentional flash to the intermediate lengths, where reduction was greater than at other parts, instead of avoiding it by roll adjustment, or in other ways, he secured control of the lengths, shunned harmful skidding of the rolls, which had been a troublesome interference, and secured rigidity of the material, in addition to aiding the annealing of the forms, after severance from the strings, by allowing the free circulation of gases about them. He did not produce a completed product. It was not his object to do so. For, after the units were separated from the strings, it was necessary for the automobile manufacturer to forge their ends or machine them somewhat for his particular uses. The primal inquiry is whether Witherow in reality accomplished something that was unknown to the art — something new and novel — a new system of die rolling complex metallic instrumentalities, and something that the skilled in the art could not make or adapt without having recourse to invention. In a broad sense, as already remarked, die rolling various specified metal articles of diverse form and length, which necessitated grooving the rolls, was very old; bufa number of witnesses, engaged in steel manufacture and long familiar with the rolling and forging art, including roll flattening and parallel rolling, testified that front axle blanks and rear axle shafts were not Tollable by any method known to them prior to Witherow’s evolvement.While, on the other hand, skilled witnesses for defendant, in some instances, challenge Witherow’s claims. In argument it is vigorously contended that the Witherow basic patent, when properly construed, does not at all disclose a new method of die rolling. It is said that axle forms had previously been rolled and marketed, meaning, I take it, by the Ajax process; and, even though first successfully die rolled in strings by Witherow, his adaptation was not patentably new, since the basis of his accomplishment wholly rests on known practical methods. This contention requires particular reference to the prior patents and publications in comparison with Witherow’s innovation. Patent No. 1,572,343. It is claimed to be basically broad and comprehensive, covering different types of complex and simple articles including (Figs. 1-3) a front axle form or blank for special uses. The specification refers to difficulties in die rolling articles owing to “nonuniformity of product,” resulting in losses from scrapping rolled material, and to expense involved in dressing rolls, especially in connection with die rolling axle blanks and the like in large quantities, where accuracy in center to center length was an important factor. Reference is made to certain earlier patents and operative failures arising from defective fin or overfill and the natural tendency of rolls to spring apart, “particularly where thin bodies or blanks are being produced.” It is pointed out that as to one form of the invention “a definitely located flash,” adapted to be sheared away, is provided; also rolling various objects is contemplated, of varying sizes to fixed dimensions, so as to produce blanks connected with each other at the ends, or connected by a gating in preparation for severance and for later forging and machining; that the blanks are prefer- . ably rolled from uniform eross-seetional leaders of rudimentary dimensions corresponding to a finished intermediate portion of the axle blank. The rudimentary leader is described as one preferably imparting to it an I-beam» shape in cross section and proportioned to form a connected string of blanks. The specification proceeds: “It will be noted that the blanks as they eome from the rolls are provided with a flash 7, the width of which varies generally in accordance with the reduction and deforming produced by the die rolls on the leader. This flash is easily trimmed away either before or after performing the finishing operations on the blank.” Exactness in length control and dimensions, measured from a point in one blank to a point on the next succeeding blank and so throughout the string, were thus attained — a result, plaintiff insists, and found to be true, that was unattainable in old adaptations and processes. It was deposed by Prof. Sauveur, an eminent metallurgist, as bearing upon utility, that front axle blanks, formed in accordance with the described process, have a grain structure superior to blanks made by forging and reforging, and were stronger and more durable, especially at the forked arm end of a complete front axle. He explained that, in forging, the grain of the metal was not so well aligned as in die rolling; that the grain runs crosswise and becomes distorted by the heavy hammering (Exhibits, photos 46, 47), subjecting the axle to breakage at the weakened point. Die-rolled blanks, he averred, possessed more firmness and evenness, resulting in decreasing or preventing porosity of the metal. Sherlock, defendant’s metallurgist, differed somewhat as to the effect of porosity, but this variance is not of material importance, for it is fairly shown that in forging operations the contour of the grain weakens the structure when put under great pressure or strain. Claims 3 to 11, covering process and articles, are in issue, of which claims 6 and 9 are typical, and read as follows: Claim 6. “As an article of manufacture, a string of die rolled ferrous blanks reduced while hot throughout at least the major portion of the length of each blank by the die rolling operation, said blanks having flash on diametrically opposite sides thereof in' substantially the median plane of the string of blanks and having accurately spaced apart portions along the string of blanks providing portions on blanks severed from the string and adapted for reworking, said blanks also having other portions of substantially finished dimensions and contour.” ' Claim 9. “The method of forming metal articles, comprising passing a leader between die rolls adapted to form a string of connected blanks, forming accurately spaced portions along the string adapted for reforging, permitting excess metal to flow sidewise in a substantially unrestricted manner to form flash integral with the blanks, separating the blanks and trimming the flash, and thereafter reheating and forging at least certain portions of the blanks.” The essential elements of claim 9 include (1) a leader forming a string of connected blanks; (2) forming accurately spaced portions along the series of blanks adapted for reforging; (3) a free sidewise flow of metal to form flash alongside the blank; (4) separating the blanks and trimming the flash; and (5) reheating and reforging parts of the blanks. To analyze all the prior patents to establish invalidity of the claims or narrowing of scope is not an easy task, and therefore those have been selected having the nearest approach to the Witherow process, or to which weight was given in argument. It is true that the word “flash” is mentioned in various prior patents, and its impediment or claimed usefulness shown. In some instances, provision was made for rolling it back to overcome its effeet, or impairment of the'rolled article. It is found in Richardson’s patent, No. 136,620, of March 11, 1873, and relates to the Ajax method for eliminating roll spring producing the flash, by rolling it back into the material, thus becoming a constituent part of the form. Richardson did not have in mind the use of flash for controlling length needs or securing straightness of his axle forms, nor to reduce roll spring, and therefore his patent is neither anticipatory nor of limiting significance. In Merriam’s patent, No. 596,899, there is a lengthwise web and spacing between the series of balls. After forming the balls, the web had to be punched out; the rolls being apart and the webs formed from a leader bar. In neither instance was the excess metal used for securing length control or center to center dimensions. The front axle forms in strings could not have been produced by these methods or without the use of the combined elements of claim 9 in issue. In fairness, the present invention must be judged by its accomplishment, rather than by bare suggestions that ordinary roller methods by modification and alterations could have achieved the successful production of front axle forms. To my mind it was not merely a contemplation of decreasing or removing roll spring difficulties by using flash gutters to keep sectional roll points in contact with each other in lieu of spacing them apart, as defendant contends. The German patent, No. 111,095, and British patent, No. 23,906, also fail to disclose or suggest Witherow’s adaptation; while the Brown patent, No. 333,106, describes a roller press pressing nuts on a flat bar. None of the ball-fabricating patents show any marked extrusion, but simply a shaping of the articles from the bar, leaving the flat side metal or connecting web which is cut off to loosen the balls; while in the Witherow process, as already stated, there is a decided functional extrusion varying in cross section along the length of the article. It was also impossible to die roll rear axle shafts in strings by any known adaptations, since the difference in the cross-sectional areas of rear? shafts prevented such methods. To die roll them 'by any known process, as defendant’s expert Wadsworth agrees, would not result in tapering lengths, but in uniform cross-sectional length, wherein flash would, not be essential. The proofs show that the strictest accuracy was required in tolerance and intermediate lengths in order to fit the terminals into connecting parts in the chassis, and smaller cross-sectional area at' the spline end was required than at the wheel end, and this needed accurate rounding of the tapered, portions. Although tolerances between the terminals were slight, they nevertheless were essentially requisite. The patent to Aldrich, No. 192,669, for picker teeth, shows a flash adjacent to the points, without reduction of portions of the body length. It has no bearing, since it shows no rudimentary leader, no length requirements from one point to another, or need of material reduction between central parts. Importance is attached to arguments by counsel in the Patent Office during the prosecution of Witherow’s application, in that unfavorable admissions were made regarding the Merriam and Aldrich patents and two British patents, Nos. 622 and 23,906 — that admissions and concessions were tendered regarding adjustment of die rolls and trimming away flash, as described in patent No. 1,572,343, which had the effect of importing into the claims a limitation by which the flash is formed in gutters to overcome roll spring, in order to save them from invalidity. The remarks of counsel, however, taken from the file wrapper, do not bear out this view. The words “flash gutter” do not appear in the claims or in the amendment from which the criticism is drawn, and the word “flash,” used in the description, in my opinion, is not limited to flash exuded or spread into flash gutters. The element of flash gutters in claims 1 and 2, not here involved, does not warrant reading that element into the claims in issue. General Electric Co. v. E. H. Freeman Elec. Co. (C. C.) 190 F. 34; Boyer v. Keller Tool Co. (C. C. A.) 127 F. 130. In any event, interpretation and construction of claims does not depend upon pro and eon arguments in the Patent Office on application for patent issuance, and, unless the inventor acquiesced in limitations imposed upon him, as evidenced by rejection of broader claims, there is no merit found in this contention. Goodwin Film Co. v. Eastman Kodak Co. (D. C.) 207 F. 357; Auto Pneumatic Action Co. v. Kindler & Collins (C. C. A.) 247 F. 328; A. G. Spalding & Bros. v. Wanamaker (C. C. A.) 256 F. 533; Campbell Metal Window Corp. v. Pomeroy & Co. (D. C.) 300 F. 872. It is next contended that the gating 6 in Figs. 1, 2, controls center to center distance instead of the flash or overfill. The specification points out that the gating serves a double purpose, to wit, showing the place of cutting or shearing adjacent blanks; and, second, to prevent roll slipping to obtain efficient rolling from center to center. Figs. 1, 2, however, in connection with Fig. 8 of patent No. 1,577,430, specifies the use of gating at parts of the blank where there is no flash. Both Witherow and Expert Sessions testified that the gating was impracticable and attributed the Figs. 1, 2, in the drawing to a draftsman’s error. But, on the whole, I think, the credit attributed to the gating for securing accuracy in lengths is not substantiated. Many prior patents were mentioned wherein separate elements appear that are included in the process of the instant patent but, since they are not shown as a combination of steps co-operating together, they are not entitled to weight or further analysis. It may be shown, of course, by the analogous art that combining old elements or steps in a process, or aggregating parts into an assemblage of elements or steps did not involve invention, but this rule does not fit this ease — a ease involving a new function and made to attain a new product. Defendant’s witnesses do not say that the prior art in any patent or publication embodies the steps or elements of the claims. Indeed, Prof. Trinks ¡and Expert Wadsworth frankly admitted that they had never heard of the various steps working together to die roll front and rear axle forms, and, though die rolling in box passes was known to prevent fin or overfill, and in open passes to produce it, yet nowhere .is shown a definitely located flash used for a particular purpose, or wherein flash was in contemplation as a step in the process. To anticipate a patent for a process, a prior patent must have contemplated the use of the process. Carnegie Steel Co. v. Cambria Iron Co., 185 U. S. 403, 423, 22 S. Ct. 698, 46 L. Ed. 968. There are, however, various prior patents that were urged by defendant as having an important bearing on the question of anticipation or the scope of the claims, which may briefly be considered. In the Loughran patent, No. 72,512, for clevis-blanks, the rolls have grooves and notches in their rounded surfaces. In pressing out the shape of the article, Wadsworth says the length of the legs requires length dimensions', and in Acheson and Ridley, and patent No. 200,603 of 1878, together with the Browif and Aldrich patents for die rolling various articles, there are articles wherein a certain standard of dimensions is required, but, since they fail, as it seems to me, in the feature of accuracy of length proportions from center to center, which, as before stated, is a prominent feature in Witherow’s conception, they are not anticipatory. The Acheson and Ridley patent for making hoes was cited against Withe-row in the Patent Office. There two box passes are shown with rolls set apart for excess metal, one roll forming the article with a web portion at the ends, the hoes being of uniform width, while the leader is shaped in the top roll and the hoes in the lower. It was not possible to form web or flash at opposite sides of the article by that patent. The Aldrich patent indicates specific length requirements in forming the rolled article* but it lacks a uniform or characteristic leader which contributed so favorably to.rolling axle forms. The skillful mechanic, in my opinion, without knowledge of Witherow’s disclosure, and having these old patents before him, could not have altered or modified the Aldrich or Acheson apparatus to die roll complex articles of the axle blank type. In Loughran patent, No. 212,241, of 1879, for a string of harrow teeth, an article of the complex class is shown, but the principle of Witherow’s disclosure is not apparent. True, there were matriced rolls geared together, and overfill for shearing from the thin sections, and the parts were later reforged, but Lough-ran was not confronted with the intricacies facing Witherow, of producing articles of varying length dimensions and within close tolerances. He says he is able to produce a more perfect article of uniform size and better texture; but, according to the evidence before me, he could not roll articles of center to center requirements with lengthwise reduction of parts and intentional flash alongside. Moreover, his use of rectangular bars and shearing fin or overfill directly on its formation would obviously have prevented string production. If automobile axle blanks could have been rolled by Loughran, or if they could have been produced by making mechanical alterations in his rolls, it is hardly conceivable that the skilled engineer and expert roller would not have taken advantage of his adaptation. Witherow’s tests and expenditure of large amounts of money, and the evidence of steel rollers and steel fabricators strongly negatives defendant’s suggestion. The Hirsch patent, No. 480,325, is said to contain all the elements except the feature of reworking the axle blanks. I do not agree with the contention. If the rolls were in contact in Hirseh’s adaptation, with the articles rolled and shaped in strings, a leader of oval or flat shape in cross section and flash to aid the production, it might justly be considered a good reference, but the description shows neither flash nor overfill in his rolling. The blanks or articles apparently are stamped out, and producing them by Witherow’s principle is not evidenced. The Roger patent, No. 180,-930, for a sucker rod, describes a device of the open pass type, and, though capable of forming flash or overfill, its practicability to roll an axle blank is not apparent. In its operation, flash was not produced, and, moreover, there were no length requirements to his article. The Refior patent was applied for April 12, 1921, after Witherow’s rolling operation. Its operativeness was seriously questioned by plaintiff’s expert witness Sessions. It. lacked a characteristic leader bar reshaped in all its parts, as Witherow used. Wadsworth agreed that, without a leader having portions’ reduced, it could not be drawn through the roll pass. Hill’s patent for axleskelps seemingly relates to a flattening operation, as the wide rolls permitted the metal to spread into the article on pressing down the rolls.' There is no rudimentary leader and no fin or flash resulted because the metal filled the groove and forging gave it shape. It is not, as contended, a disclosure of Witherow’s type of die rolling. The German patent No. 111,095, for ball machine, and British patent to Grant, No. 23,906, cited at the Patent Office against Witherow’s application, the two Rowley axle rolling patents, the Odwalt ma^ ehine, the Ajax type of rolls, the two Baehr patents, and British patent to Klatte for rudimentary leader, and-Ludvigsen for toe calk patent relating to gating between calks, in view of what has already been said in relation to other citations, may be passed without special comment other than to remark that in none is found the combined steps for practicing the Witherow process or sufficient reason for hmiting the scope of the claims. Nor do I find anything in either the Tunner Book or the Beek Book, “Die Gesehiehte des Eisens,” for die rolling coins, of a limiting character with relation to any of the patents considered herein. That the Stahl nnd Eisen article of 1912 recognized for many years that endeavors were made in the art to roll complex articles or articles of various types, which had to be forged or stamped out of the metal, would seem to be a tribute to Witherow's invention. However, the described rolls were differently arranged and were incapable of rotating continuously, and consequently articles could not be produced in strings. The described method is not unlike the Ajax process,. and the mode of operation, not unlikely, was based on Odwalt’s method of rocking the lower roll, and, owing to the mechanism of the gearing, a single blank only can be inserted in the pass. It was necessary in some instances to forge the piece twice — once at each end. Sessions describes the method as one of rock forging; the article is not formed with flash, and a rudimentary leader bar is not used. The dies, cut in the roll, form only a part of a rounded blank, and evidently there was no extrusion of metal. The Withe-row invention was not of a roller forging type, as suggested by defendant; and claims referring to the word “forging” in claim 3, or “reforging” in claim 9, interpreted in the light of the specification, are thought simply to mean forging various parts of the blank. The term “forging,” I agree with plaintiff’s counsel, can be given no broader import than the term “reworking” used in claims 4 to 8, inclusive. The entire record sufficiently shows that Witherow’s patented method was one of unique die rolling of axle blanks and for later reworking their ends by either forging or maehining, or machining only, as shown oh page 2, lines 7 to 10 and lines 17 to 39, of the specification. None of the prior patents or publications could successfully die roll front or rear axle blanks in strings for cutting and finishing, and the presumption obtains that none of the citations were practicable or useful. Cimiotti Unhairing Co. v. American Unhairing Mach. Co. (C. C. A.) 115 F. 498, affirmed 198 U. S. 399, 25 Ct. 697, 49 L. Ed. 1100. I again make reference to the Merriam patent, inasmuch as defendant, after the ease was closed, moved to take additional testimony in relation thereto. The description of the Merriam patent reveals a circular rod which is passed between the rolls B having suitably grooved recesses shown in Eig. Bf. The rounded depressions in the rolls are filled with molten metal, while thin portions of the metal flow out lengthwise on opposite sides of the balls, as already mentioned, and in spaces between them. The web does not perform a definite function as does the flash in plaintiff’s method. There are two rolls positioned in the rear of the forming rolls (Eigs. C and D) which punch out the balls and sever them from the web or flash. It has small spacing along the string, but lacks the accurate spacing adapted for reforging, and is without length or center to center requirements. In plaintiff’s reply brief, it was said that the rolls, when placed in the mill, are so placed that their peripheral surfaces are iri contact so that, in passing the leader through, they jump apart to enable the roll-set to,form flash. This statement, however, was withdrawn by plaintiff’s counsel at the hearing on the motion, it being admitted that the evidence does not show defendant’s rolls in peripheral contact. The limitation reads, “A set or pair of die rolls having matrix portions therein in peripheral contact with the line of contact of the rolls lying in substantially the median plane”; but the other claims do not contain this restriction. It is therefore conceded that defendant does not, in its operation, infringe claim 8, since the limitation is no,t readable into the other claims; but this has no bearing on other claims, and does not negative infringement’ if defendant uses the particular process described in the other combinations. In such a situation, the presumption alises that the limitation , was designedly ■ omitted from the other claims. Ryder v. Sehlichter (C. C. A. 3d Cir.) 126 F. 487. The Exhibit A, introduced on the motion, was not rolled by a Merriam ball machine or mill. It does not possess the Witherow characteristics, and obviously an axle form cannot be rolled by following the directions of the Merriam patent. Indeed, the specification states that in molding the balls, he adapts blunt, flat-faced dies instead of sharp cutting edges, and that he places the rolls a short distance apart to enable the web formed thereby to keep the balls in continuous strings. Plaintiff’s rear axle forms are produced by an entirely different method. Insufficient reason appears for taking additional evidence, and the motion is overruled. There was much testimony bearing on prior use, but none covering die rolling of complex axle forms, except the Ajax method, which the experts agree was on a different principle and incapable of operation, as shown in Witherow’s patent. The prior use testimony does not successfully negative novelty of the claims with which we are now concerned. I have in mind the various exhibit die-rolled articles by the Cleveland Hardware Company, particularly Exhibits 2-B, 2-A, X, R, and Y, consisting of so-called complex or deformed blanks and shapes. I have noted the flat and circular sections of the exhibits, and, in some instances, the cross-sectional areas and formation of fins or overfills, including the evidence that certain blanks were connected together for string production. I have noticed the evidence relating to the arrangement of the rolls or so-called roll-set; but I do not regard such articles or method of operation by which they were rolled as anticipatory. The underlying process in suit is distinguished from the processes by which the exhibit articles were fabricated. It is true that it was not necessary that they should closely conform to the exact shape of Witherow’s product, but, when it is convincingly established that the latter required rigid dimensions as to length and thickness and spaeings or tolerances to impart a distinctive shaping to meet industrial demands for fitting them into connecting parts, then a close examination of the prior use testimony is demanded. It required more than the skill of an ordinary mechanic or engineer to successfully roll axle forms by the old mill and instrumentalities. The accomplishment of the inventor strengthened the presumption of novelty from the grant of the patent, and, before overthrowing this presumption, the evidence relating to prior use of complex articles must be “so persuasive as to leave no room for doubt or controversy.” It is evidenced that, on one occasion before the present invention, the Cleveland Hardware Company contemplated rolling Eord front axle blanks having a varying I-beam cross section in the middle, and was told by its skilled foreman, Kappelle, that it could not be done successfully, and Eord orders for axle blanks were produced by its common method. The witness Gotseh said that the mill could not roll with flash; and Lewis, that it was regarded as a sin to allow fin or flash on a bar, that in his practice, when fin or flash appeared at portions requiring certain length, a redesigning of the rolls for eliminating it was necessary; while Rae of the Oliver Iron & Steel Company, declared that on the occurrence of overfill the material was scrapped as being unserviceable. The language of the Supreme Court in Eibel Process Co. v. Paper Co., 261 U. S. at page 60, 43 S. Ct. 327, 67 L. Ed. 523, is apposite: “Indeed when we consider the indisputable fact that Eibel’s successful experiment at Rhinelander and his application for a patent surprised the whole paper trade, and that for a short time many held back from risking so radical a change and then all adopted it, oral evidence that some persons had discovered the source of trouble and the means of remedying it some years before Eibel is incredible.” See, also, Dubilier Condenser Corp. v. N. Y. Coil Co. (C. C. A.) 20 F.(2d) 723. Plaintiff challenges the origin of the exhibit prior uses articles, asserting that in all probability they were recently rolled or rolled after the invention, and the alterations and erasures in certain book entries seem to support this criticism. Some of the exhibit complex articles were taken from the rollers’ huts at the factory where they were kept, and a few others were rolled a short time before the trial, duplicates of the originals whieh had been destroyed. Whether the duplicates were ever used for commercial purposes does not appear. The dates of rolling the originals depend wholly on the recollection of witnesses, without any corroboratory record entries to inform the court whether they were fabricated by the ordinary die rolling without overfill or flash. The Bailey hanger (Exhibit 2-a) shows a variation of length at the ends, but it has no rigid length between ends as in axle blanks, and there is no flash or fin or heavy reduction of intermediate parts of the hanger. Defendant asserts that at one time the Bailey hanger was rolled with flash, but this method was abandoned for rolling without flash. It was testified that a test roll by defendant of open hearth steel, shortly before the trial, showed flash on the hangers, while the hanger rolled out of Bessemer steel was of different condition than those used formerly for rolling the hangers. Tank strap (Exhibit 2-B) is claimed to have been rolled 7 or 8 years ago, since Witherow’s invention, but tank straps are rolled without flash or accurate center requirements, or reduced parts between the ends, while the grooving was in one roll only. The Q-shaft iron was produced without flash in geared, matrieed rolls; the sample, Kappelle testified, being 32 years old, its body portion of uniform cross section with slight tapering at the end. By what proeess the shaft was rolled does not appear. Exhibit R, a buggy loop, was rolled, it is said, many years ago without flash. Neither the Exhibits S, a clevis-iron; V, a comer iron for truck; W, Eord pedal breakdown bar; or Y, for Eord speed notch — come within the terms of the Witherow invention; each lacking length requirements or reduction of intermediate parts or definitely located flash. They evidently were produced either by roll flattening, or failed to impart length requirements. Actual sales or commercial use of the rolled articles, other than the Bailey hanger and the buggy loop, were not satisfactorily shown; and force is given to plaintiff’s contention that these articles represented abandoned experiments or were rolled by old processes. The prior use or fabrication by Jones & Laughlin, Carnegie Steel Company, Republic Steel Company, and Scranton Company, of guy wire clamps, concrete bars, etc., is subject to a similar criticism of either having been rolled without flash, length requirements, or reduction of intermediate portions, or being produced by means of ordinary roll flattening. Generally speaking, testimony unaccompanied by corroboratory writing as to die rolling the articles in substantially the same way as plaintiff’s method fails to meet the severe test of prior use. Douglas Pectin Corp. v. Armour & Co. (C. C. A.) 27 F.(2d) 814. Defendant has not sustained the burden of proof placed upon it by the patent laws. The articles relied upon are materially different from articles and forms that were roll-'able only by the process in suit. It has long been the rule in patent causes that every reasonable doubt must be resolved against a, defendant relying on prior use to defeat a patent. This rule applies to the characteristics •of the articles and means for their production. The patent, in my opinion, is not anticipated by prior patents or publications, oral testimony of old die-rolling processes, or prior use. The claims in controversy are entitled to a broad construction — one that will cover the generic invention and protect the inventor from appropriation and infringement by equivalent means. Witherow was the first, in my opinion, to die roll articles with an elemental leader bar, designated in the patent as a rudimentary leader, in combination with other steps and elements. He was the first to design and use peripheral grooves in the rolls of varying cross-sectional dimensions for rolling strings of articles, including front axle blanks and rear axle shafts, which were later cut into units; the said articles having intermediate portions of varying dimensions requiring reduction in diameter and with fixed center to center tolerances, and a definitely located flash extending along the intermediate portions of the blank, thereby securing length control and contour, and adapted to be cut or shorn off. It was not merely the application of an old process to a new subject. On the contrary, the old adaptations, I repeat, were unable to die roll front and rear axle blanks. The evidence establishes that his achievement, though perhaps akin to an analogous subject and method, nevertheless required changes and alterations in their application which developed a new and distinctive result; one that not only was beyond the contemplation of the old process, but one which in fact required the exercise of the inventive faculty to make practicable new and original ideas and conceptions. It is claimed that the patents granted subsequent to the first patent are of progressive growth and evolution, and that they contemplate betterment by standardizing the structure, simplifying production, facilitating supply in accordance with the demands of the trade, and devising improvements based on experiments at various stages of operation of the earlier process, which was of practical and economic value, but which became more serviceable by the latter achievements. Patent Ho. 1,577,430. This patent relates to a special form of rear axle shaft, for use in automobiles, which is subject to torsional strains from applying the clutch of the car or from other vehicular stresses. The specification states: “Heretofore shafts of this nature have been taper rolled by eccentric rolls, haVe been turned individually, or forged singly to give the varying cross-sectional dimensions ip different parts as required by the automobile manufacturers. Such methods are only capable of a limited production and necessarily the cost of the shafts is considerable. “I have found that die-rolling as herein disclosed is particularly desirable in articles such as axles or other bodies adapted for torsional strains. The die-rolling process gives rise to an unusually fine and compact grain structure and etched sections of articles produced by my process show that the grains adjacent the exterior of the article are disposed in a substantially continuous manner extending longitudinally of the article and closely following the general contour of the blank. The line of the grains changes gradually and toward the center of the bar merges into a grain structure showing lines substantially parallel to the axis of the bar. The absence of spots showing porous grain structure such as are generally found in forged articles of this character is marked.” The object of the patentee was to produce, by die rolling in a single pass — the rolls being suitably graved — a compact and uniform grain structure such as prior rear axle shafts did not possess. The shafts were rolled in long strings connected at tlieir ends, or by a grating1, as described in Witlierow’s first patent, from a specially shaped leader bar, which was rolled from a square billet in the usual way, preferably of uniform erossseetional area throughout its length and wider than portions of the pass. Each shaft has flash on opposite sides, its varying width corresponding to the quantity of metal reduction. The description says that when the leader is of elongated form in one cross-sectional dimension, its major axis should be slightly larger than the largest cross-sectional dimension of the rolled blank. After the string is produced, the units are severed and the flash sheared off for finishing. In' patent No. 1,572,343, which, in the main, was designed to give correct dimensional spaeings, broad reference was made to a so-called type of rudimentary leader; while here we are dealing with an improvement covering a special sort of leader, a rudimentary oval leader; round leaders being inferior and unsatisfactory. The patent was granted on a separate application to conform to the rules of the Patent Office, which prohibited claiming two species in the basie application. The eleven elaims are involved, but it will suffice to set out article claim 8, and process claim 10. “8. As a new article of manufacture, a die-rolled blank adapted for torsional strains when subjected to rotation, the body of said blank being of substantially circular cross-section throughout the major portion of its length and having a changing cross-sectional area, the blank having longitudinal flash zones on its opposite sides for at least a portion of its length, the grain structure of the article near the periphery thereof being substantially uniform around the article, the line of the grains from the peripheral portions of the article toward the center thereof gradually merging into lines substantially parallel to the axis of the article, substantially as described.” “10. In the manufacture of bodies of different cross section and at least in part of substantially circular cross section, the steps consisting in forming a leader having a major axis and a minor axis with its major axis larger than the largest portion of the body to be finished therefrom, supplying the leader to a die rolling mill in such manner that the direction of compression of the leader in the mill is along a long cross sectional dimension of the leader, and die rolling .the leader to form a series of similarly connected blanks, each of the several blanks having a varying cross sectional area along its length but of substantially circular section throughout a portion thereof, and' then severing the blanks.” The new form of leader is of the essence of the invention. The disputes arise over the questions of lack of novelty and double patenting, in that the present invention is covered in the earlier, and is not distinguishable therefrom. There was rejection of broader elaims filed than those finally allowed. An amendment relating to grain structure imparted from rolling the shaft was filed while the allowance of the patent, or some of its elaims, was in abeyance. The amendment of new matter, in my opinion, does not enlarge the scope of the invention. There was no claim filed at the outset for improved grain structure, but, upon rolling the forms, their shapes were distorted, and there was added a claim to remedy the defect. This amendment does not, in my opinion, amount to taking an essential element of the first patent and including it in the patent subsequently granted, as was the ease in Palmer Pneumatic Tire Co. v. Lozier (C. C. A.) 90 E. 733, or embodying an addition in violation of anything contained in section 4888 of the R. S. (35 USCA § 33). Indeed, the changes apparently gave a more complete explanation of the invention without enlarging its scope. Boyce v. Stewart-Warner, etc. (C. C. A.) 220 F. 124. The assertion that the described die rolls were taken from the prior Baehr patent of 1918 is likewise untenable, since the Baehr patent shows neither flash gutter, nor Witherow’s method of die rolling, nor the production of a torsion shaft. In effect, the reference in the specification to Baehr simply states that the leader bar may be passed through the sectional rolls of a construction embodying certain features there disclosed. All the involved elaims are for a new combination, ineluding the process of the earlier patent, and the addition of specific types of leader bars (Eigs. 1, 2), ellipse or oval shapes, of uniform grain structure — the essential element of the combination. Oval leaders, it is true, were not new in the known art, but their use was restricted to parallel rolling; while in Witherow’s torsion concept there are two kinds of oval leaders, each having “uniform cross-sectional area throughout their entire length,” usable for rolling a plurality of rear axle shafts in strings with flash on opposite sides of each blank. The point is made that the art is not informed as to the particular kind of leader that should be used. But Engineer Sloane, when interrogated, aptly said that the skilled die roller would know at once that he “must have a leader of suitable size to make his product.” And, even though its suitable size is not indicated, the mechanic would know, from the figures in the drawing and the characteristics of the leader, which kind to- use to achieve the purpose of the invention. His reply finds complete support in numerous patent adjudicar tions. See, for example, Goodwin Film Co. v. Eastman Kodak Co. ,(C. C. A.) 213 F. 237. Moreover, in addition to the specific rudimentary leader from which the shafts were rolled, the description plainly instructs that the purpose of the leader is to form compact and advantageous grain structure for use in a new combination covering a new method of die rolling axle shafts. Much reliable testimony was adduced to show its superiority over rear axle shafts produced by old methods and processes. It is also evidenced that a round leader is generally narrower than the grooves on the roll, while an oval leader is wider in parts than various portions of the groove, and in operation a round leader was objectionable owing to the formation of too much flash. The benefits, however, of the oval type resided in its reduction in width and height in the roll, together with a grain structure which definitely followed the contour of the article, eliminated distortion, and imparted a durability not inherent in forged axle shafts or in those made by the old Ajax method. Prof. Sauveur, an eminent metallurgist, attributed great value to the patented structure, owing principally to the parallel run of the grain with the surface of the shaft; while Metallurgists McGee and Stewart swore that its use greatly decreased breakage from strains. Expert Wadsworth alluded to the Lenz patent of 1912, as one of 21 exhibits in Exhibit List E submitted by him, wherein deformed articles were rolled with the object of securing improved grain structure, but his citations are not closely related to the present process. The torsional strains and stresses upon an automobile rear axle are widely different from those on a rolled tie plate, for example. It is emphasized in the patent that the feature of the grain structure resulting from use of the specific leader was not included or suggested in the earlier patent, and therefore the defense of double patenting, in my opinion, is not maintainable. The expert metallurgists were contradictory in their opinions as to the superiority of Witherow’s axle over the Ajax axles. Spears, metallurgist of the Chevrolet Motor Company, caused a comparative test of plaintiff’s, defendant’s and the Ajax shafts (12 in all) to be made, and concluded from the test that the Ajax shaft twisted more before breaking in the testing machine than the die-rolled shafts, and in his opinion the Ajax was superior. He admitted, however, on cross-examination, that he did not personally make the test, and was unaware of the origin of the samples tested, or whether the metal in the different shafts was of like analysis.' In criticism of his experiment, plaintiff contends that in reality it shows that Witherow’s shafts are structurally superior, since the amount of twisting at the yield point was greater than in the Ajax axles. Prof. Sauveur explained that in metallurgical engineering it is understood that the yield 'point in a torsional test is the chief element in ascertaining resistance, and that the yield point should never be exceeded or severely stressed to produce permanent deformation, since then the usefulness of the article for accurate testing ceases. The evidence in its entirety satisfies me that the test report relating to greater ductility and resistance to fatigue of the Ajax shafts was erroneous, and that the twist at the yield point of the die-rolled shafts used in the test was of superior quality. Objection was also urged to the test of Prof. Menefee on the ground that it was secretly made without notice to plaintiffs and should therefore be disregarded. It appears that Chevrolet shafts and round machine bars were tested after die rolling them at defendant’s plant, and that round bars were sent to the Chevrolet Company for testing with Ajax shafts. After subjecting them to heat, they were tested in the presence of defendant’s witnesses Woods and Peterson. These tests are criticized, not' only as having-been made in plaintiff’s ■ absence, but also because the shafts were Brinnelled after separately heating them, thereby producing separate Brinnell hardnesses, and for failure to correctly caliper them. The Brinnell reading was embodied in a report, which, however, has not been exhibited. It was important to know, according to the evidence, before a correct conclusion could be formed by the testing metallurgists, whether the shafts had different Brinnell hardness, since they would have different torsional strength. It was therefore agreed by the parties that another test should be made, open to both sides, by Mr. Comstock whose report (plaintiff’s Exhibit 128) eorroboratively indicates that the Ajax shafts and machine shafts had a higher Brinnell hardness than the die-rolled articles. Notwithstanding the conceded verity of the Comstock report, plaintiff contended that by the first test the shafts were treated to reduce the torsion by improper heating which tended to impart to the Ajax and machine shafts a better quality. To omit properly measuring the diameters of the shafts used for comparison, before heating them, was, no doubt, an oversight; but I think it tended to becloud and make uncertain the verity of the experiment. Upon this point, Prof. Sauveur gave cogent reasons for his conclusion that the reported tests are not entitled to probative weight. However that may be, the experts on both sides agree that laboratory tests must yield to experience resulting from real endurance and practical service of the article. The fact that motorcar makers no longer make their shafts by forging or machining, but instead approve and place reliance on Witherow’s productions, is an impelling reason for ignoring the tests. It is next said that the Witherow shafts are susceptible to cracking along the flash lines, owing to the failure of the grain lines to accurately follow the contour of the shaft near the flash. These asserted structural defects, however, are not sufficiently proven, and, in my opinion, the evidence preponderatingly shows the utility of plaintiffs’ rear axle shafts over prior structures. There was testimony by two witnesses relating to prior uses of oval leaders in ordinary rolling of round bars and tie rods and in flattening operations, and it is argued that it did not involve invention to use an oval-shaped leader in the rolling operation in question. The testimony as to prior use, assuming that it bears upon the torsion structure, is believed insufficiently corroborated. In any event, it is not anticipatory of the plaintiffs’ combination. The oval leader used in ordinary rolling had never before been included as an element in a combination for die rolling torsion rear axle shafts, with flash. In this connection it should be remembered that the patent is particularly directed’ to axle shafts subjected to torsional strains and stresses, as'distinguished from other articles, and that the leader has a major and minor axis, the former larger than the latter, for rolling a shape of varying cross section with flash along portions of its length. The combination of the steps and