Chester N. Weaver, Inc. v. American Chain Co.

MORROW, Circuit Judge. In the opinion just filed in this case, denying a motion for leave to file a bill in the nature of a bill of review, the facts are sufficiently stated to indicate generally the questions in controversy in this appeal. As the discussion of the issues proceeds, other relevant facts in evidence will be stated as required. The appeal is by the defendant from a final decree in favor of the plaintiff, charging defendant with infringement of claims 1, 3, and 6 of letters patent No. 1,191,306 issued to Thomas A. Hoover on» July 18, 1916, on application filed January 24, 1912. The patent was sold, assigned, and transferred by Hoover to plaintiff on July 1,1920, since which time the plaintiff has been and still is the owner and holder of the patent and all rights and privileges thereon. In the ax^plication for the patent, the inventor declared that the invention related to improvements in bumpers for vehicles, particularly to bumpers used upon self-propelled vehicles such as automobiles; that an object of the invention was to provide an automobile bumper which would be simple in construction, comparatively cheap in manufacture, efficient and durable in use, and susceptible of ready repair in case it was broken or otherwise injured; that another object of the invention was to provide a bumper of the character described, which would not rattle, arid which would be susceptible o£easy adjustment, and which might be readily connected with the proper parts of the automobile, such as the frame members or chassis; that a third object of the invention was the provision of a bumper which would yield in all directions, and which would absorb the shock of impact upon striking an obstacle. The inventor stated further that this new bumper, being made of spring steel throughout, would yield readily to forces exerted upon it in any direction; and, if from any cause it became bent or broken, it might be readily repaired, which was not the ease with the square, tubular, and channel bumpers then in use. The new bumper was stated to be attractive in appearance. The specification and drawing describe a bumper consisting of a single piece o£ spring metal, having a front or body portion in the form of a bar slightly curved at its ends, which arc bent around the rivets or pins, and from which extend, parallel to the bar, the branches integral with which are formed the rearwardly extending arms, having suitable ear plates. The bar portion of the spring is reinforced by means of the strip which may, if desired, be made integral with the body portion in the form of a rib. A modified form of the bumper is made up of a pair of bars, centrally disposed, between which is a separator block. The patent contains six claims, of which claims 1, 3, and 6 are at issue in this case. They are as follows: (1) In a fender, the combination with a vehicle frame, of a continuous spring arranged transversely in front of the frame, the ends of the spring being bent upon itself to the rear and inwardly, the bends in said spring forming the ends of the fender, and means secured to the frame and to the said ends of the spring at a distance from the said bends and in rear of the main portion of the spring whereby the spring is supported. (3) A bumper for vehicles comprising a continuous spring buffer bar for extension transversely of the vehicle, said bar having integral spring-supporting members, the said supporting members constituting a continuation of the body member, and being extended laterally therefrom, and then rearwardly. (6) A bumper for vehicles comprising a spring buffer bar for extension 'transversely of the vehicle, said bar being of continuous spring material throughout the entire length thereof, the ends of the bar being bent first upon itself to the rear and then inwardly to form integral spring supporting members, the bends in said bar forming the ends of the bumper. The infringing device owned by the defendant was made by the Metal Stamping Company of New York at its plant at Long Island City, N. Y., as a licensee under patent No. 1,198,246, issued to George Albert Lyon on September 12,1916, on an original application filed April 21, 1913, and on a divided application filed on January 30, 1916, for a motor vehicle buffer composed of resilient parts or springs, each of which may comprise a transverse member, such transverse members overlapping and being secured together by attachment clips or connections, these transverse members extend — 374 — ing longitudinally and secured to the corresponding side frame of the vehicle. It will be observed that the Hoover patent has two important priorities in the proceedings in the Patent Office and in the official action of the officers of the Patent Office with respect thereto: First, priority in the application for a patent wherein'the invention is sufficiently described; and, second, priority in the issuance of a patent upon sueh application. The Hoover application was filed January 24, 1912, and the patent was issued July 18, 3916. The Lyon original application was filed April 21, 1913; and a divided application filed June 30, 1916. The patent was issued September 12,1916. Both of these applications were pending in the Patent Office at the same time, and were the subject of interference proceedings upon the question of priority under section 4904 of the Revised Statutes (U. S. Comp. Stat. 1918, § 9449). That section provides as follows: “Whenever an application is/ made for a patent which, in the opinion of the Commissioner, would interfere with any pending application, or with any unexpired patent, he shall give notice thereof to the applicants, or applicant and patentee, as the case may be, and shall direct the primary examiner to proceed to determine the question of priority of invention. And the Commissioner may issue a patent to the party who is adjudged the prior inventor, unless the adverse party appeals from the decision of the primary examiner, or of the hoard of examiners in chief, as the case may be, within sueh time, not less than twenty days, as the commissioner shall prescribe." It appears from the file wrapper and contents in the Hoover application that on June 6,-1913, the Commissioner of Patents sent to the attorneys for Lyon and Hoover identical copies of an examiner’s report to the effect that a claim in a cop ending application (claim No. 1 in the Hoover application) was deemed allowable; and, if the applicant for the Lyon patent desired to contest the interference, he should submit the claim and put the case into condition for allowance on or before 30 days; that a failure to file the claim within that time would be taken without further action as a disclaimer of the subject-matter disclosed therein, and a patent therefor would not be withheld from the other applicant (Hoover). In the Lyon specification of his original application is a figure 8 referred to as another form''of the buffer. This figure '8 shows with other connecting elements a single transverse member extending across the front of the ear. In response to the notice of the Commissioner of Patents, the attorney for Lyon, on September 13, 1913, refused to contest the Hoover claim No. 1, and notified the Commissioner that figure 8 “is hereby erased from the drawing of the above application.” He also canceled the following description of a broad form of his invention. “My invention in its broader embodiment, however, is not limited to a buffer composed of opposite like but independent parts, as said buffer may, if desired, consist of a single bar presenting a transverse member la, longitudinal members 2a, rearwardly bent yokes 8a, and curved members 4a, as shown, for instance, in figure No. 8.” He also erased claims 2, 15, and 16 relating to certain structural elements of the transverse member. These proceedings in the Patent Office are in effect a formal admission on the part of Lyon with the concurrence and approval of the Patent Office officials, and upon their adjudication that he was not the first inventor of the single spring bar bumper for automobiles described in claim No. 1 of the Hoover patent, and that Hoover was sueh inventor, thus determining that Hoover was the prior inventor of a fender, in combination with a vehicle frame, “of a continuous spring arranged transversely in front of the frame, the ends of the spring being bent upon itself to the rear and inwardly, the bends in said spring forming the ends of the fender and means secured to the frames and to the said ends of the spring at a distance from the said bends and in rear of the main portion of the spring whereby the spring is supported.” Had Lyon been the first inventor in the single bar structure described in claims No. 3 and No. 6 of the Hoover application,,an interference 'would unquestionably have been declared with respect to those claims, as it was with respect to claim No. 1; but no interference was declared except as to claim No. 1, and we must conclude that it was the opinion of the Patent Office officials that Lyon had no claim of priority of invention in his application as against the single bar — 375 — structure in claims No. 3' and No. 6 of the Hoover patent. In the recent ease of Lyon v. Boh, 1 F.(2d) 48, 51, 53, Judge Learned Hand, commenting with his usual ability on this same interference with the Lyon structure involved in claim No. .1 of the Hoover patent, said: “At the very outset the examiner declared an interference with Hoover’s application, and, according to the practice of the Patent Office, drew up a claim for ‘a continuous spring’ supported by figure 8” of the Lyon application “as it then was in the drawings, and by lines 8-13 of page 5 of the specifications. Claims 2, 3, and 4” of the Lyon application “as originally drawn, were upon this form of the invention. This was in effect Hoover’s disclosure, figures 2 and 4, except for the absence of the reinforcing strip, Lyon did not accept the claim, and it became Hoover’s claim 1. Lyon lost it for all purposes as part of his invention, whichever of the two was in fact prior in time. His claims cannot cover such a structure; as patentee, it is as to him a part of the prior art, though not, of course, in any suit in which he does not sue on the patent. “Thereupon Lyon erased claim 2 and figure 8, along with the corresponding text. Claims 3 and 4 were soon rejected, and the application stood without any cla.ims for Hoover’s disclosure. On November 10, 1913, the applicant tried again, with five new claims, to get a single spring buffer, and these were all rejected, because of his withdrawal of the Hoover figure, and his failure to contest the interference. He did the same thing again on January 29, 1914 (claims 8-10), on July 23, 1914 (claims 7 and 8), and on June 23, 1915 (claims 8 and 9), and each timo ho was unsuccessful. When the whole application was renewed, he was bound by his former estoppels. “It is clear at a glance that Lyon’s invention, as it was allowed, is a species of the genus so rejected; he himself calls it a 'broader embodiment* of it. The form is the same — a continuous spring bent into loops and turned at right angles to the front till it reaches the frame to which the ends are fastened. What then did he do? How did ho improve it? By making it adjustable, and by reinforcing the front, nothing more. These things he accomplished by the simple and admirable expedient of splitting the spring in two and allowing enough extra length to each half for an overlap. It remains the same open buffer as before, but it was more adaptable, easier to make, and stronger in front. In function it was a single strip buffer, with attaching legs at right angles; it was an improvement on Hoover, but it was of the same type. I do not, of course, suggest that Hoover was its origin, but only that, for the purposes of this suit, 'by his surrender in the Patent Office, Lyon put himself in the position of an improver upon Hoover. If he meant to claim larger rights, then was the time to assert them; I take him now at his own estimate then.” Commenting further upon Lyon’s claim for a broad interpretation of his claims, the court said: “I think it clear that there is no intimation in word or in function which should give to Lyon’s invention so wide a scope. It has its place, and a very important place, as any one may see who walks the streets; but, while he was prosecuting it in the Patent Office, he did not suppose that it was even for an open spring buffer goneraliter. He is now trying to make it occupy even more, the place of a spring buffer with continuous perimeter, fixed in width, and operating as a unit.” Judge Hand’s opinion is of persuasive force in interpreting the scope of the Lyon invention and the authority of Patent Office proceedings in determining such a controversy. Interferences were subsequently declared by the examiner involving the Hoover application with other patent applications, namely with Prank M. Kramer, October 18, 1913, Rollie B. Pageol, May 29, 1914, and C. H. Duffy, October 23, 1914. These interference proceedings show that the Hoover application was vigorously contested in all its elements in its progress through the Patent Office. We are, however, only concerned for the present with the applications of Hoover and Lyon. With respect to these two applications, we find from the proceedings in interference that the issuance of letters patent to Hoover on July 18, 1916, and to Lyon on September 12, 1916, was an adjudication on the part of the Patent Office officials that Hoover was prior in time and superior in right as to all matters in controversy before them over which they had jurisdiction. To understand the full significance of this adjudication, it is necessary to compare the two structures, and for that purpose to insert here figures taken from the Hoover and Lyon patente diagrammatically identifying the structural elements of the two devices. — 376 — The impact member of the Hoover patent is shown in figure 4 of the patent as follows: In the specification for the Hoover patent, these structural elements have been given letters of identification showing a single piece of spring metal /, having a front body portion g, in the form of a bar slightly curved at its ends, which are bent around the rivets or pins e, and from which extend parallel to the bar the branches h, integral with which are formed the rearwardly extending arms i, having suitable ear plates j. The reinforced strip is identified by the letter g. The form of the buffer bar in the original Lyon application, shown in figure 8, with corresponding claims, has been withdrawn but it will contribute to a better understanding of the elements of these two patents as issued, if we here give figure 2 of the Lyon patent with the corresponding references to detail. The impact member of the Lyon patent, showing the overlapping spring sections to afford adjustability, is shown in figure 2 of his patent as follows: In the application for the Lyon patent, the structural elements corresponding to the Hoover patent have been given figures of identification showing: (1) Transverse or front members extend-, ing across the front of the motor vehicle and a longitudinally or rearwardly extending member. (2) The extended member substantially parallel with the side member 6, of the automobile to which it is secured. (3) A rearwardly bent loop having an open inner end. (4) An inwardly and rearwardly extending curved member. (5) Clips holding together the transverse front members 1 of the opposite overlapping parts of the buffer. The figures of identification for the structural elements Nos. 6 to 13, inclusive, are immaterial here. (16) Bolt clamping the transverse members- 1 in both a vertical and horizontal direction. Anticipations. It is alleged in the defendant’s answei that Hoover was not the original or first inventor of the invention described in his patent No. I, 191,306, but that the invention had been known to and was in public use or had been invented by persons named in a list set out in the answer. In this list we find the following mentioned patents named as anticipations, to wit: No. 554,525, issued-February 11, 1896, on an application filed December 13, 1895, to Robert Muir. No. 873,544, issued December 10, 1907, on an application filed March 29, 1906, to Ray W. Harroun. No. 955,624, issued April 19, 1910, on an application filed December 8, 1909, to Dante J. Wei ton. No. 969,143, issued August 30, 1910, on an application filed February 18,1910, to Ralph N. Harris. No. 1,202,690, issued October 24, 1916, on an application filed June 6, 1910, renewed November 21, 1913, to Rollie B. Fageol. Was the Hoover Invention Anticipated? In the Book of Exhibits, the application and specification with drawings and claims for the elements of each device is set out and described for the information of the public and those skilled in the art. With this information before us, we have diligently searched for the elements alleged to anticipate the essential elements of the Hoover invention. The Muir patent, No. 554, 525, is for a new and improved attachment for car fenders. The specification shows clearly that the device is for a trolley car attachment, and has, for its declared purpose, the providing of a plurality of very elastic springs for the fender, which should be adapted to neutralize the shock of impaet if the bumper strikes a person, and thus avoid any injury to the person struck. On the fender frame a wire netting may be secured to afford an elastic bed whereon the person struck by the fender may fall and be earned without material injury. It is specified that the invention comprises the plurality of “very elastic, thin, plate springs, which are arched in front of the frame” and lie in the same plane therewith. The springs have different degrees of — 377 — curvature, and have their ends secured in clamping boxes at each side o£ the fender frame, which will space the springs apart and allow each spring to preserve its resilience. We find nothing in this device that anticipates the Hoover structure, unless the “very elastic thin plate springs arched in front of the frame” he admitted to a place of structural invention in the art, and this we cannot do under the patent law. The element is in and of itself an old and well-known device, performing no new function in the combination to which it is assigned as an improved attachment for a ear fender. The Newcomb patent, No. 969,143, is for a new and useful ear buffer fender. It is specifically stated that it is an invention to provide means for preventing the destruction of life by trolley cars. The buffer fender is to throw standing persons off the track. It is specified that, in case the car strikes any person standing on the track, the tendency of the buffer will be to relieve the stress of the blow, and to shunt the person to one or the other side of the track. The bumper is collapsible to a minimum space for the storage of the car in the ear barn, and to provide a fender of the character stated which will not interfere with the use of the drawbar or other bar. The combination includes bendable; ribs, mounted in front of the ear, and means for bending the ribs to bow them out in front. We find nothing in this structure anticipating the Hoover invention. The Harroun patent, No. 873,544, is for an automobile bumper placed in front of the car to protect the same in case of collision. If consists of a rigid horizontal bar, extending across the front of the car at, a proper distance from the ground. When the ear collides with an object, the bar or rail, being in advance of anything on the car, will receive the shock. Coiled compression springs are interposed between the bar and the frame of the car. When the collision is of sufficient force, these springs will be partly compressed. There is nothing in this device calling for comment. It has obviously no anticipating structure. f£he Welton patent No. 955,624, is for a fender for automobiles. The impact member consists of a tube of rubber and fabric, tbe material of which is of such thickness and rigidity as not to require inflation to maintain it in distended form, but of such flexibility as to permit it to yield when in forcible collision with other objects. The tube is plugged at its ends with rubber plugs, and, tlie tube is made of such length that it shall extend across the front of the vehicle and wheels, and it is mounted upon a suitable curved seat of metal, wood, or other rigid material. The tube and its seat are mounted on a metallic supporting bar, having its ends bent to 8 and revei-sed S form constituting springs; the seat being bolted or riveted to the free terminal portions of said springs. It is claimed for this structure that the resilience of the tube is supplemented by tbe resilience of the springs behind it, and that, upon collision, danger of injury to persons or the vehicle is greatly lessened. There is nothing in this structure requiring comment. Clearly it does not anticipate the Hoover invention. The Harris patent, No. 981,260, is for an improvement in fenders for motor vehicles. Its purpose is to form a partial protection to pedestrians who may be struck by tbe fender; the fender being placed about tbe height of the knee, furnishing sufficient support to carry the person struck in case he falls upon the fender. The fender includes a bar of channel iron which forms the base of the fender, and which extends transversely of the car and attached thereto. At the opposite ends of the bar of channel iron are clips. Mounted between the clips is a bowed spring; this spring being formed precisely like the usual carriage spring and consisting of superposed leaves, each leaf being shorter than the leaf immediately beneath it. Supported on the exterior face of the said spring leaves is the buffer piece, preferably made of rubber with a rounded outer face. This device is in no sense an anticipation of the Hoover invention. The Paged patent, No. 1,202,690, is for an improvement in automobile bumper, relating generally to improvements in means for protecting vehicles against damage resulting from collisions with moving or stationary objects. The affidavit of Prof. Chas. Moser, a civil engineer by profession, and an instructor in mechanics of materials in tbe Stanford University, was taken under equity rule No. 48. The witness qualified himself as an expert by stating that the problems that he was called upon to consider in connection with his work related to the internal stresses, the deformations and stability of the various elements of structures subject to loads. He says, among other things, that he is familiar with the bumper illustrated and described in Paged patent, No. 1,202,690, and understands the action of such a device and the results obtainable therefrom. In bis opinion the — 378 — Fageol device is not a practical bump absorber for the-following reasons: “(a) The loop end spring arms are made of stock having a circular cross-section; and (b) the larger £>art of the impact bar is composed of a substantially rigid tu,be which is not intended to deflect under impact. The round stock, of which the loop ends are made, seriously interferes with any spring action on the part of the bumper.- The circular cross-section is admittedly an excessively stiff form of structure horizontally and too weak vertically. As compared with a rectangular cross-section of the proportions used in the Hoover bumper, the circular section is one-fourth as stiff in the vertical direction 'and eight times as stiff in the horizontal direction. ' “The round stock is also objectionable because the flexing of the same at the ends of the loop and at the points of attachment is likely to cause crystallization and breakage. The round stock is further highly objectionable because a bumper made of such stock is necessarily as capable of vertical vibration as of lateral deflection, and it is constantly subjected to severe vertical vibrations in usage over rough or even smooth roads. This vertical vibration must necessarily bring severe strain on the metal constituting the supporting arms, and, if this metal be of good spring quality, it is likely to crystallize, become brittle, and crack, thus resulting in the destruction of the bumper in case of impact at the time when it is called upon to perform its intfended function. ■ “Another very marked effect of the round stock used for the spring arms in the Fageol bumper lies in the fact that, when the bumper has been deflected under impact to such extent that the loops have closed, the front member of the loop will not bottom on the rear member, but will pass over or under the rear member. This fact presents two serious objections-: First, the danger of the metal at the ends of the loops parting under the very severe strain that will be imposed as the front membey of the loop passes beyond the rear member; and, second, a building up of resistance of 1;he bumper under further deflection will not take place as in the ease of the flat strip stock, where the loop ends bottom when they close and thus build up the desired resistance. “Finally, the use of the long rigid member extending over the greater portion of the impact bar very seriously interferes with the resiliency, deflection, and bump-absorption value of the structure as- a whole. “While this Fageol device shows the use of loop ends, the form of the material selected for the ends is the worst that could possibly have been found, and seriously affects the deflection possibilities of the device. The use of the rigid central bar also seriously diminishes the deflection possibilities of the bumper, and'" the failure to provide any means for the bottoming of the loops under severe deflections prevents the building up of the necessary resistance under heavy impacts, and the use of round stock in the loops and' supporting arms permit crystallization under vibration and strain, and in iny opinion renders the device impractical.” In Lyon Non-Skid Co. v. Edward V. Hartford, Inc. (D. C.) 247 F. 524, 533, 534, the suit was brought in the Southern District of New York by the Lyon Non-Skid Company on the Lyon patent against the Edward Y. Hartford, Inc. The defendant brought a cross-suit in defense on the Fageol patent. In August and September, 1916, Fageol had filed amendments to his application in the Patent Office, enlarging the scope of his device as a shock-absorbing bumper. Complainant claimed that these amendments were made by Fageol as a part of the plan of adoption of the Lyon bumper. The court, referring to anticipating devices, mentions the Hoover bumper as showing the general lopped end form of bumper of one single strip, and says the Hoover patent was granted after the Lyon patent. This was an obvious inadvertence on the part of the court, 'as the Hoover patent was issued July 18, 1916, and the Lyon patent on September 16, 1916. The court also says that the Lyon dates of invention are carried back of the Hoover application date. It will appear later, that one of the questions in this ease is whether the evidence. has carried the Hoover invention, not only back of the Hoover application, but back of the Lyon invention. But for the present we are dealing with the Fa— geo 1 patent, describing the Fageol bumper. With respect to that device, Judge Mantón said: “It is a rigid bar type of bumper in which a round rigid bumper bar is used, extending across the entire distance between the automobile frames. This was a decided reversion from the Lane bumper. It is claimed that the end loops, because they were necessarily made of round stock to secure the adjustable connection with the bumper bar, vibrated freely in response to the successive vertical shocks under running conditions of the automobile, with a consequential breaking. This bumper was not successful. Fa — 379 — geol turned out about 200 in California during the first half of 1910; about 17 or 18 were sold as bumpers and the *balanee as junk. It did not serve the requirements of an automobile bumper. This, together with the experiments in California, demonstrate that the Fageol bumper was a failure commercially, as well as practically.” The record discloses an interference between the Hoover and Fageol applications declared by the Commissioner of Patents May 26, 1914. The result of this interference will be considered when we come to the question of infringement. We conclude that the Fageol invention was not an anticipation of the Hoover invention. We do not find that the functional element of the fenders described in the foregoing patents yielded effectively or in a practical way tó a collision coming from any direction, and by continuous spying action across the front of the automobile absorbed the shock of such a collision by bringing the car to a standstill over an appreciable period of time. This is the soul, of the Hoover invention in its advance in the art, and distinguishes it from all the fenders and bumpers cited in anticipation or referred to in the record. It distinguishes the Hoover from the Abresch-Cramer fender, which is referred to but is not cited in anticipation. This device was made for use on heavy trucks weighing as much as 8,000 pounds, and adapted to carry a load of 10,000 pounds, making an aggregate of 18,000 pounds. The fender extended across the frame of the truck, and a few inches only on each side, but did not extend from wheel to wheel. It was in all respects a rigid bar bumper, made of nonresilient steel or iron, and intended to knock obstacles out of the way. Louis Schneller, a stockholder and vice-president of the Abresch-Cramer Auto Truck Company, testified that the bumper was made of heavy stock, “so anything the truck might meet would be damaged more than the truck itself” — had no intention of producing a spring bumper. This rigid bumper was generally the dominating feature of the prior art, and we may as well say here, without further discussion of prior patents or devices, that from the evidence before us either Hoover or Lyon was. a pioneer in the effective and practical shock-absorbing feature of the bumper art. The Lyon invention is described in the specification and claims of patent No. 1,198,246. The controversy in the Patent Office, Lyon’s withdrawal in interference, and the adjudication of its officers upon the question of priority between Hoover and Lyon, submitted in the applications for these two patents, have reduced the present controversy to very narrow limits. “It is well known,” said the Supreme Court in Burns v. Meyer, 100 U. S. 671, 672 (25 L. Ed. 738), “that the terms of the claim in letters patent are carefully scrutinized in the Patent Office. Over this part of the specification the chief contest generally arises. It defines what the office, after a full examination of previous inventions and the state of the art, determines the applicant is entitled to.” When Lyon refused to contest the Hoover claim No. 1, and notified the Commissioner that he erased figure No. 8 and canceled the broad description of his invention, and also erased claims No. 2, No. 15, and No. 16, he withdrew his claim of priority to the essential elements of the invention described in the Hoover patent, No. 1,198,246. This was the opinion of Judge Hand in Lyon v. Boh, 1 F.(2d) 48, 51, 53, which we have heretofore quoted. This opinion is entitled to respectful consideration. The Priority of Invention as Between Hoover and Lyon. An effort has beeh made to extend the question of priority between the Hoover and Lyon device in this suit back to the time when these two inventors first conceived the idea of their respective structures and reduced that idea to practice. The question turns largely upon oral testimony relating to the claim of Hoover that he conceived the idea of his invention in April, 1910, and reduced the idea to practice in July, 1910. The defendant claims that Hoover did not reduce his conception to practice until August or October, 1911, and that in the meantime Lyon conceived his invention and reduced it to practice in January, 1911. The evidence of the witnesses was taken before the master in chancery, and he had the opportunity of weighing their testimony and judging of their credibility where there were material contradictions. The master reviewed this testimony fully in his report to the court, and found that the letters patent to Hoover were valid, and that Hoover was the original and first inventor of the deviee described therein. There is a presumption in favor of the correctness of this finding. The District Judge, who heard this report on exceptions, reviewed this evidence independently in a memorandum and in a supplemental opinion- In both opinions he states — 380 — the question in controversy in the following language: “Defendant, undertook to show a reduction to practice by Lyon as of a date anterior to Hoover’s application. This defendant was called upon to do ‘only by a fair preponderance of the evidence,’ and, assuming its success in that regard, plaintiff was then entitled to show, if possible, by a similar preponderance of evidence, a reduction to practice by Hoover antedating that of Lyon., This was attempted, and, as before stated, assuming the sufficiency of Lyon’s showing, under the rule, the real question in the case is, Did plaintiff show by fair preponderance of the evidence addressed to that point, a reduction to practice by Hoover antedating the reduction to practice proven by Lyon? From a careful consideration of the evidence introduced before the master relevant to that particular issue, I am constrained to find and hold that this question should be answered in the affirmative.” ' The court concludes both opinions as follows : “The findings of the special master as upon the evidence, together with his report and recommendations as to the validity of the Hoover patent, are approved.” In Cantrell v. Wallick, 117 U. S. 689, 695, 6 S. Ct. 970, 973, 29 L. Ed. 1017, the Supreme Court had before it a controversy relating to prior use by the defendant of a structure in which the invention described in the patent of the plaintiff was embodied. Upon a charge of infringement, the court had this to say concerning the proof in such a 'case: “The burden of proof is Upon the defendant to establish this defense, for the grant of letters patent is prima facie evidence that the patentee is the first inventor of the device described in the letters patent and of its novelty. Smith v. Goodyear Dental Vulcanite Co., 93 U. S. 486 [23 L. Ed. 952]; Lehnbeuter v. Holthaus, 105 U. S. 94, 26 L. Ed. 939. Not only is the burden of proof to make good this defense upon the party setting it up, but it has been held that ‘every reasonable doubt should be resolved against him.’ Coffin v. Ogden, 18 Wall. 120, 124, [21 L. Ed. 821]; Washburn v. Gould, 3 Story, 122, 142, Fed. Cas. No. 17214.” In Parker v. Stebler, 177 F. 210, 213, 101 C. C. A. 380, 383, this court said: “In brief, the courts have recognized the rule that the oral testimony of witnesses, speaking from memory only in respect to past transactions and old structures claimed to anticipate a patented device, physical evidence of which is not produced, is very unreliable, and, that it must be so clear and satisfactory as to convince the court beyond a reasonable doubt, before it will be accepted as establishing 'anticipation.” The defendant has failed to show by preponderance of proof where it was required to show beyond a reasonable doubt the prior conception and reduction to practice of the Lyon structure. In our opinion, the weight of the evidence is clearly against this defense. Infringement. Upon the question of infringement we have reached a different conclusion from that of the court below. We are of opinion that Lyon has made an invention in -the bumper art which we,may treat as improvements upon the Hoover structure, and for which Lyon has received a patent. The Lyon bumper is made of two similar parts, each one forming one-half of the whole transverse structure ,of the bumper across the front of the vehicle; the two parts having” sufficient extra length to meet and' overlap in front and detachably connected together with suitable dips. This structure has a reinforcement and slippage in the two overlapping parts operating more or less effectively as a spring in ease of collision. The Hoover bumper has no such element in its one single continuous buffer bar. We think the Lyon structure is in this respect an improvement on the Hoover structure in the spring, shoek-absorbing quality of the humper. The two parts of the Lyon bumper have an adjustability to the side frame members of automobiles located at different distances apart that the Hoover bumper has not, and is therefore an improvement on the Hoover bumper in its usefulness in adjustment to automobiles of different widths. The Lyon bumper has, at the end of each transverse member, a rearwardly curved open loop, acting, as additional springs and as wheel guards of the vehicle. The Hoover structure has, at the end of its transverse member, elongated closed loops which, in the specification and diagrams, are bent around rivets or pins. The Lyon structure has,'in the open loops, another spring or yielding element we do not find in the Hoover structure, and which we believe to be an improvement in the latter. All these improvements in the Lyon bumper add to the shock-absorbing quality of that bumper, increasing its usefulness and efficiency. — 381 — We understand that the rule of law applicable to such improvements is stated by the Supreme Court in Cantrell v. Wallick, supra, where the court said: “The great majority of patents are for improvements in old and well-known devices, or on patented inventions. Changes in the construction of an old machine which increase its usefulness are patentable. Seymour v. Osborne, 11 Wall. 516 [20 L. Ed. 33]. So a new combination of known devices, whereby the effectiveness of a machine is increased, may be the subject of a patent. Loom Co. v. Higgins, 105 U. S. 580 [26 L. Ed. 1177]; Hailes v. Van Wormer, 20 Wall. 353 [22 L. Ed. 241]. Two patents may both be valid when the second is an improvement on the first, in which event, if the second includes the first, neither of the two patentees can lawfully use the invention of the other without the other’s consent. Star Salt Caster Co. v. Crossman, 4 Cliff. 568, Fed. Cas. No. 13321.” This principle of law is well illustrated in Westinghouse v. Boyden Power Brake Co., 170 U. S. 537, 18 S. Ct. 707, 42 L. Ed. 1136, where Westinghouse brought suit for infringement of the Westinghouse patent, No. 360,070, issued to Mm March 29, 1887, for new and useful improvements in fluid pressure automatic brake mechanism, the object of which, it was said in the specification, was “to enable the application of brake shoes to ear wheels by fluid pressure to be effected with greater rapidity and effectiveness than heretofore, more particularly in trains of considerable length, as well as to economize compressed air in the operation of braking, by utilizing in the brake cylinders the greater portion of the volume of air which in former practice was directly discharged into the atmosphere.” The defendant’s mechanism was covered by patent No. 481,135 for improvements in fluid pressure brakes, for a more powerful and also a quicker application of each brake at the same time, and No. 481,136 for improvements in valves for automatic air brakes, to produce a quick application of the brakes, both issued to Boyden on August 16, 1892. Boyden improved upon the Westinghouse air brake mechanism by a different arrangement of the train pipe and valve system in venting the air of the train pipe directly and more promptly into air brake cylinder for the purpose of securing quick action in emergency work. The ease has points in common in both law and fact with the Hoover and Lyon controversy concerning the bumper mechanism on an automobile required in the quick work of an emergency in absorbing the shock of a sudden collision. The Supreme Court sustained the Boyden improvements, using language appropriate here, as did the Circuit Court of Appeals, whose opinion was affirmed by the Supreme Court. We refer particularly to the language of the Circuit Court of Appeals concerning the respect duo to the decisions of the Patent Office officials in such a case. The question was whether the Boyden device, successfully accomplishing the function of discharging train pipe air into the brake cylinder simultaneously with the triple valve discharge of auxiliary reservoir air into that cylinder, was the mechanical equivalent of the prior Westinghouse device, having another stem, another valve, and by passages peculiar to itself leading from the additional to the air brake cylinder. The Circuit Court of Appeals said: “This question was presented, necessarily, to the Patent Office of the United States when Boyden applied for a patent for the device under consideration. That office employs the best experts in mechanics which it can secure in this and other countries. Its examinations are indeed ex parte in form, but they are nevertheless conducted under hot and skilled contestation in every ease of importance; and its decisions, though not conclusive, are entitled to great respect. That office, after full examination, awarded a patent to Boyden on the 16th day of August, 1892, for his quick-action improvement” on the device of patent No. 220,556, and that action by the office was in effect a ruling “that the Boyden device did not infringe Westinghouse’s quick-action patent, No. 360,070. That ruling takes rank here as the testimony of experts of the Mghest experience, skill, and knowledge in mechanics. That ruling was subsequent to the issuing to Westinghouse of both the patents, Nos. 360,-070 and 376,837, four years after the latter patent, when the Patent Office had full knowledge of them.” 70 F. 816, 827, 17 C. C. A. 430, 441, 25 U. S. App. 475, 569. In the present case the Hoover and Lyon applications for patents were pending in the Patent Office at the same time, and were subjected to the close scrutiny of a hot contest by the officials of the Patent Office in interference proceedings, and the Lyon patent was issued one month and twenty-four days after the Hoover patent. That action by the Patent Office was in effect a ruling that the Lyon patent was not an in — 382 — fringement of the Hoover patent-, and that ruling is entitled to great consideration here. The Fageol interference, declared by the Commissioner of Patents on April 29, 1914, involved claim No. 3 of the Hoover application as follows: “A bumper for vehicles- comprising a buffer bar for extension transversely of the vehicle, spring-supporting members for bar, the said supporting members constituting a continuation of the body member and being extended laterally therefrom.” The interference resulted in a decision of priority in favor of Fageol on February 4, 1916, and became the subject-matter of elaim No. 6 in the Fageol patent, No. 1,202,690, issued October 24, 1916. Thereupon Hoover amended his application by inserting in claim No. 3 “continuous spring” before “buffer” in line 1, “said bar having integral” before “spring” in line 3, canceled “for said bar” after “members” in line 3, changed the period at the end .of the claim to a comma, and added “and then rearwardly,” and the claim so amended became claim No. 3 of the Hoov-, er patent. It is contended now on behalf of the Hoover patent that the words “continuous spring,” as used by Hoover in amending claims No. 1, No. 3, and No. 6, was intended to qualify the word “springs” and not the word “bar,” requiring a structure having “continuous spring action.” This was the decision of the special master and the Judge of the District Court, but we cannot concur in that construction of the words “continuous spring.” We think the qualifying words were intended to qualify the words “buffer bar.” The words will have the same meaning then in claim's No. 1 and No. 6. Upon the other hand, if we should construe the words “continuous spring” as* meaning “continuous spring action,” we would give the words the element of a “function,” and, as the “function” of a machine cannot be patented, we are confronted with the.decision of the Supreme Court in Westinghouse v. Boyden Power Brake Co., 170 U. S. 537, 557, 18 S. Ct. 707, 42 L. Ed. 1136, that the performance of a “function” must be limited to the particular means described in the specification, which we find to be in the Hoover specification “a single piece of spring metal having a front or body portion in the form of a bar slightly curved at .its ends, which are bent around the rivets or pins, and from which -extend parallel to the bar the branches integral with which are formed the rearwardly extending arms.” This specification we must read into the claim, and, being so read, we must hold that the Hoover amended application does not cover the Lyon structure. - We are therefore of the opinion, upon all the considerations we have stated, that the Lyon invention does not infringe the Hoover invention. A decree will be entered in accordance with the view expressed in this opinion, affirming the decree of the court below adjudging that plaintiff is the owner of letters patent No. 1,191,306, that Thomas A. Hoover was the first original and sole inventor of the invention set forth and described in said letters patent, that said letters patent are good and valid; reversing the decree in adjudging that the defendant, Chester N. Weaver Company, Inc., has infringed the said letters, patent by the sale of the bumpers described in the complaint.