POPE, Circuit Judge.

The appellant, owner of Harwin Patent No. 2,740,739, for Carpet Underlay, brought this action against the appellee charging infringement of its patent and seeking damages therefor. After issue joined, the case was tried to a jury which returned a verdict for the plaintiff. Thereafter the court granted defendant’s motion for judgment notwithstanding the verdict under Rule 50(b) of Federal Rules of Civil Procedure, 28 U.S.C.A., and entered judgment accordingly for the defendant.

The action of the court was based upon the findings of the trial judge that the claimed patent lacked novelty because of anticipation of its claims by developments in the prior art. By this appeal the plaintiff challenges the propriety of the court’s rejection of the contrary finding of the jury and its own finding of anticipation.

Plaintiff’s patent relates to a type of product which is used to underlay, anchor and cushion floor carpets. In this general area of product, such carpet underlays are useful in cushioning the carpet and making it more resilient. For this reason the main constituent of the underlay is made of rubber, usually foam rubber, because of its cushioning effect. The rubber serves to anchor the carpet and prevent slipping along the floor.

Such devices are usually completed by attaching to the upper surface of the rubber pad which lies next to the floor a fabric material sealed or bonded in some manner to the rubber pad. Prior to *732 plaintiff’s product, this was done by using a cement to join the fabric to the rubber. This fabric serves two purposes: It furnishes a fairly flat surface on which the carpet may lie and it holds firmly in place the rubber pad to which it is affixed preventing it from spreading by lengthening or widening as a result of use, as it might otherwise do. Some of these carpet underlays are made of foam rubber formed in corrugations so as to facilitate air reaching all parts of the pad and thus aiding in retention of its resiliency. When the rubber pad is formed in that manner the anti-spreading action of the attached fabric is particularly useful as a corrugated rubber pad is more likely to spread apart than would be the case with a solid rubber one.

Plaintiff’s claimed invention, as the stated claims disclose, relates solely to the manner in which the fabric portion of the underlay is attached and fastened to the rubber which lies beneath it. The patented structure discloses that the coarse fabric, with interstices approximately one-sixteenth of an inch square, has been bonded to the rubber through the fact that the rubber has been caused to penetrate the interstices of the fabric so that after passing through such openings the penetrated rubber has spread out and formed protrusions with knob-like heads larger than the interstices. The result is a key-locked connection between the fabrics and the sponge rubber which has been accomplished without the use of any adhesive.

In further explanation of the structure, the specifications state that this key-locking mode of attachment of the fabric to the rubber pad with its waffle-like corrugations is accomplished by heating and blowing the sheet of sponge rubber into a corrugated mold whereby the rubber is held in intimate contact, with the fabric so that in blowing or expanding, the rubber is forced to enter the interstices of the fabric and swell around adjacent threads of the fabric and form the headed knob-like projections that are somewhat larger in size than the areas of the fabric interstices. The claims as stated appear in the margin. 1

In this case, as in many other cases where a claimed invention is attacked on the ground that the claims of patented devices were anticipated by earlier patents, the decision as to validity may turn upon a question of fact. Leishman v. General Motors Corp., 9 Cir., 191 F.2d 522, 529. (See other cases cited in Bergman v. Aluminum Lock Shingle Corp. of America, 9 Cir., 251 F.2d 801, 809-813 concurring opinion.) Our first inquiry necessarily is: what was the prior art as disclosed by the prior patents, which were part of the record in this case? The question is whether insofar as this factor entered into the verdict of the jury, the evidence of anticipation was so overwhelming and compelling that it can be said that as a matter of law there was a total failure of proof of the essential novelty to sustain the plaintiff’s case *733 so that the district court was justified in granting judgment notwithstanding the verdict.

The Judge’s conclusion, as disclosed by his memorandum of decision, was based upon his views of two prior patents, Blair and Bodle, No. 2,480,316 and Coyte, No. 562,281. This he stated in the memorandum as follows: “The Patent in suit does not disclose invention in the light of the teaching of Blair and Bodle United States Letters Patent No. 2,480,316, particularly in view of Coyte, United States Letters Patent No. 562,281.” So far as those two patents are concerned, we have no difficulty in arriving at a conclusion that neither one of them was of such character as to establish anticipation. As we have said, the novelty claimed for the patent here in suit relates to the manner of key-locking or bonding the woven fabric to the rubber pad; the Blair and Bodle patent was a process patent stated to be for a method of making laminated cushion material. That patent described a product made of a sheet or lamination of rubber which might be shaped or formed into corrugations or hollow ribs which had applied to it a backing sheet of material such as burlap to reinforce the rubber material. However, the method of attachment of the backing sheet is there described as follows : “The backing sheet 12 touches and is secured to the peaks or tops of the embossments or corrugations at one side of the rubber sheet as by a layer or coating 13 of cement, rubber or the like, or by being vulcanized directly thereto.” 2

Another prior patent relied on by the trial court as an anticipation has even less resemblance to the claimed patentable feature of the patent here in suit. Claims 1 and 2 of the plaintiff's patent described a project which has rubber “formed and forced by pressure from a blowing agent and extending into and through the interstices of the fabric.” (Emphasis added.) The portions of the projections that extend through the interstices of the fabric are described as “being larger than said interstices to effect a key-locking engagement of the fabric and sponge rubber.” Coyte contains nothing which remotely resembles this.

The Coyte process is for the purpose of making gutta-percha cloth. The process involves the passing of a sheet of guttapercha with a cloth or fabric laid upon it between heated rolls, firmly pressed together, so that portions of the guttapercha sheet will “project into the meshes of the fabric” and “cause the guttapercha to be held to the fabric.”

It would appear to be obvious that this process could not possibly cause the gut-ta-percha to extend “through” the interstices of the fabric or to accomplish an attachment by a key-locking process.

If all that this record contained with respect to anticipation of the patent in suit were the two patents we have just discussed, namely, Blair and Bodle, and Coyte, it would be hard to say that there was sufficient evidence even to permit a submission of that question to a jury. With that alone in the record it could well be said that no valid challenge of the novelty of plaintiff’s device had been presented. But there is in the record another patent not alluded to by the court in its *734 findings or opinion which in our view does present a question of fact with respect to anticipation, although a question which we think it was proper for the jury to decide.

That patent is Alderfer No. 2,648,619. It discloses an undertaking by the inventor to produce a very different product having very different qualities than that contemplated by the patent in suit. That patent was for what was called a “fabric coated sponge rubber and method of making same.” The indicated objectives related solely to the procuring of desirable surfaces both on the rubber and on the fabric. The patent undertook to disclose a novel product where the “outer surface of the foam rubber is smooth, planished and porous having uniformly distributed microscopic pores.” For the fabric side of the product the object was to see to it that “the fabric be a knitted textile fabric, rather than woven, to increase the softness and enhance the appearance of the fabric side of the product.” As illustrations of the use of the product the inventor stated “automobile seat coverings foam rubber backed with finely knitted fabric presents a softer texture and a more pleasing appearance than can be obtained with a woven textile fabric. Also, foam rubber backed with knitted textile fabric such as tricot is desired in making certain products such as brassieres, because of the soft sheen and delicate appearance of the fabric.”

For his means of accomplishing these surface objectives the inventor introduced between the layer of the sponge rubber and the knitted fabric “an intermediate layer of square mesh fabric which was substantially non-elastic.” To put this product together he spread freshly foamed latex upon a smooth surface which traveled in one direction under a doctor roll or doctor bar which was adjusted to give the latex the desired depth or thickness. The two fabrics, the woven and the knitted, were then carried under the doctor bar so that the open mesh woven fabric was in direct contact with the top surface of the reduced layer of foapa latex and the knitted fabric was introduced on top of the woven fabric. The specification stated that by this process “the foamed latex will pass or ‘strike’ through the open meshes (of the woven fabric) into contact with the inner surface of the knitted fabric layer which also absorbs some of the water in the latex to cause the latex to firmly embed itself in the interstices of the knitted fabric and become securely anchored to the fabric layer.”

Plainly enough this discloses a product and a process where foam latex passed through a woven fabric and embedded itself in the interstices of the knitted fabric. It appears to us that the passing of heated sponge rubber through a woven fabric does not in itself anticipate a key locking device such as that of the patent in suit. In Alderfer the passing of the latex through the interstices is described, but it simply passes through and into interstices of knit fabric which is shown in the Alderfer drawings as being of substantial thickness as in all knit fabrics. There is no suggestion whatever of a bonding by key-lock as in the patent before us. 3

In describing his product the inventor of plaintiff’s device set forth how the key locking knobs were formed. When the uncured sponge rubber sheet comes from the rubber mill it is placed on a traveling mold and the net fabric is placed thereover. A web is placed over the fabric and another beneath the mold. This five-ply assembly is then moved slowly between heated platens, fixedly spaced without pressure. The heat causes the sponge rubber sheet to expand and since the rubber is held in intimate con *735 tact with the fabric, the expansion forces the rubber to enter the interstices and “swell around adjacent threads of said fabric to form headed knob-like projections that are somewhat larger in size than the areas of the fabric interstices.”

We are of the opinion that on the evidence here the jury were justified in finding as a fact that the product thus produced had a novelty not anticipated by anything shown in Alderfer. Its verdict should not have been set aside.

One thing that confirms us in this view is that the portion of the patent office file received in evidence discloses that Alderfer was cited in connection with the application for plaintiff’s patent and. that this problem of anticipation by Alderfer was considered and a resolution of the matter arrived at only after considerable controversy and an amendment of the claims originally presented. 4

As noted in S. H. Kress & Co. v. Aghnides, 4 Cir., 246 F.2d 718, the usual presumption of validity of a patent (Mumm v. Jacob E. Decker & Sons, 301 U.S. 168, 57 S.Ct. 675, 81 L.Ed. 983; Patterson-Ballagh Corp. v. Moss, 9 Cir., 201 F.2d 403, 406), may be strengthened by the history of the patent in the patent office, particularly when it has been granted only after considerable controversy.

This brings us to the question whether the new improvement disclosed in the patent in suit measures up to the established standards for invention. This is a question of law, Bergman v. Aluminum Lock Shingle Corp. of America, supra, which requires a discussion of whether the device here shown required more ingenuity than the work of a mechanic skilled in the art. Cuno Engineering Corp. v. Automatic Devices Corp., 314 U.S. 84, 90, 62 S.Ct. 37, 86 L.Ed. 58; Leishman v. General Motors Corp., 9 Cir., 191 F.2d 522, 530; 35 U.S. C. § 103.

The record here shows that leading up to the development of the product here in question was a long period of study, effort and experiment in an attempt to produce a practical means of attaching fabric to sponge rubber without the use of cement, which was the previously accepted method of attachment. The cementing process added approximately 10% to the cost of such a product. The device had commercial success ; beginning with gross sales of $350,000 in 1951 they increased in three years to $2,370,000. All of these things, though not controlling, are indicia of actual invention and properly considered in determining that question. It is of no significance that “viewed after the event, the means * * * adopted seem simple and such as should have been obvious to those who worked in the field, but this is not enough to negative invention.” Goodyear Tire & Rubber Co. v. Ray-O-Vac Co., 321 U.S. 275, 279, 64 S.Ct. 593, 594, 88 L.Ed. 721. “Now that it has succeeded, it may seem very plain to anyone that he could have done it as well. This is often the case with inventions of the greatest merit.” Carnegie Steel Co. v. Cambria Iron Co., 185 U.S. 403, 446, 22 S.Ct. 698, 715, 46 L.Ed. 968. See Patterson-Ballach Corp. v. Moss, supra.

*736 It is inferable from the facts in this case, including the rapid success which the new product found in the market, that this development for the first time, of a key-lock bond system for attaching the rubber to the fabric, was not only highly useful and a significant improvement over former products which utilized cement for attachment, but that it measured up to the established standards required for invention.

The trial court, without dealing in any specific manner with the question of the development of a key-locking engagement between the fabric and sponge rubber so as to bond the same, complained that plaintiff’s evidence too much emphasized the method and apparatus used in the construction of the product. It is true that the specifications include a description of the manner and process of making and using the invention; but the precise structure of the patent is also described with clarity and the drawings referred to in the specifications, particularly Figures 2 and 3, adequately disclose the nature of the claims. The addition to the specifications of a description of the manner in which the carpet underlay described may be made, merely adds to the other description but in no manner invalidates the patent. 5 Furthermore, the claims adequately and sufficiently covered and described the improvement regarded as the invention. 6

We hold that the patent was valid.

With respect to the question of infringement, little need be said. A mere examination of the exhibits disclosing the patented article produced pursuant to plaintiff’s patent and the accused device adequately demonstrates the claimed infringement. In setting aside the verdict the court below did not purport to find any fault with the jury’s finding of infringement. It would be impossible to do so.

The judgment is reversed and -the cause is remanded for further proceedings not inconsistent with this opinion.

Notes

. “1. A carpet underlay comprising a coarse-mesh fabric, and a sponge rubber facing on one side of said fabric and formed with bulges that are spaced from the fabric, said facing between the bulges, being in contact with the fabric and having a multiplicity of projections formed and forced by pressure from a blowing agent and extending into and through the interstices of the fabric.

“2. A carpet underlay according to claim 1: the portions of said projections that extend through the interstices of the fabric being larger than said interstices to effect a key-locking engagement of the fabric and the sponge rubber.

“3. A carpet underlay comprising a coarse-mesh fabric, a sponge rubber facing on one side of said fabric and formed with bulges that are spaced from the fabric, said facing, between the bulges, being in contact with the fabric, a multiplicity of integral projections on the bulges and extending through the interstices of the fabric to bond the bulges of the facing and the fabric, the latter being generally flat.

“4. In a carpet underlay according to claim 3: the projections being disposed along spaced parallel lines that are on a bias relative to the weave of the fabric.”

. Appellee suggests that since this patent contemplated the use of blowable cellular or sponge rubber, that the blowable sponge rubber would project into and through the open mesh fabric thereafter widening and forming the key-lock referred to in plaintiff’s claims. There is no evidence to support such a contention, and nothing in the .Blair-Bodle patent to suggest that any such thing was contemplated. On the contrary, the patent suggests quite the opposite. It states that the heated rubber “is soft enough so that it will gravitate into the channels of the shaping drum.” It suggests that a suction pull may be added to that of gravity. The quoted description refers to a pull by gravity or suction precisely opposite to that which is described in plaintiff’s patent. This gravitating into the channels of the shaping drum is a process to form the waffle-like ridges or projections, the push or pull being away from and not toward the fabric.

. Nothing in Alderfer suggests that the inventor intended or appreciated any key-locking development from the passage of the latex through the woven fabric. “It is well settled that prior accidental production of the same thing, where character and function arc not recognized, does not anticipate.” McKee v. Graton & Knight Co., 4 Cir., 87 F.2d 262, 264, and cases there cited. Walker on Patents, Deller’s Edition, § 57.

. An original claim “6” was rejected, it being stated by the Examiner “In Alderfer the fabric 16 is considered to have the interstices through which the rubber extends and fastens itself as claimed.” Claim 6 to which this referred read as follows: “In a carpet underlay-comprising a coarse-mesh fabric and a sponge rubber facing thereon, a multiplicity of headed projections integral with the facing and extending from the facing through the interstices of the fabric with the heads thereof on the side of tbe fabric opposite to the facing.” This claim was then amended to read as in Claim 3 of the patent, (footnote 1, supra,).

It will be noted that in the amended claim the product is described not merely as one involving projections through the fabric, but with one having “projections on the bulges” and extending through the interstices “to bond the bulges of the facing and the fabric”. So amended the claim No. 3 was allowed.

. Trussell Mfg. Co. v. Wilson-Jones Co., 2d Cir., 50 F.2d 1027, 1029: “But, if the article is new, it may be patented, and the claim will not be invalidated merely because it describes the article with some reference to the method of producing it.” See 85 U.S.C. § 112.

. In Winslow Engineering Co. v. Smith, 223 F.2d 438, this court, following Graver Tank & Mfg. Co. v. Linde Air Products Co., 336 U.S. 271, 69 S.Ct. 535, 93 L.Ed. 672, although it found invention, lield a patent invalid for insufficiency of the statements in the claims and that the deficiency could not be supplied by reference to the specifications. A similar point has not been raised in this case; but it is plain that the claims here are stated adequately to meet the requirements in the Graver Tank case, supra. Apparently in Bergman v. Aluminum Lock, supra, this court overlooked the rule of the Graver case and its own decision in the Winslow case.