Karl Isac Joel Rosen, the inventor, and AB-Iro, the co-owner of the patent in suit (hereinafter we will refer to the plaintiffs collectively as "Iro"), instituted this action to enjoin Lawson-Hemphill, Inc. (Lawson) from infringing U.S. Patent No. 3,648,939 (Rosen patent) and to obtain an accounting. The district court found that the Rosen patent was valid and had been infringed, and it granted the requested relief, staying the questions of an accounting and of attorney fees pending appeal. 399 F. Supp. 532 (D.R.I.1975). Lawson now appeals, alleging that the district court committed reversible error in making the determinations that gave rise to the finding of liability. We affirm. In considering the issues raised on appeal, we will frequently simply refer to the careful, well reasoned opinion of District Judge Pettine. Many of the arguments presented to this court are identical to those presented below, and where we agree with the district court's analysis, we will simply say so.

Before discussing the facts and the legal issues, there is a preliminary matter that requires discussion. A recurring theme in Lawson's appellate brief and in its prosecution of this appeal generally is that the voluminous record that was prepared below is somehow not adequate to permit the correct resolution of the issues raised in this case. Lawson now attempts to supplement that record by making factual representations as to matters not in the record. We, of course, must disregard all such representations. It is elementary that the rights of the parties are, for purposes of this appeal, determined on the basis of the materials that constitute the record on appeal. See Fed.R.App.P. 10(a). Although nothing further need be said, we note that Lawson had every opportunity below to test and challenge Iro's evidence and to present whatever evidence it thought relevant to the issues at hand.

The Rosen patent, which was granted in March, 1972, is for a "yarn storing device" that ensures that the yarn that is fed to a knitting machine has a low and uniform tension. The production of such yarn tension is critical in the fabric knitting industry. Knitting machines employ a plurality of needles to form loops and stitches of yarn, and the quality of the fabric produced is greatly affected by the uniformity of these loops or stitches. A knitting machine will knit fabrics containing loops and stitches of the desired length only if the yarn tension is very low and constant. When the tension is not low, the needles will "rob" yarn back from previously formed loops and stitches. When the yarn tension is not constant, each loop or stitch will not have substantially the same length. To control yarn tension in knitting machines, the industry developed automatic yarn feeders which, as technology improved, developed into devices which stored the yarn at a station between the yarn packages and the machine and these achieved the proper tension control for plain knitting. But these proved inadequate for pattern knitting, which is done on circular knitting machines and in which the knitting needles at the different knitting stations have varying or intermittent demands for yarn.

The Rosen invention essentially solves the problem of achieving low, constant yarn tension on the knitting machines that are used to knit patterned fabrics. It contains eleven claims, eight of which were allegedly infringed by Lawson's device. The first claim is independent, and it is the focal point of this suit. It reads:

"The embodiments of the invention in which an exclusive property or privilege is claimed are defined as follows:

1. In a yarn storing device having a storing drum adapted to have yarn wound tangentially thereon and a retarding ring surrounding the drum and adapted to have the yarn passed thereunder for removal axially from the drum, comprising the improvement wherein said retarding ring includes a base ring surrounding the periphery of the drum and spaced a small distance therefrom and a plurality of thin elongated resilient fingers extending from the base ring toward the surface of the drum, the resilient fingers being inclined inwardly from the base ring along an imaginary conical surface, said resilient fingers extending in the direction of relative rotation of the yarn during its withdrawal from the drum and disposed so as to overlap one another in the direction of the periphery of the drum, and said drum having a shoulder thereon disposed in engagement with the free ends of said resilient fingers."

Figure 1 of the patent, which is reproduced as an appendix to this opinion, illustrates the various embodiments of the invention. We cannot improve upon the district court's description of the manner in which the preferred embodiment of the Lawson invention works:

"Referring to figure 1, the yarn (3) runs off the bobbin (2), through the guide elements (4, 5, 6, 7), onto the storing drum (8); it is wrapped around the storing drum a number of turns and then removed therefrom by coming down from the storing drum under the retarding ring (10) (11) (12); as a consequence the yarn as it is withdrawn is subjected to the tension of a series of fingers. There is a holding and releasing action as the yarn goes from one finger to the other. The next finger tends to bend slightly sideways. Tension is imparted to the yarn as it passes between the free end of the finger and the drum surface. The amount of sideways bending of the finger is a function of the amount of friction created. Any irregularity in the yarn itself that tends to increase the tension is thus compensated. This tension can be predetermined. The point is, the fingers having a circumference less than that of the drum create a tension on the drum holding the ring on it, and since the base of the ring (as set on the drum it is the upper part of this one-piece construction) has a greater circumference than the drum, the effect is a self-adjusting activity. The fingers respond to the tug of the yarn as it passes under each little tip. These fingers are inclined so they overlap each other in the direction of the removal of the yarn. (For example, the fingers of a comb (teeth) are at right angles from the base and run parallel to each other. If these same fingers were slanted from their base they would overlap each would overlap the vertical area occupied by its neighbor, and if this same comb was bent to form a circle it would, in a crude way, resemble the retarding ring). The motor driven storing drum pulls the yarn from the bobbin, but from the storing drum to the knitting needles the knitting machine does the pulling. This action is by design so free from tension that a tension imposing device is added so that the yarn can come off as needed at the proper predetermined tension and at the rate the knitting function demands." 399 F. Supp. at 536 (footnote omitted).

Iro's commercial embodiment of the invention operates in precisely this manner. The alleged infringing device of Lawson differs in no significant respect from Iro's except that the Lawson device omits the shoulder (13) and instead uses a supporting ring.

The key witness below was a Dr. Harry K. Wignall, who appeared as Iro's expert and whom Lawson stipulated to be a "well-known knitting expert". Dr. Wignall had conducted elaborate tests on the Rosen device, the Lawson device, and devices based upon the prior art, and these tests were introduced into evidence without objection.1 The tests revealed that only the yarn storage devices which employed the Rosen retarding ring produced the desired yarn tension and that there was no difference of any significance between the Lawson and Rosen devices. At trial Dr. Wignall explained his tests and then offered his expert opinion as to how the Rosen device worked and why it produced results superior to the prior art.

Relying upon Dr. Wignall's testimony, the district court found that the Rosen shoulder had no effect on the yarn tensioning and was simply a device to keep the retarding ring from being drawn off the storage drum. Since the Lawson supporting ring was found to be simply an alternative means of holding the retarding ring on the drum, the district court found infringement, applying the doctrine of equivalents. 399 F. Supp. at 550-52. The district court also rejected Lawson's claim that the file wrapper estopped Iro from relying upon the doctrine of equivalents. Id. at 553. Lawson attacks these findings on appeal, raising what appear to be the same arguments that were rejected below. We have considered the record and are satisfied that the district court was correct in finding that, if valid, the Rosen patent was infringed.

The substantial issue on appeal concerns the validity of the Rosen patent. Below Lawson maintained that the patent in suit was invalid, arguing that it was simply a combination of old elements and, alternatively, that the one arguable contribution to the art the Rosen retarding ring itself was an obvious improvement over the prior art. The district court rejected both arguments. Lawson now maintains that the district court committed fundamental errors of law in evaluating the patent in suit and that, properly evaluated, it is invalid. Even taking Rosen's device on the district court's terms, Lawson maintains the district court committed reversible error.

In attacking the district court's conclusion that the patent is invalid, Lawson does not quarrel in any meaningful fashion with the district court's summation of the general principles governing patentability of a combination like the one in suit.2 See 399 F. Supp. at 538-40. Lawson does not dispute the proposition that, notwithstanding the fact that combinations of retarding rings and storage drums are quite old, the Rosen combination is patentable if it is based upon an inventive improvement in one of the elements that permits the combination device to produce a beneficial result never previously obtained. See Carnegie Steel Co. v. Cambria Iron Co., 185 U.S. 403, 445-46, 22 S. Ct. 698, 46 L. Ed. 968 (1902); cf. Sakraida v. Ag Pro Inc., 425 U.S. 273 at 281, 96 S. Ct. 1532, 47 L. Ed. 2d 784 (1976). The record amply supports the district court's finding that the embodiments of the Rosen device produced a low, constant yarn tension which was not obtainable by combining the prior art retarding rings and a storage drum.3 Lawson does not now contend that the Rosen device is neither useful under 35 U.S.C. § 101 nor novel under § 102. The principal issue on appeal is whether the district court committed reversible error in finding that the development of the Rosen retarding ring for use in combination with a yarn storage drum would not have been obvious to one with ordinary skill in the relevant art. See id. § 103.

The question of obviousness is determined on the basis of several basic factual inquiries: "the scope and content of the prior art are (to be) determined; the differences between the prior art and the claims at issue are (to be) ascertained; and the level of skill in the pertinent art is (to be) resolved." Graham v. John Deere Co., 383 U.S. 1, 17, 86 S. Ct. 684, 694, 15 L. Ed. 2d 545 (1966). In this circuit, the question of obviousness vel non is essentially one of fact, see ITT v. Raychem Corp., 538 F.2d 453, 456 (1st Cir. 1976) and cases cited; cf. Spound v. Mohasco Industries, Inc., 534 F.2d 404 (1st Cir. 1976), and we will not reverse a district court's applications of the correct legal principles unless we believe them clearly erroneous. See Forbro Design Corp. v. Raytheon Co., 532 F.2d 758, 763 (1st Cir. 1976).

In finding the patent in suit nonobvious, the district court relied principally upon the "finger action" of the Rosen device and the manner in which the fingers, the yarn, and the storage drum interacted to ensure a low, constant yarn tension. The district court derived much of its understanding of the subtle operations of the Rosen device and how it compared to the relevant prior art from Dr. Wignall. Dr. Wignall identified several tensioning factors which arose from the combined actions of the Rosen retarding ring and the storage drum. The first of these occurs sometime before the yarn actually reaches the ends of the fingers. Inside the tensioning ring, the resilient fingers form a type of a funnel. As the yarn, which is "ballooning" due to the centrifugal force, approaches the finger tips, there is increasing contact between the yarn and the fingers and the result is a slight drag on the yarn before the finger tips are engaged.4 A second, more significant, effect occurs when the yarn reaches the finger tip. As the yarn passes beneath the ends of the resilient elements, each finger, acting individually and independently, one after another, clamps or "nips" the yarn slightly, lifting upwards slightly as the yarn passes. When the yarn leaves the extremity of each resilient finger, the finger "snubs" the yarn, causing the yarn to coil around the base of the finger and a more severe drag on the yarn. Thirdly, as the yarn comes off the fingertips, the tips cause the yarn to vibrate, reducing the area of contact between the yarn and any succeeding object.

The most significant of the tensioning effects is the fourth. The greatest threat to the uniformity of tension arises when there is a knot or snub in the yarn. The danger is that the retarding ring will impose greater tension on such portions of the yarn, destroying the needed uniformity. On the Rosen ring, a tug on the yarn will lift the finger up slightly and move it sideways, reducing the angle of contact between finger and yarn, causing the imposition of less tension on that portion of the yarn, and compensating for the defect in the yarn.

The successful operation of these tensioning factors arises in no small part because the ring is self-centering that is, the ring is not held rigidly, but floats, the fingers controlling the horizontal position of the ring. The result is that every individual fingertip on the ring has the same pressure on the drum, and each one is independent. Each segment of the yarn possibly every millimeter is treated independently.

There were two principal sets of prior art references before the district court: the Gift patent, which was a device for the prevention of the snarling of twisted yarn as it is withdrawn from a yarn supply package, and a series of patented retarding ring-storage drum combinations which employed bristles or bristle like substances to impart tension on yarn the Pfarrwaller, Mulhauser, Noton patents and the Sulzer retarding ring, and the Savi ring. The district court relied upon the aforementioned effects of the Rosen device to support its conclusion that the patent in suit was not obvious. Lawson attacks these findings on two grounds.

Lawson's most substantial argument is that the comparison of the patent in suit with the prior art was infected with legal error because the district court did not confine itself to the patent which was granted in evaluating the Rosen device. First, Lawson submits that the details of construction which are contained in the Rosen device which was tested by Dr. Wignall and which are required to produce the finger action attributed to the patent in suit are not disclosed in claim 1 of the patent and may not be relied upon to support a finding of nonobviousness. Appellant is, of course, correct that the claims of the patent, which disclose the physical structure of what has been invented, "measure the invention", see General Electric Co. v. Wabash Co., 304 U.S. 364, 369, 58 S. Ct. 899, 82 L. Ed. 1402 (1939), cf. 35 U.S.C. § 112, and that the district court was obliged to compare the prior art with the invention that Rosen claimed, not with its commercial embodiment. By the same token, it is accepted that, although the claims limit the invention and the specifications cannot be utilized to expand the patent monopoly, "the claims are to be construed in light of the specifications and both are to be read with a view toward ascertaining the invention." United States v. Adams,383 U.S. 39, 48-49, 86 S. Ct. 708, 713, 15 L. Ed. 2d 572 (1966) and cases cited. A court is not required to conceptualize the claims of a patent in the abstract. Here, we strongly doubt that claim 1 did not disclose a structure capable of producing the tensioning effects attributed to the device. To the extent the claims themselves are ambiguous, we think the specifications and drawings leave no doubt as to how the claims should be interpreted.

Lawson's first argument regarding the adequacy of the disclosure in claim 1 is that it does not describe the specific finger direction required to produce the various tensioning effects. This seems meritless. Claim 1 calls for (1) a retarding ring surrounding the drum and spaced a small distance from the drum, it being clear from the specifications and drawings that the ring is perpendicular to the drum axis, (2) a plurality of thin elongated resilient fingers extending from the base ring towards the surface of the drum, that is, inwardly toward the drum surface, (3) the fingers extending in the direction of relative rotation of the yarn during its withdrawal from the drum and disposed to overlap one another in the direction of the periphery of the drum, and (4) the fingers inclined inwardly from the ring alone an imaginary conical structure, it being clear, given the supporting materials, that each finger curves inwardly from the base ring to its tip engagement with the drum. We find no support whatsoever in the record for the proposition that the specific finger direction here described will not produce the required tensioning effects. But even if these disclosures were insufficient in some technical respect, we would think the specifications' careful description of the finger action and finger construction, see infra, would cure any ambiguity.

In a similar vein, Lawson argues that some of the features of the Rosen device which permit it to produce the low, uniform yarn tension the self-centering character of the ring and several of Dr. Wignall's tensioning factors were not disclosed in the patent specifications and could not be relied upon to support a finding of nonobviousness. It is not clear to us precisely what Lawson is arguing. If it is submitting that a district court must confine itself to the patent's disclosures regarding the scientific principles at work when it sets out to compare the patent in suit with the prior art, then its argument is meritless. The extent to which a patentee who, like Rosen, adequately disclosed the beneficial use of his device and the structure that is required to produce it5 has apprehended the scientific principles at work in his invention does not affect his ability to demonstrate the nonobviousness of the structure that has been claimed. As the Court has stated:

"A patentee may be baldly empirical, seeing nothing beyond his experiments and the result; yet if he has added a new and valuable article to the world's utilities, he is entitled to the rank and protection of the inventor. And how can it take from his merit that he may not know all of the forces which he has brought into operation? It is certainly not necessary that he understand or be able to state the scientific principles underlying his invention, and it is immaterial whether he can stand a successful examination as to the speculative ideas involved." Diamond Rubber Co. v. Consolidated Rubber Tire Co., 220 U.S. 428, 435-36, 31 S. Ct. 444, 447, 55 L. Ed. 527 (1911). Accord Reese v. Elkhart Welding & Boiler Works, 447 F.2d 517 (7th Cir. 1971).

On the other hand, we recognize that Lawson's argument may simply be a gloss on Lawson's earlier submission that the structure Rosen claimed will not produce a low, constant yarn tension. Lawson may be suggesting that the claims, when considered alone, do not sufficiently describe the invention and that, because the specifications do not describe all the tensioning factors and the self-centering character of the ring, the claims and the specifications read together do not teach the world how to construct a device which will produce this new, highly beneficial result. We reject this argument as well. Assuming arguendo that the claims considered alone are not sufficiently detailed, the specifications supply whatever additional information is needed. There is no doubt that a device constructed in view of the specifications would be self-centering.6 The specifications, moreover, contain a detailed description of the manner in which the fingers operate.7

Having found no error in the district court's assessment of the Rosen patent, we now turn to consider Lawson's further contention that the district court, in any case, committed reversible error in finding that the Rosen device was nonobvious as against the prior art. We find no clear error. To the extent that Lawson relies upon the brush and brushlike devices alone, we cannot say that a clear mistake was made when the district court found that Rosen's improvement over the brush art was sufficient to entitle him to the status of an inventor. There is ample support in the record for the district court's conclusion that the "bristle action", which consisted of the production of continuous pressure, was a far cry from the independent, individual nip and tuck of the Rosen fingers. See 399 F. Supp. at 541-44.

The prior art upon which Lawson relies most heavily now is the Gift patent (No. 2,838,933). It argues that Gift contains all the tensioning factors attributed to the Rosen ring and that the use of a retarding ring which employed these tensioning factors was obvious. At least on this record, there is no support for appellant's submission. Gift discloses a device to prevent twisted yarn from snarling as it is withdrawn from a supply package; it is not designed to control yarn tension. It consists of a circular rubber disc with a small central aperture and annularly spaced long slits and short slits which define a plurality of inwardly extending tongues. The rubber tongues are not thin but rather taper from relatively wide bases to pointed tips. When the disc is impaled downward over a yarn package, the tongues flex upwardly into frictional engagement with the surface of the yarn. As the yarn is withdrawn from the package, the yarn is controlled against snarling by the progressive lengthwise wiping action of the tongues. As the yarn package becomes smaller and smaller, the contact pressure between the package surface and the tongues diminishes progressively.

We find no support in the record for Lawson's assertion that the Gift device contains the tensioning factors of the Rosen ring. All the "retarding" action of Gift results from the progressive wiping action of the sides of the tongues, not by action of the ends or tips of the fingers. We frankly do not see how the subtle tensioning factors of the Rosen ring can be said to be at work in Gift, and certainly we cannot say the district court clearly erred in concluding they were not. We also find no error in the district court's other reasons for distinguishing Gift. See 399 F. Supp. at 541.

Finally, Lawson suggests that, regardless of what the prior art disclosed individually, an examination of the prior art brush retarding rings together with Gift compels the conclusion that the district court was clearly mistaken in holding that the Rosen device was not "obvious to try". See Mandel Bros. v. Wallace, 335 U.S. 291, 69 S. Ct. 73, 93 L. Ed. 12 (1948); Contour Saws, Inc. v. L. S. Starrett Co., 428 F.2d 314, 319 (1st Cir. 1970) (subsequent history omitted). There is no question but that the Rosen device's finger size, shape, positioning and ring arrangement are collectively new and that the finger action was a far cry from anything which had been seen in the prior art. Putting aside "infallible hindsight" and noting that Dr. Wignall testified that he would not have believed it obvious to try the Rosen device, we see not basis for concluding that the district court clearly erred in rejecting Lawson's claim. We note, moreover, that if we perceived this as a close question on appeal, which we do not, the district court's findings regarding long felt but unsolved needs and Iro's commercial success, each of which is amply supported by the evidence, would tip the scales. See Graham v. John Deere Co., supra, 383 U.S. at 17, 86 S. Ct. 684.

Affirmed.

NOTE: OPINION CONTAINS TABLE OR OTHER DATA THAT IS NOT VIEWABLE