Fred Whitaker Company v. E. T. Barwick Industries, Inc.

551 F.2d 622, 194 U.S.P.Q. (BNA) 113, 1977 U.S. App. LEXIS 13622

• Court: Court of Appeals for the Fifth Circuit
• Decided: April 28, 1977
• Docket: 74-3112
• Status: Published

Opinion

GODBOLD, Circuit Judge:

Plaintiff Fred Whitaker Co. (Whitaker) charges that E. T. Barwick Industries, Inc. (Barwick) has infringed the ex- elusive rights Whitaker owns under claims 4 1 and 10 2 of U.S. Patent No. 3,012,303. The patent in suit is a combination process patent. It describes a process of dyeing and treating yarn for use in carpets and upholstery fabrics. The district court held that the patent is invalid and, even if valid, uninfringed. We agree that the patent is invalid, and, for this reason, hold that the district court properly'dismissed the action.

I. Description of the patent

There are five basic elements which, in combination, comprise the process Whitaker claimed as its invention when it applied for a patent in December 1959.

1. Use of continuous filament yarn. In recent years carpet and upholstery manufacturers have had available to them two kinds of yarns, staple yarns and continuous filament yarns. The district court explained the difference between them: “Spun staple yarns are made from short fibers called staple fibers, for example cotton, wool and chopped nylon fibers. Continuous filament yarn is made from individual filaments which are formed in a continuous, unbroken length.” Whitaker sought patent rights for a general process (claim 1) and specific embodiments of it in both staple (claim 5) and continuous filament (claim 4) yarn. Continuous filament yarn is the only kind at issue in this case. 3

In this litigation Whitaker has defended the validity of its patent by arguing that use of its techniques on bulked continuous filament nylon yarn (known as “BCF yarn”) would not have been obvious in light of the prior art. Bulked continuous filament yarn is manufactured with its individual filaments “crimped,” creating air spaces between the filaments. The claims of the patent do not, however, limit the invention to bulked continuous filament yarn. Claim 4 simply refers to “the process of claim 1, in which the yarn is continuous filament yarn.” Nor do the specifications even imply that the process should be employed only with bulked yarn. Indeed, they suggest the contrary by noting that “[t]he invention may be applied to yarn previously texturized or bulked by some other process” (emphasis added). Much trial testimony was devoted to an explanation of how Whitaker restores bulk put in by the manufacturer, but the record supports the construction that the 1959 patent application intended to make use of both kinds of yarns. 4 Accordingly, we interpret claims 4 and 10 as describing a process for unbulked (or “feeder”), as well as bulked, continuous fiber yarn.

2. The knit-deknit process. Whitaker’s patent describes a process by which yarn is knit into an endless tube called a prefabric. The yarn is dyed in that form and then unwound again. The district court ex plained: “Use of the knit-deknit prefabric, ungainly though it seems to the uninitiated, so facilitates the handling of the yarn during the dyeing operations that it offsets the added cost of knitting and deknitting.”

3. Multicolor dyeing. In initial experimentation with the knit-deknit process, Whitaker encountered problems when it had dyed its yarn with a single color. Streaks appeared because of the uneven dye affinity of the yarn along its length. 5 In response to this difficulty, Whitaker began to dye its yarn by printing a multicolored striped pattern onto the prefabric. When unravelled, the yarn would display a variegated pattern. With many shades blending into each other, the undesirable streaking effect would be somewhat disguised.

4. Heat-setting. One of the approaches Whitaker describes (embodied in claim 10) 6 includes a heat-setting step. Before being unravelled, the yarn is heated in an autoclave 7 at 180° to 270° for a period of time ranging from one to 30 minutes.

Heat-setting appears to serve two purposes. When used after the multicoloring step, it can set the dye permanently and dry the yarn. Also, heat-setting imparts a curl to the yarn by fixing it in the position into which it is formed when knitted into the tubular prefabric. As seen immediately after the heating process, this curl (referred to also as a “kink”) does more harm than good with respect to the objective of imparting (or re-establishing) bulk. While the bundle of individual filaments (i. e., the yarn) becomes kinked within the bundle, the individual filaments also become more tightly bound together:

Q. And when the yarn is knitted or in a captured position then and it’s heat-set, how does that affect the yarn? A. Well, it closes the stitches together.

Q. Is that desirable or undesirable? A. It’s undesirable.

Q. Why is it undesirable? A. Well, you want to maintain the original bulk as it comes in from a producer or as you start with and, therefore, you don’t want to shrink it and lock those curls together.

Q. If it’s undesirable, why do you do it? A. Well, we can’t help from doing it because we set the color and it also sets the yarn at the same time. Simultaneously.

Q. Suppose you didn’t heat it enough to heat-set it, what would that do to the colors? A. Well, the colors would wash right off.

Q. Would that make a satisfactory salable yarn? A. Oh, no. No. 8

Additionally, the dye becomes concentrated in the apexes of the resulting kinks. 9 However, the subsequent opening up process, explained below, turns this relative disadvantage into an advantage. Opening up preserves the heat-setting curl but spreads apart the individual filaments and distributes their uneven dye concentration:

Q. You have also told us about the curl of the knit. A. Yes.

Q. Which is formed where the curl is formed into a loop. A. Yes.

Q. Was that curl desirable or undesirable? A. Well, I would say it was desirable. Once we reoriented it.

Whether it was more desirable than the original crimp of DuPont’s — that area of the loop — we didn’t take out all of the DuPont crimp but we must have taken some out. The tension took some of it out and we superimposed this curl on it.

Now when that’s fluffed open it must be quite desirable because everybody wanted it and they didn’t complaint about it.

Q. Was it desirable until you fluffed it open? A. No. We had to — oh, no. If you didn’t fluff it open that marring of the fibers, that group of fibers around the loop would stick together and you’d have to reorient them, you'd have to open them up to open the yarn up and bring the yarn back to approximately its original condition. 10

For our purposes, the important thing to understand about heat-setting is that, apart from setting the dye, it only has net worth after the yarn is subsequently opened up. As- noted below, this later spreading out of the filaments is what we understand to be the real objective of “opening up,” whether or not straightening out the curl occurs to some degree in the process.

5. Opening up. After the unravelling of the prefabric comes a process called “opening up,” or “fluffing.” The definition of “opening up” is perhaps the most hotly contested issue in this litigation, and we will say more about it later. For now, we note that the claims refer to it as a process which will “re-orient the fibers in the yarn and distribute the color.” This is accomplished either by running the yarn over a corner or series of corners, or by rubbing the yarn along its axis between rotating pulleys. The specifications explain that the goal is to “disarrange the loops so that the loops on individual continuous filaments or fibers will not be side by side.” Thus it appears that individual filaments are no longer pressed into a tight bundle but are spread apart from each other. The patent pictures this effect in two illustrations (Figures 9 and 9a), showing that the opening up process tends to widen the overall diameter of the yarn. It thus has the effect of restoring part of the bulk lost in heat-setting. A natural concomitant of this dispersal of the individual strands is the complementing of the multicoloring technique. The lines of demarcation between the stripes of color become softened or blended. “[W]hen this yarn is opened up,” the specifications say, “the color becomes distributed laterally and the zones of demarkation [sic ] of the several strands are graduated since there are different orientations of different strands laterally and the colors are even shifted longitudinally of the yarn from strand to strand.”

After the opening up step, the processed yarn is formed into a finished fabric by Whitaker or by its customers, like Barwick, who send their yarn to Whitaker to obtain the above service. The manner in which this is done does not appear to be relevant here.

II. Validity

The sole challenge to the validity of the Whitaker patent is on grounds of obviousness, as established in 35 U.S.C. § 103:

A patent may not be obtained . if the differences between the subject matter sought to be patented and the prior art are such that the subject matter as a whole would have been obvious at the time the invention was made to a person having ordinary skill in the art to which said subject matter pertains. Patentability shall not be negatived by the manner in which the invention was made. 11

Obviousness is a question of law that “necessarily entails several basic factual inquiries.” Sakraida v. Ag Pro, Inc., 425 U.S. 273, 280, 96 S.Ct. 1532, 1536, 47 L.Ed.2d 784, 790 (1976); Yoder Bros., Inc. v. California-Florida Plant Corp., 537 F.2d 1347, 1379 (CA 5, 1976), cert. denied,-U.S.-, 97 S.Ct. 1108, 51 L.Ed.2d 540 (1977). These factual inquiries are three-fold:

Under § 103, the scope and content of the prior art are to be determined; differences between the prior art and the claims at issue are to be ascertained; and the level of ordinary skill in the pertinent art resolved.

Graham v. John Deere Co., 383 U.S. 1, 17, 86 S.Ct. 684, 15 L.Ed.2d 545, 556 (1966); Sakraida, 425 U.S. at 280, 96 S.Ct. at 1536, 47 L.Ed.2d at 790. Our review of the district court’s findings on these three inquiries is limited by Rule 52(a), F.R.Civ.P. Parker v. Motorola, Inc., 524 F.2d 518, 531 (CA 5, 1975), cert. denied, 425 U.S. 975, 96 S.Ct. 2175, 48 L.Ed.2d 799 (1976). 12

Putting aside for a moment the more troublesome problem of opening up, we agree with the district court’s finding that Whitaker’s other steps were well known from the prior art. The history of Whitaker’s application to the Patent Office is instructive in this regard. The patent examiner rejected Whitaker’s initial application by referring to a 1958 article by Needle and Romer describing the knit-deknit operation. In response Whitaker pointed out that this article made no mention of the multicolor dyeing approach. But the examiner again rejected the application, observing that a 1933 British patent issued to Henry Giesler 13 had already instructed the public as to how these two techniques might be combined. In order to secure the patent Whitaker had to amend its application, stressing the advantages of opening up and rewriting the claims so that all claims would refer to an opening up step. Thus, in response to questions propounded by the court, Dr. Hamby, Whitaker’s expert wit ness, agreed at trial that the knitting, printing, deknitting, heat-setting, and winding of the processed yarn onto a cone were all known from the prior art. 14

. [4] However, Whitaker has claimed that the knit-deknit and multicoloring processes, while widely accepted in other contexts, had not previously been known as useful techniques for treatment of continuous filament yarn. 15 The patent examiner and the district court were justified in concluding that to experiment with these old processes on newly developed yarns would have been .obvious to the yarn industry in 1958, the year in which Whitaker began developing its process. Dr. Hamby conceded that by this time continuous filament yarn was a standard yarn. App. 276, 282-83. There is also evidence that at least two of Whitaker’s competitors, Pharr and Rossville, actually did try using continuous filament yarn in similar processes before learning about the Whitaker process and before the latter was patented in December 1961. App. 434-35, 444-460. Such contemporaneous independent development can be evidence of obviousness. Concrete Appliances Co. v. Gomery, 269 U.S. 177, 185, 46 S.Ct. 42, 45, 70 L.Ed. 222, 226 (1925); accord, Lerner v. Child Guidance Products, Inc., 547 F.2d 29, 31 (CA 2, 1976).

Finally, as early as 1943, the Imperial Chemical Industries Ltd. patent 16 had applied a crimping technique to continuous filament yarn. The Imperial patent is neither an application of the knit-deknit process nor an answer to the problem of dyeing continuous filament yarns. But Imperial does develop its technique for crimping, a means of imparting bulk, in order to make a “useful . . replacement for wool in carpets.” In view of the broad test of pertinence adopted by the Court in John Deere, 383 U.S. at 35, 86 S.Ct. at 702, 15 L.Ed.2d at 566, we think that the Imperial patent qualifies as prior art. See also Tapco Products Co. v. Van Mark Products Co., 446 F.2d 420, 427 (CA 6, 1971); Gerner v. Moog Industries, Inc., 383 F.2d 56, 59 (CA 8, 1967); cf. Ling-Temco-Vought, Inc. v. Kollsman Instrument Corp., 372 F.2d 263, 268 n. 1 (CA 2, 1967). 17 And we think it would be obvious for someone reasonably skilled in the carpet yarn handling business, seeking to use the knit-deknit process, to consider use of continuous filament yarn in that process after he had read of the Imperial patent’s imparting of kinks to such yarn. 18

Since every other aspect of Whitaker’s yarn processing was either anticipated by the prior art or obvious in light of it, it falls to the opening up step to save the patent. In order to assess what, if anything, the opening up step adds to the combination, we elaborate further on our understanding of what the patent meant by that term. 19 All that can be gleaned from the claims is that the patented process involves the step of “opening up the kinked yarn to re-orient the fibers in the yarn and distribute the color.” 20 Whitaker would have us look no further. It cites cases for the proposition that a patent’s specifications and examples cannot serve to limit the claims. E. g., Edward Valves, Inc. v. Cameron Iron Works, Inc., 286 F.2d 933, 944 (CA5), modified on other grounds, 289 F.2d 355 (CA5), cert. denied, 368 U.S. 833, 82 S.Ct. 55, 7 L.Ed.2d 34 (1961). Here, however, “reorientation” is a vague term, and various methods of treating yarn will “reorient” its fibers in different ways. Thus it is necessary to resort to the specifications to ascertain what the claims mean. Maclaren v. B-I-W Group, Inc., 535 F.2d 1367, 1372-73 (CA2), cert. denied, - U.S. -, 97 S.Ct. 531, 50 L.Ed.2d 612 (1976); Interdent Corp. v. United States, 531 F.2d 547, 550, 209 Ct.Cl. 201 (1976); Rosen v. Kahlenberg, 474 F.2d 858, 860 (CA 5, 1973). 21 -

In light of the patent’s discussion, which we have already cited in part I above, we can only interpret the “opening up” step to be a method of widening the yarn’s diameter by separating the individual filaments. It “causes the bulk of the yarn to increase greatly,” as stated in the specifications. The district court properly laid stress on this phrase. 22 This interpretation of the specifications is amplified by the testimony of those familiar with the patent. As Ralph Whitaker observed, for example:

Q. Then what happens to it when the yarn is fluffed or reoriented? A. Well, that is disbursed [sic] and becomes part of the bulk of the yarn. It is a superimpose^] bulk on top of the bulk of Du Pont or whosever yarn we were using.

If it is a bulk yarn, which it would be, a texture yarn, the curl is superimposed on top of their bulk, and when it is dephased it probably adds a little bit more bulk to the yarn.

Q. Actually what happens to the filaments when they are fluffed? A, What happens to them when they are fluffed? They open up and disburse [sic], that is what I would say..

Q. In what position are the filaments after they are knitted and dyed and printed? A. They are compacted and compressed in the curl, you see, and then when you reorient them they fall in their proper position, although they are carrying with them the curl of the knit, some of it, plus the crimp that was imposed.

Q. They are no longer immediately adjacent to each other? A. No, no, they are opened up. They have air spaces and they are fluffed open. 23

Thus, “opening up” refers to a result achieved by the particular process of running the yarn over an edge. 24

Some confusion has arisen from the fact that a few of the witnesses appeared to use the phrase “opening up” in close association with the phrase “dephasing,” which refers to the process of removing from the yarn curls that were imparted by the knit-deknit and heat-setting steps. 25 We think that the two terms are not synonymous. Such a construction would confuse a physical consequence of the Whitaker process with a necessary technique. Perhaps in running its yarn over an edge under tension after the yarn comes off the prefabric, Whitaker does eliminate some of the amplitude in the kinks that heat-setting has rigidified in the yarn. And perhaps when the yarn is being wound onto a cone after the edging process, it assumes a straighter physical condition. Whether these by-products exist is unimportant to our discussion here. It is the edging technique that “re-orient[s] the fibers in the yarn and distribute^] the color.”

Having described more precisely our understanding of the technique for and the results of opening up, we turn to an appraisal of the prior art concerning this particular step. The district court pointed to the Newton 26 and Imperial patents, among others, as demonstrating that opening up would have been obvious in light of the relevant prior art. These two patents discuss techniques for “edge crimping,” 27 a process of inducing curls into yarn by running it over an edge. As the district court found, it is based on a principle similar to that employed in curling ribbon over a scissors blade. We agree with the finding of the district court that the physical act of drawing yarn over a corner in the edge crimping process is identical to that employed in the Whitaker patent. 28 The results of the two processes are different. Edge crimping creates a different orientation of the crystallites 29 on the underside of the filaments. 30 Opening up simply spreads the various filaments apart. The only reason for this difference in physical results appears to be that in edge crimping one uses either a sharper edge, a greater tension, or both. 31 And an end product in both cases appears to be a bulking of the yarn. 32

Having examined these differences from the relevant but undisclosed 33 prior art, we look to whether the Whitaker process is made non-obvious by virtue of its opening up step. What we have here is a combination process patent. It is not exact-' ly a combination of old processes, which we should scrutinize . . with a care '

proportioned to the difficulty and improbability of finding invention in an assembly of old elements. ... A patent for a combination which only unites old elements with no change in their respective functions . . . obviously withdraws what already is known into the field of its monopoly and diminishes the resources available to skillful men.

Sakraida, 425 U.S. at 281, 96 S.Ct. at 1537, 47 L.Ed.2d at 790-91, quoting Great Atlantic & Pacific Tea Co. v. Supermarket Equipment Corp., 340 U.S. 147, 152-53, 71 S.Ct. 127, 130, 95 L.Ed. 162, 167 (1950). Rather, Whitaker’s combination uses the old edging technique to achieve a somewhat different result than that found in Imperial and Newton. Nevertheless, we agree with the district court that Whitaker’s idea is not a sufficient departure from these two as to constitute a non-obvious application of an identical technique. “It is not invention to use an old process for a new and analogous purpose.” 2A WALKER, PATENTS § 124, at 294 (A. Deller 1964 ed., Supp.1975). See also, e. g., Research Gorp. v. Nasco Industries, Inc., 501 F.2d 358, 360-61 (CA7), cert. denied, 419 U.S. 1096, 95 S.Ct. 689, 42 L.Ed.2d 688 (1974). 34 Simply because Whitaker uses the process to achieve a different result does not answer whether the difference is qualitative enough to be non-obvious. The goals of the two processes are similar enough to find Whitaker’s advance an obvious one. The Newton patent already had spoken of the possibilities for increased bulking. And the results are different only because of a difference in the degree of tension used. 35 With these two patents as prior art, it would have been obvious to one reasonably skilled in the yarn handling business that the edging technique, which “reoriented” the structure of the filaments, could be used to “reorient” the filaments vis-a-vis each other and therefore open up the yarn. Phrased another way, we think that the “skillful mechanic” would see fit to tinker slightly with the edging technique already in existence.

Because Whitaker’s use of the edging technique was obvious, its combination consists entirely of old or obvious elements. We find no synergistic interplay of these in the patented process, which is but a consecutive series of discrete and independent steps. Such steps here do not “result in an effect greater than the sum of the several effects taken separately.” Anderson’s-Black Rock, Inc. v. Pavement Salvage Co., 396 U.S. 57, 61, 90 S.Ct. 305, 308, 24 L.Ed.2d 258, 261 (1969). Claims 4 and 10 of the patent are invalid for obviousness. 36

The district court’s judgment is AFFIRMED.

1 . Claim 4 reads: “The process of claim 1, in which the yarn is continuous filament yarn.” Claim 1 reads: “The process of making a house furnishing fabric, which comprises producing a prefabric from face yarn, dyeing localized areas of the prefabric so that different lengths of the yarn will have different colors when the prefabric is unravelled, unravelling the prefabric, opening up the yarn to re-orient the fibers in the yam and distribute the color, and forming a fabric of the opened up yarn having the yarn and its colors exposed at the face of said fabric.”

2 . Claim 10 reads: “The process of making a house furnishing fabric, which comprises knitting a prefabric from face yam, dyeing localized areas of the prefabric so that different lengths of the yam will have different colors when the prefabric is unravelled and setting the prefabric to form kinks in the yarn, unravelling the prefabric, opening up the kinked yarn to re-orient the fibers in the yarn and distribute the color, and forming a pile on a backing by tufting said opened up yarn through said backing and exposing said yarn at the pile face.”

3 . In 1972 Whitaker wrote to the Patent Office disclaiming claim 5, which concerned use of staple yarn in the claim 1 process. Admittedly, the language of claim 10 does not confine itself to the use of any particular kind of yarn, and so staple yarn may still be included within the claims in litigation here. However, Whitaker tells us that staple yarns cannot be “opened up.” Opening up, a step we describe below, is the only technique that may possibly save the patent from obviousness. As a result, we need not consider the patent’s use as applied to staple yarns.

4 . See app. 238, 274, 275-276.

5 . Ralph Whitaker, the president of Whitaker, testified that the reason for this uneven dye affinity was uneven bulking. Whether bulked by the manufacturer at the manufacturing stage, or bulked by Whitaker prior to dyeing through use of a “stuffer box” or through the knit process itself, the yarn cannot maintain complete uniformity in bulk. This uneveness results in different light reflectancy, thus different color. App. 88-90. See also id. at 236 (testimony of Dr. Hamby, Whitaker’s expert).

6 . According to the specifications, heat-setting • is the “preferred embodiment” of the patent. We do not clearly understand how the claim 4 process sets the dye without heat-setting, but such an understanding is unnecessary to our resolution of the obviousness question.

7 . As described in the specifications and drawings, an autoclave is a chamber with a pressure-tight cover and a suitable support for the prefabric. It heats by using superheated steam introduced at the interior.

8 . App. 145 (testimony of James Painter, Vice-President and General Manager of Whitaker). See also id. at 154 (court’s characterization of this testimony).

9 . Id. at 93 (testimony of Ralph Whitaker).

10 . Id at 94-95. The language of the patent specifications confirms the idea that heat-setting holds out the advantage of fixing the curl in the yarn. The specifications note that heat-setting may precede the multicoloring step, from which statement one may infer that heat-setting can serve some purpose other than merely setting the dye. This idea is evident as well from the description contained in Whitaker’s April 10, 1961, amendment to the patent submitted to the Patent Office. “In those instances where the yam in the prefabric is set prior to unravelling, the kinks in the unravelled yarn also encounter the effects of opening up the yarn and have a worthwhile physical contribution to fluff and color distribution.” App. 50.

Cf. id at 276-77 (testimony of Dr. Hamby, Whitaker’s expert witness, on cross-examination):

“Q. Isn’t it apparent that the patent speaks of its purpose being to impart the curl of the knit to the yarn that is being treated?

A. Yes, and it further states another purpose is to open that up and to randomize the color.

“Q. But it says it wants to put that curl in there; it doesn’t say that the curl is a necessary evil, does it? A. The feeder yam is certainly where you want to put the curls. If you could process the BCF yarn without the curl, you certainly wouldn’t object to doing that.

“Q. Where does it say that in the patent? A. It doesn't say, but I think it is, again, implicit in the understanding of the patent.

“Q. But the patent nowhere says that it is desirable to avoid imparting this knitted curl; on the contrary, it repeatedly said that it is desirable to impart the knitted curl; isn’t that so? A. Yes, but again I say that an executive of any caliber who would pay a manufacturer to get a yarn of this type and then destroy it would not—

“Q. We are talking about the disclosure of the patent, not some executive, and the disclosure of the patent repeatedly talks about imparting the curl to the knitted yarn, does it not? A. It does.

“Q. And it states that that is a desirable aspect of the invention, does it not? A. It states that, yes.”

11 . This standard is a response to the constitutional imperative that “Congress shall have power ... To promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries,” U.S.Const. art. I, § 8, cl. 8. Sakraida v. Ag Pro, Inc., 425 U.S. 273, 279, 96 S.Ct. 1532, 1536, 47 L.Ed.2d 784, 789 (1976).

12 . In this case the district court adopted verbatim the proposed findings of fact and conclusions of law submitted by Barwick (except that the court omitted Barwick’s contentions regarding alleged patent misuse by Whitaker). While this court has often expressed its disapproval of this procedure, it does not eliminate the “clearly erroneous” scope of review. See, e. g., Keystone Plastics, Inc. v. C & P Plastics, Inc., 506 F.2d 960, 962-63 (CA 5 1975).

13 . British Patent No. 396,219.

14 . App. 299-300.

15 . Which, after Whitaker’s 1972 disclaimer, was arguably the only type of yarn contemplated under its patent. See note 3 supra.

16 . British Patent No. 558,297.

17 . We reject Whitaker’s argument that the Giesler patent was the “best” prior art and therefore that looking to other art such as the Imperial patent is subject to criticism, Eli Lilly & Co. v. Generix Drug Sales, Inc., 460 F.2d 1096, 1103 (CA 5 1972); cf. United States Gypsum Co. v. National Gypsum Co., 440 F.2d 510, 514 (CA7), cert. denied, 403 U.S. 923, 91 S.Ct. 2233, 29 L.Ed.2d 702 (1971). Whitaker’s patent is really a combination of several techniques. To say that it approaches one of these (Giesler) more than the others is hard to accept, especially when Whitaker premises its difference from Giesler on having added another step (opening up) which the Imperial patent most closely resembles.

18 . Whitaker contends that other yam handlers would have been surprised to learn that bulked continuous filament yarn could benefit from a knit-deknit treatment, since such treatment would tend to destroy the very bulk that is expensively introduced into such yarn by its manufacturer. This contention cannot stand because, as already discussed, we do not construe the patent as limited to pre-bulked yarn. As a result, Whitaker cannot lay claim to the protection of Ingersoll-Rand Co. v. Brunner & Lay, Inc., 474 F.2d 491 (CA5), cert. denied, 414 U.S. 865, 94 S.Ct. 125, 38 L.Ed.2d 117 (1973). At any rate, Pharr and Rossville in fact limited their knit-deknit operations to' the use of bulked continuous filament yarn.

19 . We have consistently treated the construction of a patent as a question of law. Ziegler v. Phillips Petroleum Co., 483 F.2d 858, 867 (CA5), cert. denied, 414 U.S. 1079, 94 S.Ct. 597, 38 L.Ed.2d 485 (1973).

20 . This is the language of claim 10. Claim 1, the language incorporated in claim 4, omits the word “kinked” because it does not include the heat-setting step.

21 . Resort to expert testimony in order better to understand the specifications is also well established. Kohn v. Eimer, 265 F. 900, 902 (CA2 1920) (L. Hand, J.); accord, Duplan Corp. v. Deering Milliken, Inc., 370 F.Supp. 769, 786-87 (D.S.C.1973). See also Universal Athletic Sales Co. v. American Gym, Recreational & Athletic Equip. Corp., 546 F.2d 530, 541 — 42 & n. 39 (CA3 1976).

22 . The district court also incorporated the following statement from the specifications into its definition: “Whether a corner or a rubbing motion is used, the effect is to open up the yarn to at least three times the initial diameter.” An identical statement appears in the filewrapper in Whitaker’s letter to the Patent Office emphasizing the opening up step as the reason why its process was a patentable advance over the prior art. App. 27. It is unclear whether the doctrine of filewrapper estoppel extends to such representations. Compare Williams Bit & Tool Co. v. Christensen Diamond Prods. Co., 399 F.2d 628, 634 (CA 5 1968), with Parker v. Motorola, Inc., 524 F.2d 518, 532-33 (CA 5 1975); see Duplan Corp. v. Deering Milliken, Inc., 379 F.Supp. 388, 400-403 (D.S.C.1974). Resolving this issue is unnecessary to our basic understanding of the patent.

23 . App. 97.

24 . As noted above, the patent also describes another technique for opening up, rubbing the yarn transversely to its axis. The parties have not developed, or even discussed, this alternative technique. From the record, unaided by any discussion, the existence of this technique does not change our conclusion on the obviousness issue.

25 . See App. 353 (Frank Whiting, one of the patentees); cf. id. at 288 (Dr. Hamby).

26 . U.S. Patent No. 3,024,518 (continuation of application of June 17, 1955). Although the Newton application was filed almost eleven months after Whitaker’s, no one here questions its status as “prior.”

27 . Frank Whiting, one of the patentees, also testified that edge crimping was known in the art in 1958. A 1956 article by N. C. Armitage in Modern Textiles Magazine reinforces this conclusion.

28 . We attach no significance, however, to the fact, noted by the district court, that the Imperial patent draws a corner similar in angle to that depicted in the Whitaker patent. We cannot limit the claims of the patent to the particular embodiments of the drawings, specifications, or examples.

29 . According to Webster’s Third New International Dictionary 549 (1967), the crystallites are the structural units in textile fibers.

30 . Newton patent, col. 2, lines 5-7; see app. 292 (“altering of the cross-section of the filament”). See also Imperial patent, p. 2, lines 92 — 100; app. 298.

31 . See app. 293-98. Whitaker’s patent does not specify the degree of tension required for opening up.

32 . The Newton patent notes that inducing crimps helps to bulk the yam:

“The pile yams are then subjected to a crimp or coil developing treatment, as by wet or dry heating of the fabric to a suitable temperature, in order to develop in situ in the pile per se the latent crimp in the pile yarns, which takes the form of randomly reversed generally out-of-phase coils in the filaments. This effects a general shortening of the effective length of the pile yarns and a substantial bulking of the individual pile formations as a result of the patent development of the latent coil formations in the individual yarn filaments and the consequent self-pushing apart of these filaments.”

Newton patent, col. 1, lines 24-35. Although the district court asserted that the older Imperi al patent made a similar observation, we cannot find such a suggestion in that patent.

33 . Because this relevant prior art was not disclosed to the patent examiner, Barwick’s statutory burden of proof regarding invalidity, see 35 U.S.C. § 282, is eased. Beckman Instruments, Inc. v. Chemtronics, Inc., 439 F.2d 1369, 1374-75 (CA5), cert. denied, 400 U.S. 956, 91 S.Ct. 353, 27 L.Ed.2d 264 (1970); Cf. White v. Mar-bel, Inc., 509 F.2d 287, 291 (CA5 1975).

34 . Perhaps the question whether Whitaker’s use of the technique of running yarn over an edge is analogous to the old uses of that technique in crimping yarn really restates the obviousness question in more precise terms. Research Corp., supra, at 360.

35 . Whitaker does its crimping at an earlier stage. But the fact that crimping is involved in both, albeit in different stages, does not negate obviousness.

36 . Having held the patent invalid, we need not examine the district court’s holding of noninfringement. Beckman Instruments, Inc. v. Chemtronics, Inc., 439 F.2d 1369, 1371-72 & n. 4 (CA5), cert. denied, 400 U.S. 956, 91 S.Ct. 353, 27 L.Ed.2d 264 (1970).