Kimberly-Clark Corporation v. Johnson & Johnson and Personal Products Company

745 F.2d 1437, 223 U.S.P.Q. (BNA) 603, 1984 U.S. App. LEXIS 15202

• Court: Court of Appeals for the Federal Circuit
• Decided: October 9, 1984
• Docket: Appeal 83-1066
• Status: Published

Opinions

RICH, Circuit Judge.

This appeal is from the February 4, 1983, March 15, 1983 (219 USPQ 214), and April 5, 1983, 573 F.Supp. 1179 (219 USPQ 217), judgments of the United States District Court for the Northern District of Illinois, Eastern Division, sitting without a jury, holding that Kimberly-Clark Corporation’s Roeder patent No. 3,672,371 (’371) issued June 27, 1972, for “Sanitary Napkin with Improved Adhesive Fastening Means” was not infringed, “unenforceable” because of “fraud on the PTO,” and invalid under 35 U.S.C. § 103. We affirm the holding of non-infringement, reverse the holdings of obviousness and fraud, and remand.

Background

Kimberly-Clark Corporation (K-C), which manufactures and sells MAXI-PADS under its KOTEX and NEW FREEDOM marks, sued Johnson & Johnson (J & J) and its wholly owned subsidiary, Personal Products Company (PPC), which manufactures and sells similar products under its STAY-FREE and SURE & NATURAL marks. The patent in suit discloses a sanitary napkin having pressure-sensitive adhesive strips to secure it temporarily to a supporting undergarment, e.g., panties (see Fig. 3, below, at 19).

The preferred embodiment of the invention is shown in Fig. 1 of the patent, reproduced below. The top side in the drawing would be the underside in use, positioned as in Fig. 3.

The ’371 patent has one independent claim. It reads (bracketed numbers from Fig. 1 inserted, emphasis ours):

1. In an elongate sanitary napkin structure [13] provided with means for attaching said napkin to a supporting garment in which said means comprises pressure sensitive adhesive disposed on the bottom surface of said napkin [15, 16] and covered by a removable protective sheet [17], and said napkin comprises a pad of absorbent material enclosed in a fluid pervious non-woven wrapper [11] comprising a substantially rectangular sheet enveloping said pad and overlapped on the bottom side thereof [12 to 12a], the improvement wherein said adhesive comprises at least two narrow lines of adhesive parallely spaced from each other and extending longitudinally of said bottom surface [15, 16], said spaced lines being centrally disposed thereon with respect to the sides and ends thereof, and said lines of adhesive penetrate both of said overlapped portions of said wrapper.

Fig. la, below, is a cross-sectional diagram taken on the line . la-la of Fig. 1, supra:

It serves to illuminate the district court’s finding of fact number 12, in which we have inserted the reference numerals of Fig. la, wherein the court set forth its view of the critical claim limitation. It reads:

12. I find that the Roeder patent covers a sanitary napkin enclosed in a non-woven wrapper [11] in which all of the multiple lines of pressure sensitive adhesive [15, 16] must penetrate through the overlap of the non-woven wrapper [12 to 12a]. I also find that each of these multiple lines of .adhesive [15, 16] must perform the two operations — attaching and sealing the non-woven wrapper — with a single application of adhesive.

We agree with this construction of the ’371 patent claims as calling for a single application of adhesive in at least two parallel lines serving the dual function of penetrating and sealing the overlap and leaving enough adhesive on the surface to provide attachment to a garment.

On February 4, 1983, the district court ruled orally that the ’371 patent was not infringed by J & J’s competing product, a schematic cross-sectional diagram of which is here reproduced. It is to be understood, of course, that this is an exploded view and that in reality the layers are in contact.

J & J’s product consists of absorbent fluff [A] partially surrounded on the bottom and sides by a sheet-plastic “poly baffle” [B], held in place by tissue wrap [C] during the cutting stage of manufacture. All these components [A, B, and C] are enclosed in a non-woven wrapper [D] which is sealed by one adhesive strip [F] represented by plural parallel lines. Components [A-D] are made in very long strips which are cut up into pads. The diagram, submitted by K-C, shows another adhesive strip [G] sealing the tissue wrap [C] within the circle [H]. J & J argues that penetration of the overlap [C-Ca] by the other adhesive strip [G] is neither necessary nor intended. Rather, it is applied to the cover [D] for the sole purpose of garment attachment.

K-C argues that the components [C] and [D] constitute nothing more than individual plies of the product’s 2-ply cover, in which each ply overlap [C-Ca] and [D-Da] is sealed by a single adhesive strip, [G] and [F], respectively.

In reaching its determination of non-infringement, the district court found as facts that J & J’s product cover [D] was made with non-woven material, e.g., paper, that tissue [C] which is used to hold the internal component parts together during manufacture is not part of the cover, and that adhesive line [G] does not function to seal the cover overlap as required by claim 1.

With respect to dependent claim 3, which requires that the adhesive “also penetrate partially into said pad,” the district court found no infringement because neither of the two adhesive strips used in defendants’ accused product penetrates into the pad because it cannot penetrate the polyethylene poly baffle [B],

On April 5, 1983, the district court issued its “Memorandum Opinion and Order” finding the Roeder ’371 patent invalid for “obviousness” and “unenforceable” because of “fraud on the Patent Office.”

As to “fraud,” the district court held that “The plaintiff [K-C] failed to disclose the Tyrrell, Beery, and Joa patents” and “prior work done by plaintiff’s research department, specifically by Carolyn Mobley and John Champaigne____”

Tyrrell discloses a disposable, plastic-backed shield to be adhered to a garment, which uses two strips of transfer adhesive tape to adhere the waterproof backing of the shield to the garment. Fig. 1, reproduced below, shows the underside of the shield with adhesive strips 28 having release-coated removable protective strips 36.

In discussing Tyrrell, the district court said:

Of the two features of the double lines, (1) being longitudinal and (2) centrally disposed with respect to the sides and ends, the Tyrrell patent discloses the first. The fact that the Tyrrell lines are not also centrally disposed with respect to the sides and ends of the pads does not sufficiently distinguish Tyrrell from the Roeder invention. The Tyrrell patent was material and more pertinent than the prior art actually cited to the Patent Office.

Both Joa and Beery disclose an adhesive composition for sealing sanitary napkin covers which is applied to the cover overlap to penetrate and seal the cover. The adhesive in Joa is applied as a fine spray which penetrates the cover. In Beery the thermoplastic adhesive is heated and subjected to pressure to promote the penetration. The trial court stated, “The use of penetration and sealing was thus fully disclosed by the Beery and Joa patents.”

As to the K-C in-house research, there are two items which the district court treated as “prior art” under § 103. First, there is the work of John F. Champaigne, Jr., exemplified by the invention disclosed in U.S. patent No. 3,665,923, discussed and illustrated infra. The district court found that

The Champaigne napkin is identical to the Roeder napkin except that Cham-paigne uses only a single, wide adhesive strip in the center of the pad and Roeder uses two, narrow, centrally disposed, parallel adhesive strips.

Also, at a later point in the opinion, the court said:

The napkin developed by Champaigne ... was a duplicate of the Roeder napkin except that [it] used only one adhesive line.

Second, the district court also treated as prior art an in-house experiment at K-C conducted by Carolyn Mobley to determine “whether the adhesives would penetrate.” The court’s conclusion was that “[t]hey did, and that [the] discovery was noted in her laboratory notebook.”

The court, in addition to finding Roeder aware of the aforesaid references and prior work, found them to be both material and the most relevant prior art, and that K-C’s failure to make disclosure thereof to the examiner constituted “fraud on the Patent Office.”

Finally, the district court held the claimed invention in Roeder’s ’371 patent obvious under 35 U.S.C. § 103, saying:

The prior art showed the use of emulsion adhesives to seal napkin covers and some of those adhesives penetrated into the pad. To use two lines rather than one is obvious. It is a mere duplication of elements.

Issues on Appeal

Whether the district court erred in holding:

1. the invention claimed in the Roe-der ’371 patent obvious from the pri- or art;

2. that K-C committed fraud in the PTO; and,

3. noninfringement by J & J or PPC.

OPINION

I. Obviousness

The district court began its discussion of obviousness by stating: “Patents are presumed to be valid, but that presumption is weakened when, as in this case relevant prior art is not cited to the Patent Office.” While such statements have often been made, this court recently addressed this point in American Hoist & Derrick Co. v. Sowa & Sons, Inc., 725 F.2d 1350, 1358-60, 220 USPQ 763, 770-71 (Fed.Cir. 1984), stating:

[N]ew prior art not before the PTO ... has no effect on the presumption [of validity] or on who has the burden of proof. They are static and in reality different expressions of the same thing — a single hurdle to be cleared____ [The effect] of new prior art ... is to eliminate or at least reduce, the element of deference due the PTO, thereby partially, if not wholly, discharging the attacker’s burden ____

... That burden is constant and never changes and is to convince the court of invalidity by clear evidence. [Emphasis except “discharging” added.]

Thus, appellees had the burden to prove by clear and convincing evidence the invalidity of the Roeder '371 patent. We find error in the holding that they did so.

The district court continued by noting the obviousness standard mandated by 35 U.S.C. § 103 and reiterated in Graham v. John Deere Co., 383 U.S. 1, 17-18, 86 S.Ct. 684, 693-694, 15 L.Ed.2d 545 (1966). While the ultimate determination is a legal conclusion, it rests on the factual determinations required by § 103. K-C argues that both the factual determinations and ultimate conclusion of obviousness were incorrect. We review K-C’s contentions, but review of the underlying factual findings is limited to determining whether they were clearly erroneous in light of the evidence. Fed.R.Civ.P. 52(a); Carl Schenck, A.G. v. Nortron Corp., 713 F.2d 782, 785, 218 USPQ 698, 700 (Fed.Cir.1983).

A. Scope and Content of the Prior Art

K-C contends that the district court incorrectly included as prior art the in-house work at K-C by Champaigne and Mobley and incorrectly determined the scope of the remaining prior art. We address these contentions separately.

The work of Champaigne found by the district court to be prior art related to a pad with a single wide line of attachment adhesive. To qualify as prior art, this work must meet the requirements of § 102(g) which, in relevant part, reads:

A person shall be entitled to a patent unless—

(g) before the applicant’s invention thereof the invention was made in this country by another who had not abandoned, suppressed, or concealed it.

Section 102(g) “relates to prior inventorship by another in this country as preventing the grant of a patent____ It ... retains the rules of law governing the determination of priority of invention____” P.J. Federico, Commentary on the New Patent Act, 35 U.S.C.A. page 1, at 19 (1954).

Thus, under § 102(g), one of the issues now before us is the question of priority as between Champaigne and Roeder. K-C argues that the district court incorrectly determined Champaigne’s work to be prior art on the ground that “Roeder’s work preceded ... Champaigne’s filing date,” citing In re Clemens, 622 F.2d 1029, 206 USPQ 289 (CCPA 1980). Clemens was an appeal from the PTO Board of Appeals (board) of a §§ 102(g)/103 rejection on an invention shown in a Barrett patent, issued to a common assignee, having an earlier filing date. The dispositive issue in Clemens, as here, was whether there had been a prior reduction to practice of Barrett’s invention relied on as invalidating art. Our predecessor court reversed the board, saying:

Because the record does not support a finding that Barrett made his invention before appellants [Clemens et al.] made the invention ..., the present 35 USC 102(g)/103 rejection ... must fall. [622 F.2d 1029, 206 USPQ at 299.]

Here, K-C relies on Roeder’s experimental conclusions on the location of the adhesive strips as proof of Roeder’s prior reduction to practice. Those Roeder conclusions, however, make no reference to the adhesive’s multiple functions which the examiner, the board and the trial court correctly determined to be critical limitations of the ’371 patent claims. We are therefore not persuaded that the record supports a reduction to practice of Roeder’s claimed invention prior to Champaigne’s filing date. Accordingly, the district court’s finding that Champaigne’s invention was a prior reduction to practice was not legally wrong. Since Champaigne’s invention meets the other requirements of § 102(g)¹ his work can be considered as prior art to Roeder under § 103.

The second item of K-C in-house work, that of Carolyn Mobley, consisted of laboratory experiments, documented in lab notebooks, designed to test various adhesive mixtures to find an adhesive composition that would alleviate the prior art problems ranging from insufficient tack to the undergarment to transfer of the adhesive to the undergarment. To qualify as prior art, this work also has to fit within the requirements of § 102(g).

The district court, in finding Mobley’s work to be prior art, made specific references to those adhesive experiments conducted “to determine whether the adhesives would penetrate,” and concluded that Mobley determined “[t]hey did, and that discovery was noted in her laboratory notebook.” The trial judge made no reference to a statutory basis for finding the Mobley work to be prior art, nor did she make a specific legal determination as to whether there had been a successful reduction to practice of Mobley’s work.

K-C argues Mobley’s experiments were not reduced to practice. J & J counters that such a suggestion is “absurd” stating: The actual making of these napkins is

recorded in numerous lab notebooks____ In fact, hundreds of the napkins were actually worn by women participating in tests conducted by Champaigne, Mobley and outside consultants____

While evidence of in-house testing may be prima facie evidence of conception, reduction to practice requires that an invention be sufficiently tested to demonstrate that it will work for its intended purpose. Barmag Barmer Maschinenfabrik v. Murata Machinery, Ltd., 731 F.2d 831, 838, 221 USPQ 561, 567 (Fed.Cir.1984); Shurie v. Richmond, 699 F.2d 1156, 1159, 216 USPQ 1042, 1045 (Fed.Cir.1983). The record shows that the usefulness of the adhesive mixtures for their intended purpose was not inherently apparent, so that utility must have been demonstrated by actual testing of various adhesive mixtures. Mobley’s experiments failed to set forth a single adhesive mixture that performed with sufficient success. As is the case here, “When the invention has not quite passed beyond experiment and has not quite attained certainty and has fallen short of demonstrating the capacity of the invention to produce the desired result, the invention itself is still inchoate.” 1 Del-ler’s Walker on Patents, § 46 at 202 (2d ed. 1964). We hold the Mobley experiments were not prima facie evidence of a reduction to practice. Under § 102(g), proof of a conception alone does not suffice to establish Mobley’s work as prior to Roe-der’s invention. We therefore agree with K-C that Mobley’s work was unavailable as § 103 “prior art” under § 102(g).

In addition, both parties have responded to the district court’s finding that Roeder was aware of the foregoing in-house research. We need consider this point only with respect to Mobley’s work since we have already decided, above, that Cham-paigne’s work is § 103 prior art under § 102(g). K-C argues that “there is no evidence that Roeder was aware of any specific Champaigne or Mobley work on which the court could have relied. In re Clemens, 622 F.2d 1029, 1039-40 (C.C.P.A. 1980).” J & J counters, stating there is “substantial evidence to support the Court’s finding that the Mobley and Cham-paigne work was known to Roeder____ Accordingly, this work by Champaigne and Mobley clearly qualifies as prior art. In re Clemens ...; In re Bass, 474 F.2d 1276 [177 USPQ 178] (CCPA 1973).”

Both parties are citing Clemens for the legal proposition that personal knowledge of non-public work is sufficient to qualify that work as § 103 “prior art.” Clemens, as previously discussed, involved a §§ 102(g)/103 rejection in which the reference, Barrett, was determined not to be prior under § 102(g) and therefore was not available as § 103 prior art. As § 102(g) contains no personal knowledge requirement, the court’s sole discussion of personal knowledge was dictum in the course of a discussion which distinguished the facts before it from those in a previous opinion of the court also dealing with §§ 102(g)/103, namely, In re Bass. In that § 103 discussion, the court said:

[W]here this other invention is unknown to both the applicant and the art at the time the applicant makes his invention, treating it as 35 USC 103 prior art would establish a standard for patentability in which an applicant’s contribution would be measured against secret prior art. [622 F.2d at 1040, 206 USPQ at 299.]

The key factor in the above statement is the reference to “secret prior art.” As Clemens points out, the use of such secret art — as § 103 “prior art” — except as required by § 102(e), is not favored for reasons of public policy.

In conclusion, the district court’s finding that Roeder was aware of Mobley’s work is irrelevant to a §§ 102(g)/103 analysis. It is still not prior art.

The in-house “work” which we find to be prior art is therefore that of Champaigne alone, whose patent is before us, the work thus having become public. Although appellees, for reasons not made clear to us, state in their brief that they do not contend that the patent is prior art,² they depend entirely on its disclosure to show what Champaigne’s “work” was, and we necessarily proceed on the same basis. The essence of their argument, adopted by the trial court (quoted supra), is, as stated in their brief:

When compared with the Roeder napkin, it can readily be seen that the only difference between the two pads is that Roeder has two lines rather than one. The dual functioning adhesive works exactly the same way in Champaigne ... as in Roeder____

They then reproduce Champaigne’s Fig. 1, as follows, to which we add Fig. 2, a cross-section taken on the line 2-2 of Fig. 1:

The flaw in appellees’ argument, and the consequent clear factual error of the trial court which accepted it, resides in the nonexistence in Champaigne’s work, as established by evidence, of a “dual functioning adhesive.” Mobley may have been working on one, but we have held her work was not successful and is not prior art under § 103.

An understanding of what Champaigne has been shown to have produced requires a brief explanation of what his patent discloses. Referring to Figs. 1 and 2, the structure comprises a pad 12 enclosed in a wrapper 14 with overlapping edges, the outer edge being shown at 15. Fluid-pervious wrapper 14 is a “non-woven fiber web bonded -by a water-dispersible binder.” The overlap at 15 is “sealed by a water-dis-persible adhesive similar to that used for binding together the fibers in the non-woven wrapper.” The disclosed “adhesive” is “a polyvinyl alcohol which is soluble in cold water.” The whole idea is that when flushed down a toilet the entire wrapper, including its overlap seal, will disintegrate because of binder/adhesive solubility. Furthermore, the patent disclosure warns against using a binder or adhesive containing too much polyvinyl alcohol (more than in a 5 to 15% solution) because “[e]xcessive amounts are inclined to become sticky in use, and are thus not as desirable.” Roe-der, we comment at this point, requires a sticky adhesive. So much for the adhesive with which Champaigne secures the edges of his wrapper, which appellee’s brief categorizes as an “emulsion adhesive.”

Champaigne also discloses, of course, garment attachment means in the form of an adhesive, which was not anything Champaigne claimed to have invented. “The prior art is replete with patents defining napkins of this type,” his patent states, and lists 9 patents. He describes his own version:

An elongate strip of pressure sensitive adhesive 18 (the dotted phantom line 18 in Fig. 1) is disposed centrally on the bottom portion of the enclosing wrapper and is covered by a removable protective strip 19.

The last paragraph of the specification reads:

The pressure sensitive adhesive used for attachment may be any one of a number of well-known conventional types and may be in the form of a two-sided tape or may be directly applied to the wrapper. It may be in the form of a single elongate strip or in separate strategically located patches.

Fig. 4 of his patent shows crosswise patches near the pad ends. Clearly, then, so far as Champaigne’s patent disclosure is concerned (and appellees tell us the “work” of Champaigne and Mobley “was eventually incorporated in the Champaigne patent”), he had two distinct and quite different adhesives, not a “dual-functioning adhesive” as stated in appellees’ brief. One of them, used to secure the overlap of wrapper 14, was the polyvinyl alcohol solution which, when dried, was required to be non-tacky. The garment-securing adhesive, on the other hand, was a strip of prior art “pressure sensitive” adhesive applied on top of the already sealed overlap. If necessary, we take judicial notice that a “pressure sensitive adhesive”³ is permanently tacky. It cannot be said, therefore, as did the trial court, that Roeder’s napkin and Champaigne’s are identical except that Roeder uses two strips of adhesive instead of one.

Champaigne’s patent further distinguishes the two adhesives he used in discussing napkin disposal. He points out that the water in a toilet “rapidly dissolves the water-dispersible adhesive which binds the fibers in the wrapper together” so that the wrapper breaks up. In contrast, he explains that “the pressure-sensitive adhesive areas will not break up at once but these areas are relatively small and should not interfere with disposal since they will eventually degrade.” This is a far cry from Roeder’s single adhesive structure and directly contrary to appellees’ interpretation which was accepted by the trial court.

B. Would Roeder’s Claimed Invention Have Been Obvious?

In this court, appellees persist in repeating their former argument. Their brief states, under the heading “Emulsion Adhesives”:

The prominence of emulsion adhesives is demonstrated by K-C’s own work. Champaigne’s patent describes the emulsion adhesive that he used as being “any one of a number of well-known conventional tapes [sic, types]” (A 220, col. 4, 1. 55; A 578).

The first specific reference cited in that quotation is to the last paragraph of the Champaigne patent’s specification which we quoted above. Clearly the cited statement does not refer to the emulsion adhesive used to secure the overlap of the wrapper but to the distinct and separate pressure sensitive adhesive applied on top of the overlap after it is sealed. Therefore, there is not the slightest suggestion of a. dual function adhesive. Moreover, the only specific adhesive disclosed for garment adhesion is two-sided tape. The second reference (“A 578”) is to testimony of Cham-paigne, under examination by appellees’ counsel, which not only fails to support appellees’ argument but flatly contradicts it.

We would not have spent so much time discussing the foregoing details of the Champaigne structure had the trial court not made it so pivotal to all of its reasoning on most of the issues in this case — obviousness, fraud in the PTO, and attorney fees— and had the trial court not been, as it seems to us, so thoroughly misled as to just what it was. The Champaigne structure may have led up to Roeder’s invention but the court was clearly wrong in concluding that the Champaigne napkin “was a duplicate of the Roeder napkin except that [it] used only one adhesive line.”

Not only does Champaigne’s “work” (as ultimately disclosed in his patent) appear to have been basic to the trial court’s thinking, it is the only prior art specifically mentioned in its discussion of the obviousness issue, which is to be found in the third and last opinion of the court in this case, 219 USPQ at 220. The rest of the prior art is referred to only in general terms and it is therefore difficult .to say just how it was applied to the Roeder claims in finding the invention they define to have been obvious. We have carefully analyzed the court’s reasoning and are constrained to conclude that its rationale is fatally flawed for the following reasons.

After some general discussion, the court came to the conclusion that “The inquiry ... is whether two lines of adhesive, both of which penetrate the napkin and seal the cover as opposed to only one penetrating and sealing the cover, is a non-obvious invention.” At the outset, that is a misstatement of the issue, which is whether the subject matter claimed would have been obvious, at the time of Roeder’s invention, to one of ordinary skill in the art. 35 U.S.C. § 103. Claim 1, the only independent claim, is set forth at the beginning of this opinion. We need consider no other claim because if the invention of claim 1 would not have been obvious the same is true as to the remaining dependent claims. Shortly after quoting claim 1, we set forth Finding of Fact 12, submitted by counsel and adopted by the court, which correctly states that the two lines of adhesive are lines of pressure sensitive adhesive for garment attachment. It is clear from the patent in suit, as at other times the trial court seems to have appreciated, that that adhesive had to serve the dual function of (1) wrapper penetration and sealing and (2) garment attachment, a concept totally lacking in Champaigne, or in any other prior art reference, which is why the PTO allowed Roeder’s claims. This essential of the claimed invention seems to have been lost sight of by the trial court in its discussion of obviousness. From the misstatement of the inquiry forward, the discussion is in terms merely of using two lines versus one line of wrapper-securing adhesive. This is manifest from such statements as: “Plaintiff did not need a napkin with two lines of adhesive”; “To use two lines rather than one is obvious”; “there was no need for both of those lines to penetrate into the napkin. This is the correct inquiry, because that is the essence of plaintiff’s invention”; “the use of two rather than one adhesive line to penetrate and seal had almost no significance.”

To have thus reduced the subject matter claimed to the oversimplified question of two lines versus one, apparently on a theory that Champaigne disclosed everything except that difference, was clear error, because it failed to deal with what is claimed, what Champaigne’s “work” was (as shown in his patent), and of how the issue of obviousness must be approached.

The proper approach to the obviousness issue must start with the claimed invention as a whole. 35 U.S.C. § 103. The court’s Finding of Fact No. 12, supra, adopted on the infringement issue, is a fair summary of what that invention was. It is true that it consists of a combination of old elements so arranged as to perform certain related functions. It is immaterial to the issue, however, that all of the elements were old in other contexts. What must be found obvious to defeat the patent is the claimed combination. The court appears not to have considered that precise question. Regardless of how patent claims may be arrived at during the vicissitudes of prosecution, the invention patented is no more and no less than what the finally issued claims, as construed by the court, define; and they must be construed in the identical way for both infringement and validity. That Roeder may have once thought he could get a claim allowed on the broad idea of merely using two lines of adhesive, or that his attorney may have made the attempt without success, is not determinative of what he finally succeeded in patenting as his invention. The court must consider, on the obviousness issue, only what was granted, not the unsuccessful — and not unusual — attempts at overreaching in the early stages of the patent application prosecution.

Since the trial court’s opinion lacks a comparison of the invention of Roeder’s claims with the prior art, we have made it ourselves, examining all of the references of record, and we fail to find a clear suggestion of the claimed subject matter. The invention which we find non-obvious is, however, that which is specifically claimed in the patent in suit, a narrow invention in a crowded art. The holding of invalidity on the ground of obviousness is therefore reversed.

II. Fraud

A. Who Is Presumed To Know The Prior Art

In ending the section of her April 5,1983, opinion, (573 F.Supp. 1179, 219 USPQ 217, 218) on “Fraud on the Patent Office,” the trial judge concluded that plaintiff was chargeable with “fraud” because prior art had not been disclosed to the PTO. This decision was amplified by footnote 3 reading (573 F.Supp. 1179, 219 USPQ p. 220):

In closing argument^] plaintiff’s counsel argued that defendant had not proved intent to defraud. As I understand the law, proof of specific intent to defraud is not necessary. What is called “fraud on the patent office” is proved by failure to disclose material prior art and the patent applicant is presumed to have knowledge of the existence of material prior art. [Our emphasis.]

Now, that understanding of “the law”— which has become a headnote in the USPQ report of the opinion — is a witches’ brew which could do untold damage to patentees if ingested into the body of patent law. Even the fumes of this brew, whoever originated it, are so stupifying that they have driven us to analyze anew this ancient presumption of knowledge of prior art on the part of inventors. Whence did it come, what does it mean, why does it exist, is it valid, and should it survive? We will treat of this presently, but first an example of why combining the “fraud by failure to disclose” concept with the “presumption of knowledge of prior art” concept is on a par with terrorism.

Suppose the inventor, as here, to be working in Neenah, Wisconsin, U.S.A. The item of material prior art is a Japanese patent issued before the inventor’s date of invention, printed in the language of that country and never translated. No copy of it had reached this country before the prosecution of the inventor’s patent application was completed; but under the statute the Japanese patent is § 102(a) prior art. The inventor knows nothing of it, however. By the trial judge’s understanding of the “law,” presumably under instruction from learned patent counsel, the inventor, with no knowledge of the Japanese patent, commits “fraud in the PTO”⁴ by not disclosing it, thereby rendering his patent invalid or unenforceable, possibly turning the lawsuit into an “exceptional case” in the mind of the trial judge, thus making his assignee liable to pay attorney fees! “Proof of specific intent to defraud is not necessary.” Knowledge by the inventor of the Japanese patent, which is unquestionably material prior art, “is presumed.”

Surely, such an absurd result is not to be condoned. It would suffice, perhaps, to say that the presumption of the inventor’s knowledge of the prior art, long treated as axiomatic in patent law, is not to be applied in connection with the duty of candor and the obligation to disclose material prior art to the PTO — limiting the obligation to disclose material prior art actually known to the inventor or his assigns, agents, and attorneys. See 37 CFR 1.56. However, since it has actually happened that the axiom has been combined with the obligation to disclose — a tactic likely to be repeated — we are moved to analyze this old axiom (as we have analyzed equally venerable dogmas from the past for the purpose of burying them) to see what, if any, validity or vitality it possesses today.

The germ of the idea is very old, going back at least to Eaton v. Evans, 16 U.S. (3 Wheat.) 210, 4 L.Ed. 454 (1818). Under rather complicated facts and without benefit of modern patent claims, questions were raised as to what Evans’ (the famous Oliver Evans) patent actually covered. The Court, by Chief Justice Marshall, said (p. 237):

The plaintiff has also contended, that it is not necessary for the patentee to be the first inventor or discoverer. That the law is satisfied, by his having invented a machine, although it may have been previously discovered by some other person____ [T]he court considers this question as completely decided by the 6th section of the general patent act [of 1790]. That declares, that if the thing ' was not originally discovered by the pat-entee, but had been in use, or had been described in some public work, anterior to the supposed discovery of the pat-entee, judgment shall be rendered for the defendant, and the patent declared void____ It may be, that the patentee had no knowledge of this previous use or previous description; still his patent is void: the law supposes he may have known it____ [Our emphasis.]

Supposing, in this context, equates with presuming. In an early opinion of the Circuit Court, D. Massachusetts, Crompton v. Knowles, 7 F. 199 (1881), Lowell, C.J., said (p. 203):

It is a presumption of law that all mechanics interested in upholding or defeating a patent were fully acquainted with the state of their art when they took out their patent, or when they built their machine. This presumption is founded upon a policy like that which imputes to all persons charged with crime a knowledge of the law. It is necessary to the safe administration of justice. Each party may then be assumed to have borrowed from the other whatever was actually first invented and used by that other. [Our emphasis.]

In the Supreme Court case of Mast, Foos & Co. v. Stover Mfg. Co., 177 U.S. 485, 20 S.Ct. 708, 44 L.Ed. 856 (1900), involving gearing for a windmill, Mr. Justice Brown, after an extensive discussion of prior art, said (p. 494, 20 S.Ct. p. 712):

While it is entirely true that the fact that this change had not occurred to any mechanic familiar with windmills is evidence of something more than mechanical skill in the person who did discover it, it is probable that no one of these [mechanics] was fully aware of the state of the art and the prior devices; but, as before stated, in determining the question of invention, we must presume the pat-entee was fully informed of everything which preceded him, whether such were the actual fact or not. There is no doubt the patent laws sometimes fail to do justice to an individual who may, with the light he had before him, have exhibited inventive talent of a high order, and yet be denied a patent by reason of antecedent devices which actually existed, but not to his knowledge, and are only revealed after a careful search in the Patent Office. But the statute (sec. 4886 [R.S.] is inexorable. It denies the patent, if the device were known or used by others in this country before his invention. [Our emphasis.]

Justice Brown, as appears to have been his habit, cited no precedent supporting the existence of the presumption. In two earlier opinions he wrote, he had made similar holdings. In Derby v. Thompson, 146 U.S. 476, 481-82, 13 S.Ct. 181, 183-84, 36 L.Ed. 1051 (1892), discussing the patentability of a highchair that converted into a sort of stroller, and speaking of the patentee, he said:

Though he may not in fact have known of these three chairs, but may have supposed that he was inventing something valuable, we are bound, in passing upon his device, to assume that he had them all before him, and with that knowledge it seems to us that it required nothing more than the skill of an ordinary mechanic to adopt the most valuable features of each in the construction of a new chair. [Our emphasis.]

In Duer v. Corbin Cabinet Lock Co., 149 U.S. 216, 223, 13 S.Ct. 850, 853, 37 L.Ed. 707 (1893), which involved a lock for a cabinet drawer or the like patented by Mr. Orum, Justice Brown said:

In view of the fact that Mr. Orum had no actual knowledge of the Gory patent, he may rightfully claim the quality of invention in the conception of his own device, but as he is deemed in a legal point of view to have had this and all other prior patents before him, his title to invention rests upon modifications of these, too trivial to be the subject of serious consideration. [Our emphasis.]

In these opinions, we witness the age-old struggle of the courts with the philosophical problem of reconciling the fact of pat-entees having truly made “inventions,” when seen from the standpoint of what their personal efforts were, while having at the same time to hold that they were not “inventors” in the eyes of the law and their production not “inventions,” because, under the law, they were not entitled to patents. In those days, the invention could not be regarded as having involved “invention,” the sine qua non of patentability.

Contemporaneously with the foregoing decisions, and in conformity with them, the textwriters were recording the law of patents for future generations, outstanding among them being Professor Albert H. Walker, whose text became the standard work. The first edition of Walker on Patents (1883), in the chapter entitled “Invention,” said, in § 43,

Every inventor or constructor is presumed by the law to have borrowed from another, whatever he produces that was actually first invented or constructed or used by that other [our emphasis],

such presumptive “borrowing” being, of course, from what is now called “prior art.” Those exact words were continued in six subsequent editions down through the first “Deller’s Edition” of Walker published in 1937 where they were printed under the sub-heading “Patentee Presumed to Know Prior Art” (§ 25 at page 117). In the second edition of “Deller’s Walker on Patents” (1964) the words are still there under the same heading, modified only by the omission of “or constructor” from the first line (§ 105 page 28). Since Mast, Foos was decided, it has been the principal decision cited in support of the statement. In all of these learned tomes through the 1937 edition, the chapter on “Invention” has been intended as a discussion of that former prerequisite to patentability known as “the requirement for invention” which had existed in the law since Hotchkiss v. Greenwood, 52 U.S. 248, 13 L.Ed. 683 (1850). Not until the current 2d Deller edition was the chapter renamed, more appropriately in view of the 1952 Patent Act, “Patentable Invention.”

Simultaneously, there ran through the Walker texts a related statement in the chapter entitled “Novelty” which was § 73 in the first edition, reading in its entirety:

Negation of novelty is not averted by the fact that the inventor had no knowledge of the anticipating matter when he made the invention covered by the patent. The patent laws do not reward people for producing things which, though new to them, are old to others in this country.

It continued unchanged through the 1937 edition (1st Deller § 59) and is still present in the 2d Deller (§ 73) but is now combined with the presumption of knowledge of prior art concept, reading thus:

Negation of novelty is not averted by the fact that the inventor had no knowledge of the anticipating matter when he made the invention covered by the patent; he is presumed to know the prior art even if he had no actual knowledge. The patent laws do not reward persons for producing things which, though new to them, are old to others in this country. It has been held that the patentee is presumed to know not only prior patents but also practices old in the art. [Our emphasis.]

Thus has the conventional wisdom been purveyed, hundreds if not thousands of lawyers and judges learning their law from these texts. It is therefore not surprising that the courts have perpetuated the law as they learned it and as, indeed, it was. By way of example, hear it in the eloquent words of Judge Learned Hand in a number of opinions, all emphasis being ours:

[Presumably he designed his machine from the ground up and without suggestion or cue; and we may well agree that, taken as a feat, that displayed enough originality to support a patent. However ... the law does not use such a subjective test in judging a patent; on the contrary it imputes to the inventor an omniscience which will again and again deprive him of the reward that his talents as an individual might otherwise deserve. [Frank B. Killian & Co. v. Allied Latex Corp., 188 F.2d 940, 943 (2d Cir.1959).]

[A]s the law stands, the inventor must accept the position of a mythically omniscient worker in his chosen field. As the arts proliferate with prodigious fecundity, his lot is an increasingly hard one. [Merit Mfg. Co. v. Hero Mfg. Co., 185 F.2d 350, 352 (2d Cir.1950).]

[W]e must suppose the inventor to be endowed, as in fact no inventor ever is endowed; we are to impute to him knowledge of all that is not only in his immediate field, but in all fields nearly akin to that field. [International Cellucotton Prod. Co. v. Sterilek Co., 94 F.2d 10, 13 (2d Cir.1938).]

Mason at best was the last of a number of inventors.... True, Mason may personally be entitled to more credit than that; he may have devised the whole machine unaided; but that is not the test; the law unsparingly charges him with an acquaintance with all the earlier developments of the art. [Foxboro Co. v. Taylor Instrument Cos., 157 F.2d 226, 233 (2d Cir.1946).]

It is therefore not surprising that our predecessor Court of Customs and Patent Appeals would say, even after the effective date of the 1952 Patent Act, in discussing its new § 103:

We think the proper way to apply the 103 obviousness test to a case like this is to first picture the inventor as working in his shop with the prior art references which he is presumed to know — hanging on the walls around him. In re Win-slow, 365 F.2d 1017, 1020, 151 USPQ 48, 51 (CCPA 1966). [Our emphasis.]

The court later found it necessary to qualify that overly picturesque statement, which has met with unfortunate popularity, in In re Antle, 444 F.2d 1168, 1171-72, 170 USPQ 285, 287-88 (CCPA 1971), and even then reiterated the statement about the presumption, saying, “As we also said in Winslow, 'Section 103 requires us to presume full knowledge by the inventor of the prior art in the field of his endeavor adding, however, that “it only requires us to presume that the inventor would have that ability to select and utilize knowledge ... pertinent to his particular problem which would be expected of a man of ordinary skill in the art to which the subject matter pertains.” That man, of course, is the hypothetical person postulated by § 103.

Therein we see the forecast of confusion to come. If we look at subsequent cases such as are collected in the E. Lipscomb, III, October 1983 cumulative supplement to Volume 2 of Deller’s second edition of Walker on Patents as annotations to § 105 under the heading “Patentee Presumed to Know the Prior Art,” what we observe is cases, beginning with the CCPA’s Winslow image, saying the inventor is presumed to know all the prior art, others saying it is the patentee to whom the presumption applies, and many, in encouragingly increasing numbers, holding, not that it is the inventor or applicant or patentee, but that it is the hypothetical person of ordinary skill in the art who is referred to in § 103 of the 1952 Patent Act who must be presumed to have, or is charged with having, knowledge of all material prior art. (See the supplement, supra, at pages 28-30.) Why the change, and what is its significance?

Before 1952, the statutes contained no such provision as § 103. Its requirement of nonobviousness as a prerequisite to pat-entability existed in the form of the case law requirement for “invention.” To be entitled to a patent, the inventor was required to have made an “invention” — inversely, he had to be “an inventor.” This requirement was in addition to the requirements of novelty and utility, which were always statutory. This was law made by the courts. In essence, it required that an applicant for patent, when presented with the problem, must have done more than would have been done by “an ordinary mechanic acquainted with the business.” Hotchkiss v. Greenwood, supra 52 U.S. at 265.

As the above collection of quotations makes clear, the courts were constantly faced with situations where the applicant or patentee had actually made a new and useful invention, and therefore was an inventor, in every sense of the word but one — he had invented something which the courts, viewing prior art, felt did not involve the quality of “invention” and a reason for denying a patent had to be rationalized. There seems to have been a felt need to find a reason for saying the applicant or patentee either had not made an “invention” or should not be called “an inventor.” In the lack of novelty cases it was easy enough to say the “statute is inexorable,” because it required novelty; but in a “lack of invention” case there was no statute and to show that “an invention” had not been made, or that the requirement for “invention” had not been met, the applicant or patentee was, in effect, viewed from on high and treated as though he knew things he really did not know in order to determine his rights from a social point of view. There was no one other than the inventor to pin it on. By presuming he knew all the prior art, he could be said not to have made “an invention” and, not being “an inventor,” he could be denied a patent or, if he had one, it could be invalidated, as well-developed law required. This line of reasoning having been once established, the courts, including our predecessor court as in Winslow, mentioned above, decided cases on the basis of the axiom — the applicant (or patentee) is presumed to have knowledge of all material prior art — without giving the situation further thought. But a bad axiom is like a noxious weed, a thriving plant in the wrong place interfering with the growth of more desirable plants.

Basing the rationalization of decisions on a fiction — which the presumption of knowledge is — it has been unusual that opinions have explained the real reason for the denial of patent rights, which is the basic principle (to which there are minor exceptions) that no patent should be granted which withdraws from the public domain technology already available to the public. Graham v. John Deere Co., 383 U.S. 1, 6, 86 S.Ct. 684, 688, 15 L.Ed.2d 545, 148 USPQ 459, 462 (1966). It is available, in legal theory at least, when it is described in the world’s accessible literature, including patents, or has been publicly known or in the public use or on sale “in this country.” 35 U.S.C. § 102(a) and (b). That is the real meaning of “prior art” in legal theory — it is knowledge that is available, including what would be obvious from it, at a given time, to a person of ordinary skill in an art. Society, speaking through Congress and the courts, has said “thou shalt not take it away.”

Since January 1, 1953, the effective date of the 1952 Patent Act, the implementation of that social policy has not required courts to use the legal fiction that an inventor must be presumed to know the “prior art.” The inventor, for the purposes of legal reasoning, has been replaced, as some courts have discovered, by the statutory hypothetical “person having ordinary skill in the art”⁵ who has been provided by 35 U.S.C. § 103. Since that date, there has been no need to presume that the inventor knows anything about the prior art.

Since we believe that progress in legal thinking is not only possible but highly desirable when it simplifies such thinking, we believe the time has come to discontinue this particular fiction of the patent law. Congress has given us in § 103 a substitute for the former “requirement for invention,” which gave rise to the presumption, and that substitute, being statutory, should be used exclusively. We hereby declare the presumption that the inventor has knowledge of all material prior art to be dead.

What controls the patentability of the fruits of the inventor’s labors are the statutory conditions of novelty, utility, and unobviousness “to a person having ordinary skill in the art to which said subject matter pertains” as stated in § 103. It should be clear that that hypothetical person is not the inventor, but an imaginary being possessing “ordinary skill in the art” created by Congress to provide a standard of patentability, a descendant of the “ordinary mechanic acquainted with the business” of Hotchkiss v. Greenwood. Realistically, courts never have judged patentability by what the real inventor/applicant/patentee could or would do. Real inventors, as a class, vary in their capacities from ignorant geniuses to Nobel laureates; the courts have always applied a standard based on an imaginary worker of their own devising whom they have equated with the inventor.

B. Was Plaintiff Chargeable With Breach of Duty to Disclose Material Prior Art?

This heading states our concept of what is really involved here. The unfortunately misnamed “fraud in the PTO” in this case resides at most in a failure to tell the examiner, in the course of examining Roe-der’s application, of the existence of three patents and Champaigne’s work for their employer K-C. There is no suggestion of anything related to real fraud, such as false or misleading statements; we deal only with non-disclosure. The question is the legal effect of nonfeasance, not malfeasance. “Fraud in the PTO” has been overplayed, is appearing in nearly every patent suit, and is cluttering up the patent system. Cases in this category vary immensely in their factual situations and must be decided on a case-by-case basis.

This court has recently held that “Fraud must be proved by clear and convincing evidence, and the party asserting it carries a heavy burden” [cases cited]. Kansas Jack, Inc. v. Kuhn, 719 F.2d 1144, 1151, 219 USPQ 857, 861 (Fed.Cir.1983); Environmental Designs, Ltd. v. Union Oil Co. of Cal., 713 F.2d 693, 698, 218 USPQ 865, 870 (Fed.Cir.1983). In American Hoist & Derrick Co. v. Sowa & Sons, Inc., 725 F.2d 1350, 1362-64, 220 USPQ 763, 772-74 (Fed. Cir.1984), we discussed “fraud in the PTO” in the context of a jury trial, commenting that it “is an area of law fraught with confusion and contradiction.” We also pointed out that it is “clear that an applicant is under no obligation to disclose ‘all pertinent prior art or other pertinent information’ of which he is aware,” citing Digital Equipment Corp. v. Diamond, 653 F.2d 701, 716, 210 USPQ 521, 538 (1st Cir. 1981), and pointing out that two controlling factors are the materiality of the information and the intent of the actor. From Digital we quoted with approval the statement:

Questions of “materiality” and “culpability” are often interrelated and intertwined, so that a lesser showing of the materiality of the withheld information may suffice when an intentional scheme to defraud is established, whereas a greater showing of the materiality of withheld information would necessarily create an inference that its non-disclosure was “wrongful.”

The converse is also true—wrongful intent may lower the materiality requirement. With these considerations in mind we turn to the facts of this case. The trial judge’s legal conclusion was: “I conclude that plaintiff’s [sic] failed to disclose material prior art to the Patent Office [sic] and such failure constitutes fraud on the Patent Office.” (To this was appended the footnote quoted at the beginning of section “A” of this part, dealing with intent.) The court also stated: “The plaintiff affirmatively misled the Examiner.” The underlying basis of these conclusions was: “The plaintiff failed to disclose the Tyrrell, Beery and Joa patents and prior work done by plaintiff's research department.” That work was the Mobley and Champaigne work,- the failure to disclose which was by itself expressly found to constitute “fraud on the Patent Office.” To explain why we do not agree with the trial court’s legal conclusion, we are obliged to consider each of these items seriatim.

1. The Tyrrell patent

We discussed this patent under “Background” and reproduced Fig. 1 from it. Over a third of the court’s “fraud” discussion is devoted to it. The court said, “The Tyrrell patent was material and more pertinent than the prior art actually cited to the Patent Office.” We cannot agree. Defendants rely on this patent only because it shows two parallel lines of adhesive and reproduce in their brief the same Fig. 1 we have reproduced above. They then construct thereon an ingenuous scenario of an attempt to mislead the examiner, which the trial court adopted, none of which, incidentally, is directed to the patented claims in suit but only to original claims which the examiner found adequate reasons for rejecting without recourse to Tyrrell and which were finally cancelled or substantially amended.

The trial judge found K-C’s failure to disclose Tyrrell to the examiner specifically in the Roeder application “puzzling” and found it fatally misleading. We do not. All Tyrrell discloses, of negligible relevance here, is a shield with a polyethylene plastic backing, along the edges of which are two narrow strips of double-faced pressure sensitive adhesive tape, one face being adhered to the plastic. No other kind of adhesive is mentioned. (The trial judge conceded in a footnote it did not disclose a sanitary napkin so equipped.)

Roeder’s patent specification is written with an unusually ample disclosure of prior art, including drawings. Of the 8 prior art patents listed, albeit Tyrrell was not among them, the examiner used two of them to reject 5 of the 6 original claims in his first action. Roeder’s Fig. 6 is reproduced below together with its description (our emphasis):

Fig. 6 shows still another prior art embodiment in which four patches of adhesive are used, two patches 31 and 32 are disposed along each edge of the napkin and another two patches 33 and 34 are disposed at the opposing ends of the napkin.

Clearly this disclosure of acknowledged “prior art” is closer to what Roeder was attempting to claim, and finally succeeded in claiming in his patent, than Tyrrell’s Fig. 1 for the following reasons: Fig. 6 is an equally good showing of two parallel lines of adhesive; the adhesive is not specified as two-faced tape and could well be an emulsion type of adhesive directly applied to the napkin wrapper, in which case it would likely penetrate it, whereas, Tyrrell’s two-faced tape adhesive would not. Of course, Tyrrell’s tape adhesive does not penetrate his plastic shield. It would be quite as useful to an examiner in support of a rejection as Tyrrell, if not more so. In view of this, we can see no justification for finding that Roeder “affirmatively misled the Examiner” or breached his duty of candor in not disclosing Tyrrell. Nor do we read any nefarious purpose into the failure to disclose it in the Roeder application while disclosing it in K-C’s Cham-paigne application, as did defendants and apparently the trial court. Consider, too, the fact that the same examiner, Mr. Rosenbaum, was examining both of these copending applications filed only 3 months apart, and had previously examined Hendricks patent 3,463,154 issued in 1969 in which, over his name, is a list of the references he cited which included Tyrrell. Roe-der’s specification lists the Hendricks patent as prior art. Tyrrell was not being concealed from Examiner Rosenbaum. Most certainly he was not “affirmatively misled.” The essential elements of materiality of Tyrrell and the intent of the patent solicitor are both exceedingly weak and we find the trial court’s conclusion with respect to that reference unsupported and therefore erroneous.

2. The Beery and Joa patents

These two patents are relied on by defendants only for their disclosures of adhesive penetration of napkin wrapper overlap to seal it. In summary, Beery applies a line of hot thermoplastic adhesive, with or without a supporting filament, along the overlap and it is caused to strike through the wrapper layers by heat and pressure. It is specified to be non-tacky when cooled. The napkin is not of the type to be adhered to a supporting garment and the sealing adhesive, being non-tacky in use, could not be used for that purpose. (The patent is assigned to defendants.)

Joa discloses a method of sealing napkin wrappers by applying a fine spray of quick-setting liquid adhesive along the overlap in such dilution and small quantity as to be “almost undetectable” in the finished product.

Since Roeder’s application for the patent in suit, no matter how broad his original claims, was solely concerned from the beginning with napkins which could be attached to garments with pressure-sensitive adhesive, and since these two references expressly teach that their sealants are either non-tacky or undetectable in the final product, we fail to see why anyone would deem either to be material prior art. Therefore, both materiality and intent are lacking.

In Roeder’s 6 original claims, claims 4 and 5 added to claim 1 a limitation that his pressure-sensitive adhesive “penetrates” the wrapper overlap. Apparently for that reason, defendants argued, and the trial court agreed, that prior art showing penetration was vital to proper consideration of the application by the PTO. On the other hand, blowing hot and cold on the significance of such references, the trial court, in discussing obviousness, made the following observations (219 USPQ at 221) with which we fully agree:

The fact that emulsion adhesives do penetrate into the typical sanitary napkin cover and pad would be a surprise to no one who thought about the issue. Sanitary napkins are designed to absorb moisture and when an adhesive is applied (as it is) with heat to the bottom of the napkin, penetration occurs. Although the court heard expert testimony on this issue, and accepts the expert’s opinion, the fact of penetration under these circumstances accords with common sense, even the court’s common sense.5

The footnote, which we have omitted, refers to “expert testimony” to the same effect. We cannot reconcile this view with the holding that it was “fraud in the PTO” or misleading to the examiner, or breach of a duty of candor, not to cite two prior art patents disclosing what would be “a surprise to no one.”

3. K-C’s in-house research

Having ruled that the work of Caroline Mobley was not prior art, we are here concerned only with the work of Roeder’s fellow-employee, Champaigne, whose patent was copending and is admitted not to be prior art but to embody the research work relied on under 35 U.S.C. § 102(g). As to said work, the trial judge said (219 USPQ at 220):

The Champaigne napkin is identical to the Roeder napkin except that Cham-paigne uses only a single, wide adhesive strip in the center of the pad and Roeder uses two, narrow, centrally disposed, parallel adhesive strips.

We have explained in our discussion of obviousness, supra, why that holding is factually clearly erroneous. It is the court’s preamble to the further holding:

Champaigne’s work and the then-pending Champaigne application were not disclosed to the Patent Office in connection with the Roeder application.

... there was an obligation to disclose the prior work performed in plaintiff’s research department. Such prior work was material and more pertinent than the prior art disclosed in the application.

The court held this to be one of the bases for its conclusion that there was “fraud on the Patent Office.” The gravamen of it, since Champaigne’s work was what he discloses in his patent, resides in the above statement that his patent application was not disclosed to the examiner in charge of the Roeder application, Mr. Rosenbaum. It suffices, we think, to note the fact that the name of the examiner printed on the Cham-paigne patent is Charles F. Rosenbaum, the same as that appearing on the Roeder patent. The same K-C patent agent was prosecuting both applications and necessarily knew that Mr. Rosenbaum was in charge of both of them and had the duty of knowing their contents so as to be sure they claimed distinct inventions.

In a case such as this, involving an issued patent attacked for breach of the duty of candor only on the basis of nonfeasance consisting of a failure to disclose known prior art, the key issues of materiality and intent should be decided with reference to the claims of the patent. What we see here, however, is an attempt by defendants to build a case of “fraud” by reason of non-disclosure of prior art material only to abandoned claims long since cancelled during prosecution after being rejected by the examiner as unpatentable for reasons not involving the uncited prior art. To base a conclusion of “fraud” on such grounds is to deal with a hypothetical situation, not with reality. Defendants and the trial judge have had little or nothing to say, in discussing the breach of duty to disclose issue, about the limited claims finally patented; they have not shown that non-disclosure of prior art would have had any effect on their allowance. Therefore, defendants have not sustained their heavy burden.

In view of the foregoing, we hold there was no breach of the duty to disclose or any “fraud in the PTO” on the part of Roeder and reverse the contrary holding of the trial court.

III. Infringement

The trial court merged its oral decision of February 4, 1983, that defendants have not infringed the Roeder ’371 patent, into its published findings and conclusions of March 15, 1983 (219 USPQ 214). Having fully considered the arguments of the parties thereon, we affirm the judgment of non-infringement of claims 1 and 3, the only claims in suit, for the reasons stated, the principal one being that defendants avoid infringement by not having at least two lines of adhesive both of which penetrate and seal both portions of the wrapper overlap. We also agree that there is no infringement by virtue of the doctrine of equivalents. We agree that defendants’ tissue wrap is not part of the “wrapper” of the Roeder patent claims. In short, defendants have successfully designed around Roeder’s claims, as they had a right to do. As the Supreme Court stated in White v. Dunbar, 119 U.S. 47, 52, 7 S.Ct. 72, 75, 30 L.Ed. 303 (1886):

The claim is a statutory requirement, prescribed for the very purpose of making the patentee define precisely what his invention is; and it is unjust to the public, as well as an invasion of the law to construe it in a manner different from the plain import of its terms.

IV. Attorney Fees

Section 285 of the 1952 Patent Act provides that “The court in exceptional cases may award reasonable attorney fees to the prevailing party.” That Act added the words “in exceptional cases” to the prior statute. P.J. Federico, “Commentary on the New Patent Act,” 35 U.S.C.A. 1, p. 56. They are not to be awarded in normal cases involving no unconscionable conduct on the part of the losing party. We note that here neither party has prevailed on all issues. The trial judge found this to be an “exceptional case” primarily, it would appear, because of the “plaintiffs fraud on the Patent Office [which] was substantial and material,” but not only for that reason. The matter of attorney fees rests, of course, within the sound discretion of the trial judge, who is in the best position to determine whether, as an equitable matter, they should be awarded. Since the trial court clearly stated that “the fraud on the Patent Office in connection with the Roeder patent is most relevant and an adequate basis for an award of attorney’s fees” (573 F.Supp. 1179, 219 USPQ at 223), and since we have reversed that “fraud” finding, we must vacate the award and remand so that the question may be reconsidered and the parties heard thereon.

SUMMARY ⁶

The trial court’s holdings of invalidity and unenforceability are reversed. The holding of non-infringement is affirmed. The court’s award of attorney fees is vacated, and the case is remanded for further proceedings consistent herewith.

Each party shall bear its own costs in this court.

REVERSED IN PART, AFFIRMED IN PART, VACATED IN PART AND REMANDED.

1 . As discussed infra, the timely filing of a patent application, and subsequent issuance of a patent thereon, covering the Champaigne pad, constitutes sufficient proof that his invention was not abandoned, suppressed, or concealed.

2 . We presume it amounts to a concession that Roeder had an actual reduction to practice prior to Champaigne’s filing date. The trial court found Roeder’s reduction to practice date was April 14, 1969, but the point has not been argued. Champaigne filed Feb. 5, 1970.

3 . Webster’s Third New International Dictionary (1971):

"adhering or sealing under the influence of pressure alone {pressure-sensitive adhesives are normally in the form of adhesive tapes....).”

4 . We prefer to say in rather than on the PTO. If fraud is perpetrated against anyone, it is the American public which is represented by the PTO. The agency is in no way defrauded and has nothing to lose.

5 . Cyril A. Soans, in "Some Absurd Presumptions in Patent Cases,” 10 IDEA 433, 438 (1966), gave this hypothetical person the name of Mr. Phosi-ta, Person Having Ordinary Skill In The Art, and had more to say about him in his book "Return to Robinson (1890),” (1970) page 61 et seq. Soans, with some justification, was highly critical of the tendency of judges and PTO examiners to ascribe too much omniscience to the mythical Mr. Phosita.

6 . Plaintiff had a claim below for breach of contract on which the court held against it, but that holding was not appealed.