OPINION

MURRAY M. SCHWARTZ, District Judge.

This is an action for patent infringement of United States Patent No. 3,510,391 brought by plaintiff Grefco, Inc. (“Grefco”) against defendant Kewanee Industries, Inc. (“Kewanee”). Kewanee has counterclaimed for a declaratory judgment of patent invalidity, unenforceability and non-infringement. 1 Jurisdiction and venue are conferred on this Court by 28 U.S.C. §§ 1338(a), 1391(b). Trial in this matter began on March 31, 1980 and concluded on April 28, 1980. Post-trial briefing and argument was completed on June 6, 1980. This Opinion constitutes the Court’s findings of facts and conclusions of law required by Rule 52(a), F.R.Civ.P.

1. BACKGROUND FACTS

Grefco, a Delaware corporation and a wholly owned subsidiary of General Refractories Co., was formerly the Mining and Mineral Products Division of Great Lakes Carbon Corp. (“GLC”) and was acquired by General Refractories on or about May 16, 1966. (Stip. No. I). 2 Although at the time of many of the events discussed throughout this Opinion plaintiff was operating as a division of GLC, it will be referred to as “Grefco” unless otherwise indicated.

*846 Kewanee is a Delaware corporation into which Apache Foam Products (“Apache”) was merged effective February 3, 1976. Historically, Apache was purchased by the Millmaster-Onyx Corporation on September 29, 1967 from Air Products and Chemical Company. It was a division of Millmaster-Onyx Corporation until September, 1974, when it was incorporated as a wholly-owned subsidiary of Millmaster-Onyx Corporation. In 1976, Apache was merged into Kewanee and Millmaster-Onyx became an independent financial group within Kewanee. (Stip. No. 2).

U.S. Patent Application Serial No. 638,-629 (“Bolster application”) (PX-1) was filed on May 15, 1967 in the United States Patent and Trademark Office (“PTO”) on behalf of Lyle R. Bolster (“Bolster”), Harlan E. Tarbell (“Tarbell") and Donald W. Mogg (“Mogg”). Armand McMillan (“McMillan”), then assistant patent counsel for GLC, prepared and filed the Bolster application with the knowledge and consent of General Refractories. On September 23, 1969, Patent Examiner William J. VanBalen (“Examiner”) issued an Office Action indicating that all claims of the application would be allowable if amended. The minor changes suggested by the Examiner were then made and the patent issued on May 5, 1970. (PX-1; Stip. No. 32). Grefco owns the Bolster patent.

The Bolster patent is entitled “Perlite Board Bonded to Organic Plastic Foam.” The claims in suit are as follows:

1. A composite board having a U value of not more than 0.2, which comprises

(a) a rigid organic plastic foam layer having a K factor not greater than about 0.4, covered on at least one of its surfaces by

(b) a perlite board having a minimum thickness of 0.6 inch, a K factor not greater than 0.45, a minimum perlite content of at least about 30% by weight and a combustible fiber and sizing content of not more than about 35% by weight, said layer being adhered to said board.

2. The board of claim 1 wherein the other surface of the foam layer is covered by a material selected from the group consisting of perlite board, roofing felt and paper.

3. The board of claim 1 wherein the rigid foam material is selected from the thermosetting and thermoplastic materials within the class consisting of polyvinyl chloride, polyurethane, polystyrene and epoxy resins.

4. The board of claim 1 wherein the rigid foam layer has a thickness within the range of xk to 2 inches.

5. The board of claim 1 wherein the perlite board has a perlite content of 50 to 90% by weight.

6. The board of claim 5 wherein the thickness of the perlite board is within the range of Vi to 1 inch.

7. A composite structure in the form of a sandwich consisting of

(a) a Vi to 1 inch thick base layer of perlite board having a perlite content of about 50 to 90% by weight and a combustible fiber and sizing content of not more than about 30% by weight:

(b) a rigid polyurethane foam layer of xk to 2 inches having a density of about 1.5 to 3.5 pounds per cubic foot, and

(c) a top layer selected from the class consisting of perlite board, roofing felt and paper, said base layer and said board layer being adhered to said foam layer.

8. [Not asserted by Grefco]

9. [Not asserted by Grefco]

10. The composite structure of claim 7 wherein the rigid foam layer is foamed in place.

(PX-2 at 7/19-8/28). 3

The objects of the Bolster patent, as set forth in the patent, are “to provide an insulating structure which overcomes the mechanical and physical disadvantages of plastic foam boards ... [,] to provide a combustion resistant structure based on ur *847 ethane foam .. . [and] to provide an insulating structure of sufficiently low heat conductance so that it can find application in conventional roof building systems, even when an unusually low U value [the overall coefficient of heat transmission through a complete roof assembly (Stip. No. 36c)] is required.” (PX-2 at 2/38-45).

Kewanee manufactures a product known as Millox (DX-A; PX — 370). Although its infringement of the Bolster patent was an issue at the start of the trial, Kewanee admitted infringement midway through the trial. See Tr. 1647. Thus, the issues remaining for disposition are the validity of the Bolster patent and whether Grefco’s conduct before the PTO was fraudulent. 4

II. VALIDITY

A. Presumption of Validity

“[T]he starting point in analyzing a challenge to the validity of a patent is the presumption that the patent is valid, with the burden of demonstrating invalidity by clear and convincing proof resting on the party asserting it.” Aluminum Co. of America v. Amerola Products Corp., 552 F.2d 1020, 1024 (3d Cir. 1977); see also 35 U.S.C. § 282; Tokyo Shibaura Electric Co. v. Zenith Radio Corp., 548 F.2d 88, 93 (3d Cir. 1977); Universal Athletic Sales Co. v. American Gym, Recreational & Athletic Equipment Corp., 546 F.2d 530, 540 (3d Cir. 1976), cert. denied, 430 U.S. 984, 97 S.Ct. 1681, 52 L.Ed.2d 378 (1977). When relevant prior art has not been considered by the PTO, however, the presumption of validity is weakened or overcome. Aluminum Co. of America, supra; Hadco Products, Inc. v. Walter Kidde & Co., 462 F.2d 1265, 1272 n.33 (3d Cir.), cert. denied, 409 U.S. 1023, 93 5. Ct. 464, 34 L.Ed.2d 315 (1972); PIC, Inc. v. Prescon Corp., 485 F.Supp. 1302, 1312 (D.Del.1980); Azoplate Corp. v. Silverlith, 367 F.Supp. 711, 717 (D.Del.1973), aff’d mem., 506 F.2d 1050 (3d Cir. 1974), cert. denied, 421 U.S. 914, 95 S.Ct. 1572, 43 L.Ed.2d 780 (1975). “[T]he degree by which the presumption is weakened depends on a balancing of the pertinence of the newly cited art against the pertinence of the art actually considered by the Patent Office.” Aluminum Co. of America, supra, 552 F.2d at 1025.

Kewanee asserts that the most relevant art was not considered by the Examiner and that therefore the presumption of validity has been overcome. Grefco, on the other hand, contends that the art cited by Kewanee was before the Examiner during his consideration of the Bolster application, or alternatively is no more relevant than the art considered by the Examiner. The Court finds that several important items of prior art under 35 U.S.C. § 102 were not considered by the Examiner and therefore the presumption of validity has been substantially weakened. Specifically, an article in the Journal of Cellular Plastics (“JCP”) (DX-GA) 5 and the Apache brochure (DX-HA), 6 both not before the Examiner, add significantly to the prior art *848 considered by disclosing a combination of urethane and perlite as roofing insulation. 7

B. Anticipation

Under 35 U.S.C. § 102, a patent is invalid if the invention was “anticipated” by the prior art. Under section 102(b), a patent is invalid if “the invention was patented or described in a printed publication in this or a foreign country or in public use or on sale in this country, more than one year prior to the date of the application for patent in the United States.” In addition, section 102(a) renders a patent invalid if “the invention was known or used by others in this country, or patented or described in a printed publication in this or a foreign country, before the invention thereof by the applicant for patent.” Thus, the critical inquiries in the determination of anticipation are the date of invention of the patent and an analysis of the pertinent art available prior to invention or more than one year before the application filing date.

1. Date of Invention

The date of invention is deemed to be the date of the filing of the patent application. U. S. Expansion Bolt Co. v. Jordan Industries, Inc., 488 F.2d 566, 568 n.3 (3d Cir. 1973). The burden rests with the inventor to demonstrate an invention or reduction to practice prior to the time of filing. See generally Rex Chainbelt Inc. v. Borg-Warner Corp., 477 F.2d 481, 487 (7th Cir. 1973); 1 Chisum, Patents § 3.08[3] (1980); 37 C.F.R. § 1.131. This burden is a relatively heavy one, comparable to the burden imposed upon an alleged infringer to demonstrate prior use under § 102(a). 1 Chisum, Patents § 3.08[3]; United Shoe Machine Corp. v. Brooklyn Wood Heel Corp., 77 F.2d 263 (2d Cir. 1935). Moreover, the uncorroborated and undocumented testimony of the patentee is insufficient to prove invention date. See Bell Telephone Laboratories, Inc. v. Hughes Aircraft Co., 564 F.2d 654, 657 (3d Cir. 1977), cert. denied, 435 U.S. 924, 98 S.Ct. 1489, 55 L.Ed.2d 518 (1978); Kardulas v. Florida Machine Products Co., 438 F.2d 1118, 1121 (5th Cir. 1971); Rooted Hair, Inc. v. Ideal Toy Corp., 329 F.2d 761, 767 (2d Cir. 1964).

The Third Circuit Court of Appeals recently set forth the requirements for a reduction to practice:

A reduction to practice ‘is not achieved until the inventor has sufficiently tested the prototype to prove its utility and to determine that no further refinements are necessary.’ In Re Yarn Processing Patent Validity Litigation, 498 F.2d 271, 279 (5th Cir.), cert. denied sub nom., Sauquoit Fibers Co. v. Leesona Corp., 419 U.S. 1057, 95 S.Ct. 640, 42 L.Ed.2d 654 (1974). Considerable confusion exists in the cases as to when an invention has been reduced to practice. It is a legal term of art and the apparent conflict has been ascribed to the differing definitions of the term. See id. at 282. Even when a working model has been produced, this does not necessarily connote a ‘reduction to practice’ within the meaning of patent law. Thus, ‘[w]hile it is true that reduction to practice does not require a device embodying the invention to be mechanically perfect, or a commercial success, its practical efficiency and utility must be demonstrated.’ Van Auken v. Cummings, 49 F.2d 490, 492 (C.C.P.A.1931).

DeLong Corp. v. Raymond International, Inc., 622 F.2d 1135 at 1143 (3d Cir. 1980) (addressing “reduction to practice” of prior use); see also Standard Oil Co. v. Montedison S.p.A., 494 F.Supp. 370 at 379-382 C.A. No. 4319 (D.Del. 1980). The evidence put forth by Grefco to establish an invention date prior to May 15, 1967 must be analyzed against these legal standards.

*849 Grefco contends that products “meeting the terms of claims 1-7 and 10 were made and successfully tested no later than March 1966.” (Plaintiffs Post-Trial Brief at 28). The credible testimony and documentary evidence proffered by Grefco in support of this contention, however, fall short of meeting its burden of establishing a prior invention of the product defined by the patent. While the Court finds that Tarbell made combinations of 1" perlite and urethane and submitted such products to Underwriters’ Laboratories (“UL”) for testing in late 1965 or early 1966 (PX-36, 37, 45), it is not convinced that these products met all the limitations of the patent’s claims 1-7 and 10. For reasons set forth more fully below, see page 865, infra, the Court does not believe that the composite board described in Example 2 of the patent made by Tarbell in January 1966 (Tr. 960) had a density of 3.2 lbs./cu.ft. Grefco produced no documents indicating the production of a composite board with such a density (see Tr. 992), and Tarbell’s uncorroborated testimony, even if credible, would be insufficient to prove prior reduction to practice.

There is evidence (DX-S) that Tarbell made a composite board with a density of 3.95 lbs./cu.ft. in January, 1966. Grefco contends that this embodiment of the patent satisfies the limitation of claim 7 that the layer of rigid polyurethane foam have a density of “about 1.5 to 3.5 pounds per cubic foot.” (PX-2 at 8/15-17). Even assuming that this language permits an upper limit of more than 3.5, see National Research Development Corp. v. Great Lakes Carbon Corp., 410 F.Supp. 1108, 1112 (D.Del.1975), Grefco has not carried its heavy burden of demonstrating that 3.95 lbs./cu.ft. is “about” 3.5. Very little evidence on this issue was presented. (Doc. No. 192 at 39—41, 80-83). The testimony indicated that 3.95 was the “average density” of the product described in DX-S; the lower portion of the foam had a density of more than 3.95, while the upper part had a density of less than 3.95. (Tr. 750-51). The parties have stipulated that foam can be used as insulation if it has a density of 4.0 lbs./cu.ft. or less, that “high density foams” are foams with a density of greater than four lbs./cu.ft. and that “[fjoams used for thermal insulation typically had a density in the 1.8 to 2.2 lb./cu ft range.” (Stip. No. 38; see also Tr. 133). Finally, Tarbell referred to the foam produced in January, 1966 as “higher density” foam made by a “vertical pour process” which would not be utilized in commercial production. (PX-37). The foregoing evidence does not demonstrate to the Court’s satisfaction that a density of 3.95 is “about” 3.5. First, there is no evidence that any portion of the foam produced by Tarbell had a density of 3.5. Second, given the stipulation concerning the range of densities for thermal insulating foam, the Court is unconvinced that foam with a density of 3.95 has sufficient utility and efficiency, as compared to foam with a density of 3.5, so as to bring it within the scope of claim 7. While reduction to practice does not require the “invention” to be mechanically perfect or a commercial success, “its practical efficiency and utility must be demonstrated.” DeLong Corp., supra, 622 F.2d at 1143. For these reasons, the Court concludes that reduction to practice of the Bolster patent did not occur prior to May 15, 1967. This conclusion, as noted supra, brings documents such as the Apache brochure (DX-HA) and UL Construction No. 9 (DX-W) within the framework of § 102(a) prior art, even though these documents were not printed publications more than one year prior to the filing date. 8

*850 2. Analysis of Prior Art

To anticipate a patent, “[a] prior art reference must teach the very invention of the patent,” Tokyo Shibaura Electric Co., Ltd. v. Zenith Radio Corp., 404 F.Supp. 547, 558 (D.Del.1975), aff’d, 548 F.2d 88 (3d Cir. 1977); see Paeco, Inc. v. Applied Moldings, Inc., 562 F.2d 870, 875 (3d Cir. 1977); Congoleum Industries, Inc. v. Armstrong Cork Co., 339 F.Supp. 1036 (E.D.Pa.1972), aff’d, 510 F.2d 334 (3d Cir.), cert. denied, 421 U.S. 988, 95 S.Ct. 1991, 44 L.Ed.2d 478 (1975), or disclose “a device substantially identical to that claimed under the terms of the patent.” DeLong Corp., supra, at 1141 (referencing “on sale” bar of § 102). Further, “it must appear that every material element of the claim in question was disclosed by a single prior art reference.” Mueller Brass Co. v. Reading Industries, Inc., 352 F.Supp. 1357, 1367 (E.D.Pa.1972), aff’d mem., 485 F.2d 1395 (3d Cir. 1973); accord, Paeco, Inc., supra, 562 F.2d at 875.

In Universal Athletic Sales Co. v. American Gym, Recreational & Athletic Equipment Corp., 546 F.2d 530 (3d Cir. 1976), cert. denied, 430 U.S. 984, 97 S.Ct. 1681, 52 L.Ed.2d 378 (1977), the Third Circuit Court of Appeals cited the following standard of anticipation:

For a prior publication to be sufficient to defeat a patent it must exhibit a substantial representation of the invention in such full, clear, and exact terms that one skilled in the art may make, construct and practice the invention without having to depend on either the patent or on his own inventive skills.

546 F.2d at 544, citing Philips Electronic and Pharmaceutical Industries Corp. v. Thermal and Electronics Industries, Inc., 450 F.2d 1164, 1169 (3d Cir. 1971). More recently, in Paeco, Inc., supra, that court stated:

[w]hen a claim incorporates an improvement over a product sold prior to the critical date, which product, except for the improvement, is identical to that described in the claim, the improvement itself must be non-obvious under § 103 or the whole claim is invalid [under § 102(b)],

562 F.2d at 877. Thus, when a patentee makes an obvious improvement over a prior art reference, that reference is still anticipatory because the patentee has not relied on his “inventive skills.”

Kewanee contends that the Apache brochure (DX-HA), the July, 1965 JCP article (DX-GA), and perlite/urethane field installations anticipate the Bolster patent. The Court turns next to these contentions.

a. Apache Brochure (DX-HA)

The Apache brochure states in pertinent part:

CLASS I ROOF DECK APPLICATIONS-Possible Opportunities

At this present moment APACHE does not share in the CLASS I roof deck business; however, as many of you have been told, we expect to have FACTORY MUTUAL rated Class I insulation in the hot too distant future.

I wish to suggest that you may be able to pick up some business in spite of the rating limitation. This may be particularly attractive in areas where moderately high “C” or “U” values are required and while rather acute shortages exist in normal Class I rated insulation materials. Factory Mutual would permit use of APACHE over a minimum thickness of rated materials installed in accordance with Factory Mutual specified procedures. This required thickness is 3A" of Perlite Type, Fiberboard, Fiberglas and Cellular Glass.

IN APPENDIX A I have listed the effective “C” values obtained by combining % " of the several rated materials with standard APACHE thicknesses. For ex *851 ample, one inch of APACHE combined with 3/i" Perlite Type board will give a combined “C” of 0.114 (ditto, Piberboard), and 0.106 for %" Fiberglas.

APPENDIX B lists required thickness of APACHE used with %" of Class I approved materials to give “C” values corresponding to U. S. Department of Commerce No. R257-55.

We suggest that you may find exploration of the combination of products potential very worthwhile.

DX-HA (emphasis added).

Grefco contends that the Apache brochure is not anticipatory solely because it “merely suggests the possible components for several field fabrications” and because “the structures suggested in DX-HA are outside the scope of the claims of the Bolster Patent” which, in its view, is limited to factory-manufactured composite boards. (Plaintiff’s Post-Trial Brief at 31). The Court finds, however, that the brochure is not limited to field fabrications and does disclose the possible combination in a factory of perlite board and urethane foam. Moreover, the Court finds that, contrary to Grefco’s assertions, the Bolster patent encompasses field fabrications of urethane and perlite combinations. The brochure itself, notwithstanding Grefco’s assertions to the contrary, does not either expressly or indirectly limit the product defined by the Apache brochure to a field fabrication of a complete roofing system. In fact, the brochure discusses the “C values” obtained by combining %" board with Apache foam. As stipulated by the parties, C value is “a measure of the thermal conductivity of a given product,” while U value, referred to in the Bolster patent (e. g. PX-2 at 1/13, 2/5, 2/45, 2/60, 2/65, 5/2, 7/19), is “the overall coefficient of heat transmission through a complete roof assembly (including insulation, built-up roofing and the roof deck).” (Stip. No. 36) (emphasis added). See also Tr. 1657-58. Therefore, the use of the term “C value” in the Apache brochure is evidence that it discloses a factory-made product of the type claimed by Grefco to be the subject of the Bolster patent.

To determine whether the Apache brochure, as construed by the Court, anticipates the Bolster patent, the Court must analyze the patent and determine whether the “invention” claimed is taught by the brochure. See page 850, supra. As noted, Grefco contends that the invention claimed in the Bolster patent is limited to a factory-made composite of urethane and per-lite boards, and does not teach the making of a field fabrication. 9 The Court finds to the contrary.

In defining the invention claimed by a patent, the patent’s “claims are to be construed in the light of the specification and both are to be read with a view to ascertaining the invention.” United States v. Adams, 383 U.S. 39, 49, 86 S.Ct. 708, 713, 15 L.Ed.2d 572 (1966). An analysis of the claims and the specification of the Bolster patent, in light of the evidence presented at trial, leads to a finding that both field and factory fabrication of perlite/urethane combinations are encompassed by the Bolster patent. Grefco contends that the terms used in claims 1 and 7 to describe the invention-“composite board” and “composite structure in the form of a sandwich”-limit the invention to factory fabrication. However, Mogg himself conceded that these terms did not take on a meaning in the roofing industry limited to factory manufacture until several years after the Bolster application was filed. (Tr. 811-13; see also Tr. 95-96, 1673; PX-8; Bolster Dep. 270; Maroni Dep. 83-90). Cf. DX-HE (Cham *852 berlain patent for “Sandwich Roofing Element”; invention is defined as “prefabricated” insulating element and patent claims cover “an integral roof deck and insulation panel”); PX-368 (1969 patent application by Coglianese in which invention is described as “a pre-fabricated panel”). Unlike the Bolster patent, the Chamberlain patent draws a distinction between field and factory fabrication at the outset (DX-HE at 1/11-15). Thus, the claims' of the Bolster patent do not limit by definition the types of structures encompassed by the patent. See Funnelcap, Inc. v. Orion Industries, Inc., 421 F.Supp. 700, 706-07 (D.Del. 1976).

An analysis of the specification in light of Mogg’s candid testimony reinforces this finding. The patent contains the following language with respect to the making of the invention:

The polyurethane foam can be formed and stabilized in situ on the perlite board; in this case, it serves as its own adhesive and nó other bonding substance is needed. Alternately, as is the case with other usable plastic foams, ready made board of polyurethane foam can be employed, necessitating recourse to an extraneous adhesive to bond the foam board to the perlite board and to the other facing material that may be used. Any conventional adhesive suitable for the surfaces involved can be used for this purpose, asphalt emulsions and phenolaldehyde adhesives such as phenol-resorcinol-formaldehyde being typical of such well known materials.

PX-2 at 4/27-38. Mogg testified that, although he and his co-inventors intended to refer to urethane foam without felt facings in the specification (Tr. 826, 1234, 1492-93), the term “ready made board of polyurethane foam” includes foam made by Apache (DX-B) and other manufacturers which contained felt facings on both sides. (Tr. 599, 840-42; see also Tr. 1234, 1492-93, 2483, 2650-51). Such polyurethane foam board (DX-B) was one of the four main insulating boards in use at the time of the Bolster application. (Tr. 63, 629, 655). Thus, the patent envisions the use of an adhesive to combine a urethane board with felt facings with a perlite board. (See Tr. 622-23). This is exactly what is done in a field fabrication. See note 9 supra. In fact, Mogg conceded, under questioning by the Court, that “as to where the patent draws the line between a factory and doing it in some cruder fashion, I am not really quite sure.” 10 (Tr. 818; see also Tr. 620). The Court therefore concludes that the Apache brochure, whether or not limited to a field fabrication, anticipates the Bolster patent. 11

b. Journal of Cellular Plastics, July 1965

The July, 1965 Journal of Cellular Plastics (“JCP”) contains an article entitled “Cellular Plastics in Construction Applications and Fire Protection.” (DX-HD). Among the authors of this article is Fred Coglianese, a Kewanee employee, who was with Johns Manville Products Co. at the time of publication. Kewanee asserts that a section of this article, written by Coglia *853 nese (Stip. No. 76; Tr. 1711-12), anticipates the Bolster patent. That section reads:

Roof Deck Insulation-Plastic foams are used [as] roof deck insulation because of low heat transmission, high impact resistance, high compressive strength, low water-vapor transmission and low installation costs. The foams are bonded to metal, wood, concrete or gypsum decks and then topped by a standard built-up roof. In many applications, this system is readily accepted and its use presents no problem. In certain large area industrial applications, however, special precautions are necessary to achieve a system with suitable fire behavior. The major consideration is when a fire exists in an area below the roof. If combustibles in the roof become available fuel, the existing fire is worsened and a spreading, self-perpetuating fire results. Because of their heavy fire resistive nature, concrete and gypsum decks post no problem in this area. Metal decks, however, must be of special construction in order to meet safety requirements.

Several large scale tests have been developed to determine suitable safety standards for metal roof deck assemblies.( * )

See Section titled Test Methods.

DX-HD at 360-61 (emphasis added).

In all of these construction tests, an insulating barrier must be imposed between deck and roof. Plastic foams that offer no resistance under fire conditions would require one inch of fiber board or other barrier over the deck and under the foam in order to qualify.

In conclusion it can be stated that:

a. In many applications, plastic foams offer excellent service in roof decks and are fully accepted.

b. Plastic foams are not suitable for steel deck unless underlaid by a barrier such as fiber board. No problem exists with concrete and gypsum decks.

Once again, Grefco contends that the JCP article “does nothing more than suggest or infer . . . that other unnamed materials could be used in combination with urethane foam in a field fabrication. . What materials could be used is left to the reader.” (Plaintiff’s Post-Trial Brief at 32 (citations omitted)). However, Coglianese, a man conceded by Grefco to be of ordinary skill in the pertinent art of rigid thermal insulation (Tr. 1717, 1723, 1730-39), testified that, although the use of perlite board is not expressly disclosed by the article, “to someone with knowledge, to me” the reference to “fiber board or other barrier” would teach that perlite board and urethane foam could be combined in a steel deck roof assembly. (Tr. 1722). The article itself refers in the footnote to a list of “suitable decks” taken from Section 360 RO of the UL Building Materials List. Construction Nos. 1 and 2, the first two such decks listed in the footnote, may contain perlite board as an approved insulation board, a fact known to those of ordinary skill in the art. (Stip. Nos. 17,18, 21, 22; see also DX-JG). 12 Earlier in the article, it is stated that test roof deck systems listed in Section 360 RO are those which “pass” the UL tunnel roof deck test. (DX-HD at 357). Given the inclusion of perlite board as an insulating component of UL Constructions 1 and 2, the Court concludes that the statement that “plastic foams that offer no resistance under fire conditions would require one inch of fiber board or other barrier over the deck and under the foam in order to qualify” discloses to a person of ordinary skill in the *854 art of rigid thermal insulation the combination of urethane and perlite boards. Finally, for the reasons stated at pages 851-852 supra, Grefco’s attempt to distinguish the JCP article because it teaches a field fabrication fails. Under the standard of anticipation in this Circuit, then, the JCP article anticipates the Bolster patent. See Universal Athletic Sales Co., supra, 546 F.2d at 544; Paeco, Inc., supra, 562 F.2d at 875; Johnson & Johnson v. W. L. Gore & Assoc., Inc., 436 F.Supp. 704, 726 (D.Del. 1977). See also Tr. 3213-16 (Testimony of C. Marshall Dann, former Commissioner of Patents and an expert trial witness for Grefco, that if one skilled in the art would read “fiber board or other barrier” as including perlite, then JCP article discloses combination of perlite and urethane).

c. Actual Field Fabrications of Perlite Board and Urethane Foam

Kewanee presented evidence at trial in an attempt to prove that perlite board was used in combination with urethane foam as roofing insulation in field fabrications prior to the Bolster application. If true, these field fabrications would constitute an invalidating “prior use” of the invention under sections 102(a) and (b). Kewanee, however, has not satisfied its “heavy burden” of proving such prior use. See page 848 supra; 1 Chisum, Patents § 3.05.[2]c. Although several witnesses testified that they were generally aware of such fabrications (Tr. 1668, 2739-44, 2503-04; Bolster Dep. 166; Sullivan Dep. 176-80), there was no direct testimony as to specific applications of urethane foam over perlite board. See PX-378 (Answer to Interrog. No. 202). Kewanee concedes that a partial roof of urethane and perlite constructed at JohnsManville in the mid-1960’s was experimental and not a prior use under section 102. (Tr. 2748). The overall evidence proffered by Kewanee falls far short of demonstrating that field fabrications anticipated the Bolster patent.

C. Obviousness

Even assuming arguendo that the Bolster patent is not anticipated by the Apache brochure and the JCP article, the Court concludes that the invention was obvious from the pertinent prior art and therefore the patent is invalid under 35 U.S.C. § 103. 13

In the landmark case of Graham v. John Deere Co., 383 U.S. 1, 86 S.Ct. 684, 15 L.Ed.2d 545 (1966), the Supreme Court addressed for the first time the tests to be employed by the courts in determining obviousness under § 103. The Court’s often-quoted language is as follows:

While the ultimate question of patent validity is one of law, . . . , the § 103 condition, which is but one of three conditions, each of which must be satisfied, lends itself to several basic factual inquiries. 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. Against this background, the obviousness or non-obviousness of the subject matter is determined. Such secondary considerations as commercial success, long felt but unsolved needs, failure [to] others, etc., might be utilized to give light to the circumstances surrounding the origin of the subject matter sought to be patented. As indicia of obviousness or nonobviousness, these inquiries may have relevancy.

383 U.S. at 17-18, 86 S.Ct. at 694 (citations omitted). See also American Sterilizer Co. v. Sybron Corp., 614 F.2d 890 (3d Cir. 1980); Sims v. Mack Truck Corp., 608 F.2d 87 (3d Cir. 1979); Paeco, Inc. v. Applied Moldings, Inc., 562 F.2d 870 (3d Cir. 1977); Tokyo *855 Shibaura Electric Co. Ltd. v. Zenith Radio Corp., 548 F.2d 88 (3d Cir. 1977).

1. Scope and Content of the Prior Art

Several items of prior art, the Apache brochure and the JCP article, have already been analyzed. At a minimum (even assuming Grefco is correct in its contention that the Apache brochure teaches only a field fabrication and that the JCP article does not disclose the use of perlite in combination with urethane), these publications teach the combination of perlite or fiber board with urethane foam in a field fabrication. Another item of prior art, UL Construction No. 9 (DX-W1, also teaches the operability of a field fabrication of urethane foam and fiber board. In addition, DX-HE, Patent No. 3,111,787, issued to Chamberlain in 1963, discloses the invention of prefabricated insulating roofing elements utilizing a “self-adhering” insulated core of foamed plastic, including polyurethane. (DX-HE at 1/29-46). According to the patent, the foamed plastic core may be adhered to the rigid facings identified (plywood, hardwood board, vegetable fiber board, cement asbestos board, aluminum, steel, gypsum board (DX-HE at 2/10-12)) either by preforming the foamed plastic core over the board or by utilizing a bonding adhesive. (DX-HE at 1/29-32, 2/3-6, 3/46 — 49). DX-GD, a DuPont brochure dated March 29, 1961, discloses the manufacture of wall panels with “cores of rigid urethane foam.” This publication states that “[ajlmost any facing material can be used with rigid urethane foam cores” and that such panels may be made by “the foamed-in-place technique” which “has been discussed in Foam Bulletins of October 28, 1960 and January 16, 1961.” (DX-GD at 1, 2).

Grefco presented evidence that the prior art was concerned with making urethane foam fire-resistant, see Tr. 72-73, 81; PX-81, PX-341; Stip. No. 57(2). The Court finds that there were those in the roofing industry who were attempting to solve the problem of foam’s susceptibility to heat damage (Stip. No. 57(1)) by the formulation of fire resistant foam. However, a second, and eventually successful, focus of the industry was the use of rigid thermal barriers to protect urethane foam from heat damage. It is this body of prior art which renders the Bolster patent obvious.

2. Differences Between the Prior Art and the Claims at Issue

Even if the claims of the Bolster patent are construed to cover only a prefabricated composite of urethane and perlite boards, 14 the differences between the teachings of the prior art and the patent are few. A field fabrication contains an extra layer of roofing felt between the urethane and per-lite boards as well as an adhesive with which to bond the two boards together. See note 9 supra (Tr. 371, 415, 607-08). However, claims 1 and 7 and dependent claims 2-6 specify adhesion of the perlite board to the foam. (PX-2 at 7/28-29, 8/19-21). Since the specification teaches the use of an “extraneous adhesive,” the claims must be interpreted to include a bonding substance, see Tr. 3153, 15 leaving the layer of felt between the urethane and perlite boards as the only difference between the Bolster patent and a field fabrication of urethane foam and perlite board as taught by the prior art. 16

*856 3. Level of Ordinary Skill in the Pertinent Art

In determining the hypothetical person skilled in the pertinent art, the Court must look to those in the industry attempting to solve the problems addressed by the invention and not to the subsequent users of the invention. See Systematic Tool & Machine Co. v. Walter Kidde & Co., Inc., 555 F.2d 342 (3d Cir.), cert. denied, 434 U.S. 857, 98 S.Ct. 178, 54 L.Ed.2d 128 (1977); Universal Athletic Sales Co., 546 F.2d 530 (3d Cir. 1976), cert. denied, 430 U.S. 984, 97 S.Ct. 1681, 52 L.Ed.2d 378 (1977); Komline-Sanderson Engineering Corp. v. Ingersoll-Rand Co., 485 F.Supp. 973 (D.Del.1980). The parties agreed at trial that the relevant art for purposes of determining validity under § 103 is the art of “rigid thermal insulation.” (Tr. 1734-38). Further, Grefco agreed that Fred Coglianese was a man of ordinary skill in this art. (Tr. 1717, 1724-25).

The Court finds that a person of ordinary skill in the pertinent art would be conversant with the chemistry involved in making and bonding foam to fire-resistant barrier boards and would have worked for a company making roofing insulation products. As noted supra, considerable attention was being devoted by the industry to reducing the fire risks attendant upon the use of urethane foam, which was recognized as an excellent thermal insulator. (Stip. No. 57(1)). Many of those in the industry were attempting to combine urethane with other, more fire-resistant barriers to form insulating structures, see Stip. Nos. 57(3), (9), and Johns-Manville had actually installed an experimental partial roof of urethane and perlite before the filing of the Bolster application. (Tr. 2748). The Court finds that these attempts to solve the problem addressed by the Bolster patent (see PX-2 at 1/13-15; 1/38-45; 2/38-45) are very probative of the level of skill in the art of rigid thermal insulation.

4. Analysis

(a) Graham Primary Tests of Obviousness

Under the “practical test of patentability” established by the Supreme Court in Graham, the Court must consider the factual findings made above and reach a conclusion of law as to obviousness. 383 U.S. at 17, 86 S.Ct. at 693. This analysis leads inexorably to a conclusion that the Bolster patent is invalid as obvious under 35 U.S.C. § 103.

As found at page 855 supra, the only possible differences between a field fabrication and a factory-made, foamed in place composite of urethane foam and perlite board are the addition of a layer of roofing felt and adhesive between the two layers in the field. These are not patentable differences. In Paeco, Inc. v. Applied Moldings, Inc., 562 F.2d 870 (3d Cir. 1977), the Third Circuit Court of Appeals addressed the validity of a patent for “replica wooden beams” made of foamed urethane. The only difference between the beams claimed in the patent and prior art beams was the removal of a skin from the surface of the beam so as to facilitate the mounting of the beam to ceilings. The court affirmed the district court’s conclusion that “the idea of removing the skin of a urethane foam beam for the purpose of making it easier to glue the beam to the ceiling would have been obvious to a person with ordinary skill in the art.” 562 F.2d at 878. Similarly, in the instant case, given the testimony and other evidence regarding the known self-adhering properties of urethane foam when applied directly to other boards (see Tr. 103, Sullivan Dep. 45, PX-8, PX-15, DX-HE, DX-IE), the prior disclosure of the “foamed-in-plaee technique” (DX-GD, DX-IP, DX-JK at 235-39), and the efforts of others in the industry to manufacture foamed-in-place composite structures of perlite and urethane in a factory prior to *857 the filing of the Bolster application (Tr. 2476-80, 2511, 2642-45; PX-70), the Court is compelled to conclude that the Bolster “invention,” even if limited to a factory-made composite as arguably defined by claim 10, was obvious.

In addition, the Court finds the use of perlite board in combination with urethane foam obvious from the prior art disclosures of fiber board and urethane composites. Bolster conceded that the industry would have looked, in 1965-66, to materials already rated for use over a metal deck as the most likely candidates for the substrate between a metal deck and a plastic foam. (Bolster Dep. 246; see also Tr. 2644, Maroni Dep. 96-97). These materials were fiber board, fiberglass, foamed glass, mineral wool board and perlite board. (DX-HG, Tr. 69). Mogg and Nash both testified that perlite was even less combustible and more fire-resistant than fiber board. (Tr. 685, 2520; see also Tr. 1061). Moreover, Coglianese, a man of ordinary skill in the art, testified that, to the roofing industry, per-lite and fiber board were “interchangeable” (Tr. 1669-70), and Mogg himself conceded that fiber board was considered equivalent to perlite in the roofing industry. (Tr. 637, 713). Thus, the Court finds it would have been obvious to a person of ordinary skill in the art to combine perlite board with urethane foam in an effort to make a fire-resistant insulating product. 17

(b) Secondary Tests

In Graham, the Supreme Court articulated several so-called “secondary considerations,” such as “commercial success, long felt but unsolved needs, failure of others,” which “may have relevancy” to the determination under § 103. 383 U.S. at 17, 86 S.Ct. at 693. Grefco relies heavily upon these secondary considerations as evidence of nonobviousness. (See e. g., Plaintiff’s Post-Trial Brief at 37, 39, 41-42). While these criteria have often engendered confusion, see generally, Witherspoon, Nonobviousness-The Ultimate Condition of Patentability, at 2:501-18 (1980), the Third Circuit Court of Appeals has articulated the weight to be accorded such factors:

[T]hese secondary considerations are entitled to ‘only measured weight’ in adjudging obviousness, U. S. Expansion Bolt Co. v. Jordan Industries, Inc., 488 F.2d 566, 572 (3d Cir. 1973); Frank W. Egan & Co. v. Modern Plastic Machinery Corp., 387 F.2d 319, 327 (3d Cir. 1967), cert. denied, 391 U.S. 966, 88 S.Ct. 2036, 20 L.Ed.2d 879 (1968), and they cannot, by themselves, support a finding of nonobviousness if it is otherwise established that a patent’s disclosures are obvious in light of the prior art. Continental Can Co. v. Crown Cork & Seal Co., 415 F.2d 601, 602 (3d Cir. 1969), cert. denied, 397 U.S. 914, 90 S.Ct. 916, 25 L.Ed.2d 94 (1970).

Tokyo Shibaura Electric Co. v. Zenith Radio Corp., supra, 548 F.2d at 94-95. Accord, Aluminum Co. of America v. Amerola Products Corp., supra, 552 F.2d at 1025-26.

In the instant case, the Court finds that the secondary considerations reinforce its previous conclusion of obviousness. Apache’s urethane/perlite board product, Millox, first joined the roofing insulation market in 1969. (Tr. 2699, DX-DX, PX-359). According to Nash and Coglianese, there was no long felt need or want at that time for a composite product of this type. (Tr. 2527, 2700-01). This was due to the fact that “very few” specifications required insulation as thick and as thermally efficient as Millox. (Id.; see also Tr. 2508-13). Similarly, the commercial success currently enjoyed by Millox (Tr. 2703-05, Stip. No. 84; PX-360) first came about in 1973 when the energy shortage created a need for more energy-efficient roofing insulation. (Tr. 2702-03). “Commercial success pro *858 vides a basis for inferring nonobviousness only when the success is attributable to the feature which is the subject of the patent claim.” 2 Chisum, Patents § 5.05[2] at 5-247; Marconi Wireless Tel. Co. v. United States, 320 U.S. 1, 35 n. 20, 63 S.Ct. 1393, 1409, 87 L.Ed. 1731 (1943). Contrary to any evidence of “failure of others,” there is evidence of prior independent development of both factory-made composite and field fabrications of perlite and urethane, see pages 854, 856 supra. While this activity may not be § 102 prior art, it is entitled to some weight as a secondary factor showing obviousness. See Graham v. Jeoffroy Mfg., Inc., 206 F.2d 769, 771 (5th Cir. 1953); Schimizzi v. Chrysler Corp., 462 F.Supp. 630, 639 (S.D.N.Y.1978); 2 Chisum, Patents § 5.05[7],

Finally, Grefco relies on both Apache’s repeated praise of its Millox product (see PX — 186, 285, 301, 350, 351) and Coglianese’s unsuccessful attempt to patent a composite of perlite and urethane (see PX-368) as “significant and persuasive evidence that the invention defined in the claims of the Bolster Patent was non-obvious.” (Doc. No. 176, Plaintiff’s Proposed Conclusions of Law ¶ 8). See also Plaintiff’s Post-Trial Brief at 29, 41-42, 45). While there is authority emanating from the Seventh Circuit Court of Appeals supporting this proposition, see Reynolds Metals Co. v. Aluminum Co. of America, 457 F.Supp. 482, 496, 507 (N.D.Ind.1978), rev’d on other grounds, 609 F.2d 1218 (7th Cir. 1979); Tracor, Inc. v. Hewlett-Packard Co., 519 F.2d 1288, 1307 (7th Cir. 1975), this Court is of the view that such evidence is entitled to very little, if any, weight on the issue of obviousness. The primary test to be employed by the courts in determining validity under § 103 is the three-prong Graham test. An alleged infringer’s statements in annual reports and advertisements or its efforts to obtain a patent on the product can not override overwhelming proof of obviousness under both the three-prong Graham test and the other secondary considerations.

c. The Patentability of Combination Patents

As conceded by Grefco, the Bolster patent claims “a composite board made from known materials.” (Tr. 1293-96; see Bolster Dep. 22). Indeed, there can be no dispute that both perlite board and urethane foam were well-known materials in the roofing insulation industry at the time of the Bolster application. There is also no dispute that a person having ordinary skill in this art would have known that urethane was an excellent thermal insulator but was susceptible to damage or degradation when exposed to heat and that perlite board was flame resistant as well as a thermal insulator. Stip. Nos. 57(1), (5). Thus, in combining these two products, the Bolster patent sought to obtain a fire resistant, thermal insulating product. (See PX-2 at 1/13-15, 2/38-45).

Although there is some confusion in the law as to whether combination patents should be analyzed for obviousness under a standard different from that set forth in Graham, see generally Witherspoon, supra, at 7:306-310; 2 Chisum, Patents, § 5.04[5], there is no question that such patents have long been subject to the most careful scrutiny. See Great Atlantic and Pacific Tea Co. v. Supermarket Equipment Corp., 340 U.S. 147, 71 S.Ct. 127, 95 L.Ed. 162 (1950). In Anderson’s — Black Rock, Inc. v. Pavement Salvage Co., 396 U.S. 57, 90 S.Ct. 305, 24 L.Ed.2d 258 (1969), the Supreme Court, in invalidating a patent on a combination radiant heat burner and a paving machine, stated:

The combination of putting the burner together with the other elements in one machine, though perhaps a matter of great convenience, did not produce a ‘new or different function,’ . . . within the test of validity of combination patents.

A combination of elements may result in an effect greater than the sum of the several effects taken separately. No such synergistic result is argued here.

396 U.S. at 60-61, 90 S.Ct. at 307-08. More recently in Sakraida v. Ag Pro, Inc., 425 *859 U.S. 273, 96 S.Ct. 1532, 47 L.Ed.2d 784 (1976), the Court invalidated a patent containing a combination of 13 old elements, and stated:

We cannot agree that the combination of these old elements . . . can properly be characterized as synergistic, that is, ‘result[ing] in an effect greater than the sum of the several effects taken separately.’ Anderson’s-Black Rock v. Pavement Co., 396 U.S. 57, 61, 90 S.Ct. 305, 308, 24 L.Ed.2d 258 (1969). Rather, this patent simply arranges old elements with each performing the same function it had been known to perform, although perhaps producing a more striking result than in previous combinations. Such combinations are not patentable under standards appropriate for a combination patent

Though doubtless a matter of great convenience, producing a desired result in a cheaper and faster way, and enjoying commercial success, [the patent] “did not produce a ‘new or different function’ . within the test of validity of combination patents.” Anderson’s-Black Rock v. Pavement Co., supra, at 60, 90 S.Ct. at 307. These desirable benefits ‘without invention will not make patentability.’ Great A. & P. Tea Co. v. Supermarket Corp., 340 U.S. at 153, 71 S.Ct. at 130. See Dann v. Johnston [425 U.S.] 219 at 230 n. 4, 96 S.Ct. 1393, 1399, 47 L.Ed.2d 692.

425 U.S. at 282-83, 96 S.Ct. at 1537-38. 18

The Third Circuit Court of Appeals recently declined to rule whether a finding of synergism is a precondition to validity in all cases involving combination patents, Sims v. Mack Truck Corp., 608 F.2d 87, 93 (3d Cir. 1979), and in light of this Court’s finding that the Bolster patent is invalid under the Graham test, this issue need not be decided. It is noted, however, that the two elements of the Bolster patent, urethane and perlite, each perform “the same function it had been known to perform.” The “new and useful result” claimed by Grefco-“a unitary product needing no further assembly at the job site that is both thermally efficient and fire-resistant” 19 (Plaintiff’s Post-Trial Brief at 42), while no doubt “a matter of great convenience,” Sakraida, supra, 425 U.S. at 282, 96 S.Ct. at 1537, is not patentable. 20

Application of the three-prong Graham tests, consideration of the secondary factors set forth in Graham and an analysis of the patentability of combination patents all lead to the inescapable conclusion that the Bolster patent is invalid as obvious under 35 U.S.C. § 103.

III. GREFCO’S CONDUCT BEFORE THE PATENT OFFICE

Kewanee contends that Grefco’s prosecution of the Bolster patent before the PTO involved fraudulent conduct, thereby rendering the patent invalid and unenforceable. Its contention addresses essentially two sections of the patent application: misrepresentation or omission of the most pertinent prior art (PX-2 at 1/29-2/36) and misrepresentation and concealment of the results of experiments and tests referenced in examples 1-6 (PX-2 at 4/45-7/8).

*860 A. The Legal Standards

Kewanee has the burden of proving Grefco’s fraud “clearly, unequivocally and convincingly.” In Re Multidistrict Litigation Involving Frost Patent, 540 F.2d 601, 603 (3d Cir. 1976) (hereinafter “Frost”); DeLong Corp. v. Raymond International, Inc., 622 F.2d 1135 at 1145-46 (3d Cir. 1980). There are two necessary elements of a finding of fraud. The alleged infringer must first prove that the omission or misrepresentation was “material” in the sense that it made it “impossible for the Patent Office fairly to assess the patent application against the prevailing statutory criteria.” Frost, supra, 540 F.2d at 604 n.9, citing Monsanto Co. v. Rohm & Haas Co., 456 F.2d 592, 600 (3d Cir. 1972). Second, the patentee must possess “an affirmative intent to deceive or at least a gross negligence or recklessness in misrepresenting the truth.” DeLong Corp., supra, 622 F.2d at 1146.

Participants in proceedings before the PTO are held to high standards of conduct because of the ex parte nature of such proceedings, see generally PIC, Inc. v. Prescon Corp., 485 F.Supp. 1302, 1309-10 (D.Del. 1980), and because of the Patent Examiner’s relative inability to test the veracity of certain representations made to him. Frost, supra, 540 F.2d at 611 n.36; Norton v. Curtiss, 433 F.2d 779, 793-94, 57 C.C.P.A. 1384 (1970). See Hercules v. Exxon, 497 F.Supp. 661, 690 (D.Del. 1980); Tr. 2951-52.

In Standard Oil Co. v. Montedison, S.p.A., 540 F.2d 611 (3d Cir. 1976), the Third Circuit Court of Appeals wrote,

one who prosecutes an application for a patent must meet the highest standards of conduct in dealing with the Patent Office. An applicant who seeks to obtain a patent for his invention is asking the government for the grant of a monopoly which concerns not only other private parties who may also be claiming the same invention, but which involves issues of great moment to the general public as well. The circumstances surrounding such a proceeding must be wholly free from fraud and other inequitable conduct.

540 F.2d at 616.

This Court has recently stated that “[a] party who withholds information does so at its own risk since it is ultimately the duty of the [Patent Office], and not of the applicants and their attorneys, to determine what information is relevant.” Standard Oil Co. v. Montedison S.p.A., 494 F.Supp. 370, 436 (D.Del. 1980). The Court went on to explain,

“[i]t is the province of a patent examiner, and not a party, to determine what data is relevant and to elicit it from the parties. When requested to do so, the parties must supply available information. Although a party may subsequently attack such information as unfair or irrelevant, it cannot attempt to perform the examiner’s function by determining the validity of its own case.”

Id. at 444 (emphasis added).

At the same time, this Court and the Third Circuit Court of Appeals have consistently recognized that patent applicants and their attorneys are not held to an absolute, inflexible duty of disclosure in their dealings with the PTO. In Johnson and Johnson v. W. L. Gore & Associates, Inc., 436 F.Supp. 704 (D.Del. 1977), Judge Wright stated,

The applicant should be accorded the right to exercise good faith judgment. Only when he is guilty of fraud, wilfulness or recklessness indicating a disregard for his duty of candor should enforcement of the patent be barred. Xerox Corp. v. Dennison Mfg. Co., 322 F.Supp. 963, 968-69 (S.D.N.Y.1971). As Judge Stapleton expressed in Tokyo Shibaura, supra [Tokyo Shibaura Electric Co., Ltd. v. Zenith Radio Corp., 404 F.Supp. 547 (D.Del.1975), aff’d, 548 F.2d 88 (3d Cir. 1977)]:

A patent attorney has no absolute duty to perceive that which one skilled in the art would have perceived if left in a room to study a display of all the prior art. His duty is one of candor *861 and this duty leaves room for the exercise of good faith judgment even if that judgment ultimately is held to have been faulty. 404 F.Supp. at 569.

436 F.Supp. at 732.

In DeLong Corp., supra, the Third Circuit Court of Appeals held that

without more the mere misstatement in the patent application that there had not been a prior sale of the invention for more than one year prior to the date of the application, particularly when the circumstances pertaining to the transaction required a sifting of the facts, will not support a finding of fraud

622 F.2d at 1146.

It is against this legal framework that Kewanee’s allegation of fraud must be examined.

B. Prior Art Section

Kewanee contends that McMillan fraudulently misrepresented the state of the art in the prior art section of the Bolster application. Specifically, it cites his alleged failure to disclose that perlite board and rigid urethane foam were among the four leading roofing insulation boards in commercial use prior to the application and his failure to disclose the substance of the JCP article which indicated that fiber board and urethane combinations “qualified” on industry construction tests and which, it has been found herein, constituted a disclosure of perlite board and urethane foam to a man of ordinary skill in the art. Grefco contends that McMillan’s description of the pri- or art was “fair and adequate.”

In the prior art section of the Bolster patent, McMillan did not refer the Examiner to specific pieces of prior art, but rather attempted to summarize developments in the roofing industry and the findings or teachings of several publications. (Tr. 1214, 1235, 1329). The discussion of urethane foam states that “[rjigid plastic foam boards are unusually good thermal insulators and have thus found wide and varied use in industry” (PX-2 at 1/29-31) and that “[ujrethane foam board, in particular, exhibits unusually low thermal conductivity and is for that reason the preferred plastic foam material of the building industry.” (PX-2 at 1/31-34). The section goes on to state that “[tjhese boards, however, have serious drawbacks for use in building construction,” in that their high flammability precludes them from qualifying for use “in the more incombustible types of steel deck roof construction.” (PX-2 at 1/36-44). After describing (and criticizing) attempts to develop more fire resistant foam (PX-2 at 1/44-52), the application notes urethane foam’s “very poor resistance ... to abrasion and other types of physical abuse” and its susceptibility to “thermal degradation and softening when exposed to even moderately elevated temperatures, such as those obtained at the surface of roof insulation board when the board is mopped with molten bitumen.” (PX-2 at 1/52-59). McMillan’s statement with regard to urethane board concludes as follows: “it will be readily seen why the use of urethane board is undesirable for roof insulation even in cases where the combustibility and the smoke generating properties of the urethane might be tolerated.” (PX-2 at 1/59-63).

Mogg testified that the conclusory statement was accurate only with respect to urethane foam boards without felt facings on either side of the foam. (Tr. 654). Indeed, Mogg testified that facings such as roofing felt were placed on urethane foam “to protect the foam from physical abuse or the weather,” thereby addressing the specific disadvantage attributed to urethane foam. (Tr. 85-86; see also, Tr. 2648). 21 *862 There is also evidence that the urethane foam boards with felt facings on both sides, made both by Apache and Barrett, were “a major factor in the market.” (Tr. 2651). Finally, there is evidence that urethane foam board was “one of the four main roof insulation boards” in use prior to the Bolster application. (Tr. 63, 629, 655). It is thus readily apparent that the discussion of urethane foam in the Bolster patent was somewhat inaccurate with respect to the use of such boards in the roofing industry as well as their drawbacks. However, “under the totality of circumstances,” Monsanto, supra, 456 F.2d at 600, Frost, supra, 540 F.2d at 611, the Court does not find that these inaccuracies were fraudulent. Attempts to address the problem of physical abuse were disclosed 22 as were the facts that rigid foam boards were widely used and that urethane foam was the preferred foam material of the building industry. The Court therefore does not believe that McMillan’s description of the use and properties of urethane foam precluded the Examiner from “assessing the patent application against the prevailing statutory criteria.”

With respect to perlite board, the patent application states that this is “another type of board much used in construction” (PX-2 at 1/63-64), praises its resistance to heat, reaction to molten bitumen, and performance on standard flame spread tests (PX-2 1/64-69), and notes the low insulating capacity of perlite board (PX-2 at 1/69-2/7). The discussion of perlite board concludes with the following statement:

The excessive thickness of perlite board that its higher conductivity requires causes building design and construction problems which have generally resulted in the exclusion of perlite board from such applications. In other words, the use of conventional perlite board is generally limited to buildings in which a moderate amount of insulating value is required.

PX-2 at 2/7-14.

While it is true that perlite board was one of the four main roofing insulation materials in use at the time of the Bolster application (Tr. 63, 629, 655), the Court does not find that the discussion of perlite in the prior art section of the Bolster patent fraudulently misrepresents the facts. On the whole, the Court finds the discussion to be reasonably balanced, although perhaps moderately exaggerative with regard to the “limited” use of perlite board.

Kewanee’s final contention with respect to Grefco’s “misrepresentations” of the pri- or art concerns the JCP article. The prior art statement of the Bolster application discussed the use of “dense, incombustible materials” in combination with urethane foam. Those materials listed, taken from a variety of sources reviewed by McMillan (e. g., DX-GA, DX-GC, DX-GD, DX-GE; see Tr. 1217-18), were “asbestos cement board, procelained or enameled steel, aluminum, tempered glass.” (PX-2 at 2/21-23). With regard to the findings of the JCP article, McMillan wrote,

it is interesting to note that plastic foams are considered unsuitable for use in roof deck insulation when a steel deck is employed; fiber board has been suggested as a barrier between the deck and the foam in such a ease. (PX-2 at 2/29-33) (emphasis added). 23

McMillan testified that the information concerning the use of fiber board came from the JCP article (Tr. 1330-31), and that, although the JCP article is nowhere referenced in the application, he intended to disclose “the substance” of the article to the *863 PTO. (Tr. 1338). However, McMillan also testified that his understanding of the statement in the JCP article that “plastic foams that offer no resistance under fire conditions would require one inch of fiber board or other barrier over the deck and under the foam in order to qualify” (DX-GA at 361) was that “whatever the problem was that prevented the use of plastic foams directly on steel decks would be solved.” (Tr. 1336). In other words, while McMillan believed that the problem addressed by the Bolster patent-the inability of urethane foam to be used as roofing insulation over a steel deck-was “solved” by the JCP article, he told the PTO that the use of fiberboard was merely “suggested as a barrier . in such a case.” Moreover, McMillan believed (erroneously) that the reference to expanded perlite in the footnote on page 361 of the JCP article referred to perlite board. (Tr. 1343). He never discussed with the inventors, however, what the term “or other barrier” referred to, despite this belief that perlite board was disclosed by the JCP article. (Tr. 133). While, as recently noted by Judge Wright, courts are “reluctant to find fraud in an attorney’s failure to cite prior art references,” Hercules v. Exxon, supra, 497 F.Supp. at 691, the Court finds that McMillan’s failure to direct the PTO’s attention to the JCP article was fraudulent.

In United States v. Standard Electric Time Co., 155 F.Supp. 949 (D.Mass. 1957), Judge Wyzanski discussed the extent of the duty to disclose prior art:

Of course, a putative inventor must disclose any printed publication which he either knows or believes describes the very invention claimed. United States v. American Bell Telephone Co., 128 U.S. 315, 355-356, 9 S.Ct. 90, 92-93, 32 L.Ed. 450. More than this, if he knows of a printed publication which plainly describes his claimed invention, or comes so close thereto that every reasonable man would say the invention claimed was not original but had been anticipated, then regardless of his personal view that he is the original inventor, he will not be excused for his failure to disclose his knowledge.

155 F.Supp. at 952, cited in Hercules v. Exxon, supra, 497 F.Supp. at 691.

The JCP article, it has been found herein, anticipates or, in the alternative, renders obvious the Bolster patent. See pages 852-854, 856-857 supra. Consequently, McMillan’s failure to cite this publication or to disclose the importance of its findings in the patent application goes to the very heart of the Examiner’s task of determining patentability under the applicable statutory criteria. 24 Thus, with respect to this allegation, the Court finds that all the elements of fraud have been proved by clear and convincing evidence. Frost, supra; Monsanto, supra.

C. Examples

1. Introduction

The Bolster application contained six examples which purported to report the re- *864 suits of various experiments and tests conducted on the claimed invention and materials suggested by the prior art. Examples 1-4 described the results of a “tunnel test” and “gas-air burner flame test” (“gas-air test”) involving composites of perlite board and plastic foam. Example 5 compared the results of a perlite/foam composite with a fiberglass/foam composite on the gas-air test, and example 6 reported the failure of a fiber board/urethane composite on the gas-air test. The purpose of these examples was “to illustrate the invention and the improved properties thereof.” (PX-2 at 4/39-40; Tr. 1245). Following the examples, it is stated:

It stands therefore demonstrated by the results of the foregoing examples that perlite boards are peculiarly successful in composite structures with thermosetting and thermoplastic rigid organic foams. It is equally evident that other materials which have been suggested in this type of application or which might be considered equivalent, have failed to achieve satisfactory performance. (PX-2 at 7/3-9).

The Court finds that all six examples had the purpose of demonstrating the patentability over prior art of the invention claimed in the Bolster application. Although only examples 5 and 6 involved direct testing of perlite composites against fiber board and fiberglass composites, it is clear from a reading of the introductory and conclusory statements regarding the examples and from McMillan’s testimony 25 that all six examples were designed to impress the Examiner with the superiority, and hence patentability, of perlite board composites over prior art substrates. See also PX-2 at 2/53-56.

2. Misrepresentations of Test Results

In example 1, it was reported that a 1" layer of polyurethane foam stabilized on a 1" perlite board “passed the tunnel test and successfully withstood a gas-air burner flame test for more than 10 minutes.” (PX-2 at 5/2-4). In this example, these two tests were described, with the “tunnel test” referring to an Underwriters’ Laboratories (“UL”) 26 test (PX-2 at 5/6-31), and the gas-air test identified as a test “devised [by Grefco (Tr. 691)] to compare the flame resistance of various composite boards.” (PX-2 at 5/32-33). 27

*865 Although the evidence presented by Grefeo with respect to the testing of examples 1 through 4 was often inconsistent, confusing and incomplete, the Court finds that the composite referenced in example 1 was never tested in the UL tunnel test and therefore could not have “passed” such a test. (Tr. 1076-79). Instead, this sample was most likely tested on a so-called “modified bunsen burner test” which was referred to by Grefeo at trial as a “simulated tunnel test.” Tarbell testified that a sample as set forth in example 1 passed the modified test. (Tr. 941). Thus, example 1 is incorrect in its report that the sample “passed the tunnel test.” Since the patent refers only to the tunnel test conducted by UL, as McMillan acknowledged, a reader of the patent could only assume that the sample passed the UL tunnel test. (Tr. 1445-46). Thus, example 1 describes the results of a test that was never run.

With respect to example 2, the Court finds that a sample' of 1" “high density polyurethane” and 1" perlite board was successfully tested at UL. Mogg and McMillan both clearly so testified (Tr. 1163, 1433, 1439, 1449, 1479); Tarbell, however, testified that example 2 was not tested at UL, but rather passed the “simulated tunnel test.” (Tr. 941, 1077). On this point, the Court credits the testimony of Mogg and McMillan over that of Tarbell and finds that example 2, a composite of 1" polyurethane and 1" perlite board, passed the UL tunnel test. (See PX-45 at 05501). 28 There is no evidence, however, that the density of the polyurethane portion of this composite had a density of “3.2 lbs. per cubic foot” as set forth in example 2. Documentation refers only to this sample as consisting of “higher density foam.” (PX-37; see Tr. 992). The parties are agreed, however, that the sample of 1" foam and 1" perlite that passed the UL tunnel test in March, 1966 had a density of 3.95 lbs./cu.ft. (Doc. No. 192 at 39, 80), and it is clear to the Court, despite Tarbell’s denial (Tr. 957), that the sample sent to UL is that referenced in DX-S containing a density of 3.95. 29 Therefore, given the credited testimony that example 2 was actually tested at UL, one can only conclude that example 2 incorrectly states the density of the foam as 3.2 lbs./cu.ft.

The efforts of the parties and the Court to ascertain the circumstances surrounding the inclusion of example 3 in the patent application produced perhaps the most mystery and drama of the trial. Upon review of the evidence, it now appears that a sample of “V*" dry process perlite board, 1" low density polyurethane foam (2.0 lbs./cubic foot) and 5-mil Kraft paper,” as set forth in example 3, was never tested at UL, although that example states that “[t]he resulting assembly . . . passed the tunnel test.” (PX-2 at 20-26). See Tr. 231, 440-42, 446-47, 960-63, 1077-78, 1163. Grefeo did, however, submit samples of Vi" *866 wet process perlite/urethane composites to UL for testing in April, 1967 and in August, 1967. Neither test result was disclosed to the PTO, either in the application or in subsequent discussions between Grefco patent counsel and the Examiner.

In the April, 1967 test, the Grefco sample received a “calculated flame spread value” of 50. (PX — 71). As noted in the Bolster patent, a rating of 25 or under is considered acceptable. (PX-2 at 5/22-24). Grefco contends that this test was invalid because the perlite board was “kerfed” or slit so as to expose the urethane foam directly to heat, and that the inventors therefore properly withheld the test results from the Examiner. 30 The kerf was placed in the per-lite board by Tarbell (Tr. 1108), who testified that he did so at the suggestion of Mr. Schmidgall, Technical Service Manager, Building Products Division, Grefco. 31 (Tr. 1019, 1108). However, it is clear that the purpose of the April, 1967 test was to see if zk" perlite board could be successfully combined with urethane foam. (Tr. 222-23, 553, 980, 1019-21, 1320, 1328; PX-63, 70). Tarbell’s testimony that Grefco ran the UL test on the %" perlite/urethane composite expecting it to fail because of the kerf defies logic and is highly unlikely (Tr. 979, 1108); the entire emphasis at Grefco at the time was to the contrary. Therefore, Grefco’s conduct in running the UL test on kerfed %" perlite board belies the inventors’ protestations at trial that the kerf rendered the entire test invalid.

Grefco’s attempts to excuse its failure to disclose the April, 1967 UL results by relying on doubts as to the test’s validity are further undermined by the fact that the August, 1967 test on a 8/i" perlite/urethane composite, conducted without a kerf, resulted in a flame spread value of 30. (PX-82). While the absence of a kerf in the August test may have reduced somewhat the flame spread value, unkerfed %" perlite board still did not sufficiently protect the urethane foam. 32 Despite these adverse test results on %" perlite composites, the patent application states unequivocally that such a composite “passed the tunnel test.” (PX-2 at 6/25). Thus, the patent Examiner was told the combination of %" perlite and urethane had passed the UL tunnel test when in fact it failed twice.

Based on the evidence, it is also apparent that example 4, which describes a sample of *867 “2" commercial polyvinyl chloride foam slab [adhered to] a 1" dry process perlite board” (PX-2 at 6/35-39), is incorrect in stating that the “composite board . . . was found acceptable” by the tunnel test. (PX-2 at 6/43-45). Such a sample was tested, if at all, only on the so-called “simulated tunnel test” and not on the UL test. (See Tr. 1078-79, 1165).

Thus, the Bolster application states that four examples passed the UL tunnel test, when McMillan and Mogg believed at the time that only example 2 did so, and when Tarbell believed that none of the examples had done so. 33 Moreover, the unsuccessful results of the April, 1967 and August, 1967 tests on %" perlite board were not disclosed to the Examiner.

3. Analysis

As mentioned at page 860, supra, patent applicants are held to a high standard of conduct before the PTO due in part to the Office’s inability to verify independently many of the representations made to it. This is particularly true when the representations pertain to the results of tests or experiments conducted by applicants. Since patent examiners are not equipped to perform their own tests and experiments, they must rely upon the candor and good faith of the applicants in reporting such results. (See Tr. 2951-52). See Norton v. Curtiss, 433 F.2d 779, 794, 57 CCPA 1384 (1970); Hercules v. Exxon, supra, at 690; CPC International, Inc. v. Standard Brands Inc., 385 F.Supp. 1057, 1067 (D.Del.1974); Monsanto Co. v. Rohm and Haas Co., 312 F.Supp. 778, 793 (E.D.Pa. 1970), aff’d, 456 F.2d 592 (3d Cir.), cert. denied, 407 U.S. 934, 92 S.Ct. 2463, 32 L.Ed.2d 817 (1972). This Court and the Third Circuit have therefore been particularly vigilant in requiring patent applicants to disclose all pertinent tests results. See e. g., Hercules v. Exxon, supra, at 698; Frost, supra, 540 F.2d at 609-11; Monsanto, supra, 456 F.2d at 599-600.

In the instant casé, Grefco failed to disclose the two unsuccessful UL tests on %" perlite composites and misrepresented to the PTO that examples 1, 3 and 4 passed the UL tunnel test. At the same time, as previously found, Grefco did inform the PTO of the successful UL testing of example 2. Even assuming, without deciding, that the PTO’s practice of permitting patent applicants to state test results in a patent application when in fact they are only predicting the results of tests that have not actually been conducted, so long as the applicants have based such predictions in good faith upon scientific foundation (Tr. 2946, 2949-54) is permissible, 34 Grefco’s conduct before the PTO still renders the Bolster patent unenforceable. McMillan and the three inventors did not “merely” predict passage of unrun tests; they exercised selectivity in failing to disclose unsuccessful tests (the UL tests of April, 1967 and August, 1967), while disclosing the results of successful tests (example 2) 35 and misled the PTO with respect to the results set forth in example 3. No case has upheld such conduct against allegations of fraud. See Frost, supra, 540 F.2d at 609-10; Mon *868 santo, supra, 456 F.2d at 599-600; Tr. 2949-50. 36

While the co-inventors’ good faith belief excuses their failure to disclose that examples 1 and 4 had not been tested on the UL tunnel test for purposes of a finding of fraudulent conduct, the Court finds that the inventors’ “doubts” as to the validity of the UL tests on %" perlite composites do not excuse their failure to apprise the Examiner of the results of these tests. 37 The doubts relating to the fact that n/i6" board may have been tested has been found to be an afterthought, see note 30 supra, the contention that the 11 minute test affected the test results has been discounted, note 31 supra, and the concerns expressed at trial about the effect of the kerf in the April test are not persuasive in light of the conduct of Grefco in running the test (with the kerf) to see whether %" perlite was in fact operable, see page 40 supra. Second, and most important, any doubts the inventors had about the validity of the tests of %" perlite board composites should have been disclosed to the Examiner for his determination. See Frost, supra, 540 F.2d at 609-10; Standard Oil v. Montedison, S.p.A., supra, 494 F.Supp. at 444. Instead, Grefco chose to usurp the function of the Examiner when it withheld test results because they were, in its view, “inconclusive” (Tr. 804) or invalid. Grefco’s argument to this Court that the April, 1967 and August, 1967 tests support, rather than harm, its theory of patentability would have been better presented to the Examiner, along with disclosure of the results.

The Court finds that the omission and misrepresentation of test results were material to the Examiner’s consideration of the Bolster application. While it is true that examples 1 — 4 did not directly compare the performance of perlite with fiber board or fiberglass as a substrate for use with urethane, and that the prior art solutions were not tested, according to the patent, 38 by the tunnel test method, these examples still were an integral part of Grefco’s overall theory of patentability that its invention worked and protected urethane foam from fire while the prior art “equivalents” did not. The patent itself states, “[i]t stands therefore demonstrated by the results of the foregoing examples that perlite boards are peculiarly successful in composite structures with thermosetting and thermoplastic rigid organic foams.” (PX-2 at 7/3-6) (emphasis added). 39 Clearly, a purpose of examples 1 — 4 was to demonstrate that perlite board of %" and 1" thickness successfully protected urethane. 40

There is no question that the UL tunnel test had great significance in the roofing industry at the time of the Bolster application. See Tr. 66, 638. In fact, the patent itself indicates that the passage of “a new type of urethane foam” on the ASTM E 84 tunnel test made it “suitable for some of the more incombustible types of construction.” (PX-2 at 1/44-49). The Examiner *869 might justifiably have placed greater weight upon the information in the patent that the perlite samples passed the UL tunnel test than he would had he known that the passage was actually on a “simulated test.” Similarly, it cannot be said that the disclosures of the unsuccessful results of the UL test run on the %" boards would have been unimportant to the Examiner. Moreover, had the Examiner been advised that “one inch of fiber board or other barrier” (DX-GA) qualified on industry construction tests and that %" perlite board did not, he might seriously have questioned whether the Bolster patent should be granted with the .6" and %" limits of claims 1 and 7, respectively, and, if not, whether a 1" per-lite board composite was really new in light of the JCP disclosures. Thus, the nondisclosure of the results of the two UL tests on % " board and the misrepresentation of the passage of example 3 prevented the Examiner from fairly assessing the Bolster application against the applicable statutory criteria of utility, anticipation and obviousness.

The Court has no difficulty in finding that Kewanee has met its burden of proving that Grefco’s conduct in its reporting of test results to the PTO was the product of “an affirmative intent to deceive or at least a gross negligence or recklessness in misrepresenting the truth.” DeLong, supra, 622 F.2d at 1146. Tarbell witnessed the April, 1967 test at UL (Tr. 938); there can be no doubt that he was aware of the results prior to the filing of the application. Mogg was also aware, prior to May 15,1967,

of the fact that the %" board received a flame spread value of 50 on the UL tunnel test. (PX-72, 492, 502). McMillan testified, however, that he was not aware of the UL test results on %" perlite board until this litigation. (Tr. 1255, 1457-58). Mogg never told McMillan the results of this test (Tr. 162), and the Court finds, notwithstanding Tarbell’s testimony (Tr. 963-64, 972-78), that Tarbell did not inform McMillan that perlite board had failed the UL tunnel test. 41 There is evidence that McMillan considered the UL results to be important with respect to the patent application; sometime after January, 1967 he wrote a note to himself to “get composite board rating as soon as available.” (DX-GB; Tr. 1309-14,1496-97). McMillan knew that UL tests were planned (Tr. 1307-08, 1497), and told Tarbell of his desire for the UL rating (Tr. 1312), yet Tarbell never provided it to him. Moreover, McMillan apparently never sought to obtain the results of the test immediately prior to writing and submitting example 3, which states that a z/t" perlite board composite “passed the tunnel test.” McMillan testified that had Tarbell reported a “rating of failure,” he would have had to “look at the rating and decide whether there was some valid invention left or not” (Tr. 1328) and might have “closed the whole case.” (Tr. 1320). 42

McMillan was also never informed of the results of the August, 1967 test on %" per-lite composite (Tr. 1255, 1457-58). Mogg and Tarbell, however, were informed of the results shortly after the test was per *870 formed. (See PX-87; Tr. 939). Despite this knowledge, the test results were not disclosed to the PTO. Grefco does not contend that simply because the test results were learned after the filing of the patent application, they need not be disclosed. Such a rule of law would be contrary to the requirement of high standards of conduct in proceedings before the PTO. In Precision Instrument Manufacturing Co. v. Automotive Maintenance Machinery Co., 324 U.S. 806, 818, 65 S.Ct. 993, 999, 89 L.Ed. 1381 (1945), the Court stated:

Those who have applications pending with the Patent Office or who are parties to Patent Office proceedings have an uncompromising duty to report to it all facts concerning possible fraud or inequitableness underlying the applications in issue.

It is immaterial whether the failure to disclose the UL test results of April, 1967 and August, 1967 was due to an affirmative intent to deceive the PTO or to gross negligence on the part of McMillan and the inventors. Consideration of all the evidence presented on this issue, including the materiality of the omissions and representations, see Hercules v. Exxon, supra, 497 F.Supp. at 696, the failure of the inventors to inform McMillan of the results while knowing of his interest in them and McMillan’s failure to secure the results given his belief that the UL results were important, leads to the conclusion that Kewanee has proved, clearly and convincingly, that Grefco’s conduct before the PTO in this regard was fraudulent. 43

IV. CONCLUSION

For all the reasons set forth herein, the Court finds that the Bolster patent is invalid under 35 U.S.C. §§ 102,103 and is invalid and unenforceable because of Grefco’s fraudulent conduct in the procurement of the patent. The Court further finds that Grefco’s conduct before the Patent Office makes this an “exceptional case” under 35 U.S.C. § 285 and therefore, attorney’s fees will be awarded to Kewanee. See Watts v. University of Delaware, 471 F.Supp. 1272, 1285 (D.Del.1979), rev’d on other grounds, 622 F.2d 47 (3d Cir.1980); W. L. Gore & Associates, Inc. v. Oak Materials Group, Inc., 424 F.Supp. 700, 704-05 (D.Del.1976); Chromalloy American Corp. v. Alloy Surfaces Co., Inc., 353 F.Supp. 429, 431 (D.Del.1973); Mueller Brass Co. v. Reading Industries, Inc., 352 F.Supp. 1357, 1380 (E.D.Pa.1972), aff’d mem., 487 F.2d 1395 (3d Cir.1973). If the litigants cannot agree as to the appropriate amount of attorney’s fees, upon receipt of advice to that effect, a hearing will be scheduled for disposition of the attorney’s fees issue.

The parties shall submit an appropriate form of order granting judgment for the defendant on notice within 20 days. 44

. Kewanee has also alleged, in Counts II and III of its amended counterclaim (Doc. No. 166), that Grefco’s enforcement of its patent violates antitrust law and constitutes unfair competition. These counts have been stayed pending resolution of the matters raised by plaintiffs complaint and Count I of defendant’s counterclaim.

. “Stip. No._” refers to stipulations entered into by the parties in the Pre-Trial Order (Doc. No. 167), “Tr.” to the trial transcript, “PX” to plaintiff’s exhibits, “DX” to defendant’s exhibits, and -Dep.” to the deposition testimony of the named witness.

. References to the Bolster patent will be as follows: PX-2 at (column)/(lines).

. The Court’s decision that the Bolster patent is invalid and unenforceable requires judgment for Kewanee on Grefco’s allegation of willful infringement as well. An invalid patent cannot be infringed. See 35 U.S.C. § 282; 4 Chisum, Patents § 19.02 (1980).

. Although McMillan attempted to summarize the conclusions of the JCP article in the prior art section of the application (Tr. 1338-39), the Court finds that the most relevant portions of the JCP article, the information that fiber board and urethane combinations would “qualify" on construction tests, and the reference in a footnote of that article to the use of perlite boards as insulating barriers, were not before the Examiner. McMillan himself testified that he believed the use of fiber board as a substrate in combination with urethane foam solved the problem that had prevented the use of plastic foams directly on steel decks (Tr. 1336), and that perlite board was identified in the footnote in the JCP article. (Tr. 1343). Regardless of whether McMillan’s failure to disclose such information was fraudulent, see pages 862-863 infra, it does affect the presumption of validity.

. The Apache brochure (and UL Construction No. 9 (DX-W)) is a § 102 item of prior art by virtue of the Court’s findings that the date of invention of the subject matter of the Bolster patent was the application filing date of May 15, 1967, see pages 847-849 infra, and that disclosures of field fabrications are pertinent prior art, see pages 851-852, 856-857 infra.

. Grefco contends that the Chamberlain patent (DX-HE), which disclosed, inter alia, fiber board/urethane composites, was before the Examiner since it belongs to Class 161-Subclass 161 which was searched by the Examiner. (PX-1; Tr. 3138). While it is more likely than not that the Chamberlain patent was before the Examiner, see Hercules v. Exxon, 497 F.Supp. 661 at 693 C.A.No.3439 & n.282 (D.Del. 1980), the presumption of validity is still substantially affected by the Examiner’s lack of access to the prior art referred to above. See Aluminum Co. of America v. Amerola Products Corp., 552 F.2d 1020, 1025 n.9 (3d Cir. 1977).

. The Apache brochure (DX-HA) contains a date of May 10, 1966. At trial, the Court admitted this exhibit into evidence as an exception to the hearsay rule, but reserved decision as to whether the May 10 date was accurate. (Tr. 2603). Upon review of the relevant portions of the deposition and trial transcripts (Tr. 2443-48; Bickford Dep. 56-60; 166-74), the Court now concludes that Kewanee has not met its burden of demonstrating that the Apache brochure was first received by a member of the public by May 14, 1966. See Protein Foundation, Inc. v. Brenner, 260 F.Supp. 519 (D.D.C.1966); 1 Chisum, Patents § 3.04[2], However, Grefco does not challenge the characterization of the Apache brochure as a “printed publication,” and, indeed, in light of the testimony regarding the distribution of the *850 Apache brochure (Bickford Dep. 57-58, 167-68; Snyder Dep. 54-59; Tr. 2451-56, 2562-63; DX-DT), it could not successfully do so. See generally 1 Chisum, Patents § 3.04[2]. Therefore, the disclosures of the Apache brochure will be analyzed in connection with the determination of anticipation under § 102 and obviousness under § 103.

. As explained several times by counsel and witnesses (Tr. 413-14, 2578-80, 2715, DX-JT; PX-377; Sullivan Dep. 76-77), a field fabrication of a urethane/perlite combination would require the following steps: placing adhesive over the steel deck of the roof, adhering perlite board to the deck, placing adhesive over the perlite board and adhering a ready made board of polyurethane foam with felt facings (DX-B) to the perlite, and applying the built-up roof (BUR) consisting of asphalt, gravel and additional layers of roofing felt. Construction of a roof utilizing a combination product such as Millox (DX-A, PX-370) or Permalite PK (Grefco’s perlite and urethane combination product) requires adhering the perlite/urethane board to the steel deck and applying the BUR over the Millox or Permalite PK board.

. Mogg testified that if perlite board and “readymade board of polyurethane foam” were glued together with asphalt emulsion “in a manufacturing operation,” it would come within the Bolster patent, despite the presence of an adhesive and a layer of roofing felt (Tr. 617-18), and despite his prior testimony that the patent was not intended to cover “taking a premade urethane board and adhering that to the perlite board with asphalt.” (Tr. 96). As Mogg conceded, however, the patent is silent with respect to place of manufacture of the product. (Tr. 659, 678-79). Moreover, inasmuch as the patent claims a product and not a process for making the product (Tr. 1348, 3142), the distinction consistently urged by Grefco-that between field and factory fabrication-is insignificant. See Johnson & Johnson v. W. L. Gore & Associates, Inc., 436 F.Supp. 704, 726 (D.Del. 1977).

. As noted, Grefco’s sole contentions with respect to the Apache brochure were that it disclosed only a field fabrication and “merely suggested” the use of perlite board. Since an anticipatory piece of prior art must disclose every material element of the patent’s claims, Grefco presumably believes either that the claims’ limitations are not “material” or that the brochure disclosed such limitations to those of ordinary skill in the art. See Stip. Nos. 12, 57(5); PX-2 at 3/10-14.

Suitable decks are listed in Section 360 RO of the UL Building Materials list. Briefly, the following constructions have been accepted:

1. Steel Deck -1" vegetable fibre board-limited, special adhesive-built -up roof.

2. Steel Deck-1" mineral fibre board, etc.

3. Steel Deck-1" expanded perlite, etc.

4. Wood Deck -3A" tongue and grooved Douglas fir plywood, 0.002" aluminum sheet 1" vegetable fibre board, etc.

5. Wood’ Deck-1" tongue and grooved Douglas fir lumber-0.002" aluminum sheet 1" vegetable fibre board, etc.

. Construction No. 3 in footnote (*) refers to “l'A" expanded perlite, etc.”; this is not a reference to perlite board, but rather to a mixture of asphalt and perlite known as “all - weather Crete.” Tr. 386, 1714; DX-JG at 456). McMillan testified, however, that he believed this footnote referred to perlite board. (Tr. 1343). See page 863 infra.

. 35 U.S.C. § 103 reads:

A patent may not be obtained though the invention is not identically disclosed or described as set forth in section 102 of this title, 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.

. It is arguable that claim 7, as limited by its dependent claim 10, excludes field fabrications because of the requirement that the rigid foam layer be “foamed in place.” See Tr. 3156. Assuming this is true, claim 10, is still invalid as obvious for the reasons set forth below. See page 856 infra.

. Mogg and Tarbell both testified that, in their testing of composite products, the use of an adhesive had no effect on the combustibility of the product. (Tr. 670, 681-82, 901.)

. It is noted that Grefco, in the section of its brief entitled “The Differences Between the Claimed Invention and the Prior Art,” devotes the entire discussion to arguing why the Bolster patent does not encompass field fabrications. (Plaintiff’s Post-Trial Brief at 37-39). It does not attempt, however, to rebut Kewanee’s assertion and evidence that the two additional components of a field fabrication as compared to foamed -in -place boards -felt and adhesive-are insignificant. See pages 856-857 infra; Tr. 2711-12.

*856 As noted by the Third Circuit Court of Appeals in In Re Multidistrict Litigation Involving Frost Patent, 540 F.2d 601, 609 (3d Cir. 1976): The advocate’s mind is quick to fasten on a distinction . . . , often forgetting or avoiding inquiry into whether that distinction makes a significant difference.

. In light of the interplay between §§ 102 and 103, the Court’s findings under § 103 that factory production of a urethane and perlite composite was obvious from the prior art field fabrications and that the use of perlite board was obvious from the prior art references to fiber board and urethane combinations renders the Bolster patent invalid under § 102 over the Apaches brochure and the JCP article even if these references do not directly anticipate the Bolster patent. See Paeco, Inc. v. Applied Moldings, Inc., 562 F.2d 870, 877-78 (3d Cir. 1977).

. See also Roamvell Corp. v. Plantronics, Inc., 429 U.S. 1004, 97 S.Ct. 538, 50 L.Ed.2d 617 (1976) (denial of certiorari) (White, J., dissenting):

to be patentable, a combination of elements must produce something more than the sum of the pre-existing elements; there must be a synergistic result that is itself nonobvious. 429 U.S. at 1006, 97 S.Ct. at 539.

. As found supra, the Bolster patent in actuality is not limited to the product described by Grefco. Seepages 851-852 supra.

. Perhaps in an effort to demonstrate “synergism,” Tarbell was asked whether “a composite product of plastic foam and perlite board [has] overall properties not obtainable by the use of its single components.” (Tr. 891). Not surprisingly, Tarbell responded by stating that a combination of perlite and urethane does contain properties not found in either perlite or urethane alone. The combination, he testified, has better insulating capability than perlite and better fire resistance than urethane. (Tr. 891-93). However, there was no evidence that per-lite board and urethane foam, when used in combination with one another, perform a function different from that which they were known to perform individually.

. The prior art section did state that “[t]hin protective skins of thermoset polyester resins containing, for instance, an adduct of a hexahalocyclopentadiene, have been used to increase both the fire resistance and the physical strength of the [foamed urethane or other foamed plastic] cores.” (PX-2 at 2/23-27). Although McMillan steadfastly denied that this information came from Robitschek Patent No. 3,158, 529 (DX-IP), (Tr. 1354-61), it is obvious that DX-IP was the source. See e. g., DX-IP at 7/10-15. That patent is classified at Class 161-Subclass 161. Inasmuch as this class was searched by the Examiner in the prosecution of *862 the Bolster patent (PX-1, Tr. 3138), the Court assumes the Examiner was aware of the Robitschek patent. See note 7 supra.

. These problems were known to those in the industry. See Stip. No. 57.

. The prior art section also referred to plywood/urethane and glass fiber board/urethane combinations. The “Disclosure of Invention” prepared in May, 1965 by Bolster, et al. contemplated the use of fiberglass as a possible facing for urethane foam. (PX-22). However, this idea was subsequently abandoned, most likely because of McMillan’s belief that such a combination was already “known” or “anticipated” in the art. (DX-JP at 64, 67; see also Tr. 715, 1402, 1420).

. This omission is especially material because the JCP article directly contradicts the statement at the conclusion of the examples that “it is equally evident that other materials which have been suggested in this type of application or which might be considered equivalent, have failed to achieve satisfactory performance." (PX-2 at 7/8-9) (emphasis added). McMillan chose to rely solely upon the results of the gas -air burner flame test set forth in example 6 without disclosing to the Examiner the results of the JCP article that indicated that fiber board/urethane combinations would “qualify” on UL and FM construction tests. (Tr. 1350-53). In fact, it was the JCP article which most likely prompted Mogg and his co-inventors to run tests on fiber board/urethane composites so as to “check” the results disclosed in the JCP article in an effort to prove the superiority of perlite/urethane composites. (Seepages 863-864 infra; Tr. 667-68). Clearly, Grefco’s theory of patentability at the time of the filing of the Bolster application was its superior performance in comparison to fiber board and fiberglass. See Tr. 1899- 1903. It is clear then that the omission or misrepresentation of the JCP article was material. See Frost, supra, 540 F.2d at 610 (statement in affidavit that patent invention produced “markedly superior results to those obtained by utilization of any of the solvent materials disclosed in any of the reference patents” found to be “extremely troubling” in light of attorneys’ knowledge that “there existed data casting some doubt on the accuracy of this statement.”).

. McMillan testified that the paragraph quoted above was not considered by him to be part of the discussion regarding example 6, but was rather a summation of the results of all six examples. (Tr. 1297, 1452).

. Underwriters’ Laboratories Inc. (“UL”), a non -profit, independent organization that conducts testing for pubic safety (Stip. No. 47), sets standards for proper roof deck construction and conducts test on steel deck assemblies to determine whether the proposed assembly is safe in the event of a building fire. The specifications of those roof assemblies that pass the applicable roof deck tests are published at regular intervals (e. g. Construction No. 9 (DX - W)). (Stip. Nos. 49, 50).

. Kewanee implies that the gas-air test was devised so as to make perlite look good in comparison to fiber board and fiberglass and that the test was inappropriate for comparisons among building materials because substrates are not subject to direct flames on a roof. See Maroni Dep. 41. Indeed, McMillan’s “interest” in the gas-air test was due to his belief that it was a good predicter of performance in the UL tunnel test (Tr. 1422), but Mogg testified that there was no correlation between the two tests. (Tr. 692-95). This Court, however, has expressed “doubt that the mere act of choosing promising experiments can, by itself, be either misleading or material.” Hercules v. Exxon, supra, 497 F.Supp. at 699; Corning Glass Works v. Anchor Hocking Glass Corp., 253 F.Supp. 461, 472 (D.Del.1966).

Grefco’s determination that “an assembly which resists the conditions of this test for a minimum period of 5 minutes is considered acceptable (PX-2 at 5/44-46) is, though, extremely disturbing. An early draft of the application was silent with respect to passage of the gas-air test. (DX-JP at 114). In response to a letter from Mogg, McMillan inserted the reference to a five-minute period sometime after March 23, 1967. (DX-JP at 24, 114; PX-67). There is no documentation in the record concerning a five-minute minimum period of passage. McMillan’s testimony was that “we somehow or other arrived at the five-minute level because obviously if it lasts long enough, everything will be destroyed.” (Tr. 1519). The “decision” to adopt a five-minute passage period is especially curious in light of Tarbell’s contemporaneous (November 22, 1966) statement that samples were tested on the gas-air test for ten minutes “as this is the time a material must withstand a flame to be approved by Underwriters’ Laboratories when *865 tested by their ‘Tunnel Flame Spread Test.’ ” (DX X). Further, William Maroni, manager of the materials section of Factory Mutual, whose duties include the “testing and qualification of building materials” (Maroni Dep. 48), testified that a five minute test was insufficient to predict the performance of an insulation board on a roof. (Maroni Dep. 43). Were Grefco’s choice and use of the gas-air test the only allegation of fraudulent conduct raised by Kewanee, the Court would be hard-pressed to find fraudulent intent and materiality in light of the disclosure in the application that a period of five minutes is considered acceptable. Presumably, the Examiner could have decided to request ten-minute tests of the samples if he doubted the validity of a five-minute test. However, since the Vi" perlite board of example 3 lasted only seven minutes under the gas-air test and therefore would have failed a ten-minute test, the decision to deem five minutes “acceptable” impacts on the overall conclusion as to the results of the examples and on Grefco’s theory of patentability. As such, the Court finds it to be an additional indicium of fraudulent intent by the inventors and their attorney.

. It is interesting to note that Tarbell’s testimony so totally confused Mogg and McMillan, both of whom sat through Tarbell’s testimony, that, after hearing Tarbell, neither was certain as to what, if anything, in examples 1-4 had actually been tested at UL. (Tr. 1164, 1248, 1254).

. The Court noted at the time that Tarbell’s testimony with respect to the sample in example 2 was “virtually worthless” because of his poor recollection. (Tr. 959).

. Grefco also presented evidence that the samples tested at UL were actually 1 Vie" perlite boards. (See Tr. 177). A stipulation entered into by the parties states, however, that the sample was 3A" perlite (Stip. No. 80), and the plaintiffs Proposed Findings of Fact reference 3A", not u/w" perlite. (See Doc. No. 176 ’' 145-49, 151-53). Although at trial, counsel for Grefco acknowledged he was bound by the stipulation (Tr. 185), at oral argument he asked the Court to make a factual finding that the perlite board tested at UL in 1967 was “/is" thick. (Doc. No. 192 at 84). Since the stipulation and Grefco’s proposed findings (filed prior to trial) led Kewanee to believe that the samples were in fact 3A" thick, the Court adheres to its ruling at trial that Grefco was bound by the stipulation. (Tr. 185). In any event, there was testimony by Mogg that he and his co-inventors were not disturbed that the board was “/is " “because we felt that a data point in the tunnel test [at] 1 Vt6th would be just as valuable to us as the data point at three-quarters of an inch and therefore, we didn’t ask the plant to reship board that was within spec thickness. We went ahead with the “/isth inch board.” (Tr. 191; see also Tr. 544). Thus, any contention by Grefco that the inventors believed, in good faith, that the 1967 UL test results were invalid partly because the perlite board was not 3A" thick is a mere afterthought.

. Tarbell’s testimony with respect to the kerf is apparently confused. Funk testified that he fired Schmidgall in 1966. (Tr. 2094). Tarbell sent the 3A" perlite composite to UL in March, 1967 (DX-AH). How he could have put the kerf in the composite at the suggestion of Schmidgall when Schmidgall was fired a year prior to that is unclear.

. Grefco argues that this test was also invalid because “Ae" board was tested and because the test was incorrectly run for 11, and not 10, minutes. The Court has already addressed the ‘Vis" contention in note 30 supra. Mogg testified that even if the test had been run for 10 minutes, the flame spread value would have been 28.2 which, according to UL procedure, would have been rounded off to 30. (Tr. 199). Thus, even if the 3A" perlite board were tested for 10 minutes without a kerf, it still would have failed the UL tunnel test. In fact, in November, 1967, Mogg stated that in “the UL tunnel tests to date on composite board . 3A" of perlite failed.” (PX-87 at 01939).

. McMillan could not deny this allegation; he testified as follows: “I am saying [in the patent] that four examples passed the tunnel test when I only knew that one passed the UL tunnel test.” (Tr. 1449-50).

. The Court can conceive of no reason for the PTO to countenance such a practice. In effect, the PTO is permitting itself to be misled by patent applicants during the process of granting a monopoly. Moreover, the public is misled by such misrepresentations upon the publication of the patent. Nonetheless, this unexplained PTO procedure would preclude a finding of fraud in the Patent Office were it not for Grefco’s omissions and misrepresentations as to the passing of certain tests. Whether and under what circumstances PTO countenanced misrepresentations could lead to judicial invalidation of a patent is left to another day.

. The Court has already found that the failure to disclose the important findings of the JCP article was fraudulent, particularly in light of the disclosure in the patent application that fiber board failed to qualify for use as a substrate when tested by the gas air method in example 6. See pages 862-863 supra.

. Grefco argues that Frost and Monsanto, in which the Third Circuit Court of Appeals found fraud in the selective reporting of favorable test results, are “readily distinguishable” in that the applicants in those cases submitted data to the PTO showing the superiority of the claimed invention over the prior art. As found herein, however, see page 864 supra, the purpose of all six examples of the Bolster patent was to demonstrate the superiority of perlite over the prior art solutions of fiberglass and fiber board. Inasmuch as the omissions and misrepresentations pertaining to the test results were material to this contention, see pages 868-869 infra, Frost and Monsanto are highly pertinent, if not controlling, cases.

. There is no question that a 1" perlite/urethane composite had actually passed the UL tunnel test by the time of the patent application. (PX-45). See page 865 supra. With respect to example 3, however, despite the passage of 3/t" perlite on the “simulated tunnel test” (Tr. 942), Grefco’s prediction of passage on the UL tunnel test cannot be condoned in light of the two failures when actually tested on the UL test.

. Tarbell testified that he ran “simulated tunnel tests” on fiber board and fiberglass, but did not recall the results. (Tr. 902, 999).

. See note 25 supra.

. If examples 1-4 were really as immaterial to patentability as now urged by Grefco, there presumably would have been no reason for their inclusion in the application.

. McMillan testified that if anyone had told him the results, it would have been Tarbell. McMillan’s testimony that he never received the results of the test at UL is credible in light of PX-59. In this document, Mogg wrote to Carl Peters, Head of the Patent Department at GLC, on January 5, 1967, that he was removing McMillan from “our distribution list for future reports and further information concerning our research activities and results will be supplied to you only on a need-to-know basis.” The inference to be drawn from this is that either Mogg or his co-inventors determined that Peters or McMillan did not have a “need-to-know” the April, 1967 test results.

McMillan also does not recall being informed by the inventors that they were dissatisfied with the procedures employed at UL or that the testing done at UL was defective. (Tr. 1314— 15; see also 494-95). This testimony tends to undermine Grefco’s contentions that the inventors felt that the UL tests were invalid and that their “good faith belief” excused their failure to disclose the tests results. Mogg knew he had a duty to bring even unfavorable test results to the attention of Mr. McMillan (Tr. 492-93), but did not.

. McMillan also testified, under questioning from the Court, that had he known that *h" perlite board failed the UL tunnel test he “would not have said ‘passed the tunnel test’ ” in example 3. (Tr. 1458).

. Kewanee also alleges that Grefco’s failure to disclose that a sample of 'h“ perlite board/urethane failed the UL tunnel test was fraudulent. (See PX-14 at 05481). There is no question McMillan knew about this test and deliberately omitted it from the patent. (Tr. 1302, 1428). While the Examiner probably could not have approved the Bolster patent with a lower limit of .6" thick had he known that 'h“ and 3A" thickness had both failed the UL tunnel test, the Court does not believe that the information that Vi" perlite board was unsuccessful would have added anything to the Examiner’s deliberations with respect to patentability if he had known 3/t" thickness failed. See Norton Co. v. Carborundum Co., 530 F.2d 435, 444 (1st Cir.1976) (“[t]he scope of a patent defines not only its enforceability but also the limits of materiality of an allegation of fraud in the procurement of the patent.”); Corning Glass Works v. Anchor Hocking Glass Corp., 253 F.Supp. 461, 474-75 (D.Del.1966), mod’d, 374 F.2d 473 (3d Cir.), cert. denied, 389 U.S. 826, 88 S.Ct. 65, 19 L.Ed.2d 80 (1967).

. Although the resolution herein renders unnecessary a determination with respect to Kewanee’s remaining allegations of invalidity-defective oath, laches and estoppel, the Court finds that Kewanee has failed to prove all the essential elements of each of these defenses. See Tr. 1393, 1499-1500, 1260-82 (defective oath allegation); Jenn Air Corp. v. Penn Ventilator Co., Inc., 464 F.2d 48 (3d Cir. 1972); Johnson & Johnson v. W. L. Gore & Associates, Inc., 436 F.Supp. 704, 733 34 (D.Del.1977); Minnesota Mining & Mfg. Co. v. Berwick Industries, Inc., 373 F.Supp. 851, 862 (M.D.Pa. 1974), aff’d, 532 F.2d 830 (3d Cir.1976); Tr. 2218-25, 2250-58, 2286-2300, 2311-17 (failure to prove prejudice to Kewanee as a result of *871 Grefco’s delay in filing infringement suit); Id. (failure to prove detrimental reliance by Kewanee upon action by Grefco in not enforcing patent).