Full opinion text
OPINION McKELVIE, District Judge. This is a patent case. Plaintiff LNP Engineering Plastics, Inc. is a Delaware corporation with its principal place of business in Exton, Pennsylvania. Plaintiff Kawasaki Chemical Holding Co., Inc. is a Delaware corporation with its principal place of business in Wilmington, Delaware. Kawasaki is the owner of U.S. Patent Nos. 4,559,262 (the ’262 patent); 5,019,450 (the ’450 patent); and 5,213,889 (the ’889 patent). LNP is the exclusive licensee of much of the technology covered by the three patents at issue. Defendant Miller Waste Mills, Inc. is a Minnesota corporation with its principal place of business in Winona, Minnesota. Miller Waste Mills trades as RTP Company. On September 16, 1996, LNP filed a complaint alleging that RTP infringes the ’262 patent, the ’450 patent, and the ’889 patent. LNP amended its complaint on December 6, 1996, adding Kawasaki as a plaintiff. On December 22, 1997, the court denied various jurisdictional challenges to the suit. On November 5, 1998, the court issued its Markman Opinion construing disputed claims of the patents at issue. The case proceeded to a nine-day jury trial beginning on November 12, 1998. The jury returned a verdict finding the three asserted claims invalid and noninfringed. Thereafter, both parties filed post-trial motions. This is the court’s decision on the motions. I. FACTUAL AND PROCEDURAL BACKGROUND The court finds the following facts from the patents at issue, their prosecution history, and the evidence presented at trial. A. The Parties Imperial Chemical Industries, PLC (“ICI”) is based in London, England. Under ICI’s employ, Frederic N. Cogswell, David J. Hezzell, and Peter J. Williams developed a number of inventions relating to fiber-reinforced plastics. The inventors applied for United States patents, and the United States Patent and Trademark Office (PTO) issued the ’262, ’450, and ’889 patents on December 17, 1985; May 28, 1991; and May 25, 1993, respectively. The three patents were assigned to ICI. On October 3, 1991, ICI granted Kawasaki a license to make, use, or sell the technology covered by the ’262 and ’450 patents. On December 1, 1992, Kawasaki granted LNP a license to develop this same technology. On September 14, 1995, ICI assigned its interest in these three patents to Kawasaki and filed a “Notice of Recordation of Assignment Document” with the PTO. And on August 1, 1995, Kawasaki and LNP amended their licensing agreement to include rights to use the technology claimed in the ’889 patent. Miller Waste Mills manufactures and sells engineering plastic products. Miller Waste has previously conducted business under the name Fiberite Corporation. Fi-berite was acquired in 1980 by Beatrice Chemical, which had previously acquired LNP. Because of antitrust concerns, in 1981 the government forced Fiberite to divest itself of its long fiber reinforced thermoplastics (LFRTs) division, RTP. RTP thus became an independent company, and it retained the intellectual property rights of Fiberite relating to LFRT production. The trial testimony mentions two other firms that produce LFRTs. Polymer Composites, Incorporated (“PCI”), a Minnesota company, was founded by a Fiberite employee named Ronald Hawley. PCI was acquired by Hoechst Celanese in 1988. Fiberfil Company, based in Evansville, Indiana, was a division of Dart Industries. Fiberfil employed inventors named Kiyo-shi Hattori and Rexford Bradt, whose patents constitute prior art for the patents in suit. In 1981, Fiberfil merged with Wilson Products to become Wilson-Fiberfil, and then changed, its name to DSM Engineering Plastics. B. The Technology 1. Development of Long-Fiber Reinforced Thermoplastics All three patents at issue pertain to a special type of plastic known as long fiber reinforced thermoplastic, or LFRT. As explained at trial by Frederic Cogswell, one of the inventors of the claimed technology, thermoplastics are materials that become malleable when heated and that can be repeatedly reprocessed. These materials can be reinforced with fibers, typically glass or carbon, to render the plastic stiffer and stronger. Ultrastrong LFRTs are employed in applications such as the flooring of military aircraft, where they must withstand the impact of ammunition cases and infantrymen’s boots. LFRTs also have more common uses, such as in snowshovel handles and manhole covers. Due to the thermoplastic qualities of LFRTs, the plastics can be salvaged from manufactured goods and reprocessed. Two factors help determine the stiffness of reinforced plastics: the length of the fiber filaments, and the degree to which the filaments are “wetted” by the thermoplastic resin. “Wetting” refers to the extent to which the plastic resin surrounds the individual reinforcement filaments. According to Cogswell, fiber-reinforced plastics have been employed since the 1950s. He explained that long-fiber plastics were known, but that the fibers were poorly wetted, with bundles of filaments just coated with polymer. Short-fiber plastics were also known, and although the filaments therein were well wetted, the shortness of the fibers did not allow for sufficiently strong products. Cogswell testified that the artisans of the time believed that well-wetted long fiber reinforced plastics could not be made, and that, even if they could be made, the resultant plastics would have bad properties, such as brittleness and the inability to pass smoothly through an injection molding machine. Cogswell testified that he and his colleagues developed new techniques to achieve highly wetted long-fiber reinforced thermoplastics. The technology at issue involves impregnating strands of fiber with thermoplastic resin. The strands, which are also known as rovings or rods, comprise thousands of filaments. Each filament is between 10 and 25 microns in diameter, barely visible to the naked eye. The strand is drawn through a “pultrusion chamber” which contains melted plastic resin. To work the resin into the filaments, the strand is pulled over spreader surfaces. When pulled through the pultrusion chamber in such a fashion, the resin is forced into the strand, and the resin “wets” the filaments. The more resin that surrounds the filaments, the more thoroughly “wetted” the filaments become, and the greater the stiffness exhibited by the resulting plastics. Once the wetted strand emerges from the pultrusion chamber, it is chopped into pellets. The pellets claimed by the inventors range in length from 2 mm to 100 mm. The pellets are subjected to a melt homogenization process, wherein they are heated, melted, and mixed. The melted plastic is then injected into molds. 2. How is “wetting” measured? The patents at issue claim plastics products wherein the filaments are highly wet-ted. The question of how to determine whether filaments are sufficiently wetted forms the basis of one of the primary disputes of this case. There are a number of techniques, direct and indirect, to determine the degree of wetting of reinforcement filaments. a. direct method: visual inspection A direct method of measuring wetting is to cut a pellet with a knife and to examine the sliced pellet visually. A microscope can be used to inspect the pellet. The more thoroughly wetted the fibers, the fewer loose, unwetted fibers that will be visible. Under a microscope, it can be determined whether polymer surrounds the individual filaments. An advantage to this technique is that it provides a direct, visible measure of how thoroughly wetted the individual fibers are. A disadvantage is that there is some subjectivity in how the fibers are viewed and in how the wetting is evaluated. Moreover, visual inspections are relatively cumbersome to perform. b. indirect method: flexural modulus test An indirect method for assessing the degree of wettedness is to measure the “flexural modulus” of a molded product. The flexural modulus test is a more convenient way of measuring stiffness than a truly scientific measurement, and can be readily performed in the laboratory. Once the pelletized polymer is melted, injected into a mold, and hardened, the resulting product can be subjected to a flexural modulus test. There is a standard protocol recognized by the industry, labeled ASTM D790-80, that stipulates how to perform the analysis. The test requires the use of a standard rectangular bar of the specimen LFRT. The specimen bar is bent, and the amount of force necessary to flex the bar is recorded. There is a direct correspondence between the flexural modulus of a specimen and the degree of wetting of the reinforcement fibers therein. Flexural modulus measurements are expressed as a percentage. The measured flexural modulus is divided by a theoretical stiffness that would be obtained if the fibers of the specimen bar were 100% wetted. Specimens, for example, may be characterized as exhibiting 90% of the theoretically attainable flexural modulus. The flexural modulus test has a disadvantage. Because the test requires that a strand be chopped into pellets and then injection molded into a specimen bar, the filaments are shorter than they are in the strand. Also, the fibers are not aligned in the same direction, as they are in the unehopped strand. As such, the specimen bar will be weaker than the unchopped, wetted strand, and the flexural modulus as measured by the standard ASTM D790-80 protocol will be significantly lower. c. indirect method: dispersal/length test A second indirect method of determining the degree of wettedness of the fibers is to examine the dispersal of the individual filaments and the distribution of the lengths of the filaments in molded articles. The court will refer to this method as the dispersal/length test. The first aspect of this test involves an analysis of the spatial distribution of filaments in a product that has been injection molded from LFRT pellets. When each filament is individually wetted in the pellet, the filaments purportedly do not clump during the melt homogenization process, and the filaments will be randomly dispersed when the plastic is molded into a product. To examine the distribution of fibers in a molded product, the product is cooked in an oven such that the polymer component of the LFRT is “ashed,” leaving behind the reinforcement filaments. The filaments can be visually inspected to determine whether they have been randomly dispersed throughout the product. If a rectangular bar is used as a specimen for the ashing test, the fibers will have a slight bias to orient themselves along the length of the bar. Aside from this bias, however, filaments that were completely wetted in the pellets should be randomly dispersed throughout the bar. A second aspect of the dispersal/length test requires an analysis of the length of the filaments in a product that has been injection molded from the LFRT pellets. This aspect of the test derives from an allegedly surprising discovery of Cogswell and his colleagues. Cogswell testified that long-fiber reinforced plastics had commonly been found to work poorly in injection molding processes. Injection molding involves forcing melted, mixed plastics through a small hole — often approximately 2 mm wide. Cogswell testified that when long filaments that have been poorly wet-ted are forced through the channels of an injection molding apparatus, the bundles of filaments scratch the machine and the filaments break into shorter pieces. According to his testimony, thorough wetting of the filaments solves these problems. By surrounding the filaments with polymer, the filaments supposedly become lubricated, which allows them to pass through the injection molding apparatus without scratching the machine and without breaking. Under this theory, the degree of wet-tedness of reinforcement filaments can be determined by analyzing the length of the filaments in a product that has been injection molded from LFRT pellets. To determine the length of the fibers in molded products, a molded specimen is cooked in an oven, which “ashes,” or removes, the polymer. A filament skeleton remains. With the aid of a computer, the fibers can be counted, measured, and sorted. The computer may sort the filaments by length, for example, separating all filaments with a length of 2 mm or greater. When the plastics claimed in the present invention are molded into a specimen, they are characterized as having at least 50% by weight of the filaments retaining a length of greater than 2 mm, or having a “weight-average length” of at least 2 mm. C. The Patents LNP and Kawasaki have asserted claims of three patents against RTP. All three patents relate to the same technology and contain the same specification. The ’262 patent claims a fiber strand that is wetted with polymer; the ’450 patent, as reexamined, claims pellets chopped from the wet-ted fiber strand; and the ’889 patent, as reexamined, claims a molded article formed from the reinforced thermoplastic. Following are the patent claims relevant to the present dispute: 1. Claim 1 of the ’262 patent A thermoformable, fibre reinforced structure comprising a thermoplastic polymer and at least 30% by volume of parallel aligned reinforcing filaments having a diameter of up to 24 microns said filaments having been wetted by the said polymer in a continuous melt protrusion process so as to give a longitudinal flexural modulus determined by ASTM D790-80 of at least 70% of the theoretical attainable flexural modulus. 2. Claim 8 of the ’262 patent A thermoformable fibre reinforced structure comprising a thermoplastic polymer and at least 30% by volume of parallel, aligned reinforcing filaments having a diameter up to 24 microns said filaments being substantially completely wetted by thermoplastic polymer characterized in that when the structure is chopped into moulding pellets between 2 mm and 100 mm in length and formed into a shaped article by a process which includes the step of subjecting the pellets to a melt homogenisation process to produce a random distribution of individual filaments in molten polymer, the fibre length is retained to the extent that at least 50% by weight of the filaments are at least 2 mm long. 3. Claim 1 of the 450 Patent, as Reexamined Pellets of reinforced thermoplastic material containing at least 30% by volume of parallel, aligned reinforcing filaments between 2 and 100 mm in length, the filaments extending through the length of the pellets, the pellets having been cut from a continuous reinforced product prepared by melt protrusion in which the filaments have been substantially completely wetted by a molten thermoplastic material, and which pellets can be injection molded into an article in which the filaments are present in the form of randomly dispersed individual filaments at least 50% by weight of the filaments of the pellets retaining a length of greater than 2 mm in the molded article. 4. Claim 6 of the 450 Patent, as Reexamined Pellets according to claim 1 which have been cut from a continuous reinforced product in which the individual filaments of the product have been wet-ted to the extent that the longitudinal flexural modulus of the product as determined by ASTM D790-80 is at least 70% of the theoretically attainable flexural modulus. 5. Claim 1 of the ’889 Patent, as Reexamined A molded article formed from a fibre reinforced thermoplastic composition in a process which includes the step of melting and homogenizing a composition containing at least 30% by weight of fibre reinforced pellets between 2 mm and 100 mm long which pellets have filaments extending the length of the pellet, characterized in that, the molded article contains reinforcing filaments in the form of individual filaments and at least 50% by weight of the filaments in the pellets being present in the molded article at a length of greater than 2 mm, the pellets having been cut from a structure of continuous, parallel, aligned, reinforcing filaments which have been substantially completely wetted by a molten thermoplastic in a melt pultrusion process. There are a few important distinctions to be made among the above claims. First, three of the claims — claim 3 of the ’262 patent; claim 1 of the ’450 patent, as reexamined; and claim 1 of the ’889 patent, as reexamined— describe the fibers as being “substantially completely wetted.” These three claims, in order to provide a measure for the term “substantially completely wet-ted,” recite the dispersal/length test. Claim 1 of the ’262 patent does not recite the term “substantially completely wetted,” and recites the flexural modulus test as the measure for determining the degree of wettedness of the reinforcement fibers. Claim 6 of the ’450 patent, being dependent on claim 1, necessarily incorporates the dispersal/length test as the measure of wettedness, but also specifies the flexural modulus test as a further limitation. The three patents at issue share a common specification. The flexural modulus test is mentioned repeatedly therein. The specification states that “[t]he efficiency of a particular process in wetting the fibres ... may be assessed by measuring the extent to which the process provides a product having a flexural modulus approaching the theoretically attainable flex-ural modulus.” D. The Lawsuit On September 16, 1996, LNP filed a complaint alleging that RTP infringes unspecified claims of the ’262 patent, the ’450 patent, and the ’889 patent. LNP amended its complaint on December 6, 1996, adding Kawasaki as a party plaintiff. On January 28,1998, RTP filed its answer and counterclaim. RTP asserted that it was not subject to personal jurisdiction in this court, that venue in this court is improper, and that LNP lacks standing to sue under the asserted patents. RTP further contended that it does not infringe any of the patents at issue, that the claims are invalid under §§ 101, 102, 103 and/or 112 of the Patent Act, that LNP’s patented products were inadequately marked, that the suit is barred by laches, that RTFs reliance on LNP’s actions and inactions estop plaintiffs from bringing suit, and that RTP has acquired intervening rights under 35 U.S.C. § 307. RTP also counterclaimed that the claims are invalid, unenforceable, and not infringed for all the reasons set forth in its affirmative defenses. In its prayer for relief, RTP requested that plaintiffs’ complaint be dismissed with prejudice, that LNP’s counter-defenses be denied and judgment entered for RTP, that the court enter a declaratory judgment that the claims at issue are invalid and unenforceable, that the court enter a declaratory judgment that the claims at issue have not been infringed by RTP, and that the court declare that this case is exceptional and warrants an award of costs and fees. The case was originally assigned to Judge Joseph J. Longobardi. Upon the retirement of Judge Longobardi, the case was reassigned to Judge Roderick R. McKelvie on June 25,1997. In response to the various jurisdictional challenges brought by RTP, the court issued a Memorandum Opinion on December 22, 1997. The court ruled that LNP’s rights as a licensee were sufficient to confer standing. The court also asserted personal jurisdiction over RTP and declined to transfer venue to Minnesota. On January 22,1998, LNP filed a second amended complaint. The complaint stated that the PTO had reexamined the ’450 patent and the ’889 patent, and that reexamination certificates had issued on October 29, 1996. The complaint asserted that the reexamination proceedings did not change the legal scope of those patents. The expert reports submitted by the parties during discovery identified only three claims as being at issue in this case. The expert report of R. Byron Pipes, submitted on behalf of plaintiffs, and the report of Lawrence J. Broutman, submitted on behalf of defendant, both state that the scope of testimony sought to be provided by the experts would relate to claim 3 of the ’262 patent; claim 1 of the ’450 patent, as reexamined; and claim 1 of the ’889 patent, as reexamined. Similarly, LNP’s discovery responses and contention interrogatories allegedly did not provide notice that LNP would assert other than these three claims. In the Joint Pre-Trial Order filed on September 1, 1998, LNP identified two additional claims — claim 1 of the ’262 patent and claim 6 of the ’450 patent, as reexamined — as being infringed by RTP. The stated reason for this belated assertion is that the declaration of Mr. Ralph Scott Charbonneau, as submitted by RTP on August 7, 1998, provided grounds for finding infringement of the two previously unasserted claims. The court allowed the parties to present evidence at trial relating to all five of the claims. On September 3, 1998, RTP filed a motion to amend its answer and counterclaim to allege the defense of inequitable conduct. RTP allegedly was not able to discover evidence necessary to its inequitable conduct defense because the documents were held by the original patentee, ICI, a British company, and because counsel for LNP purportedly could not procure these documents. At the pre-trial conference on the same day, the court granted the motion. The court stated that the issue of inequitable conduct would not be submitted to the jury, but that evidence relating to the issue could be presented to the court at another time. On September 8,1998, the court granted RTP’s request for international judicial assistance pursuant to the Hague Convention to compel discovery of the documents. On September 15, 1998, the Queen’s Bench Division of the High Court of Justice of England and Wales issued a summons, and discovery proceeded. On November 6, 1998, the court held a final pre-trial conference during which the court permitted RTP to supplement the pretrial order to add a best mode defense. In justifying this belated assertion of the best mode defense, RTP contended that it did not receive the notebooks of the British inventors until discovery had proceeded pursuant to the Hague Convention. Once the notebooks were produced, material therein purportedly provided RTP with grounds to support its best mode defense. RTP sought to present the best mode defense to the jury with the remainder of the issues. Rather than postpone the trial to permit the presentation of evidence to the jury on the best mode defense, the court invoked Rule 16 to. bar litigation of the issue in front of the jury, as the defense had not been raised in the pre-trial order. The court stated that it would hold a hearing after the close of the jury trial to resolve issues relating to the inequitable conduct and best mode defenses. E. Claim Construction The court conducted a Markman hearing on September 15, 1998. On November 5, 1998, the court issued its Markman opinion. The parties had disputed the meaning of four claim terms: 1) “pultrusion;” 2) “thermoformable;” 3) “substantially completely wetted;” and 4) “a composition containing at least 30% by weight of fibre reinforced pellets.” Because RTP has challenged the court’s construction of the term “substantially completely wetted” in its post-trial motions, the court will briefly review its findings regarding this limitation as presented in its Markman opinion. During claim construction, the parties contested the meaning of the claim term “substantially completely wetted.” LNP sought a claim construction that reflected the literal wording of the claims, and that did not refer to the flexural modulus test. LNP’s proposed construction, which was largely adopted by the court, stated that “substantially completely wetted” meant that the filaments be “largely, but not necessarily wholly, surrounded by resin.” As a measure to quantify the degree of wetting, LNP’s proposed construction, as is recited in the claim language, recites use of the dispersal/length test. RTP, on the other hand, sought to construe the term “substantially completely wetted” in reference to the flexural modulus test. RTP noted that the specification and the prosecution history of the patents in suit used the flexural modulus test to characterize the claimed plastics, and to distinguish them from products in the pri- or art. As such, RTP advocated incorporating language from the specification into the construction of the claims, and asserted that “substantially completely wetted” means “the product must attain at least 90% of the theoretically attainable flexural modulus.” The court rejected RTP’s proposed construction. The court reasoned that the flex-ural modulus test was only an indirect test that did not directly measure the characteristics of the claimed wetted strand and pellets. Moreover, the court found that the prosecution history of the patents did not compel a finding that the term “substantially completely wetted” is tied to the flexural modulus test. Thus, the court concluded that the term “substantially completely wetted” should be defined in terms of the dispersal/length test, as follows: Largely, but not necessarily wholly, surrounded by resin. In the context of LFRT pellets, it is surrounding the individual filaments by resin to the extent that in articles injection molded from such pellets, the individual filaments are randomly dispersed and at least 50% by weight of the filaments retain a length of 2 millimeters or greater. LNP Engineering Plastics, Inc. v. Miller Waste Mills, Inc., No. 96-462 (D.Del. Nov. 5, 1998) (claim construction opinion). F. The Trial The case was tried to a jury for nine days beginning on November 12, 1998. 1. LNP’s opening statement Thomas Kenworthy delivered the opening statement on behalf of LNP. He explained the genesis of the claimed inventions and introduced the patents at issue to the jury. Kenworthy focused on claim 3 of the ’262 patent, and explained that the limitation “substantially completely wet-ted” would be met if the individual filaments in a product injected molded from the plastic pellets were randomly dispersed, and if the length of the filaments was retained in the molded product such that at least 50% by weight of the filaments were at least 2 mm in length. Kenworthy told the jury that the evidence would show that LNP’s Verton product, which embodies the claims at issue, became a successful product. RTP, he continued, sought to gain entry into the LFRT market, and conducted a patent search. Upon learning of the ’262 and ’450 patents, RTP allegedly gave no indication that the patents did not read upon their products; rather, RTP purportedly believed that it was free to produce LFRTs, as it was merely practicing the prior art. 2. RTP’s opening statement Ray Lupo delivered the opening statement on behalf of RTP. He explained to the jury the nature of the plastics that RTP manufactures. He indicated that the evidence would show that the technology relied upon by RTP derives from devices invented by Ron Hawley, and that Hawley had granted RTP a license to develop this technology. This technology, Lupo emphasized, was developed in 1979 and patented in 1982, years before the issuance of the patents at issue in this litigation. Lupo stated that Hawley worked for a company, Fiberite, that acquired rights in Hawley’s technology, and that later spun off the firm of RTP. The evidence would show, Lupo continued, that when RTP sought to produce LFRTs, it built an apparatus according to Hawley’s teachings. The products obtained from this process, Lupo explained, typically exhibited less than 70% of the theoretically attainable flexural modulus, and had less than 30% by volume of fibers. Given these physical characteristics and the fact that RTP was merely practicing the Hawley art, Lupo told the jury that the evidence would not permit a finding of infringement. Lupo emphasized that the evidence would show that the patents at issue are invalid. He explained that Hattori and Bradt, two inventors, had worked together at Fiberfil Company for many years and had patented numerous developments on thermoplastics. Lupo explained that when Kawasaki had previously attempted to assert its patents against companies, the Hattori patent was raised as a prior art reference that potentially invalidated Kawasaki’s patents. In response, Kawasaki submitted the ’450 and ’889 patents to the PTO for reexamination, reconstructed the Hattori apparatus, and convinced the examiner that the Hattori method produced plastic inferior to Kawasaki’s claimed technology. Lupo said that the evidence would show that Kawasaki improperly recreated the Hattori process, misleading the examiner to believe that Kawasaki’s technology was distinct and superior. Lupo told the jury that the evidence would show that willful infringement is inapplicable because RTP obtained an opinion stating that RTP was merely practicing the prior art, and that it was permissible to continue doing so. 3. Robert E. Schultz LNP called Robert E. Schultz, who was then serving as President and CEO of Kawasaki LNP, the holding company to which LNP reports. Schultz testified that LNP has rights as a licensee to the patents at issue. Schultz discussed the history of LNP, and described that RTP is one of LNP’s three principal competitors. When the patents at issue emerged from reexamination proceedings, Schultz continued, LNP sued all three of these competitors. Schultz concluded his testimony with financial data of LNP. 4. Frederic Neil Cogswell LNP next called Frederic Neil Cogswell. Cogswell is a native Englishman who began working at ICI in 1959 as a laboratory assistant, and who rose through the ranks as an experimental chemist in ICI’s Plastics Division, eventually becoming a technical officer of the company. Cogswell is currently an adjunct professor at the University of Delaware, and lectures at other universities as well. He told the jury about the physical properties of plastics. He testified that the prior art at the time of his invention taught that long-fiber reinforced thermoplastics could not be made, and that if they were made, they would be undesirably brittle. The widespread understanding in the industry, he indicated, was that the long fibers would scratch injection molding machines, and break. Using a pultrusion technique, Cogswell said that he came to the surprising result that by thoroughly wetting the filaments of a fiber strand, the individual filaments would become sufficiently lubricated by the polymer that they would pass freely through an injection molding apparatus, remaining intact. The resulting products, he described, had surprisingly excellent properties that were found desirable in applications such as manhole covers, propeller blades, and snowshovel handles. He testified that the molded products, when burnt in an oven, leave behind a skeleton of randomly distributed fibers. He testified that the product’s commercial name, Ver-ton, derives from this vertebrate structure. He displayed for the jury the skeletal remainder of a pellet that had been ashed in an oven. On cross-examination, Cogswell acknowledged that the prior art had disclosed that improved wetting of reinforcement filaments can be achieved by spreading the filaments of the strand over a series of “spreader bars” in a pultrusion chamber. Cogswell also recognized that the flexural modulus test is a common, convenient method of determining the degree of wetting of the reinforcement filaments. Cogswell acknowledged that his co-inventors had experimented on how best to place spreader surfaces in the pul-trusion chamber, and that none' of the patents at issue contained drawings disclosing these spreader bar systems. 5. Hugh Miller LNP next called Hugh Miller, who at the time of trial was the Chief Executive Officer for RTP, and who had held that post since 1982, following his father and grandfather in their leadership of Miller Waste Mills. Miller acknowledged that an internal RTP memorandum relating to the development of RTP’s own long-fiber production line stated that LNP’s Verton product had achieved success through excellent physical properties and good marketing. Miller recognized that a letter written on February 12, 1992 by Peter McCamley, the Vice President of RTP, to McDermott, Will & Emery, recites that the Cogswell ’450 patent “would fall in line” with products made by RTP. As acknowledged by Miller, this letter also states that RTP’s wet-out process “is different in concept/design to the ICI patent and to the PCI/Celanese or AKZO patents.” Miller testified that in 1994 he had received a letter from William Schwarze, a patent attorney retained by LNP, and that the letter apprised RTP of the patents at issue in this case. Miller emphasized, however, that he believed RTP had the right to manufacture LFRTs according to the Hawley process. Miller acknowledged that he had no knowledge of any effort to get an opinion from McDermott, Will & Emery before April 1994. Miller testified that when that opinion was received on November 4, 1994, he never read the opinion letter, but understood through discussions with counsel that RTP had the authority to make LFRTs. The opinion letter suggested that RTP analyze its products to determine which ones had less than 80% by volume of parallel aligned reinforcing filaments, as is claimed in the ’450 patent. Miller testified that since he is not a lawyer, he would not understand the letter, and did not read it. Miller testified that in September 1995 he received a letter from Dr. George Niz-nik, Vice President of LNP, which invited RTP to participate in the reexamination of the ’450 and ’889 patents. Miller stated that he had no knowledge as to whether RTP participated in these proceedings. On cross-examination, Miller reviewed the opinion letter from McDermott, Will & Emery and testified that the letter stated that the ’450 and ’889 patents are invalid. 6. Thomas Markey LNP next called Thomas Markey, Vice President of Business Management for RTP. Markey testified that RTP was planning to “toll manufacture” LFRTs for DuPont, whereby DuPont would provide the raw materials, and RTP would produce reinforced plastic pellets for DuPont. He stated that during negotiations with DuPont in 1991, DuPont inquired into RTP’s situation with respect to patents in this area of technology. Markey acknowledged that during a series of correspondences between RTP and McDermott, Will & Emery, Peter McCamley, Vice President of RTP, wrote to the law firm and stated that the claims of the ’450 patent “would fall in line” with products made by PCI Celanese, Wilson Fiberfil, and RTP. Markey did not recall what he did with this letter, but presumably “just filed it.” Markey testified that RTP’s sales of LFRTs increased annually from 1992 through 1997. Markey testified that he had requested that McDermott, Will & Emery prepare an opinion regarding RTP’s potential liability under all three patents in suit and one non-asserted patent (U.S. Patent No. 4,549,920, or the ’920 patent). The opinion was issued in September 1994. Markey acknowledged that the non-infringement section of the opinion only deals with the ’920 patent and the ’262 patent, and that the opinion letter does not state that RTP does not infringe the ’450 and ’889 patents. The letter, as noted by Markey, stated that all four patents being considered were invalid in view of the prior art. Markey then reviewed a document entitled “RTP Company VLF Program,” which was issued on November 12, 1991. This document states its objective as illustrating the status of RTP’s LFRT product line. The document, as acknowledged by Markey, states that LNP has done an excellent job marketing their Verton product line, and that the Verton products have excellent physical properties. RTP, the document continues, “has been almost exclusively replacing Verton product.” Mar-key declined to offer a contrary opinion. On cross-examination, Markey explained that most of the growth in sales of RTP’s long fiber products is attributed to products having less than 30% by volume fiber reinforcement. On redirect, Markey testified that when the present lawsuit was filed, he issued a guideline to RTP personnel indicating that they should only sell long fiber products with less than 30% fiber, by volume. 7. George E. Niznik LNP next called George E. Niznik, Vice President and Director or Research and Development at LNP. Niznik testified that LNP began analyzing the products of its competitors in 1994, when LNP became aware of a customer that was using RTP materials. According to Niznik, LNP conducted a major project to determine which, if any, of the competitors’ products infringed the patents at issue. The testing has an eight-step methodology, directing the analysts to: (1) examine pellets for “parallel and aligned” glass fibers with filament diameters of less than 24 microns; (2) extract thermoplastic from pellets and fiber; (3) burn-out pellets for volume % glass determination; (4) compression mold pellets to determine probability that the pul-truded rods were greater than 70% of theoretical modulus; (5) determine that filaments of glass are for the most part completely wetted; (6) mold competitive samples side by side with comparable VERTON product; (7) conduct burn-out and document that filaments are randomly dispersed; (8) measure filament size in molded article to verify that 50% by weight of the glass is greater than 2mm in length. The testing was supervised by Dale Grove, and later by Lee Glen, and performed by Daria Miller. Niznik testified that one sample of RTP product, RTP type 84009, was analyzed in 1994. This sample yielded 30.229% glass by volume. LNP acquired specimens of a second RTP product, designated No. 68903B, in 1996. This product yielded 30.0285% glass by volume, Niznik testified that a number of letters were exchanged between Schwarze, patent counsel for LNP, and Duncan, patent counsel for RTP. Niznik stated that Duncan, at no time in the letters, made any statement to the effect that RTP did not infringe the patents in suit. According to Niznik, Duncan’s writings targeted the potential invalidity of the patents, particularly in light of the Hawley and Bradt patents. Niznik said that, in response to the allegations of invalidity, LNP submitted the ’450 and ’889 patents to the PTO for reexamination. Niznik stated that he invited Miller, the CEO of RTP, to participate in the reexamination proceedings, but that Miller did not do so. When LNP received notification from the PTO that the reexamination certificates were going to issue, Niznik continued, LNP sued RTP and two other parties, PCI and DSM. On cross-examination, Niznik conceded that he received a document dated March 5, 1990, from Dale Grove, that had concluded that RTP’s products do not infringe the patents at issue. The document recites that “RTP is not infringing ICI patents because they produce a 16% fiber volume (26% by weight) product that must employ a means of impregnation similar to PCI’s approach.” The document further states that the -RTP pellets yielded a filament dispersion pattern similar to that found in the PCI process created by Hawley. Niz-nik admitted that Grove did not discuss this testing with him when the 1994 testing project began. Niznik also conceded that the pen-knife test is subjective, and that experts can disagree on whether a pellet is substantially completely wetted when the pen-knife test is employed. And, Niznik admitted that LNP had only tested two of RTP’s products to determine if they infringe the patents at issue. On redirect examination, Niznik noted that the testing performed by Grove in 1990 was conducted on pellets containing carbon fiber, not glass fiber. Niznik also pointed out that the testing was performed in 1990, prior to the issuance of the ’460 or ’889 patents. And, Niznik stated that the language in the specification discussing the flexural modulus test is irrelevant for determining infringement of three of the claims at issue — claim 3 of the ’262 patent; claim 1 of the ’450 patent, as reexamined; and claim 1 of the ’889 patent, as reexamined — because these claims had been defined by the court in terms of the dispersal/length test. 8. Peter J. McCamley LNP next called Peter J. McCamley. At the time of trial, McCamley was the Vice President of Research and Development for RTP. He conceded that “the whole advantage” of the long fiber pellets is their ability to retain their fiber length in the injection-molded article, and that RTP had limited technological capacity for producing LFRTs in 1990. McCamley was shown an internal document that he had written in 1991, in which he disclosed data on LNP’s Verton “6/6” compounds, and in which he wrote that “I am confident we can produce % nylon VLF glass fiber materials similar to what is available currently.” McCamley testified that in 1992 RTP had one LFRT production line in operation, and that it had identified numerous large prospective customers that would potentially buy LFRT products in 1993. McCamley also acknowledged that an internal RTP document (PX 158) listed various RTP polymers, with the correlation between weight percentage of fiber and volume percentage. He stated, moreover, that all the fibers used in RTP LFRT products have a diameter of less than 24 microns. 9. R. Byron Pipes LNP next called R. Byron Pipes. Pipes, at the time of trial, was the President of Rennslaer Polytechnic Institute, having had extensive background in polymer science. Pipes gave his opinion that a number of RTP products infringe three of the claims at issue: claim 3 of the ’262 patent; claim 1 of the ’450 patent, as reexamined; and claim 1 of the ’889 patent, as reexamined. Pipes stated that he had six grounds for his opinion. First, Pipes testified that he had visually inspected the pellets, having cut them open and found no loose unwetted fibers which would be typical of poorly wetted materials. He stated that he then had examined the pellets under a microscope, which substantiated his earlier observations. He then testified that he had inspected injection-molded articles, like a manhole cover, looking for aggregations of fibers, and had seen a uniform distribution of fibers on the surface of the article, which he stated is indicative of well-distributed fibers under the surface. Second, Pipes stated that he had examined specimen bars which had been ashed, and that the fibers retained a vertebrate structure, and that the fibers were inter-meshed with one another. Third, Pipes testified that he visited the RTP production site. He played a video for the jury, which showed a continuous fiber structure being introduced into the polymer impregnation chamber and then chopped into pellets. Fourth, Pipes reviewed a number of’ RTP documents describing the properties of their LFRT products. One such document listed a number of RTP products, with the percent glass fiber, by weight. Pipes testified that these weight percentages could be readily converted to volume percentages. Pipes agreed with McCam-ley’s testimony that all but two of the products having 50% by weight of fiber or greater have a volume fraction of at least 30% fiber. Fifth, Pipes stated that McCamley’s deposition testimony provided a basis for his infringement opinion. McCamley therein admitted that in the injection molding of RTP’s LFRT pellets, the individual filaments are randomly dispersed. The pellets, McCamley had testified, had fiber lengths greater than 2 mm for up to 50% of the fiber by weight, and this fiber length is retained, McCamley had stated, in the molded product. Pipes recounted McCam-ley’s deposition testimony, in which McCamley agreed that RTP’s internal documents directed to its marketing personnel stated that “RTP Company completely wets out individual glass fibers and pellets.” Sixth, Pipes testified that he had reviewed the laboratory notebooks and the testing performed by RTP scientists and engineers, and that these sources corroborated his own findings. He also testified that he read the legal opinions of Duncan, and that Duncan had not therein stated that the RTP LFRT pellets containing greater than 30% by volume did not infringe claim 1 of the ’450 patent; and that she had not therein written that injection-molded articles made from RTP LFRT pellets did not infringe claim 1 of the ’889 patent. On cross examination, Pipes admitted that, with respect to the video that he had shown to the jury, he had not seen inside the impregnation chamber used by RTP in wetting its fibers. Pipes conceded that his opinion does not inform whether RTP infringes claim 1 of the ’262 patent. He admitted that he was not present when Daria Miller had performed testing for LNP on RTP’s accused products, and that some of the test procedures she had used are subjective in nature. He refuted, however, Lupo’s suggestion that the average weight length determination tests performed by Miller are subjective. 10. Ralph Scott Charbonneau LNP next published a declaration of Ralph Scott Charbonneau, who at the time of the trial was Technical Coordinator for RTP’s Western Region. Charbonneau therein recited the flexural modulus properties of various RTP products. 11. Richard J. Bums LNP next called Richard J. Burns, who at the time of the trial was the President of LNP Engineering Plastics, Inc. He testified that, in 1992, LNP wasmot interested in licensing its technology to RTP, but rather that LNP’s intention was to exploit its patent position. RTP, he testified, was LNP’s chief competitor across all LNP’s other product lines. Burns’ testimony concluded LNP’s case-in-chief. 12. Frederic N. Cogswell a. direct examination As its first witness, RTP called Frederic N. Cogswell, who had previously been called by LNP, and who is an inventor of the patents at issue. Lupo began by questioning Cogswell as to his belief that he was the first to achieve substantially complete wetting of long reinforcement fibers in a melt protrusion process. Lupo questioned Cogswell on U.S. Patent No. 2,877,-501, issued to Rexford Bradt (“Bradt ’501 patent”), which was assigned to Fiberfil, and which claims technology relating to emulsion impregnation of fibers to produce reinforced plastics. The specification of the Bradt ’501 patent recites opening up a glass strand before the polymer is applied thereto “so that each filament in the granule may be uniformly coated.” Cogswell acknowledged that this language could be interpreted to disclose the technology at issue, but stated that Bradt never disclosed how to perform this impregnation. Rather, Cogswell stated, the specification can be interpreted as disclosing the method of production that Fiberfil actually employed, which was to coat the entire strand with polymer without wetting the individual filaments. Lupo next questioned Cogswell on U.S. Patent No. 3,042,570, also issued to Bradt and assigned to Fiberfil (“the Bradt ’570 patent”). Cogswell acknowledged that the patent specification recites that it may be desirable “to work or flex the [glass] strands as they pass through the molten resin in the coating chamber in order to vary the extent to which such resin penetrates the strand,” but Cogswell refused to agree that this language teaches that each individual filament of the strand becomes substantially completely wetted thereby. This language, Cogswell said, states that it is possible to flex the fibers, without giving an indication of how to do so. Cogswell next commented on the reexamination proceedings for the ’450 and ’889 patents, in which the examiner initially rejected the claims based in part on UK Patent No. 1,167,849, issued to Kiyoshi Hattori et al. Cogswell noted that the Hattori patent recites that, in order to chop pellets from a wetted strand, the strand must first be heated. Otherwise, Cogswell stated, the polymer would shatter, leaving a “mass of fiberglass,” because the individual filaments had not been wet-ted. Cogswell testified that he participated in replicating the Hattori process for the aid of the patent examiner, and that he tried to faithfully reproduce the process as taught in the ’849 patent. Cogswell acknowledged that the replicated device did not comprise spreader bars, which were disclosed in the Bradt ’570 patent and incorporated into the Hattori ’849 patent by reference. Cogswell testified, however, that he believed that Hattori’s preferred embodiment, as disclosed in Figure 2 of the Hattori patent, did not employ such spreader bars. Cogswell also acknowledged that the replicated device was run at a speed of 30 meters per minute, whereas Fiberfil had operated its production line at a speed of 3 meters per minute. This difference, Cogswell testified, is due to the difference in polymers employed. He stated that for polypropylene, which LNP used in the replicated process, a faster speed is appropriate due to the polymer’s low viscosity. Lupo next questioned Cogswell on U.S. Patent Nos. 4,439,387 and 4,312,917, both issued to Ronald C. Hawley (“the Hawley patents”). Cogswell acknowledged that he had assumed the products produced by PCI embodied the Hawley patents. He also acknowledged that he had represented to Mr. Scott (who was the attorney representing LNP in the reexamination process) that the Verton products are distinct from the products made by Hawley because Hawley’s products only exhibited 70% of the theoretically attainable flexural modulus, whereas the Verton products attained a level of 90%. Lupo questioned Cogswell on the notebooks of his co-inventor, Peter J. Williams. Williams’ notebook states that “JI]t was also found that the lace path through the melt was crucial in that one lacing up pattern was superior to another, giving rise to an indication of the extra wetting achieved in conjunction with the number of spreader bars engaged,” which Cogswell paraphrased as stating that a greater number of spreader bars yields improved wetting. Cogswell acknowledged that no drawing in the patent illustrated this concept, and that no spreader bars were illustrated. Lastly, Cogswell addressed the language of claim 1 of the ’889 patent, which recites the “30% by weight of fiber reinforced pellets” limitation. Cogswell acknowledged that the preamble of the patent recites that the technology relates to thermoplastic structures containing “at least 30% by volume of reinforcing filaments.” He recognized that the words “30% by weight of pellets” are not recited in the patent specification. Lupo introduced into evidence an internal ICI document written by Dale Grove in 1990 that describes three patents- — U.S. Patent .Nos. 4,541,884; 4,559,262; and 4,549,920 — owned by ICI and that states that the pellets claimed therein: contain at least 30% by volume (not weight) fibers. There has been a recent attempt to amend this claim to read 30% by weight fibers, but until this new amendment is in effect (if it ever is) the law currently stands at this volume percentage. Thirty percent by volume fibers corresponds to roughly 50% by weight fiber in the RF series. Cogswell stated that he was not aware of this prior attempt to change the language of the claims, as he was in ill health at the time. Lupo presented deposition testimony of Hezzell, who stated that he interpreted the claim to mean that the composition “contains at least 30% by weight of fiber.” Hezzell also stated therein that “I don’t understand the concept of a percent by weight of pellets.” Lupo then presented a 1991 letter by J.M. DdWner, an attorney in ICI’s patent department, that was addressed to the European Patent Office. This letter relates to the European counterpart of the ’889 patent. The letter attempts to correct an error that Downer discovered in the application. The letter states: An error has been noticed in claim 9 (now claim 8) in that the requirement should be that the composition contains “at least 30% by weight of fibres in the form of reinforced thermoplastics pellets”, rather than “at least 30% by weight of.pellets”. Cogswell declined to speculate as to what Downer intended by this language. b. cross-examination On cross-examination, Cogswell first addressed LNP’s efforts to replicate the Hat-tori process for the PTO reexamination proceedings. He stated that it was LNP’s goal to replicate the Hattori process, not the Bradt process that is referenced in the Hattori patent. Kenworthy next directed Cogswell’s attention to claim 1 of the ’889 patent. Cogs-well testified that Example 28 in the specification gives support to the “30% by weight” claim limitation. c. redirect examination On redirect examination, Cogswell acknowledged that the process he replicated for the PTO during reexamination was based in part on the Bradt ’570 patent. 13. Richard J. Bums LNP next called Richard J. Burns to complete his earlier testimony. Burns testified that, in an October 1992 negotiation between LNP and RTP, there was discussion about potentially granting RTP a license. Burns testified that LNP would have regarded the license as an incremental sale, allowing RTP to cover their costs, and allowing RTP a 10% margin. LNP would expect the rest of the margin, Burns testified, as a reasonable royalty. On cross-examination, Burns contested the suggestion that there was a discrepancy between his formulation of the damages, and that of Niznik, who was deposed on behalf of LNP during discovery. 14. George E. Niznik RTP next called George E. Niznik, who had previously testified for LNP. Regarding the replication of the Hattori device for the reexamination proceeding at the PTO, Niznik testified that LNP based the replication on Example 2 of the Hattori patent. Niznik acknowledged that a number of figures in the Hattori patent illustrate the use of bars, but that LNP chose not to include bars in their replica. Niznik also acknowledged that the replica used a relatively fast rate of pultrusion, and that the rate of pultrusion does have an effect on the efficiency of the wetting process, particularly when spreader bars are employed. On cross-examination, Niznik stated that, when replicating the Hattori device in accordance with Example 2 of the patent, it would be improper to build the device in light of the teachings of the Bradt ’570 patent, as Example 2 contains no reference to the Bradt patent. He also stated that neither the Bradt patent nor the Hattori patent give an indication of the proper line speed to use in the replica. Niznik testified that, in arriving at a line speed of 30 meters per minute for the replica, LNP looked to the prior art, and Japanese reference 52-112653, in particular. 15. Ronald C. Hawley RTP next called Ronald C. Hawley, who had served as Director of Technical Service for Fiberite Corporation in Minnesota, and who had invented the technology claimed by U.S. Patent Nos. 4,439,387 and 4,312,917. He testified that he left employment at Fiberite in September of October of 1981 to help start a new company, Polymer Composites, Inc. (“PCI”). He stated that during a particularly heavy snowstorm in January 1979, while still working for Fiberite, his wife had asked him if he couldn’t design a better, stronger plastic snowshovel. He conceived of the idea of designing the snowshovel handle from a thermoplastic pultrusion, in which he sought to coat the individual filaments of a strand of glass fibers with thermoplastic resin. Hawley testified that he contacted a local patent attorney, Malcolm Moore, about patenting the idea and to analyze the relevant prior art. Hawley testified that his employment with Fiberite posed a problem to his independent development of this technology, as he had previously entered into an agreement with Fiberite that inventions made by employees become the property of the company. Hawley stated that he met with William Laurie, who was then Vice President of Fiberite. Laurie agreed to allow Hawley to retain rights to the technology, and to assist Hawley in its development, in exchange for “shop rights” in the technology, whereby Fiberite would retain a royalty-free, nonexclusive right to practice the technology. Hawley testified that he designed a prototype process whereby he could force thermoplastic resins into the center of a bundle of glass fiber, thereby coating each individual filament. He stated that his method comprises running the fiber strand over rods, which causes the strand to open up and expose filaments within the strand to the resin. Hawley showed the jury members samples of plastics resembling those produced by the prototype device. Hawley testified that he and Moore filed a patent application on September 13, 1979, and that U.S. Patent No. 4,312,917 issued on January 26, 1982. He recited language from the patent, and stated that it discloses that the glass fiber material and resin would flow under one strand, one lobe, thereby flattening the fiber strand, exposing resins, and then flow over another and another lobe, in an effort to dry or cause resin to cover or coat each filament. Hawley stated that lobes, bars, or rods could be interchangeably used as spreader surfaces. Hawley testified that he obtained incorporation status for PCI on October 29, 1980. He solicited working capital from friends and family, and moved into a new facility in Winona, Minnesota in November 1980. Hawley testified that he achieved his first product run in April 1981. He stated that the process could produce pultruded bars, rods comprised of several bars, and tapes, all trademarked under the name Fiberod. PCI was not allowed to produce pellets, pursuant to a two-year covenant-not-to-compete with Fiberite. He stated that PCI’s products yielded a 65% by weight fiber content, which is approximately 40% fiber by volume. Hawley testified that he provided engineering drawings to Fiberite in compliance with the shop right agreement, and that he gave Ben Miller, then-President of Fiber-ite, an opportunity to observe the equipment at PCI. Hawley stated that he spoke with Peter McCamley, and that he agreed to supply McCamley with lengths of rods, so that McCamley could pelletize the rods and provide Hawley with test data. Haw-ley discussed the data sheet that PCI published in July or August of 1982. He stated that the glass content of one his products (PPGL-65) was greater than 80% by volume. Hawley stated that PCI became a successful company, with a 1988 valuation of between 15 and 20 million dollars, when the company was sold to Hoechst Cela-nese. a. cross-examination On cross-examination, Hawley acknowledged that in his deposition testimony he had stated that Fiberfil products, in the early 1980s, were impregnated with resin on the outside of the glass pellets, but that the inside of the pellets was dry and void of resin. Hawley reviewed the disclosure statement he and Fiberite signed in 1981, in which he identified twenty-three drawings that represented the “know how” of his invention. Hawley insisted that the enumerated materials did disclose the use of spreader surfaces, and that spreader bars were in place by the time representatives from Fiberite inspected the invention. Hawley acknowledged that one embodiment of his invention was to produce reinforced plastic insert rods that could be put into injection-molded products to provide localized reinforcement, but he insisted that his invention was not limited to this embodiment, and that the rods he produced could be pelletized. Hawley acknowledged that he had not reduced to practice all the inventions he had conceived for achieving a complete wet-out of filaments. A PCI business plan, dated November 20, 1981, recites that PCI had difficulty in achieving a complete wet-out within the Fiberod products. Hawley recognized that the business plan stated that there were many variables within the control of PCI, and that PCI had to find the right combination of these parameters to achieve an optimal wet-out. 16. June Hawley RTP next called June Hawley, wife of Ronald Hawley. She testified that she helped her husband build