William A. Marshall, Merriam, Marshall & Bicknell, Chicago, Ill., for defendants-appellants.

Roy E. Hofer, Hume, Clement, Brinks, Willian & Olds, Ltd., Chicago, Ill., for plaintiff-appellee.

Before SWYGERT and SPRECHER, Circuit Judges, and CRABB, Chief District Judge.*

CRABB, Chief Judge, Sitting by Designation.

This case is before the court on appeal from a judgment and injunction entered by the district court upon its findings that a patent held by plaintiff-appellee Novo was valid, that it had been infringed by defendants-appellants, and that plaintiff-appellee Novo was entitled to increased damages and attorney's fees because the infringement was wilful, deliberate, and in bad faith.

1. Was the district court correct in its conclusion that the Novo invention was not obvious?

2. Was the district court correct in its conclusion that the Novo patent satisfied the "best mode" requirement of 35 U.S.C. § 112?

3. Was the district court correct in its conclusion that Novo had not misused its patent?

4. Did the district court act within its discretion in awarding Novo increased damages and attorney's fees?

For the reasons that follow, we answer these questions "yes" and accordingly affirm the district court in all respects.

The patent at issue, United States patent, No. 3,988,207, concerns the identification of a fungus species which, when cultivated by a previously known process, produces a milk-coagulating enzyme needed for the production of cheese. It is owned by plaintiff-appellee, Novo Industri, a Danish biochemical and pharmaceutical concern. Novo does not claim to have discovered the fungus species; rather it claims the discovery of the fungus' ability to produce a milk-coagulating enzyme in commercially profitable amounts.

Defendant-appellant, Travenol Laboratories, Inc., is a Delaware corporation; through its Wallerstein division, Travenol manufactured the milk-coagulating enzyme known as Fromase, which was found by the district court to infringe Novo's patent because the production of Fromase involves the use of the fungus species covered by Novo's patent.

Defendant-appellant GB Fermentation Industries, Inc. (GBFI) is a Delaware corporation and a United States subsidiary of defendant-appellant Gist-Brocades N.V., a Dutch corporation. In July, 1977, GBFI and Gist acquired Travenol's Wallerstein division and thereafter continued to make and sell Fromase in the United States; they have been added to this litigation pursuant to Rule 25(c), Federal Rules of Civil Procedure. For simplicity the defendants-appellants will be referred to collectively as Travenol throughout this opinion.

To place the parties' claims in context, we set forth the factual background against which the claims arise.1

The process of cheesemaking begins with the coagulation of milk into curds and whey. Curds contain the casein protein from which the cheese is made; whey is the liquid by-product of the coagulation process that occurs when an enzyme acts on the milk to precipitate the casein.

Until the 1970's, the principal milk-coagulating enzyme was calf rennet, an enzyme found in the fourth stomach of unweaned calves. Because of the fluctuation in the supply and cost of calves' stomach rennet, cheesemakers have been searching for years for commercially viable substitutes. Many of them have focused their research efforts upon microbial rennets; that is, milk-coagulating enzymes produced by bacterial or fungal microorganisms. As of 1965, researchers had identified microbial substitutes for calf rennet produced by microorganisms from various genera, including Bacillus, Penicillum, Aspergillus, Endothia, and Mucor. At that time, however, no one had discovered a microorganism of any commercial value. Most of the microorganisms studied were found to be unsuitable for good milk-coagulation because their high proteolytic activity reduced cheese yield and produced decomposition products that resulted in bitter tasting cheese.

In 1964 and 1965, two Japanese researchers working together obtained two United States patents, Nos. 3,151,039 and 3,212,905, for milk-coagulating enzyme "microbial rennets" made from cultures of a species of fungus known as Mucor pusillus Lindt, used in combination with other fungus genera. These Japanese patents are referred to as the Arima patents. Although they claim several methods of fungus cultivation and microbial rennet recovery, their primary novel aspect is the identification of the properties of the species, Mucor pusillus Lindt, in milk-coagulation.2

The Arima patents identified M. pusillus as producing an effective microbial rennet; the patents did not claim any strains of M. pusillus by name.3 However, the patents described the M. pusillus culture used as one in which no zygospores are observable, thus indicating that it was "heterothallic."4 The file history of the Arima '039 patent contains a reference to an article written by George Smith in which Smith discusses a homothallic strain of M. pusillus. From this reference, it can be inferred that Arima was aware of two different strains of M. pusillus and that his failure to refer to both in his patent evinced his intention to claim only the heterothallic strain of M. pusillus.5

In 1964, the same year in which the first Arima patent was obtained, Donald Cooney and Ralph Emerson published a book on the taxonomy of certain types of fungi, in which they set forth their opinion that homothallic strains of M. pusillus, including the one identified by George Smith and known as the Smith strain, belonged to a separate species. In their book, Cooney and Emerson reclassified these homothallic strains as the species Mucor miehei.6 By definition, members of the species M. pusillus would not be capable of interbreeding with members of the species M. miehei.

As of 1964, the catalog of at least one culture collection, the Commonwealth Mycological Institute (CMI), listed two homothallic strains, the Smith strain and another known as the Gregory strain, under the heading, M. pusillus.7 The CMI is one of a small number of culture collections from which researchers from all over the world can obtain samples of fungi species or even separate strains of a certain species.

In 1965, Karl Aunstrup, an employee of Novo, discovered that an unidentified fungus produced an excellent milk-coagulating enzyme. Earlier, he had replicated Arima's work, using a strain of M. pusillus purchased from a culture collection. He had found the results "not interesting," and had turned to natural sources for microorganisms. It was while screening samples from hay, compost, and molds that Aunstrup discovered the species that proved to be so remarkable in coagulation. Aunstrup did not know the identity of the fungus; it was only after performing detailed studies that he determined that it belonged to the species to which Cooney and Emerson had referred, M. miehei.

In December, 1965, only a few months after Aunstrup's discovery, his employer, Novo, filed a provisional patent application in Great Britain. On November 21, 1966, Novo filed the United States patent application that matured into the patent at issue in this case. The novel aspect of the Novo patent was the use of M. miehei; the process in which it was used to produce enzymes was essentially that described in the Arima patent '039, which itself was primarily novel in its use of the heterothallic strain of M. pusillus.

Meanwhile, in the United States, Louis Feldman was making his own search for a calf rennet substitute on behalf of his employer, Travenol. In March, 1966, Feldman learned about a newly-identified Mucor species, M. miehei, from a Mucor specialist, C. W. Hesseltine, a United States Department of Agriculture employee associated with the Northern Regional Research Laboratory (NRRL), an American culture collection. In July, 1966, Feldman wrote to Hesseltine asking him for samples of any "candidate organism (Hesseltine) may deem likely for a rennin screen, particularly any Mucor mehie (sic) cultures that might be available." Two months later, Feldman wrote again to Hesseltine asking for all available cultures of homothallic M. pusillus.

In response to these requests, Feldman received from the NRRL some strains of M. miehei and the Smith strain of M. pusillus. This is exactly what he would have received had he ordered only "homothallic M. pusillus" from the culture collection of the Commonwealth Mycological Institute because he would have been sent the Smith strain of homothallic M. pusillus and the Gregory strain, which has now been identified as M. miehei.8 The M. miehei strains produced unexpectedly good results, better than anything Feldman had tried before. He believed he had a patentable invention, and Travenol's patent counsel agreed with him. Feldman and Travenol filed a patent application in April, 1967. Six months later, Travenol learned of the Novo application and shortly thereafter, Travenol filed a second patent application claiming certain specific strains of M. miehei ; however, it did not abandon the first application until more than two years later. Travenol has filed patent applications on the invention in eighteen foreign countries.

In July, 1971, the Patent and Trademark Office (PTO) declared an interference between the Novo application and the second Travenol application. On November 16, 1973, the Board of Patent Interferences awarded priority of invention to Novo's application. Its decision was affirmed by the Court of Customs and Patent Appeals, Feldman v. Aunstrup, 517 F.2d 1351 (C.C.P.A. 1975).9 The United States Supreme Court denied Travenol's petition for a writ of certiorari in February, 1976. Travenol continued to market Fromase microbial rennet.

Travenol is in the anomalous position of asserting the obviousness of a discovery which it and its patent counsel believed to be so revolutionary as to warrant both an application for a patent and extended litigation for the patent rights. Despite its own efforts to patent the same discovery, Travenol now defends against plaintiff-appellee's claim of patent infringement by asserting the obviousness of the discovery of M. miehei for producing an enzyme capable of coagulating milk to the person of ordinary skill in the art, who has knowledge of all of the prior art disclosed at the time of the invention.

The standard for obviousness derives from the Supreme Court's opinion in Graham v. John Deere, Inc., 383 U.S. 1, 17, 86 S. Ct. 684, 693, 15 L. Ed. 2d 545 (1966):

Under (35 U.S.C.) § 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 resolved; and the level of ordinary skill in the pertinent art resolved.

On the issue of obviousness, the trial court made findings of facts which we paraphrase as follows:

With respect to the prior art, the Arima '039 patent did not, and never intended to, claim a homothallic M. pusillus. As a general rule, inventors do not claim less than they think they have invented, particularly when they are represented by counsel; had the Arima patent applicants wanted to claim a homothallic strain, they would have done so.

The Arima file history contains a reference to the Smith article describing a homothallic M. pusillus. The homothallic M. pusillus described in the article is not a M. miehei. From reading the file wrapper, a person could infer that Arima believed that the Smith article referred to a homothallic M. pusillus. The wrapper is not prior art which suggests a M. miehei.

In saying that there was no such thing as a homothallic M. pusillus, Cooney and Emerson were wrong. The Smith strain is a homothallic strain of M. pusillus. Although Travenol denies this, its position is not supported by any credible evidence.

The 1964 CMI catalog listed both the Smith and the Gregory strains under the species heading M. pusillus. We know now that the Gregory "strain" is actually a separate species, M. miehei.

From reading the 1964 catalog and Cooney and Emerson, the person of ordinary skill would have concluded that the catalog was in error, since, according to Cooney and Emerson, there was no such thing as homothallic M. pusillus, and the person might have guessed that the Smith and the Gregory strains were really M. miehei. However, neither the Arima patent nor its file wrapper would have taught that a homothallic strain of M. pusillus would be effective in producing milk-coagulating enzymes; therefore, it certainly would not have indicated that there was a whole separate species of Mucor that would be successful. No person of ordinary skill in the art would have expected the dramatic results obtained with the use of M. miehei.

The art involved is that of fermenting useful enzymes from microbes. The level of skill would be a bachelor's or master's degree, with no expert knowledge of fungal taxonomy.

Travenol makes two alternative arguments in support of its contention that the discovery of the value of M. miehei was obvious. Both arguments rest on the assertions that the Arima patents and file history suggested the use of homothallic M. pusillus in the production of microbial rennets and that all homothallic M. pusillus strains are M. miehei.

The first argument focuses on the erroneous 1964 CMI catalog listing of M. miehei as homothallic M. pusillus. According to this argument, the hypothetical enzyme researcher would have discovered the value of M. miehei when he or she ordered all of the available homothallic M. pusillus from the CMI culture collection and received both the Smith strain and the Gregory strain, now known to be the separate species, M. miehei.

The alternative argument is directed at the taxonomic classification M. miehei as a separate species. Travenol maintains that M. miehei was carved out of the species M. pusillus after taxonomists had decided that what appeared to be a homothallic strain of M. pusillus should be classified as a separate species. Travenol appears to be arguing that there is no actual difference between homothallic M. pusillus and M. miehei, only a difference in nomenclature; therefore the Arima suggestion to use homothallic M. pusillus would lead a researcher directly to M. miehei because it is the same organism, reclassified under a new name.10

With respect to the teachings of the file wrapper, the trial court recognized that the actual import of the file history would have been to deter intelligent researchers from trying the homothallic strains of M. pusillus, because they would have inferred from the fact that the Arima patent did not claim the homothallic Smith strain that the strain had been tried and found wanting.11 Given the thousands of fungal species (Mucor alone has 150 to 200 separate species) and the finite time available to any researcher, there would be little reason to devote any time to working with what Arima had disdained.

Moreover, being aware of all of the extant writing in the field at the time, the hypothetical researcher would have known of the Cooney and Emerson opinion that there was no such thing as a homothallic strain of M. pusillus and that anything appearing to be such was in reality a separate species, M. miehei. Far from leading the researcher to an investigation of M. miehei, the Cooney and Emerson evidence would seem to explain Arima's apparent lack of success with the homothallic strain of M. pusillus : presumably, like other Mucor species, M. miehei did not produce milk-coagulating enzymes.

With respect to Travenol's assertion that M. miehei is nothing more than a reclassification of homothallic M. pusillus, the trial judge found to the contrary; that additional scientific testing has determined that there is a homothallic strain of M. pusillus (the Smith strain mentioned by Arima) and a separate species, M. miehei (the Gregory strain, never mentioned by Arima). Additionally, the trial judge found that Travenol was asserting the position that all homothallic M. pusillus strains are M. miehei without ever having performed the examination of the Smith strain that would uphold or destroy its argument.12 As did the trial judge, we find Travenol's failure to examine the Smith strain highly suspect in light of the significance the company places upon the assertion that all homothallic strains of M. pusillus, including the Smith strain, are actually M. miehei under a different name.

Finally, even if we were to accept Travenol's assertion that the Arima file history taught the use of homothallic M. pusillus, and to accept, as well, the assertion that homothallic M. pusillus is only another name for M. miehei, we would not conclude that the discovery of M. miehei's enzyme-producing qualities was obvious. Even if it had been obvious to try M. miehei, it does not follow that the results of such a trial could have been anticipated.

"Obvious to try" is not the same as "obviousness." In re Goodwin, 576 F.2d 375, 377 (C.C.P.A.1978), aff'd, 599 F.2d 1061 (C.C.P.A.1979); In re Mercier, 515 F.2d 1161, 1167 (C.C.P.A.1975). The test of obviousness is whether the invention as a whole would be obvious, including the nature of the results obtained. The results from using M. miehei in place of M. pusillus were totally unexpected. At best, Travenol's argument demonstrates that in hindsight there existed a route to a particular result. To argue that either the route or the result was obvious is comparable to arguing the obviousness of Columbus' discovery of America from the fact of its discovery.

We adopt the district court's conclusion that neither the Arima '039 patent and its file wrapper nor the 1964 CMI catalog teach the use of M. miehei, and that Aunstrup's invention would not have been obvious to a person of ordinary skill in the art of enzyme fermentation.

This finding of non-obviousness is buttressed by Travenol's efforts to claim domestic and foreign patent protection for the same discovery. McKee Door Co. v. Forest Door Co., 284 F.2d 809, 814 (7th Cir. 1960). See also LaSalle Street Press, Inc. v. McCormick & Henderson, Inc., 445 F.2d 84, 87-88 (7th Cir. 1971); University of Illinois Fdn. v. Block Drug Co., 241 F.2d 6, 13 (7th Cir. 1957).

Because this is not a close case, we do not need to consider any secondary evidence of non-obviousness, such as the long felt but unsolved need for calf rennet substitute or the commercial success of M. miehei. Republic Industries, Inc. v. Schlage Lock Co., 592 F.2d 963, 975-76 (7th Cir. 1979).

Additionally, we conclude that because the PTO had before it all of the pertinent prior art (the Arima '039 patent and its file wrapper) a presumption of validity attends Novo's patent which Travenol did not overcome by clear and convincing proof.13 35 U.S.C. § 282; L. E. Sauer Machine Co. v. Corrugated Finishing Products, 642 F.2d 203, 206 (7th Cir. 1981); Chicago Rawhide Mfg. Co. v. Crane Packing Co., 523 F.2d 452 (7th Cir. 1975).

Travenol contends that Novo's patent violates the "best mode" requirement of 35 U.S.C. § 112, which requires both that the patent disclose an invention in a manner that will enable a worker skilled in the art to reproduce the invention (enablement) and that the patent disclose as well the best method of using the invention known to the patentee at the time of the patent application (best mode). Additionally, Travenol argues that the district court erred in considering "best mode" and "enablement" as indistinguishable from each other in this case.

On this issue, the district court made findings of fact which we summarize as follows:

Prior to the filing of the British patent, Aunstrup had deposited in the CBS culture collection a specimen of the M. miehei strain that he was using in his experiments and that he considered to be the best of the M. miehei strains he had tested. At the time he and Novo filed their United States patent application, Aunstrup was using M. miehei strains no better than those he had deposited in 1965.

The British patent application disclosed an ethanol recovery step; the United States application omitted this recovery step. The ethanol recovery step is a conventional enzyme recovery step well known to workers skilled in the art. Its disclosure was not essential to the replication of the patented process. At the time of filing of the patent applications, Aunstrup had no preferred recovery step.

As of the filing of the British application, Aunstrup had used two different defoaming agents in his M. miehei work: soybean oil and Pluronic L-61. (A defoaming agent is a substance added to the shake flask or fermentor to reduce the foam produced by the aeration of the fermentation medium.) Aunstrup did not disclose the use of Pluronic L-61. Both the use of a defoaming agent and the kind of defoaming agents used by Aunstrup were well known in the art at the time of the filing of the application.

The use of Pluronic L-61 was irrelevant to the success of the newly-invented process. Travenol's own records show that within a fairly short time after reading Novo's patent, Feldman was able to reproduce the patent without using Pluronic L-61 or any other defoaming agent, despite Feldman's trial testimony to the contrary.

From these findings the district court concluded that Novo had not omitted anything from either the British application or the patent at issue herein that would prevent anyone with ordinary skill in the art from enjoying the benefits of Aunstrup's invention, and that therefore the patent was in compliance with the requirement. 35 U.S.C. § 112.

The district court's findings are not clearly erroneous or subject to attack because the court collapsed the best mode and enablement requirements of 35 U.S.C. § 112 into a single inquiry. In the circumstances of this case, the two requirements merge with each other.

The significance of the patent at issue is its disclosure of a new agent, the fungus M. miehei, for producing a cheese-making enzyme. The patent does not claim to disclose a new process for producing the enzyme; rather, it identifies M. miehei as an agent which produces significantly greater amounts of the enzyme than previously known agents by way of a process that the district court found was already familiar to ordinary workers skilled in the art. Feldman's testimony bolsters this point. He stated that the only difference between his own claim and Arima's was in the species cited. The preparation and recovery processes were no different from those set out in the Arima '039 patent. As the fungus does all the work that makes this discovery patentable, the patent application required only that the patent applicant provide samples and enough taxonomic information that an ordinary worker skilled in the art could order the proper fungus species from a culture collection. Thus, the accurate identification of M. miehei enables the ordinary worker skilled in that art to reproduce Aunstrup's discovery and simultaneously discloses the best way to carry out the invention. Flick-Reedy Corp. v. Hydro-Line Mfg. Co., 241 F. Supp. 127 (N.D.Ill.1964), rev'd on other grounds, 351 F.2d 546 (7th Cir. 1965).

As an additional defense to Novo's claim of infringement, Travenol alleged that Novo had misused its patent rights by entering into a licensing agreement with Miles Laboratories which extended Novo's monopoly power to control the sale of unpatented products in countries where Novo had no patent protection.14

On the issue of patent misuse the trial court found that the agreement between Novo and Miles was ambiguous; that in one paragraph, (16), Miles agreed that in any country where Novo had a patent or pending patent application, it would not sell M. miehei microbial rennet manufactured by Miles' Mexican subsidiary; that in another paragraph, (24), of the same agreement, Miles agreed that in any country other than the United States and Canada where it had not established its own production of M. miehei microbial rennet, Miles would buy all of its rennet from Novo; that the contract was in need of external aids to construction, such as evidence of actions taken by the parties pursuant to the contract; that there was no such evidence; that, since Travenol had the burden of proof on the issue, the lack of evidence meant that Travenol had failed to carry its burden of proving patent misuse.

In finding that the contract was ambiguous and that the record lacked the evidence to support Travenol's view of the restrictive nature of the contract terms, the trial judge did not err. He was correct in his conclusion that Travenol had failed to carry its burden of proof on this point. Having reached that conclusion, it never became necessary for him to reach the question whether the agreement between Miles and Novo would constitute misuse of Novo's patent if construed as Travenol asserted it should be.IV. Increased Damages and Attorney's Fees

Sections 284 and 285 of Title 35 permit the trial court to award increased damages (up to three times the amount found) and, in exceptional cases, attorney's fees. In this case, the trial judge concluded that Travenol's infringement of Novo's patent was deliberate and wilful, and that this was an exceptional case warranting an award of increased damages and attorney's fees.

In reviewing this decision, we must determine whether the district court abused its discretion in making these awards, bearing in mind that such awards are not to be given to a prevailing party as a matter of course but only upon an "unambiguous showing of extraordinary misconduct." Airtex Corp. v. Shelley Radiant Ceiling Co., 536 F.2d 145, 155 (7th Cir. 1976). An award of increased damages under § 284 is directed at acts constituting "conscious and wilful infringement of the patent." Deere & Co. v. International Harvester Co., 658 F.2d 1137, 1146 (7th Cir. 1981); Union Carbide Corp. v. Graver Tank & Mfg. Co., 282 F.2d 653, 660 (7th Cir. 1960), cert. denied, 365 U.S. 812, 81 S. Ct. 695, 5 L. Ed. 2d 691 (1961). An award of attorney's fees focuses upon misconduct in the prosecution or defense of the patent claim, such as fraud on the patent office or the initiation of suit with unconfirmed data to support infringement, Loctite Corp. v. Fel-Pro, Inc., 667 F.2d 577, 584 (7th Cir. 1981); failure to respond to discovery requests, L. F. Strassheim Co. v. Gold Metal Folding Furniture Co., 477 F.2d 818, 824 n.9 (7th Cir. 1973); testimony of a defendant "contrary to all documentary evidence, amounting to 'an unconscionable act done with deceptive intent,' " Loctite Corp., 667 F.2d at 584, (quoting Skil Corp. v. Lucerne Products, Inc., 503 F.2d 745, 750 (7th Cir. 1974), cert. denied, 420 U.S. 974, 95 S. Ct. 1398, 43 L. Ed. 2d 654 (1975)); or misrepresentation of the character of exhibits, together with the unnecessary prolongation of the taking of depositions, and a refusal to obey a subpoena for documents, Ashcroft v. Paper Mate Mfg. Co., 434 F.2d 910 (9th Cir. 1970).

With respect to the wilfulness of the infringement, the trial judge found that Travenol had no bona fide belief or any reasonable basis for a bona fide belief that the Novo patent was invalid, unenforceable, or not infringed, either at the time it commenced its infringement of the patent in October, 1976, or later when this suit was filed, or in July, 1977, when Travenol sold its Wallerstein division to Gist and GBFI; that Travenol paid $80,000 to Gist and GBFI at the time of the sale for an agreement from them to hold Travenol harmless for infringement of the Novo patent; that in August, 1972, Travenol sought oral opinions from four patent law firms in Chicago and New York on the outcome of the interference proceeding and subsequent infringement suit by Novo; that among the questions on which Travenol sought advice was the possibility of an award of treble damages and attorney's fees; and that Travenol adduced no evidence that it had obtained a legal opinion justifying its continued sale of Fromase in the United States, either at the conclusion of the interference proceeding or afterwards. Further, the trial judge found that once the interference proceeding had been concluded and the patent had been issued to Novo, Travenol was under an affirmative duty to exercise due care to determine whether it had a reasonable legal basis for its continued sale of Fromase; that Gist and GBFI assumed this same duty upon their acquisition of the Wallerstein division; and that the record is devoid of any evidence that might indicate that any of the defendants-appellants ever obtained an opinion of legal counsel that would have justified their infringement of the Novo patent.15 From this, the trial judge found Travenol's (and later, Gist's and GBFI's) infringement of the Novo patent to be deliberate and wilful.

The trial court's findings provide adequate support for the conclusion that increased damages are proper in this case. Although the size of the increase has yet to be determined, we are satisfied from our own review of the record that the trial court would be acting within its discretion in awarding Novo any increase in the amount of damages up to and including the maximum allowed under the statute.

With respect to the award of attorney's fees, the trial court found that the defenses asserted by Travenol at trial were not bona fide defenses; that they were not asserted in good faith, but as an afterthought; that Travenol had deliberately avoided conducting any detailed examination or testing of the Smith strain, despite the weight it placed on the assertion that this strain was a M. miehei; and that Travenol refused to concede the fact of infringement prior to trial although it made no effort to contest it once trial was under way.

We conclude that these findings are not clearly erroneous. From them, and from our own review of the record, we conclude that the trial court was correct in its assessment of this case as an exceptional one warranting an award of attorney's fees to the prevailing party.

As our extensive review indicates, Travenol's "obviousness" argument seems to be an after-the-fact argument designed primarily to dazzle the district judge and take advantage of his presumed lack of familiarity with the current state of the art of biology and taxonomy. As judges, we are not so naive as to believe that such an approach is uncommon, although we do not condone it. Neither are we so naive as to think that "obviousness" arguments are not derived primarily from the inventiveness of patent lawyers. In finding this case exceptional, we rely neither upon the esoteric nature of Travenol's argument nor upon the obvious fact that the argument is an artificial construct of counsel (rather than a description of how an actual researcher might have discovered the value of M. miehei). What we do rely upon is what underlies Travenol's argument: mischaracterization of the evidence and the deliberate avoidance of knowledge that might be disadvantageous to its position.

In upholding an award of attorney's fees in Loctite Corp. v. Fel-Pro, Inc., 667 F.2d at 585, we noted that "responsible action on the part of Loctite and its attorneys could have greatly alleviated the cost and extent of this suit. Indeed, with more care, parts of it may never have been brought at all." The same may be said of this case. Had Travenol not maintained its meritless defenses of obviousness and best mode, the trial would have never consumed twelve days of trial time. Had Travenol not continued to assert two additional contentions of invalidity and a denial of infringement right up until trial, it is probable that Novo's trial preparation time would have been reduced.

That Travenol's actions were not responsible is shown by its deliberate avoidance of any examination or testing of the Smith strain and by its continued mischaracterization of the Arima '039 file wrapper as "teaching" the use of homothallic M. pusillus, a mischaracterization that is compounded in this court by Travenol's misrepresentation of the trial judge's finding on that point. We conclude that the trial court will be acting within its discretion in awarding reasonable attorney's fees to Novo.

In the past, we have noted the need to examine "more critically" district court findings that are adopted verbatim from findings proposed by the prevailing party. Machlett Laboratories, Inc. v. Techny Industries, 665 F.2d 795 (7th Cir. 1981); Lektro-Vend Corp. v. Vendo Co., 660 F.2d 255, 263 (7th Cir. 1981), cert. denied, --- U.S. ----, 102 S. Ct. 1277, 71 L. Ed. 2d 461 (1982); Photovest Corp. v. Fotomat Corp., 606 F.2d 704, 731 (7th Cir. 1979), cert. denied, 445 U.S. 917, 100 S. Ct. 1278, 63 L. Ed. 2d 601 (1980). In this case, the trial court adopted findings submitted by the prevailing party, Novo. However, these findings only supplemented and clarified the thirty-one pages of findings made orally from the bench at the conclusion of the trial and following argument on the questions of increased damages and attorney's fees. It is obvious from reading the transcribed findings that Judge Grady had a firm grasp of the essential elements of his decision in this case, including the elements material to an award of increased damages and attorney's fees.

For this reason and for the other reasons discussed previously, the judgment and injunction of the district court are, in all respects,

AFFIRMED.