Davis-Edwards Pharmacal Corporation, now in an arrangement proceeding under Chapter XI of the Bankruptcy Act in the Eastern District of New York, appeals from an order of that court granting a preliminary injunction against its conceded infringement of the United *870States Patent 2,724,720 (Claim 4) for a drug known generally as meprobamate, issued on November 22, 1955, to Frank M. Berger and Bernard J. Ludwig, and assigned by them to plaintiff’s predecessor, Carter Products, Inc. The pharmacological compound described in the patent has been made into pills by plaintiff and widely marketed as a tranquilizer under the trade name “Miltown.” For many years the only other authorized user and seller of the single compound was the Wyeth Division of American Home Products Corporation, which manufactured it into pills marketed under the trade name “Equanil.” As a result of a consent decree in 1962 in an action brought by the Government under the antitrust laws, United States v. Carter Products, Inc., 211 F.Supp. 144 (S.D.N.Y.1962), Carter Products’ agreement with American Home was terminated, and it agreed to sell the compound to anyone at a price not to exceed $20.00 per pound, with a protective provision for increase based on the Consumer Price Index. Davis-Edwards purchased the compound from plaintiff at $20.00-$21.40 per pound from 1963 to 1968. Since that time it has been acquiring the compound from other sources, mainly foreign, at about 87 cents per pound. Defendant has not been by any means alone in following this course. Indeed, the largest infringer has been the United States, which began purchasing the compound from other sources at least as early as 1960 but was not sued by plaintiff until 1968.
Although Carter-Wallace was notified by the Bureau of Customs on June 11, 1968 that Davis-Edwards was purchasing meprobamate from foreign sources, plaintiff did not bring this action until October 9, 1969, and did not move for a preliminary injunction until December 22, 1970, some six weeks after this court had sustained the validity of the patent, against limited attack, in Carter-Wallace, Inc. v. Riverton Laboratories, Inc., 433 F.2d 1034 (1970).1 Extensive affidavits, exhibits and memoranda were filed. Judge Dooling heard argument for three days. On the following day, February 25, 1971, he granted the injunction, after delivering an oral opinion which, including colloquy with counsel who were allowed to comment on his observations as he proceeded, fills 167 typewritten pages.2 On the basis of defendant’s representation that the loss of its meprobamate business and the effect of this on its sales of other generic drugs not only would preclude any possibility of a Chapter XI arrangement but would prevent its proceeding to trial in this action, a panel of this court, on March 11, 1971, stayed the temporary injunction on condition that the appeal be argued during the week of March 29. We appreciate the assistance rendered by counsel in the excellent briefs filed and the arguments prepared within the severe time limits thus imposed.
I.
Davis-Edwards attacked the patent on three general grounds, the first two of which have some tendency to merge: invalidity over prior art, most of which was not previously cited; inequitable conduct of the applicants before the Patent Office; and patent misuse predicated on violations of the antitrust laws. Examination of the Riverton opinion, 433 F.2d at 1037, shows that none of the *871specific points raised here was there presented or considered.3
Davis-Edwards’ case for reversal rests fundamentally on the rule succinctly stated by Judge Learned Hand in Simson Bros. v. Blancard & Co., 22 F.2d 498, 499 (2 Cir. 1927), “We have often said that an injunction pendente lite in a patent suit should not go except when the patent is beyond question valid and infringed.” He cited four cases, Newhall v. McCabe Hanger Mfg. Co., 125 F. 919 (2 Cir. 1903); Hildreth v. Norton, 159 F. 428 (2 Cir. 1908); George Cutter Co. v. Metropolitan Electric Mfg. Co., 275 F. 158, 164 (2 Cir. 1921); and A. B. Dick Co. v. Barnett, 277 F. 423 (2 Cir. 1921). He might also have cited Dickerson v. De la Vergne Refrigerating Machine Co., 35 F. 143, 147-148 (C.C.S.D.N.Y. 1888), where Judge Lacombe referred to “the wholesome and well-settled rule which requires an adjudication in court, or public acquiescence, as a necessary prerequisite to granting the relief [a preliminary injunction in an action for patent infringement] here prayed for,” and Reed Mfg. Co. v. Smith & Winchester Co., 107 F. 719, 720 (2 Cir. 1901).
Perhaps because the rule, imposing on a patentee seeking a preliminary injunction against infringement a burden substantially heavier than the usual ones of showing irreparable injury and probability of success, was regarded as “well-settled” as long back as 1888, most of the later decisions shed little light on the reasons for it. In Rosenberg v. Groov-Pin Corp., 81 F.2d 46, 47 (2 Cir. 1936), Judge L. Hand addressed himself to its rationale:
The doctrine that in the absence of long acquiescence or adjudication an injunction [i. e., a preliminary injunction in a suit for patent infringement] will not go, is at first blush anomalous in the light of the presumption of validity which courts generally grant to a patent once issued. * * * The theory is * * * practical. Examiners have neither the time nor the assistance to exhaust the prior art; nothing is more common in a suit for infringement than to find that all the important references are turned up for the first time by the industry of a defendant whose interest animates his search. It is a reasonable caution not to tie the hands of a whole art until there is at least the added assurance which comes from such an incentive.4
Indeed, a unanimous Supreme Court in Graham v. John Deere Co., 383 U.S. 1, 18, 86 S.Ct. 684, 694, 15 L.Ed.2d 545 (1966), noted the “notorious difference between the standards applied by the Patent Office and by the courts,” and suggested that the “discrepancy” was due in part to “the free rein often exercised by Examiners in their use of the concept of ‘invention,’ ” a result, perhaps, *872of the enormous number of patent applications (nearly 100,000) which are filed each year. Without in any way disparaging the skill and devotion of patent examiners, to which our dissenting brother pays extensive tribute, and despite the statutory presumption of validity, the bald fact is that more than 80% of patent infringement actions on appeal result in a determination that the patent sued upon is invalid. Milgrim, Sears to Lear to Painton: Of Whales and Other Matters, 46 N.Y.TJ.L.Rev. 17, 31 & n.62 (1971). In apparent recognition of the potential unfairness of allowing one armed with letters patent obtained in an ex parte proceeding to prevent the conduct of business by others before he has submitted all fairly disputed issues of fact to a full adversary hearing and has won a favorable decision by some court, or the industry has in effect conceded the validity of his patent, the “wholesome and well-settled rule” has been followed in other circuits, although with differences in precise formulation. Packard Paper Box Co. v. O. B. Andrews Co, 67 F.2d 783, 784 (1 Cir. 1933); Leavitt v. McBee Co, 124 F.2d 938, 939-940 (1 Cir. 1942); Hoeme v. Jeoffroy, 100 F.2d 225, 226 (5 Cir. 1938); Flintkote Co. v. Philip Carey Co, 13 F.2d 850 (7 Cir. 1926); Stoody Co. v. Osage Metal Co, 95 F.2d 592, 593 (10 Cir. 1938): Pacific Cage & Screen Co. v. Continental Cage Corp, 259 F.2d 87, 88 (9 Cir. 1958) (dicta).5 Carter-Wallace does not seriously question the existence or propriety of the rule, but rather contends that it comes within the exceptions relating to acquiescence and adjudication, and also that Davis-Edwards’ insolvency constitutes a basis for a further exception.
Amplifying the explanation of what on its face seems a curious doctrine that acquiescence by others can affect one who does not, Judge Hand stated, Rosenberg v. Groov-Pin Corp, supra, 81 F.2d at 48:
Its only rational justification is that those whose interest would lead them to contest validity, have upon examination concluded that contest would be fruitless, and that that examination is some assurance of the truth.
*873There are two reasons why that assumption is unwarranted here. First it is questionable whether there has been long acquiescence. The United States began to infringe some five years after the patent issued, and Judge Dooling characterized this as “significant breaches in the picture of acquiescence,” despite his ultimate conclusion that there was acquiescence. Despite the considerations adduced in the dissent, one can hardly assume that the Government deliberately infringes patents it considers valid, and Carter-Wallace agreed to waive recovery for any use prior to November 9, 1962, Carter-Wallace, Inc. v. United States, 167 U.S.P.Q. 667, 669 n. 4 (Ct.Cl. 1970). More important, a considerable number of companies in the generic meprobamate market began to infringe the patent soon after the Government, an infringer certain to mount a vigorous defense, filed its answer to Carter-Wallace’s complaint in the Court of Claims. However matters might have stood when the United States and Riverton were the only known infringers, there is no longer acquiescence in the generic meprobamate market. When general acquiescence ceases, any inference of validity which depends upon it is greatly weakened. Second, Davis-Edwards submitted an affidavit of its executive vice president, George R. Squibb, who had formerly held various offices with E. R. Squibb & Sons, a leading drug manufacturer, that the acquiescence of major pharmaceutical producers in the meprobamate patent had been due to factors other than belief in its validity,6 a question to which the judge did not seriously address himself and an allegation which, if substantiated, would deprive industry acquiescence of its usual significance here.
Carter-Wallace’s reliance on our adjudication of validity in Riverton is even more misplaced. Obviously an adjudication of validity in an action against another infringer is not res adjudicata against a different one. The theory of the prior adjudication exception is rather that a determination of validity after a full trial adds enough to what is now the statutory presumption, 35 U. S.C. § 282, to justify dispensation from the rule that a preliminary injunction will not be granted “except when the patent is beyond question valid.” But this theory rests primarily on the basis that the objections urged in the second suit have been raised and overruled in the first. Our cases have so held. George Cutter Co. v. Metropolitan Electric Mfg. Co., supra, 275 F. at 164; National Electric Prods. Corp. v. Grossman, 70 F.2d 257, 258 (2 Cir. 1934). The Riverton decision would indeed justify a temporary injunction against a conceded infringer like Davis-Edwards if its only substantial objections were those urged by Riverton, namely, that the patent application did not satisfy the statutory requirement of showing utility, 35 U.S.C. § 101; that, by failing to disclose that the drug was intended primarily for humans and that tests had been carried out on them, the patentee did not comply with the requirement that the application disclose “the best mode contemplated by the inventor of carrying out his invention,” 35 U.S.C. § 112; and that the same failure constituted a fraud on the Patent Office, see fn. 3 and accompanying text. Davis-Edwards relies not on those grounds but on others neither raised nor considered in Riverton. Perhaps also there may be instances where prior adjudication would arguably have an effect similar to acquiescence even on issues not litigated if it afforded sufficient basis for inference that there had been a “vigorous search of the art” and *874of other defenses and that the prior infringer’s counsel had selected only those issues he considered to be substantial. On the other hand, we cannot blind our eyes to the fact that researching prior art and such defenses as concealment and misrepresentation is a costly and difficult business; the selection of only a few issues, most of them apparent on the face of the patent, as in Riverton, on which an infringer thinks he can win, thus affords scant rational basis for inferring that he deliberately discarded others as lacking merit. In any event, contrary to the suggestion in Carter-Wallace’s brief that “competent counsel [in Riverton] was fully aware of such ‘new prior art’ and discarded it * * we now have an affidavit from Riverton’s counsel averring that “[t]hat suggestion is categorically untrue.”
While at first blush Carter-Wallaee’s argument that defendant’s insolvency constitutes ground for another exception in its favor is appealing, this also does not withstand analysis. In the first place, we cannot wholly accept the statement that because Davis-Edwards is in a Chapter XI proceeding, plaintiff will be unable to recover damages from continued infringement, 35 U.S.C. § 284, if the validity of the patent is sustained. Section 64, sub. a of the Bankruptcy Act gives first priority to “the costs and expenses of administration, including the actual and necessary costs and expenses of preserving the estate subsequent to filing the petition * * There can be no doubt that this section applies to Chapter XI proceedings or that “the words ‘preserving the estate’ include the larger objective, common to arrangements, of operating the debtor’s business with a view to rehabilitating it,” Reading Co. v. Brown, 391 U.S. 471, 475, 88 S.Ct. 1759, 1762, 20 L.Ed.2d 751 (1968). The Court there recognized potential tort damages of $3,500,000, substantially more than the assets of the debtor, due to losses in adjoining premises from a fire cause by the alleged (and, for purposes of the decision, assumed) negligence of a receiver for the debtor in possession, as priority claims subject only to administration expenses in a subsequent proceeding in straight bankruptcy, see § 64, sub. a. Damages for infringing a patent in the course of sales made for profit would seem an a fortiori case for priority. Cf. Diners Club, Inc. v. Bumb, 421 F.2d 396, 402 (9 Cir. 1970); 3A Collier on Bankruptcy [[62.15 at 1540 (14th ed. 1969). While the schedule filed in the arrangement proceeding does show a rather thin equity over secured debts, it is not unreasonable to believe that Davis-Edwards will have sufficient assets to respond to plaintiff’s recoverable damages for infringement if, as we direct, this action is promptly tried.6a. Moreover, this claim is partly drained of its force by the defendant’s offer of security, which we discuss in the concluding portion of this opinion.
In light of these factors, we do not feel compelled to reach a firm conclusion concerning Davis-Edwards’ contention that because Car ter-Wallace delayed sixteen months after learning of Davis-Edwards’ infringement before filing suit and another fourteen months before it sought relief pendente lite, it cannot fairly complain that another three or four months delay until a trial on the merits would cause it irreparable injury. While plaintiff might respond that it assumed it could collect damages *875until Davis-Edwards filed its petition for an arrangement (an assumption that cannot be verified on the present record), and that in any event it could not have hoped to win a temporary injunction until the patent had been sustained by this court in Riverton, cf. Clair v. Kastar, Inc., 148 F.2d 644, 646 (2 Cir.), cert. denied, 326 U.S. 762, 66 S.Ct. 143, 90 L.Ed. 459 (1945), defendant would probably reply that plaintiff must have known at least since January 1, 1969, see Marked Pleadings, Carter-Wallace, Inc. v. Riverton Laboratories, Inc., 65 Civ. 4, that the Riverton decision would settle only a few points although at least some of those here raised by Davis-Edwards had already been interposed by the United States in Carter-Wallaee’s action in the Court of Claims in March, 1968. See Carter-Wallace, Inc. v. United States, 167 U.S.P.Q. 667 (1970). The record does not clearly indicate, however, the precise nature, extent, and timing of the prior art defenses raised by the Government — although Carter-Wallace alleges that “ [defendant's counsel got prior art defenses from the Government.” In any event, we do not mean to imply that this rather inconclusive exchange over the question of delay would disentitle plaintiff to a preliminary injunction if it met the appropriate tests; at most it reflects adversely on the special equities plaintiff claims to relieve it of the general rule.
II.
Davis-Edwards contends that, once this point is reached, reversal is called for without more since, on its view, the judge, giving an undue effect to the Riverton decision, did not address himself to the question “whether the patent is beyond question valid” but rather applied the less rigorous standard of probability of success and, by taking into account the statutory presumption of validity, 35 U.S.C. § 282, in effect placed upon Davis-Edwards the burden it will have to carry at a trial, when, however, it will have had the benefits of discovery, cross-examination, and the presentation of live testimony. We do not read plaintiff’s brief as seriously disputing this, although at argument counsel pointed out that some of the judge’s remarks on which Davis-Edwards particularly relied were made after he had considered the substantive issues, and was weighing the more general considerations germane to the propriety of granting a preliminary injunction. In any event we cannot ignore the judge’s unequivocal statement that were this not a case in which “the Court of Appeals had held the patent valid, on whatever bases, I would not think there was any basis on which an injunction could be granted.” This is but one of many statements in the opinion which convince us that on this issue the defendant is right.
Nevertheless we are not disposed to reverse on this ground alone. While, for the reasons stated, we do not regard it as certain that plaintiff will not be able to recover damages if it establishes the validity of the patent, we should not subject Carter-Wallace to that risk if the record shows that a temporary injunction would have been justified under the appropriate standard. We are here reviewing not the decision of an administrative agency, but an order of a district court, and in view of the absence of oral testimony, are in as good a position as it was to determine whether a preliminary injunction would be justified under the proper standard. See Concord Fabrics, Inc. v. Marcus Bros. Textile Corp. 409 F.2d 1315, 1317 (2 Cir. 1969).
III.
The opinions in Riverton, 304 F.Supp. 357 (S.D.N.Y. 1969) and 433 F.2d 1034 (2 Cir. 1970), give the highlights in the history of the patent application, which was first filed on July 29, 1950. However, in view of the limited issues raised, they were not required to state the molecular structure of the invention. *876For this we cannot do better than quote from plaintiff’s brief:
The basic molecule, from which the invention is derived, is propane :
When the end H atoms are replaced by OH, the result is 1,3 propanediol: *
The same compound can be written as
or as
The key to meprobamate’s outstanding performance is the substitution of particular atoms or groups for X and X1 and for Ri and R2 in the 1,3-propanediol structure. In meprobamate, both X and X1 are carbamate groups
0
(NH2 — C—0). Ri is a methyl group
(CH3) and R2 is a propyl group
(CaH7) :**
The only references cited in the patent are a British patent issued to Baird in 1948 and three articles, the only relevant one of which is a 1949 article by one of the inventors, Dr. Berger, to which the parties refer as Berger I.
Davis-Edwards points to five other prior art references which, when considered along with the two just mentioned, allegedly make the invention of meprobamate obvious within 35 U.S.C. § 103. When the patent examiner rejected the initial application he did so on the basis of Baird and Berger I. Baird taught a process by which diols, that is those compounds with two hydroxyl groups (such as X and- X1) could be dicarbamated. While the particular compounds Baird was dealing with were not anticonvulsants, but rather were “useful as intermediates for the manufacture of compounds which are useful for the treatment of textile materials,” Baird’s substitution of carbamate groups for the hydroxyl groups over a broad range of diol compounds (to produce the dicarbamate ester of the parent diol) by implication embraced the particular 1,3 propanediol at issue here.
Berger I was a study of the pharmacological properties of a group of 1,3 disubstituted propanediols. A previously known drug, myanesin, a 1,2 propanediol, distinguishable from the 1,3 propanediol in that there is only one terminal hydroxyl group, the other being bonded to the middle carbon atom, was known to possess two important pharmacological properties: It caused transient paralysis of the skeletal muscles and had an anticonvulsant action. Berger I compared the pharmacological properties of a group of 12 disubstituted 1,3 propanediols with those of myanesin and found that the former “differ from myanesin in having a weaker paralyzing action and a more powerful anticonvulsant action.” More specifically, Berger injected mice with metrazol — a substance “which caused convulsions and deaths in 99% of the animals” — along with graded doses of the compound under test. The animals were observed for 90 minutes following the injection and the incidence of convulsions and deaths noted. Using myanesin as a *877standard of comparison, Berger found that a number of the compounds had better or roughly comparable anticonvulsant properties when compared with myanesin and that others showed little or no anticonvulsant activity. The most effective was 2,2-diethyl-l,3-propanediol. The next most effective was 2-methyl-2-n-propyl-1,3-propanediol, the parent propanediol of the patent in suit.
The significance of Berger I was that it identified six or seven 1,3 propanediols having effective anticonvulsant properties. After Berger I, a scientist attempting to develop an anticonvulsant would probably not continue to investigate various Ri and R2 disubstitutions but would start with the particular Ri and R2 disubstitutions in the propanediols which Berger I had found possessed effective anticonvulsant properties and work from there. It was on this basis that the initial application was rejected. The examiner reasoned that since the process of dicarbamation was known as shown by Baird and since Berger had already revealed a few compounds (the homologues of which were disclosed by Baird) with anticonvulsant properties, a chemist familiar with this would try out the Baird process of diearbamate esterification on the Berger compounds and come up with the invention in question, since the diearbamate ester of Berger’s compounds “would be expected to have similar [i. e., anticonvulsent] properties.”
Berger then set out to show that dicarbamation of the 1,3 propanediol in question was not so obvious as all that — indeed that the result was unexpected in terms of its unique anticonvulsant properties and perhaps even counterindicated —which he ultimately did to the satisfaction of the patent examiner. Whether he was fair in his efforts to demonstrate this will be considered below. The question here is whether the other alleged prior art references cited by Davis-Edwards, when considered in conjunction with Baird and Berger I, raise a substantial issue whether the invention was obvious.
Three of the five important references are concededly prior art: Berger II, a 1949 article in a technical journal; Kraft, a 1945 article; and a 1901 patent to Bonhoeffer. Berger II dealt with myanesin, whose anticonvulsant properties were compared with those of the 1,3 propanediols discussed in Berger I. Previous studies had shown that the oxidation of the terminal hydroxyl group rendered a major part of the myanesin molecule ineffective. Berger sought to prevent this by protecting the terminal hydroxyl through esterification and in particular by the substitution of the acid succinate for the terminal hydroxyl to produce myanesin acid succinate (MAS). Two conclusions in the summary in Berger II are of particular interest: that the succinate esterification of the terminal hydroxyl gave MAS a much longer duration than the parent propanediol (myanesin) although the intensity of the pharmacological effects (paralyzing action) was lower than in the parent; and that “the study of other organic esters of Myanesin therefore appears warranted.” 7 Berger II was received for publication two months prior to the like receipt of Berger I.
The importance of Berger II must be considered in light of what the patent application claimed. The principal claim was that meprobamate (i. e., the dicarbamate ester of 2-methyl-2-n-propyl-l,3-propanediol) had qualities of both inten*878sity and duration much greater (and most significantly, unexpectedly greater) than its parent propanediol. Berger II shows that the succinate esterification of a 1,2 propanediol has both a negative and a positive effect in terms of what Berger was looking for and found in meprobamate. The parent compound can be distinguished from the 1,3 propanediols which are the subject of the patent both in terms of its primary quality as a paralyzing agent rather than an anticonvulsant and in terms of its chemical structure, having only one terminal hydroxyl rather than two. Moreover, the esterification was through the use of acid succinate rather than through the use of a carbamate group, and the essential claim to novelty is that carbamate esterification produced unique and unexpected properties in the resultant compound that other organic esters would not have. On the other hand, as noted above, Berger II suggested that the study of other organic esters of myanesin appears warranted; and it is clear the article does reveal that the esterification of this propanediol lengthens its effect. Moreover, the fact that Berger II involved monosuceination does not deprive it of significance, since it is the esterification of the terminal hydroxyl to prolong its effect which is indicated, and myanesin had only one such hydroxyl, while the 1,3 propanediols had two. Hence to the extent that esterification of the terminal hydroxyl was seen as a process to have some usefulness in terms of enhancing the duration of a parent propanediol’s activity, it cannot be deemed necessarily to lead away from the invention of meprobamate and might well have suggested the actual method which was used to discover it. The facts that dicarbamation of myanesin actually “ruined it” and that there was a cleavage effect in the myanesin succinate which was not present in the meprobamate lose a good deal of the force urged by Carter-Wallace when it is recognized that dicarbamation of myanesin apparently was not discovered until after the discovery of meprobamate and that the precise activity of meprobamate, could not be learned until it was produced. In sum, while Berger II was less than a beacon illuminating the path to meprobamate, its teachings might well have been helpful to one who was experimenting in the field and was armed with the knowledge that Berger I and Baird imparted.
Kraft and Bonhoeffer can be considered together. Both indicate that the carbamate esters of a number of compounds, which are chemically distinct from the propanediols, will have a soporific, sedative or hypnotic effect, in terms of the resultant activity of the compound in question. The principal argument for the relevance of these references is that they show that carbamates generally produce the kind of effect that meprobamate successfully embodied, and that in light of the teaching of Berger I, the carbamate esters of the anticonvulsants revealed therein would be an obvious choice for one seeking to enhance the anticonvulsant effect of a parent compound through esterification, as suggested by Berger II. The principal argument to the contrary is that both in chemical structure and in terms of the precise “anticonvulsant” effect sought, these compounds were a far cry from anything closely related to the 1,3 propanediol in question. This seems sound if but only if there is a significant distinction between soporifics, hynotics, anticonvulsants, and muscle relaxants — an issue which cannot be properly determined on the present record. While there are distinctions among the various effects described, many of these compounds produced a number of them, and inquiry into one would not preclude simultaneous inquiry into another.
The two other references cited by Davis-Edwards are characterized as Yale I and Yale II. Yale was a member of a group at E. R. Squibb & Sons, which was doing research on muscle relaxants. Dr. Berger, then a professor at the University of Rochester, had acted at one time as a consultant to this group although he claims to have terminated this role in *879March 1949 and joined Carter Products Inc. in June.
The initial question with respect to these references is whether the art is prior at all. While the publication of Yale I and II concededly was after the initial patent application, both were delivered as presentations before a meeting of the American Chemical Society in September 1949, before the patent application was filed. It now appears from further discovery, as it did not at the argument before Judge Dooling, that during the late fall of 1948, Lott, one of the coauthors of the Yale articles, who was providing Berger with some of the basic anticonvulsants he was investigating, as Berger acknowledged in Berger I, had in fact provided him with the monocarbamate of the 2,2 diethyl propanediol, discussed below, and that Berger’s co-inventor, Ludwig, did in fact attend the conventions where the papers were read, although he claims to have no recollection of their presentation. While this is an area for further fact-finding, there is legitimate ground for scepticism that Ludwig, who was working in the very same areas as Yale et al. and whose co-inventor was receiving compounds from one of the Yale I and II authors, did not at least hear about the substance of these papers from his fellow scientists. In any event, whatever question the district court had about Yale, and these were serious ones, the post-hearing revelations give added support to Davis-Edwards’ position,8 particularly since it is now clear that the Yale I and II references were not part of Berger’s own work, a possibility on which Judge Dooling had relied in part in discounting them.
The significance of the Yale articles, particularly Yale II, is that they show that someone (with ordinary skill in the art) was systematically investigating essentially the same compounds whose pharmacological properties Berger I had explored, compare Berger I at 271 (table) with Yale II at 3717 (table), and indeed that the most effective anticonvulsant taught by Berger I (the diethyl 1,3 propanediol) had been monocarbamated. Yale I, which reported on the synthesis of a number of compounds similar to myanesin, discussed above, discloses myanesin monocarbamate. In light of Berger II’s disclosures, with respect to myanesin succinate and its suggestion that the study of other organic esters of myanesin was warranted, the fact that the carbamate ester, which the prior art might have suggested would produce activity of the sort Berger was investigating, was one of those synthesized cannot be regarded as by any means irrelevant. It is true that the activity of the carbamates are not reported, but that was not the specific goal of either Yale I or Yale II. Rather they were concerned with the synthesizing of a number of propanediols (Yale II, in particular, describing most of those studied in Berger I), and, as mentioned, with the monocarbamation of myanesin and the diethyl-1,3-propanediol.9 What gives Yale II particular significance is that esterification in general (see Berger II) and carbamate esterification in particular (see Yale I) was in fact being applied to a 1,3 propanediol which was known to have good anticonvulsant properties. In other words, once *880someone had synthesized the carbamate ester of a propanediol known to have good anticonvulsant properties, as Yale II did (although admittedly without indicating the pharmacological implications), it became much more likely that one interested in the effects would attempt to use carbamate esters in the relatively few propanediols with known anticonvulsant properties to test out their effects, given what the prior art suggested with respect to esterification in general and carbamation in particular.
Again, it is true that Yale II did not teach a dicarbamate; that many of the paths it might have suggested would be dead ends; and that it is a jump from the particular monocarbamated propanediol there described to the dicarbamated propanediol, meprobamate. Nevertheless, Yale II did teach the carbamation of a propanediol listed in the same table with the parent propanediol in question and having similar properties. And it was the very carbamate disclosed by Yale II which Berger had requested of Lott and received from him. That in itself goes a long way towards refuting the contention that Yale II was irrelevant to Berger’s investigations. This added information enhances the force of Judge Dooling’s statement that given Yale II “it certainly would be difficult to say * * * that there wouldn’t be an enormously important block to granting [a] product patent suddenly standing athwart the prosecution, if, as we are satisfied, the patent is critically a product patent and not a patent on undiscovered utilities within a disclosed product.”
We agree that a pharmacological patent should not be invalidated as obvious simply because, years after its issuance, an infringer has discovered recondite references in seemingly unrelated fields of chemistry, which hindsight shows might have come sufficiently close to the solution that this would have been obvious “to a person having ordinary skill to the art” to which the invention pertains, 35 U.S.C. § 103. But, except for Bonhoeffer, the prior art here discussed dated back at most five years from the patent application and was in the very field of pharmacology wherein Berger and Ludwig were working. The subject was surely one where even, perhaps particularly, a judge so thorough and so expert in patent matters could derive enlightenment from hearing the testimony of scientists, listening to their cross-examination, and even questioning them himself, rather than only the argument of lawyers based on their affidavits.10 Contrast G. B. Lewis Co. v. Gould Products, Inc., 436 F.2d 1176, 1179 (2 Cir. 1971). We would not wish at all to be understood as instructing the district judge to hold the meprobamate patent invalid as obvious in light of what could be found to be the prior art. We hold only that, as he recognized, “there was a great deal more to the point of validity than appeared at first blush” — enough more than that the district court could not properly regard the patent as valid “beyond question.”11
*881IV.
We reach the same conclusion with respect to the defense of misconduct before the Patent Office. By way of preliminary it is useful to note that, as the Court of Customs and Patent Appeals has recently written, the concept of misconduct now goes considerably beyond the classical definition of “fraud,” see Prosser, Torts, §§ 100-05 (3d ed. 1964), and includes “a wider range of ‘inequitable’ conduct found to justify holding a patent unenforceable,” Norton v. Curtiss, 433 F.2d 779, 793 (1970). The Court continued, id. at 793-794:
we do subscribe to the recognition of a relationship of trust between the Patent Office and those wishing to avail themselves of the governmental grants which that agency has been given authority to issue. The ex parte prosecution and examination of a patent application must not be considered as an adversary proceeding and should not be limited to the standards required in inter partes procedings. With the seemingly ever-increasing number of applications before it, the Patent Office has a tremendous burden. While being a fact-finding as well as an adjudicatory agency, it is necesarily limited in the time permitted to ascertain the facts necessary to adjudge the patentable merits of each application. In addition, it has no testing facilities of its own. Clearly, it must rely on applicants for many of the facts upon which its decisions are based. The highest standards of honesty and candor on the part of applicants in presenting such facts to the office are thus necessary elements in a working patent system. We would go so far as to say they are essential. It follows, therefore, that we do approve- of the indicated expansion of the types of misconduct for which applicants will be penalized.
See also Precision Instrument Mfg. Co. v. Automotive Maintenance Machinery Co., 324 U.S. 806, 816, 65 S.Ct. 993, 89 L.Ed. 561 (1945); Beckman Instruments, Inc. v. Chemtronics, Inc., 428 F.2d 555, 565 (5 Cir.), cert. denied, 400 U.S. 956, 91 S.Ct. 354, 27 L.Ed.2d 264 (1970).
Davis-Edwards’ claim of impropriety in failing to bring Berger II, Yale I and Yale II to the attention of the patent examiner hinges on the question of their importance as prior art which we have just discussed including, of course, the issue — with respect to which recent disclosures have strengthened Davis-Edwards’ position — whether Yale I and Yale II are prior art at all. It is true that none of these references specifically deals with the subject of the patent, at least with respect to its qualities of added intensity and duration as compared to its parent propanediol. On the other hand, each reveals the esterification (and in two instances carbamate esterification) of either the related compound of myanesin (myanesin monosuccinate in Berger II, myanesin monocarbamate in Yale I) or of one of the very 1,3 propanediols shown by Berger I to have good anticonvulsant properties (the 2,2-diethyl-1,3-propanediol monocarbamate in Yale II). Berger was obviously glad to learn that Lott had synthesized the carbamic acid ester of the diethyl propanediol (a 1,3 propanediol), although only in its monocarbamate form. There is no reason to think the patent examiner would not also have liked to know this, and also of Berger II and Yale I which tended to reinforce the notion that carbamate esterification was not a shot in the dark. We agree with the statement in Xerox *882Corp. v. Dennison Mfg. Co., 322 F.Supp. 963 (S.D.N.Y. 1971), that in order for nondisclosure to constitute inequitable misconduct there must be something more than negligence — that there must be “some element of wrongfulness, wilfulness or bad faith that transgresses the basic concept of doing equity.” But we do not regard it as foreclosed that after a full trial, a court might conclude that Berger’s failure to disclose any of the three articles was just that, rather than the “exercise [of] good faith judgment in deciding what matters are and are not of sufficient relevance and materiality to require disclosure.” Id. at 968-969. Here again it is important that we are dealing not with articles. Berger may have read in the remote past but with very recent publications, one by himself and two by his former colleagues at Squibb, which (were he in fact aware of the latter) must have been in the forefront of his mind. It is no answer that the patent examiner could also have found them. The paucity of references cited, with the only prior art patent relating to chemicals useful in the textile industry, raises an issue with respect to the failure to supply the examiner with more relevant material, concerning which the inventors, and perhaps also the examiner if he is available, should testify and be cross-examined. It may be said that this issue is not a true question of fact since the non-disclosure is conceded, and the problem is the proper characterization of the inventors’ conduct, which will here be decided by the judge since we gather no jury has been demanded. But it is still true, as with the similar problem concerning summary judgment, see United States v. Diebold Inc., 369 U.S. 654, 82 S.Ct. 993, 8 L.Ed.2d 176 (1962), that oral testimony, particularly demean- or under cross-examination, might powerfully assist in determining what inference to draw.
The major claim with respect to misrepresentation concerns an affidavit of July 1952 in which Berger asserted that the dicarbamates possess greater protective action than their parent propanediols (which was true) and that other esters, in particular the diacetate ester of one of the propanediols (the 2,2-diethyl-1,3-propanediol), disclosed in Berger I as having very effective anticonvulsant properties, was “less effective than the parent.” This was directed at the patent examiner’s rejection of the initial application on the ground that the carbamate esters would be expected to have the same properties as their parents and hence did not constitute invention. Berger was trying to show that not only did the dicarbamates have an unexpectedly better anticonvulsant properties but also that esters other than carbamates actually produced compounds that were less effective than their parents, thereby enhancing his claim to the inventive quality of his use of carbamates. Although the patent application was again rejected, the affidavit was nonetheless material since he had thereby demonstrated the superiority and unexpectedness of the dicarbamate esters in comparison to other esters noted in his affidavit and all he then had to do was to narrow his claim to those esters including unexpected properties and to verify the fact that the dicarbamate was unexpected in comparison to Baird, the “textile” compound.12
*883The question is whether Berger was fair in basing his representation to the Patent Office as to the inefficacy of the diacetate ester only on the 30-minute interval (which he stated) without making reference to the fact that the diacetate ester was more effective than the parent propanediol at the 60, 90, and 120 minute intervals (which he did not state), as an article co-authored by Berger before the time he made the affidavit had shown. Two things should be pointed out: The parent propanediol in question did have a stronger intensity than the diacetate ester at 30 minutes, but then significantly decreased in intensity from 60-120 minutes while the diacetate maintained some of its strength. Second, when Berger spoke in his affidavit of the superiority of the dicarbamate as compared with the parent propanediol in terms of duration, he failed to mention that the diacetate ester also showed prolongation of effects. In short, what Berger did was to compare the diacetate ester with the parent in terms of intensity (to show that it was “less effective” than its parent, unlike the dicarbamate) and then compare the dicarbamate with the parent propanediol in terms of duration, to show the greater prolongation properties of the dicarbamate, not mentioning that the diacetate also had these properties. Had this been disclosed, it certainly would have shown that other esters as well as the dicarbamate had certain desirable properties. (Cf. Berger II, with respect to myanesin succinate.) This would have detracted somewhat from the claim of the uniqueness of the carbamate ester and perhaps have led the examiner to conclude that the claims of unexpectedness with respect to carbamate were not justified. While the judge recognized that the question was “close” and the point “substantial” he thought it was not
one which at this stage of the case and for present purposes should be resolved in favor of the patentee and the plaintiff, since to make a finding * * * even for today’s purpose would require rejection of the inference that substantial people don’t do treacherous and insubstantial things when advised by counsel of standing, and that what seems at this point an unsatisfactory lack of forthcomingness may readily lend itself to explanations of greater depth.
The many cases in which patents have been held unenforceable because of inequitable conduct by “substantial people” show that the inference is a rebuttable one, as the judge clearly recognized. While the inference, and the explanation, might well suffice to sustain a negative finding as to misrepresentation if the standard here were the ordinary one of probability of success or under the rule that would be applicable after a trial, the long-standing rule in this circuit required that “explanations of greater depth” should be presented before issuing a temporary injunction in a patent infringement case. Our comments with respect to the value of oral testimony in resolving the significance of Berger’s seeming “lack of forthcomingness” concerning Berger II, Yale I and II, and the receipt of the carbamate from Lott apply equally in resolving the issue of his failure to give the patent examiner the entire story concerning the diacetate ester.
In view of the conclusions we have reached on the issues of prior art, concealment and misrepresentation by the *884Berger affidavit, we find it unnecessary to consider Davis-Edwards’ other claim of misrepresentation or its arguments with respect to patent misuse.
Y.
Our action in vacating the temporary injunction is predicated on defendant’s offer to go to trial on July 15, conditioned only on plaintiff’s giving discovery without delay. We read that condition to mean reasonable discovery; defendant must not be allowed to prolong its freedom from injunction by discovery on matters of merely peripheral significance. Rule 65(a) (2) gives the judge ample power to prevent undue delay if need for this should arise.
We also condition our order on Davis-Edwards’ carrying out its offer to pay into escrow under the supervision of the referee in bankruptcy 5% of the gross proceeds of its sales of meprobamate. This should date from February 25, 1971, when the temporary injunction issued. While normally no security would be required from a defendant against whom a preliminary injunction was issued in error, this condition is not unreasonable when the reversal is grounded on the especially severe rule applicable to the issuance of such injunctions in patent infringement cases. The equity of such a condition is all the more evident when the ability of the defendant to respond in damages in the event of judgment against it is questionable. Plaintiff points out that, although 5% was the royalty on American Home’s sales of Equanil tablets, the percentage was applied to a much higher base than the $5-$7 per pound at which defendant sells tablets under the drug’s generic name; it claims its damages would be the $20 per pound paid by manufacturers who buy from it. Defendant challenges this, arguing that, in the face of competition by other infringers, it could not have afforded to purchase the compound from Carter-Wallace; in light of the doubt whether defendant’s disappearance from the market might not have meant simply its replacement by other infringers, we are by no means convinced of the validity of plaintiff’s claim.13 It will be time to consider such questions when and if the validity of the patent has been sustained. At least plaintiff will be better off with the escrow payment than without it.
Upon these conditions we vacate the temporary injunction and leave the case in the hands of the capable and diligent district judge. We repeat that we have not in any way decided that the defenses, so ably criticized by our dissenting brother, have such merit as to entitle defendant to prevail; we have decided only that they have sufficient merit to require decision after discovery and the taking of evidence on material issues where there is fair dispute concerning the facts or the inferences to be drawn therefrom. The mandate shall issue forthwith.
. Defendant filed under Chapter XI two days after the motion was served. After having stayed prosecution of the motion, the bankruptcy court vacated the stay.
. The remarkable knowledge of the relevant chemistry and pharmacology displayed in Judge Dooling’s oral opinion may have been due in part to his having rendered a decision, a few weeks earlier, granting a preliminary injunction against another infringer, Carter-Wallace, Inc. v. Wolins Pharmacal Corp., 326 F.Supp. 1299 (E.D.N.Y.1971). Wolins apparently settled its damages for past infringement on what terms we do not know, and agreed that the injunction be made permanent.
. Carter-Wallace cites against this the statement in Riverton, 433 F.2d at 1040, The Patent Office was extremely careful and deliberate in reaching its decision of patentability in this case and there is nothing in the record to suggest that it was not fully substantiated.
But the record in Riverton does not at all reflect the issues raised here, and Judge Dooling was entirely correct in saying “I do not know what Judge Cannella [the district judge in Riverton] and Judge Anderson [who wrote for this court] would have said had they had the advantage of this material before them.”
. The presumption of validity accorded a patent regularly issued, referred to in the quoted passage and long recognized in judicial precedent, see, e.g., Radio Corp. of America v. Radio Engineering Laboratories, Inc., 293 U.S. 1, 7-8, 55 S.Ct. 928, 79 L.Ed. 634 (1934), was codified in 1952, 35 U.S.C. § 282. The Reviser’s Note states: “The first paragraph declares the existing presumption of validity of patents.” We have recently noted that the presumption “has no independent evidentiary value; rather, it serves to place the burden of proof on the person who asserts invalidity. Reasonable doubt on the issue of validity must be resolved in favor of the patent holder, but in the usual case a preponderance of the evidence determines the issue.” Rains v. Niaqua, Inc., 406 F.2d 275, 278 (2 Cir.), cert. denied, 395 U.S. 909, 89 S.Ct. 1751, 23 L.Ed.2d 222 (1969).
. While we do not suggest that all these eases would have been decided differently if they had not relied on the special rule with respect to preliminary injunctions in patent litigation, it cannot seriously be argued that the rule has been employed simply as a makeweight. Indeed, the only case to which we have been referred (and we have found no others) that rejects the general principle is Ryan v. Ideal Toy Corp, 260 F.Supp. 828 (C.D. Cal.1966). The district court, though acknowledging the enunciation of the generally accepted rule by the court of appeals of its own circuit, see Pacific Cage & Screen Co. v. Continental Cage Corp, supra, nevertheless felt its decision compelled by two earlier Supreme Court opinions, Radio Corp. of America v. Radio Engineering Laboratories, Inc, 293 U.S. 1, 55 S.Ct. 928, 79 L.Ed. 634 (1934), and Mumm v. Jacob E. Decker & Sons, 301 U.S. 168, 57 S.Ct. 675, 81 L.Ed. 983 (1937), which considered the presumption of validity that attends a patent regularly issued. Neither Supreme Court case either involved or adverted to the proper standard to be employed in the grant or denial of a preliminary injunction. There was no appeal from the district court’s decision as the parties promptly settled the case after a motion to stay the injunction pending appeal was denied by the court of appeals. Though not yet disapproved, the Ryan case has never been cited. It is for the Ninth Circuit to determine whether its rather strong statement of the rule in the Pacific Gage t£ Screen case is still to be followed by district courts in that circuit.
In addition to the special rule discussed in the text, there are objections to the grant of a preliminary injunction on the basis merely of affidavits in any case where there are disputed issues of fact. F.R.Civ.P. 65(b), and 52(a). While this court has not gone so far on this score as Sims v. Greene, 161 F.2d 87, 88-89 (3 Cir. 1947) (Biggs, J.), see contra, Ross-Whitney Corp v. Smith, Kline & French Laboratories, 207 F.2d 190, 198 (9 Cir. 1953), we have stated in no uncertain terms that where there are disputed issues of fact, as we believe there are here, a temporary injunction should not issue on the basis of affidavits save in instances of extreme urgency. SEC v. Frank, 388 F.2d 486, 490-493 (2 Cir. 1968). See also Cerruti, Inc. v. McCrory Corp, 438 F.2d 281, 284 (2 Cir. 1971).
. These were that the industry thought it “better to have an invalid patent on meprobamate than to open up the widespread manufacture of meprobamate in the absence of a patent” and that Carter-Wallace’s high pricing for meprobamate provided “a floor for the price structure of all of the subsequently developed tranquilizers.” Mr. Squibb’s affidavit also asserted that E. R. Squibb & Sons had decided not to pursue research on meprobamate since this “would not be a medically acceptable product” for inclusion in Squibb’s line and its patent counsel believed “that a strong position on the patentability of meprobamate could not be maintained.”
. The equity shown is some $111,000. The $4,500 figure noted in the dissent represents only assets not encumbered at all. While the dissent downgrades this as a “paper balance,” see fn. 1 thereto, there is no suggestion that the schedule overstates Davis-Edwards’ assets. Moreover, Davis-Edwards’ profits on meprobamate would not be the “only source to which Carter could look for satisfaction of its damage claim,” as the dissent suggests, since Davis-Edwards manufactures between 50 and 100 drug products. The crucial importance of meprobamate to the continued operation of Davis-Edwards’ business is that it is a high-volume “lead” product, without which Davis-Edwards’ ability to sell other profit-producing products, even at a low volume, would be seriously impaired.
1,3 refers to the first and third carbon atoms of the propane part of the molecule where a hydrogen atom (H) has been replaced by a hydroxyl (OH) group.
Note that the Ri and R2 substitutions both occur on the second carbon atom. For this reason, the chemical name of meprobamate is 2-methyl-2-propyl-l,3-propanediol dicarbamate.
. Berger II also noted that MAS was inferior to myanesin “because of the high paralyzing dose and smaller margin of safety.” This follows from the fact that MAS has a less intense effect than myanesin, hence larger doses are necessary to produce paralyzant activity, and these doses may prove to be lethal. But, of course, Berger II was attempting to find a means to prolong the effects of myanesin, which MAS did. Hence, as Berger II noted: “The low paralyzant activity of MAS did not discourage our interest in the drug since the therapeutic effects of Myanesin itself are exerted by subparalytic doses. There is the possibility that the paralyzing dose as measured experimentally is not a true index of the potential clinical value of a drug of this type.”
. While we could hardly use this material as a basis for reversal if the district judge had applied the proper standard, we may appropriately take note of it when we are considering whether to affirm despite his failure to do so. The production of such important material on discovery illustrates the wisdom of “the wholesome and well-settled rule” recognized by our predecessors.
. We regard it as not insignificant, however, that both Yale I and Yale II refer to other articles for the pharmacological evaluation of the properties of the compounds under consideration, which suggests that those who were reporting on the synthesizing of these compounds were not unaware of their possible use. Indeed, the reference in Yale I is to a study by Lott, one of the co-authors of the Yale articles, which tends to undermine the suggestion in Carter-Wallace’s brief that the synthesizers can somehow be distinguished from the evaluators.
. Defendant submitted an affidavit of a professor of chemistry at the Stevens Institute of Technology that meprobamate was obvious to one skilled in the art in light of the references now cited. Plaintiff responded with a counter-affidavit by a professor at Columbia.
. Although we do not rest our conclusion on Indiana General Corp. v. Krystinel Corp., 421 F.2d 1023 (2 Cir.), cert. denied, 398 U.S. 928, 90 S.Ct. 1820, 26 L.Ed.2d 91 (1970), a case on which defendant’s brief places too much reliance and plaintiff’s unwisely ignores, Judge Medina’s statement, with respect to a chemical patent for a soft ferrite manifesting a high Q-factor at high frequencies, bears repeating:
Thus the only difference between the prior art and Zerbe’s alleged invention is the ascertainment of the proportions of manganese oxide and cobalt oxide that will produce a high quality factor and a sufficient permeability. Just what will be the exact proportions to supply the needs of particular products may require a large amount of experimentation. But this very experimentation is of the essence of the art involved in this lawsuit; it is on the level of ordinary skill in the art. This is all the more true because the practitioners of
*881this art are, and of necessity must be highly educated, sophisticated persons who generally have at their disposal laboratory facilities and competent assistants.
Id. at 1030-1031. While it would appear —and we say this without intimating that it will so appear after trial — that the
prior art delineated a substantially clearer path to the result sought and ultimately achieved in Indiana General than here, the suggestion that “experimentation is of the essence of the art involved in this lawsuit,” is as applicable to the instant case as it was to Indiana General.
. The Examiner’s letter stated that “the applicants’ affidavit of July 9 fails to establish patentability on the basis of unexpected properties since
1. Broad claim 1 is inclusive of members not showing unexpected properties. For example, when Rx and R2 are ethyl and phenyl it appears the dicarbamate is inferior to the diol.
2. Claim 10 covers a member which has no unexpected properties. The mean protective dose of the ethyl ethyl dicarbamate does not appear to be sig-
nificantly different from the mean protective dose of the ethyl ethyl diol compound. This is particularly true in view of the standard error.
3. No showing has been made with respect to the homologues and isomers disclosed by Baird. Homologues and isomers are unpatentable in the absence of a verified showing of unexpected properties.
As noted in the text, all Berger had to do to meet the first two objections was to eliminate from his claim those diearba-
*883mates which were not unexpectedly effective. The other objection simply required him to show that those dicarbamates, whose properties he had persuaded the Examiner were unexpected in comparison with the anticonvulsant compounds with which he had been working (see Berger I), were similarly unexpected in comparison to the compounds disclosed by Baird. The initial rejection
in May, 1954 of the continuation-in-part application, which acknowledged a “showing of superiority,” stated only that “how much of a dose must be given before a compound is labeled ineffective has not been specified” and “no showing of superiority over the propanediols diurethanes of Baird has been presented.” Berger then submitted affidavits to meet these objections and the patent was granted.
. Cf. Power Specialty Co. v. Connecticut Light & Power Co., 80 F.2d 874, 875 (2 Cir. 1936); Electric Pipe Line, Inc. v. Fluid Systems, Inc., 250 F.2d 697, 699 (2 Cir. 1957) (lost profits awarded on findings which “justify the conclusion that but for Electric Pipe’s infringement, Fluid Systems would have made all these installations”). In American Safety Table Co. v. Schreiber, 415 F.2d 373, 378 (2 Cir. 1969), cert. denied, 396 U.S. 1038, 90 S.Ct. 683, 24 L.Ed.2d 682 (1970), cited by the dissent, lost profits were awarded on the basis of the district court’s finding that “but for defendants’ sale * * * purchasers would have bought from plaintiff in order to supply their need.”