This is an appeal from a judgment holding invalid two patents owned by the plaintiff Newburgh Moire Company, Inc. (Newburgh) and from an order holding license agreements issued by Newburgh under the patents in suit to be in violation of the antitrust laws.

Newburgh sued Superior Moire Company, Inc. (Superior) for infringement of United States Patents Nos. 2,448,145 and 2,513,646, both for processes to produce moire pattern effects in cloth, and for unfair competition. In answer Superior alleged that the patents were invalid and that there was no basis for the charge of unfair competition. Superior also filed two counterclaims. In the first, Superior sought a declaratory judgment that the patents were invalid and that Newburgh through its licenses under the patents and various other practices had violated the antitrust laws. In its second counterclaim, Superior asserted that Newburgh had breached an agreement not to announce the settlement of an earlier patent suit and asked the return of $1,000 paid in settlement. Newburgh replied, denying the material allegations of the counterclaims.

Superior then moved to dismiss the complaint, alleging that Newburgh had violated the antitrust laws and had misused its patents by setting up price-fixing provisos in the licenses issued by it. In an opinion reported at D.C.1952, 105 F. Supp. 372, the trial court held that Newburgh's license agreements did violate the antitrust laws. The opinion, however, specifically pointed out that Newburgh's cause of action for alleged trade-mark infringement and Superior's two counterclaims still remained for disposition. The opinion was handed down on June 5, 1952. Following a hearing on August 12, on December 12, 1952, the court below ordered Newburgh's cause of action for infringement dismissed but, finding that Newburgh had purged itself of the antitrust violations by renegotiations of its license agreements, omitting therefrom the price-fixing clauses but leaving therein the royalty charges, reinstated that cause of action "as of June 25, 1952."

The case then proceeded to trial. The Court below held that the first claim of each patent was invalid for lack of invention and that Superior had not been guilty of unfair competition. The court also dismissed Superior's two counterclaims, one for the return of the consideration for the settlement of the previous suit and the other for damages for Newburgh's alleged violations of the antitrust laws, D.C.1953, 116 F. Supp. 759. Newburgh appealed to this court. We dismissed the appeal as premature because of the absence of any adjudication of the other claims of the two patents, put in issue by Superior's counterclaim for a declaratory judgment that the patents were invalid. See 3 Cir., 1955, 218 F.2d 580.

The court below then held further hearings and, on January 31, 1956, entered a judgment identical with its previous one except that the two patents in suit were held invalid as to all claims. See 1955, 136 F. Supp. 923. Newburgh now appeals from that part of the judgment dismissing its cause of action for infringement because of the invalidity of the patents and from the order of December 12, 1952 dismissing and reinstating the above cause of action because of the finding that provisions of Newburgh's license agreements violated the antitrust laws.

Both patents relate to improvements in the production of moiré pattern effects in textiles. The trial court described moiré as "the process of incorporating into textiles a luster, with or without a pattern or design, by applying heat and pressure to the goods. If the weave is shifted or distorted prior to the application of heat and pressure, patterned luster effects are achieved." 116 F.Supp. at page 760.

Newburgh's patents embody the so-called "H" process, named after the inventor August Holterhoff, assignor and president of the plaintiff. Patent — 145 has two claims both of which are in issue. Claim 1 is as follows: "The method of producing moiré pattern effects in fabrics comprising moistening the fabric in confined pattern areas, applying tension to the fabric, drying the fabric while maintaining tension, folding the fabric double, and applying heat and pressure to produce the finished moiré effect."1

Patent — 646 has five claims, all of which are in issue. Claim 1 is typical. It is for "The method of producing moiré effects on fabrics comprising moistening the fabrics within confined pattern areas while applying tension to the same, drying the fabrics while maintaining the tension, doubling the thus treated fabrics and applying heat and pressure to produce the finished moiré effect."2

After examination of the prior art, the trial court concluded that "one skilled in the art would, with a general knowledge of moire and with the disclosures of the Lardy, Milhomme and Dreyfus Patents, must find the Holterhoff processes obvious advances in the art, but that is not invention." 116 F.Supp. at page 764.

Prior to the development of the plaintiff "H" process, there were four methods of producing moiré in commercial usage. In the simplest of these, mirror moiré dry fabric is doubled and passed through calender rollers, subjecting the goods to high pressure. The inside surfaces of the doubled goods attain an even luster, but no pattern is imparted. In bar moiré, before the doubled goods are fed into the calender, they are separated by a separate bar, the teeth of which distort the weave at the points of contact. As the material runs through the calender, the threads in the distorted areas intersect the threads of the opposing fabric, adding lines and shadows to the luster. In calender moiré, one of the rollers is engraved with a pattern. As a result, pressure is applied to the dry fabric only at the raised portions of the engraved roller, which imparts a design to the goods. Scratch moiré is a more complicated process. The goods, usually in a wet state, are run double through two rollers. One has a pattern, generally carved in rubber. The second roller carries scratch blades of flexible metal running along its width. These blades strike the fabric at high speed, forcing the fabric against the raised pattern of the pattern roller and distorting the weave in the pattern areas. The fabric is then sent through smooth calender rollers to impart the luster.

Only two of these processes, calender and scratch, create moiré pattern effects of the type which result from the plaintiff's "H" process. The engraved cylinder and its core used in the calender process weigh several hundred pounds. The cylinder is expensive, and engraving it is a time-consuming process. The same can be said of the pattern roller used in the scratch process. The scratch roller itself costs over $1,000; its blades must be very carefully machined and honed and replaced every few months. The smallest burr or nick in one of the blades will ruin yards of cloth. Only heavier fabrics can be used in the scratch process; and the pressure between the two rollers must be very carefully regulated — too much will damage the cloth, too little will result in an uneven pattern. In addition, "no match" cloth, in which the threads are improperly aligned, cannot be used in these traditional moiré processes.

These disadvantages and high costs are absent from the methods embodied in the patents in suit. In the "H" process, the pattern is formed by selective wetting of the cloth. Initially done by sponging, the wetting is now accomplished by passing the cloth between a dry roller and a wet pattern roller. The pattern roller is light and economical, consisting simply of pieces of leather cut in the desired pattern shapes and tacked onto a wooden roller. In Patent — 145, tension is not applied until after this step; in Patent — 646, the wetting as well takes place under tension. The cloth is then dried under tension, and distortion of the fabric occurs in the wetted pattern areas. If the cloth were either all dry or all wet, the fabric would stretch uniformly and there would be no distortion. After drying, the cloth is doubled and passed through ordinary calender rollers in the conventional fashion, producing the finished moiré effect. "No match" cloth can be used. The potential damage to the cloth inherent in the scratch process is eliminated. And unskilled operators can run the machines with a minimum of instruction.

We think it clear that the "H" process discloses improvements involving "more ingenuity * * * than the work of a mechanic skilled in the art." Cuno Engineering Corp. v. Automatic Devices Corp., 1941, 314 U.S. 84, 90, 62 S. Ct. 37, 40, 86 L. Ed. 58, citing inter alia, Hotchkiss v. Greenwood, 1851, 11 How. 248, 267, 13 L. Ed. 683. Not only do we find the improvements over the existing practice to be considerable, but in our opinion the "H" process could not be deduced from the prior art by one having "merely the skill of the calling." Cuno Engineering Corp. v. Automatic Devices Corp., supra, 314 U.S. at page 91, 62 S.Ct. at page 41. The trial court found some of the prior patents relied on by Superior not to be pertinent. We agree and will therefore restrict our consideration of the prior patents to those on which the trial court based its finding of invalidity.

Finding wetting in confined pattern areas, tension, and calendering doubled fabric separately disclosed in prior patents, the trial court considered these disclosures to be anticipatory of the "H" process. In both of the patents in suit, the operation of folding the fabric double and calendering it are described as "conventional." This operation is described in the claims only because necessary to complete the description of the "H" process. Newburgh does not contend that Holterhoff invented this step. We therefore limit our consideration to the elements of selective wetting and to the application of tension.

On the issue of selective wetting, the trial court relied on the Dreyfus and Blume Patent, No. 1,780,645, the Lardy Patent, No. 1,803,672, and the Dreyfus Patent, No. 1,860,456. The first of these patents involves the production of matte (dull) effects on cloth made of cellulose derivatives. Various types of matte paste may be applied to the whole or part of the cloth, resulting in a dull finish all over or in patterns, respectively, when the fabric is subjected to a relustering agency. The only similarity between this process and the "H" process is that both can be used to obtain a differential luster in cloth. The nature of the wetting, if the application of matte paste can be called "wetting," is quite different in the two processes. The distortion of the fabric sought in the "H" process and in other methods of obtaining moiré pattern effects plays no part in the Dreyfus-Blume process. The other aspects of moiré production are absent from that patent. In short, the patent has nothing to do with moiré. Compare Globe Knitting Works v. Segal, 3 Cir., 1917, 248 F. 495, 497.

The Lardy patent discloses a process for "restoring the luster to rayon" by wetting the fabric and then hot-calendering it. The specification points out that by moistening only certain areas, a patterned luster effect can be obtained. Although this patent can be said to involve selective wetting as does the "H" process, it does not involve distortion of the fabric nor does it produce moiré effects. We do not consider the method of the Dreyfus patent sufficiently different from the Lardy process to require separate discussion, except to point out "imparting differential lustre effects" is its sole aim.

In respect to tension, the trial court pointed out that the second Milhomme Patent, No. 1,821,392, relating to the scratch process of producing moiré, describes "a method of applying tension to the fabric as it is drawn around the pattern roll." 116 F.Supp. at page 762. The trial court also states that, although the Lardy patent, for example, does not mention tension "Yet tension must certainly be applied in order to run the fabric through the ovens and calenders * * *." Id. at page 763. The trial court then goes on to state: "If the application of tension [in the "H" process patents] is not novel, then is the degree and manner of application novel. For if not, then certainly the Lardy Patent anticipates the patents in suit. The patents in suit do not specify how or how much tension is to be applied. It is asserted that the Holterhoff Patents rely on the skill of the mill worker to judge the application of tension; that experienced mill hands are expected and able to feel the requirements as the fabric is run. The court agrees with this assertion. But here plaintiff seeks to prove novelty of patents on the ground that the patent directs tension to do more than merely force the fabric through roller and calender, namely, to dislocate the warp and the weft, and this in combination with moisture and quick calender to produce the moire effect. But if the tension were a key factor of the process, i. e., if tension were here used to do more than force the fabric to flow through the mill, would the patent be devoid of more specific direction as to amount of tension or the manner of application of that tension? I think not. And if the tension required were to be no more than that normally employed to run fabrics through processing mills, does not the Lardy Patent anticipate the patents in suit? I think yes. Though Lardy does not specify tension, is not positive tension necessary to cause the fabric to flow through the mill, and cannot Lardy leave the implication as well as the application of tension to the skilled artisan? I feel he can." Id. at pages 763-764.

The proposition is aptly stated, but we think that the dilemma posited by the trial court is not a real one. The holding that the amount of tension used in the "H" process is no more than that required to pull the cloth through the machinery does not find support in the record. We are of the view that it is not necessary to specify in the patents the exact amount of tension to be applied. Since, unlike the scratch process, it is possible to use a wide variety of fabrics in the "H" process, at least the maximum tension, ex necessitate, must vary with the strength of the cloth. We think that the amount of tension required can readily be determined for each type of cloth by experimentation. The amount of tension to be applied is neither so obscure nor of such a limited range that the failure to specify it renders the patent invalid for nondisclosure.

We are of the view that the trial court erred in failing to take into account the necessary connection in the "H" process between selective wetting and the application of tension in obtaining the distortion of the fabric necessary to create the moiré pattern effects when the fabric was doubled and calendered. The separate use of selective wetting and tension for other purposes is irrelevant unless it points implicitly to the "H" process. Otherwise, the patents in suit would have been anticipated the first time a washerwoman sprinkled water on clothes before ironing them. We can see no foreordination of the "H" process in the prior art. See Loom Co. v. Higgins, 1882, 105 U.S. 580, 591-592, 26 L. Ed. 1177. As to the standards by which invention is to be tested see R. H. Palmer Company v. Luden's, Inc., 3 Cir., 1956, 236 F.2d 496.

We are confirmed in our view by the fact that fifteen years elapsed between the date of the most recent patent relied upon by Superior and the date of filing of the first of the two "H" process patents. The industry's acceptance of the "H" process and its drastic simplification in creating moiré pattern effects support our conclusion that the "H" process was a patentable invention. See Minerals Separation, Ltd. v. Hyde, 1916, 242 U.S. 261, 270, 37 S. Ct. 82, 61 L. Ed. 286; Crozier-Straub, Inc., v. Reiter, D.C.E.D.Pa.1929, 34 F.2d 577, 578.

From this generic consideration of the "H" process, we turn to the specific patents in suit. For the reasons stated above, we find Claim 1 of Patent — 145 valid. We also find Claim 2 of that patent valid as "a patentable species under the genus" of Claim 1. See In re Brogden, 1946, 156 F.2d 82, 86, 33 C.C.P.A. Patents 1181, 1186.

As to Patent — 646, Superior contends that it is invalid and is merely an attempt to prolong the patent monopoly. We agree with this contention. The only difference between the two patents is that — 646 specifies that the wetting, as well as the drying, is to be done under tension. This addition was not essential to the successful utilization of the "H" process nor was it itself an invention but simply an improvement which naturally would suggest itself as the disclosures of — 145 were put to use. See Miller v. Eagle Manufacturing Co., 1894, 151 U.S. 186, 14 S. Ct. 310, 38 L. Ed. 121. The fact that the two patents were copending does not save — 646. See In re Copeman, 1943, 135 F.2d 349, 30 C.C. P.A. Patents 962, and cases cited therein.

As to the valid patent, No. — 145, the evidence clearly supports a finding of infringement. It is unnecessary to take the space to review the record on this issue. So much for the appeal from the dismissal of Newburgh's cause of action for infringement on the ground of invalidity of the patents in suit.

In its opinion, 105 F.Supp. at page 376, handed down on June 5, 1952, the court stated: "The first cause of action [for patent infringement] alleged in the complaint is hereby dismissed but without prejudice to renewal upon plaintiff's renegotiating his [sic] license agreements, omitting those clauses incompatible with the anti-trust laws."

The court below, however, entered no judgment as to the first cause of action on June 5, 1952. On August 12 the court discussed with counsel the kind of order which was to be entered. On December 12, 1952 it entered an order which reads in part as follows: "This cause having again come on for hearing on August 12, 1952, * * * and having received evidence of plaintiff's [Newburgh's] renegotiation of its license agreements on June 25, 1952, omitting the clauses which were held to be incompatible with the Anti-Trust Laws, and defendant * * * having conceded that the evidence constituted sufficient proof of a renegotiation of the license agreements in accordance and conformity with the court's opinion, and the court having accepted the evidence and having held that the plaintiff had thereby fully purged itself of its violations of the Anti-Trust Laws as of June 25, 1952, * * * [it is] Ordered that the motion to dismiss the complaint be and the same is hereby granted as to the first cause of action effective only to June 25, 1952, but said first cause of action is hereby reinstated as of June 25, 1952, the date upon which plaintiff fully purged itself of its violations of the Anti-Trust Laws, and said motion is denied as to the second cause of action * * *." It will be observed that the court below adopted the unusual course of dismissing the complaint and reinstating it in one and the same order, both the dismissal and the reinstatement operating nunc pro tunc as of the same prior day, June 25, 1952, Newburgh having, according to the order, "purged itself" of its violations of the antitrust law. The court below cited no decision in support of the cleansing effect of such purgation but there is authority for such a proposition in general.3 Moreover, it would appear both from the transcript of the hearing of August 12, 1952 and from the order of the court entered December 12, 1952 that Superior conceded that Newburgh had complied with the court's opinion. Superior conceded further that Newburgh had purged itself.

But the antitrust phase of the case is not rendered moot by the purging as Superior on its brief and in its oral argument contends. The licenses granted by Newburgh to Hollywood Piece Dye & Finishing Works, Inc. (Hollywood) and to Progressive Silk Finishing Works, Inc. (Progressive) contained price-fixing provisos, the nature of which is revealed by the deposition referred to in the footnote.4 Since Patent No. — 145 is valid we must determine whether the price-fixing provisos as contained in the licenses prior to the renegotiation on June 25, 1952 were in violation of the antitrust law.5 This determination is necessary because if the price-fixing provisions were not in violation of the antitrust laws Newburgh would be entitled to injunctive relief and to an accounting from Superior because of Superior's use of the "H" process and the order of the court below dismissing the first cause of action would have to be reversed as Newburgh desires.

On this antitrust aspect of the case, it is important to note that there appear to be only five moiré finishers in the United States; Newburgh itself, Superior, Hollywood, Progressive, and a newcomer in the field, unidentified on this record.6 By granting Hollywood and Progressive licenses containing minimum price-fixing provisos Newburgh thereby involved three of the five moiré finishers in price fixing. There is no Supreme Court authority as to whether such a degree of price-fixing licenses within an industry violates the antitrust laws. In United States v. United States Gypsum Co., 1950, 340 U.S. 76, 71 S. Ct. 160, 95 L. Ed. 89, the Court, referring to its decision in an earlier case of the same name, said: "It was not necessary to reach the issue as to whether a mere plurality of licenses, each containing a price-fixing provision, violates the Sherman Act. It is not necessary now." Id., 340 U.S. at page 85, 71 S.Ct. at page 167. Cf. Westinghouse Elec. Corp. v. Bulldog Electric Prod. Corp., 4 Cir., 1950, 179 F.2d 139, where this specific issue was not in focus. The court below in the case at bar took the position that under Section 1 of the Sherman Act, 15 U.S.C.A. § 1, a licensor cannot be permitted to impose a price-fixing proviso upon its licensees though there is a "mere plurality" of licensees containing price-fixing provisions, the licensor and the licensees constituting what is a bare majority of the manufacturers or processors in an industry. 105 F. Supp. 372.

Some of the authorities pertinent to the issue of price-fixing licenses in the patent field are discussed in the thoughtful opinion of the court below. The principle enunciated in Bement & Sons v. National Harrow Co., 1902, 186 U.S. 70, 22 S. Ct. 747, 46 L. Ed. 1058, and followed in United States v. General Electric Co., 1926, 272 U.S. 476, 47 S. Ct. 192, 71 L. Ed. 362, permitted a patentee to fix the price the licensee of the patent may charge for the device. It is difficult to say how much of this principle remains in the law. This question came to a head but was not resolved in United States v. Line Material Co., 1948, 333 U.S. 287, 312, 68 S. Ct. 550, 92 L. Ed. 701. The Line decision dealt with arrangements made between two patentees for the cross-licensing of their interdependent product patents and for the licensing by one of them of other manufacturers under both patents. The Court held that price maintenance by the various types of agreements constituted a violation of Section 1 of the Sherman Act. Four Justices thought that the General Electric decision should be overruled. Three thought otherwise and dissented. One Justice did not participate. Mr. Justice Reed, voicing an opinion for a divided Court, distinguished the General Electric decision and construed the patent statutes as giving a patentee a right to license "another to make and vend at a fixed price." (Emphasis added.) Id., 333 U.S. at page 312, 68 S.Ct. at page 563. The court below in the case at bar construed the word "another" employed by Mr. Justice Reed as meaning "one." 105 F.Supp. at page 374.

On the same day as the Line decision a unanimous Supreme Court, albeit with two Justices not participating, decided United States v. United States Gypsum Co., 1948, 333 U.S. 364, 68 S. Ct. 525, 92 L. Ed. 746. In Gypsum there was no cross-licensing, but rather industry-wide licensing under several patents held by Gypsum Company. Mr. Justice Reed stated for the Court: "We think that the industry-wide license agreements, entered into with knowledge on the part of licensor and licensees of the adherence of others, with the control over prices and methods of distribution through the agreements and the bulletins, were sufficient to establish a prima facie case of conspiracy." Id., 333 U.S. at page 389, 68 S.Ct. at page 539. The licensing agreements extended to unpatented as well as to Patented Gypsum products but the Supreme Court did not seem to rely on this fact. Mr. Justice Reed went on to say: "Even in the absence of the specific abuses in this case, which fall within the traditional prohibitions of the Sherman Act, it would be sufficient to show that the defendants, constituting all former competitors in an entire industry, had acted in concert to restrain commerce in an entire industry under patent licenses in order to organize the industry and stabilize prices. That conclusion follows despite the assumed legality of each separate patent license, for it is familiar doctrine that lawful acts may become unlawful when taken in concert. Such concerted action is an effective deterrent to competition; * * *." (Emphasis added.) Id., 333 U.S. at page 401, 68 S.Ct. at page 544.

In the second Gypsum case, 1950, 340 U.S. 76, 71 S. Ct. 160, 95 L. Ed. 89, the Supreme Court explained that conspiracy to restrain commerce was the basis of the first Gypsum decision. Mr. Justice Reed stated: "There was no holding in our first opinion in Gypsum that mere multiple licensing violated the Sherman Act." Id., 340 U.S. at page 84, 71 S.Ct. at page 167. Footnote 4 cited to the text, to the sentence just quoted, seems to indicate that the dissenters in Line Material would not have concurred in the first Gypsum opinion were it not for the conspiracy element in that case. The footnote should be repeated here. "The dissenters in Line * * * joined in the United States Gypsum opinion, since the concerted action in the United States Gypsum case was thought to violate the Sherman Act, despite their view that the mere multiplication of licenses, as in Line, `produces a repetition of the same issue (as in General Electric) rather than a different issue.' 333 U.S. [at page] 354, 68 S.Ct. [at page] 583."

In United States v. New Wrinkle, Inc., 1952, 342 U.S. 371, 72 S. Ct. 350, 96 L. Ed. 417, a unanimous Supreme Court, one Justice not participating, held invalid patent license agreements fixing prices "throughout substantially all of the wrinkle finish industry * * *". Id., 342 U.S. at page 372, 72 S.Ct. at page 351. Cf. Times-Picayune Publishing Co. v. United States, 1953, 345 U.S. 594, 73 S. Ct. 872, 97 L. Ed. 1277. The facts of the case at bar are different from those of New Wrinkle for here only three-fifths of an industry, not "substantially all" of it is involved in the price-fixing. But it is difficult to distinguish logically between a situation whereby 85% of an industry, as in the General Electric case, is gobbled up by General Electric and a price-fixing license granted to a single company, Westinghouse, and where at least 90% of the wrinkle finish industry is engrossed by price-fixing licenses issued to two hundred companies.7 It can be argued with much persuasivenses that the difficulties in controlling two hundred licensees are greater than in controlling one. It must be conceded that prices to the public are controlled to some degree at least by the amount of the royalties which licensees must pay and that therefore even ordinary royalty licenses smack of price-fixing. Of course the patentee who licenses two hundred licensees on a royalty basis as a practical matter can exercise less control over them than he could over one licensee so licensed. The difficulty of control rises with the number whether the licensees require the payment of royalty or contain price-fixing provisos, for legal expenses rise with the number of suits to be brought. We can find no very sound basis for distinguishing between General Electric and New Wrinkle. The explanation probably lies more in history than in logic for history is an untidy housekeeper.

As early as 1927, the Supreme Court held price-fixing in commerce to be a per se violation of the Sherman Act. United States v. Trenton Potteries, 273 U.S. 392, 397, 47 S. Ct. 377, 71 L. Ed. 700. Newburgh's original patent licenses with Hollywood and Progressive were price-fixing agreements and therefore on the face of it were illegal per se. The Patent and Commerce Clauses of the Constitution present no conflict.8 Were it not for the General Electric decision one could argue persuasively that a line could be drawn between the patent laws and the antitrust laws. The division would permit an inventor on the one hand to license his product for manufacture and sale or his process for use for such payments to be made by the licensee to the licensor as should be agreed upon. On the other hand, the inventor would be prohibited from fixing the price at which his product should be sold by his licensee to the public or the rate at which the licensee could employ the process in the public market.

But it is not necessary to go so far as to set up such a division in the case at bar. Here we reach no further than the issue reserved by the Supreme Court in each of the Gypsum cases. At worst, we think that the patent laws were not intended to empower a patentee to grant a plurality of licenses, each containing provisions fixing the price at which the licensee might sell the product or process to the company, and that, if a plurality of licenses are granted, such provisions therein are prohibited by the antitrust laws. The course pursued by Newburgh in the case at bar transcended the authority of the General Electric decision.

The order of December 12, 1952 dismissing the first cause of action will be affirmed. That portion of the judgment of the court below declaring Patent No. — 145 to be invalid will be reversed. That part of the judgment declaring Patent No. — 646 to be invalid will be affirmed. The case will be remanded with directions to enter judgment in accordance with this opinion.