United States of America v. Studiengesellschaft Kohle, M.b.h., Director Max Planck,institut Fur Kohlenforschung, Appellant,hercules Incorporated, et al, 670 F.2d 1122 (D.C. Cir. 1981)

Appeal from the United States District Court for the District of Columbia (D.C.Civil Action No. 1255-70).

Arnold Sprung, New York City, with whom Nathaniel Kramer, New York City, and Barry E. Cohen, Washington, D.C., were on the brief, for appellant.

Bruce E. Fein, Atty., Dept. of Justice, Washington, D.C., with whom John J. Powers, III and Kurt Shaffert, Attys., Dept. of Justice, Washington, D.C., were on the brief, for appellee.

Before ROBINSON, Chief Judge, McGOWAN, Senior Circuit Judge, and LOUIS F. OBERDORFER,^*  United States District Judge.

Opinion for the Court filed by District Judge OBERDORFER.

OBERDORFER, District Judge:

This is a civil antitrust enforcement action brought by the United States against Studiengesellschaft Kohle m.b.H. (S.K.) and its licensees. Essentially, the complaint challenged certain arrangements which granted an exclusive license to sell the product of a patented process as an unreasonable restraint of trade and an attempt to monopolize a part of trade or commerce in violation of Sections 1 and 2 of the Sherman Act, 15 U.S.C. §§ 1, 2. Before trial, the district court denied defendants' motion for summary judgment, holding that the patents did not immunize the challenged arrangements from antitrust scrutiny. United States v. Studiengesellschaft Kohle, m.b.H., 426 F. Supp. 143 (D.D.C. 1976). After a trial without a jury, the district court found that the license provisions in question, stripped of any patent law protection, violated Sections 1 and 2 of the Sherman Act. The court entered a decree enjoining defendant S.K.¹  from enforcing any agreement limiting sales of the product of its process patent, and requiring defendant to license the patented process to all applicants at a reasonable royalty. This appeal followed. For reasons stated below, we reverse the judgment of the district court and remand with instructions that it enter judgment for the defendant.

Between 1953 and 1954, Dr. Karl Ziegler, the Director of the Max Planck Institute in Mulheim, West Germany, developed a new process for the production of aluminum trialklys (ATAs), a catalytic agent and chemical reactant. For this and other related discoveries of organic-metallic catalysts and processes, Dr. Ziegler was awarded the 1963 Nobel Prize in Chemistry. He was also awarded with a number of U.S. patents on his process for producing ATAs.

Prior to Ziegler's invention, ATAs were known but had no commercial uses. J.A. 163. Experience with the invention shows that ATAs produced by the Ziegler process cost an estimated 5% of what it cost to produce ATAs using the prior art. As the district court found, Ziegler's process "is so economical that no other process can be commercially competitive with it." J.A. at 95. Primarily as a result of these economies, the number of uses for ATAs increased dramatically. ATAs are now consumed in large quantities as a reactant in the manufacture of biodegradable household detergents and as a catalyst in the manufacture of synthetic rubber for tires. J.A. 90, 95. But since Ziegler did not discover the product ATAs, he was not awarded a patent on ATAs, but only on the process which he invented for their more economical manufacture.

In 1954 Dr. Ziegler concluded an agreement with Hercules Incorporated (formerly Hercules Powder Company) to exploit his process patents. Hercules had followed his research for a number of years and had previously expressed interest in obtaining licenses for the exploitation of his inventions. This Ziegler/Hercules agreement, designated by the parties as the "Technical Field Contract," granted Hercules a nonexclusive license to manufacture ATAs by the Ziegler process for use in Hercules' own manufacturing operations. It also granted to Hercules "an exclusive license to sell in the United States the aluminum trialkyl produced within the scope of the Technical Field." J.A. 143. Ziegler's agreement with Hercules permitted him to grant nonexclusive licenses to a number of other manufacturers as long as they used, but did not sell, the ATAs manufactured pursuant to the license. Accordingly, Ziegler granted licenses to Ethyl Corporation (Ethyl), Continental Oil Co. (Conoco) and others to use his process to manufacture ATAs for internal use.

In 1959 Hercules entered into a joint venture with Stauffer Chemical Co. in an effort to exploit the exclusive license to manufacture ATAs for sale. Together they formed Texas Alkyls, Inc. (Texas); Stauffer contributed capital while Hercules contributed its exclusive license. This transaction purported to give Texas the exclusive right to sell ATAs in the United States, diluted only by licenses authorizing ATA users such as Conoco and Ethyl to manufacture ATAs for internal consumption. Several nonexclusive licensees sought licenses from Ziegler to use his patented process to produce ATAs for sale in competition with Texas, but he consistently denied these applications.

Ethyl was one such disappointed licensee. In 1959 it reacted to Ziegler's denial of its application for a license to sell by filing a declaratory judgment action in the United States District Court for Delaware challenging Ziegler's right to enforce the license provisions banning ATA sales by Ethyl. Before that case came to issue, the parties agreed that, whatever the outcome, Ethyl would be permitted to sell ATAs-by right if the result was in favor of Ethyl, and if otherwise, according to the terms of a special license that would require Ethyl to pay Hercules an additional 2% royalty on sales of ATAs.²  The Delaware Court ultimately found that the license restrictions were a valid exercise of the monopoly power inherent in Ziegler's process patent. Ethyl Corp. v. Hercules Powder Co., 232 F. Supp. 453 (D. Del. 1964). Ethyl took no appeal, but began selling ATAs, subject only to an obligation to pay the 2% royalty agreed upon in partial settlement of the Delaware litigation. Only Hercules and Ethyl have sold industrial quantities of ATAs since that time.

On April 24, 1970, the United States filed the complaint in this case against Dr. Ziegler, Hercules, Stauffer, and Texas. After extensive discovery, defendants moved for summary judgment on the ground that the exclusive license to market ATAs manufactured by the Ziegler process was protected by the patent laws and thus immune from attack under the antitrust laws. The district court denied this motion, finding, contrary to the Delaware decision, that the patent laws did not protect the Ziegler licensing agreement. J.A. 77; 426 F. Supp. 143, 149 (D.D.C. 1976). The court emphasized that S.K. (which was substituted as a defendant upon the death of Dr. Ziegler) held a process patent and not a product patent. The court concluded that a restriction on sales of the product by the nonexclusive process licensees exceeded the scope of a process patent. According to the district court, Ziegler's

"... patent claim gave him the right to exclude others from marketing, using, or selling his process. He therefore never had protection under the patent laws when he sought by whatever means to extend his process claim to restrict use or distribution of the unpatented product, the ATAs.... The holder of a process patent could not ... license several companies to use the process and attempt to limit the manner in which some of those companies decided to use the ultimate product."

426 F. Supp. at 148, 149 and n.3. Having denied defendants' motion for summary judgment, the court set the case for trial on the issue of whether the agreements limiting sales violated the Sherman Act.

With trial pending, defendants Hercules, Stauffer and Texas agreed to consent judgments and decrees.³  After a trial in which S.K. was the sole remaining defendant, the district court entered findings of fact and conclusions of law. The court found essentially four anticompetitive effects from the restrictions at issue: (1) the nonexclusive licensees, prospective sellers of ATAs, were excluded from the market, (2) the price of ATAs exceeded competitive levels, (3) the development of new uses for ATAs was retarded because of their supracompetitive price, and (4) trade in certain aluminum alkyls other than ATAs was restrained as a result of the restrictions on the Ziegler process. J.A. 125-27. Without discussion of relevant authorities, the district court concluded that the agreements constituted an attempt to monopolize the sale of the ATAs in the United States in violation of Section 2 of the Sherman Act, 15 U.S.C. § 2, and were illegal restraints of trade under Section 1 of the Sherman Act, 15 U.S.C. § 1, both under a per se rule and as measured by the rule of reason.⁴  On March 15, 1979, the court entered an amended final decree that, inter alia, prohibited defendants from enforcing or granting limited licenses and mandated compulsory licensing at reasonable and nondiscriminatory rates of the Ziegler process, including the right to sell any ATAs so produced. See J.A. at 130.

On appeal, the parties make a number of contentions, not all of which need be addressed in view of the conclusion we reach. Appellant renews its argument below in support of its motion for summary judgment that its conduct is protected by the patent laws because (1) it effected no greater restraint than the patent granted, and (2) it could have legally imposed a much greater restraint by granting Hercules an absolutely exclusive license to use the Ziegler process, thereby automatically vesting Hercules with the exclusive right to sell ATA's manufactured by that process. In addition, appellant argues that the restraints challenged here should be upheld as permissible quantity restrictions on the use of the process, citing Q-Tips, Inc. v. Johnson & Johnson, 109 F. Supp. 657 (D.N.J. 1951), aff'd, 207 F.2d 509 (3rd Cir. 1953), cert. denied, 347 U.S. 935, 74 S. Ct. 630, 98 L. Ed. 1080 (1954).⁵  Finally, appellant argues that the government's failure to prove a relevant market is fatal to its claim under both Section 1 and Section 2.⁶ 

The government grounds its position to the contrary on the summary judgment ruling by the district court below, that the agreements were designed to expand a legal monopoly of the process into an impermissible monopoly of the unpatented product. Arguing that a process patentee has no authority to control the sales of the unpatented product of the process, the government maintains that such an attempt is per se illegal as well as effecting an unreasonable restraint of trade. It also disagrees with the analogy to quantity limitations which defendant asserts, arguing that the license here restrains the use of the products and not the amount of the product which the licensee may produce. Finally, the government argues that it is not required to prove a relevant market, and that in any event the district court made adequate findings that ATAs constitute a relevant market. We conclude that, even assuming arguendo that ATAs constitute a relevant market and that the agreements are not defensible as quantity restrictions, the judgment must be reversed because defendant did not expand its monopoly or impose restraints beyond the scope of the monopoly which its patent gave it.III. ANALYSIS

The patent laws, authorized by the Constitution, were enacted by Congress to stimulate invention and reward innovation by granting the patentee a 17-year monopoly of the making, using, and selling of the patented invention. U.S.Const., art. I, sec. 8; 35 U.S.C. § 154. Such a grant is in inevitable tension with the general hostility against monopoly expressed in the antitrust laws, 15 U.S.C. § 1 et seq. Therefore, courts normally construe patent rights narrowly in deference to the public interest in competition. Mercoid Corp. v. Mid-Continent Inv. Co., 320 U.S. 661, 665-66, 64 S. Ct. 268, 271, 88 L. Ed. 376 (1944); United States v. Masonite Corp., 316 U.S. 265, 278-80, 62 S. Ct. 1070, 1077-78, 62 S. Ct. 1302 (1942). As a result, there has been a stream of litigation down through the years flowing from the conflict between the monopoly rights created by the patent laws on one hand and the national policy favoring competition expressed in the antitrust laws on the other.⁷ 

The essential rights of a patentee may be briefly summarized. A patentee has the right to exclude others from profiting from the patented invention. Zenith Radio Corp. v. Hazeltine Research Inc., 395 U.S. 100, 135, 89 S. Ct. 1562, 23 L. Ed. 2d 129 (1969); Continental Paper Bag Co. v. Eastern Paper Bag Co., 210 U.S. 405, 424-25, 28 S. Ct. 748, 753-54, 52 L. Ed. 1122 (1908). This includes the right to suppress the invention while continuing to prevent all others from using it, Continental Paper Bag Co., supra, to license others, or to refuse to license, and to charge such royalty as the leverage of the patent monopoly permits. See Brulotte v. Thys Co., 379 U.S. 29, 33, 85 S. Ct. 176, 179, 13 L. Ed. 2d 99 (1964) (dictum); W.L. Gore & Assoc. v. Carlisle Corp., 529 F.2d 614, 622-23 (3d Cir. 1976). A patentee may grant one exclusive license or may grant many licenses. See E. Bement & Sons v. National Harrow Co., 186 U.S. 70, 22 S. Ct. 747, 46 L. Ed. 1058 (1902); United States v. Westinghouse Elec. Co., 648 F.2d 642 (9th Cir. 1981).

A license is an agreement by the patentee, usually for a consideration, not to sue the licensee of the patent for infringement of the patent. A patentee has the right to an injunction barring use of the patented process or product without his permission and the additional right to recover damages caused by such an infringement. The license waives this right to judicial relief against what, but for the license, would be an infringement. Frequently, a patentee grants licenses on certain conditions, in addition to the requirement that the licensee pay royalties. The validity of various restrictions in licensing agreements has been the focus of much patent-antitrust litigation. See generally P. Areeda, Antitrust Analysis P 410, 411 (3d ed. 1981).

A patent may be awarded for either a product or a process. A product patent creates a monopoly over the manufacture, use, and sale of a product; a process patent creates a monopoly over the manufacture, use, and sale of a process. The essential difference between the two relates to scope. A product patent gives the patentee the right to restrict the use and sale of the product regardless of how and by whom it was manufactured. A process patentee's power extends only to those products made by the patented process. Merrill v. Yeomans, 94 U.S. 568, 24 L. Ed. 235 (1876); see pages 1133-1134, infra. A process patent thus "leaves the field open to ingenious men to invent and to employ other processes." 1 A. Walker, Patents § 23 at 140 (2d Deller ed. 1964).

A sale of a product made by a patented process does not itself infringe the patent; it is the unauthorized use of the process that infringes the patent. See, e.g., Koratron Co. v. Lion Uniform, Inc., 449 F.2d 337, 338 (9th Cir. 1971); In re Amtorg Trading Corp., 75 F.2d 826 (C.C.P.A. 1935), cert. denied 296 U.S. 576, 56 S. Ct. 102, 80 L. Ed. 407 (1935).⁸  Here the several licenses contained conditions restricting sales by the nonexclusive licensees of the product of the patented process. Such sales would violate the license agreement and therefore expose the seller to infringement claims by Hercules and S.K., unless the antitrust laws rendered those conditions unenforceable. See Ethyl Corp. v. Hercules Powder Co., supra. The validity of these restrictions in the nonexclusive licenses under the antitrust laws is the issue which we must resolve.

The district court treated the question of the interaction between the patent laws and the antitrust laws here in two stages: it first determined that the restriction imposed was outside the scope of the patent protection, then examined the unprotected restriction to determine if it constituted an attempt to monopolize or an unlawful restraint of trade. See 426 F. Supp. at 149. Although in the second step of its analysis the court purported to assess the reasonableness of the competitive effects of the restriction, in fact, its method of analysis had the effect of applying a per se rule. This is so because once the protection of the patent was removed, the license conditions, like the patent itself, inevitably had the effect of restricting competition.

Such a formalistic, two-step analysis forecloses adequate consideration of the fundamental fact that a patent by definition restrains trade, and in effect makes most exclusive patent licenses per se violations of the antitrust laws. But as the Supreme Court noted in E. Bement & Sons v. National Harrow Co., 186 U.S. 70, 91, 22 S. Ct. 747, 755, 46 L. Ed. 1058 (1902); "(t)he very object of (the patent laws) is monopoly.... The fact that the conditions in the contract keep up the monopoly does not render them illegal." Thus, as appears more fully below, we conclude that a rule of reason rather than a per se rule applies here. Under our analysis, the protection of the patent laws and the coverage of the antitrust laws are not separate issues. Rather, the conduct at issue is illegal if it threatens competition in areas other than those protected by the patent, and is otherwise legal.⁹  The patentee is entitled to exact the full value of his invention but is not entitled to endanger competition in other areas by manipulating his patent monopoly. It was thus error to consider the scope of the patent protection irrespective of any competitive effects in the first phase of the case, and then rule separately on the anticompetitive effects of the arrangement without consideration of the protection of the patent.

If the four anticompetitive effects found by the district court are examined, it is clear that all of them were restraints on what the patent lawfully protects: ATAs manufactured by the Ziegler process. The effects as stated were (1) the exclusion from the market of potential sellers of ATAs manufactured by that process, (2) supracompetitive prices, (3) retarded development of new uses for ATAs because of supracompetitive prices, and (4) restraint on certain other aluminum alkyls which resulted from restrictions on the Ziegler process. None of these restraints go beyond what the patent itself authorizes. Such an exclusion of competitors and charging of supracompetitive prices are at the core of the patentee's rights, and are legitimate rewards of the patent monopoly. Brulottee v. Thys Co., supra; Zenith Radio Corp. v. Hazeltine Research, Inc., supra. Similarly, the restraint on aluminum alkyls other than ATAs produced by the Ziegler process was a legitimate reward of the patent. All of these "anticompetitive effects" of the restriction on sales by licensees would have resulted from a conventional grant of an exclusive license to Hercules to practice the patented process. In fact, the anticompetitive effects of such an exclusive license would be even greater.

The government contends and the district court concluded that defendant's conduct in this case was an attempt to extend its monopoly of the process for making ATAs into a monopoly of the product, ATAs, which are themselves unpatented. As already noted previously, the defendant was not given a product patent on the ATAs because he did not invent them. S.K. thus has no right to prevent others from selling ATAs which are manufactured by another process. An attempt by defendant to expand its monopoly to cover such an unpatented product manufactured by another process would be an attempt to enlarge its monopoly beyond what the patent law gives it and would thus be subject to antitrust attack. For example, if S.K. had required its licensees to refrain from selling any ATAs, even if they were made by other processes, it might well be in violation of the antitrust laws. Cf. Compton v. Metal Prods., Inc., 453 F.2d 38 (4th Cir. 1971), cert. denied, 406 U.S. 968, 92 S. Ct. 2414, 32 L. Ed. 2d 667 (1972). A requirement by defendant that any licensee who discovered an alternative process grant back to S.K. a license under that process might similarly be illegal, since such a license might dampen the incentive to invent around the patent. But S.K. did neither of these things. It merely restricted the sale of ATAs which were manufactured by its process. Such a restriction does not in any way operate to limit competition between ATAs manufactured pursuant to the Ziegler process and ATAs manufactured by other processes, now or hereafter available. Nor does this restriction make it less likely that such process will be discovered.

Defendant has thus sought nothing beyond what the patent itself gave it. The patent gives it the unlimited right to exclude others from utilizing its process. This process is, in fact, so superior to other processes that a monopoly over the process gives its holder a de facto monopoly over the product. But there is no danger that defendant can, by manipulating its process patent, "convert a process patent into a product patent." 426 F. Supp. at 149. S.K. has no monopoly over ATAs not produced by the Ziegler patent. Its monopoly of the product can continue only so long as its process remains "so superior to other processes that ATAs made by those other processes could not compete commercially..." J.A. 92. The district court's ruling that Ziegler could not restrict sales of products made by the patented process, without a showing that trade in some article other than ATAs made by that process was restrained, amounted to a rule that restraints on sales by a process patentee, though not by a product patentee, are per se unlawful.

Even were we otherwise attracted to this purely formalistic distinction between a process patent and a product patent, however, it is clear that such a per se approach is no longer acceptable in the wake of the Supreme Court's decision in Continental TV, Inc. v. GTE Sylvania, Inc., 433 U.S. 36, 97 S. Ct. 2549, 53 L. Ed. 2d 568 (1977). Prior to Continental TV, restrictions by a seller upon the territories in which dealers could sell were per se illegal when title to the relevant goods had passed from the seller to the dealer, but were governed by the rule of reason when the seller retained title. United States v. Arnold Schwinn & Co., 388 U.S. 365, 87 S. Ct. 1856, 18 L. Ed. 2d 1249 (1967).¹⁰  This distinction, based as it was essentially upon a "restraints on alienation" theory, was criticized by the Continental TV Court as overly formalistic and insufficiently attentive to the real economic effects of the particular challenged restraint. Continental TV is a message to lower courts that antitrust violations should be based upon economic effects rather than upon formal distinctions. As the Supreme Court said:

(W)e do not foreclose the possibility that particular applications of vertical restrictions might justify per se prohibition under Northern Pac. R. Co. But we do make clear that departure from the rule-of-reason standard must be based upon demonstrable economic effects rather than-as in Schwinn-upon formalistic line drawing.

433 U.S. at 58-59, 97 S. Ct. at 2561-62.

What the district court did below and the government urges here is subject to similar criticism. The government would have us declare that any restraint on products produced by a process patent is outside the protection of the patent laws and illegal per se regardless of the effects of the restraint, on the theory that licensing of the process exhausts the patentee's rights in that process. In the absence of "demonstrable economic effects," however, such a conclusion would be wholly unrelated to any realistic distinction between a process patent and a product patent, and would be just the sort of "formalistic line-drawing" which the Supreme Court condemned.

This conclusion is further supported by the Supreme Court's decision in Broadcast Music, Inc. v. Columbia Broadcasting System, Inc., 441 U.S. 1, 99 S. Ct. 1551, 60 L. Ed. 2d 1 (1979). There the Court again cautioned against a willingness to invoke per se rules lightly, stating that it is only after considerable experience with a restraint that courts should classify it as per se illegal. This is particularly true where, as here, the restraint is not a "naked restraint of trade with no purpose except stifling of competition." 441 U.S. at 10, 99 S. Ct. at 1557, citing White Motor Co. v. United States, 372 U.S. 253, 263, 83 S. Ct. 696, 702, 9 L. Ed. 2d 738 (1963). In addition, the Court noted in the related context of copyrights that "we would not expect that any market arrangements reasonably necessary to effectuate the rights that are granted would be deemed a per se violation of the Sherman Act." 441 U.S. at 19, 99 S. Ct. at 1562.

While it is possible that some restraints in a patent license, such as tying restrictions, may be illegal per se after Continental TV and Broadcasting Music, Inc., it would be necessary at least to show that the restraints involved had no purpose except restraining trade, and had unequivocally anticompetitive effects in the vast majority of cases. As we noted in Smith v. Pro Football, Inc., 193 U.S.App.D.C. 19, 27, 593 F.2d 1173, 1181 (1978), per se rules risk sweeping reasonable, procompetitive activity within a general condemnation, and a court will run this risk only where dictated on the basis of unambiguous experience. As appears below, we conclude that, far from being a "naked restraint of trade," the restraint challenged here is "reasonably necessary to effectuate the rights that are granted" by the Ziegler process patent. Cf. Broadcast Music, Inc., supra, 441 U.S. at 19, 99 S. Ct. at 1562.

Treatment of this sui generis arrangement as a per se violation is particularly inappropriate when six years before this suit was filed, the United States District Court for the District of Delaware had ruled that the very series of agreements and transactions at issue here did not constitute patent misuse,¹¹  because defendant was not reaching out for any new monopoly. Ethyl Corp. v. Hercules Power Co., supra at 458. In our view these authorities, along with numerous patent licensing cases discussed below, require that the result turn on a careful analysis of whether the restriction imposed here constituted an unreasonable restraint of trade. We proceed to consideration of that question.

Defendant's reasonableness claim rests primarily on its assertion that it effected no more control over ATAs than it could have exercised if it had given Hercules an absolutely exclusive license. The law is settled that S.K. could legally have licensed Hercules alone to use the patented process. See, e.g., E. Bement & Sons v. National Harrow Co., supra; United States v. Westinghouse Elec. Co., 648 F.2d 642 (9th Cir. 1981); Rail Trailer Co. v. ACF Indus. Inc., 358 F.2d 15 (7th Cir. 1966); Benger Laboratories Ltd. v. R.K. Laros Co., 209 F. Supp. 639 (E.D. Pa. 1962), aff'd, 317 F.2d 455 (3d Cir.), cert. denied, 375 U.S. 833, 84 S. Ct. 69, 11 L. Ed. 2d 64 (1963). The actual licensing arrangement here is less restrictive than an exclusive license. Because other nonexclusive licensees are free to manufacture for their own use, demand for ATAs produced by Texas, and therefore the benefit of its exclusive license, is lessened. This lesser restraint, which has fewer anti-competitive effects than a lawful exclusive license, gives the Ziegler arrangement an important badge of reasonableness.

The government attempts to respond by pointing out that antitrust scrutiny of a patent arrangement is not foreclosed simply because the patentee had the power to license less widely than it did. Of course, it is not the case that merely because the patentee has the right to refuse to license anyone, he has the right to license upon any condition he chooses; that position was rejected as early as Motion Picture Patents Co. v. Universal Film Mfg. Co., 243 U.S. 502, 37 S. Ct. 416, 61 L. Ed. 871 (1917). But in all of the cases which the government cites, and which are discussed below, the restriction imposed itself had competitive dangers which were not present in purely legal arrangements. The government has cited no case, and we are not aware of any, which found antitrust infirmities in a restriction that posed only the restraint on competition imposed by the patent itself or by exercise of legal rights corollary to those created by the patent.

Of course, where the restraint imposed has the danger of extending the patent monopoly, courts have not hesitated to prohibit it. For example, tying arrangements, in which the patentee agrees to license the patent only in exchange for the licensee's agreement to purchase other goods from the patentee, are illegal per se because they tend to foreclose competition in markets other than those in the patented product or process. See, e.g., International Salt Co. v. United States, 332 U.S. 392, 68 S. Ct. 12, 92 L. Ed. 20 (1947); Morton Salt v. G.S. Suppiger Co., 314 U.S. 488, 62 S. Ct. 402, 86 L. Ed. 363 (1942); Motion Picture Patents Co., supra.

Similarly, courts have sometimes struck down attempts by a patentee to restrict the prices at which products may be resold, because of the danger that such a restriction can be used as a means of organizing a cartel. Cases involving price-fixing restrictions illustrate well the controlling considerations. At one end of the spectrum, United States v. General Electric Co., 272 U.S. 476, 47 S. Ct. 192, 71 L. Ed. 362 (1926), approved a restriction on the prices at which the licensee could sell the patented products, stating that "a patentee may grant a license... upon any condition the performance of which is reasonably within the reward which the patentee by the grant of the patent is entitled to secure." Id. at 489, 47 S. Ct. at 196. The Supreme Court concluded that the patentee's reward included the right to fix prices at which products were sold and that a price restriction in a patent license was thus not illegal. While this toleration of price-fixing by the patentee has been seriously questioned,¹²  and has survived twice only by the grace of an equally divided Court,¹³  over the years the General Electric formulation has been the verbal frame of reference for testing the validity of a license restriction in many subsequent decisions. See e.g., United States v. Westinghouse Elec. Co., 648 F.2d 642 (9th Cir. 1981); Hensley Equipment Co. v. Esco Corp., 383 F.2d 252 (5th Cir. 1967); Turner Glass Corp. v. Hartford-Empire Co., 173 F.2d 49 (7th Cir.), cert. denied, 338 U.S. 830, 70 S. Ct. 57, 94 L. Ed. 505 (1949).

At the other end of the spectrum, United States v. Univis Lens Co., 316 U.S. 241, 62 S. Ct. 1088, 86 L. Ed. 1048 (1942) held invalid an attempt by a patentee of lenses to control the price of finished lenses produced by others from the lens blanks after it had sold the lens blanks it produced. The Supreme Court ruled that sale of a patented article exhausts the patentee's monopoly of that article and that the patentee could not thereafter control the use or disposition of the article to any greater extent than could a seller of an unpatented article. The Court relied on a number of earlier cases which predated the Sherman Act and which held that the patent gave its owner no power over the product once it was sold. See, e.g., Bauer & Cie v. O'Donnell, 229 U.S. 1, 33 S. Ct. 616, 57 L. Ed. 1041 (1913); Adams v. Burke, 84 U.S. (17 Wall.) 453, 21 L. Ed. 700 (1873).

This early Supreme Court case law tended to rely on per se rules and contained little analysis of possible anticompetitive effects. In Ethyl Gasoline Corp. v. United States, 309 U.S. 436, 60 S. Ct. 618, 84 L. Ed. 852 (1940), however, there began to emerge a tendency to focus more on the economic effects of the challenged license restriction. The patentee in Ethyl owned a patent on a gasoline additive which improved performance and which was used by almost the entire gasoline industry. The patentee required all jobbers who sold gasoline with this additive to adhere to its prices and refused to license jobbers with a history of selling for less. This practice gave the patentee the power, not authorized by the patent grant, to fix prices in the retail gasoline industry. Ethyl was using its patent to obtain a reward not contemplated by the patent laws: the cartelization of the gasoline market. The Supreme Court proscribed this practice.

United States v. Masonite Corp., 316 U.S. 265, 62 S. Ct. 1070, 86 L. Ed. 1461 (1942) further clarified the distinction between legal and illegal restraints on prices. Masonite had entered into del credere agency agreements with its competitors under which it consigned to them its patented hardboard, for sale at agreed prices. Title to the patented hardboard remained with Masonite under the agency agreement; and the case would have seemed to be governed by General Electric, supra, which allowed price-fixing where title to the patented product had not yet passed rather than by Univis, which did not allow price fixing with respect to patented goods once they were sold. Nonetheless, the Supreme Court, refusing to be bound by the form of the transaction, found the arrangement illegal. It noted:

(I)f the del credere agency device were given broad approval, whole industries could be knit together so as to regulate prices and suppress competition. That would allow the patent holder under the guise of his patent monopoly not merely to secure a reward for his invention but to secure protection from competition which the patent law unaided by restrictive agreements does not afford.

316 U.S. at 278-79, 62 S. Ct. at 1078-79 (emphasis supplied). The Court found that the arrangement gave Masonite far more than just a reward for its invention:

The power of Masonite to fix the price of the product which it manufactures and which the entire group sells ... is a powerful inducement to abandon competition... Active and vigorous competition then tend to be impaired not from any preference of the public for the patented product, but from the preference of the competitors for a mutual arrangement for price-fixing.

Id. at 281, 62 S. Ct. at 1079.¹⁴ 

The so-called "field-of-use" cases, which are perhaps most closely related to the case at bar, illustrate well the distinction between lawful use of the patent monopoly and uses which operate to extend the patent monopoly. General Talking Pictures v. Western Elec. Co., 304 U.S. 175, 58 S. Ct. 849, 82 L. Ed. 1273 aff'd on rehearing, 305 U.S. 124, 59 S. Ct. 116, 83 L. Ed. 81 (1938), involved a typical "field-of-use" restriction. The patentee there licensed patents relating to amplifiers which could be used both in motion picture exhibition equipment and in private radio reception equipment. The defendant was licensed only to practice the patent in equipment for noncommercial use, while others were licensed to use the patent in equipment used commercially. The Supreme Court upheld this restriction, concluding that it was well within the scope of the patentee's monopoly and therefore legal under the General Electric test. Courts have generally followed General Talking Pictures in holding legal such field-of-use restraints as a restriction on classes of customers to which licensees could sell and a restriction on the kinds of objects on which the process could be used.¹⁵  But courts have occasionally distinguished General Talking Pictures and held the restraint illegal where they perceived that the field-of-use restriction was being used to extend the patent into areas not protected by the patent monopoly, such as a requirement that a patented strain gauge only be sold with the licensee's machines.¹⁶ 

Viewed in perspective, these authorities, underscored by Continental TV, supra, expose the vice of the government's position and the decision below. The government urges and the district court essentially adopted a formalistic approach. Under the government's theory, a process patentee has only the right to restrict the use of his process. By analogy to Univis and similar cases, it suggests that the process patent's protection is exhausted once the process is used, and the patentee may not restrict disposition of the product once the process is completed. The government accepts the proposition that if Ziegler held a product patent, the license arrangement permitting nonexclusive licensees to use but not to sell ATAs would be legal. Thus, according to the government, the arrangement should be held illegal here solely because the defendant holds only a process patent instead of a product patent.

No functional difference between a process patent and a product patent supports the purely formal distinction the government urges. The government and the district court assume correctly that a process patent gives no power over products not manufactured by a patented process, such as unpatented salt, see International Salt, supra, or in this case, ATAs not manufactured by the Ziegler process. But the difference between a product patent and a process patent does not justify either the government's contention or the decision below that the process patentee has no more power over the product of his process than he does over the products not produced by that process. Ziegler did not invent ATAs, and thus he was not awarded any right to prevent others from making, using, or selling ATAs manufactured by a process other than his own. Any discoverer of an equally (or more) efficient process was and is free to compete in the market for ATAs despite S.K.'s patent. Ziegler's nonexclusive licensees are completely free to manufacture, use and sell any ATAs manufactured by a process not patented by Ziegler. Thus it is true, as the government asserts, that the sale of a product does not itself infringe a process patent, since the product may have been made by another process. This does not impair the patentee's right to impose restrictions in a license that affect products that were made by its process.

The government attempts to support its position by citing a number of decisions holding that the owner of a patent on a machine may not control the price of products made or handled by that machine. See, e.g., Cummer-Graham Co. v. Straight Side Basket Corp., 142 F.2d 646 (5th Cir.), cert. denied, 323 U.S. 726, 65 S. Ct. 60, 89 L. Ed. 583 (1944); American Equipment v. Tuthill Bldg. Material Co., 69 F.2d 406 (7th Cir. 1934). However, there is an obvious difference between a patentee of a brick-handling machine seeking to control the price of ordinary bricks, see American Equipment, supra, or a patentee of an attachment to a basket-making device seeking to fix the price of baskets, see Cummer-Graham, supra, and the restriction here on the unique product, ATAs. If Ziegler were attempting to restrain a generic product primarily in the public domain, the government's position here might have some merit. But, in fact, the district court found the Ziegler process to be the only presently feasible method of manufacturing ATAs.

Only one case cited by the government draws the distinction between a process patent and a product patent which the government urges here. In Barber-Coleman Co. v. National Tool Co., 136 F.2d 339 (6th Cir. 1943), the court held that a process patentee did not have the right to fix prices on sales of the product of the process, even though under General Electric a product patentee would have that right. The decision in Barber-Coleman, is, however, probably best regarded as a transparent attempt by a court to avoid the result it would have reached by literally following General Electric, which as previously noted¹⁷  has been widely criticized and frequently distinguished on dubious grounds. See L. Sullivan, supra, § 185c (1977). Indeed, the General Electric case itself involved process patents as well as product patents, and the Court there did not draw any distinction between the two. As a number of commentators have noted, the difference between a process patent and a product patent makes no difference to the potential for competitive harm inherent in a given restraint. See L. Sullivan, supra ; Adelman & Juenger, supra, at 273, 286, 288, 303. In addition, in Barber-Coleman, the patent was on a process for manufacturing hobs, which, like the bricks in American Equipment, and unlike ATAs, were also made by a number of other processes. Further, there is no issue presented here as to whether Ziegler could restrict the prices at which licensees sold ATAs manufactured with its process. Finally, courts have applied General Electric and similar cases numerous times without distinguishing process patents from product patents. See, e.g., Extractol Process v. Hiram Walker & Sons, 153 F.2d 264, 266 (7th Cir. 1946) (dictum); Barr Rubber Prods. Co. v. Sun Rubber Co., 277 F. Supp. 484, 506 (S.D.N.Y. 1967), modified on other issues, 425 F.2d 1114 (2d Cir.), cert. denied 400 U.S. 878, 91 S. Ct. 118, 27 L. Ed. 2d 115 (1970); Deering-Milliken & Co. v. Temp-Resisto Corp., 160 F. Supp. 463 (S.D.N.Y. 1958), rev'd on other grounds, 274 F.2d 626 (2d Cir. 1960); Ethyl Corp. v. Hercules Powder Co., 232 F. Supp. 453 (D. Del. 1964).

The situation here is thus quite different from those obtaining where the patentee threatened to extend its monopoly beyond those rights accruing under the patent. None of the anticompetitive effects found by the district court¹⁸  involve this danger. Moreover, these same anticompetitive effects would be created if Ziegler had given Hercules an exclusive license. Thus, none of the effects found will support a finding that the license restraints at issue violated Section 1 of the Sherman Act under the rule of reason. Since we have rejected both a per se rule and the district court's rule-of-reason approach which, in effect, applied a per se rule, the district court's conclusion that defendant violated Section 1 must be reversed. The conclusion that the restraints violated Section 2 because they constitute an attempt to monopolize must similarly be reversed: defendant cannot be guilty of attempting to monopolize the market for ATAs manufactured by the Ziegler process. That is the very market which the patent authorizes its grantee to monopolize. See, e.g., Handgards, Inc. v. Ethicon, Inc., 601 F.2d 986 (9th Cir. 1979), cert. denied, 444 U.S. 1025, 100 S. Ct. 688, 62 L. Ed. 2d 659 (1980); Commissioner of Patents v. Deutsche-Gold-Und-Silber-Scheideanstalt Vormals Roessler, 130 U.S.App.D.C. 95, 397 F.2d 656, 663 (1968).

The issue remains whether, having reversed the decision of the district court, we must remand for a new trial or whether we may judge the reasonableness of the restriction on sales by licensees on the basis of the findings and the record before us. We conclude, based on an examination of the restraint at issue and the findings and the record under the rule of reason, that a remand is unnecessary and inappropriate.

None of the anticompetitive effects of the challenged restriction found by the district court exceed the anticompetitive effects which the patent authorized. There is no basis in the record for a finding of other anticompetitive effects, and with one exception, which is discussed below, the government is unable to point to any other possible effect on competition. Finally, analysis of the arrangement itself demonstrates that it is no more restrictive than a legal exclusive license, and in fact has certain procompetitive effects not created by such a license.

As stated earlier, Dr. Ziegler lawfully could have licensed Hercules alone to practice the patented process. Such an exclusive license would, by definition, have given Hercules a monopoly over ATAs manufactured by his process, and would have effectively granted Hercules a monopoly over the sale of ATAs, since no other process is now commercially competitive with the Ziegler process. Exclusive licenses are tolerated because they normally threaten competition to no greater extent than is threatened by the patent itself. See L. Sullivan, supra, at 534. Equally important from the perspective of the rule of reason, many potential licensees might be unwilling to undertake the expense necessary to develop and promote a product but for assurances against attempts by later licensees to exploit the early licensee's development and promotion. See Adelman & Juenger, supra, at 298-99; P. Areeda, Antitrust Analysis P 411, at 585-86 (3d ed. 1981). An exclusive license protects licensees against such "free rider" problems, and thereby serves the interests of both the patentee and the public by facilitating more rapid and widespread use of new inventions.

The license restriction at issue here has similar virtues with somewhat fewer vices. It prevents other licensees from taking advantage of any promotional or developmental efforts by Hercules, but leaves them free to utilize the patented technology in their own operations. The prohibition on sales is thus a more narrowly tailored version of the exclusive license, having most of its benefits but fewer of its liabilities. The same considerations that lead courts to validate exclusive licenses lead us to approve the restriction at issue here.

This conclusion might not follow if the restriction here caused or threatened anticompetitive effects not present in an exclusive license. Price-fixing licenses create the possibility of using the patent to mask a cartel, as was the case in Ethyl and Masonite, supra. Other impermissible market divisions by means of patent arrangements might give each competitor an economic incentive to continue the arrangement which insures its loyalty to the arrangement. For example, in Ethyl Gasoline Corp., each gasoline refiner benefited from the price-fixing arrangement which was incidental to the patent license. In Masonite, the Court found that the defendant's product would be preferred for the price-fixing capacity of the patent rather than because of the competitive merits of the product. Similarly, Barber-Coleman, American Equipment, and Cummer-Graham, supra, all involved price-fixing and the consequent danger that competitors would be attracted to the product because the patent facilitated price-fixing rather than because of the product's competitive merits. But nothing in this record or the district court findings reveals or portends such effects.

A number of scholars have also noted that some patent arrangements, such as field-of-use restrictions, have a potential for facilitating market division by giving each participant a stake in the patent in the form of an exclusive territory, or by parceling out to each competitor exclusive access to particular customers. Thus, such agreements give potential competitors incentives to remain in cartels rather than turning to another product, inventing around the patent, or challenging its validity. See, e.g., L. Sullivan, supra, § 184; Adelman & Juenger, supra, at 305-08; Buxbaum, supra. In this case by contrast, only Hercules¹⁹  enjoys any advantage from the limitation on sales, and the advantage it enjoys is far less than the advantage which a conventional exclusive license would give it. All other competitors, bound as they are by the prohibition on sales, have every incentive to compete or challenge the defendant's patent and thereby become entitled to sell ATAs.

In addition, market division is a much more serious problem where the product restricted is also produced by other means. Each of the cases cited involved a product which was generic: other processes existed and competed with the patented process. If one of twenty competitors makes a slight improvement in existing technology and proceeds to divide the country into exclusive territories for the purpose of exploiting the patent, the danger that his competitors might abandon the existing technology in favor of the patented technology in order to divide markets is manifest. If the patentee invents a clearly superior technology, whether a product or process, and divides the country into exclusive territories, the case is completely different; there would be a de facto monopoly irrespective of whether the territories were divided. In such a case, as Professor Sullivan argues, "(v)irtue is not attacked, let alone undone .... competition is restricted to no greater degree by the assignment than is inherent in the patent grant itself." L. Sullivan, supra, at 534. The commercial superiority of the Ziegler process is clear. The strength of the Ziegler patent and the inherent value of the process itself give the patentee a monopoly of all ATAs. The licensing arrangement adds nothing.²⁰ 

A patent license might also restrict competition by undermining incentives to attack the validity of the patent or to invent around it.²¹  The government briefly contends, without support from the record, that the restriction at issue here may have undermined Hercules' incentives to challenge the Ziegler patent. However, even if this were true, it would be functionally indistinguishable from the effect of a conventional exclusive license. See Adelman & Juenger, supra, at 299. And, as we have demonstrated, all other licensees have ample incentives to attack the validity of the patent. Indeed, the chief advocates of the theory that restraints that undermine incentives to challenge the patent should be illegal, Professors Adelman and Juenger, have directly and persuasively discounted the possibility of such a disincentive in this very case:

This arrangement did not lessen incentives to attack the patent, since all other licensees presumably smarted under the prohibition against selling ... (T)he Government ... seemed to assert quite simply that since the unpatented products were "outside" the scope of the patent monopoly, the attempt to impose such restraints on sales must be illegal. As noted, such an approach sidesteps practical analysis of the issues.

Adelman & Juenger, supra, at 303 (footnotes omitted). See also P. Areeda, supra, at 586; American Bar Ass'n., U.S. Antitrust Law in International Patent and Know-How Licensing, 25-28, 45-47 (1981).

Indeed, it may well be that striking down the restraint at issue here would itself injure competition. There is a reasonable likelihood that had the defendant been unable to issue licenses for internal use only, it would have issued only an exclusive license to Hercules. The record tends to show that Ziegler took his contract with Hercules seriously, and considered himself bound to refuse to grant any other licenses for sale. See J.A. at 106-09. We should hesitate before we impose a legal rule that would force a patentee to follow the more anticompetitive route of a single exclusive license.

Our conclusion is strengthened by the fact that the potential competition even arguably restrained is competition from manufacturers who are able to make and use the product solely because Ziegler invented his process and licensed them to use it. Other courts have emphasized this fact in reaching the same conclusion we reach here. A & E Plastic Pak Co. v. Monsanto Co., 396 F.2d 710 (9th Cir. 1968) upheld an agreement in which the defendant licensed its unpatented know-how to plaintiff in exchange for a promise by plaintiff not to compete with it using the licensed know-how. Here, as in A&E, the restraint "does not appear to be an agreement between competitors not to compete, for absent the licensed know-how, (the licensee) is in no position to compete." Id. at 714-15.

Our conclusion is further corroborated by United States v. Westinghouse Elec. Corp., 471 F. Supp. 532 (N.D. Cal. 1978), aff'd 648 F.2d 642 (9th Cir. 1981) where both courts upheld against antitrust challenges an agreement in which defendant licensed its patents to Mitsubishi only for use outside the United States. The courts found no evidence that the agreement operated to extend the scope of the patents. They firmly rejected the government's theory that it was wrongful for Westinghouse to license Mitsubishi to practice Westinghouse's patents while not allowing it to compete with Westinghouse. Judge Weigel in the district court noted:

What the government is really proposing is this: Since all monopolies, including patent monopolies, undesirably limit competition, every patent licensing contract should, if at all possible, be viewed as a "combination in restraint of trade." Taken to its logical limits, this argument would find almost every patent licensing agreement to be illegal. For one example, an exclusive license to manufacture would be invalid as excluding competition from the unlicensed manufacturers... This is a demand calculated to alter and substantially reduce the scope of the patent monopoly. It should be addressed to Congress, not to the courts.

471 F. Supp. at 542.²² 

The district court rested its decision that Dr. Ziegler's license arrangement was illegal on a formalistic distinction between a process patent and a product patent, without undertaking a functional comparison between the restraints on a license permissible to a process patentee and the restraints effected here. The two-step analysis conducted by the district court led it to what is tantamount to a decision that any process patentee who licenses to use, but not to sell, the product manufactured by the process has committed a per se violation of the antitrust laws. Under the district court's formulation, various considerations critical to a rule of reason, such as the fact that the patentee could have lawfully licensed no one, or licensed only Hercules to use the patented process, would be entirely irrelevant. Nor did the district court's analysis permit it to give adequate weight to the fact that Ziegler's patented process commanded the market because it was such an efficient process, and not because of some side arrangement such as price fixing which might attract ATA users interested in organizing a cartel. Furthermore, the district court's analysis did not permit adequate recognition of the critical fact that Ziegler's license affected only ATAs manufactured by his process, and did not purport to affect the use, sale, price, or any other aspect of commerce in ATAs manufactured by some other process. The profits from the Ziegler process apparently enjoyed by Hercules and S.K., as Ziegler's successor, were and are at risk from competition from anyone who invents around the Ziegler patent and develops a process to produce ATAs more economically. Those profits are strong incentives for potential competitors to do so. Nothing in the arrangements at issue here would deter such competition.

Nor did the district court confront the Supreme Court's teaching that condemnation of a practice as a per se violation is reserved for egregious violations such as price-fixing agreements and tying arrangements whose pattern is monotonously familiar and whose anticompetitive effects are obvious and long outlawed. There was no clear authority warning Ziegler, Hercules, and their advisors that the combination of an exclusive sales license and nonexclusive-use licenses would be invalid. Indeed, a transaction involving only one exclusive license for sales and use would have been beyond question. The distinction drawn between the power of a process patentee over the product of the process and the power of a product patentee over all such products (regardless of how manufactured) has been recognized only tangentially in an isolated and questionable decision. Finally, the specific arrangement at issue here had been carefully examined and found valid by the district court in Delaware only six years before the government brought this case.

For these reasons, we are satisfied that under the principles which we have outlined, the district court would have no occasion on remand to receive further evidence or to make further findings before concluding that the Ziegler license arrangements were not unreasonable restraints of trade. Accordingly, we reverse the decision of the district court and remand with directions to enter judgment for defendant.

It is SO ORDERED.