BROWNING, Chief Judge:
The issue presented for review is whether Data General’s refusal to license its NOVA operating system software except to purchasers of its NOVA central processing units (CPUs) is an unlawful tying arrangement under section 1 of the Sherman Act, 15 U.S.C. § 1 (1976) and section 3 of the Clayton Act, 15 U.S.C. § 14 (1976). We conclude that it is.
I.
Defendant Data General manufactures a computer system known as NOVA. The system consists of a NOVA CPU designed to perform a particular “instruction set” or group of tasks, and a copyrighted NOVA operating system called RDOS containing the basic commands for operation of the system. Not all operating systems work with all CPUs. Plaintiffs produce emulator NOVA CPUs designed to perform the NOVA instruction set and thus to make use of defendant’s RDOS.
Data General refuses to license its RDOS to anyone who does not also purchase its NOVA CPU. Plaintiffs allege that this constitutes an unlawful tying arrangement; the defendant’s RDOS being the tying product, the NOVA instruction set CPU being the tied product.
Plaintiffs filed a number of actions alleging violations of section 1 of the Sherman Act and section 3 of the Clayton Act. The actions were consolidated. The issues of liability and damages were segregated for trial. This appeal is from a judgment on liability.
After extensive discovery, the parties filed cross-motions for summary judgment. The district court denied the motions, but found certain facts to be uncontroverted under Fed.R.Civ.P. 56(d). Trial, limited to the issue of defendant’s economic power, resulted in a jury verdict for plaintiffs. Defendant’s motion for judgment n.o.v. or for a new trial was granted. Plaintiffs appealed.
II.
A tying arrangement is illegal if it is shown to restrain competition unreasonably or is illegal per se, without such a showing, if certain prerequisites are met. Fortner Enterprises v. U.S. Steel Corp., 394 U.S. 495, 498-500, 89 S.Ct. 1252, 1256-1257, 22 L.Ed.2d 495 (1969) (Fortner I). The prerequisites of per se illegality are: (1) separate products, the purchase of one (tying product) being conditioned on purchase of the other (tied product); (2) sufficient economic power with respect to the tying product to restrain competition appreciably in the tied product; and (3) an effect upon a substantial amount of commerce in the tied product. Fortner I, 394 U.S. at 499, 89 S.Ct. at 1256, Moore v. Jas. H. Matthews & Co., 550 F.2d 1207, 1212 (9th Cir.1977). These prerequisites were satisfied in this case. We therefore do not consider whether competition was in fact unreasonably restrained.
The district court properly granted summary judgment on the first and third of the required elements of a per se violation, holding that on the undisputed facts [1339]*1339the NOVA instruction set CPU and defendant’s RDOS are separate products and the volume of commerce in NOVA instruction set CPUs tied to the purchase of defendant’s RDOS is substantial. In re Data General Corp. Antitrust Litigation, 490 F.Supp. 1089, 1104-07, 1116-17 (N.D.Cal.1980).
We adopt the district court’s reasoning on these issues, adding that the court’s analysis of defendant’s “single product” claim is supported by the Supreme Court’s recent discussion in Jefferson Parish Hospital District No. 2 v. Hyde, — U.S. -, 104 S.Ct. 1551, 1561-65, 80 L.Ed.2d 2 (1984). The undisputed facts summarized in the district court’s opinion establish that a demand existed for NOVA instruction set CPUs separate from defendant’s RDOS, and that each element of the NOVA computer system could have been provided separately and selected separately by customers if defendant had not compelled purchasers to take both. See also Klamath-Lake Pharmaceutical Association v. Klamath Medical Service Bureau, 701 F.2d 1276, 1289 (9th Cir.1983).1
The remaining element necessary to establish a per se violation — defendant’s possession of sufficient economic power with respect to the tying product, defendant’s RDOS — was tried to a jury and resolved in plaintiffs’ favor. The district court erred in setting aside this verdict or, alternatively, ordering a new trial.
III.
One of the purposes of a per se rule is to avoid an “incredibly complicated and prolonged economic investigation ... to determine at large whether a particular restraint has been unreasonable.” Northern Pacific Railway Co. v. United States, 356 U.S. 1, 5, 78 S.Ct. 514, 518, 2 L.Ed.2d 545 (1958). See also Jefferson Parish Hospital, 104 S.Ct. at 1560 n. 25. Although not requiring as extensive an inquiry as would be necessary to determine whether the tie-in violated the general standard of reasonableness, the district court held that plaintiffs “could not recover on the alleged tie-ins unless they identified and proved the relevant market for the tying and tied products.” In re Data General Corp. Antitrust Litigation, 529 F.Supp. 801, 809 (N.D.Cal.1981). The trial that followed “focused upon the definition of the relevant markets” for the two products, which the Court characterized as the “critical issue,” (id. at 806) and consumed forty-five days. Id. at 804.
The district court recognized that detailed market analysis was not required in a per se tying case prior to United States Steel Corp. v. Fortner Enterprises, Inc., 429 U.S. 610, 97 S.Ct. 861, 51 L.Ed.2d 80 (1977) (Fortner II), but read that opinion as rejecting this approach in favor of a requirement of “some degree of market analysis even in a per se case.” 529 F.Supp. at 808. The court relied particularly upon language in Fortner II, which states the question to be:
whether the seller has the power, within the market for the tying product, to raise prices or to require purchasers to accept burdensome terms that could not be exacted in a completely competitive market. In short, the question is whether the seller has some advantage not shared by his competitors in the market for the tying product.
429 U.S. at 620, 97 S.Ct. at 867.
From the district court’s analysis of the asserted deficiencies in plaintiffs’ proof, it appears the court read this statement as requiring proof of power to fix the price of the tying product in the whole of the relevant market as defined by the inquiry described in United States v. E.I. du Pont de Nemours & Co., 351 U.S. 377, 76 S.Ct. 994, 100 L.Ed. 1264 (1956), a monopolization case. In this the district court erred. Possession by the seller of such monopoly pow[1340]*1340er is sufficient to establish per se illegality, but it is not required.
As the Supreme Court said in United States v. Loew’s, Inc., 371 U.S. 38, 45, 83 S.Ct. 97, 102, 9 L.Ed.2d 11 (1962):
Market dominance — some power to control price and to exclude competition — is by no means the only test of whether the seller has the requisite economic power. Even absent a showing of market dominance, the crucial economic power may be inferred from the tying product’s desirability to consumers or from uniqueness in its attributes.4
This position was re-affirmed in the Fortner cases. In Fortner I:
The standard of “sufficient economic power” does not, as the District Court held, require that the defendant have a monopoly or even a dominant position throughout the market for the tying product. Our tie-in cases have made unmistakably clear that the economic power over the tying product can be sufficient even though the power falls far short of dominance and even though the power exists only with respect to some of the buyers in the market. . . .
... [T]he presence of any appreciable restraint on competition provides a sufficient reason for invalidating the tie. Such appreciable restraint results whenever the seller can exert some power over some of the buyers in the market, even if his power is not complete over them and over all other buyers in the market. . . . [D]espite the freedom of some or many buyers from the seller’s power, other buyers — whether few or many, whether scattered throughout the market or part of some group within the market — can be forced to accept the higher price because of their stronger preferences for the product, and the seller could therefore choose instead to force them to accept a tying arrangement that would prevent free competition for their patronage in the market for the tied product. Accordingly, the proper focus of concern is whether the seller has the power to raise prices, or impose other burdensome terms such as a tie-in, with respect to any appreciable number of buyers within the market.
394 U.S. at 502-04, 89 S.Ct. at 1258-59 (emphasis added).
In Fortner II the Court reiterated that its prior decisions “do not require that the defendant have a monopoly or even a dominant position throughout the market for a tying product,” 429 U.S. at 620, 97 S.Ct. at 867, and approved a commentator’s summary of the holding in Fortner I: “Whenever there are some buyers who find a seller’s product uniquely attractive, and are therefore willing to pay a premium above the price of its nearest substitute, the seller has the opportunity to impose a tie to some other good.” 429 U.S. at 620 n. 14, 97 S.Ct. at 868, n. 14. (quoting Note, The Logic of Foreclosure: Tie-In Doctrine after Fortner v. U.S. Steel, 79 Yale L.J. 86, 93-94 (1969) (emphasis added). The language from Fortner II relied upon by the district court is not inconsistent with this interpretation; it required only “power, within the market for the tying product, to raise prices or to require purchasers to accept burdensome terms that could not be exacted in a completely competitive market.” 429 U.S. at 620, 97 S.Ct. at 867 (emphasis added).
In its most recent decision on the question (filed after the ruling of the district court in this case) the Supreme Court again made it clear that a tying arrangement is illegal per se if the seller of the tying product has the capacity to force [1341]*1341some buyers to purchase a tied product they do not want or would have preferred to purchase elsewhere. When such forcing occurs “competition on the merits in the market for the tied item is restrained.” Jefferson Parish Hospital, — U.S. -, 104 S.Ct. at 1558 (1984). Thus, what is required in a per se case is not power over the whole market for the tying product, but only, as the Court said, a “type of market power [that] has sometimes been referred to as ‘leverage ... defined here as a supplier’s ability to induce his customers for one product to buy a second product from him that would not be purchased solely on the merit of that second product.’ ” Id. at 1559 n. 20 quoting V.P. Areeda & D. Turner, Antitrust Law, ¶ 1134a at 202 (1980). “[W]e have condemned tying arrangements,” the Court said, “when the seller has some special ability — usually called ‘market power’ — to force a purchaser to do something that he would not do in a competitive market.” Id. at 1559.
Nor is a restraint on competition that is substantial in terms of the entire market for the tied product required. “If only a single purchaser were ‘forced’ with respect to the purchase of a tied item, the resultant impact on competition would not be sufficient to warrant the concern of antitrust law.” Id. at 1560. Beyond that, however, it need only appear that “a substantial volume of commerce is foreclosed,” id., which the court earlier defined as “substantial enough in terms of dollar-volume so as not to be merely de minimis.” Fortner I at 501, 89 S.Ct. at 1258. See also Moore v. Jas. H. Matthews & Co., 550 F.2d at 1216.
In accordance with these holdings, we review the record not for what it may reveal as to defendant’s position in a defined market in which defendant’s RDOS was sold, but only to determine whether the jury reasonably could have concluded defendant’s RDOS was sufficiently unique and desirable to an appreciable number of buyers to enable defendant to force those buyers also to buy a substantial volume of defendant’s NOVA instruction set CPUs they would have preferred not to buy.
IV.
There was abundant evidence that defendant’s RDOS was distinctive and particularly desirable to a substantial number of buyers, and could not be readily produced by other sellers. There was also substantial evidence that defendant’s insistence upon licensing its RDOS only to purchasers of defendant’s NOVA instruction set CPU, led buyers to purchase defendant’s NOVA CPUs who would not have bought them or would have bought them elsewhere absent the tying requirement.
Although expressing some doubt as to the sufficiency of the evidence, the district court assumed defendant’s RDOS was superior to competing operating systems and was viewed as uniquely desirable by buyers. 529 F.Supp. at 816. We do not share the court’s hesitancy about the adequacy of the proof of the strong preference of many customers for RDOS. It was a most popular product. Experts, customers, and even competitors testified to its many advantages over competitive products.2 Defendant’s own officials expressed the same opinion in pre-litigation documents.
Defendant's RDOS has copyright protection. Defendant also claimed the production of RDOS required use of defendant’s trade secrets. The RDOS copyright established both the distinctiveness of RDOS and a legal bar to its reproduction by competitors. “The requisite economic power is presumed when the tying product is patented or copyrighted.” United States v. Loew’s, Inc., 371 U.S. at 45, 83 S.Ct. at 102. The copyright confers upon defendant “some advantages not shared by his competitors in the market for the tying product.” Fortner II, 429 U.S. at 620, 97 [1342]*1342S.Ct. at 868. “[T]he copyright monopolies in United States v. Paramount Pictures, Inc., 334 U.S. 131 [68 S.Ct. 915, 92 L.Ed. 1260] and United States v. Loew’s Inc., 371 U.S. 38 [83 S.Ct. 97, 9 L.Ed.2d 11] ... represented tying products that the Court regarded as sufficiently unique to give rise to a presumption of economic power.” 429 U.S. at 619, 97 S.Ct. at 867. “[P]er se prohibition is appropriate if anticompetitive forcing is likely. For example, if the government has granted the seller a patent or similar monopoly over a product, it is fair to presume that the inability to buy the product elsewhere gives the seller market power.” Jefferson Parish Hospital, 104 S.Ct. at 1560. See also Moore v. Jas. H. Matthews & Co., 550 F.2d at 1215-16.
There is abundant evidence, including testimony of defendant’s own executives, customers, and plaintiffs’ expert witnesses, that defendant’s RDOS could not be reproduced without infringing defendant’s copyright and utilizing defendant’s trade secrets.3 Defendant vigorously pursued those who assertedly violated defendant’s proprietary rights. Additionally, there was evidence that creating and testing a compatible system would require millions of dollars and years of effort. One of defendant’s officers testified that the passage of the time required to reproduce RDOS would render the completed software obsolete.
The power to coerce that RDOS gave the defendant was enhanced by the fact that many of defendant’s customers were “locked in” to the use of RDOS. Briefly, defendant sells RDOS and NOVA CPUs primarily to original equipment manufacturers (OEMs) who combine them with application software (a set of instructions that allows the system to accomplish a particular task) to create a complete computer system for resale. Application system software for particular uses is developed by OEMs at substantial expense. Once developed, application software for a particular use may be used by an OEM in producing any number of computer systems for that use for resale to different customers. However, application software is designed to function only with a particular operating system. OEMs who construct their application software to function with defendant’s RDOS therefore must purchase an RDOS for each computer system they assemble using that application software. Because of the tying condition, they also must purchase one of defendant’s NOVA instruction set CPUs for each such computer system they sell.
An OEM can free itself from this “lock in” only by abandoning its application software compatible with defendant’s RDOS, in which it has a substantial investment, or converting the software so that it may be used with another operating system. There was abundant testimony that conversion was not economically feasible.4
The defendant argues that “lock-in” is irrelevant in determining its market power because OEMs are aware of the tie when they select an operating system for the computer system they are assembling. At that point, defendant argues, the OEM has made no investment in application software and, as a result, chooses freely among competing systems. 529 F.Supp. at 821. This characterization of the market is not accurate. As the evidence in this case establishes, the initial choice is not free of forcing. Defendant’s operating system has [1343]*1343been shown to be unique as a matter of law and distinctively attractive as a matter of fact. Defendant’s initial leverage is magnified by the lock-in. By 1979, 93 percent of defendant’s NOVA CPU sales were made to locked-in customers. These buyers were not only forced to buy defendant’s CPUs initially to acquire the operating system they found most attractive, they were thereafter forced to buy defendant’s CPUs for their subsequent needs in order to acquire the only operating system they could economically use. Not even a decision by CPU manufacturers to broaden their base and compete in the operating system market would have alleviated the problem, for the locked-in customers were not free to choose among competing operating systems. RDOS was the only operating system that would allow them to realize the benefit of their investment in application software, an investment that in some cases totalled millions of dollars.
OEM testimony confirmed defendant’s potential power to coerce arising from the lock-in. For example, one OEM witness testified “[without [the RDOS operating system] I can’t operate”; and another: “economically I was in a position where I had to use RDOS. I had no choice at that point.”
The power arising from the special attraction of RDOS, coupled with the copyright protection, the trade secret barrier, and the lock-in, was evidenced by defendant’s minimum equipment configuration (MEC) program. To obtain defendant’s RDOS all licensees were required to purchase not only defendant’s CPU but also a set quantity of other peripheral hardware, or pay a program license charge. Defendant’s national accounts manager accurately referred to the charge as a “penalty.” Customers testified they were forced to buy peripherals from defendant they otherwise would not have purchased. An OEM testified he purchased defendant’s fixed disc because he “had to ... or pay a $5,000 fine.” He also testified he could have bought a superior disc drive for his purposes from another source at half the price. Another customer testified “we’ve had to take equipment that we either couldn’t use, or equipment that, for one reason or another, might, in our opinion, have been best— best obtained from another source.” Still another OEM called the MEC “arbitrary” because it required the purchase of “items of hardware specified in the Minimum Equipment Configuration for certain products that have no functional bearing or are not required, or not used necessarily by the programs themselves.” Defendant’s senior vice-president testified the complaints were received from customers about the MEC program “all the time.”
Defendant retained the MEC program despite buyer resistance because, as defendant’s president testified, if required to forgo the program defendant would have suffered a loss in revenues. The tie-in of RDOS to defendant’s NOVA instruction set CPU was an equally conscious exercise of economic power in one market to gain an advantage in others. As one of defendant’s managers wrote in an intra-company memorandum, “[protection from knock-off products still lies in software licensing restrictions.”
The district court properly rejected defendant’s argument, vigorously renewed in this court, “that it must bundle its software together with its CPUs in order to recover its substantial investment in software research and development,” (490 F.Supp. at 1121), and that “it would be unfair to permit emulator-CPU manufacturers to reap the benefits of [defendant’s] software [research and development] when they sell their competing CPUs for use with [defendant’s] software.” 490 F.Supp. at 1121-22.
Defendant’s president testified the tie was devised to ensure recovery of RDOS development costs. He testified the decision to tie was made after a competitive manufacturer of NOVA emulator CPUs requested permission to use RDOS. Rather than sell the software separately at a price that would reflect research and development, defendant chose to restrict availability to its own CPU customers, thus restricting competition for the tied product. As the district court said, “Recovery of investment costs has been explicitly excluded [1344]*1344from the narrowly-construed exceptions to the per se rule against tie-ins.” Id. at 1122. Defendant “has not shown, nor has it raised a genuine issue of fact with respect to its ability to show at trial, that it is any less capable than was Jerrold Electronics [United States v. Jerrold Electronics Corp., 187 F.Supp. 545 (E.D.Pa.1960), aff'd per curiam, 365 U.S. 567, 81 S.Ct. 755, 5 L.Ed.2d 806 (1961)] of adopting the less restrictive alternative of restructured prices in order to recoup its investment costs and maintain its incentive for further innovation.” 490 F.Supp. at 1122.
If the tie were allowed, competing manufacturers of CPUs would be forced
not only to match existing sellers of the tied product in price and quality, but to offset the attraction of the tying product itself. Even if this is possible through simultaneous entry into production of the tying product, entry into both markets is significantly more expensive than simply entry into the tied market....
Jefferson Parish Hospital, 104 S.Ct. at 1558 n. 19 (quoting Fortner I, 394 U.S. at 513, 89 S.Ct. at 1263 (White, J., dissenting)). In short, defendant must recover the cost of RDOS development by pricing RDOS appropriately, not by tying it to a separate product.
Evidence regarding the potential sources of power with respect to RDOS (copyright, trade secret, and “lock-in”), was submitted to the jury under appropriate instructions. The jury found as a fact that defendant possessed and used the power by means of the tying arrangement to appreciably restrain competition in the market for NOVA instruction set CPUs. The evidence outlined above fully supported the jury’s verdict.
V.
Most, although not all, of the trial court’s reasons for setting aside the verdict are traceable to the court’s view that the legality of a tying arrangement must be tested by the seller’s economic power throughout the market for the tying product, and by the relative substantiality of the restraint on competition in the tied product market considered as a whole.
As we have said, the trial court assumed customers regarded RDOS as “uniquely desirable and that it in fact possesses various features which render it superior to other software,” but concluded that plaintiffs had failed to prove that defendant’s “competitors were prevented from developing functionally equivalent software.” 529 F.Supp. at 816. Conceding that the copyright on RDOS and the trade secrets involved in its creation precluded development by defendant’s competitors of “compatible” software, the court held plaintiffs had failed to prove the effect of defendant’s copyright and secrets on the development of software “comparable” to RDOS. Id. at 816-17.
The court erroneously imposed the burden of proof on plaintiffs. The RDOS copyright created a presumption of economic power sufficient to render the tying arrangement illegal per se.5 The burden to rebut the presumption shifted to defendant.
More basically, the court was misled by its conception that power throughout the product market for the tying product was required. Earlier in its opinion the court indicated the relevant market must be defined to permit the jury to determine whether defendant’s competitors “were prevented from developing competitive software.” Id. at 809. The court continued:
The focus in defining a relevant product market must be upon recognizing those firms and products which present relevant alternatives to the defendant’s product, to the extent that they are “reasonably interchangeable” for the same or similar uses, and thus restrict the de[1345]*1345fendant’s power to raise prices. See United States v. E.I. du Pont de Nemours & Co., 351 U.S. 377 [76 S.Ct. 994, 100 L.Ed. 1264] (1956).
Id. Thus the court’s concern was whether there were reasonably interchangeable substitutes for RDOS in the operating systems market as a whole, a question of critical importance if the question were whether that market had been monopolized.
As the authorities cited earlier establish, the focus of the prohibition against tying arrangements is quite different. The concern is not with the restraint on competition in the tying product but on competition in the market for the tied product. What is required is not monopoly power in the tying product market, but only sufficient power to enable the seller to restrict competition in the tied product. If a seller’s product is distinctive, not available from other sources, and sufficiently attractive to some buyers to enable the seller by tying arrangements to foreclose a part of the market for a tied product, the adverse impact on competition in the tied product is not diminished by the fact that other sellers may be selling products similar to the tying product.
As the Supreme Court said in Northern Pacific Railway v. U.S., 356 U.S. at 10 n. 8, 78 S.Ct. at 520 n. 8 “the defendant in International Salt [332 U.S. 392, 68 S.Ct. 12, 92 L.Ed. 20] offered to prove that competitive salt machines were readily available which were satisfactory substitutes for its machines (a fact the Government did not controvert), but the Court regarded such proof as irrelevant.” And, again, in United States v. Loew’s, Inc., 371 U.S. at 49, 83 S.Ct. at 104, “[T]he mere presence of competing substitutes for the tying product, here taking the form of other programming material as well as other feature films, is insufficient to destroy the legal, and indeed the economic, distinctiveness of the copyrighted product.”
The law was succinctly summarized in Carpa, Inc. v. Ward Foods, Inc., 536 F.2d 39, 48 (5th Cir.1976), a trademark tying case, in a manner particularly pertinent here:
What is required is a factual assessment of the tying product’s uniqueness and desirability, not its market power in the sense of a Section 2 Sherman Act violation. United States v. Loew’s, Inc., 371 U.S. 38, 45, 83 S.Ct. 97 [102], 9 L.Ed.2d 11 (1962). Uniqueness, of course, presupposes that competitors are in some way foreclosed from offering the distinctive product. Fortner points out at 505, 89 S.Ct. 1252, note 2, that such barriers may be legal, as in the cases of patented or copyrighted products. Trademarks surely may be included in the list of such legal restraints, and, as with copyrighted material, the mere presence of competing substitutes is insufficient to destroy the legal, and more importantly the economic, distinctiveness of the trademark. See Loew’s, 371 U.S. at 49, 83 S.Ct. 97.
(emphasis added). See also Seigel v. Chicken Delight, Inc., 448 F.2d 43, 49-50 (9th Cir.1971).
Fortner II is not to the contrary, as the district court thought. There must, of course be power to coerce. Fortner II holds only that a seller lacks such power if buyers may choose between fungible products offered by different sellers. In Fortner II the tying item was favorable credit terms, the tied product U.S. Steel’s prefabricated homes. U.S. Steel’s credit was not unique; money is fungible. Anyone willing to accept less profit could have offered credit terms similar to those offered by U.S. Steel. As the Supreme Court said, “The unusual credit bargain offered to Fortner proves nothing more than a willingness to provide cheap financing in order to sell expensive houses.” Fortner II, 429 U.S. at 622, 97 S.Ct. at 868. In contrast, no one but defendant could offer RDOS. It was not fungible, but rather in many ways unique.
The question is not whether other operating systems with which RDOS competed [1346]*1346were as good as RDOS or better in the eyes of some buyers, but rather whether RDOS, available only from defendant, was sufficiently attractive to some customers to enable defendant to require those who wished to obtain it also to buy from defendant NOVA instruction set CPUs they might otherwise have purchased from others.6 As we have seen, evidence of the defendant’s possession of such power was ample.
Clearly the availability of “comparable” or “functionally equivalent” operating systems would not have freed “locked-in” OEMs of the pressure, imposed by their investment in application software “compatible” only with RDOS, that compelled them to accede to defendant’s condition that they purchase defendant’s NOVA CPU in order to obtain RDOS.
The district court held that the “lock-in” did not confer upon defendant any “legally cognizable power over price” because OEMs who purchased RDOS and defendant’s NOVA instruction set CPUs for use in assembling computer systems for resale were constrained by competition in the resale market from paying non-competitive prices for components of their systems, and that, in fact, defendant’s prices to OEMs were fully competitive. 529 F.Supp. at 814-15, 817-18. There are several answers. Some OEMs are insulated from strict price sensitivity. As defendant’s marketing manual stated: “[M]any OEM’s have an effective monopoly or near-monopoly for their product, because of patent position, market share dominance, control of distribution channels, or whatever. This guy isn’t forced to go the lowest possible unit cost.” Moreover, to the extent that end-users considered defendant’s RDOS to be superior to other operating system software, the OEM’s price to the end-user and thus defendant’s price of hardware to the OEM, could exceed that of other suppliers. Most important, the passage in Fortner II to which defendant refers (“whether the seller has the power, within the market for the tying product, to raise prices or to require purchasers to accept burdensome terms that could not be exacted in a completely competitive market” (429 U.S. at 620, 97 S.Ct. at 867) (emphasis added), recognizes that a seller’s power to impose an onerous tying arrangement is sufficient to invoke per se condemnation. Even assuming defendant’s package of RDOS and defendant’s NOVA instruction set CPU was competitively priced, the record establishes that defendant could, and did, force that package upon some buyers who, if free to choose, would have bought RDOS from defendant but NOVA instruction set CPUs from others.
“Fundamentally,” the district court held, “plaintiffs have not presented evidence demonstrating that their alleged inability to compete with [defendant’s] NOVA CPUs is attributable to the software licensing restrictions rather than this failure to meet” defendant’s standards of quality and service. Even from the brief summary presented here, it is evident there was ample direct and circumstantial evidence to support the jury’s verdict to the contrary.
VI.
The district court also concluded that plaintiffs failed to prove an appreciable restraint in the market for the tied product, the NOVA instruction set CPU. The court noted that the general market for CPUs was highly competitive, “characterized by a wide range of competitive hardware offerings, intense price competition, ease of entry, and rapid growth.” 529 F.Supp. at 818. The court noted many competitors had entered the CPU market after the introduction of defendant’s RDOS tied to defendant’s CPU. Id. Even assuming the tie-in appreciably restrained competition “in the NOVA instruction set sub-market,” the court said, “the portion of the broad market conceivably affected by this phenomenon is so small that it cannot be character[1347]*1347ized as indicative of power ‘appreciably to restrain competition’ in the general market.” Id.
But as we have already seen, a detailed analysis of competitive conditions in the tied product market is inappropriate in a per se case. Indeed as the district court held in its first opinion, see 490 F.Supp. at 1116-17, all that is required in respect to the extent of the restraint in the market for the tied product is that a “substantial volume of commerce be foreclosed,” Jefferson Parish Hospital District No. 2 v. Hyde, 104 S.Ct. at 1560 (emphasis added); and “substantial volume” in this context means only an amount greater than de minimis, a requirement clearly satisfied here. See 490 F.Supp. at 1117.
VII.
Defendant argues that it is entitled to a new trial both on liability and damages even if the judgment notwithstanding the verdict is overturned.
Although the district court’s ruling on the alternative motion for new trial involved the exercise of a measure of discretion, a stringent standard applies when the motion is based on insufficiency of evidence. A motion for new trial may be granted on this ground only if the verdict is against the “great weight” of the evidence, J & H Auto Trim Co. v. Bellefonte Ins. Co., 677 F.2d 1365, 1373 (11th Cir.1982); Conway v. Chemical Leaman Tank Lines, Inc., 610 F.2d 360, 363 (5th Cir.1980), or “it is quite clear that the jury has reached a seriously erroneous result”. Coffran v. Hitchcock Clinic, Inc., 683 F.2d 5, 6 (1st Cir.1982). In this case, the jury’s verdict, rather than being clearly contrary to the weight of the evidence, was a more-than-defensible resolution of a difficult issue.
Neither do we think the single issue submitted to the jury (defendant’s power to restrain competition) was so intertwined with the issue of damage that the latter cannot be submitted to a new jury without such confusion and uncertainty that the separate trial would amount to denial of a fair trial. See Gasoline Products Co. v. Champlin Refining Co., 283 U.S. 494, 499-500, 51 S.Ct. 513, 514-515, 75 L.Ed. 1188 (1931). Defendant has offered nothing but a general assertion to support its contention to the contrary. Greenwood Ranches, Inc. v. Skie Construction Co., 629 F.2d 518, 522 (8th Cir.1980). See C. Wright & A. Miller, Federal Practice and Procedure, § 2814 (1973).
Reversed on the appeal, affirmed on the cross-appeal, and remanded for further proceedings consistent with this opinion.7