California Computer Products, Inc. (“Cal-Comp”) appeals from the judgment entered on a directed verdict in favor of appellee International Business Machines Corp. (“IBM”) as to all counts of its complaint charging IBM with violations of § 2 of thé Sherman Act, 15 U.S.C. § 2 . We affirm.

I. Background and Proceedings Below

IBM is one of the largest industrial corporations in the world. It achieved technical leadership in the computer industry over other early entrants, such as Sperry Rand, in the mid-1950’s and thereafter pioneered the development of many electronic data processing products, including the disk products involved in this litigation.

Disk products are part of a broader category of what is known as peripheral equipment, such as disks, tapes, printers, and terminals, which is connected to the central processing unit (“CPU”) to enable the data processing system to perform particular functions. Included in the reference to disk products are disk drives, devices using magnetic disks similar in appearance to phonograph records to store information, and controllers, used for communication between disk drives and the CPU. Occasionally these devices are built into the CPU; alternatively, they exist as external components that may be “plugged into” the CPU. As a general purpose computer systems manufacturer, IBM sells both CPUs and peripherals, including disk products.

CalComp began manufacturing computer products in 1960, when it made plotting devices — peripheral equipment that provides graphic, printed or pictorial output. CalComp claims no injury with respect to these products. With the acquisition of Century Data Systems in 1969, CalComp entered the disk products market, manufacturing disk drives and controllers that were “plug compatible” with IBM’s and other suppliers’ CPUs. CalComp’s business strategy with respect to IBM-compatible disk products was straightforward: copy and, where possible, improve upon an IBM design, and undersell IBM to its own customers. By the “reverse engineering” of simply buying a device from IBM, taking it apart, and building a similar one, CalComp was able to avoid IBM’s expenditures for research and development and pass the savings on through lower prices.

CalComp commenced this lawsuit on October 3, 1973. The complaint alleged that IBM’s introduction of new CPUs and disk products, its price cuts on existing disk products, its leasing policies, and other marketing practices prevented CalComp from effectively competing with IBM for disk product sales and thus violated § 1 and § 2 of the Sherman Act. CalComp alleged and attempted to prove that these acts by IBM took place within a ten year span, from late 1963 to 1972, resulting in treble damages of $306 million. Following over three years of discovery and pretrial, trial to a jury began on November 15,1976. At the conclusion of fifty-four days of trial covering three months, the district court granted IBM’s motion for directed verdict on February 11, 1977.

The records and transcript on this appeal comprise 132 volumes. Voluminous briefs and supplemental briefs by the parties and amicus briefs were permitted. We have considered all of the arguments advanced and scrutinized pertinent parts of the record, particularly in view of the nature of the appellate task on review of a directed verdict.

II. Antitrust Standing

CalComp has asserted that IBM’s actions created anticompetitive effects on three classes of IBM competitors: (1) general purpose computer systems manufacturers, (2) leasing companies and (3) IBM-compatible peripheral equipment manufacturers. We believe that CalComp, an IBM-compatible peripheral equipment manufacturer, lacks antitrust standing as to the first two categories of claims.

Section 4 of the Clayton Act, 15 U.S.C. § 15 , authorizing private antitrust suits for damages, provides in part:

This statute confers standing to sue only upon those persons causally injured by antitrust violations. Kapp v. National Football League, 586 F.2d 644, 648-49 (9th Cir. 1978); John Lenore & Co. v. Olympia Brewing Co., 550 F.2d 495, 498-99 (9th Cir. 1977). Moreover, in order to prevail the plaintiff must prove not only injury causally linked to the asserted violation, but also that the injury is of the type the antitrust laws were intended to prevent. Brunswick Corp. v. Pueblo Bowl-O-Mat, Inc., 429 U.S. 477, 489 , 97 S.Ct. 690 , 50 L.Ed.2d 701 (1977); John Lenore & Co. v. Olympia Brewing Co., 550 F.2d at 498-99 ; In re Multidistrict Vehicle Air Pollution M.D.L. No. 31, 481 F.2d 122, 125 (9th Cir. 1973), cert. denied, 414 U.S. 1045 , 94 S.Ct. 551 , 38 L.Ed.2d 336 (1975). The plaintiff’s burden of proving the former is satisfied by proof of some damage flowing from the antitrust violation. Zenith Radio Corp. v. Hazeltine Research Inc., 395 U.S. 100 , 114 n.9, 89 S.Ct. 1562 , 23 L.Ed.2d 129 (1969). Satisfying the latter burden is dependent on a showing that the injury was caused by a reduction, rather than an increase, in competition flowing from the defendant’s acts, since “[t]he antitrust laws . . . were enacted for ‘the protection of competition not competitors,’ ” Brunswick Corp. v. Pueblo Bowl-O-Mat, Inc., 429 U.S. at 488 , 97 S.Ct. at 697 , quoting Brown Shoe Co. v. United States, 370 U.S. 294, 320 , 82 S.Ct. 1502 , 8 L.Ed.2d 510 (1962). See Oreck Corp. v. Whirlpool Corp., 579 F.2d 126, 133 (2d Cir. 1978). Accordingly, the plaintiff must demonstrate that the defendant’s conduct was intended to or did have some anticompetitive effect beyond his own loss of business or the market’s loss of a competitor. See Knutson v. Daily Review, Inc., 548 F.2d 795, 803 (9th Cir. 1976). Moreover, it is not sufficient for an antitrust plaintiff to allege an indirect ripple effect. As this court wrote in John Lenore & Co.:

550 F.2d at 499 .

In the present case CalComp has alleged that IBM’s actions injured general purpose computer systems manufacturers and leasing companies. But CalComp does not include itself among these two classes of IBM competitors. Nor does CalComp’s evidence demonstrate a direct causal injury which would afford it standing. Rather, at best CalComp argues that injury to these two groups has had an indirect ripple effect upon it. As John Lenore & Co. indicates, such an indirect ripple effect is not sufficient to allow CalComp to sue for treble damages on its first two categories of claims.

III. CalComp’s Claims as to IBM-Compatible Peripheral Equipment Manufacturers

A. Standard of Review on Appeal from Directed Verdict

As a general rule, the district court has the power to direct a verdict if “the evidence permits only one reasonable conclusion as to the verdict.” Fountila v. Carter, 571 F.2d 487, 489-90 (9th Cir. 1978), quoting Kay v. Cessna Aircraft Co., 548 F.2d 1370, 1372 (9th Cir. 1977); see Syufy Enterprises v. National General Theatres, 575 F.2d 233 , 235 (9th Cir. 1978). The district court must consider all the evidence — both favorable and unfavorable. But in order to avoid passing on the credibility of witnesses and weighing contradictory evidence, the court must resolve all inferences in favor of the party with the burden of persuasion, because

Fount-Wip, Inc. v. Reddi-Wip, Inc., 568 F.2d 1296, 1301 (9th Cir. 1978), quoting Cockrum v. Whitney, 479 F.2d 84, 86 (9th Cir. 1973) and Tennant v. Peoria & Pekin Union. Ry., 321 U.S. 29, 35 , 64 S.Ct. 409 , 88 L.Ed. 520 (1944); see Marquis v. Chrysler Corp., 577 F.2d 624, 639 (9th Cir. 1978); Kay v. Cessna Aircraft Co., 548 F.2d at 1372 .

Thus, this court in Maheu v. Hughes Tool Co., 569 F.2d 459, 464 (9th Cir. 1977), upholding the denial of a directed verdict against the party with the burden of persuasion, made it clear that application of the general standard of Fountila and Kay, supra, required it to view the evidence “in the light most favorable to the party opposing the motion,” and that it “must examine all the evidence” (emphasis added). See also id. at 481 (concurring and dissenting opinion); Wescott v. Impresas Armadoras, S.A., 564 F.2d 875, 882 (9th Cir. 1977); Santa Clara Valley Distributing Co. v. Pabst Brewing Co., 556 F.2d 942, 944 (9th Cir. 1977); Kay v. Cessna Aircraft Co., 548 F.2d at 1372; Chisholm Brothers Farm Equipment Co. v. International Harvester Co., 498 F.2d 1137, 1140 (9th Cir.), cert. denied, 419 U.S. 1023 , 95 S.Ct. 500 , 42 L.Ed.2d 298 (1974).

In order to benefit from the favorable inferences available under this standard, the party against whom the motion is made must present “substantial evidence.” As stated in Rutledge v. Electric Hose & Rubber Co., 511 F.2d 668 (9th Cir. 1975):

Id. at 677 (emphasis added), quoting Cleary v. National Distillers & Chemical Corp., 505 F.2d 695, 696 (9th Cir. 1974). The “sole issue” in an appeal from a directed verdict against the party with the burden of persuasion is thus the sufficiency of the evidence of the appellant’s claim. Cleary v. National Distillers & Chemical Corp., 505 F.2d at 696 . As stated in Chisholm Brothers Farm Equipment Co. v. International Harvester Co., 498 F.2d at 1140 , “the correct standard is whether or not, viewing the evidence as a whole, there is substantial evidence present that could support a finding . . . for the nonmoving party.” Accord, Janich Bros., Inc. v. American Distilling Co., 570 F.2d 848 , 853 & n.2 (9th Cir. 1977) (substantial evidence is “more than a mere scintilla” and consists of “such relevant evidence as a reasonable mind might accept as adequate to support a conclusion”), ce rt. denied, 439 U.S. 829 , 99 S. Ct.. 103, 58 L.Ed.2d 122 (1978).

The standard for determining the propriety of a directed verdict — identical to that for determining the propriety of a judgment n. o. v., Fountila v. Carter, 571 F.2d at 489 ; Cockrum v. Whitney, 479 F.2d at 85—is the same for district and appellate judges. Maheu v. Hughes Tool Co., 569 F.2d at 481 (concurring and dissenting opinion).

In its amicus brief the Department of Justice argues that:

However, in Santa Clara Valley Distributing Co. v. Pabst Brewing Co., 556 F.2d at 944 -45 n.1, this court dealt extensively and definitively with the meaning of Poller and Hospital Building Co. in the context of a directed antitrust verdict. That case, emphasizing that the Supreme Court’s disapproval of “summary proceedings” dealt with “pre trial motions” (emphasis original), concluded that “the same circumspection is not required in the case of directed verdicts.” Rather, the usual standard described above applies:

Id., quoting Rutledge v. Electric Hose & Rubber Co., 511 F.2d at 677 , and Cleary v. National Distillers & Chemical Corp., 505 F.2d at 696 . See Marquis v. Chrysler Corp., 577 F.2d at 639 & n.25

B. General § 2 Doctrine

CalComp contends that the evidence was sufficient to show that particular conduct on the part of IBM, detailed below, violated either or both the monopolization and attempt to monopolize clauses of § 2 of the Sherman Act. In order to reverse the district court’s judgment as to either of these claims, it must be that after viewing the evidence in the light most favorable to Cal-Comp, there is substantial evidence of every essential element of that claim.

1. Monopolization

There are three essential elements to a successful claim of § 2 monopolization:

The first two elements are derived from § 2 itself, and were explicated in United States v. Grinnell Corp., 384 U.S. 563, 570-71 , 86 S.Ct. 1698 , 16 L.Ed.2d 778 (1966). See Greyhound Computer Corp. v. IBM, 559 F.2d 488, 492 (9th Cir. 1977); Moore v. Jas. H. Matthews & Co., 550 F.2d 1207, 1218 (9th Cir. 1977).

As reiterated in Greyhound, monopoly power — the first element — “is the power to control prices or exclude competition.” 559 F.2d at 496 , quoting United States v. E. I. du Pont de Nemours & Co., 351 U.S. 377, 391 , 76 S.Ct. 994 , 100 L.Ed. 1264 (1956). The du Pont definition of monopoly power has been applied principally with reference to the defendant’s share of the relevant product and geographic markets. Greyhound, 559 F.2d at 496 . See United States v. Grinnell Corp., 384 U.S. at 571 , 86 S.Ct. 1698 ; United States v. E. I. du Pont de Nemours & Co., 351 U.S. at 399, 404 , 76 S.Ct. 994 ; American Tobacco Co. v. United States, 328 U.S. 781, 797 , 66 S.Ct. 1125 , 90 L.Ed. 1575 (1946); Twin City Sportservice, Inc. v. Charles O. Finley & Co., 512 F.2d 1264, 1274 (9th Cir. 1975). But see Moore v. Jas. H. Matthews & Co., 550 F.2d at 1219 (“the requisite power also can be demonstrated by evidence of the exercise of actual control over prices or exclusion of competitors”).

The second element of a successful monopolization claim requires that the conceded monopolist have engaged in “willful” acts directed at establishing or retaining its monopoly, “as distinguished from growth or development as a consequence of a superior product, business acumen, or historic accident.” United States v. Grinnell Corp., 384 U.S. at 571 , 86 S.Ct. at 1704 . That case found the defendant’s acts sufficient to meet the § 2 conduct requirement because they constituted “unlawful and exclusionary practices.” Id. at 576, 86 S.Ct. 1698 .

The plaintiff need not show that the conceded monopolist’s acts were of a kind that would be unlawful for an ordinary enterprise. Greyhound Computer Corp. v. IBM, 559 F.2d at 498 . Rather, the plaintiff must show that the defendant’s acts “unnecessarily excluded competition” from the relevant market. Id. Nor is it necessary to show a specific intent to eliminate a competitor. Sunkist Growers, Inc. v. Winckler & Smith Citrus Products Co., 284 F.2d 1, 26 (9th Cir. 1960), modified, 289 F.2d 933 (1961), rev’d on other grounds, 370 U.S. 19 , 82 S.Ct. 1130 , 8 L.Ed.2d 305 (1962). The defendant’s acts are properly analyzed analogously to contracts, combinations and conspiracies under § 1 of the Sherman Act: the test is whether the defendant’s acts, otherwise lawful, were unreasonably restrictive of competition. See Gough v. Rossmoor Corp., 487 F.2d 373, 376 (9th Cir. 1973). While this “in large measure” has the effect of making acts of monopolization “merely the end products of conduct which violates § 1,” “that is not always true.” United States v. Griffith, 334 U.S. 100, 106 , 68 S.Ct. 941 , 92 L.Ed. 1236 (1948). Section 1 is limited to concerted activity and contractual restraints, while under § 2, individual activity may also give rise to liability. See Moore v. Jas. H. Matthews & Co., 473 F.2d at 332-33.

The third element — plaintiff’s resultant “antitrust” injury — has already been discussed, Part II supra.

2. Attempt to Monopolize

There are four elements to a successful claim of § 2 attempt to monopolize:

The first element was explained in Times-Picayune Publishing Co. v. United States, 345 U.S. 594, 626 , 73 S.Ct. 872 , 97 L.Ed. 1277 (1953), wherein the Supreme Court differentiated the requisite levels of intent under the monopolization and attempt to monopolize clauses of the Sherman Act:

The intent to “build monopoly,” given the du Pont definition of monopoly power, is logically synonymous with the intent to control prices or exclude competition in the relevant market. See Gough v. Rossmoor Corp., 585 F.2d 381, 390 (9th Cir. 1978); Marquis v. Chrysler Corp., 577 F.2d 624, 641 (9th Cir. 1978); Janich Bros., Inc. v. American Distilling Co., 570 F.2d 848, 853 (9th Cir. 1977), cert. denied, 439 U.S. 829 , 99 S.Ct. 103 , 58 L.Ed.2d 122 (1978); Greyhound Computer Corp. v. IBM Corp., 559 F.2d at 504 ; Knutson v. Daily Review, Inc., 548 F.2d at 813-14; Lessig v. Tidewater Oil Co., 327 F.2d 459, 474 (9th Cir.), cert. denied, 377 U.S. 993 , 84 S.Ct. 1920 , 12 L.Ed.2d 1046 (1964). “Direct evidence” of specific intent to control prices or destroy competition, however, is not always necessary when the attempt claim is “founded upon a substantial claim of restraint of trade” — i. e., a § 1 violation. In these circumstances the requisite specific intent may be inferred. Trixler Brokerage Co. v. Ralston Purina Co., 505 F.2d 1045, 1051-52 (9th Cir. 1974); see Gough v. Rossmoor Corp., 585 F.2d at 390 ; Janich Bros., Inc. v. American Distilling Co., 570 F.2d at 854 ; Knutson v. Daily Review, Inc., 548 F.2d at 814; Bushie v. Stenocord Corp., 460 F.2d 116, 120-21 (9th Cir. 1972). Market power is relevant to determining whether such an inference is proper; but where a § 1 violation “clearly” exists, proof of market power is unnecessary to support an inference of specific intent. Janich Bros., Inc. v. American Distilling Co., 570 F.2d at 854 n.4; Hallmark Industry v. Reynolds Metals Co., 489 F.2d 8, 12-13 (9th Cir. 1973), cert. denied, 417 U.S. 932 , 94 S.Ct. 2643 , 41 L.Ed.2d 235 (1974).

Conversely, “even though the restraint effected may be reasonable under § 1, it may constitute an attempt to monopolize forbidden by § 2 if a specific intent to monopolize may be shown.” United States v. Columbia Steel Co., 334 U.S. 495, 531-32 , 68 S.Ct. 1107, 1126 , 92 L.Ed. 1533 (1948). However, “[o]rdinarily specific intent is difficult to prove,” Hallmark Industry v. Reynolds Metals Co., 489 F.2d at 12 ; see, e. g., United States v. Columbia Steel Co., 334 U.S. at 532-34, 68 S.Ct. 1107 , and thus more commonly it is shown indirectly by proof of illegal conduct and, where necessary, market power.

The “predatory or anticompetitive conduct” element of § 2 attempt, like the conduct element of monopolization, encompasses more than violations of § 1. See,. e. g., Knutson v. Daily Review, Inc., 548 F.2d at 814 (attempt requires proof of “some illegal [under Section 1] or predatory activity”) (emphasis added). The reason for this was stated in Moore v. Jas. H. Matthews & Co., 473 F.2d at 332:

Additionally, “[s]ection 1 also prohibits ‘contracts’ that restrain trade,” id. at 333, and since individual actions may violate § 2, no contractual agreement is required. Nonetheless, under § 2 attempt — as with § 1 monopolization — individual conduct is measured against the same “reasonableness” standard governing concerted and contractual activity under § 1. See note 7 infra; Greyhound Computer Corp. v. IBM Corp., 559 F.2d at 505 n.37 (“§ 2 was intended to prohibit unreasonable restraints of trade that exclude competition even when they are imposed by a single trader”) (emphasis added).

The third element of § 2 attempt, the requirement that a defendant’s demonstrated specific intent to control prices or destroy competition have a “dangerous probability of success,” may be satisfied either by direct proof of market power, see Janich Bros., Inc. v. American Distilling Co., 570 F.2d at 853 , or by inference from the proven specific intent itself, see Greyhound Computer Corp. v. IBM Corp., 559 F.2d at 504 . Because this element may be inferred from the existence of a specific intent in proper cases, it is not an “essential” element of an attempt claim, Lessig v. Tidewater Oil Co., 327 F.2d at 474 . For the same reason, neither is proof of any particular degree of market power necessarily an “independent” element of such a claim. Hallmark Indus try v. Reynolds Metals Go., 489 F.2d at 12 n

In addition to the first three elements necessary to establish a prima facie case of attempt, the fourth element — causal “antitrust” injury — is as necessary to confer standing to sue and to support a claim for damages under § 2 attempt as it is under § 2 monopolization. 15 U.S.C. § 15 ; see Greyhound Computer Corp. v. IBM Corp., 559 F.2d at 505 & n.38; Part II supra.

C. CalComp’s Evidence

The foregoing analysis of the elements of § 2 monopolization and attempt to monopolize demonstrates that if CalComp presented sufficient evidence to go to the jury on the issue of IBM’s monopoly power in a relevant market, both of its § 2 claims may be analyzed together. The only remaining elements necessary to establish the monopolization offense in this case would be monopolizing conduct and causal “antitrust” injury. If CalComp failed to show either of these, its attempt claim must necessarily also fail: first, because as discussed supra, “conduct lawful for a monopolist must, a fortiori, be excluded as a basis for the attempt offense,” and second, because the requirement of causal “antitrust” injury is common to both.

1. Market Evidence

Earlier it was noted that CalComp sought to define three relevant product markets: (a) a general purpose computer systems market; (b) an all disk drive and associated controller market; and (c) a plug-compatible disk drive and associated controller market (which excluded disk products for use with the CPU’s of other manufacturers). Several experts and industry executives testified, and a great deal of documentary evidence was admitted, in an effort to support each of these market definitions. Not only did the evidence concerning the various market definitions often conflict, but also the testimony and documents purportedly supporting a single market were sometimes ■ internally inconsistent. Still, we assume arguendo that the third category is an appropriately defined product market. As noted earlier, it is the only market about which CalComp has standing to sue. See Part II supra.

The evidence of IBM’s share of the product markets was similarly irresolute. For example, CalComp introduced the testimony of employees of IBM and other industry members concerning IBM’s share of the loosely defined general purpose computer systems market; their estimates ranged from as low as 60% to as high as 80%, while IBM’s proof was designed to show that none of these figures was even based on the particular market alleged. CalComp offered expert testimony that in 1970, 1971 and 1972, IBM’s share of the all disk drive and associated controller market was 79.4%, 70.1% and 67.6% respectively, while IBM’s evidence was that its share was under 30% during these years: CalComp’s figures reflected cumulative shipments of disk products, while IBM’s data was based on annual shipments. IBM contended that because its share of market was rapidly declining (from 100% when it invented disk products in 1960 to 25% in 1975), cumulative measures were misleading. In the plug-compatible market, IBM — again, because it invented these products — began with a 100% share; here too, CalComp’s own evidence showed that IBM’s share of market declined steadily.

While the district judge expressly made no finding on the adequacy of proof of share of a relevant market, he noted “in passing” that he was not inclined to direct a verdict solely on that ground. Because we conclude that there was no substantial evidence that any of IBM’s alleged acts both constituted unreasonable conduct for a monopolist and produced causal “antitrust” injury to CalComp, we need not become enmeshed in the hundreds of pages of conflicting and complex evidence on this element of CalComp’s claim. For purposes of this decision, we assume that IBM possessed monopoly power in the third purported market during the period 1963 to 1972.

We may therefore accept, with one important addition, CalComp’s statement in its brief that the main issue of this appeal is “whether the acts and practices of IBM . constitute a violation of either the monopolization or attempt to monopolize clauses of Section 2 of the Sherman Act, or both.” Should we decide this question in the affirmative, the additional question we must answer would be whether any loss CalComp suffered actually resulted from the asserted § 2 violation (as opposed to lawful competitive practices).

2. Evidence of § 2 Violations

CalComp argues that IBM directly injured it in three ways. Principally, Cal-Comp contends that IBM engaged in “predatory” pricing by cutting its peripheral equipment prices in response to competition from CalComp and other manufacturers. CalComp also attempts to show that IBM made design changes on certain of its CPUs, disk drives and controllers of no technological advantage and solely for the purpose of frustrating competition from plug-compatible manufacturers. Finally, it urges that IBM raised CPU prices in an effort to offset revenue losses caused by its price reductions on peripheral equipment, and that these price increases constituted impermissible conduct for a monopolist.

a. Price cuts on peripherals

When IBM introduced its new System 370 Model 145 in September, 1970, it designated as the standard disk drive for use with this CPU a reworked version of earlier disk drives that had been replaced by competitive equipment and returned to IBM. The monthly rental for this three-drive disk product, the 2319A, was 30% below that for IBM’s other four-, two-, and single-drive disk products on a per-drive basis — -in part because the reuse of earlier disk drives permitted reduced development and manufacturing costs. The integrated control function in the Model 145-2319A system was also priced 60% lower than earlier stand-alone controllers, also due at least in part to cost-saving design changes. As opposed to the System 370 Models 155 and 165, which were “high-end” processors using the newly-designed 3300 high-speed disk product, the Model 145 was to be the lower-priced mainstay of System 370 for commercial applications. Viewing all of the relevant evidence in the light most favorable to CalComp, IBM’s purpose in offering the lower-priced 2319A disk drive in conjunction with its Model 145 was twofold: primarily, to regain market share from peripheral equipment manufacturers who were flourishing under IBM’s high price umbrella; and secondarily, to reduce the overall price of the 145 system, thereby providing the lower-priced alternative to the higher-performance, higher priced models using the 3300 disk product that was to be the 145’s market niche. CalComp introduced no evidence that the lower-priced 2319A was not “substantially profitable,” as IBM asserted.

Three months after the announcement of the 2319A, IBM introduced the 2319B disk drive for use with all System 360 models. Like the 2319A, this was also a “retread” manufactured from older IBM disk drives that had been replaced with other equipment and returned by customers. IBM similarly priced the three-drive 2319B more than 30% below its other disk drives on a per-drive basis. Also, it made no “additional use” charges (for use above a fixed number of hours per month) for the 2319B as it did with its other disk drives, in effect a further reduction in price vis-a-vis these other products. The 2319B, viewing all of the evidence in the light most favorable to CalComp, was introduced for the express purpose of abating competitive inroads by other peripheral manufacturers. Cal-Comp introduced no evidence that the 2319B, like the 2319A, was not “substantially profitable.”

Following IBM’s 2319B announcement, virtually all of the plug compatible manufacturers reduced their own prices below the 2319B. As a further response to this competition, IBM reduced its prices on peripheral products in still another way with the introduction of its Fixed Term Plan (“FTP”) on May 27,1971, described by CalComp as “the single most important act in the CalComp litigation and the principal basis for its damage claim.” Prior to the introduction of FTP, IBM offered its products only for sale or 30-day lease, although virtually all of its competitors offered reduced prices for longer leases ranging from one to eight years. Under FTP, IBM customers were given the additional option of an 8% discount for signing a one-year lease and a 16% discount for a two-year lease on certain peripheral equipment, principally disk drives, tapes and printers. “Additional use” charges were also eliminated on FTP leases, consistent with the practice of most of IBM’s competitors. Customers — but not IBM — could cancel the leases in return for a termination charge. Purchase prices on FTP products were also reduced by -15%. The evidence at trial was uncontroverted that FTP price reductions were expected to return a profit of 30% of revenue before taxes. Lease revenues would be reduced through 1972, but through 1975, FTP was expected to generate $165 million more profit than would have been the case without FTP lease options. This increased profitability due to FTP stemmed not only from an expected increase in market share, but also from longer average lease lives and reduced sales, reconditioning and reinstallation costs.

Finally, following its adoption of FTP, IBM introduced two other new products: the Integrated Storage Controller (“ISC”) and the Integrated File Adaptor (“IFA”). ISC and IFA were options available for use with certain System 370 models, allowing direct attachment of disk products to the CPU. Customers choosing one of these “inboard” storage control units could save money by avoiding purchase or lease of a separate controller for some of their disk products, because the price of the ISC and IFA options were lower. Thus, ISC and IFA represented price cuts vis-a-vis IBM’s “outboard” controllers. However, there is little evidence in the record that the ISC and IFA inboard storage control units were priced in response to competition. Rather, the evidence indicates that these products were less expensive to produce. CalComp’s chairman, for example, testified that IFA enabled the CPU to perform functions “which had been performed in the standalone controller,” at a “much cheaper” cost of design and manufacture. For purposes of analysis, however, we will assume that the IFA and ISC options, like the 2319A, 2319B and FTP products, were priced primarily in response to competition from the plug compatible manufacturers.

The test of the reasonableness of the foregoing pricing actions, and the principal question facing us in this case, is whether IBM — which was the inventor and dominant supplier of the disk products in question — had the right to respond to the lower prices of its competitors with re duced, but still substantially profitable, prices on its own products. We conclude that it did.

CalComp’s principal damages claim is for lost revenues as a result of price reductions it made following IBM’s 2319B and FTP announcements. But since these price reductions admittedly resulted from competition by IBM — and since, as both CalComp’s and IBM’s evidence clearly demonstrates, IBM’s stimulus to price competition was in turn competition from peripheral equipment manufacturers such as Cal-Comp — it is impossible to say that Cal-Comp’s losses represent compensable “injury” from acts of IBM unnecessarily excluding or restricting competition. See Brunswick Corp. v. Pueblo Bowl-O-Mat, Inc., 429 U.S. at 488 , 97 S.Ct. 690 . Rather, IBM’s price cuts were a part of the very competitive process the Sherman Act was designed to promote. To accept CalComp’s position would be to hold that IBM could not compete if competition would result in injury to its competitors, an ill-advised reversal of the Supreme Court’s pronouncement that the Sherman Act is meant to protect the competitive process, not competitors. See id.

In Grinnell, as noted supra, the Supreme Court excepted from monopolizing conduct those actions directed toward establishing growth by means of “a superior product, business acumen, or historic accident.” 384 U.S. at 571 , 86 S.Ct. at 1704 . IBM’s dominance in disk products, of course, was not due to historic accident. CalComp witnesses repeatedly testified that IBM’s position and success were due to its capable management, technological leadership, market orientation and superior products. Particularly relevant is the fact that IBM invented the disk products that Cal-Comp and other manufacturers copied. As one of CalComp’s witnesses, a former chairman of Scientific Data Systems, acknowledged, “IBM pioneered in the use of moveable head disk technology . . . .” More bluntly, according to CalComp’s Chairman:

Granted that IBM’s technological innovations resulted in “growth as a consequence of a superior product,” it was entitled to maintain its consequent dominant position in the market it created through “business acumen,” which we take to include shrewdness in profitable price competition. The Sherman Act does not draw a distinction between competition on the bases of price and of performance: the two are inseparable parts of any competitive offering. Where the opportunity exists to increase or protect market share profitably by offering equivalent or superior performance at a lower price, even a virtual monopolist may do so. Cf. United States Steel Corp. v. Fortner Enterprises, Inc., 429 U.S. 610 , 612 n.1, 97 S.Ct. 861, 864, n.1 , 51 L.Ed.2d 80 (1977) (in attempt context, purposes of “ ‘increasing sales’ and ‘increasing market share’ are normal business goals, not forbidden by § 2” without more). As Judge Aid-rich observed in Dehydrating Process Co. v. A. O. Smith Corp., 292 F.2d 653, 657 (1st Cir. 1961), “the antitrust laws do not require a business to cut its own throat.”

The boundaries of reasonable price competition have recently been defined in this circuit. In Hanson v. Shell Oil Co., 541 F.2d 1352, 1359 (9th Cir. 1976), cert. denied, 429 U.S. 1074 , 97 S.Ct. 813 , 50 L.Ed.2d 792 (1977), we affirmed a directed verdict for a defendant charged with attempted monopolization, holding that the plaintiff’s failure to show that the defendant’s prices were below its marginal or average variable costs “was a failure as a matter of law to present a prima facie case under § 2.” More recently in Janich Brothers, also an attempt case, we stated that “ ‘pricing at marginal cost is the competitive and socially optimal result’ ” of § 2 enforcement. 570 F.2d at 857 , quoting Areeda & Turner, Predatory Pricing and Related Practices Under Section 2 of the Sherman Act, 88 Harv.L.Rev. 697, 711 (1975). The analysis in Hanson and Janich Brothers is precisely apposite in the monopolization context. Indeed, the preceding passage quoted in Janich Brothers dealt specifically with pricing by a monopolist. The thrust of this analysis is that price reductions up to the point of marginal cost are consistent with competition on the merits, since in this ease only less efficient firms will be disadvantaged, while a firm pricing below marginal cost by definition incurs losses, so that competition on the basis of efficiency in this situation is frustrated.

We recognize that refinement of the marginal or average variable cost test will be necessary as future cases arise. For instance, limit pricing by a monopolist might, on a record which presented the issue, be held an impermissible predatory practice. See, e. g., P. Areeda & D. Turner, Antitrust Law ¶ ¶ 711-722 (1978); Williamson, Predatory Pricing: A Strategic and Welfare Analysis, 87 Yale L.J. 284 (1977); Areeda & Turner, Williamson on Predatory Pricing, 87 Yale L.J. 1337 (1978); Williamson, A Preliminary Response, 87 Yale L.J. 1353 (1978); Scherer, Predatory Pricing and the Sherman Act: A Comment, 89 Harv.L.Rev. 868 (1976); Areeda & Turner, Scherer on Predatory Pricing: A Reply, 89 Harv.L. Rev. 891 (1976); Scherer, Some Last Words on Predatory Pricing, 89 Harv.L.Rev. 901 (1976). And we do not foreclose the possibility that a monopolist who reduces prices to some point above marginal or average variable costs might still be held to have engaged in a predatory act because of other aspects of its conduct. On this record, however, IBM’s pricing policies were not predatory.

The case before us presents an a fortiori situation as compared to Hanson and Janich Brothers. CalComp has not only failed to produce evidence of pricing below marginal or average variable cost, but it has also failed as well to introduce any evidence to controvert IBM’s substantial proof that its price cuts were highly profitable. Moreover, the evidence of both parties established that IBM’s disk price reductions were a response to lower-priced competition to which IBM was rapidly losing its disk business. Were § 2 interpreted not to exempt price cuts from attack under these circumstances, there could be no adequate guidelines for a jury to decide the issue of whether the prices at issue were “reasonable.” The directed verdict as to CalComp’s claims based on IBM’s price competition was therefore proper.

b. Design changes

As noted above, when IBM introduced its System 370 Model 145 in September, 1970, it announced the 2319A as the standard disk product for use with the 145. The control function for the 2319A disk drive was integrated into the Model 145 CPU, and thus the interface between the disk drive and its control function was different from earlier models. CalComp claimed that it was competitively disadvantaged as a result of these design changes, because it could not legally begin to copy the 2319A until IBM shipped the first of these disk drives, thereby disclosing the design requirements.

In February, 1971, IBM introduced its optional Integrated Storage Controller for use with its System 370 Models 158 and 168 which integrated the disk control function into the CPU. CalComp claimed it was injured by the introduction of ISC and the similar IFA, described above, because it was thereby precluded from replacing the control functions on CPUs with these options.

CalComp characterized these design changes as “technological manipulation” which did not improve performance. It also complained of the fact that the newly integrated functions were priced below their non-integrated counterparts. But as we have stated, price and performance are inseparable parts of any competitive offering; and equivalent function at lower cost certainly represents a superior product from the buyer’s point of view. The evidence at trial was uncontroverted that integration was a cost-saving step, consistent with industry trends, which enabled IBM effectively to reduce prices for equivalent functions. Moreover, there was substantial evidence as well that in the case of Models 145,158 and 168 the integration of control and memory functions also represented a performance improvement.

One of CalComp’s witnesses stated: “I think in general the manufacturer will try and minimize his costs and where he integrates the control unit the assumption must be that he is achieving a lower cost solution.” Other of CalComp’s evidence showed that among the reasons a separate control unit is more expensive than integrated control circuitry are that the former requires its own cabinet, frames, power supply, additional cabling and electronics. According to an IBM witness, the monolithic systems technology that preceded the 145-2319A system required a large standalone controller, whereas the new generation technology represented by the 145-2319A system produced a comparable control function “which was in the area of ten times smaller . . . . [Y]ou could now put that into the 145 system, utilizing its frames and its covers and then passing on the advantages of that to the customer in a price reduction.” CalComp’s Chairman stated that as a result of integration, the customer uses less floor space which “tends to be relatively expensive in a computer room.”

IBM, assuming it was a monopolist, had the right to redesign its products to make them more attractive to buyers— whether by reason of lower manufacturing cost and price or improved performance. It was under no duty to help CalComp or other peripheral equipment manufacturers survive or expand. IBM need not have provided its rivals with disk products to examine and copy, see 3 P. Areeda & D. Turner, Antitrust Law ¶738, at 286 (1978), nor have constricted its product development so as to facilitate sales of rival products. The reasonableness of IBM’s conduct in this regard did not present a jury issue.

c. Increases in CPU prices

In July, 1971, IBM announced price increases on certain of its CPUs of up to 8%. The weighted average price increase of all CPUs was 1.5%. CalComp argued that these price increases, instituted two months after FTP was announced, were designed to offset the short-term revenue losses expected from FTP price cuts. After FTP users were “locked into” IBM peripherals for one or two year lease terms, according to Cal-Comp, they would be unable to change CPUs, thus enabling IBM to raise CPU prices without appreciably affecting demand or profitability.

However, CalComp’s own evidence indicated that the net effect of FTP and the CPU price increases was not expected to be a “wash” with respect to either CPU demand or overall profitability. CalComp introduced an IBM memo written after the CPU price increases went into effect stating that “the net effects of the Fixed Term Plan and [CPU] price change will probably be a wash insofar as business volumes are concerned.” (Emphasis added). The author of this memo went on to explain that FTP would have the effect of increasing peripheral volume, while the CPU price increases would probably induce “adverse reaction on the part of some customers . .” In terms of overall business volume, the author concluded, “[a]ny decrease in system acceptances which occurs will probably be offset by . the Fixed Term Plan.” In other words, there was not a single demand function for CPUs and peripherals; IBM did not expect to raise CPU prices “with impunity,” as CalComp asserts, by virtue of the effects of FTP.

Moreover, there was no evidence that CPU price increases actually offset the FTP price cuts. According to CalComp’s amended complaint, “[peripheral devices . account for between 50% and 75% of the total value of a system configuration.” The trial testimony of Richard Whitcomb, Director of Product Planning at Itel and a former IBM employee, and Edwin McCollister, Director of Market Development at Burroughs, confirmed this. Since, as discussed supra, the FTP price cuts amounted to 8% and 16%, a jury would be hard pressed to conclude from CalComp’s evidence that the CPU price increases in fact offset FTP losses — i. e., that a weighted average price increase of 1.5% on CPUs made up for price cuts of 8% and 16% on peripheral products of equal or greater value.

Other evidence portrayed IBM’s decision to raise CPU prices as motivated by inflation and cost concerns. IBM had not raised its mainframe prices in over four years; its 1971 increases were approved as cost-justified by the newly-formed Price Commission under the Economic Stabilization Program then in effect.

Viewed in the light most favorable to CalComp, this evidence does not provide support for its “offset” theory. Especially since we have already determined that the FTP price cuts allegedly facilitated by the CPU price increases were themselves profitable and reasonable even for a monopolist in any event, it was proper to take the case from the jury on this issue.

IV. Conclusion

In its amicus brief, the Department of Justice questioned that the district court could have directed a verdict in view of the complexity of the facts in this case. Cal-Comp, too, contends that “this Court has [neither] the power [nor] obligation, to sift through the record to determine who should prevail” on appeal from the district court’s directed verdict.

The complexity of a plaintiff’s case, however, does not entitle it to avoid a di rected verdict if the evidence — no matter how much of it there may be — is weak. It is in fact the duty of both the trial and appellate courts to “sift through the record” to make this determination. Having closely analyzed each of CalComp’s arguments and the evidence on both sides in support thereof, drawing all reasonable inferences in Cal-Comp’s favor, we conclude that the district judge was correct in taking the case from the jury.

Nor does viewing the various acts of IBM collectively change our conclusion. The number of legal and evidentiary issues has required us to consider each instance of IBM’s alleged monopolizing conduct separately for purposes of analytical clarity. However, we are mindful of the fact that “plaintiffs should be given the full benefit of their proof without tightly compartmentalizing the various factual components and wiping the slate clean after scrutiny of each.” Continental Ore Co. v. Union Carbide & Carbon Corp., 370 U.S. 690, 699 , 82 S.Ct. 1404, 1410 , 8 L.Ed.2d 777 (1962); see Knutson v. Daily Review, Inc., 548 F.2d 795, 814 (9th Cir. 1976). But there can be no synergistic result such as CalComp claims from a number of acts none of which show causal antitrust injury to CalComp.

In Continental Ore, the Supreme Court held that the plaintiffs were entitled to introduce evidence “that there was a conspiracy and monopolization in existence when they came into the industry, and that they were eliminated in furtherance thereof.” 370 U.S. at 710, 82 S.Ct. at 1416 (emphasis added). Such proof connecting the acts allegedly directed against systems manufacturers and leasing companies to the acts affecting CalComp was lacking in this case. Moreover, even assuming that IBM gained monopoly power in a relevant market by virtue of the former categories of acts, the latter acts constituted reasonable, pro-competitive conduct for a monopolist. CalComp suffered no “antitrust” injury from these, nor indeed any injury at all as that term is defined in § 4 of the Clayton Act. See Brunswick Corp. v. Pueblo Bowl-O-Mat, Inc., 429 U.S. 477, 488 , 97 S.Ct. 690 , 50 L.Ed.2d 701 (1977).

Accordingly, the judgment of the district court entered on its directed verdict is

AFFIRMED.

. CalComp has dropped its § 1 claim on this appeal. . Although the district court may direct a verdict either against the party who does not bear the burden of persuasion or the party who does bear that burden, the amount of evidence required for a directed verdict differs. The party seeking a directed verdict must make a stronger showing of evidence if he bears the burden of persuasion. See United California Bank v. THC Financial Corp., 557 F.2d 1351, 1356 (9th Cir. 1977); Juhnke v. EIG Corp., 444 F.2d 1323, 1325 (9th Cir. 1971); Comment, “Directing the Verdict in Favor of the Party with the Burden of Proof,” 50 N.C.L.Rev. 843 (1972). Thus directed verdicts for defendants are more frequent than for plaintiffs. See C. Wright, Federal Courts 464 (3d ed. 1976). . This test is a hybrid of the so-called “new trial” and “most favorable evidence” tests. Shortly after the power of the federal courts to grant directed verdicts consistently with the seventh amendment was established in Galloway v. United States, 319 U.S. 372 , 63 S.Ct. 1077 , 87 L.Ed. 1458 (1943), two polar tests were proposed for determining whether the evidence in a particular case is sufficient to warrant a directed verdict against the bearer of the persuasion burden. The first, the “setting aside” or “new trial” test, permitted a directed verdict against the party with the burden of persuasion if the judge, looking at ail the evidence, both favorable and unfavorable, determined that he would be duty bound to set aside a verdict for that party. In the second, the “most favorable evidence” test, the court was to consider only the evidence favorable to the party with the persuasion burden, completely disregarding all unfavorable evidence, and determine whether a reasonable jury, viewing the evidence in the light most favorable to that party, could find every essential part in that party’s favor. If it could not, a directed verdict for the other party would be proper. See McBaine, “Trial Practice; Directed Verdicts; Federal Rule,” 31 Calif.L.Rev. 454, 460-61 (1943). The principal differences between these polar tests were that the “setting aside” test involved passing upon the credibility of witnesses and weighing contradictory evidence, while the “most favorable evidence” tesi did not; and the former test permitted consideration of all the evidence, while the latter did not. . That directed verdicts in antitrust cases have not been “generally disapproved” is evident from the fact that Santa Clara Valley, like Syufy, Janich Brothers, Cleary and Chisholm Brothers, affirmed a directed verdict in favor of the antitrust defendant. In addition, see Hanson v. Shell Oil Co., 541 F.2d 1352, 1357-59 (9th Cir. 1976), cert. denied, 429 U.S. 1074 , 97 S.Ct. 813 , 50 L.Ed.2d 792 (1977); Peltier v. Exxon Corp., 527 F.2d 757, 759 (9th Cir. 1975); Hallmark Industry v. Reynolds Metals Co., 489 F.2d 8, 11-13 (9th Cir. 1973), cert. denied, 417 U.S. 932 , 94 S.Ct. 2643 , 41 L.Ed.2d 235 (1974). As the Supreme Court has noted, “[t]he same rule” generally applicable in civil cases “governs in ruling upon motions for directed verdict in treble damage suits under the antitrust laws.” Continental Ore Co. v. Union Carbide & Carbon Corp., 370 U.S. 690 , 696 n.6, 82 S.Ct. 1404 , 1409 n.6, 8 L.Ed.2d 777 (1962). . Sherman Act § 2, 15 U.S.C. § 2 , provides: Every person who shall monopolize, or attempt to monopolize . . . any part of the trade or commerce among the several States, or with foreign nations, shall be deemed guilty of a felony . . “The Act was deemed by its sponsors not to be applicable to one ‘who merely by superior skill and intelligence . got the whole business because nobody could do it as well.’ [T]he exception was an indication that power obtained or maintained by the kind of behavior that competition is thought to foster, if not compel, was immune even though businesses and business opportunities were destroyed in the process. In short, . . protection of incentives to competitive behavior would prevail over dispersion of market power.” C. Kaysen & D. Turner, Antitrust Policy 20 (1959), quoting 21 Cong.Rec. 3151-52 (1890) (remarks of Sen. Hoar). . Thus, appellant’s contention that “any act taken by the monopolist satisfies the literal ‘monopolizes’ language of the statute” is overly broad. See Sunkist Growers, Inc. v. Winckler & Smith Citrus Products Co., 284 F.2d at 23 (after noting that jury instruction susceptible of similar interpretation was overly broad since under it, it was “difficult ... to see how a monopolist can make any business decision nor how it can make any sale or purchase (for example) at an advantageous price,” court held its use not reversible error in view of other instructions which “modified and tempered its broadness”). . Rossmoor involved claims under § 1 and both the monopolization and attempt to monopolize clauses of § 2. 487 F.2d at 376 . We held that whether the conduct requirement of each of these statutory proscriptions was met depended, under all three clauses, on the effect on “commercial competition: whether the defendant’s conduct . unreasonably restrains competition in the market place.” Id. See also FTC v. Motion Picture Advertising Service Co., 344 U.S. 392, 397 , 73 S.Ct. 361, 365 , 97 L.Ed. 426 (1953) (“when the Sherman Act is involved the crucial fact is the impact of the particular practice on competition”); Appalachian Coals, Inc. v. United States, 288 U.S. 344, 360 , 53 S.Ct. 471 , 77 L.Ed. 825 (1933) (both § 1 and § 2 “set up the essential standard of reasonableness”). . Conversely, not all violations of § 1 constitute acts of monopolization under § 2. The monopoly must be “acquired or maintained . by means of” the § 1 violation, United States v. Griffith, 334 U.S. 100, 106 , 68 S.Ct. 941 , 92 L.Ed. 1236 (1948) (emphasis added), “and not all violations of § 1 have the requisite causal effect.” 3 P. Areeda & D. Turner, Antitrust Law H 626, at 81 (1978). . Since “under the rule of reason market definition is required to establish a § 1 violation,” such evidence is necessary where specific intent is sought to be shown by inference unless a per se violation of § 1 is made out. Gough v. Rossmoor Corp., 585 F.2d at 390 ; see Chisholm Brothers Farm Equipment Co. v. International Harvester Co., 498 F.2d 1137, 1140 (9th Cir.), cert. denied, 419 U.S. 1023 , 95 S.Ct. 500 , 42 L.Ed.2d 298 (1974). . See generally Janich Bros., Inc. v. American Distilling Co., 570 F.2d at 854-58 ; Greyhound Computer Corp. v. IBM Corp., 559 F.2d at 505 ; Chisholm Brothers Farm Equipment Co. v. International Harvester Co., 498 F.2d at 1145 . Because monopoly power is a necessary element of the completed offense of § 2 monopolization but is not necessary for attempt, it follows that the same conduct may provide a basis for the monopolization offense and yet be excluded as a basis for the attempt offense. This is so since a defendant’s monopoly power — or lack thereof — is part of the “reasonableness” calculus; conduct reasonable for other firms is not necessarily reasonable for the monopolist. . Of course, proof of market power may be necessary to establish specific intent. See note 10 supra; Janich Bros., Inc. v. American Distilling Co., 570 F.2d at 854 n.4. . See note 11 supra. . 3 P. Areeda & D. Turner, Antitrust Law ¶ 828, at 321 (1978). As stated there, “[i]t would be perversely illogical to hold a defendant without [monopoly] power liable for conduct practiced with legal immunity by a monopolist.” . The data presented by both parties assume that the relevant geographic market is the United States. . CalComp also claims that prior to IBM’s price cuts on disk products, it offset losses on CPUs with “monopolistic profits” from its then high-priced disk equipment. Even if we were to accept that CalComp’s evidence was sufficient to go to a jury on the issue of whether such an antitrust violation occurred, a directed verdict would still be required because of the lack of any showing of causal injury. CalComp did not introduce substantial evidence (and indeed, would have had difficulty if it had attempted to do so) that it was “injured in [its] business or property by reason of’ IBM’s high prices on disk products. 15 U.S.C. § 15 . Cf. Greyhound Computer Corp. v. IBM Corp., 559 F.2d at 499 (“Greyhound can hardly complain of IBM’s increase in its rental rates for main frame equipment, since this change could only work to the advantage of leasing companies”). . CalComp quoted the chief financial officer of IBM’s Data Processing Group as saying, “[w]e have strategically put ourselves into a weak position as a result of IBM’s very high profit on disk products.” CalComp’s Chairman stated that IBM’s loss of disk products business “was heading straight toward 100%, like in two or three years,” and that IBM “had to react because they were losing so much business that no other choice was open to them. I wouldn’t call that a punitive reaction. I’d call it a defense of market position.” Another of CalComp’s witnesses, a securities analyst who followed the computer industry in general and IBM in particular, testified that IBM “was facing very strong competition in the marketplace” from plug-compatible manufacturers, and that there was “an acceleration in terms of migration to non-IBM equipment.” The same witness had earlier described IBM’s disk price cuts in one of his reports as “dropping the traditional price umbrella for drive manufacturers.” Still another CalComp witness, a former IBM employee, testified that if IBM had not responded to the “accelerating” competition for peripheral equipment sales, it would have lost “most or all” of this business. . Other evidence indicated that by reducing the price of the system, the 2319A also helped IBM in competing with systems, manufacturers such as RCA. . IBM’s evidence was that the profit on each of its disk products was greater than 20% of revenue before taxes, and that the 2319A in particular was expected to yield a profit of 33%. In its amicus brief, Memorex Corp., another peripheral products manufacturer, contended that IBM’s cost accounting system is too “inexact” to rely on for such determinations of profitability. Its argument is that by allocating the fixed cost burden for a particular product in part according to the product’s revenues, “costs” are artificially reduced when a product’s price is cut, and hence profitability is artificially increased. The evidence at trial, however, supports the single conclusion that IBM indeed fairly expected significant profits from its disk products even after price reductions. In response to IBM’s price cuts, Cal-Comp and other manufacturers reduced their prices again below IBM’s, with the effect that, according to CalComp, they “received much smaller profit margins” as opposed to incurring losses. CalComp’s own evidence established that prior to its price cuts, IBM’s projected profits on its disk products ranged from 47% to 58%. In addition, IBM’s evidence on profitability that Memorex challenges consisted of its internal calculations relied on at the time the decision to go forward with each product was made. As one of CalComp’s experts testified, profit expectations at announcement or introduction of a product at a particular price are the primary consideration in analyzing a firm’s pricing behavior. Finally, there is absolutely no evidence that IBM priced its disk products below marginal cost, which ordinarily is required to show predatory pricing as a means of obtaining or maintaining a monopoly in violation of § 2. See Janich Brothers, Inc. v. American Distilling Co., 570 F.2d at 855-58 , and text infra. . CalComp argues on appeal that IBM’s failure to cut prices on existing disk products for use with System 360 CPUs to the level of the 2319B constituted “discrimination.” Since the 2319B “retread” was available to all IBM System 360 users, however, this fact is of little consequence. . See note 17 supra. . See note 19 supra. IBM’s evidence was that the 2319B was expected to return a profit before taxes equal to 32% of revenues. . According to CalComp’s Chairman, Cal-Comp, like other plug-compatible manufacturers, generally had to offer discounts from the systems manufacturers’ prices in order to convince users to take their copy instead of the original. . CalComp introduced evidence of IBM studies reviewed by IBM management shortly before the announcement of FTP that included an analysis of the extent to which two of IBM’s competitors, Telex and Memorex, could afford to increase manufacturing and sales volume and reduce prices on disk products. In particular, the studies concluded that Memorex would remain viable despite IBM’s introduction of its 2319 disk drives, and that by 1976 IBM would lose 28.8% of the plug compatible disk market. IBM’s financial group also reported, however, that if Memorex achieved 60% of the plug-compatible disk market, such rapid growth in marketing overhead and working capital requirements would require substantial additional financing — particularly if Memorex met IBM price cuts in the area of 17%-22%. The obvious concern in these reports to management was the increasing gains by peripheral manufacturers and the lengths to which price competition by Memorex in particular might be carried. Accepting CalComp’s inferences from this evidence, FTP was a direct response to these competitor gains, and the level of FTP price cuts was determined by IBM’s estimation of the extent of its competitors’ willingness to lower prices still further. . CalComp’s argument that ISC and IFA, by placing part of the control function “inboard,” were designed solely to preclude plug-compatible manufacturers from part of the controller market is discussed under the heading “Design Changes” infra. . CalComp attributed its price reductions and consequent lost revenues solely to the 2319B and FTP price reductions. We need not discuss the resultant question of causation with respect to IBM’s other price cuts in view of our result. . “Marginal cost” refers to the variable cost of the last unit of output, while “average varia ble cost,” appropriately, refers to the average variable cost of all units. See Janich Brothers, 570 F.2d at 857 n.8, 858 n.11. Hanson determined that “because marginal cost is often impossible to ascertain” average variable cost was a sufficient judicial surrogate, 541 F.2d at 1358 ; Janich Brothers concurred that “[ajverage variable cost is likely to approximate marginal cost.” 570 F.2d at 858 . Given the variation in cost accounting systems among business firms and the likelihood of the significant differences in capacity utilization from case to case, we expect that acceptance of this rule of thumb would first require careful scrutiny of the particular facts in a case where the defendant’s prices are very near average variable cost. That is not the case here. . Unlike the 2319A, the 2319B — introduced for use with all System 360 models — attached to IBM’s 2314 stand-alone controller. . The System 370 Models 158 and 168 also featured substantial minimum memory integrated within the CPU. CalComp claimed that this redesign injured add-on memory manufacturers by making it impossible for them to compete for the integrated portion of these models’ memories. . This same witness reported that the inclusion of a minimum memory capacity within the CPU was made possible by the reduction in memory size, “reflecting the product evolution from relatively large magnetic cores to high density semi-conductor packaging.” He continued: As the basic cycle speed of the memory increases, it becomes important to minimize the distance between components. It may well be impractical to keep the new memories external to the processor . . In Telex Corp. v. IBM Corp., 367 F.Supp. 258, 304 (N.D.Okl.1973), rev’d on other grounds, 510 F.2d 894 (10th Cir. 1975), dealing with a different record developed from the same underlying facts, the district court found with respect to IFA and ISC that there had been a “preponderant showing that these integrations represented a legitimate technological and performance advance consistent with trends in the industry and at significant decreased cost.” No appeal was taken from this ruling. . In its Telex decision, the Tenth Circuit held that IBM’s conduct in this regard was not violative of § 2, focusing on the profitability of both the CPU and FTP prices rather than on IBM’s possible motivation. Regardless of intent, the court implied, § 2 condemns only the monopolist’s “use of monopoly power,” and the facts in Telex showed that IBM’s lowering peripheral prices and raising CPU prices was not a case of “an economic giant [dropping] prices to a level where it is not receiving an adequate return and . . instead [relying] on its reserves or other activities to continue producing and marketing the particular product.” See 510 F.2d at 925-26.