MEMORANDUM OF OPINION AND ORDER

Plaintiffs Red Lion Medical Safety, Inc. et al. bring this antitrust action against defendant Ohmeda, Inc. Ohmeda moves for summary judgment.

I.

Ohmeda, Inc. is one of the nation’s leading manufacturers of medical anesthesia equipment. Plaintiffs are independent service organizations (“ISOs”) who service anesthesia equipment, including Ohmeda machines. Plaintiffs accuse Ohmeda of trying to exclude them from the servicing market for Ohmeda equipment.

Until recently, when it was partitioned and sold, Ohmeda was the' health care division of The BOC Group, a British conglomerate. (See Willig Decl. at 7.) Ohme-da manufactures and sells anesthesia systems. (See Brandmeier Decl., ¶ 8.) An anesthesia system consists of (1) a gas delivery platform, including vaporizers and ventilators, that mixes gases with an anesthetic agent and induces the mixture into the patient; (2) respiratory gas monitors that track the presence of gases in the patient; and (3) physiological monitors that track the patient’s vital signs, such as heart rate. (See id., ¶ 9.) Ohmeda provides service for its anesthesia systems and also manufactures and sells replacement parts for its machines.

Ohmeda classifies the parts it manufactures for its hardware into three categories. (See Grosse Deck, ¶¶ 19-21.) Some parts are “restricted” and are not for sale by themselves; instead, these parts are incorporated into larger “subassemblies” and sold in that form. (See id., ¶ 19.) Other parts are “unrestricted” and may be sold to anyone. Ohmeda asserts that “the vast majority of [its] repair and maintenance parts and subassemblies” fall within the unrestricted category. (See id., ¶ 20.) Finally, about 12% of Ohmeda parts are “service restricted.” (See id., ¶ 24.) This category includes parts “which require installation, replacement or adjustment by trained personnel using specialized procedures to help maintain patient safety and proper operation of the equipment.” (Id., Exh. A.)

From at least 1984 to mid-1997, Ohme-da refused to sell service restricted parts or provide any training or manuals to ISOs. (See Lewitzke Decl., ¶ 9.) If a hospital did not use Ohmeda for service, Ohme-da would only sell the hospital service restricted parts in two situations: first, if the hospital had its own in-house, Ohmeda-trained biomedical technicians to do the servicing, or second, if the hospital signed a waiver letter absolving Ohmeda of liability arising out of the use of the machine, including liability unrelated to service by the ISO. (See id.; Burke Deck, ¶ 22.) Plaintiffs allege that Ohmeda maintained these restrictions on the sales of replacement parts in order to coerce buyers of its equipment into using Ohmeda service rather than ISO service.

Ohmeda’s restrictive parts policy is the heart of plaintiffs’ claim that Ohmeda has sought to dominate the servicing market in violation of the antitrust laws. In addition to their attack on the availability of Ohme-da parts, plaintiffs make a number of other related factual allegations: they assert that Ohmeda engaged in price discrimination by selling identical parts at different prices under separate part numbers, (see Ardrey Deck, ¶ 12); levied hidden price increases on its customers by decreasing the amount of service provided while keeping prices constant, (see Lemanek Deck, ¶ 9; Lewis Deck, ¶ 14); struck back at customers who used ISOs by making unnecessary repairs, (see Lesko Deck, ¶¶ 24-25), and even sabotaging its own machines, (see Burke Deck, ¶ 31; Garrett Deck, ¶ 21); and generally disparaged the quality of ISO service, (see Lemanek Deck, ¶ 6; Les-ko Deck, ¶ 9; Lewis Deck, ¶ 28). Plaintiffs also assert that Ohmeda forbade independent original equipment manufacturers (“OEMs”) of Ohmeda parts from selling service restricted parts to ISOs. (See Foster Deck, ¶ 16; Garrett Deck, ¶ 18.)

Plaintiffs filed this lawsuit on November 1, 1996. In June 1997, Ohmeda changed its parts and servicing policy and established the Qualified Independent Service Organization (“QISO”) program. Under this program, Ohmeda offers the same service training to ISOs that it requires of its own service technicians and hospital biomedical technicians. (See Willig Deck at 60; Yeager Deck, ¶¶ 3-4.) Any ISO with a technician who has passed the Ohmeda training course becomes eligible to order manuals and parts — even service restricted parts — directly from Ohmeda for any equipment on which the technician has been trained. (See Willig Deck at 60-61.)

According to plaintiffs, however, Ohme-da’s QISO training is infrequently given, prohibitively expensive, and unhelpful to most experienced technicians. Plaintiffs assert that it costs approximately $40,000 to train a single technician on the full line of Ohmeda equipment, (see Ardrey Deck, ¶ 24; Coholan Deck, ¶ 23; L. McBride Deck, ¶ 24), not including lost income and travel expenses. Plaintiffs allege that the classes are given only a few times a year and are often full. (See Coholan Deck, ¶ 25; Foster Deck, ¶ 23). As a result, it may take “up to three years” to fully train a single technician. (Id.) Even after an ISO technician completes QISO training, the technician is required to take periodic refresher courses, a requirement Ohmeda does not impose on its service employees. (See id.) Moreover, plaintiffs contend that if a trained technician leaves one ISO for another, Ohmeda requires that the technician be retrained. (See Coholan Deck, ¶ 23; L. McBride Deck, ¶ 24.) Finally, plaintiffs allege that Ohmeda is free to cancel an ISO’s QISO status at any time without justification or penalty. (See Co-holan Deck, ¶ 26; Foster Deck, ¶ 24.)

Based on the above policies and practices, plaintiffs make eight antitrust claims against Ohmeda: (1) monopolization of the market for service on all anesthesia equipment in violation of § 2 of the Sherman Antitrust Act; (2) monopolization of the market for service on Ohmeda equipment in violation of § 2; (3) attempted monopolization of the market for service on all anesthesia equipment under § 2; (4) attempted monopolization of the market for service on Ohmeda equipment under § 2; (5) per se illegal tying of service to parts under § 1 of the Sherman Act; (6) illegal tying of service to parts under the rule of reason in violation of § 1; (7) a per se illegal group boycott under § 1; and (8) an illegal group boycott under the rule of reason in violation of § 1. Ohmeda now moves for summary judgment on all claims.

II.

Ohmeda first argues that plaintiffs’ claims are time-barred. Under the Clayton Act, 15 U.S.C. § 15b, private antitrust claims are subject to a four-year statute of limitations. See Hennegan v. Pacifico Creative Serv., Inc., 787 F.2d 1299 , 1300 (9th Cir.1986). “A civil cause of action under the [antitrust laws] arises at each time the plaintiffs interest is invaded to his damage, and the statute of limitations begins to run at that time.” Id. (quoting AMF, Inc. v. General Motors Corp., 591 F.2d 68, 70 (9th Cir.1979)). Plaintiffs filed this suit on November 1, 1996, so any antitrust claim that accrued before November 1, 1992 is untimely. It is undisputed that Ohmeda’s restrictive parts policy was put into effect no later than 1984, eight years before the start of the limitations period. (See Lewitzke Decl., ¶¶ 8-9.) Plaintiffs’ industry expert, Michael Brink-man, conjectures that the Ohmeda policy dates from the late 1970s. (See Brinkman Depo. at 28.) All but one of the plaintiffs had been in business for more than four years at the time the suit was filed.

In response, plaintiffs argue that the continuing violations doctrine saves their claims. In antitrust law, “[a] continuing violation is one in which the plaintiffs interests are repeatedly invaded and a cause of action arises each time the plaintiff is injured.” Pace Industries, Inc. v. Three Phoenix Co., 813 F.2d 234, 237 (9th Cir.1987) (citing Hennegan, 787 F.2d at 1300-01). When a continuing violation exists, the limitations period runs from the “last overt act” by the defendant, which must be (1) a new and independent act that is not merely a reaffirmation of a previous decision that (2) inflicts “new and accumulating” injury on the plaintiff. Id. at 238.

“New and independent acts” may include active enforcement of policies first put into place outside the limitations period. In Columbia Steel Casting Co. v. Portland General Elec. Co., 111 F.3d 1427 (9th Cir.1996), a power company refused to sell power to the plaintiff under an 18-year-old horizontal market share division agreement. The Ninth Circuit held that the refusal was a new and independent act because the original agreement “was not a permanent and final decision that controlled the later act.” 111 F.3d at 1444. Similarly, in Hennegan, tour operators diverted customers from plaintiffs’ souvenir shop. Although the tour operators had originally agreed to do so more than four years before plaintiffs filed suit, the Ninth Circuit held that the original agreement “did not immediately and permanently destroy the Hennegans’ business, nor [was it] ‘irrevocable, immutable, permanent and final.’ ” 787 F.2d at 1301. New and accumulating injury was inflicted upon the plaintiffs each time a defendant diverted customers from plaintiffs’ business. See also Airweld, Inc. v. Airco, Inc., 742 F.2d 1184, 1190 (9th Cir.1984) (holding that active enforcement of an illegal tie is an independent act); Aurora Enterprises, Inc. v. National Broadcasting Co., Inc., 688 F.2d 689, 694 (9th Cir.1982) (holding that active enforcement of the terms of an illegal contract is a new and independent act).

In contrast to Columbia Steel and Hen-negan, the Ninth Circuit has found no continuing violation only in situations in which the act outside the limitations period “completely and permanently excluded [the plaintiff] from the market.” Hennegan, 787 F.2d at 1301. In AMF, 591 F.2d 68 , four automobile manufacturers collectively refused to buy afterburners from the plaintiff outside the limitations period, deciding instead to manufacture their own afterburners. The AMF court reasoned that because the defendants required considerable lead time to integrate afterburners into new cars, the market for plaintiffs new car afterburners “effectively disappeared” then and for the foreseeable future at the point defendants decided to develop their own afterburners. In practice, the decision not to buy from plaintiff was “irrevocable, immutable, permanent, and final.” 591 F.2d at 72 . Because defendants’ decision had the effect of permanently excluding plaintiff from the afterburner market, plaintiffs injury was complete at that time, and it suffered no additional harm within the limitations period. See id.

This case is more like Columbia Steel and Hennegan than AMF. Although Ohmeda’s parts policy has been in place for at least 15 years, the policy is not a “permanent and final” decision that forever more compels Ohmeda to refuse to sell service restricted parts to ISOs. Ohmeda could always change its policy, as indeed it did in 1997 by establishing the QISO program, or make an exception for a particular ISO or a particular part. Unlike AMF, there is no technological impediment or necessary lead timé associated with an amendment to the service restricted parts policy. Moreover, unlike the auto manufacturers’ decision in AMF, the mere adoption of the parts policy did not “immediately and permanently destroy” plaintiffs’ businesses or “completely and permanently exclude” plaintiffs from the relevant market. Hennegan, 787 F.2d at 1301. Instead, Ohmeda’s policy incrementally limits plaintiffs’ ability to expand their businesses. Thus, the effects of the policy are felt not all at once and for the foreseeable future, as was the case in AMF, but each time an ISO is unable to sign a hospital customer to a service contract because it cannot readily obtain Ohmeda parts.

Because Ohmeda’s enforcement of its parts policy allegedly has caused new injury to plaintiffs within the four-year limitations period, plaintiffs’ claims are not time-barred.

III.

Plaintiffs claim that Ohmeda has monopolized the market for service on all anesthesia equipment, or, alternatively, on Ohmeda anesthesia equipment in violation of § 2 of the Sherman Act. A monopoly-claim under § 2 has two elements: “(1) the possession of monopoly power in the relevant market and (2) the willful acquisition or maintenance of that power as distinguished from growth or development as a consequence of a superior product, business acumen, or historic accident.” Eastman Kodak Co. v. Image Technical Services, Inc., 504 U.S. 451, 481 , 112 S.Ct. 2072, 2089 , 119 L.Ed.2d 265 (1992) (hereinafter Kodak I) (quoting United States v. Grinnell Corp., 384 U.S. 563, 570-71 , 86 S.Ct. 1698, 1704 , 16 L.Ed.2d 778 (1966)).

A. Monopoly Power in the Relevant Market.

Proof of monopoly power, which the Supreme Court has defined as “the power to control prices or exclude competition,” see Grinnell, 384 U.S. at 571 , 86 S.Ct. at 1704 (quoting United States v. E.I. du Pont de Nemours & Co., 351 U.S. 377, 391 , 76 S.Ct. 994, 1005 , 100 L.Ed. 1264 (1956)), may be by direct or circumstantial evidence. See Image Technical Services, Inc. v. Eastman Kodak Co., 125 F.3d 1195, 1202 (9th Cir.1997) (hereinafter Kodak II). Direct evidence of monopoly power is “evidence of restricted output and supracompetitive prices.” Rebel Oil Co., Inc. v. Atlantic Richfield Co., 51 F.3d 1421, 1434 (9th Cir.1995) (citing Federal Trade Comm’n v. Indiana Fed’n of Dentists, 476 U.S. 447, 460-61 , 106 S.Ct. 2009, 2018-19 , 90 L.Ed.2d 445 (1986)). To demonstrate monopoly power by circumstantial evidence, “a plaintiff must: (1) define the relevant market, (2) show that the defendant owns a dominant share of that market, and (3) show that there are significant barriers to entry and show that existing competitors lack the capacity to increase their output in the short run.” Kodak II, 125 F.3d at 1202 (quoting Rebel Oil, 51 F.3d at 1434 ). Plaintiffs rely on circumstantial evidence to establish their monopoly power claim.

In Kodak I, the Supreme Court addressed the antitrust implications of “aftermarkets,” derivative markets for replacement parts and servicing. The Kodak plaintiffs were ISOs who serviced Kodak photocopiers and micro-graphics equipment. They sued Kodak for monopolizing the market for service on Kodak equipment by restricting access to replacement parts. See Kodak I, 504 U.S. at 456 , 112 S.Ct. at 2076 . The Court held that as an abstract matter of law and economics, a manufacturer could have monopoly power in the servicing of its own equipment even if it had no such power in the sale of that equipment, and further, that a manufacturer could achieve monopoly power in servicing by tying service to parts. See id. at 481-82 , 112 S.Ct. at 2089-90 ; see also Shapiro, “Aftermarkets and Consumer Welfare: Making Sense of Kodak,” 63 Antitrust L.J. 483, 483-84 (1995). Two related principles follow from the Kodak analysis: first, parts and service can constitute separate markets, see Kodak I, 504 U.S. at 462-63, 481-82 , 112 S.Ct. at 2079-80, 2090 , and second, a single brand, by itself, can constitute a separate market for parts or service, see id. at 482 , 112 S.Ct. at 2090 .

In making its ruling, the Court reaffirmed that “[t]he proper market definition ... can be determined only after a factual inquiry into the ‘commercial realities’ faced by consumers.” Kodak I, 504 U.S. at 482 , 112 S.Ct. at 2090 . The crucial inquiry in determining whether related products— such as equipment and service — constitute one or multiple markets is whether there is “sufficient consumer demand so that it is efficient for a firm to provide” one separately from the others. Id. at 462 , 112 S.Ct. at 2080 (citing Jefferson Parish Hospital Dist. No. 2 v. Hyde, 466 U.S. 2, 21-22 , 104 S.Ct. 1551, 1563 , 80 L.Ed.2d 2 (1984)).

1. The Relevant Market

Plaintiffs argue that the relevant market is for service of all anesthesia gas platforms, excluding monitors, or, alternatively, of Ohmeda gas platforms. Ohmeda contends that the relevant market is much broader, encompassing sales of gas systems, including monitors, as well as post-sale service. (See Willig Deck at 11-12.) Ohmeda argues that there is no separate and distinct market for service of anesthesia gas systems, apart from sales, and thus that service cannot be a “relevant market” for antitrust purposes.

The Ninth Circuit has defined the “relevant market” as the group of sellers or producers who have the “actual or potential ability to deprive each other of significant levels of business.” Id. at 1434 (quoting Thurman Indus., Inc. v. Pay ‘N Pak Stores, Inc., 875 F.2d 1369, 1374 (9th Cir.1989)). Market definition thus “focuse[s] on ... whether consumers view the products as substitutes for each other.” Id. at 1435 (citing Sullivan & Harrison, Understanding Antitrust and Its Economic Implications § 6.04[2] (1988)). “If consumers view the products as substitutes, the products are part of the same market.” Id.

Plaintiffs have introduced sufficient evidence to create a triable issue as to whether service is a separate and distinct product from sales of anesthesia equipment and parts. Plaintiffs do not sell gas machines and monitors themselves, and have only recently begun to sell parts in conjunction with Ohmeda’s QISO program. There is plainly enough consumer demand so that it is efficient for competitors such as the plaintiffs here to provide service without selling equipment or parts. See Kodak I, 504 U.S. at 462 , 112 S.Ct. at 2080 (noting that “service and parts have been sold separately in the past”). Ohmeda’s claim that hospitals view the equipment and service markets as one because hospitals engage in life cycle pricing is placed in dispute by plaintiffs’ evidence to the contrary.

Moreover, contrary to Ohmeda’s assertion, that consumers may often buy equipment and service as a package does not necessarily combine those products into a single market. The Supreme Court held in Kodak I that parts and service could constitute separate markets even though “there is no demand for parts separate from service.” Kodak I, 504 U.S. at 463 , 112 S.Ct. at 2080 . Otherwise, “we would be forced to conclude that there could never be separate markets, for example, for cameras and film, computers and software, or automobiles and tires.” Id. And in tying cases under § 1 of the Sherman Act, the Court has “often found arrangements involving functionally linked products at least one of which is useless without the other to be prohibited tying devices.” Jefferson Parish, 466 U.S. at 19 n. 30, 104 S.Ct. at 1562 n. 30.

Plaintiffs have also created a triable issue as to whether the relevant market is service of Ohmeda machines rather than anesthesia systems in general, including those manufactured by other companies. In Kodak I, the Court held that the relevant market was service of Kodak copiers and not all copiers:

504 U.S. at 481-82 , 112 S.Ct. at 2090 . As was the case in Kodak I, service and parts for Ohmeda equipment are not interchangeable with other manufacturers’ service and parts; presumably this is the reason Ohmeda insists on specialized training through the QISO program. In sum, the jury may find that from the perspective of an owner of Ohmeda equipment the relevant market for service consists only of organizations that service Ohmeda machines. If so, service of Ohmeda machines could constitute a relevant market for antitrust analysis.

Finally, the parties dispute whether service on the monitors attached to the gas delivery platform should be included as part of the relevant service market. There seems to be evidence pointing either way on this issue. On the one hand, all ISOs who service Ohmeda platforms also service the monitors attached to the platforms. Ohmeda also services both its gas platforms and its monitors. On the other hand, monitors are often sold separately from gas machines, (see Leath Decl., ¶ 8; Bernick Deck, ¶ 5; Boldan Deck, ¶ 5), and are normally upgraded more frequently than machines, (see Leath Deck, ¶ 8). Most Ohmeda gas machines are designed specifically to function with any monitor, not just monitors manufactured by Ohme-da. (See id.) Ohmeda gas machine owners do not always purchase Ohmeda monitors. (See id.). Finally, many monitor manufacturers, such as Hewlett Packard and SpaceLabs, do not make gas machines. It is a reasonable inference that there are companies who service monitors but not gas platforms.

For these reasons, plaintiffs' have created a triable issue as to whether the relevant antitrust market here is the market for service on Ohmeda anesthesia gas machines whether with or without monitors.

2. Market Power: Dominant Share of the Market

Next, plaintiffs must show that a jury could find that Ohmeda owns a dominant share of the relevant market for service on Ohmeda gas platforms. See Kodak II, 125 F.3d at 1206 (citing Rebel Oil, 51 F.3d at 1434 ). “A dominant share often carries with it the power to control output across the market, and thereby control prices. Courts generally require a 65% market share to establish a prima facie case of market power.” Id. (citing American Tobacco Co. v. United States, 328 U.S. 781, 797 , 66 S.Ct. 1125, 1133 , 90 L.Ed. 1575 (1946) (internal citation omitted)).

Plaintiffs rely on three kinds of evidence to establish Ohmeda’s share of the market for service of Ohmeda machines. First, they cite a 1994 survey which reported that 69% of hospitals that own Ohmeda machines contracted with Ohmeda for service. (See Pis’ Exh. 469.) As plaintiffs concede, (see Opp’n at 45 n. 220), this figure is not as impressive as it appears because some customers used more than one service provider. Ohmeda’s actual market share for all available accounts according to this document is 58.06%. (See id.) Second, plaintiffs cite to internal Ohmeda reports which state that “of the 24,000 Ohmeda machines in use, 17,000 are on Ohmeda maintenance contracts,” (Opp’n at 45 (citing Pis’ Exh. 141)), which works out to a market share of 70.83%. Third, they offer the testimony of various former Ohmeda employees as to Ohmeda’s market share in regional service markets. (See, e.g., Lewis Deck, ¶ 22 (“Our service market share on Ohmeda machines was nearly 100%”).)

Ohmeda disputes the accuracy of its internal market share data, stating that the figures were casually compiled through broad estimates. (See Spence Depo. at 394, 405-06; see also Brandmeier Depo. at 104-05, 174-75; Brandmeier Decl., ¶ 27-30.) Instead, Ohmeda offers three reports — one compiled by an outside source — showing that Ohmeda has achieved only between 40% and 48% of the market for service on all anesthesia equipment. (See Stoll Deck, ¶ 32 & Exh. 8 (presentation by president of Ohmeda to potential purchasers describing 40% market share); Stoll Deck, ¶ 26 & Exh. 3 (Bain & Company report commissioned by Ohmeda showing 48% market share); Wil-lig Deck at 78 & Exh. 13 (hospital survey prepared for Dr. Willig showing 47% share).) Ohmeda has not offered any evidence on its share of the market for service on Ohmeda machines alone.

Ohmeda’s market share figures for servicing on all machines are actually consistent with plaintiffs’ data concerning servicing on Ohmeda systems. Because Ohmeda only performs service on Ohmeda machines, its share of the market for service on those machines will equal its share of the total service market divided by its share of the equipment market. Ohme-da’s share of the equipment sales market is in dispute, but plaintiffs offer evidence that Ohmeda controls between 61%, (see Pis’ Exh. 424 (1996 Booz-Allen and Hamilton study)), and 65%, (see Millstein Deck, Exh. A (1993 Frost & Sullivan report)), of the market. Ohmeda’s own evidence, gathered in more informal surveys, shows a market share somewhere between 46%, (see Pis’ Exh. 210 (blind survey conducted at 1992 American Society for Anesthesiology convention)), and “about” 50%, 25% (Willig Deck at 84 & Exh. 11 (1997 Willig telephone survey)).

Adopting a provisional figure of 55% for Ohmeda’s share of the equipment market — roughly halfway between plaintiffs’ top figure of 65% and Ohmeda’s bottom figure of 46% — plaintiffs’ claim that Ohme-da possesses 71% of the service market for Ohmeda machines is consistent with Ohmeda’s assertion that it controls 40% of the service market for all anesthesia machines. In light of this evidence, plaintiffs’ claim that Ohmeda possesses a 71% share of the service market on its own machines, and thus controls a “dominant share of the market” under Kodak II, is triable.

Despite this market share evidence, Ohmeda insists that an examination of market realities reveals that it has no mar ket power. First, it asserts that it is losing money on its service business. (See Bourgart Decl., ¶ 37 (showing 6.1% loss for fiscal 1997).) By itself, the fact that an accused monopolist is operating at a loss is not dispositive; an aspiring monopolist might accept short-run losses as the price of eliminating competition in a given market. Ohmeda also claims that it has not raised its prices appreciably in recent years, (see Shrago Deck, ¶ 30), behavior it asserts is inconsistent with monopoly power. Plaintiffs, however, may prove market power through circumstantial evidence and need not submit direct evidence of supra-competitive prices. See Kodak II, 125 F.3d at 1202 . Moreover, plaintiffs have introduced evidence suggesting that Ohmeda has levied hidden price increases on its customers by keeping its prices constant while reducing the amount of service its technicians actually provide. (See Lemanek Deck, ¶ 9; Lewis Deck, ¶ 14.) A triable issue thus exists as to whether and to what extent Ohmeda has increased its prices.

Although plaintiffs have come forward with sufficient evidence as to whether Ohmeda possesses a monopoly share of the service market for its own machines, they have not shown that Ohmeda possesses a dominant share of the market for service of anesthesia machines in general. Even if plaintiffs’ most favorable market share data is accurate, Ohmeda possesses only 71% of the market for service on its own machines, and only 65% of the equipment sales market. When the service-market percentage is multiplied by the overall equipment percentage, Ohmeda’s share of the overall service market is only 46.15%, well short of the 65% threshold announced in Kodak II. Plaintiffs’ monopolization claim can thus succeed only if the trier of fact accepts plaintiffs’ position that the relevant market is the market for service on Ohmeda machines.

S. Market Power: the Kodak Aftermarket Issues

Ohmeda asserts the same defenses to the allegation of market power offered by Kodak in Kodak I: Ohmeda argues that even if it has monopoly power in the service market, it cannot exercise that power because of competition in the equipment market. Ohmeda contends that if it were to flex its muscles in the service market by raising its service prices above a competitive level, any resulting increase in profits would be more than offset by losses in the equipment market as customers switched to equipment with lower service costs. In Kodak I, the Supreme Court accepted the plausibility of this general principle, noting that the “extent to which one market prevents exploitation of another market depends on the extent to which consumers will change their consumption of one product in response to a price change in another, i.e., the ‘cross-elasticity of demand.’ ” Kodak I, 504 U.S. at 469 , 112 S.Ct. at 2083 (citing United States v. E.I. du Pont de Nemours & Co., 351 U.S. at 400 , 76 S.Ct. at 1010 , and 2 Areeda & Kaplow, Antitrust Analysis ¶ 342(c) (1988)).

The Kodak I ISOs offered two possible and related reasons why equipment sales and service prices might not be “cross-elastic,” and thus why equipment sales might not respond to an increase in service prices, as apparently they had not. First, consumers looking to purchase equipment might not have enough information or sophistication to engage in “life cycle pricing,” the estimation of the total cost of a piece of equipment over the life of the machine, including parts and service. See id. at 473 , 112 S.Ct. at 2085 . Second, once the consumer has purchased the equipment, the cost of switching to another brand of equipment in response to high servicing charges might be prohibitive. See id. at 476 , 112 S.Ct. at 2087 . Consumers who cannot readily switch to another brand — whether because of the high capital outlay or for other reasons — are “locked in” and “will tolerate some level of service-price increases.” Id. “Under this scenario, a seller profitably could maintain supracompetitive prices in the aftermarket if the switching costs were high relative to the increase in service prices, and the number of locked-in customers were high relative to the number of new purchasers.” Id.

Several circuit courts have found in Kodak I an implicit limitation on aftermarket antitrust claims to situations involving a change of policy or pricing as to after market parts and services. In PSI Repair Servs., Inc. v. Honeywell, Inc., 104 F.3d 811 (6th Cir.1997), the Sixth Circuit held that “an antitrust plaintiff cannot succeed on [an aftermarket] theory when the defendant has not changed its policy after locking in some of its customers, and the defendant has been otherwise forthcoming about its pricing structure and service policies.” 104 F.3d at 820 ; see also Digital Equip. Corp. v. Uniq Digital Techs., Inc., 73 F.3d 756, 762-63 (7th Cir.1996); Lee v. Life Ins. Co. of N. Am., 23 F.3d 14, 20 (1st Cir.1994). Ohmeda asserts that, unlike Kodak, it did not surprise its equipment customers by changing its parts policy or raising its servicing prices. It contends that it has not raised its prices appreciably in recent years, (see Shrago Decl., ¶ 30), its parts policy remained constant between 1984 and 1997, (see Lewitzke Deck, ¶¶ 8-9), and the policy was public and widely-known by hospital customers, (see Taylor Deck, ¶ 12).

The Honeywell court’s interpretation of Kodak I, however, is not supported by the text or reasoning of that opinion. Kodak I does not hold that an aftermarket claim is contingent on a change in a manufacturer’s parts or service policy; it simply acknowledges that Kodak’s ability to make a policy change without suffering losses in the equipment market was evidence that the service market was not disciplined by competition in the equipment market. See Kodak I, 504 U.S. at 477 , 112 S.Ct. at 2087-88 (“It is clearly reasonable to infer that Kodak has market power to raise prices and drive out competition in the aftermarkets, since respondents offer direct evidence that Kodak did so.”).

The Honeywell opinion assumes that information costs “were particularly high in Kodak because Kodak adopted its parts-restrictive policy after numerous customers had already purchased Kodak copiers, thus creating a ‘lock-in’ effect.” Honeywell, 104 F.3d at 818 . But this has it backwards. The policy change did not create lock-in; instead, the existence of lock-in — high switching costs — made it both possible and economically desirable for Kodak to change its policy and exploit aftermarket consumers. See Kodak I, 504 U.S. at 476-77 , 112 S.Ct. at 2087 . Thus, the policy change did not create monopoly power; it was merely persuasive evidence that Kodak had market power in parts and engaged in monopolistic conduct in the aftermarket despite competition in the equipment market. To insist on a showing of a policy change confuses a symptom of market power and a lack of cross elasticity with the underlying condition itself.

It is noteworthy that in the Kodak I Court’s lengthy discussion of information costs it never once makes reference to Kodak’s change in policy. Information costs may be high, and a manufacturer may thus have considerable market power in the aftermarket, even in the absence of a change in policy. As the Kodak I Court pointed out,

Kodak I, 504 U.S. at 474-75 , 112 S.Ct. at 2086 .

Moreover, to the extent that the Ninth Circuit has considered the Honeywell aftermarket theory, it appears to have re-jeeted it. In Datagate, Inc. v. Hewlett Packard Co., 60 F.3d 1421 (9th Cir.1995), Hewlett Packard allegedly refused to provide software support to its customers unless they agreed also to buy hardware service. Various ISOs sued, alleging an illegal tie. The Ninth Circuit never examined whether competition in the equipment market might discipline the service aftermarkets. Neither party offered evidence that Hewlett Packard had recently changed its software service policy, and the court never indicated that a policy change was the crucial element of an aftermarket claim. Instead, in the marked absence of such evidence, the court held that Hewlett Packard’s conduct could amount to an illegal tie. See id. at 1426 . And when Kodak returned to the Ninth Circuit on remand, the Ninth Circuit did not identify Kodak’s policy change as an essential element of the plaintiffs’ aftermarket claim.

Thus, the mere fact that Ohmeda has not recently changed its parts policy does not entitle it to summary judgment.

Applying Kodak I to the facts here, plaintiffs have presented sufficient evidence of information costs and “lock-in” to defeat Ohmeda’s contention that competition in the equipment market restrains any market power that it might have in the aftermarket. Plaintiffs offer evidence that consumers in the anesthesia equipment market do not engage in lifecycle pricing. {See Lemanek Decl., ¶ 19; Naylor Decl., ¶ 19; Perkins Deck, ¶ 8.) There is no dispute that service costs are small compared to the cost of a new anesthesia machine, which reduces a purchaser’s incentive to collect life cycle information. And plaintiffs offer an additional reason why hospitals might not collect life cycle information, or might even buy Ohmeda machines despite high service costs: most gas machine users are anesthesiologists, who generally prefer to use the same brand of equipment on which they were trained. {See Leath Deck, ¶ 6; Lesko Deck, ¶ 19; Perkins Deck, ¶ 9.) A hospital whose purchasing anesthesiologist was trained on Ohmeda equipment might not gather life cycle price information because the convenience and comfort of the familiar Ohmeda machine is more important than some measure of service market savings. Finally, plaintiffs contend that Ohmeda did not inform its equipment customers of its restrictive parts policy prior to sales of equipment. {See Lesko Deck, ¶ 16; Naylor Deck, ¶ 21.)

The disparity in cost between anesthesia machines and service on those machines also increases the likelihood of lock-in. Anesthesia machines cost anywhere from $35,000 to $100,000, while a year’s preventive maintenance contract costs approximately $940. {See J. McBride Deck, ¶ 15). A hospital is unlikely to switch brands and incur thousands of dollars in capital expenses in order to save $20 to $40 per hour, {see Burke Deck, ¶ 13; Foster Deck, ¶ 13; Garrett Deck, ¶ 15; Lueders Deck, ¶ 13), in service costs, {see Naylor Deck, ¶ 22; Horner Deck, ¶ 8; Lemanek Deck, ¶ 20; Lowe Deck, ¶ 9). This is particularly true given the important role physician preference plays in brand selection. And because anesthesiology machines are expensive, each purchase must be approved months ahead of time as part of a hospital’s capital budget for a given year. {See Naylor Deck, ¶ 19.) In his deposition, Mark Jones, an Ohmeda employee, conceded that it “can often take a number of years” for a hospital to switch all of its gas machines from one brand to another, (Jones Depo. at 519). Finally, plaintiffs assert that the lack of a market for used anesthesia equipment makes it harder for customers to switch brands. They aver that the trade-in value of a $40,000 anesthesia system is just $3,000. {See Horner Deck, ¶ 6.)

Ohmeda offers evidence that points in the other direction regarding life cycle pricing by customers and the existence of lock-in, {see, e.g., Bernick Deck, ¶ 8; Bol-dan Deck, ¶ 9; DeMur Deck, ¶ 9 (consumers regularly calculate life cycle cost); Bernick Deck, ¶ 12; DeMur Deck, ¶ 13; Rongey Deck, ¶ 15 (consumers would consider switching to another brand in the event of a substantial price increase)), but these are disputed issues for trial. Kodak I does not entitle Ohmeda to summary judgment on this claim.

J. Barriers to Entry and Potential to Increase Output

Even if a § 2 plaintiff shows that the defendant has a 100% share of a given market, it cannot make out a valid § 2 claim without showing barriers to market entry and expansion. See Kodak II, 125 F.3d at 1208 . “Barriers to entry ‘must be capable of constraining the normal operation of the market to the extent that the problem is unlikely to be self-correcting.’ ” Id. (quoting Rebel Oil, 51 F.3d at 1439 ). Common entry barriers include patents and other intellectual property licenses, control of essential or superior resources, entrenched buyer preferences, high capital entry costs and economies of scale. See id.

In their declarations, plaintiffs assert that hospitals will not buy service from ISOs who do not maintain a full inventory of Ohmeda parts. (See Burke Deck, ¶¶ 19-21 & Exh. B; Garrett Deck, ¶ 20; Hoyt Deck, ¶¶ 16-17; Horner Deck, ¶ 19; Wier Deck, ¶ 7; Lueders Deck, ¶ 14.) Robert Burke, the owner of plaintiff Safety Anesthesia Equipment Services (“SAES”), states that “[a]s a result of Ohmeda’s restrictive and anticompetitive policy, SAES has been prohibited from expanding its business beyond those customers with whom I have established a business relationship.” (Burke Deck, ¶ 20.) Burke names six specific hospitals that have refused to contract with SAES because it cannot guarantee a full inventory of Ohme-da parts, and estimates that Ohmeda’s parts policy costs him at least $130,000 per year. (See id., ¶ 21.) Burke also submits a bid request from a hospital that requires ISO bidders to “show proof of parts stock relevant to each item listed on bid requisition” and “provide exact model loaners to the hospital if the hospital machine is down past one day of operation” in order to obtain the hospital’s service contract. (Id., Exh. B at 2.)

Similarly, Daniel Coholan, president of plaintiff De-Tec, Inc., contends that De-Tec has been unable to expand its business into the Ohmeda service market because of Ohmeda’s policy. (See Coholan Deck, ¶ 17.) Coholan also names six hospitals that would have contracted with De-Tec for service on Ohmeda machines were it not for Ohmeda’s restrictive parts policy. (See id., ¶ 18.) Gerald Hoyt, president of plaintiff Bio-Medic, Inc., also alleges that Ohmeda’s parts policy has prevented Bio-Medic from expanding its service business beyond its existing customer base, (see Hoyt Deck, ¶ 21), and names five hospitals that would have contracted with Bio-Medic had Ohmeda been willing to sell parts directly to ISOs, (see id., ¶ 22).

Plaintiffs concede they are generally able to obtain service restricted parts on the gray market, although they are required to pay up to 25% more than Ohme-da’s list price. (See Garrett Deck, ¶ 19; L. McBride Deck, ¶ 19; see also Foster Deck, ¶ 17; Keogh Deck, ¶ 16.) Plaintiffs assert, however, that spending so much time, energy and money locating parts through alternative channels prevents them from increasing their output. (See Garrett Deck, ¶ 19; Keogh Deck, ¶ 16; Lueders Deck, ¶ 18.) And although Ohmeda has recently begun to allow ISOs to purchase parts directly under the QISO program, plaintiffs assert that the program takes many years to complete, is prohibitively expensive, and can be canceled by Ohmeda at any time without notice. (See Ardrey Deck, ¶ 24; Coholan Deck, ¶¶ 23, 25-26; Foster Deck, ¶¶ 23-24; L. McBride Deck, ¶ 24.) This is enough evidence of barriers to expansion to avoid summary judgment.

B. Willful Acquisition or Maintenance of Monopoly Power

A § 2 monopoly plaintiff must also show monopolistic conduct by the defendant, which is the use of monopoly power “to foreclose competition, to gain a competitive advantage, or to destroy a competitor.” Kodak I, 504 U.S. at 482-83 , 112 S.Ct. at 2090 (quoting United States v. Griffith, 334 U.S. 100, 107 , 68 S.Ct. 941, 945 , 92 L.Ed. 1236 (1948)). Plaintiffs assert that the following actions taken by Ohmeda have injured competition in the Ohmeda service market: (1) refusing to sell service restricted parts to hospitals without a signed waiver letter; (2) refusing to sell parts or manuals to ISOs and insisting that OEMs who manufacture Ohmeda parts do the same; (3) disparaging ISOs to end users; and (4) establishing the allegedly unfair QISO program. As Ohmeda concedes, plaintiffs’ evidence here is similar to the evidence of monopolistic acts introduced by the Kodak I plaintiffs. Where a plaintiff presents evidence of exclusionary action designed “to maintain [a] parts monopoly and [use of] control over parts to strengthen [a] monopoly share of [a] service market,” liability turns on whether legitimate business justifications can explain the defendant’s actions. See id. at 483, 112 S.Ct. at 2091 .

C. Legitimate Business Justifications

Even where a plaintiff shows monopoly power and monopolistic conduct, the defendant may not be liable under § 2 if its conduct is excused by a legitimate business justification. See Kodak II, 125 F.3d at 1212 (citing Kodak I, 504 U.S. at 483 , 112 S.Ct. at 2090-91 ). When the defendant offers a legitimate business justification, the plaintiff may rebut that justification “by demonstrating either that the justification does not legitimately promote competition or that the justification is pretextual.” Id. Here, Ohmeda offers four justifications for its restrictive parts policies: (1) promoting patient safety, (2) reducing potential liability, (3) preserving its reputation, and (4) protecting its investment in its equipment and parts business. (See Mot. at 62-68.)

All four of these justifications are essentially restatements of each other. By promoting patient safety, Ohmeda can avoid liability. By avoiding liability, it can avoid a bad reputation. By avoiding liability and maintaining its good reputation, Ohmeda can avoid commercial damage to its equipment and parts business. In Kodak I, Kodak offered a similar business justification — maintaining high quality service and avoiding blame for equipment breakdowns. See Kodak I, 504 U.S. at 483 , 112 S.Ct. at 2091 . The Ninth Circuit has held that the quality-control justification “is suspect.” High Technology Careers v. San Jose Mercury News, 996 F.2d 987, 991 (9th Cir.1993); see also Mozart Co. v. Mercedes-Benz of North America, 833 F.2d 1342, 1349-52 (9th Cir.1987) (describing “skepticism” with which courts have viewed the justification). In Kodak I, in rebuttal to Kodak’s legitimate business justifications, the ISOs presented evidence that they provided quality service and that some Kodak equipment owners preferred ISO service to Kodak service. The Court found that this evidence was “sufficient to raise a genuine issue of fact” as to the legitimacy of the justification. Id. at 483-84, 112 S.Ct. at 2091 . In this case, plaintiffs have offered similar evidence that some Ohmeda hardware owners consider ISO service to be both safe and superior to Ohmeda service. (See Leath Deck, ¶ 11; Lowe Deck, ¶¶ 9-10; Naylor Deck, ¶¶ 15-16; Perkins Deck, ¶¶ 9-12; Wier Deck, ¶ 7.)

The respondents in Kodak I also argued that Kodak’s quality-control justification was pretextual. Specifically, they pointed out that the quality justification

Kodak I, 504 U.S. at 484 , 112 S.Ct. at 2091 .

Ohmeda echoes Kodak’s argument in its discussion of the need to preserve its reputation, stating that “[i]n any given case, the cause of the problem may be inadequate service by a third party service provider, but to physicians and other clinicians whose work depends on the availability of the system, the problem may appear to be with the equipment or its design rather than with the service provider.” (Mot. at 66-67.) In the anesthesia equipment market, however, it is these “physicians and clinicians” who are primarily responsible for making purchasing and service decisions regarding equipment. (See Bernick Deck, ¶ 3; Boldan Deck, ¶4; DeMur Deck, ¶4; Rongey Deck, ¶ 4.) Like Kodak, Ohmeda portrays its customers as sophisticated when making buying decisions but undiseerning when apportioning blame for breakdowns.

On this record, plaintiffs have established a triable issue as to the legitimacy of Ohmeda’s business justifications.

D. Antitrust Injury

Finally, a § 2 plaintiff must establish antitrust injury by showing that its “injury ‘flows from that which makes defendants’ acts unlawful.” Lucas Automotive v. Bridgestone/Firestone, Inc., 140 F.3d 1228, 1233 (9th Cir.1998) (quoting Cargill, Inc. v. Monfort of Colorado, Inc., 479 U.S. 104, 113 , 107 S.Ct. 484, 491 , 93 L.Ed.2d 427 (1986)). Plaintiffs have done so. To begin with, the asserted antitrust violation in this case is Ohmeda’s restrictive parts policy, and plaintiffs have submitted adequate evidence that they have lost service business as the result of that policy. Plaintiffs have introduced testimony from Ohmeda equipment owners who, because of Ohmeda’s parts policy, use Ohmeda service despite their preference for ISO service, (see Lowe Deck, ¶6-9; Naylor Deck, ¶¶ 13, 16; Oltz Deck, ¶¶ 12-13; Wier Deck, ¶ 7), and who believe that ISO service is of higher quality than Ohmeda service, (see Leath Deck, ¶ 11; Naylor Deck, ¶ 16; Wier Deck, ¶ 7). Plaintiffs also offer testimony from ISOs who claim that they have lost substantial, specified business because of Ohmeda’s parts policies. (See Coholan Deck, ¶¶ 18-19; Foster Deck, ¶¶ 18-19; Garrett Deck, ¶¶ 19-20; Migliaccio Deck, ¶ 10; Nichols Deck, ¶ 21.) This is a sufficient showing that plaintiffs’ injuries are caused by Ohmeda’s challenged parts policies.

In response, Ohmeda argues that, far from harming competition in the service market, its policies have actually had pro-competitive effects. Because of these positive effects, Ohmeda contends that, even if its parts policy has harmed plaintiffs, it has not done the sort of harm that the antitrust laws were enacted to prevent. Specifically, Ohmeda argues that even if its policies have increased prices to supra-competitive levels and restricted competition in the service market, these alleged harms actually benefit plaintiffs: if Ohme-da charges artificially high prices, the argument goes, plaintiffs are able either to make more money by charging higher prices themselves or to boost their market share by undercutting Ohmeda, and if Ohmeda’s policies weed competitors out of the market by decreasing competition, the remaining service firms, such as plaintiffs, will be able to increase their market share and raise their prices. (See Mot. at 86-87.)

If courts adopted such backwards economic logic, a plaintiff would never be able to establish antitrust injury. Plaintiffs allege that Ohmeda is only able to increase its own prices and restrict competition by excluding ISOs such as plaintiffs from large segments of the service market. This is precisely the sort of harm to consumer welfare and competition that the antitrust laws address. Plaintiffs have established a triable issue as to whether they suffered antitrust injury.

E. Summary

Ohmeda is entitled to summary judgment on plaintiffs’ claim of monopolization with respect to the market for service on all anesthesia systems because plaintiffs have not established that Ohmeda has mo nopoly power in that market. Ohmeda is not entitled to summary judgment on plaintiffs’ monopolization claim regarding the market for service on Ohmeda machines.

IV.

Next, plaintiffs contend that Ohmeda attempted to monopolize the market for service on anesthesia equipment, or, alternatively, Ohmeda anesthesia equipment. To make out an attempted monopolization claim under § 2, plaintiffs must show “(1) specific intent to control prices or destroy competition; (2) predatory or anticompetitive conduct directed at accomplishing that purpose; (3) a dangerous probability of achieving “monopoly power”; and (4) causal antitrust injury.” Rebel Oil, 51 F.3d at 1433 (citing McGlinchy v. Shell Chem. Co., 845 F.2d 802 , 811 (9th Cir.1988)). As noted above, plaintiffs have established triable issues as to three of the four elements of the claim: anticompetitive conduct, monopoly power in the service market, and antitrust injury.

The remaining element, specific intent, can be inferred where the defendant’s asserted anticompetitive conduct is “predatory or clearly in restraint of competition such as a per se violation under section 1.” Thurman Indus., 875 F.2d at 1378 (citing Blanton v. Mobil Oil Corp., 721 F.2d 1207, 1214 (9th Cir.1983)). Plaintiffs have come forward with evidence that Ohmeda refused to sell service restricted parts to ISOs, refused to sell such parts directly to end users without a signed waiver of liability, and disparaged the quality of ISO service. Faced with similar evidence, the Kodak I Court permitted an attempt claim to go forward at the summary judgment stage, see 504 U.S. at 485-86 , 112 S.Ct. at 2092 , and the Kodak II panel upheld the subsequent jury verdict on that claim, see 125 F.3d at 1212 (upholding both § 2 verdicts). Accordingly, plaintiffs have offered enough evidence to support an attempt claim, and Ohmeda is not entitled to summary judgment.

Moreover, the minimum required showing of market share on an attempt claim is “a lower quantum than the minimum showing required in an actual monopolization case.” Rebel Oil, 51 F.3d at 1438 . Although most courts have held that a market share of 30% is too low to establish market share for purposes of an attempt claim, see id., the Rebel Oil court found a triable issue as to market share where the defendant had a 44% share of a market with high barriers to entry and expansion, see id. at 1438 & n. 10.

As noted above, plaintiffs have introduced evidence showing that Ohmeda’s share of the market for service on all anesthesia machines is about 46.15%. Plaintiffs have also offered evidence that Ohmeda’s parts policy prevents them from expanding their output. (See Burke Deck, ¶¶ 19-21 & Exh. B; Garrett Deck, ¶ 20; Hoyt Deck, ¶¶ 16-17; Horner Deck, ¶ 19; Wier Deck, ¶ 7.) This is sufficient under Rebel Oil to create a triable issue as to whether Ohmeda attempted to monopolize the market for service on anesthesia machines generally as well as the service market for Ohmeda machines.

Accordingly, Ohmeda is not entitled to summary judgment on either of plaintiffs’ attempt claims.

V.

Plaintiffs allege that Ohmeda has illegally tied a customer’s purchase of Ohmeda service to the sale of Ohmeda equipment and replacement parts in violation of § 1 of the Sherman Act. “A tying arrangement is a device used by a competitor with market power in one market (for the “tying” product) to extend its market power into an entirely distinct market (for the “tied” product). To accomplish this, the competitor agrees to sell the tying product ... only on the condition that its customers also purchase the tied product.” Data- gate, 60 F.3d at 1423 .

To make out a tying claim, plaintiffs must establish that: (1) Ohmeda has tied two products or services sold in different markets; (2) Ohmeda has market power in the tying product, and (3) the tie affects a “not insubstantial volume of commerce.” Id. at 1423-24 . Plaintiffs argue that Ohmeda has illegally tied the purchase of Ohmeda service to the sale of Ohmeda parts. Ohmeda concedes that the alleged tie affects a not insubstantial volume of commerce, but contends that plaintiffs have not established the other two elements of their tying claim.

A. Tie-In Between Ttuo Products

To defeat Ohmeda’s motion for summary judgment on the tying claims, plaintiffs must come forward with evidence that (1) service is a product distinct from equipment and parts, and (2) Ohmeda has tied the sale of those products. See Kodak I, 504 U.S. at 462 , 112 S.Ct. at 2079-80 . Plaintiffs have presented enough evidence to establish a material issue as to whether the market for service is distinct from the market for parts.

Ohmeda argues, however, that plaintiffs have not established that it tied the purchase of Ohmeda service to the sale of Ohmeda parts. Ohmeda contends that it has only provided incentives for customers to buy Ohmeda service by refusing to sell parts directly to ISOs or to hospitals without trained in-house biomeds or a signed waiver letter. It asserts that these incentives, although potentially powerful, do not amount to a tie because it has not explicitly limited the sale of parts to purchasers of Ohmeda service.

But as Datagate makes clear, the “ ‘essential characteristic’ of a per se illegal tying arrangement is that the seller makes use of its market power in the tying product to coerce the buyer to purchase the tied product.” 60 F.3d at 1426 (citing Jefferson Parish, 466 U.S. at 12-13 , 104 S.Ct. at 1558 ). And in an earlier opinion during the Datagate litigation, the Ninth Circuit described a tie as a relationship in which the seller “coerced to some extent the purchaser into buying the tied product.” Datagate, Inc. v. Hewlett-Packard Co., 941 F.2d 864, 870 (9th Cir.1991) (emphasis added). In Datagate, summary judgment was inappropriate because the plaintiff had shown that the defendant’s restrictive policies “precluded [customers] from considering other providers of hardware service.” Datagate, 60 F.3d at 1426 . Here, plaintiffs have offered testimony from customers who would prefer to use ISOs to service their Ohmeda machines but do not because of defendant’s restrictive parts policy. (See Leath Deck, ¶ 11; Lowe Decl., ¶¶ 9-10; Naylor Decl, ¶¶ lb-16; Perkins Decl., ¶¶ 9-12; Wier Decl., ¶ 7.) This is a sufficient showing to establish that Ohmeda “coerced to some extent” hospitals into buying Ohmeda service instead of ISO service, and is thus sufficient evidence of a tying relationship under Datagate. See also Kodak I, 504 U.S. at 462-63 , 112 S.Ct. at 2079-80 .

B. Market Power in the Tying Product

Plaintiffs contend that the tying product in this case is replacement parts for Ohme-da machines and that Ohmeda has market power in this market. (See Compl., ¶¶ 101-02.) Market power under § 1 is a more lenient standard than monopoly power under § 2. See Kodak I, 504 U.S. at 481 , 112 S.Ct. at 2089-90 . The Supreme Court defines market power as “the power ‘to force a purchaser to do something that he would not do in a competitive market.”’ Id. at 464 , 112 S.Ct. at 2080 (quoting Jefferson Parish, 466 U.S. at 14 , 104 S.Ct. at 1559 ). Generally, market power is inferred from a seller’s possession of a dominant share of the market. Id. at 464, 112 S.Ct. at 2081 .

Both parties have assumed throughout the litigation, and the court sees no reason to doubt, that Ohmeda has a dominant share of the market for its own parts. Although plaintiffs are able to buy service restricted parts from sources other than Ohmeda on a secondary gray market, (see Shrago Decl., ¶ 12; Pepper Decl., ¶ 8; Lewitzke Deck, ¶ 9), there is no indication that the gray market occupies anything more than a small fraction of the overall parts market. It is undisputed that Ohme-da has a monopoly over service restricted parts and that it restricts its OEMs from selling parts directly to customers. (See Coholan Decl., ¶ 16; Hoyt Decl,, ¶ 18; Lueders Decl., ¶,16.) This is sufficient evidence to establish market power for purposes of § 1. Cf. Kodak II, 125 F.3d at 1207 (holding that defendant had power in market for its own parts where it manufactured 30% of parts itself, controlled an additional 20-25% through tooling clauses and discouraged self-servicing and resale of parts by end users, even though replacement parts were not “absolutely unavailable” from other sources). Summary judgment is inappropriate as to plaintiffs’ tying claims.

VI.

Finally, plaintiffs contend that Ohmeda and its OEMs engaged in a group boycott in violation of § 1 by conspiring to refuse to sell parts to ISOs. Plaintiffs base this claim on the tooling clause provision in the contracts between Ohmeda and its OEMs that bars the OEMs from manufacturing service restricted parts for anyone other than Ohmeda.

Plaintiffs advance both a per se boycott claim and a boycott claim under the rule of reason. Under the Supreme Court’s recent decision in NYNEX Corp. v. Discon, Inc., 525 U.S. 128 , 119 S.Ct. 493 , 142 L.Ed.2d 510 (1998), however, the per se rule is limited “in the boycott context to cases involving horizontal agreements among direct competitors.” 119 S.Ct. at 498 . The asserted boycott in this case does not involve a horizontal agreement among competitors; instead, the allegation involves a purely vertical agreement between Ohmeda and each of its various OEMs. Ohmeda is not a competitor of its OEMs. And the OEMs who may be competitors have not come to an agreement with one another but individually with Ohmeda. In this circumstance, plaintiffs’ per se boycott claim must be dismissed.

Plaintiffs may still assert a boycott claim under the rule of reason. Ohmeda argues, however, that plaintiffs have not offered sufficient evidence that Ohmeda conspired with OEMs to deny ISOs access to replacement parts. Instead, Ohmeda contends, it unilaterally imposed the tooling clause on its OEMs. A plaintiff asserting a § 1 claim must offer evidence “that reasonably tends to prove that the manufacturer and others ‘had a conscious commit ment to a common scheme designed to achieve an unlawful objective,’ ” Monsanto Co. v. Spray-Rite Serv. Corp., 465 U.S. 752 , 764, 104 S.Ct. 1464, 1471 , 79 L.Ed.2d 775 (1984), including evidence that “tends to exclude the possibility that the manufacturer and [alleged co-conspirators] were acting independently.” Id. at 766, 104 S.Ct. at 1471.

Plaintiffs do not submit any direct evidence that Ohmeda and the OEMs jointly agreed to adopt the tooling clause and boycott parts sales to ISOs, such as declarations from the OEMs. Plaintiffs’ only evidence of a group boycott is the tooling clause itself. While this may be circumstantial evidence of an agreement to a common scheme, where a § 1 plaintiff presents only circumstantial evidence of a conspiracy, “defendant may discharge its initial summary judgment burden by proffering a plausible and justifiable alternative interpretation of its conduct that rebuts the plaintiffs allegation of conspiracy.” American Ad Mgmt., Inc. v. GTE Corp., 92 F.3d 781 , 788-89 (9th Cir.1996) (quoting T.W. Elec. Serv., Inc. v. Pacific Elec. Contractors Ass’n, 809 F.2d 626 , 632 (9th Cir.1987)) (internal quotation marks omitted).

Ohmeda’s assertion that the tooling clause was the result of its independent action and not a conspiracy is a “plausible and justifiable alternative interpretation” of the facts of this case. Ohmeda has every incentive to protect its proprietary parts by requiring its OEMs to agree not to sell those parts.to ISOs. The OEMs, on the other hand, would gain nothing by agreeing to boycott ISOs and forego thousands of dollars in potential sales every year. Commercial common sense supports an inference that Ohmeda presented its OEMs with a.form contract, including a tooling clause, that they were free either to sign or to reject but not to modify. Simple acquiescence to a form contract is normally not the sort of concerted action that constitutes an “agreement” under § 1. See Matrix Essentials v. Emporium Drug Mart, Inc., 988 F.2d 587 , 594 (5th Cir.1993); American Airlines v. Christensen, 967 F.2d 410, 412-15 (10th Cir.1992). Plaintiffs offer no evidence to support the contrary — and counterintuitive — conclusion that the OEMs bargained for or participated in the drafting or adoption of a tooling clause that has the direct effect of limiting their potential profits.

Plaintiffs are free to offer evidence of the tooling clause and its effect on competition in the service market in support of their monopolization, attempted monopolization and tying claims. But because they have not established that the tooling clause was the product of concerted action between Ohmeda and its OEMs, plaintiffs may not bring an independent group boycott claim based on the clause.

VII.

In Kodak I, the Supreme Court acknowledged that, in the end, Kodak’s aftermarket arguments might prove to be correct. “It may be that [Kodak] parts, service, and equipment are components of one unified market, or that the equipment market does discipline the aftermarkets so that all three are priced competitively overall, or that any anti-competitive effects of Kodak’s behavior are outweighed by its competitive effects.” Kodak I, 504 U.S. at 486 , 112 S.Ct. at 2092 . The same may be true here, but the court cannot conclude that Ohmeda is entitled to summary judgment as a matter of law on all of plaintiffs’ claims.

The motion for summary judgment is GRANTED as to plaintiffs’ claim of monopolization of the service market for all anesthesia machines and as to the per se and rule of reason group boycott claims. The motion is DENIED as to all other claims.

IT IS SO ORDERED.

. Fred Nichols of Anatech, Inc. did not enter the industry until January 1996. (See Nichols Deck, ¶ 6.) Although Nichols was an Ohmeda employee prior to and during the limitations period and was aware of Ohmeda’s parts policy when he formed Anatech, (see id., ¶¶ 5-6, 16-17), his late entry into the market means Anatech did not suffer antitrust injury until some point within the limitations period. Thus, even if the other plaintiffs’ claims were time-barred, Anatech’s claims would go forward. . Defendant also relies on Pace Industries, 813 F.2d 234 , but that case is limited to a particular kind of allegation. In Pace Industries , the only alleged antitrust violation was the defendants’ prosecution of a lawsuit seeking to enforce an assertedly illegal contract. The plaintiff argued that "each phase” of the lawsuit "from the filing of [the] complaint through resolution of the final appeal constituted discrete overt acts which restarted the statute of limitations.” Id. at 237 . The Ninth Circuit rejected this argument, holding that the filing of the suit was the last overt act for statute of limitations purposes "where the alleged antitrust violation is the attempted enforcement of an illegal contract through the judicial process.” Id. . However, as counsel for plaintiffs acknowledged at oral argument, plaintiffs’ recovery is limited to damages suffered within the four-year limitations period. See Hanover Shoe, Inc. v. United Shoe Mach. Corp., 392 U.S. 481, 502 , 88 S.Ct. 2224, 2236 , 20 L.Ed.2d 1231 (1968); LaSalvia v. United Dairymen of Ariz., 804 F.2d 1113, 1119 (9th Cir.1986). . Although ''[u]ltimately what constitutes a relevant market is a factual determination for the jury,” Kodak II, 125 F.3d at 1203 , a court may grant summary judgment on a § 2 claim where no reasonable jury could find in favor of the plaintiff on the relevant market issue, see Rebel Oil, 51 F.3d at 1435 . . Both parties assume for this motion that the relevant geographic market is the United States. (See Mot. at 33 n. 100.) . Life cycle pricing is discussed later in greater detail as part of Ohmeda’s attempt to distinguish Kodak I on the basis that its customers (1) were not surprised by any sudden changes to Ohmeda's parts sales policy, (2) make purchasing decisions in the equipment market based on costs throughout the life cycle of the equipment, and (3) were not locked in to continued use of Ohmeda machines by high switching costs. These factors are considered infra in Part III.A.3 as part of the evaluation of Ohmeda’s market power, although they also relate to market definition. See Kodak I, 504 U.S. at 469 n. 15, 112 S.Ct. at 2083 n. 15 (‘‘Whether considered in the conceptual category of 'market definition' or 'market power,' the ultimate inquiry is the same — whether competition in the equipment market will significantly restrain competition in the service and parts markets.”). . These statistics include monitor service as part of the relevant market; excluding monitors would likely increase Ohmeda’s market share. .Plaintiffs argue that Ohmeda’s share of the service market on its machines is actually higher than the data suggest because the data include customers who service their machines in-house and thus, according to plaintiffs, do not participate in the service market at all. (See Opp'n at 44-46.) But because the proper definition of an antitrust market turns on "whether consumers view the products as substitutes for each other,” Rebel Oil, 51 F.3d at 1435 , in-house technicians may be viewed as part of the overall service market. In-house technicians are a substitute for outside service contracts, whether with Ohmeda or an ISO. . In other words, A (Ohmeda’s share of the Ohmeda service market) will equal B (Ohme-da's share of the total service market) divided by C (Ohmeda’s share of the equipment market, or the percentage of all machines that were made by Ohmeda). . Forty percent (Ohmeda's share of the service market on all machines, according to Ohmeda) divided by .55 (Ohmeda’s share of the market for sales of anesthesia equipment) equals 72.73% (.40 / .55 = .7273). This figure is very close to the 71% share of the market for service on Ohmeda anesthesia machines that plaintiffs contend Ohmeda controls. . Acknowledging that the Kodak I Court “did not explicitly state the extent to which Kodak's change in policy affected the Court's analysis,” the Sixth Circuit in Honeywell bases its interpretation of Kodak I on the following cryptic statement in a footnote: [t]he dissent disagrees [with our conclusion in this case] based on its hypothetical case of a tie between equipment and service. "The only thing lacking” to bring this case within the hypothetical case, states the dissent, “is concrete evidence that the restrictive parts policy was generally known.” But the dissent's "only thing lacking” is the crucial thing lacking — evidence. 504 U.S. at 477 n. 24, 112 S.Ct. at 2087 n. 24. This enigmatic comment, designed as a brisk rejoinder, is too slender a reed on which to base a significant restriction on the Court's holding particularly when in its very next sentence, the Kodak I majoritj' emphasizes that liability "depends on whether the equipment market prevents the exertion of market power in the [aftermarket].” The Sixth Circuit also attempts to limit Kodak I in light of Jefferson Parish, on the theory that unless a change of policy is understood as a critical component of the Kodak I decision, then Kodak I would implicitly overrule Jefferson Parish. But the two cases are so factually dis tinct that there is little or no tension between them. In the factual setting of Jefferson Parish, where there was no evidence of an increase in price or decrease in quality of the tied product — anesthesiology services — and no evidence that any consumer was forced to accept unwanted anesthesiology’services, the Court was unwilling to base a finding of an illegal tying arrangement solely on alleged market imperfections, including information costs. Jefferson Parish is not addressed to the question of information costs in the context of after market services and the effect of competition in the equipment market. Nor is much of the discussion in the Seventh Circuit’s opinion in Digital consistent with Kodak I. The Digital court noted that Kodak I could not stand for the proposition that "there is a market in a firm’s own products, even if it sells in vigorous competition," 73 F.3d at 762 , because then Kodak I would have effectively overruled the Supreme Court's earlier decision in Jefferson Parish. In the first place, Kodak I does hold, and unambiguously, that a single brand can be a relevant antitrust market where "service and parts for [that brand] are not interchangeable with other manufacturers’ service and parts.” See Kodak I, 504 U.S. at 482 , 112 S.Ct. at 2090 . It is doubtful that this holding is inconsistent with Jefferson Parish, in which the Court found that a hospital — the “single brand” — did not have market power because consumers could easily travel to other hospitals. In any event, Digital does not involve the specific question of aftermarket servicing that was the issue in Kodak I and is the issue here. . Declining to reverse the jury’s verdict for plaintiffs, the court stated that [t]he record reflects the following [evidence of monopoly power in the service market]: Kodak has a 95% share of the Kodak high volume photocopier service market and an 88% share of the Kodak micrographic service market. Several ISOs withdrew from the Kodak service market or substantially restricted their service due to the lack of available parts, although other ISOs made substantial profits and continued to grow steadily. Some ISOs could not obtain any parts for newer models, indicating that the current ISO service market may last only as long as the pre-1986 models survive. Kodak equipment customers experienced the “lock-in” and information imperfections as described in [Kodak /]. Kodak’s market share in the equipment market further limits choices by consumers. Finally, although Kodak criticizes their methodology, the ISOs presented evidence that Kodak earns supracompetitive profits in service, and overall. Substantial evidence supports the jury's verdict on the issue of Kodak’s service market monopoly. Kodak II, 125 F.3d at 1212 . . Ohmeda replies by noting that this figure applies only to machines traded in at the end of their useful lives. {See Horner Deck, ¶ 6.) Ohmeda asserts that a newer used machine would command significantly more on the trade-in market, but they provide no testimony or evidence to support this specific assertion. Ohmeda does offer general testimony about the "lively” market for used anesthesia systems. (See Brandmeier Deck, ¶ 16; Willig Deck at 89.) This is an issue for the fact finder to resolve. . Kodak also refused to sell parts to ISOs, controlled sales of parts by OEMs, pressured L its customers not to purchase service from ISOs, and restricted the availability of used Kodak machines. See Kodak I, 504 U.S. at 456-58 , 112 S.Ct. at 2076-78 . . In their complaint, plaintiffs assert both a per se tying claim and a tying claim under the rule of reason. (See CompL, ¶¶ 100-111.) As the Sixth Circuit noted in Honeywell, however, these theories have merged in recent years. See Honeywell, 104 F.3d at 815 n. 2. As applied by the Supreme Court in Kodak I, per se lying analysis involves detailed factual inquiries into the defendant’s market power and whether the defendant's challenged conduct has procompetitive effects, see Kodak I, 504 U.S. at 478-79 , 112 S.Ct. at 2088-89 , which looks more like traditional rule of reason methodology than a per se rule. The Supreme Court never mentions the terms "per se” or "rule of reason” in its Kodak I opinion. Professors Areeda, Elhauge and Ho-venkamp explain the apparent merger of the two tying theories by stating that "the per se rule against tying is ‘per se' in only one respect — namely, dispensing with proof of anti-competitive effects....” 10 Areeda, Elhauge & Hovenkamp, Antitrust Law ¶ 1760e, at 372 (1996). The two theories may be analyzed together on this motion. . As was the case with the distinction between the equipment sales market and the service market, there is sufficient customer demand for market participants to provide service but not parts and vice versa. Like Kodak I, there is evidence in the record in this case "that service and parts have been sold separately in the past and still are sold separately to self-service equipment owners.” Kodak I, 504 U.S. at 462 , 112 S.Ct. at 2080 . Many, if not most, ISOs do not sell parts and would presumably not need to if Ohmeda would make service restricted parts readily available to equipment owners. . In fact, the inadmissible hearsay evidence plaintiffs have offered supports the opposite conclusion. Several plaintiffs testify that various OEMs have told them that although they would like to sell service restricted parts directly to ISOs, Ohmeda will not let them. (See Foster Deck, ¶ 16; Garrett Deck, ¶ 18; J. McBride Deck, ¶ 5.) The alleged statements by OEMs are out-of-court statements offered for their truth — i.e., to prove that Ohmeda coerced OEMs into a group boycott — and as such are hearsay, see Fed.R.Evid. 801(c), and are inadmissible to defeat a motion for summary judgment, see Anheuser-Busch v. Natural Beverage Distributors, 69 F.3d 337 , 345 n. 4 (9th Cir.1995), unless they fall within an exception to the hearsay rule, see, e.g., Fed. R.Evid. 801(d), 803, 804. Plaintiffs do not suggest any potentially applicable hearsay exceptions and none is readily apparent. In any event, although plaintiffs offer the statements to prove the existence of a common scheme, they actually point to the opposite conclusion. The fact that OEMs would sell service restricted parts directly to ISOs but for Ohmeda's insistence on the tooling clause tends to prove that the tooling clause was Ohmeda's unilateral creation rather than the result of concerted action between Ohme-da and its OEMs. As noted below, it would make no economic sense for OEMs to agree to the tooling clause if they had any choice in the matter, because the tooling clause restricts their ability to sell their products and earn revenue. . In one case, the Ninth Circuit appeared to create another requirement for a defendant seeking summary judgment in a § 1 case: proof that "permitting an inference of conspiracy would pose a significant deterrent to beneficial procompetitive behavior.” In re Coordinated Pretrial Proceedings in Petroleum Products Antitrust Litig., 906 F.2d 432, 440 (9th Cir.1990). The Ninth Circuit has not always imposed the requirement in subsequent cases. See American Ad Management, 92 F.3d at 788-89. The significant deterrent requirement has its origins in the Supreme Court's opinion in Matsushita Elec. Indus. Co. v. Zenith Radio Corp., 475 U.S. 574 , 106 S.Ct. 1348 , 89 L.Ed.2d 538 (1986). In Matsushita, the Court "emphasized that courts should not permit factfinders to infer conspiracies when such inferences are implausible, because the effect of such practices is often to deter procompetitive conduct.” 475 U.S. at 593 , 106 S.Ct. at 1360. "Thus, mistaken inferences in cases such as this one are especially costly, because they chill the very conduct the antitrust laws are designed to protect.” - Id. at 594, 106 S.Ct. at 1360. Here, the inference plaintiffs ask the court to draw — that OEMs agreed with Ohmeda to adopt a tooling clause that substantially limited the OEMs' sales opportunities — is implausible on its face, and plaintiffs have offered no evidence to suggest otherwise. Moreover, plaintiffs do not dispute that tooling clauses are standard throughout the medical anesthesia equipment industry and have the effect of protecting Ohmeda's proprietary service restricted parts from counterfeiting. There is no question that Ohmeda’s ability to protect its proprietary parts enhances its ability to compete in the equipment sales market. Inferring a conspiracy from the tooling clause would restrict Ohmeda's ability to protect its proprietary parts and thus hamper this beneficial procompetitive behavior. Similarly, it would interfere with an OEM’s ability to accede to unilaterally imposed conditions. This is a sufficient showing to satisfy the Petroleum Products standard.