Metrix Warehouse, Inc., (Metrix) brought the underlying action in this case against Mercedes-Benz of North America (MBNA) *1035under Section 1 of the Sherman Act.1 MBNA filed a counterclaim grounded on Section 2(c) of the Robinson-Patman Act.2 This is an interlocutory appeal from three of the district court’s rulings in the case.
MBNA is the exclusive franchisor of Mercedes-Benz dealerships in the United States and, among other things, has been responsible for the sale and distribution of Mercedes-Benz replacement parts to its franchised dealers. Metrix, an independent automobile parts distributor, sells Mercedes-Benz replacement parts to independent garage repair shops and, when it can, to Mercedes-Benz dealers. MBNA’s franchise agreements with its dealers, however, required the dealers to buy Mercedes-Benz replacement parts only from MBNA. In its complaint, Metrix claimed that MBNA’s enforcement of this requirement was a tying agreement that constituted a per se violation of Section 1 of the Sherman Act or alternatively that it constituted a violation of the Act under the rule of reason. MBNA charged in its Robinson-Patman Act counterclaim that it suffered lost profits as a result of Metrix’s “incentive” program, in which Metrix paid cash and prizes to personnel at Mercedes-Benz dealerships to induce them to purchase replacement parts from it.3
The jury awarded Metrix approximately $2.3 million in damages ($7,005,930 after trebling) on its Sherman Act claim. It also resolved MBNA’s Robinson-Patman Act counterclaim in favor of Metrix finding specifically that MBNA suffered no antitrust injury from Metrix’s “incentive” program.
The district court upheld the jury’s liability verdict on the Sherman Act claim, but set aside its damage award to Metrix as “contrary to the weight of evidence.” The court also granted judgment notwithstanding the verdict (or alternatively a new trial) to MBNA on its Robinson-Patman Act counterclaim. It concluded that on the basis of the evidence presented “the jury could not reasonably return a finding of no [antitrust] injury” on the counterclaim. The district court certified questions of law underlying its rulings on both Metrix’s claim and MBNA’s counterclaim, and we granted both parties’ petitions and consolidated them for interlocutory review.4 We *1036affirm both the district court’s judgment upholding MBNA’s Sherman Act liability and its judgment granting a new trial on Metrix’s damages. We reverse the district court’s judgment on MBNA’s counterclaim. We discuss sequentially the issues relating to Sherman Act liability, Sherman Act damages, and the Robinson-Patman Act counterclaim.
I. Sherman Act Liability
In this interlocutory appeal, we, of course, do not address the ultimate issue of MBNA’s Sherman Act liability. The interlocutory question we consider relates to the Supreme Court’s decision in Pick Manufacturing Co. v. General Motors Corp., 80 F.2d 641 (7th Cir.1935), aff'd per curiam, 299 U.S. 3, 57 S.Ct. 1, 81 L.Ed. 4 (1936), and the issue is whether Pick established an automobile exception to Sherman Act liability or, if it did not, whether MBNA proved a business justification for its tying arrangement as a matter of law. Since our consideration is limited in this manner, it is not necessary to decide whether Metrix proved the requisite elements of an illegal tying arrangement. An understanding of the interlocutory liability issues, however, requires at least a summary of facts bearing on all the Sherman Act liability issues tried by the district court.
Since 1965, MBNA has been the exclusive franchisor of Mercedes-Benz dealerships in the United States. As a wholly-owned subsidiary of Daimler-Benz Aktiengesellschaft (DBAG),5 the West German manufacturer of Mercedes-Benz vehicles, MBNA has also been responsible for the wholesale distribution of Mercedes-Benz passenger cars and replacement parts to its franchised dealers in this country. At the time of the events in question here, all of the cars and parts MBNA distributed to its dealers were sold pursuant to a standard franchise agreement that it executed with each of its dealers.
Under the terms of the franchise agreement, MBNA retained almost complete control over the allocation of new cars to the dealers. Its relationship with its franchisees, however, was not limited to allocating new cars and overseeing their sale. The agreement required each dealer to establish a customer-service department for the repair of Mercedes-Benz automobiles. It also governed specific repair practices. Most germane to this appeal is Section 9(c) of the agreement, which restricted the dealers’ right to choose the source of replacement parts used in the repair of Mercedes-Benz automobiles. Entitled “Quality of Operational Parts Used and Sold by Dealer,” this provision stated that:
Dealer shall neither sell or offer to sell for use in connection with MB passenger cars nor use in the repair or servicing of MB passenger cars any parts other than genuine MB parts or parts expressly approved by DBAG if such parts are necessary to the mechanical operation of such MB passenger cars.
At least eight independent wholesale distributors of Mercedes-Benz automobile replacement parts competed with MBNA for the sale of these parts. Metrix was one of these competitors. It had commenced operations in New Jersey in 1969 as an independent wholesale distributor of replacement parts for foreign cars. After three years of selling parts for many different models, Metrix restricted its inventory to *1037“fast-moving” replacement parts for Mercedes-Benz passenger cars.6 Although Metrix initially had limited its sales area to the northeastern United States, by 1975 it was competing nationwide with MBNA for sales to Mercedes-Benz dealerships. Unlike MBNA, however, Metrix also sold Mercedes-Benz parts to independent garages specializing in the repair of Mercedes-Benz automobiles. Many of the parts Metrix sold were of identical make and manufacture as those distributed by MBNA. Although Metrix also obtained some parts from sources that MBNA did not use, these parts were designed for the Mercedes-Benz automobile and in many cases were functionally identical to those supplied by MBNA.
The majority of MBNA’s “fast-moving” replacement parts were produced by independent manufacturers in West Germany according to DBAG’s standards and specifications. DBAG subjected samples of the manufactured parts to quality control tests and, if satisfied, exported them to MBNA for resale to the dealerships. Although some of the parts Metrix imported were manufactured by the same West German independent manufacturers to DBAG standards and specifications, it did not employ after-purchase quality testing procedures.
Some of the West German parts manufacturers established distribution centers in the United States. MBNA purchased at least twenty percent of its total parts requirements directly from these outlets. However, MBNA subjected none of these parts to DBAG’s quality control procedures and, like Metrix, performed no independent quality testing. Metrix also purchased parts from these domestic facilities and resold them to dealerships and to independent garages. Dealers who acquired parts from the American distribution centers therefore received products of identical quality, whether they purchased the part from MBNA, Metrix, or another wholesaler.
Uncontroverted trial evidence established that Metrix’s lower prices, quicker deliveries, and ability to supply in many instances the identical product that MBNA offered were a profitable combination in Metrix’s early years. Other independent warehouse distributors enjoyed similar competitive advantages. This rivalry took its toll on MBNA’s replacement part sales to its dealerships. In the early 1970s, the market inroads of the warehouse distributors became significant. MBNA still possessed a high market share, but competition in the replacement parts market had cut into MBNA’s profits.7 According to Metrix, MBNA began enforcing the Section 9(C) requirements in the mid 1970s.
Although its business with independent garages had increased steadily, in 1976 Metrix’s share of the dealer market began to slip. At the same time the evidence reflected that retail sales of Mercedes-Benz automobiles had reached a high point. MBNA had cut prices on some parts, but Metrix attempted to maintain an average twenty-percent margin between its prices and those of MBNA and had matched many of MBNA’s price cuts. Nevertheless, Metrix’s replacement parts sales to dealers fell in 1977 and 1978.8 In 1979, Metrix filed this action.
In Metrix’s case-in-chief, it principally tried to prove that MBNA used its franchise power generally, and the allocation of *1038new cars specifically, to force its dealers to comply with Section 9(C) of the franchise agreement. Its trial design was to show that MBNA officials linked new car access and other favorable treatment to the dealers’ purchase of MBNA’s “genuine” parts, creating a classic tying arrangement that was per se unreasonable. Alternatively, Metrix attempted to prove that MBNA’s tying arrangement violated the Sherman Act under the rule of reason.
MBNA presented multiple defenses, but the ones that we consider in this appeal are those grounded on its interpretation of the Supreme Court’s decision in Pick. It argued below and contends here that Pick established an “automobile exception” to the Sherman Act’s proscription against tying arrangements. It also insists on appeal that even if Pick stopped short of establishing such a blanket exception, the Supreme Court’s opinion still recognized a business justification defense to liability for automobile tie-ins. MBNA argues that its evidence established that defense as a matter of law, i.e., its tying arrangement was essential to preserve the goodwill of its customers, as well as to protect itself from its dealers’ fraudulent behavior. As an independent ground for reversal, MBNA asserts that the jury instructions on the business justification issue improperly shifted to it the burden of proving reasonableness in Metrix’s rule of reason claim.
A. MBNA’s Claim that Pick Established an “Automobile Exception” to Liability for Tying Arrangements
MBNA contends that the Supreme Court’s affirmance of the Seventh Circuit’s decision in Pick carved out an “exception” to standard Sherman Act analysis of tying arrangements.9 According to MBNA, this exception permits automobile manufacturers to prohibit their dealers from purchasing replacement parts from independent sources.
In Pick, an independent manufacturer of automotive replacement parts, proceeding under Section 3 of the Clayton Act, 15 U.S.C. § 14, unsuccessfully challenged provisions in General Motor’s (GM) dealer contracts that required the dealers to use only “genuine” GM replacement parts in their repairs of Chevrolet and Buick automobiles. The trial court found, and the Court of Appeals agreed, that the provisions were “reasonable and necessary to further [GM and its subsidiaries’] business as well as the interests of buyers of cars of their manufacture.” Pick, 80 F.2d at 642. The Court of Appeals observed that the restriction preserved consumer goodwill toward the new car manufacturer and protected new car warranties (which would be invalidated by the use of “non genuine” parts). Id. at 643. The Court of Appeals discussed both the “reasonableness” of GM’s actions under the rationale of United Shoe Ma*1039chinery Corp. v. United States, 258 U.S. 451, 42 S.Ct. 363, 66 L.Ed. 708 (1922), and the statutory element of substantial lessening of competition under Section 3 of the Clayton Act.10 Pick, 80 F.2d at 643, 644. It concluded both that the tying provision was reasonable and that no substantial lessening of competition resulted from it. Id. at 644.
The Supreme Court affirmed the Seventh Circuit in a per curiam opinion. Pick Manufacturing Co. v. General Motors Corp., 299 U.S. 3, 57 S.Ct. 1, 81 L.Ed. 4 (1936). Its decision, however, referred only to the issue of substantial lessening of competition under the Clayton Act:
Upon the evidence adduced at the trial the District Court found that the effect of the clause had not been in any way substantially to lessen competition or to create a monopoly in any line of commerce. This finding was sustained by the Court of Appeals.
Under the established rule, this Court accepts the findings in which two courts concur unless clear error is shown.
Id. at 4, 57 S.Ct. at 2 (citation omitted).
The Supreme Court’s holding in Pick suggests no more than that it found no clear error in the lower courts’ factual finding, i.e., the plaintiff failed to prove that GM’s tying arrangement substantially lessened competition, as required to establish a violation of the Clayton Act. Contrary to MBNA’s contention, the Court simply did not consider or address the “reasonableness” of the tie-in as a matter of law, and its opinion bears none of the marks of an intended broad-based exception to the Sherman Act.11 In contrast, there was evidence in the present case to support the jury’s conclusion that MBNA’s tying arrangement had a substantial anticompetitive impact. We therefore not only disagree with MBNA’s expansive interpretation of Pick, but conclude that the evidence of anticompetitive impact forecloses any comparison with the Supreme Court’s ruling in Pick.
There is a final reason why Pick cannot control the outcome of this case. Pick was decided solely under the rule of reason; it predated the application of per se rules to tying arrangements by more than ten years. To interpret Pick as creating an “automobile exception,” we would have to ignore our own precedent, see Osborn v. Sinclair Refining Co., 286 F.2d 832 (4th Cir.1960), cert. denied, 366 U.S. 963, 81 S.Ct. 1924, 6 L.Ed.2d 1255 (1961), as well as *1040four decades of Supreme Court precedent. Recently in Hyde, a majority of the Court expressly reaffirmed that “[i]t is far too late in the history of our antitrust jurisprudence to question the proposition that certain tying arrangements pose an unacceptable risk of stifling competition and therefore are unreasonable ‘per se.’ ” Hyde, 466 U.S. at 9, 104 S.Ct. at 1556 (footnote omitted). In our view, not only does Pick not establish the reasonableness of MBNA’s tying arrangement as a matter of law, but we perceive nothing in Pick justifying a departure from standard per se analysis in this case.
B. MBNA’s “Business Justification” Defense
We turn to MBNA’s contentions that it was entitled to a directed verdict or judgment notwithstanding the verdict on liability because it proved “every conceivable element” of a business justification defense. In this context, MBNA advances three related arguments. In the first two, MBNA contends that its tying arrangement was necessary to preserve consumer goodwill because (1) it ensured the quality of the replacement parts sold and (2) it fulfilled customers’ expectations of receiving “genuine parts” at MBNA dealerships. MBNA’s third argument is that the tie-in protected it from unscrupulous dealers who submitted warranty claims on non-MBNA parts. In light of the evidence presented at trial, none of these arguments persuades us. An asserted business justification cannot salvage a tying arrangement that is otherwise per se unlawful without proof that means less restrictive than the tie-in were not feasible to achieve the desired protection. See Standard Stations, 337 U.S. at 306, 69 S.Ct. at 1058; see also Betaseed, Inc. v.U & I Inc., 681 F.2d 1203, 1227 (9th Cir.1982); Carpa, Inc. v. Ward Foods, Inc., 536 F.2d 39, 46-47 (5th Cir.1976); ABA Antitrust Section, Monograph No. 8, Vertical Restrictions Upon Buyers Limiting Purchases of Goods from Others 18-20 (1982).
MBNA’s first goodwill argument— that the tie-in was necessary to ensure the quality of the replacement parts sold by its dealerships — fails because its evidence did not meet the standard for a directed verdict or judgment notwithstanding the verdict. There was evidence of such quality and weight that reasonable and fair-minded persons in the exercise of impartial judgment could find that feasible means less restrictive than the tie-in were available to maintain quality control. See Wyatt v. Interstate & Ocean Transport Co., 623 F.2d 888, 891 (4th Cir.1980); cf. Betaseed, 681 F.2d at 1227-28. It was undisputed that DBAG issued product design specifications to the West German companies that manufactured over half of the fast-moving replacement parts that MBNA sold to the dealers. This evidence dispelled the notion that it was not feasible to designate preferred manufacturers and provide them with design specifications.12 The court therefore acted properly in submitting the quality control question to the jury and in refusing to overturn its verdict.13
MBNA next contends that its customers expected to receive “genuine parts” from MBNA-franchised dealerships and that this expectation constituted a valid justification for the tying arrangement. The evidence, however, suggests that MBNA could have satisfied its customers’ expectations through less-restrictive means. Sufficient *1041evidence was presented, for example, to support a finding that MBNA could have required its dealers to inform their customers of the origin of the parts sold14 and marshaled its efforts toward informing the consumer of the need to request MBNA’s parts. In this manner, a customer who expected MBNA parts could have purchased them.15 It is not necessary, of course, to speculate on the jury’s reasoning for rejecting MBNA’s goodwill justifications. However, the evidence suggests that MBNA’s judgment of its dealers and consumers ran contrary to the Sherman Act’s policy that competition rule the markets of trade. As the Supreme Court observed in Times-Picayune Publishing Co. v. United States, 345 U.S. 594, 605, 73 S.Ct. 872, 878, 97 L.Ed. 1277 (1953), “any intrinsic superiority of the ‘tied’ product would convince freely choosing buyers to select it over others, anyway.” Perceived consumer expectation, without more, will rarely justify an unlawful tie-in because it is “the public, acting through the market’s impersonal judgment ... [that] allocate[s] the Nation’s resources and thus directfs] the course its economic development will take.” Id.
In its final business justification argument, MBNA contends that the tie-in was necessary to protect it against “free-riding” on its replacement-parts warranty. It maintains that, absent the restraint, its dealers would submit non-MBNA parts to MBNA for warranty reimbursement and asserts that its “right to protect itself against this fraudulent practice should be self-evident.” While MBNA, of course, may protect itself against unfounded warranty claims, it cannot impose a tying arrangement when less-restrictive alternatives are reasonably available. MBNA failed to show that its “genuine” parts could not be distinguished from those of its competitors and that it could not avoid reimbursing dealers for its competitors’ defective parts. Even assuming the parts were indistinguishable, MBNA presented no evidence that procedures such as stamping, inscribing, or numbering were infeasible alternatives to protect it against warranty “free-riding.” Finally, MBNA did not establish that the tie-in was necessary to police dealer misconduct. In its brief, MBNA asserts that its franchisees “commingle aftermarket parts with Mercedes parts in the same bins and do not distinguish between them in making repairs.” This assertion, however, fails to establish that a replacement parts tie-in was the least restrictive method of avoiding unfounded warranty claims.16
In summary, we conclude that the district court correctly denied MBNA’s motions for a directed verdict and judgment notwithstanding the verdict on its affirma*1042tive business justification defense. The evidence was sufficient to preclude a grant of either motion.
C. Jury Instruction on the Burden of Proving Business Justification
We next consider MBNA’s contention that the district court erroneously instructed the jury on the burden of proving business justification in Metrix’s rule of reason claim. MBNA argues that the court misallocated the burden of proof in instructing the jury that:
If you find the tying agreement in this case were [sic] unlawful as to the elements I have previously explained, either because they [sic] were a per se violation of the Sherman Act or because they were unreasonable under the rule of reason, and that MBNA caused Metrix’s injury through its unlawful acts, you will then determine whether they [sic] were, in any event, proper because it was known [sic] as [a] legitimate business justification. The burden is on MBNA to demonstrate the existence of such a business justification.
MBNA asserts that this instruction improperly placed the burden on it to prove that its tie-in was reasonable under the rule of reason. We disagree.
The trial court’s instruction did not shift the burden of proof. The instruction directed the jury to make the business justification inquiry only if it first determined that the tying arrangement violated the rule of reason. Thus, the instruction could not have affected the jury's deliberations until after it had determined that Metrix had proved its rule of reason case.17 In effect, it gave MBNA a second chance to avoid an adverse outcome — an opportunity to which it was not technically entitled under normal proof allocations in rule of reason cases.
II. Sherman Act Damages
The trial court set aside the jury’s award of damages to Metrix on its Sherman Act claim, holding that MBNA was entitled to a new trial on damages because the jury ignored evidence clearly demonstrating that a part of the lost profits Metrix claimed as damages resulted from MBNA’s legitimate procompetitive business activities and not from the tying arrangement. The damages question certified by the district court, as properly applied to the issues developed at trial, is whether the jury’s calculation of Metrix’s damages was against the clear weight of the evidence.
Metrix presented its evidence of lost profits through the testimony of an economist, Dr. Edward Heiden, who concluded that Metrix lost approximately $2.3 million in profits as a result of MBNA’s enforcement of Section 9(C). He theorized that MBNA’s increase in sales of “fast-moving” replacement parts between 1977 and 198118 resulted only from its unlawful tie-in. Heiden identified warehouse distributors’ sales to independent garage repair shops as a “yardstick” against which to measure what Metrix’s dealership sales would have been absent MBNA’s tying arrangement. Through a comparative analysis of sales to independent garages during that time period, Heiden determined that, but for the tie-in, the independent warehouse distributors would have experienced a general market share increase of 13.9 percentage points in their dealership sales between 1977 and 1981. He then translated this projected growth figure into Metrix’s lost profits, calculated the lingering effect of the tie-in on 1982 and 1983 sales, and concluded that Metrix lost slightly more than $2.3 million overall as a result of MBNA’s unlawful conduct.
In defense, MBNA attempted to show that Heiden’s analysis failed to account for the impact of price reductions and lawful *1043marketing programs MBNA introduced while the tie-in was in effect. After several dealers testified that MBNA’s procompetitive efforts influenced their replacement parts purchasing, Heiden testified on rebuttal that MBNA’s lawful attempts to increase its sales had no overall impact on the dealers’ purchasing practices. The jury-rejected MBNA’s evidence of lawful activities and returned a verdict for $2,335,301— the precise amount calculated by Heiden.
The district court set aside the jury’s damage award to Metrix. It held that MBNA was entitled to a new trial on damages because the jury ignored the clear weight of the evidence, which established that at least part of Metrix’s claimed losses were attributable to MBNA’s lawful conduct. Metrix now contends that this was error and requests us to reinstate the jury’s verdict. We disagree and decline. In Wyatt v. Interstate & Ocean Transport Co., this court stated that in considering a new trial motion:
[A] trial court may weigh the evidence and consider the credibility of the witnesses. Indeed, a trial judge has a duty to set aside a verdict and grant a new trial even though it is supported by substantial evidence, “if he is of the opinion that the verdict is against the clear weight of the evidence, or is based upon evidence which is false or will result in a miscarriage of justice____”
623 F.2d at 891-92 (quoting Williams v. Nichols, 266 F.2d 389, 392 (4th Cir.1959)). Here, the clear weight of the evidence supports the trial court’s conclusion that at least some of Metrix’s claimed losses were attributable to lawful conduct. Although Metrix was not obligated to prove its damages with great particularity, we conclude that the trial court did not abuse its discretion in granting MBNA a new trial on damages.19
As Metrix correctly notes, plaintiffs generally enjoy a lenient burden of proof in establishing the amount of damages caused by an antitrust violation. The Supreme Court observed in Zenith Radio Corp. v. Hazeltine Research, Inc., 395 U.S. 100, 89 S.Ct. 1562, 23 L.Ed.2d 129 (1969), that a reasonable amount of uncertainty and lack of precision is tolerable because “damage issues in [antitrust] cases are rarely susceptible of the kind of concrete, detailed proof of injury which is available in other contexts.” Id. at 123, 89 S.Ct. at 1576. Moreover, a wrongdoing defendant should not profit at the expense of his victim where it is his own anticompetitive conduct that precludes direct and specific proof of damage. See Bigelow v. RKO Radio Pictures, Inc., 327 U.S. 251, 265, 66 S.Ct. 574, 580, 90 L.Ed. 652 (1946); see also Story Parchment Co. v. Paterson Parchment Paper Co., 282 U.S. 555, 563, 51 S.Ct. 248, 250, 75 L.Ed. 544 (1931).20 Therefore, when the violation has been proven, “the jury may make a just and reasonable estimate of the damage based on relevant data, and render its verdict accordingly. In such circumstances ‘juries are allowed to act on probable and inferential, as well as upon direct and positive proof.’ ” Bigelow, 327 U.S. at 264, 66 S.Ct. at 580 (quoting Story Parchment, 282 U.S. at 564, 51 S.Ct. at 251); see also Charlotte Telecasters, Inc. v. Jefferson-Pilot Corp., 546 F.2d 570, 573 (4th Cir.1976).
In contrast with the “leniency” principle, our decisions also recognize that the *1044antitrust plaintiff must demonstrate a causal connection between the defendant’s violation and the damages claimed. See Burlington Industries, Inc. v. Milliken & Co., 690 F.2d 380, 385 (4th Cir.1982); Lee-Moore Oil Co. v. Union Oil Co. of California, 599 F.2d 1299, 1306 (4th Cir.1979). As stated in Allegheny Pepsi-Cola Bottling Co. v. Mid-Atlantic Coca-Cola Bottling Co., Inc., 690 F.2d 411, 415 (4th Cir.1982), “[a]n antitrust plaintiff is entitled to recover only for that portion of .... [its losses] caused by the defendants’ unlawful conduct.” See also Dolphin Tours, Inc. v. Pacifico Creative Service, Inc., 773 F.2d 1506, 1510 (9th Cir.1985); MCI Communications Corp. v. American Telephone & Telegraph Co., 708 F.2d 1081, 1161, 1187 (7th Cir.) (majority and dissenting opinions), cert. denied, 464 U.S. 891, 104 S.Ct. 234, 78 L.Ed.2d 226 (1983).
Applying these principles, we agree with the district court that although Metrix presented a reasonable method for estimating damages,21 the jury ignored the clear weight of the evidence that at least part of Metrix’s claimed losses were attributable to lawful conduct. Heiden, Metrix’s expert witness, conceded in rebuttal testimony that his calculations were based on his assumption that MBNA’s legitimate marketing programs introduced in these years had no overall impact on Metrix’s lost sales. This assumption, however, is contrary to the clear weight of the evidence.
MBNA presented direct evidence that at least part of Metrix’s claimed losses were attributable to lawful competitive efforts commenced by MBNA during the damage years. It established that it had reduced prices on popular parts and instituted various procompetitive marketing programs during the years in question. It then presented the testimony of a number of dealers who stated that MBNA’s price reductions and new programs influenced them to purchase more from MBNA and less from the warehouse distributors.22 This direct evidence unequivocally contradicted Heiden’s assumption that MBNA’s *1045lawful conduct had no overall impact.23 We remain sensitive to the complexities of proving hypothetical market behavior. From our review of the record, however, we do not conclude that the district court abused its discretion in holding that the clear weight of the evidence established that at least some of Metrix’s claimed losses were attributable to lawful attempts by MBNA to recapture dealer sales. The district court correctly set aside the verdict and granted MBNA a new trial on damages.
III. Robinson-Patman Act Injury
At trial, the court submitted a special interrogatory to the jury on MBNA’s Robinson-Patman Act counterclaim.24 It asked “[d]o you find that Mercedes-Benz of North America was injured by reason of the illegal incentive program of Metrix.” The jury answered “No.” The trial court, however, granted judgment notwithstanding the verdict to MBNA, holding that MBNA had suffered injury from Metrix’s program. The court also conditionally granted MBNA a new trial, should the judgment notwithstanding the verdict be overturned. The question certified by the district court, as properly applied to the issue developed at trial, is simply whether the court correctly applied the proper standards in granting judgment notwithstanding the verdict or, alternatively, a new trial.
In granting the judgment notwithstanding the verdict, the district court reasoned that “Metrix, acting rationally, would not have persisted in such a program unless it took sales from competitors,” and that it was “simply irrational [for the jury] to find that Mercedes-Benz would not have made any extra sales but for the illegal incentive program.” On appeal, Metrix contends that the jury’s finding of no antitrust injury was clearly supported by the evidence. We agree. While injury usually results from per se violations of the antitrust laws, Lee-Moore Oil, 599 F.2d at 1303, this is not always true. The discrete characteristics of the replacement parts market simply do not lead inevitably to such a conclusion.
*1046At trial, the evidence on MBNA’s Robinson-Patman Act counterclaim was undisputed as to one issue. It clearly established that Metrix offered and awarded cash and prize “incentives” to encourage personnel at Mercedes-Benz dealerships to purchase replacement parts from Metrix. The conflicting evidence related to the effects of Metrix’s program. MBNA attempted to show that absent the unlawful “incentives,” some of Metrix’s sales would have gone to it and that Metrix’s program caused it to suffer approximately $2.3 million in lost profits. Metrix’s evidence showed that “incentive” programs were standard among independent warehouse distributors and that if it had abolished its program, dealership personnel simply would have turned to other warehouse distributors for parts they otherwise would have bought from Metrix. According to the testimony of Metrix’s witnesses, the dealerships that purchased parts from warehouse distributors were motivated by a combination of lower prices, quicker deliveries, and the availability of “incentives;” since MBNA offered none of these amenities, it lost no sales as a result of Metrix’s program.
MBNA’s evidence of antitrust injury principally consisted of the testimony of two dealership personnel and an economist. A dealer testified that contrary to instructions, an employee purchased parts from Metrix instead of MBNA to receive Metrix’s commissions. In deposition testimony, however, the employee had stated that absent Metrix’s program, he would have purchased from other warehouse distributors offering similar “incentives.” At trial, the parts manager at another dealership attributed his purchases of Metrix’s parts to a combination of factors, including lower prices, quicker deliveries, and the “incentive” payments. He did not state, however, that absent Metrix’s illegal program, he would have purchased a greater percentage of parts from MBNA. Finally, an economist testified that if Metrix sold more parts because of its incentive program, it followed that MBNA sold less.
Metrix presented evidence that even without “incentive” payments, dealership personnel would have continued to purchase from warehouse distributors because their prices were lower than MBNA’s prices. Virtually every dealer who testified at trial stated that lower prices were the primary reason he or she purchased parts from the warehouse distributors. The lower prices offered by the distributors translated into a higher profit margin for the dealerships. Metrix also introduced evidence that if it had abolished its program, dealership parts managers simply would have purchased from other warehouse distributors who offered similar “incentives.” Finally, Metrix produced evidence that its average monthly sales levels did not decline when it stopped paying “commissions” in 1983.
As plaintiff, it was MBNA's burden to prove by a preponderance of the evidence that Metrix’s “incentive” program caused it to suffer actual injury of a type that Section 2(c) was designed to prevent. J. Truett Payne Co., Inc. v. Chrysler Motors Corp., 451 U.S. 557, 562, 101 S.Ct. 1923, 68 L.Ed.2d 442 (1981); Larry R. George Sales Co. v. Cool Attic Corp., 587 F.2d 266, 270-72 (5th Cir.1979); see Brunswick Corp. v. Pueblo Bowl-O-Mat, Inc., 429 U.S. 477, 489, 97 S.Ct. 690, 697, 50 L.Ed.2d 701 (1977); Allegheny Peysi-Cola Bottling Co., 690 F.2d at 414 n. 3. MBNA would have had the jury believe that any sales Metrix gained by virtue of the “incentive” program necessarily caused it to lose sales. This argument assumes that MBNA and Metrix were the only participants in the affected market and ignores the seven other independent warehouse distributors. Any sales increases attributable to Metrix’s program may well have come from sales previously enjoyed by the warehouse distributors only.25 For that matter, benign factors, including generally lower prices and quicker deliveries, could have determined whether a dealer made a particular purchase from an independent distributor or from MBNA. These were questions for the jury and, to reiterate, it was MBNA’s burden to demonstrate a causal connection *1047between Metrix’s unlawful program and its lost profits. Zenith Radio Corp., 395 U.S. at 114 n. 9, 89 S.Ct. at 1571 n. 9; Advance Business Systems, 415 F.2d at 63.
In our view, MBNA’s evidence failed to sustain that burden. We cannot accept MBNA’s argument that the mere payment of “commissions” to dealership personnel raised a necessary inference of antitrust injury to MBNA that was disregarded by the jury. Nor can we conclude, as did the trial court, that it was “irrational” for the jury to have made the subsidiary findings necessary to support a conclusion of no antitrust injury. Metrix’s evidence was easily sufficient to support the jury’s finding that MBNA suffered no lost sales from the “incentive” program. Put simply, the jury’s verdict was supported by substantial evidence and was not against the clear weight of the evidence. We therefore reverse the district court’s grant of judgment notwithstanding the verdict or, in the alternative, a new trial to MBNA on its Robinson-Patman Act counterclaim.
In view of the above, the district court’s judgment is affirmed in part and reversed in part. The case is remanded for further proceedings consistent with this opinion.
AFFIRMED IN PART, REVERSED IN PART, AND REMANDED.
. 15 U.S.C. § 1. The section provides in pertinent part: “Every contract, combination in the form of trust or otherwise, or conspiracy, in restraint of trade or commerce among the several States, or with foreign nations, is declared to be illegal....”
. 15 U.S.C. § 13(c). Section 2(c) of the Robinson-Patman Act provides:
It shall be unlawful for any person engaged in commerce, in the course of such commerce, to pay or grant, or to receive or accept, anything of value as a commission, brokerage, or other compensation, or any allowance or discount in lieu thereof, except for services rendered in connection with the sale or purchase of goods, wares, or merchandise, either to the other party to such transaction or to an agent, representative, or other intermediary therein where such intermediary is acting in fact for or in behalf, or is subject to the direct or indirect control, of any party to such transaction other than the person by whom such compensation is so granted or paid.
. In a prior interlocutory appeal, we determined that Metrix’s "incentive” program constituted a per se violation of Section 2(c) of the RobinsonPatman Act. Metrix Warehouse, Inc. v. Daimler-Benz Aktiengesellschaft, 716 F.2d 245 (4th Cir.1983). The factual questions relating to antitrust injury and damages remained for trial.
. As presented by the parties on appeal, and as interpreted by this court, the issues raised by the certified questions are whether Pick Manufacturing Co. v. General Motors Corp., 80 F.2d 641 (7th Cir.1935), aff’d per curiam, 299 U.S. 3, 57 S.Ct. 1, 81 L.Ed. 4 (1936), established an automobile exception to Sherman Act liability for tying arrangements; whether MBNA proved as a matter of law that its tying arrangement was based on a legitimate business justification; whether the jury’s Sherman Act damage award was against the clear weight of the evidence; and whether the district court misapplied the standards for granting a judgment notwithstanding the verdict or, alternatively, a new trial on MBNA’s Robinson-Patman Act counterclaim.
In its motion for certification on the Pick question, MBNA concentrated solely on the issue of reasonableness as a matter of law, reserving the right to raise other issues on appeal after final judgment. The district court said in its certification order:
This question addresses an issue which has permeated this case for more than four years; namely, whether the Supreme Court’s affirmance in Pick Manufacturing Co. v. General Motors Corp., 80 F.2d 641 (7th Cir.1935), aff'd per curiam, 299 U.S. 3 [57 S.Ct. 1, 81 L.Ed. 4] (1936) operates to bar Metrix’s claim under § 1 of the Sherman Act____ *1036The Pick issue is clearly a controlling question of law for its determination materially effects [sic] the outcome of the litigation.
Our consideration of the Sherman Act liability issue then is limited to this question. The parties on appeal treat the issue as bifurcated, i.e., does Pick establish an automobile exception that bars this action and even if not barred as a matter of law did the business justification evidence in this case require a directed verdict for MBNA. MBNA contends that it proved every conceivable element of a business justification defense as delineated by Pick and its progeny and was therefore entitled to a directed verdict. MBNA also contends that the district court gave an erroneous instruction on the burden of proving business justification in its rule of reason instructions. Strictly speaking, that contention does not fit into the question certified by the district court, but it approaches sufficiently close to the principal business justification issue that we will determine it.
. Originally a named defendant in this case, summary judgment was granted in favor of DBAG and that ruling is not subject to review in this interlocutory appeal.
. “Fast-moving" parts are non-warranty, quick-turnover items such as filters, plugs, hoses, and belts. Metrix competed with MBNA for the sale of approximately 3,000 "fast-moving" replacement parts. According to Metrix, “fast-moving” parts accounted for 80% of MBNA’s total parts sales to dealerships.
. Statistical evidence presented at trial demonstrated that MBNA always possessed a controlling share of the dealer market, a share greatly surpassing that of all the independent warehouse distributors combined.
The "dealer market," as that term is used here, refers to the market for wholesale sales of "fast-moving” replacement parts to Mercedes-Benz dealerships. Whether the dealer market constitutes a legally distinct market for the purpose of antitrust analysis is a question not subject to review in this appeal.
. Metrix’s evidence showed that in 1975 it had a 4.22% share of the dealer market, but over the next four years its share dropped steadily to 1.04%.
. In Advance Business Systems and Supply Co. v. SCM Corp., 415 F.2d 55 (4th Cir.1969), we recognized that tying arrangements tend to impair competition by ‘'preventing buyers] from seeking alternative sources of supply for the tied product ... [and by foreclosing] competing suppliers of the tied product ... from that part of the market which is subject to the tying arrangement.” Id. at 60. While these consequences might arise from any exclusive-dealing arrangement, the principal economic purpose of the legal prohibition against tying arrangements is to prevent the extension of market power from one product to another by anticompetitive means. See Jefferson Parish Hospital District No. 2 v. Hyde, 466 U.S. 2, 14-16, 104 S.Ct. 1551, 1557-60, 80 L.Ed.2d 2 (1984); Will v. Comprehensive Accounting Corp., 776 F.2d 665, 669 (7th Cir.1985).
The Supreme Court has held consistently over the past four decades that tying arrangements that manifest certain threshold characteristics are unreasonable per se. See, e.g., Jefferson Parish Hospital District No. 2 v. Hyde, 466 U.S. 2, 104 S.Ct. 1551, 80 L.Ed.2d 2 (1984); United States Steel Corp. v. Fortner Enterprises, Inc., 429 U.S. 610, 97 S.Ct. 861, 51 L.Ed.2d 80 (1977) (Fortner II); Northern Pacific Railway Co. v. United States, 356 U.S. 1, 78 S.Ct. 514, 2 L.Ed.2d 545 (1958); International Salt Co. v. United States, 332 U.S. 392, 68 S.Ct. 12, 92 L.Ed. 20 (1947).
This court, of course, has applied the per se rule to tying arrangements where the relevant criteria were present. See Osborn v. Sinclair Refining Co., 286 F.2d 832, 840 (4th Cir.1960), cert. denied, 366 U.S. 963, 81 S.Ct. 1924, 6 L.Ed.2d 1255 (1961); see also Phillips v. Crown Central Petroleum Corp., 602 F.2d 616, 628-29 (4th Cir.1979); Advance Business Systems, 415 F.2d at 62-63.
. 15 U.S.C. § 14. This section prohibits certain tying arrangements where the effect of the tie-in "may be to substantially lessen competition ... in any line of commerce.”
. The Court’s citations to Pick in the years immediately following the decision stood only for the Court’s longstanding practice of accepting factual findings in which two lower courts concur. See, e.g., Comstock v. Group of Institutional Investors, 335 U.S. 211, 214, 68 S.Ct. 1454, 1456, 92 L.Ed. 1911 (1948); United States v. Appalachian Electric Power Co., 311 U.S. 377, 403 n. 11, 61 S.Ct. 291, 297 n. 11, 85 L.Ed. 243 (1940); Virginian Railway Co. v. System Federation No. 40, 300 U.S. 515, 542, 57 S.Ct. 592, 596, 81 L.Ed. 789 (1937). Moreover, nothing in the Court’s subsequent antitrust decisions suggests that Pick established, even in its own day, a broad antitrust exemption for automobile manufacturers. In Standard Oil Co. of California v. United States, 337 U.S. 293, 302-04, 69 S.Ct. 1051, 1056-57, 93 L.Ed. 1371 (1949) (Standard Stations), the Court identified Pick as a case that predated International Salt, which established the per se rule against tying arrangements. The Court noted that the decision in Pick ”len[t] support to the view” that a "showing as to the actual or probable economic consequences of [tying] agreements” is necessary to establish a violation of § 3 of the Clayton Act. In Times-Picayune Publishing Co. v. United States, 345 U.S. 594, 607, 73 S.Ct. 872, 879, 97 L.Ed. 1277 (1953), the Court cited Pick in a discussion of the development of tying law as a "goodwill” case falling within a doctrine whose boundaries were drawn by International Business Machines Corp. v. United States, 298 U.S. 131, 56 S.Ct. 701, 80 L.Ed. 1085 (1936). Later citations to Pick in Fortner Enterprises, Inc. v. United States Steel Corp., 394 U.S. 495, 514 n. 9, 89 S.Ct. 1252, 1264 n. 9, 22 L.Ed.2d 495 (1969) (White, J., dissenting) (Fortner I), and in Hyde, 466 U.S. at 41, 104 S.Ct. at 1573 (O’Connor, J., concurring) (quoting Fortner I, 394 U.S. at 514 n. 9, 89 S.Ct. at 1264 n. 9 (White, J., dissenting)), similarly do not refer to the decision as an exception that has endured the evolution of tie-in analysis. If anything, these latter citations suggest that the Court of Appeals opinion in Pick properly may be viewed as an early example of a possible business justification defense. See United States v. Mercedes-Benz of North America, Inc., 517 F.Supp. 1369, 1376 (N.D.Cal.1981).
. In Standard Stations the Supreme Court observed that “[t]he only situation ... in which the protection of good will may necessitate the use of tying clauses is where specifications for a substitute would be so detailed that they could not practicably be supplied.” 337 U.S. at 306, 69 S.Ct. at 1058.
. In light of the evidence presented, we cannot agree with MBNA that it proved that "product specification is no substitute for quality control” because DBAG, in its after-purchase testing procedures, rejects 47,000 shipments of parts annually for failure to meet its specifications. Although DBAG’s testing procedures are documented in the record, MBNA’s argument fails to account for the uncontroverted evidence that 20% of its parts underwent no after-purchase quality control procedures. These were the replacement parts MBNA purchased directly from the United States distribution arms of the West German manufacturers. As previously noted, Metrix also purchased identical parts from these outlets and supplied them to the dealerships.
. Indeed, Section 9(D) of the franchise agreement provided that "[djealer shall not represent as new, genuine MB parts or as parts approved by DBAG any parts used by it in the repair or servicing of MB passenger cars, ... which are not in fact new, genuine MB parts or parts expressly approved by DBAG or by MBNA.”
. Preserving consumer goodwill obviously was important to the dealers; their businesses depended on it. It follows that the dealers voluntarily would have purchased only MBNA parts if they considered it necessary to ensure consumer goodwill. The evidence established, however, that the dealers frequently obtained parts from non-MBNA sources.
. Contrary to MBNA’s contention, neither Monsanto Co. v. Spray-Rite Service Corp., 465 U.S. 752, 104 S.Ct. 1464, 79 L.Ed.2d 775 (1984), nor Continental T. V., Inc. v. GTE Sylvania Inc., 433 U.S. 36, 97 S.Ct. 2549, 53 L.Ed.2d 568 (1977), hold that tie-ins are reasonable if implemented to prevent unfounded warranty claims. The "free-rider effect” mentioned in Sylvania referred to the practice of returning defective goods of the same brand to a seller who did not make the sale in question. Sylvania, 433 U.S. at 55, 97 S.Ct. at 2560. This is a far cry from the "free-riding” of which MBNA complains.
Similarly, we cannot accept MBNA’s alternative contention that it was entitled to judgment in its favor because it did not require its franchisees to stock "genuine" parts in excess of consumer demand. The district court correctly instructed the jury that a franchisor may require its franchisees to maintain a stock of the franchisor’s products coordinate with consumer demand. In light of the evidence presented, however, the jury may well have concluded that MBNA’s purposes for imposing the tie-in had little to do with anticipated consumer demand for its products. Indeed, the evidence suggests that MBNA did not enforce Section 9(C) until it lost significant numbers of sales to the independent warehouse distributors.
. Business justification, of course, is an affirmative defense to claims of per se illegality. Under the rule of reason, however, the plaintiff has the overall burden of negating any business justification by demonstrating the unreasonableness of the defendant’s anticompetitive conduct.
. Although Metrix claimed that it lost sales between 1977 and 1981 as a result of MBNA’s tie-in, it sought damages for 1977-1983, alleging that the effect of MBNA’s unlawful conduct lingered for two years beyond its cessation.
. Our review is limited to the trial court’s ruling that the jury's damage verdict was against the clear weight of the evidence. Contrary to MBNA’s argument and the authorities it cites, the district court did not rule that Metrix failed to prove "antitrust injury,” i.e., “injury of the type the antitrust laws were intended to prevent and that flows from that which makes defendants’ acts unlawful.” Brunswick Corp. v. Pueblo Bowl-O-Mat, Inc., 429 U.S. 477, 489, 97 S.Ct. 690, 697, 50 L.Ed.2d 701 (1977); compare Burlington Industries, Inc. v. Milliken & Co., 690 F.2d 380, 386 (4th Cir.1982), and Lee-Moore Oil Co. v. Union Oil Co. of California, 599 F.2d 1299, 1303-05 (4th Cir.1979), with Allegheny Pepsi-Cola Bottling Co. v. Mid-Atlantic Coca-Cola Bottling Co., Inc., 690 F.2d 411, 414-15 n. 4. (4th Cir.1982).
. This rationale takes on special force in market foreclosure cases because probable and inferential proof is often all that is available to hypothesize lost profits. Cf. Bigelow, 327 U.S. at 264, 66 S.Ct. at 579 (“The tortious acts ... precluded ascertainment of the amount of damages more precisely, by comparison of profits, prices and values as affected by the conspiracy, with what they would have been in its absence under freely competitive conditions.”).
. We decline MBNA's invitation to hold that the independent garage “yardstick" method for proving damages is inappropriate in this case. If a plaintiff can demonstrate that its assumptions about market behavior rest on adequate bases and there is a reasonable comparability between the businesses or markets in question, it may properly rely on a "yardstick” measure of proof. See Dolphin Tours, 773 F.2d at 1511; Malcolm v. Marathon Oil Co., 642 F.2d 845, 858 (5th Cir.1981). Here, it appears that the independent garage market was the only available index by which unfettered economic behavior could be measured. We cannot conclude that the "yardstick" method was flawed to the point of irrelevancy or was overly speculative. Cf. Burlington Industries, 690 F.2d at 386; Crown Petroleum, 602 F.2d at 632.
Competition for sales to Mercedes dealerships took place between warehouse distributors and MBNA. Competition for sales to independent garages took place between warehouse distributors and Mercedes’ dealerships, who sold parts they obtained from MBNA and the warehouse distributors. MBNA contends that this distinction in market participants severely undercuts the validity of Heiden’s assumptions. It contends that as the warehouse distributors sold more parts to the independent garages, they took business away from Mercedes-Benz dealers. Therefore, the dealers purchased less from the warehouse distributors than they otherwise would have bought. On the record before us, however, we cannot assess the effect, if any, of the horizontal relationship between warehouse distributors and Mercedes-Benz dealers in the independent garage market on the validity of that market as an index of behavior in an unfettered vertical market. Since it appears that the question goes more to the weight than to the admissibility of Metrix’s theory, we leave it for the district court’s consideration on remand.
. MBNA presented evidence that between 1975 and 1978 it reduced prices on hundreds of parts. MBNA’s purchasing manager testified that MBNA’s sales of oil filters increased 300% as a result of price reductions ranging from 34% to 51%. Indeed, Metrix’s president admitted in his testimony that Metrix’s sales had declined when MBNA reduced its prices and that some of MBNA’s reduced prices were below Metrix’s costs for the same product.
In 1980, MBNA increased its standard “stock order” discount providing lower prices to dealers on certain regular purchases. Moreover, during 1980 and 1981, MBNA established an "obsolescence allowance" that provided discounts of up to 29% off the original list price. Charlotte Mancine, owner of a small Mercedes-Benz dealership in Southern California, testified that her dealership purchased replacement parts from whomever offered the lowest price. She further stated that in 1978 she began purchasing more parts from MBNA because its “prices were lowered and the incentives or stock order allowances were better.” Jules Barsotti, operator of *1045three Mercedes-Benz dealerships in the San Francisco area, testified that in the early 1970s Metrix and other warehouse distributors supplied him with a substantial portion of his replacement parts needs, but in the late 1970s his purchases decreased because "Mercedes-Benz was getting more competitive and their prices were starting to be as good, if not better" than the prices offered by the warehouse distributors. John Burnett, a Mercedes-Benz dealer in Maitland, Florida, testified that MBNA’s price reductions caused him to reduce his purchasing from warehouse distributors in the mid-1970s. MBNA presented the deposition of Mehmet Kara, parts manager of a Mercedes-Benz dealership, who testified that he analyzed the effect on "real price” of MBNA’s obsolescence program, yearly stock-order bonus, and stock allowance and determined that his dealership made a higher profit selling MBNA parts than warehouse distributor parts, despite the higher raw cost of MBNA’s parts. Finally, Rudolf Kraft, dealership owner in Pompano Beach, Florida, testified that although Metrix had a more efficient delivery system, his purchases from Metrix decreased in the late 1970s because of more competitive pricing by MBNA.
. The only procompetitive factors relevant to our analysis, of course, are those that were newly instituted by MBNA during the damage years. Since Heiden’s formula compared purchasing behavior for the pre-damage years with purchasing during the years the tie-in was enforced, factors that remained constant over the period of comparison need not be considered in the damage calculation.
. As previously noted, we held in a prior interlocutory appeal that Metrix’s incentive program violated Section 2(c) of the Robinson-Patman Act. Metrix Warehouse, Inc. v. Daimler-Benz Aktiengesellschaft, 716 F.2d 245 (4th Cir.1983). We described the salient facts of Metrix’s program as follows:
Metrix[’s] ... incentive program involves the awarding of points redeemable for either merchandise or cash or the payment of cash directly to the parts managers of the Mercedes-Benz dealers. These payments are based on a percentage of total parts purchased from Metrix. As consideration for the payments, the payees perform no services other than placing their employers’ purchase orders with Metrix.
Between February 1974 and June 1980, Metrix paid at least $119,980.80 in cash and $394,551.53 in cash and/or merchandise to parts managers of Mercedes-Benz dealers for the placement of approximately $13,000,000 in spare parts orders with Metrix. Payments are mailed monthly by Metrix to the parts managers at their home address. The value of the points is approximately 3!/% of the purchase price.
Id. at 246-47.
. MBNA presented no direct evidence that Metrix’s sales increased because of its program.