Grumman Systems Support Corporation ("Grumman") assigns error to the district court's handling of litigation arising from Grumman's acquisition, duplication, and use of MV/Advanced Diagnostic Executive System ("ADEX"), a sophisticated computer program developed by Data General Corporation ("DG") to diagnose problems in DG's MV computers. DG claimed that Grumman had infringed DG's ADEX copyrights and misappropriated trade secrets embodied in ADEX. A jury agreed, awarding DG $27,417,000 in damages (excluding prejudgment interest and attorney's fees). Grumman contends that the district court prematurely dismissed its affirmative defenses and counterclaims and committed several errors during and after the trial.

While this case raises numerous issues touching on copyright law, Grumman's most intriguing argument--presented below as both a defense and a counterclaim--is that DG illegally maintained its monopoly in the market for service of DG computers by unilaterally refusing to license ADEX to Grumman and other competitors. The antitrust claims are intriguing because they present a curious conflict, namely, whether (and to what extent) the antitrust laws, in the absence of any statutory exemption, must tolerate short-term harm to the competitive process when such harm is caused by the otherwise lawful exercise of an economically potent "monopoly" in a copyrighted work.

After a careful analysis, we affirm on all but one relatively minor issue concerning the calculation of damages.

DG and Grumman are competitors in the market for service of computers manufactured by DG, and the present litigation stems from the evolving nature of their competitive relationship. DG not only designs and manufactures computers, but also offers a line of products and services for the maintenance and repair of DG computers. Although DG has no more than a 5% share of the highly competitive "primary market" for mini-computers, DG occupies approximately 90% of the "aftermarket" for service of DG computers. As a group, various "third party maintainers" ("TPMs") earn roughly 7% of the service revenues; Grumman is the leading TPM with approximately 3% of the available service business. The remaining equipment owners (typically large companies in the high technology industry) generally maintain their own computers and peripherals, although they occasionally need outside service on a "time and materials" basis.

Support service for DG computers entails a variety of activities and a corresponding array of goods and services. The principal activities are maintenance and repair of computer equipment. Maintenance includes care of parts subject to failure as well as replacement of hardware components to bring equipment up to date. Repair involves the diagnosis and correction of hardware failure. Service technicians remedy equipment problems either by actually mending a malfunctioning part (e.g., reformatting a "broken" disk drive) or replacing the part.

Each of these support service "outputs" benefit from a range of "inputs." For example, engineering change orders, along with certain documentation and parts, allow service technicians to make technological updates to computer hardware. In order to identify the existence and location of a malfunctioning part, a service technician may use diagnostics (now increasingly sophisticated software), schematics (maps of the location and function of hardware elements), and various types of documentation, together with the technician's own experience acquired by diagnosing equipment problems. In order to actually mend a malfunctioning part, a technician might fix the part on the spot with routine tools or sophisticated software (e.g., a software diagnostic that can reformat a disk drive), or send the part to a repair depot run either by the technician's employer or another service organization. The repair of a malfunctioning part often requires very detailed information about the part (such as the information provided by schematics and other documentation), and may in turn require smaller replacement parts. Finally, replacement of parts naturally requires the availability of spares. At the core of this litigation is a dispute about Grumman's access to software diagnostics and other service "tools" produced by DG for use in the repair, upgrading, and maintenance of DG equipment.

DG's policies concerning TPM access to DG's service tools have developed over time. As described below, DG's policies have evolved through three stages.

TPMs made their debut in the 1970s while DG was still relatively new to the computer manufacturing market. DG was suspicious of the ability of TPMs, often run and staffed by former DG technicians, to service DG computers without running afoul of DG's intellectual property rights or confidentiality agreements binding on former DG employees.

In 1975, DG converted its suspicions into legal claims, filing suit against Lloyd Root and Robert Montgomery, two of its former employees, as well as Computer Systems Support Corporation ("CSSC"), the TPM that Root and Montgomery had founded after leaving DG.2 DG's principal allegations were that Root and Montgomery had breached their employment agreements by taking DG information with them when they left DG, and that CSSC personnel had been making unauthorized use of DG proprietary information. It was unclear, however, whether the proprietary items that CSSC was using were items sold or licensed to equipment owners (pursuant to agreements which arguably permitted some use by TPMs),3 or items taken directly from DG by Root and Montgomery.

Lacking promising proof to support its claims, DG proposed a settlement whereby CSSC would agree to return any proprietary information that Root and Montgomery unlawfully took from DG, and DG would expressly authorize CSSC (and its successors) to use DG proprietary information in the maintenance and repair of DG computers.4 CSSC accepted, and the parties signed a settlement agreement in 1976 ("the Settlement Agreement").5

From 1976 until some point in the mid-1980s, DG affirmatively encouraged the growth of TPMs with relatively liberal policies concerning TPM access to service tools. DG sold or licensed diagnostics directly to TPMs, and allowed TPMs to use diagnostics sold or licensed to DG equipment owners. DG did not restrict access by TPMs to spare parts manufactured by DG or other manufacturers. DG allowed (or at least tolerated) requests by TPMs for DG's repair depot to fix malfunctioning circuit boards, the heart of a computer's central processing unit ("CPU"). DG sold at least some schematics and other documentation to TPMs. DG also sold TPMs engineering change order kits. And finally, DG training classes were open to TPM field engineers. Grumman suggests that DG's liberal policies were beneficial to DG because increased capacity (and perhaps competition) in the service aftermarket would be a selling point for DG equipment.6

In the mid-1980s, DG altered its strategy. With the goal of maximizing revenues from its service business, DG began to refuse to provide many service tools directly to TPMs. DG would not allow TPMs to use the DG repair depot, nor would it permit TPMs to purchase schematics, documentation, "change order" kits, or certain spare parts. DG no longer allowed TPM technicians to attend DG training classes. Finally, DG developed and severely restricted the licensing of ADEX, a new software diagnostic for its MV computers. The MV series was at once DG's most advanced computer hardware and an increasingly important source of sales and service revenue for DG.

A number of items unavailable to TPMs directly from DG were either available to all equipment owners (even customers of TPMs) from DG, or were available to TPMs from sources other than DG. For example, DG depot service, change order kits, and at least some documentation were available to all equipment owners. There is also evidence that Grumman had its own repair depot and that Grumman could make use of repair depots run by other service organizations (sometimes called "fourth party maintainers"). Likewise, there is evidence that TPMs could purchase at least some spare parts from sources other than DG.

The situation was different with respect to ADEX. DG service technicians would use ADEX in performing service for DG equipment owners. DG would also license ADEX for the exclusive use of the in-house technicians of equipment owners who perform most of their own service.7 However, DG would not license ADEX to its own service customers or to the customers of TPMs. Nor was ADEX available to TPMs from sources other than DG. At least two other diagnostics designed to service DG's MV computers may have become available as early as 1989, but no fully functional substitute was available when this case was tried in 1992.

Grumman found various ways to skirt DG's ADEX restrictions. Some former DG employees, in violation of their employment agreements, brought copies of ADEX when they joined Grumman. In addition, DG field engineers often stored copies of ADEX at the work sites of their service customers, who were bound to preserve the confidentiality of any DG proprietary information in their possession. Although DG service customers had an obligation to return copies of ADEX to DG should they cancel their service agreement and switch to a TPM, few customers did so. It is essentially undisputed that Grumman technicians used and duplicated copies of ADEX left behind by DG field engineers. There is also uncontroverted evidence that Grumman actually acquired copies of ADEX in this manner in order to maintain libraries of diagnostics so that Grumman technicians could freely duplicate and use any copy of ADEX to service any of Grumman's customers with DG's MV computers.

In 1988, DG filed suit against Grumman in the United States District Court for the District of Massachusetts.8 DG patterned its suit after a similar action it brought against Service & Training, Inc. ("STI") in the United States District Court for the District of Maryland. See Service & Training, Inc. v. Data General Corp., 737 F. Supp. 334 (D.Md.1990), aff'd on other grounds, 963 F.2d 680 (4th Cir.1992) ("STI "). STI was another TPM in the DG aftermarket and a successor to Montgomery's interest in the 1976 Settlement Agreement. In one count, DG alleged that Grumman's use and duplication of ADEX infringed DG's ADEX copyrights, and requested injunctive relief, 17 U.S.C. Sec. 502 (1988), as well as actual damages and profits, 17 U.S.C. Sec. 504(b) (1988). In another count, DG alleged that Grumman had violated Massachusetts trade secrets law by misappropriating copies of ADEX in violation of confidentiality agreements binding on former DG employees and DG service customers. On December 29, 1988, the district court issued a preliminary injunction prohibiting Grumman from using ADEX. See Data General Corp. v. Grumman Sys. Support Corp., No. 88-0033-S, 1988 WL 159936 (D.Mass. Dec. 29, 1988) ("Grumman I ").9 The parties then prepared for trial.10

Grumman raised a host of affirmative defenses and counterclaims, all eventually rejected by the district court in response to DG's motions for partial summary judgment. Three of these issues play a pivotal role in Grumman's appeal.

a. 1976 Settlement Agreement

Grumman alleged that the 1976 Settlement Agreement authorized it (as a successor to CSSC) to "acquire, possess, copy and use" all DG diagnostics, including ADEX. Liberally construed, Grumman's allegation of a right to "copy and use" ADEX fairly includes an allegation that Grumman has a right to copy and use DG diagnostic software in the possession of DG equipment owners. Judge Skinner rejected the Settlement Agreement defense by adopting the reasoning of the STI courts, which had rebuffed the same arguments on a nearly identical record. See Grumman V, 834 F. Supp. at 482-83. In the district court decision in STI, Judge Motz analyzed the language of the Settlement Agreement, testimony from the lawyers who negotiated it, and evidence of the parties' subsequent conduct. 737 F. Supp. at 339-41. On the basis of this evidence, Judge Motz concluded that the Settlement Agreement did not require DG to license any proprietary information to CSSC or its customers, nor did the Settlement Agreement prevent DG from prohibiting CSSC from copying and using proprietary information in the custody of DG service customers. Id.11

b. Antitrust Defenses

Grumman also claimed that DG could not maintain its infringement action because DG had used its ADEX copyrights to violate Sections 1 and 2 of the Sherman Antitrust Act, 15 U.S.C. Secs. 1 and 2 (1988 & Supp. IV 1992).12 Specifically, Grumman charged that DG misused its copyrights by (1) tying the availability of ADEX to a consumer's agreement either to purchase DG support services (a "positive tie") or not to purchase support services from TPMs (a "negative tie"), and (2) willfully maintaining its monopoly in the support services aftermarket by imposing the alleged tie-in and refusing to deal with TPMs.

Concerning the tying claim, the district court again adopted the reasoning of the Fourth Circuit in STI, this time for the proposition that there was insufficient proof of a tying agreement to withstand summary judgment. Grumman V, 834 F. Supp. at 484-85. The Fourth Circuit held that there was no positive tie for two independent reasons. First, the court noted that DG did not actually license ADEX to its service customers. STI, 963 F.2d at 686-87. Second, the court held that there was not enough evidence to prove that any license to use ADEX was conditioned on the purchase of DG support services. Id. at 687. The court noted that there was no explicit tying condition in any written agreement. Id. The court also noted that there was insufficient evidence of unwilling purchases of DG support service so as to justify an inference of an implicit condition; customers may simply prefer service supported by ADEX diagnostics over service that is not. Id. at 687-88. The court further held that there was insufficient evidence of a negative tie because, on the record before the court, "[t]he fact that CMOs do not purchase repair services ... is at least as consistent with the legitimate and independent business decision not to purchase unneeded services as it is with an agreement not to purchase such services." Id. at 686.

Judge Skinner conducted his own exhaustive analysis of the monopolization claim, concluding that Grumman failed to "assert[ ] any facts that would indicate that DG has engaged in any unlawful exclusionary conduct." Grumman II, 761 F. Supp. at 192. The court essentially narrowed the question to whether DG's restrictive policies with respect to TPMs constitute unlawful unilateral refusals to deal, reasoning that DG's actions do not rise to the level of unlawful exclusionary conduct for several reasons. The court agreed with Grumman that this case, like Aspen Skiing Co. v. Aspen Highlands Skiing Corp., 472 U.S. 585, 105 S. Ct. 2847, 86 L. Ed. 2d 467 (1985), raises "the issue of prior promotion of competition in a market that is later halted," Grumman II, 761 F. Supp. at 190. The district court nonetheless concluded that Grumman had failed to demonstrate that DG's restrictive policies have unreasonably harmed the competitive process. In particular, the court noted that DG's policies with respect to most service products do not prevent TPMs from competing in the service market because "DG will sell its service products, except [ADEX and schematics], to any ultimate consumer regardless of whether [the consumer] now or later use[s] a TPM." Id. at 191. The court also observed that "TPMs have demonstrated the ability to develop diagnostics [without schematics], even if they are not as efficient as MV/ADEX." Id. Lastly, the court suggested that the Sherman Act would not compel DG to disclose its schematics, in part because such compulsory disclosure would undermine the incentives of copyright and patent laws. Id. at 192.13

In rejecting Grumman's motion for reconsideration of the grant of summary judgment on the monopolization claim, the district court also directly addressed Grumman's contention that DG's refusal to license ADEX to TPMs constitutes exclusionary conduct. The court stated that DG's refusal to license ADEX to TPMs was not exclusionary because "DG offers to the public a license to use MV/ADEX on any computer owned by the customer," and therefore DG " 'did not withhold from one member of the public a service offered to the rest[.]' " Grumman III, slip op. at 5 (citing Olympia Equip. Leasing Co. v. Western Union Tel. Co., 797 F.2d 370, 377 (7th Cir.1986), cert. denied, 480 U.S. 934, 107 S. Ct. 1574, 94 L. Ed. 2d 765 (1987)).

c. Federal Preemption of State Trade Secrets Claim

Grumman unsuccessfully sought to convince the district court that Section 301 of the Copyright Act of 1976, 17 U.S.C. Sec. 301 (1988 & Supp. IV 1992), preempts DG's state law action for misappropriation of trade secrets. The district court held that DG's trade secrets claim was not preempted because DG did not simply allege conduct equivalent to the copying and use which form the basis of an infringement claim; instead, DG's trade secrets claim was based on Grumman's acquisition of ADEX in violation of confidentiality agreements binding on former DG employees and service customers. Grumman IV, 795 F. Supp. at 507.

Stripped of its affirmative defenses, Grumman proceeded to trial. Grumman focused its defensive energies in two areas. Grumman attacked DG's proof of the prima facie elements of copyright infringement and misappropriation of trade secrets, and attempted to undermine DG's broad-gauged request for compensation for lost profits and disgorgement of Grumman's MV-related profits.

a. Validity of Copyright Registration

During the trial, it became evident that DG had made several errors in registering its ADEX copyrights. After Edward Gove, a DG official, testified that DG had deposited with the Copyright Office the correct excerpts of human-readable "source code,"14 Grumman introduced evidence that there were some errors in the deposits for the first three versions of ADEX. In rebuttal testimony, Gove confirmed that there were a number of minor, inadvertent errors in the deposits that would not affect the operation of the programs.15

Grumman argued to the district court that any error in a copyright deposit renders the registration invalid, and requested that the court so instruct the jury. The district court refused, instructing the jury instead that minor, inadvertent errors in the deposit of excerpts of computer code do not threaten the validity of the copyright registration. As a fall-back tactic, Grumman renewed its previous request that the district court compel DG to produce the entire human-readable source code for each version of ADEX so that Grumman could more effectively cross-examine Gove about the significance of the errors. The district court refused to do so, and later explained its discretionary decision by finding that "Grumman had an adequate opportunity to explore the errors contained in the initial copyright deposits, to challenge Data General's explanation of those errors, and to argue these issues before the jury." Grumman VI, 825 F. Supp. at 352. Using a special verdict form, the jury found that DG had properly registered each of the ADEX copyrights.

b. Actual Damages and Profits

Grumman argued that the jury should identify and ignore that portion of Grumman's profits which was not attributable to Grumman's use of ADEX. To this end, Grumman introduced evidence that some of its revenues were derived from servicing DG computers that cannot or need not be serviced with ADEX, and that the value of Grumman's use of ADEX to service customers with MV computers was distinct from the value of other products and services Grumman provided to those customers.

In contrast, DG offered evidence that because equipment owners prefer to purchase all service from one vendor, equipment owners with both MV computers and other DG computers ("mixed-equipment customers") would not have purchased service from Grumman if Grumman had lacked access to ADEX. DG also offered evidence tending to show that, even if Grumman did not always use ADEX in servicing a computer, Grumman could not have attracted and retained its MV-related business had it not been for Grumman's use of ADEX. DG's expert witness opined that DG's damages totaled $28,003,000-$26,364,000 in DG's lost profits and $1,639,000 in nonduplicative profits16 earned by Grumman as a result of its acquisition and use of ADEX.

Attempting to blunt at least part of DG's sweeping "but for" theory, Grumman asked the district court to instruct the jury to discount that portion of Grumman's profits which was not attributable to the infringement. The court instructed the jury that DG could recover that portion of Grumman's profits that was "attributable to the infringement," but did not elaborate on the jury's task in this regard. Left to choose between the parties' theories, the jury apparently accepted the essence of DG's theory, though the total award of compensatory damages was $27,417,000, somewhat less than DG requested.17

Grumman sought relief from the judgment on a number of grounds, two of which are most relevant to this appeal.

a. Actual Damages and Profits

Claiming that the jury's award was speculative and excessive, Grumman moved for a new trial or, in the alternative, remittitur. See Fed.R.Civ.P. 59(a). As the district court related:

Grumman complains that the jury awarded speculative and excessive damages because it uncritically adopted the plaintiff's damage analysis in its entirety which was built on theoretically unsound and factually inaccurate assumptions. More specifically, defendant contends that the plaintiff's damage analysis failed to identify relevant revenues, failed to apply a reasonable profit margin, and failed to apportion service profits between infringing and non-infringing activities.

Grumman VI, 825 F. Supp. at 349 (footnote omitted). The district court denied the motion, ruling in essence that DG's theory of damages was proper and that the jury was free to weigh the testimony of DG's experts more heavily than that of Grumman's experts. Id. at 349-51.

b. Attorney's Fees

The district court included in its judgment order an award of attorney's fees under the Copyright Act, although it appears that the court has not yet fixed the amount. Grumman argued that the court should not award attorney's fees because DG had "elected" only those remedies available under Massachusetts trade secrets law, which does not allow an award of attorney's fees.18 The district court denied the motion, finding that DG had merely sought to maximize the judgment by selecting the most generous body of law for each element of its recovery. Grumman VI, 825 F. Supp. at 346. The district court reasoned further that because DG would not receive a double award of attorney's fees, the judgment was in no need of correction. Id. at 346-47.

Grumman renews its arguments concerning the pretrial, trial, and post-trial issues described above. Grumman claims that the district court erred in entering summary judgment on its affirmative defenses, questions the propriety of certain of the district court's jury instructions, maintains that the jury's award of damages lacks evidentiary support, and insists that DG is not entitled to recover attorney's fees. After reviewing the procedural rules that govern this appeal, we address each of Grumman's arguments in turn.

Because this appeal turns largely on questions of law, we outline the corresponding standard of review. Although the reasoning of the court below may provide a useful starting point for analysis, the district court's view of the law is not binding on a court of appeals. See Williams v. Poulos, 11 F.3d 271, 278 (1st Cir.1993) (citing Dedham Water Co. v. Cumberland Farms Dairy, Inc., 972 F.2d 453, 457 (1st Cir.1992)). Thus, we exercise our independent judgment in evaluating the legal correctness of the district court's jury instructions. Likewise, we must reach our own conclusion as to a statute's correct construction. See FDIC v. Keating, 12 F.3d 314, 316 (1st Cir.1993).

Similarly, in reviewing a district court's entry of summary judgment, we determine anew whether the moving party has shown "that there is no genuine issue as to any material fact and that [it] is entitled to judgment as a matter of law." Fed.R.Civ.P. 56(c). See also Bird v. Centennial Ins. Co., 11 F.3d 228, 231 (1st Cir.1993). "In this context, 'genuine' means that the evidence about the fact is such that a reasonable jury could resolve the point in favor of the nonmoving party and 'material' means that the fact is one that might affect the outcome of the suit under the governing law." Pagano v. Frank, 983 F.2d 343, 347 (1st Cir.1993) (citations, internal quotation marks, and brackets omitted). Although "we read the record and indulge all inferences in a light most favorable to the non-moving party," Rivera-Ruiz v. Gonzalez-Rivera, 983 F.2d 332, 334 (1st Cir.1993), the adverse party cannot defeat a well-supported motion by "rest[ing] upon the mere allegations or denials of [its] pleading," Fed.R.Civ.P. 56(e). If the nonmovant bears the ultimate burden of persuasion with respect to its claim or defense, it may avert summary judgment only if it identifies issues genuinely in dispute and advances convincing theories as to their materiality. See Pagano, 983 F.2d at 347 (citing Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 247-48, 106 S. Ct. 2505, 2509-10, 91 L. Ed. 2d 202 (1986)). Of course, it may be difficult for a trial court to forecast the reaction of a reasonable jury to an intricate array of complex theories. Nonetheless, Rule 56 applies equally to simple cases as well as cases involving complicated legal principles and theories of recovery. See, e.g., Amerinet, Inc. v. Xerox Corp., 972 F.2d 1483, 1490 (8th Cir.1992) ("In complex antitrust cases, no different or heightened standard for the grant of summary judgment applies."), cert. denied, --- U.S. ----, 113 S. Ct. 1048, 122 L. Ed. 2d 356 (1993).

Finally, we note that we are at liberty to affirm a district court's grant of summary judgment " 'on any ground supported in the record even if the issue was not pleaded, tried or otherwise referred to in the proceedings below.' " de Casenave v. United States, 991 F.2d 11, 12 n. 2 (1st Cir.1993).

We first examine the two arguments that strike at the heart of DG's right to pursue its claims: DG's alleged failure to comply with the copyright registration requirements and the possible preemption of the state trade secrets claim by Section 301 of the Copyright Act. We then discuss Grumman's two affirmative defenses--the 1976 Settlement Agreement defense and the "misuse" defense--each of which is intended to undermine both the copyright claim and the trade secrets claim. Finally, we review Grumman's challenges to the award of actual damages, infringer's profits, and attorney's fees.

Registration of a work with the Copyright Office provides several benefits to a plaintiff in an infringement action. First, although copyright protection attaches the day original expression is fixed in a tangible medium, see 17 U.S.C. Sec. 102(a) (1988 & Supp. IV 1992), and thus an infringer may be liable for infringement from that day forward, see 17 U.S.C. Sec. 408(a) (1988 & Supp. IV 1992) (providing that "registration is not a condition of copyright protection"), registration of the copyright is a prerequisite to suit under the Copyright Act, 17 U.S.C. Sec. 411(a) (1988 & Supp. IV 1992). Second, upon accepting the registrant's application, fee, and deposit of a representative copy of the work, see 17 U.S.C. Sec. 408, the Copyright Office issues a certificate of registration, which is admissible in an infringement action as "prima facie evidence of the validity of the copyright and of the facts stated in the certificate," 17 U.S.C. Sec. 410(c) (1988).19 In the case of computer programs which, like ADEX, are either unpublished or published only in machine-readable form, the copyright owner must deposit "identifying portions of the program," generally the first and last 25 pages of the human-readable source code. 37 C.F.R. Sec. 202.20(c)(2)(vii) (1993).20 By questioning DG's compliance with the registration requirements, Grumman is effectively claiming that (1) DG may not claim infringement of those ADEX copyrights for which DG tendered a defective deposit; and (2) even if DG is free to bring such claims, it is not entitled to a presumption as to the validity of the copyrights at issue.

Essentially, Grumman's argument is that the district court erred in instructing the jury that minor, inadvertent errors in material deposited with a registration application do not affect the validity of the registration.21 DG admits that there were inadvertent errors in the material deposited with the registration application for ADEX Revisions 0.0 to 2.0,22 but maintains that the errors are inconsequential.23 Grumman does not quibble with DG's denial of intent, but argues in effect that any error, however minor, precludes a finding that the plaintiff complied with Section 408(b). Alternatively, Grumman argues that an unintentional error in the deposit may still invalidate a copyright registration if the error is material. Grumman contends further that the district court erred in refusing Grumman's request for production of the entire source code for each of the first three versions of ADEX, a decision which allegedly prejudiced Grumman's ability to demonstrate that the defects in the deposit were not minor. We address these contentions seriatim.

a. Immaterial Errors in the Copyright Deposit

It is well established that immaterial, inadvertent errors in an application for copyright registration do not jeopardize the validity of the registration. See Masquerade Novelty, Inc. v. Unique Indus., Inc., 912 F.2d 663, 667-68 & n. 5 (3d Cir.1990); Whimsicality, Inc. v. Rubie's Costume Co., 891 F.2d 452, 456 (2d Cir.1989) (citing Eckes v. Card Prices Update, 736 F.2d 859, 861-62 (2d Cir.1984)); Harris v. Emus Records Corp., 734 F.2d 1329, 1335 (9th Cir.1984); Original Appalachian Artworks, Inc. v. Toy Loft, Inc., 684 F.2d 821, 828 (11th Cir.1982); 2 Nimmer Sec. 7.20, at 7-201 ("[A] misstatement or clerical error in the registration application if unaccompanied by fraud will not invalidate the copyright nor render the registration certificate incapable of supporting an infringement action."). In general, an error is immaterial if its discovery is not likely to have led the Copyright Office to refuse the application. See Eckes, 736 F.2d at 861-62.24

Grumman observes that the cases approving substantial compliance with registration requirements concern errors in the application, not the deposit, and suggests that we adopt a rule demanding strict compliance with the deposit requirement. Although a different rule for deposit errors might be warranted if the language and underlying purposes of the deposit requirement were of a significantly different character than that of the application requirement, we do not find that to be the case.

In the first place, the registration application described in Section 409, as well as the deposit described in Section 408(b), are both equally mandatory components of the registration process outlined in Section 408(a). Likewise, just as Section 409 sets forth what an application "shall include," (emphasis added), Section 408(b) uses the same phrase to prescribe the contents of the deposit. There is nothing in this language that would prevent our interpreting both the application requirements and the deposit requirements in a consistent and practical manner.

Nor do the apparent purposes of the deposit requirement counsel a different result. Although related to the deposit requirement in Section 407, which is designed to further the acquisitions policy of the Library of Congress, the deposit required by Section 408(b) serves the separate purpose of providing the Library's Copyright Office with sufficient material to identify the work in which the registrant claims a copyright. See H.R.Rep. No. 94-1476, 94th Cong., 2d Sess. 5 (1976), reprinted in 1976 U.S.C.C.A.N. 5659, 5766-70; see also 37 C.F.R. Sec. 202.20(c)(2)(vii) (requiring deposit of "identifying portions" of programs that are unpublished or published only in machine-readable form). In other words, a key purpose of the Section 408(b) deposit requirement is to prevent confusion about which work the author is attempting to register.

A second apparent aim of Section 408(b) is to furnish the Copyright Office with an opportunity to assess the copyrightability of the applicant's work. Pursuant to the Copyright Act, the Register of Copyrights must register a copyright claim and issue a registration certificate "[w]hen, after examination, the Register ... determines that ... the material deposited constitutes copyrightable subject matter." 17 U.S.C. Sec. 410(a) (1988).25 Some provisions of the copyright regulations seek to preserve the same opportunity for examination in relation to the deposit of a relatively small subset of a computer program. In adopting regulations encouraging source code deposits for computer programs, the Copyright Office explained that "[i]n registering all copyright claims, the Copyright Office examines the deposit to determine the existence of copyrightable authorship." 54 Fed.Reg. 13,173 (1989). In order to allow the Office to continue this practice, the new regulations provide, for example, that when the applicant's deposit contains portions of the source code of an unpublished computer program with blocked-out trade secrets the deposit must still "reveal[ ] an appreciable amount of original computer code." See 37 C.F.R. Sec. 202.20(c)(2)(vii)(A)(2) (emphasis added). On the other hand, where there are no blocked-out portions in the deposited portions of a computer program, the regulations do not specifically require that the deposit contain "an appreciable amount of original computer code." In other words, the Copyright Office seems to have assumed that in such cases the deposited pages are likely to contain sufficient elements of original expression to determine the copyrightability of the work at issue. At any rate, it appears that Congress viewed the deposit requirement as a means of collecting information that the Copyright Office may use in resolving the question of copyrightability for the purposes of Section 410.26

Neither of these objectives differs so significantly from those of the application requirement as to justify a departure from the rule governing application errors. Quite naturally, one important function of a registration application is to identify the work in which the applicant claims a copyright. See 17 U.S.C. Sec. 409 (1988 & Supp. IV 1992) (requiring application to include, inter alia, title of work, dates of completion and publication, along with "any other information ... bearing upon the ... identification of the work"). Furthermore, like the deposit, the application also provides some evidence of copyrightability, because it must identify any preexisting work from which the author borrowed in creating a compilation or derivative work. See 17 U.S.C. Sec. 409(9). Indeed, the Copyright Office may often be in a better position to assess the originality of the work being registered by reviewing a list of preexisting works than by conducting a cursory inspection of the deposited material. And yet, an inadvertent failure to identify preexisting works on an application is treated no differently from any other application error. See, e.g., Toy Loft, 684 F.2d at 828 (analyzing in similar fashion failure to mention co-author and failure to mention preexisting works).

We conclude that there is no support in law or reason for a rule that penalizes immaterial, inadvertent errors in a copyright deposit.27 Accordingly, we find no flaw in the district court's instruction that such errors "do not impeach the validity and effect of the registration."

b. Material Errors in the Copyright Deposit

The law is not quite as settled as to the effect of an application error that is inadvertent but nonetheless material. No court has suggested that a registration premised in part on an unintentional material error would fail to satisfy the jurisdictional requirement of Section 411(a). At the same time, at least one court has suggested that in such instances the proper approach might be to prevent the plaintiff from exploiting the presumption of validity that ordinarily attaches to a registered copyright under Section 410(c). Masquerade Novelty, 912 F.2d at 668 n. 5 (dictum). We assume for argument's sake that a material error in a copyright deposit, even if unintentional, may destroy the presumption of validity.

c. Refusal to Compel Production of Source Code

Grumman next argues that it was unfairly deprived of an opportunity to prove that the errors in the deposits were material. Specifically, Grumman claims that the district court abused its discretion when, during the trial, it refused to compel DG to produce roughly 40,000 pages of source code (on approximately 33,000 floppy disks) for each of the first three versions of ADEX (0.0 to 2.0). See Geremia v. First Nat'l Bank, 653 F.2d 1, 5-6 (1st Cir.1981) (reviewing denial of mid-trial discovery motion for abuse of discretion).

Grumman renewed its unsuccessful pre-trial requests for the source code after Edward Gove, a DG witness, admitted on cross-examination that there were some discrepancies between the source code deposited with the Copyright Office and the actual source code for ADEX 0.0 to 2.0, and then explained in rebuttal testimony that those errors were minor and of no consequence to the operation of the diagnostic program as a whole. In response to the renewed request, DG provided Grumman with those portions of the source codes for ADEX Revisions 0.0 to 2.0 necessary to conduct a character-by-character comparison of the intended deposits of source code with those portions of source code actually deposited.28 Nonetheless, Grumman insisted that it was entitled to the entire source code for all three versions.

Grumman apparently sought the three sets of source code because it believed that analysis of the entire source code would permit a more effective cross-examination of the DG witness about the magnitude of the discrepancies during DG's rebuttal. It seems that Grumman had one main goal: it believed it might be able to show that, although the discrepancies were few in number and seemingly minor in character, ADEX would not function properly if the source code deposited with the Copyright Office had been inserted into the versions of ADEX DG intended to register.

The marginal benefit to Grumman of obtaining the balance of the source code was at best highly uncertain, and all indications were that such a test would produce no compelling results. Even if Grumman could demonstrate that inserting the errors would impair the operation of ADEX, it is extremely unlikely that this would establish the materiality of the errors. Grumman does not allege that any of the errors, if discovered, would have led the Copyright Office to refuse registration of DG's copyright claims. Nor does Grumman contend that the Copyright Office would have been unable to use the correct portions of the deposits to identify the works DG intended to register or make a preliminary determination concerning the copyrightability of those works.29 In contrast, DG produced evidence that production of the requested material would be an extremely cumbersome process, a point Grumman does not contest. We find no abuse of discretion in the district court's decision to deny Grumman's mid-trial discovery request.

Seeking to avoid the additional damages associated with the trade secrets remedies selected by DG, Grumman argues that the state claim is preempted by Section 301 of the Copyright Act, 17 U.S.C. Sec. 301(a).

Section 301(a) precludes enforcement of any state cause of action which is equivalent in substance to a federal copyright infringement claim.30 See generally Gates Rubber Co. v. Bando Chem. Indus., Ltd., 9 F.3d 823, 846-47 (10th Cir.1993); Trandes Corp. v. Guy F. Atkinson Co., 996 F.2d 655, 658-60 (4th Cir.), cert. denied, --- U.S. ----, 114 S. Ct. 443, 126 L. Ed. 2d 377 (1993); 1 Nimmer Sec. 1.01[B][h], at 1-35 to 1-36.1. Courts have developed a functional test to assess the question of equivalence. "[I]f a state cause of action requires an extra element, beyond mere copying, preparation of derivative works, performance, distribution or display, then the state cause of action is qualitatively different from, and not subsumed within, a copyright infringement claim and federal law will not preempt the state action." Gates Rubber, 9 F.3d at 847 (citing Computer Assocs. Int'l, Inc. v. Altai, Inc., 982 F.2d 693, 716 (2nd Cir.1992)).

Not every "extra element" of a state claim will establish a qualitative variance between the rights protected by federal copyright law and those protected by state law. For example, a state claim of tortious interference with contractual relations may require elements of awareness and intentional interference not necessary for proof of copyright infringement. And yet, such an action is equivalent in substance to a copyright infringement claim where the additional elements merely concern the extent to which authors and their licensees can prohibit unauthorized copying by third parties. Harper & Row, Publishers, Inc. v. Nation Enters., 723 F.2d 195, 201 (2d Cir.1983), rev'd on other grounds, 471 U.S. 539, 105 S. Ct. 2218, 85 L. Ed. 2d 588 (1985). Similarly, a state law misappropriation claim will not escape preemption under Section 301(a) simply because a plaintiff must prove that copying was not only unauthorized but also "commercial[ly] immoral[,]" a mere "label attached to [the same] odious business conduct." Mayer v. Josiah Wedgwood & Sons, Ltd., 601 F. Supp. 1523, 1535 (S.D.N.Y.1985). Nonetheless, a trade secrets claim that requires proof of a breach of a duty of confidentiality stands on a different footing. Such claims are not preempted because participation in the breach of a duty of confidentiality--an element that forms no part of a copyright infringement claim--represents unfair competitive conduct qualitatively different from mere unauthorized copying. See Gates Rubber, 9 F.3d at 847-48; Trandes Corp., 996 F.2d at 660; Computer Associates, 982 F.2d at 717; S.O.S., Inc. v. Payday, Inc., 886 F.2d 1081, 1090 n. 13 (9th Cir.1989).31

DG's trade secrets claim fits comfortably within this category. To demonstrate misappropriation of trade secrets under Massachusetts law, DG must prove that "(1) MV/ADEX is a trade secret; (2) Data General took reasonable steps to preserve the secrecy of MV/ADEX; and (3) Grumman used improper means, in breach of a confidential relationship, to acquire and use the trade secret." Grumman VI, 825 F. Supp. at 357 (citing, inter alia, J.T. Healy & Son, Inc. v. James A. Murphy & Son, Inc., 357 Mass. 728, 260 N.E.2d 723, 729-31 (1970)). The district court instructed the jury that "wrongful acquisition" is an element of a Massachusetts trade secrets claim, and that "[a]cquisition of a trade secret is wrongful ... if it is by theft of property known to belong to another, or by knowing participation in the breach of an express or implied confidentiality agreement by, for instance, a former employee or customer of Data General." (Emphasis added.) Grumman does not assign error to this portion of the charge, which thus becomes the law of the case. See United States v. Connell, 6 F.3d 27, 30 (1st Cir.1993) (explaining that unchallenged legal decisions are ordinarily unassailable at later stages in litigation). Furthermore, DG's theory was precisely that Grumman acquired ADEX by participating in the breach of confidentiality agreements binding on former employees and service customers of DG.32 Because the Copyright Act does not prevent the states from imposing liability for such conduct, the district court was correct to spare DG's trade secrets claim from preemption under Section 301(a).

Grumman denies its liability for copyright infringement and misappropriation of trade secrets, arguing that the Settlement Agreement contains a license allowing Grumman to copy and use ADEX in the maintenance and repair of DG computers. The district court granted DG's motion for partial summary judgment on this issue, and Grumman now appeals that decision on two alternative grounds: (1) the Settlement Agreement unambiguously grants Grumman a license to use ADEX; or (2) the Settlement Agreement is at least ambiguous, and conflicting extrinsic evidence about the scope of the license presents a factual dispute worthy of resolution by a jury.a. Maryland Contract Law

The parties agree that the Settlement Agreement, executed in Maryland, is governed by Maryland contract law. Maryland courts do not follow the subjective theory of contracts, which aims to discover the actual intent of the parties even at the expense of unambiguous language to the contrary. See Hershon v. Gibraltar Bldg. & Loan Ass'n, 864 F.2d 848, 851 (D.C.Cir.1989) (applying Maryland law). Instead, Maryland subscribes to the objective approach. See id. Under that approach, a court may consider extrinsic evidence only in determining whether contract language is ambiguous. See id. at 852. However, as long as the result is objectively reasonable, a court may not use extrinsic evidence "to interpret facially explicit contractual terms." Id. at 851-52. See also General Motors Acceptance Corp. v. Daniels, 303 Md. 254, 492 A.2d 1306, 1310 (1985).

Where contract terms are ambiguous, a court may look to extrinsic evidence in order to ascertain the intention of the parties and, if successful, interpret the contract as a matter of law. See Collier v. MD-Individual Practice Ass'n, 327 Md. 1, 607 A.2d 537, 541 (1992); Truck Ins. Exch. v. Marks Rentals, Inc., 288 Md. 428, 418 A.2d 1187, 1190 (1980). If, after such examination, the meaning of the ambiguous terms remains in genuine dispute, and the dispute is material to the outcome of the claim or defense at issue, the ambiguity must be resolved by the trier of fact. See id.; Monumental Life Ins. Co. v. United States Fidelity & Guar. Co., 94 Md.App. 505, 617 A.2d 1163, 1174 ("Only when there is a bona fide ambiguity in the contract's language or legitimate doubt as to its application under the circumstances is the contract submitted to the trier of the fact for interpretation."), cert. denied, 330 Md. 319, 624 A.2d 491 (1993).33

b. Areas of Agreement

In order to focus our analysis of DG's entitlement to summary judgment, we first determine the reach of Grumman's contentions in light of the existing areas of agreement.

In the first place, the parties agree that the existence and scope of a license turn on the interpretation of the "maintenance or repair" exception to the general prohibition of paragraph four of the Settlement Agreement, which provides that Grumman's predecessor "will not, directly or indirectly, copy or utilize 'Proprietary Information' of DG for the design or manufacture of computers or any other purpose."34 In addition, DG admits that the Settlement Agreement gives Grumman a right to use some of DG's proprietary information for some purposes. Although DG denies that the Settlement Agreement allows Grumman to use ADEX itself, DG nonetheless admitted in its answers to Grumman's request for admissions "that, as part of the settlement of the CSSC litigation, Data General agreed that CSSC could use Data General proprietary information that was defined in the Agreement and the nature of which was then understood by and agreed to by the parties, to maintain or repair Data General computers." While the Settlement Agreement does not bear many of the traits of a traditional licensing agreement, it does grant some permission to use DG's intellectual property, at least in certain circumstances, and therefore creates some type of "license."35 Consequently, Grumman's defense turns on the scope of the license.36

Also worthy of note are undisputed facts concerning the nature of Grumman's acquisition and use of ADEX. Grumman did not simply gain access to copies of ADEX left at the sites of former DG service customers solely for the purpose of using on-site maintenance tools to service computers at that site. Rather, Grumman acquired copies of ADEX from former service customers in an effort to expand its own library of MV diagnostic software, which Grumman technicians freely copied and used in servicing the computers of any Grumman customer with MV equipment. Moreover, there is no evidence that Grumman acquired ADEX from equipment owners at a time when those equipment owners were also customers of DG. Nor is there evidence that Grumman acquired ADEX directly from DG, or from current or former CMO customers. In addition, the record reveals that DG service customers were contractually bound both to prevent ADEX from falling into the hands of third parties such as TPMs and to return copies of ADEX to DG after the termination of the relevant service agreement.37 Thus, Grumman acquired ADEX from those customers who no longer had lawful possession of the program and had no right to transfer it.38

The question we must resolve is whether the "maintenance or repair" exception authorized Grumman both to gain access to and acquire copies of ADEX in the possession of former DG service customers despite the fact that these customers had agreed not only to prevent such third-party access but also to return copies of ADEX to DG after the termination of their service contract.

c. Scope of the License

The plain language of the Settlement Agreement does not answer our question. Despite the fact that the exception anticipates that Grumman will "copy or utilize" DG proprietary information for the "maintenance and repair of DGC equipment," the Settlement Agreement does not specify whether it merely refers to Grumman's right to gain access to maintenance tools it finds at a customer site (including the routine copying and use inherent in the operation of a computer program), or whether the exception somehow allows Grumman to acquire such tools for the service of DG computers at other sites. Similarly, the Settlement Agreement contains no prescription for resolving potential conflicts between the "maintenance or repair" exception and provisions in DG's Service Agreement prohibiting third-party access during the term of the Agreement and retention of DG proprietary information thereafter. Accordingly, we turn to the extrinsic evidence in the record in an attempt to resolve the ambiguity.

Even when viewed in a light most favorable to Grumman, the record evidence makes clear that the parties to the Settlement Agreement intended the "maintenance and repair" exception to function as what we shall call a "third-party access agreement," allowing CSSC, Grumman's predecessor in interest, to gain access to proprietary information that DG sold, licensed, or otherwise entrusted to owners of DG equipment. For example, when called to testify in the STI litigation, Edward Canfield, CSSC's attorney at the time, used these words to describe his contemporary understanding of the "maintenance and repair" exception: "If the customer had it, [CSSC] had a right to use it."39 In addition, the language of DG licensing agreements in the 1970s as well as the pleadings in the 1975 litigation strongly corroborate the view that the settlement negotiations primarily concerned CSSC's right to use proprietary information in the hands of DG equipment owners. As late as 1976, DG licensed proprietary maintenance information to equipment owners under an agreement which specifically allowed licensees to grant access to third parties "on LICENSEE's premises with LICENSEE's permission for purposes specifically related to LICENSEE's use of the Licensed Program." Moreover, in its 1975 counterclaim, CSSC intimated that DG had begun to undermine the ability of TPMs to gain access to maintenance information in the hands of equipment owners, alleging that DG had attempted "to prevent owners of DGC Mini-computers from having their equipment serviced and maintained by any competitor of DGC ... by restricting the use those owners make of their owner maintenance information."

There is also specific evidence that the parties to the Settlement Agreement were not negotiating about the ongoing transfer of proprietary information directly from DG to CSSC. For example, during the STI trial, counsel for DG asked Canfield whether, under the Settlement Agreement, DG had an "obligation to give [CSSC] something." "No sir," replied Canfield, "Data General was not offering to give us anything."

The nature of the "maintenance and repair" exception as a third-party access agreement has several ramifications. As a provision designed to ensure access to Grumman, the exception was arguably intended to override contrary restrictions in proprietary legends and confidentiality agreements. Indeed, there is evidence that this was the case. A letter to Canfield from Carl Kaplan, a lawyer who represented DG in the settlement negotiations, outlined the proposed settlement, stating that improper utilization of DG proprietary information "would be the use of that information other than as marked by DGC or without DGC's express written permission." (Emphasis added.) Kaplan added that "[u]se of DGC proprietary information for the maintenance of DGC equipment would expressly be permitted the defendants." Id. In addition, Canfield's deposition testimony suggests that his primary concern was for DG to guarantee CSSC's right to use proprietary information distributed to DG equipment owners, notwithstanding future restrictions on third-party access to such information. Thus, a jury could reasonably conclude that the Settlement Agreement allowed Grumman to gain access to information in the hands of DG equipment owners for the purpose of maintaining DG computers, even if equipment owners generally could not allow third parties access to DG proprietary information.

Characterizing the exception as a third-party access agreement also means that Grumman's right to use copies of ADEX in the possession of DG equipment owners is necessarily derivative of the rights of those equipment owners. As a consequence, Grumman only has the right to operate a customer's copy of ADEX for the benefit of that customer; there is no basis for the proposition that Grumman can use its third-party access rights to acquire copies of ADEX for unlimited copying and use in the service of any MV computer. Indeed, this was the import of Canfield's testimony in STI. Referring to a CSSC customer as a "party," Canfield stated that he understood the Settlement Agreement to allow "[CSSC] to use whatever [was] on the party's equipment ... for the repair and maintenance of that party's equipment." (Emphasis added).40 Furthermore, to the extent that an equipment owner no longer has the right to possess copies of ADEX, as in the case of a former DG service customer, Grumman's rights as a third party are extinguished.41

In summary, we conclude that the Settlement Agreement merely grants Grumman the right to gain access to copies of ADEX lawfully possessed by a DG equipment owner in order to service the computers of that particular equipment owner. Because Grumman's copying and use of ADEX does not fall within this category, the Settlement Agreement does not serve as a defense either to the infringement action or the trade secrets claims. The district court did not err in granting partial summary judgment for DG on Grumman's Settlement Agreement defense.

Grumman claims that DG is not entitled to enforce its copyrights or its rights under state trade secrets law because it has "misused" those property rights by engaging in anti-competitive behavior in violation of federal antitrust laws. DG argues that there is no "copyright misuse" defense to a federal copyright infringement claim and no applicable "unclean hands" defense to the state claim for misappropriation of trade secrets. Alternatively, DG argues that it did not violate the antitrust laws.

A "copyright misuse" defense is not without legal support. In a carefully reasoned opinion, the Fourth Circuit recently approved such a defense after noting that it has long been recognized in the analogous context of patent infringement. See Lasercomb America, Inc. v. Reynolds, 911 F.2d 970, 976 (4th Cir.1990) ("[S]ince copyright and patent law serve parallel public interests, a 'misuse' defense should apply to infringement actions brought to vindicate either right."); see also 3 Nimmer Sec. 13.09[A], at 13-269 to 13-276 (collecting conflicting decisions of lower courts); Ramsey Hanna, Note, Misusing Antitrust: The Search for Functional Copyright Misuse Standards, 46 Stan.L.Rev. 361, 404-10 (1994) (charting the development of the copyright misuse defense). Although DG correctly notes that the misuse in Lasercomb (conditioning a copyright license on a noncompetition agreement) is not identical to the misuse alleged in this case (tying access to ADEX to the purchase of DG service and refusing to license ADEX to TPMs), the reasoning of Lasercomb does not turn on the particular type of anti-competitive behavior alleged. DG also suggests that the policy rationale for a copyright misuse defense is weaker than in the case of patent misuse because an exclusive right to express an idea in a particular way (a copyright) is a lesser threat to competition than an exclusive right to use the idea itself (a patent). We acknowledge that it is often more difficult to prove an antitrust violation when the claim rests on the questionable market power associated with a copyright, but that would not be a reason to prohibit a defendant from attempting to meet its burden of proof, and would be a poor reason to refrain entirely from recognizing a copyright misuse defense.

Nevertheless, this case does not require us to decide whether the federal copyright law permits a misuse defense. Nor need we determine whether Massachusetts recognizes an unclean hands defense to a claim for misappropriation of trade secrets. Grumman does not claim that DG misused its copyright or acted inequitably in any fashion other than through its alleged violations of the Sherman Act.42 And, because we conclude infra, Section III.B., that there is insufficient evidence to justify a trial on either of Grumman's antitrust counterclaims, Grumman's misuse and unclean hands defenses are equally devoid of merit.43

Grumman's principal assault on the jury's award of $27,417,000 in damages (DG's lost profits and Grumman's nonduplicative profits) is that the district court failed to give the jury adequate guidance to find the necessary causal connection between Grumman's infringement and DG's damages. Because the calculus of causation is partly a function of the particular theory of damages advanced by the plaintiff, we divide our discussion accordingly.

a. Actual Damages

A successful plaintiff in an infringement action is entitled to "actual damages suffered by [it] as result of the infringement." 17 U.S.C. Sec. 504(b). Actual damages are generally calculated with reference to the loss in the fair market value of the copyright, often measured by the profits lost as a result of the infringement. See, e.g., Eales v. Envtl. Lifestyles, Inc., 958 F.2d 876, 880 (9th Cir.), cert. denied, --- U.S. ----, 113 S. Ct. 605, 121 L. Ed. 2d 541 (1992); see generally 3 Nimmer Sec. 14.02[A], at 14-8 to 14-9.

The plaintiff bears the burden of proving that the infringement was the cause of its loss of revenue. See Harper & Row, Publishers, Inc. v. Nation Enters., 471 U.S. 539, 567, 105 S. Ct. 2218, 2234, 85 L. Ed. 2d 588 (1985); Frank Music Corp. v. MGM, Inc., 772 F.2d 505, 514 n. 8 (9th Cir.1985) (citing Shapiro, Bernstein & Co. v. 4636 S. Vermont Ave., Inc., 367 F.2d 236, 241 (9th Cir.1966)). In defining that burden, it is useful to borrow familiar tort law principles of causation and damages. See Deltak, Inc. v. Advanced Sys., Inc., 574 F. Supp. 400, 403 (N.D.Ill.1983) (Posner, J., sitting by designation) (referring to "normal tort damages principles" in discussion of copyright damages), vacated on other grounds, 767 F.2d 357 (7th Cir.1985); 3 Nimmer Sec. 14.02[A], at 14-11, 14-20 to 14-21 n. 49.8 (alluding to notions of "but for" and proximate causation). Thus, the plaintiff should first establish that the infringement was the cause-in-fact of its loss by showing with reasonable probability that, but for the defendant's infringement, the plaintiff would not have suffered the loss. See, e.g., Robert R. Jones Assocs. v. Nino Homes, 858 F.2d 274, 281 (6th Cir.1988); 3 Nimmer, Sec. 14.02[A], at 14-9; cf. Harper & Row, 471 U.S. at 567, 105 S.Ct. at 2234 (noting that in rebuttal defendant may "show that this damage would have occurred [anyway] had there been no taking of copyrighted expression"); Aro Mfg. Co. v. Convertible Top Replacement Co., 377 U.S. 476, 507, 84 S. Ct. 1526, 1543, 12 L. Ed. 2d 457 (1964) (noting that actual damages in patent infringement case are based on "what [the patent holder's] condition would have been if the infringement had not occurred") (citation and internal quotation marks omitted). The plaintiff must also prove that the infringement was a proximate cause of its loss by demonstrating that the existence and amount of the loss was a natural and probable consequence of the infringement. See Big Seven Music Corp. v. Lennon, 554 F.2d 504, 509 (2d Cir.1977) ("[D]amages may be recovered only if there is a necessary, immediate and direct causal connection between the wrongdoing and the damages."). A plaintiff may seek compensation for both direct and "indirect" losses, as long as the losses claimed are not unduly speculative. See Business Trends Analysts, Inc. v. Freedonia Group, Inc., 887 F.2d 399, 404 (2d Cir.1989) (recognizing possibility of recovery for loss of "enhanced good will" and "market recognition"); Abeshouse v. Ultragraphics, Inc., 754 F.2d 467, 471 (2d Cir.1985) (ruling that claimed harm to "reputation" and "marketability" of copyrighted poster was "too speculative to support any award of actual damages"); Sunset Lamp Corp. v. Alsy Corp., 749 F. Supp. 520, 524-25 (S.D.N.Y.1990) (recognizing possibility of recovery for lost sales of noninfringed items); 3 Nimmer Sec. 14.02[A], at 14-11 to 14-21. At the same time, the plaintiff need not prove its loss of revenue with mathematical precision. See, e.g., Stevens Linen Assocs. v. Mastercraft Corp., 656 F.2d 11, 14 (2d Cir.1981) ("In establishing lost sales due to sales of an infringing product, courts must necessarily engage in some degree of speculation.").

DG argued at trial that ADEX capability was essential both to service MV computers and attract customers, and therefore nearly all of Grumman's MV customers would have remained with DG (or would have switched back to DG) had Grumman not touted its possession and use of ADEX. In opposition, Grumman introduced evidence that ADEX was of little use to Grumman's field engineers and only a minor factor in consumer's selection of a service vendor. In effect, Grumman argued that, even without ADEX, customers would have switched to (or remained with) Grumman in order to take advantage of its lower prices and allegedly higher-quality service.

In its objections to the jury charge, Grumman expressed concerns about the court's instructions on causation in the lost profits context. Grumman asked the court to instruct the jury that it was free to consider whether factors other than Grumman's infringement enabled Grumman to win customers from DG. On appeal, Grumman continues to challenge the adequacy of the district court's instructions on causation, and raises several questions about the sufficiency of the evidence.

(1) Jury Instructions

The district court's charge, relevant portions of which are set forth in the margin, invited the jury to consider the "diverse factors" that make up a customer's choice of a service organization, and properly allowed the jury to consider whether the majority of MV equipment owners would have turned to DG for service had Grumman not possessed and used ADEX.44 The instructions also introduced the jury to the concept of proximate cause. The charge not only mentioned the concept by name but also gave it content by explaining, among other things, that the plaintiff "bears the burden of proving its damages to a reasonable degree of certainty," may not be compensated for "purely speculative" damages, and is entitled only to "reasonable" damages. We conclude that the charge adequately equipped the jury to determine whether or not DG had established the requisite causal link between Grumman's infringement and the profits DG claimed to have lost.45

(2) Sufficiency of the Evidence

Grumman's challenge to the evidentiary basis for the jury's award of actual damages is less developed and equally unavailing. Upsetting a jury's damage award is a daunting task for any appellant, for we must draw all reasonable inferences in favor of the verdict, upholding the award if it derives from "any rational appraisal or estimate of the damages that could be based on the evidence before the jury." Anthony v. G.M.D. Airline Servs., 17 F.3d 490, 493 (1st Cir.1994) (citations and internal quotation marks omitted). The likelihood of a victorious appeal is especially remote in the absence of rigorous argumentation. Cf. Chakrabarti v. Cohen, 31 F.3d 1, 4 (1st Cir.1994) (suggesting that, when challenging the sufficiency of the evidence, a defendant-appellant must make a "serious effort ... to analyze the evidence taking it in the light most favorable to [the plaintiff] and resolving credibility issues in [the plaintiff's] favor"). Moreover, the calculation of lost profits will always involve "some degree of speculation." Stevens Linen Assocs., 656 F.2d at 14. As a result, we rely on the appellant to specify with some precision the manner in which unduly speculative reasoning is likely to have infected the jury's verdict.

Grumman raises several specific concerns. First, Grumman complains that the jury had no basis to conclude that Grumman would not be in the MV business because DG's damage expert, Alan Friedman, did not consider the relative infrequency of Grumman's use of ADEX, or the value Grumman added to its product through "substantially lower prices, superior service and higher level of customer satisfaction." Grumman's ultimate concern is that "no attempt at apportionment was made." But Friedman did not set out to show that there was nothing attractive about Grumman service apart from its possession and use of ADEX. Instead, he reported--and the jury apparently believed--that, for most owners of MV equipment, ADEX capability was the critical attribute in a service vendor. As a result, Friedman concluded, ADEX capability was the sine qua non of Grumman's success in its chosen niche as a national vendor of MV service. Drawing all reasonable inferences in favor of DG, we conclude that a reasonable jury was free to agree.

Grumman also argues that the jury must have improperly followed Friedman's lead in adding to the lost profits figure all of the service and hardware needs of Grumman's MV customers that DG was capable of filling. Grumman notes that Friedman based his testimony on evidence that customers prefer to have a single vendor of service, but claims that this evidence deserves little weight because "customers that had gone to [Grumman] had already demonstrated their particular price/service sensitivity." Viewed in a light most favorable to the verdict, however, the record evidence adequately supports the inference that Friedman invited the jury to draw. For example, while MV equipment owners may have switched to Grumman in search of lower prices and better service, the evidence suggests that none of them had to give up a preference for single sourcing to do so. Indeed, the evidence suggests that Grumman's drawing power was due in part to its ability to be a single source of service, particularly for customers with multiple brands of computer equipment. Nor did Grumman attempt to rebut Friedman's view with evidence of MV equipment owners who sacrificed their preference for single sourcing in certain circumstances.46 More importantly, Friedman did not presume that customers would be entirely insensitive to issues of price and quality. In calculating DG's lost profits, he reduced the figure by an estimate of the business DG would itself have lost to competition from TPMs.47 Finally, we note that DG did not seek compensation for a loss in "goodwill" or "market recognition" that was difficult to ascertain, cf. Business Trends, 887 F.2d at 404, but rather for the loss of a reasonably verifiable number of customers with a limited and predictable set of service needs and a demonstrated tendency to satisfy those needs by turning to a single vendor. In short, the evidence does not suggest that the jury's award of actual damages falls outside the " 'wide range of arguable appropriateness.' " Toucet v. Maritime Overseas Corp., 991 F.2d 5, 11 (1st Cir.1993) (quoting Wagenmann v. Adams, 829 F.2d 196, 216 (1st Cir.1987)).

b. Infringer's Profits

In addition to actual damages, a copyright plaintiff may also recover the infringer's nonduplicative profits, i.e., "any profits of the infringer that are attributable to the infringement and are not taken into account in computing the actual damages." 17 U.S.C. Sec. 504(b). In the context of infringer's profits, the plaintiff must meet only a minimal burden of proof in order to trigger a rebuttable presumption that the defendant's revenues are entirely attributable to the infringement; the burden then shifts to the defendant to demonstrate what portion of its revenues represent profits, and what portion of its profits are not traceable to the infringement. See id.; Frank Music, 772 F.2d at 514; Cream Records, Inc. v. Jos. Schlitz Brewing Co., 754 F.2d 826, 828 (9th Cir.1985). Specifically, Section 504(b) provides:

In establishing the infringer's profits, the copyright owner is required to present proof only of the infringer's gross revenue, and the infringer is required to prove his or her deductible expenses and the elements of profit attributable to factors other than the copyrighted work.48

DG introduced evidence that, of Grumman's gross revenue from MV-related business during the period 1984 to 1990, $5.4 million consisted of business eliminated from the calculation of DG's lost profits. Although no further proof was required, DG accepted Grumman's estimates of its profit margin, and concluded that Grumman's nonduplicative profits amounted to approximately $1.6 million.49 Anticipating Grumman's attempt to prove the need for apportionment, DG also argued that, without ADEX, Grumman would not have been in the MV service business on a national scale, and that therefore Grumman would not have earned the remainder of its MV-related profits. In other words, DG's theory was that because ADEX capability was generally essential to attract customers with MV computers, few such customers would have chosen Grumman as a service vendor, causing Grumman to leave (or perhaps never enter) the national market for service of MV-related equipment. Thus, according to DG, Grumman's nonduplicative profits were the indirect result of consumer choices distorted by Grumman's infringement.

It is unclear whether Grumman contested DG's theory on the merits, although Grumman did introduce some expert testimony that owners of MV equipment were relatively indifferent to the ADEX issue in their choice of service vendors. As amplified by its arguments on appeal, however, Grumman's primary strategy was to invite the jury to take Grumman's infringement as a given, and focus instead on why its customers were willing to pay for Grumman service. Grumman argued that factors other than its possession and use of ADEX contributed to its customers' willingness to pay, and that it was entitled to retain a corresponding share of the resulting profits. Grumman introduced some evidence tending to show that its customers attached high value to the price and quality of Grumman service, as well as Grumman's ability to service non-DG equipment in a mixed-equipment system.

Recognizing that DG's "but for" theory focused on a different aspect of consumer behavior than Grumman's "contributing factors" theory, Grumman argued below that the court's instructions should leave the jury free to adopt either line of reasoning. Grumman's suggested method of doing so was for the court to instruct the jury on the concept of apportionment of infringer's profits set forth in Section 504(b). The district court agreed that the jury could adopt the approach best suited to the circumstances, but refused to give an explicit instruction on apportionment.

Assuming for the moment that Grumman was entitled to invite the jury to adopt its analytical framework, we do not believe that the court's instruction "properly apprise[d]" the jury of the validity of such an approach. Joia v. Jo-Ja Serv. Corp., 817 F.2d 908, 912 (1st Cir.1987), cert. denied, 484 U.S. 1008, 108 S. Ct. 703, 98 L. Ed. 2d 654 (1988). Although the district court instructed the jury to include among infringer's profits only those revenues "attributable to the infringement," at no point did the court fully reveal or explain the relatively difficult statutory concept of "elements of profit attributable to factors other than the copyrighted work." 17 U.S.C. Sec. 504(b). As noted in the preceding section, the court did refer (at least in its instruction on actual damages) to "diverse factors" which might have influenced customers' choice of Grumman over DG, but the court did not inform the jury that there may have been many reasons for customers' willingness to pay for Grumman service apart from the fact that Grumman possessed and used ADEX. Cf. Walker v. Forbes, Inc., 28 F.3d 409, 416 (4th Cir.1994) (praising district court's "rich and detailed instructions ... explaining ... the correct apportionment of profit attributable to the infringement, [and] faithfully explaining the rules and procedures set out in the statute").

It is unclear why, if the district court chose to reject Grumman's proposed instruction, it did not simply read to the jury the language of Section 504(b). We may overlook its failure to do so only if there is no basis in law or fact for the application of Grumman's theory. See Joia, 817 F.2d at 912 (holding that "all parties are entitled to an adequate jury instruction upon the controlling issues"); cf. Allen v. Chance Mfg. Co., 873 F.2d 465, 470 (1st Cir.1989) (holding that remand on basis of instructional error is required only if error "may have unfairly affected the jury's conclusions"). For the reasons set forth below, we believe that Grumman's theory is firmly rooted in the law of copyright and the record of this case.

The defendant's burden under the apportionment provision of Section 504(b) is primarily to demonstrate the absence of a causal link between the infringement and all or part of the profits claimed by the plaintiff. See Walker, 28 F.3d at 412 (describing Section 504(b) as "a rule of causation"). Because the rebuttable presumption of causation represents a presumption as to both cause-in-fact and proximate cause, there are two avenues of attack available to a copyright defendant. First, the defendant can attempt to show that consumers would have purchased its product even without the infringing element. See, e.g., id. at 413 (holding that district court properly allowed the defendant to show that an unauthorized reproduction of a photograph in an issue of its magazine had no causal relation to "amounts of revenue ... committed to the issue sight unseen").50 Alternatively, the defendant may show that the existence and amount of its profits are not the natural and probable consequences of the infringement alone, but are also the result of other factors which either add intrinsic value to the product or have independent promotional value. See, e.g., Sheldon v. Metro-Goldwyn Pictures Corp., 309 U.S. 390, 407-08, 60 S. Ct. 681, 687-88, 84 L. Ed. 825 (1940) (approving apportionment where profits of defendant's film were largely attributable not to the plaintiff's pirated story but rather to the "drawing power" of the star performers and the artistry of others involved in the creation of the film); Abend v. MCA, Inc., 863 F.2d 1465, 1480 (9th Cir.1988) (remanding for apportionment where factors other than the underlying story--particularly the talent and popularity of Alfred Hitchcock, Jimmy Stewart, and Grace Kelly--"clearly contributed" to the success of the film "Rear Window"), aff'd on other grounds, 495 U.S. 207, 110 S. Ct. 1750, 109 L. Ed. 2d 184 (1990); Sygma Photo News, Inc. v. High Soc'y Magazine, Inc., 778 F.2d 89, 96 (2d Cir.1985) (apportioning profits from sales of "Celebrity Skin" magazine where promotional cover contained not only infringing photograph of Raquel Welch but also a list of other nude celebrity photographs contained within); Cream Records, 754 F.2d at 828-29 (upholding apportionment of profits from malt liquor sales apparently based on popularity of noninfringing product and promotional value of noninfringing elements of defendant's commercial); cf. USM Corp., 467 N.E.2d at 1277 (trade secrets; recognizing that apportionment would have been proper if defendant had demonstrated that factors such as "management skill" or "capital investment" had contributed to the success of its product). Grumman apparently wished to tread the second path, and it was unquestionably entitled to do so.

Grumman also suggests on appeal that the jury should have been instructed that it could not accept DG's theory on the apportionment issue because DG gave little or no weight to Grumman's contributions. But the only argument presented to the district court was that the court should add an instruction to inform the jury that it was permitted to apportion Grumman's profits. It is usually imprudent for a court of appeals to pass on an issue not presented to the district court in the first instance, and we decline to do so in these circumstances. See, e.g., Mariani v. Doctors Assocs., 983 F.2d 5, 8 n. 4 (1st Cir.1993) ("We have repeatedly warned that we will not entertain arguments made for the first time on appeal.") (citing FDIC v. World Univ., Inc., 978 F.2d 10, 13 (1st Cir.1992)); United States v. Zannino, 895 F.2d 1, 17 (1st Cir.) ("[A] litigant has an obligation to spell out its arguments squarely and distinctly, or else forever hold its peace.") (citations and internal quotation marks omitted), cert. denied, 494 U.S. 1082, 110 S. Ct. 1814, 108 L. Ed. 2d 944 (1990).

We are compelled to add, however, that an instruction on apportionment would not rob DG's theory of all possible meaning. In the first place, DG was free to argue that Grumman's infringement was a "but for" cause of Grumman's nonduplicative profits, even if the court should have explained to the jury that Grumman could still satisfy its burden by demonstrating the absence of proximate causation. In addition, DG was entitled to argue that Grumman's infringement should be viewed as the sole or overriding cause of Grumman's profits. Cf. Frank Music, 772 F.2d at 518 (noting that "no one element was the sole or overriding reason" for the success of defendant's infringing "Hallelujah Hollywood" stage show).

Moreover, although apportionment primarily depends on questions of causation, it is ultimately a delicate exercise informed by considerations of fairness and public policy, as well as fact. The doctrine of apportionment was "established upon equitable principles" in the analogous context of patent infringement. Sheldon, 309 U.S. at 401, 60 S.Ct. at 684. And, in adopting the principle of apportionment for copyright cases, the Court observed that "[e]quity is concerned with making a fair apportionment so that neither party will have what justly belongs to the other." Id. at 408, 60 S.Ct. at 688 (emphasis added). See also 3 Nimmer Sec. 14.03[C], at 14-42 (noting that Copyright Act of 1976 "expressly adopted" the apportionment principle announced in Sheldon ). In fact, the burden-shifting rule in Sheldon (and Section 504(b)) is itself an equitable response to an infringer who has frustrated the task of apportionment by co-mingling profits. See Sheldon, 309 U.S. at 401, 60 S.Ct. at 685 ("[T]he defendant, being responsible for the blending of the lawful with the unlawful, had to abide the consequences, as in the case of one who has wrongfully produced a confusion of goods.") (referring to Callaghan v. Myers, 128 U.S. 617, 9 S. Ct. 177, 32 L. Ed. 547 (1888)). Equitable factors may also affect the substance of the apportionment analysis. For example, where the plaintiff cannot prove actual damages and the defendant's profits are only from the sale of a noninfringing product, the only way to prevent unjust enrichment may be to place more weight on the profit-generating effect of an infringing sales tool used to promote that product. See, e.g., Konor Enters. v. Eagle Publications, Inc., 878 F.2d 138, 140 (4th Cir.1989) (suggesting that defendant may not be entitled to retain any of the profits from sale of advertising space where it is "plausible ... that all profits were a direct result" of infringing marketing information distributed to potential advertisers).

Similarly, the policies underlying the Copyright Act may play some role in the apportionment of profits. For example, Sheldon and its progeny suggest that apportionment is almost always available in the context of infringing derivative works, perhaps in part because original expression added by the infringer is itself entitled to copyright protection. Furthermore, where the plaintiff is seeking to vindicate its right to exclude others rather than its right to collect a licensing fee, see 17 U.S.C. Sec. 106 (1988 & Supp. IV 1992) (describing rights of copyright owner), it may be more appropriate to view the infringement as an "overriding" cause of the defendant's profits. In such cases, rigid isolation of the value of the infringement to the defendant (which would approximate a "reasonable" licensing fee) would effectively condone a license the plaintiff never wished to grant. Lastly, we note that an unjust enrichment theory aims to strip the defendant of its ill-gotten gains, see, e.g., 3 Nimmer Sec. 14.01[A], at 14-6, encourage compliance with the Copyright Act, see, e.g., Walker, 28 F.3d at 412 (noting that an award of infringer's profits "makes the infringer realize that it is cheaper to buy than to steal"), and perhaps "compensate" a plaintiff unable to prove actual damages, see Sheldon, 309 U.S. at 399, 60 S.Ct. at 684 (describing the goal of an award of infringer's profits as "just compensation for the wrong"). Therefore, apportionment of infringer's profits may be particularly appropriate where a concurrent award of actual damages significantly serves all three purposes.51

In light of the discussion above, we hold that Grumman was entitled to an instruction on apportionment in order to allow the jury to determine whether and to what extent apportionment of its nonduplicative profits was reasonable under the circumstances of this case.

Whether a remand is necessary is a different question, but one readily resolved. Grumman clearly introduced evidence that would have permitted a jury to find that Grumman's customers were willing to pay for Grumman service for reasons beyond its possession and use of ADEX. Indeed, we believe that Grumman's evidence is sufficiently compelling that Grumman is entitled to some apportionment as a matter of law. Because the absence of an explicit instruction on apportionment "may have unfairly affected the jury's conclusions," Allen, 873 F.2d at 470, we remand the case to the district court for an appropriate resolution of the issue of apportionment of Grumman's nonduplicative profits.52

Because it appears that an award of attorney's fees has not been quantified, see Grumman VII, 825 F. Supp. at 370 (ordering DG to resubmit its application for attorney's fees), the merits of such an award are not before this court. Nonetheless, Grumman mounts a procedural attack that does appear to be ripe for review. Grumman claims that (1) DG "elected" the state trade secrets remedy over any remedy available under the Copyright Act, and (2) since attorney's fees are only available under the Copyright Act, and not state trade secrets law, DG is not entitled to any attorney's fees. Grumman is wrong in both respects. DG did not simply elect state law remedies. DG proposed a judgment form, wholly adopted by the district court, that included (1) the compensatory damages awarded by the jury,53 (2) state law statutory damages, (3) state law prejudgment interest, and (4) federal law attorney's fees. Nor was DG required to forsake nonduplicative elements of the various federal and state law remedies. See Foley v. City of Lowell, 948 F.2d 10, 17 (1st Cir.1991) (suggesting that, as long as the damages are "segregated into federal and state components," plaintiff need not choose one body of law under which all damages will be paid); cf. Freeman v. Package Mach. Co., 865 F.2d 1331, 1343-45 (1st Cir.1988) (holding that plaintiff may not receive award based on federal and state law so as to receive double recovery for same element of relief); Schroeder v. Lotito, 747 F.2d 801, 802 (1st Cir.1984) (per curiam) (approving judgment for state law accounting of profits and federal law attorney's fees). Because DG has not requested a double award of attorney's fees, there was no error in the award of attorney's fees under federal law.

The district court granted DG's motions for summary judgment with respect to Grumman's tying claim under Section 1 of the Sherman Act as well as its monopolization claim under Section 2. We affirm both rulings, although on somewhat different grounds.

Section 1 of the Sherman Act prohibits a seller from "tying" the sale of one product to the purchase of a second product if the seller thereby avoids competition on the merits of the "tied" product. See 15 U.S.C. Sec. 1 ("Every contract ... in restraint of trade or commerce ... is declared to be illegal."); Jefferson Parish Hosp. Dist. No. 2 v. Hyde, 466 U.S. 2, 9-18, 104 S. Ct. 1551, 1556-1561, 80 L. Ed. 2d 2 (1984); Lee v. Life Ins. Co. of N. Am., 23 F.3d 14, 16 (1st Cir.1994); Grappone, Inc. v. Subaru of New England, Inc., 858 F.2d 792, 794-97 (1st Cir.1988) (Breyer, J.); Wells Real Estate, Inc. v. Greater Lowell Bd. of Realtors, 850 F.2d 803, 814-15 (1st Cir.), cert. denied, 488 U.S. 955, 109 S. Ct. 392, 102 L. Ed. 2d 381 (1988). In addition to outlawing "positive" ties likely to restrain competition, Section 1 also forbids "negative" ties--arrangements conditioning the sale of one product on an agreement not to purchase a second product from competing suppliers. See Eastman Kodak Co. v. Image Technical Servs., Inc., --- U.S. ----, ----, 112 S. Ct. 2072, 2079, 119 L. Ed. 2d 265 (1992) (citing Northern Pac. Ry. Co. v. United States, 356 U.S. 1, 5-6, 78 S. Ct. 514, 518-519, 2 L. Ed. 2d 545 (1958)); Lee, 23 F.3d at 16.

There are essentially four elements of a per se54 tying claim: (1) the tying and tied products are actually two distinct products; (2) there is an agreement or condition, express or implied, that establishes a tie; (3) the entity accused of tying has sufficient economic power in the market for the tying product to distort consumers' choices with respect to the tied product; and (4) the tie forecloses a substantial amount of commerce in the market for the tied product. See, e.g., Kodak, --- U.S. at ---- - ----, 112 S. Ct. at 2079-81; Grappone, 858 F.2d at 794; see also STI, 963 F.2d at 683.

Grumman claims that DG unlawfully restrained competition in the sale of MV service by tying access to ADEX (the tying product) to an equipment owner's promise to either purchase service from DG (a positive tie) or not purchase service from any other vendor (a negative tie). While a substantial amount of commerce is potentially involved, DG's motions for summary judgment claimed that there was no proof of any of the first three elements of a tying claim. The district court denied DG's first motion for summary judgment but then granted its renewed motion, stating in a sparse opinion that, as in STI, there was "no evidence which would warrant a finding of the existence of a tying agreement." Grumman V, 834 F. Supp. at 485. See also STI, 963 F.2d at 686 ("[STI's] evidence at bottom shows nothing more than a unilateral decision by Data General to license MV/ADEX to CMOs but not to others."). We agree with the district court's conclusion that there is insufficient evidence of a negative tying arrangement, but believe that the allegation of a positive tie falters at an earlier step.

a. Two Products

To establish the existence of two separate products, Grumman must identify the products at issue in each tie and demonstrate that "there is 'sufficient demand for the purchase of [the tied product] separate from [the tying product] to identify a distinct product market in which it is efficient to offer [the tied product] separately from [the tying product].' " STI, 963 F.2d at 684 (brackets in original) (quoting Jefferson Parish, 466 U.S. at 21-22, 104 S.Ct. at 1562-1564). See also Jefferson Parish, 466 U.S. at 40, 104 S.Ct. at 1573 (O'Connor, J., concurring) ("When the economic advantages of joint packaging are substantial the package is not appropriately viewed as two products, and that should be the end of the tying inquiry."); Lee, 23 F.3d at 16 n. 6 (noting that there must be evidence of "sufficient consumer demand for each individual product, and not merely as part of an integrated product 'package' ") (emphasis in original).

While Grumman has characterized the tying product in general terms as "access to ADEX," Grumman actually identifies two different tying products: ADEX service (a service) and ADEX software (a good). With respect to the positive tie, Grumman alleges that DG will not provide ADEX service (i.e., use of ADEX by a DG service technician) to equipment owners unless they also purchase DG support services. With respect to the negative tie, Grumman alleges that DG will not license ADEX software to equipment owners unless they agree not to purchase support services from a TPM.

Grumman has not introduced evidence that ADEX service is a product separate from other components of service. There is no evidence that any customer has purchased, or would wish to purchase, ADEX service separately from the purchase of other components of service. Nor is there evidence that it would be efficient for any entity to provide ADEX service separately from other components of service.55

In contrast, the record does contain evidence that ADEX software is a product separate from support services. It is undisputed that CMO customers wish to licenseand have licensed--ADEX software without purchasing support services from DG or a TPM. There is also evidence that some of Grumman's customers would consider licensing ADEX from DG so that Grumman could continue to service their MV computers. In addition, the summary judgment record would support a finding that for many years DG provided diagnostics and other service "tools" to computer purchasers as part of a computer equipment package, regardless whether the owner performed self-maintenance or hired DG or a TPM to maintain the computers. In fact, there is evidence that through the early 1980s, DG provided service "tools"--including diagnostic software other than ADEX--directly to TPMs. Finally, there is some evidence that other computer manufacturers (IBM, Digital Equipment Corporation, and Wang) have licensed or sold diagnostics to those other than their service customers. Viewed in a light most favorable to Grumman, the record reveals a genuine dispute as to whether ADEX software and support services for DG computers are distinct products for the purposes of a tying analysis.56 Consequently, we may proceed to determine whether Grumman has introduced sufficient proof that DG has conditioned the licensing of ADEX to CMOs on the agreement of these customers not to purchase service from TPMs.

b. Tying Arrangement

Proof of a tying arrangement generally requires evidence that the supplier's sale of the tying product is conditioned upon the unwilling purchase of the tied product from the supplier or an unwilling promise not to purchase the tied product from any other supplier. See, e.g., Jefferson Parish, 466 U.S. at 12, 104 S.Ct. at 1558 ("[T]he essential characteristic of an invalid tying arrangement lies in the seller's exploitation of its control over the tying product to force the buyer into the purchase of a tied product that the buyer either did not want at all, or might have preferred to purchase elsewhere on different terms."); Wells Real Estate, 850 F.2d at 814 ("Tying arrangements involve the use of leverage over the market for one product ... to coerce purchases of a second product...."). In the absence of an explicit tying agreement, conditioning may be inferred from evidence indicating that the supplier has actually coerced the purchase or non-purchase of another product. See Amerinet, Inc. v. Xerox Corp., 972 F.2d 1483, 1500 (8th Cir.1992), cert. denied, --- U.S. ----, 113 S. Ct. 1048, 122 L. Ed. 2d 356 (1993); Advanced Computer Servs., Inc. v. MAI Sys. Corp., 845 F. Supp. 356, 368 (E.D.Va.1994) (citing John H. Shenefield & Irwin M. Stelzer, The Antitrust Laws: A Primer 72 (1993) ("In the absence of an explicit agreement requiring the purchase as a condition of the sale, courts will accept proof suggesting any kind of coercion by the seller or unwillingness to take the second product by the buyer.")); see also Tic-X-Press, Inc. v. Omni Promotions Co., 815 F.2d 1407, 1418 (11th Cir.1987) ("It is well established that coercion may be established by showing that the facts and circumstances surrounding the transaction as a practical matter forced the buyer into purchasing the tied product."). In essence, whether the conditioning is explicit or implicit, we will not consider the anti-competitive effects of a tie to be unreasonable per se unless there is evidence that the supplier of the tying product has actually used its market power to impose the condition.

Grumman points to only one alleged negative tying arrangement, asserting that DG's Cooperative Maintenance Agreement ("CMO Agreement") contains an explicit tying condition. The CMO Agreement indeed states that DG designed the CMO program for "customers who perform their own maintenance" and that one qualifying criterion for CMO status is that the customer "[m]aintain[ ] systems which were purchased either for itself or for resale to its customers [as an official DG distributor]." And, although the record suggests that CMOs may still purchase DG service (and presumably TPM service) on a "time and materials" basis,57 CMOs do not enter contracts with either DG or TPMs for ongoing support services. DG and Grumman both agree that CMOs cannot allow TPMs to use copies of ADEX licensed by a CMO.

Grumman's allegation of an illegal tie cannot go to a jury on a record so sparse. Before turning to the principal flaw in Grumman's case, we note that these facts lend only modest support to the accusation that CMOs have actually promised not to purchase support services from TPMs. The CMO Agreement does require that participants maintain their own computers, but nowhere does the agreement define self-maintenance status in detail or elaborate on the consequences to a CMO if it enters a service contract with a TPM. In fact, as we have noted, there is some evidence in the record that CMOs are free to purchase support services from others without adverse consequences, at least on a "time and materials" basis.

More importantly, there is virtually no evidence that any CMO has unwillingly chosen to maintain its own computers. Although there is some evidence that DG officials designed the CMO program in part to prevent loss of DG revenue to TPMs, there is no evidence that consumers became CMO customers for any reason other than their belief that the CMO program was a "product" superior to TPM service. Indeed, while Grumman has argued tirelessly that DG service customers are forced to swallow overpriced and inferior support service, Grumman has offered no evidence that CMO customers are similarly disadvantaged. There is not a single affidavit in the record in which a CMO customer expresses either displeasure with the CMO program or an unfulfilled desire to switch to a TPM.58 Nor is there any other type of evidence that DG equipment owners capable of maintaining their own equipment would be more satisfied as TPM customers than as CMO customers. Consequently, the evidence in the record would not allow a reasonable jury to find that the CMO program is "an inferior or overpriced product," Amerinet, 972 F.2d at 1501, protected from competition by DG's exploitation of its control over ADEX.

In conclusion, Grumman's allegation of a positive tie between ADEX service and DG support services fails in the absence of proof that these services are truly two distinct products. Grumman's allegation of a negative tie between ADEX software and non-purchase of TPM support services fails in the absence of proof that DG coerced consumers to accept such an arrangement. Accordingly, the district court did not err in granting DG's motion for summary judgment on Grumman's tying counterclaim.

In addition to alleging unlawful tying, Grumman accused DG of willfully maintaining its monopoly in the aftermarket for service of DG computers in violation of Section 2 of the Sherman Act, 15 U.S.C. Sec. 2, which prohibits the monopolization of "any part of the trade or commerce among the several States." To survive summary judgment on its willful maintenance claim, Grumman must demonstrate a genuine dispute about the existence of two elements: (1) DG's possession of monopoly power in the market59 for support services of DG computers; and (2) DG's maintenance of that power through "exclusionary conduct." Town of Concord v. Boston Edison Co., 915 F.2d 17, 21 (1st Cir.1990) (Breyer, C.J.) (citing, inter alia, United States v. Grinnell Corp., 384 U.S. 563, 570-71, 86 S. Ct. 1698, 1703-04, 16 L. Ed. 2d 778 (1966)), cert. denied, 499 U.S. 931, 111 S. Ct. 1337, 113 L. Ed. 2d 268 (1991). The district court assumed the existence of monopoly power but granted summary judgment on the grounds that Grumman had not demonstrated the need for a trial on the element of exclusionary conduct. We follow suit.60

"Exclusionary conduct" is defined as " 'conduct, other than competition on the merits or restraints reasonably "necessary" to competition on the merits, that reasonably appears capable of making a significant contribution to creating or maintaining monopoly power.' " Town of Concord, 915 F.2d at 21 (quoting Barry Wright Corp. v. ITT Grinnell Corp., 724 F.2d 227, 230 (1st Cir.1983) (Breyer, J.), and 3 Phillip Areeda & Donald F. Turner, Antitrust Law p 626, at 83 (1978) (hereinafter "Areeda & Turner ")). We label as improper that conduct which harms the competitive process and not conduct which simply harms competitors. Id. That process is harmed when conduct "obstructs the achievement of competition's basic goals--lower prices, better products, and more efficient production methods." Id. at 21-22. Cf. Olympia Equip. Leasing Co. v. Western Union Tel. Co., 797 F.2d 370, 375 (7th Cir.1986) (describing shift in the emphasis of "antitrust policy ... from the protection of competition as a process of rivalry to the protection of competition as a means of promoting economic efficiency"), cert. denied, 480 U.S. 934, 107 S. Ct. 1574, 94 L. Ed. 2d 765 (1987). In contrast, exclusionary conduct does not include behavior which poses no unreasonable threat to consumer welfare but is merely a manifestation of healthy competition, an absence of competition, or a natural monopoly. See, e.g., United States v. Grinnell Corp., 384 U.S. 563, 570-71, 86 S. Ct. 1698, 1704, 16 L. Ed. 2d 778 (1966) (holding that Section 2 punishes only "willful acquisition or maintenance [of monopoly power] as distinguished from growth or development as a consequence of a superior product, business acumen, or historic accident").

Grumman's primary contention is that DG's unilateral refusal to license ADEX to anyone other than qualified self-maintainers constitutes exclusionary conduct.61 Grumman also attacks as exclusionary DG's refusal to provide other service tools directly to TPMs. We first review the principles governing the analysis of a monopolist's unilateral refusal to deal, and then discuss whether a unilateral refusal to license a copyrighted work might ever deserve to be condemned as exclusionary. We hold below that the desire of an author to be the exclusive user of its original work is a presumptively legitimate business justification for the author's refusal to license to competitors. We hold further that Grumman has not presented sufficient proof to rebut this presumption and thereby avert summary judgment. In particular, we find no merit in Grumman's contention that DG acted in an exclusionary fashion in discontinuing its liberal policies allowing TPM access to diagnostic software. Finally, we conclude that no reasonable jury could find that DG's restrictions on TPM access to other service tools amount to exclusionary conduct.

a. Unilateral Refusals to Deal

Because a monopolization claim does not require proof of concerted activity, even the unilateral actions of a monopolist can constitute exclusionary conduct. See 15 U.S.C. Sec. 2 (referring to "[e]very person who shall monopolize ... or combine or conspire with any other person ... to monopolize") (emphasis added); Moore v. Jas. H. Matthews & Co., 473 F.2d 328, 332 (9th Cir.1973) (observing that "section 2 is not limited to concerted activity"). Thus, a monopolist's unilateral refusal to deal with its competitors (as long as the refusal harms the competitive process) may constitute prima facie evidence of exclusionary conduct in the context of a Section 2 claim. See Kodak, --- U.S. at ---- n. 32, 112 S. Ct. at 2091 n. 32 (citing Aspen Skiing Co. v. Aspen Highlands Skiing Corp., 472 U.S. 585, 602-05, 105 S. Ct. 2847, 2857-59, 86 L. Ed. 2d 467 (1985)). A monopolist may nevertheless rebut such evidence by establishing a valid business justification for its conduct. See Kodak, --- U.S. at ---- n. 32, 112 S. Ct. at 2091 n. 32 (suggesting that monopolist may rebut an inference of exclusionary conduct by establishing "legitimate competitive reasons for the refusal"); Aspen Skiing, 472 U.S. at 608, 105 S.Ct. at 2860 (suggesting that sufficient evidence of harm to consumers and competitors triggers further inquiry as to whether the monopolist has "persuade[d] the jury that its [harmful] conduct was justified by [a] normal business purpose"). In general, a business justification is valid if it relates directly or indirectly to the enhancement of consumer welfare. Thus, pursuit of efficiency and quality control might be legitimate competitive reasons for an otherwise exclusionary refusal to deal, while the desire to maintain a monopoly market share or thwart the entry of competitors would not. See Kodak, --- U.S. at ----, 112 S. Ct. at 2091 (discussing the validity and sufficiency of various business justifications); Aspen Skiing, 472 U.S. at 608-11, 105 S.Ct. at 2860-62 (same); see generally 7 Areeda & Turner p 1504, at 377-83; 9 Areeda & Turner pp 1713, 1716-17, at 148-61, 185-239. In essence, a unilateral refusal to deal is prima facie exclusionary if there is evidence of harm to the competitive process; a valid business justification requires proof of countervailing benefits to the competitive process.

Despite the theoretical possibility, there have been relatively few cases in which a unilateral refusal to deal has formed the basis of a successful Section 2 claim. Several of the cases commonly cited for a supposed duty to deal were actually cases of joint conduct in which some competitors joined to frustrate others. See Associated Press v. United States, 326 U.S. 1, 65 S. Ct. 1416, 89 L. Ed. 2013 (1945); United States v. Terminal R.R. Ass'n, 224 U.S. 383, 32 S. Ct. 507, 56 L. Ed. 810 (1912). Prior to Aspen Skiing, the case that probably came closest to condemning a true unilateral refusal to deal was Otter Tail Power Co. v. United States, 410 U.S. 366, 93 S. Ct. 1022, 35 L. Ed. 2d 359 (1973), which condemned the refusal of a wholesale power supplier either to sell wholesale power to municipal systems or to "wheel power" when Otter Tail's retail franchises expired and local municipalities sought to supplant Otter Tail's local distributors. The case not only involved a capital-intensive public utility facility--which could not effectively be duplicated and occupied a distinct separate market--but the Supreme Court laid considerable emphasis on "supported" findings in the district court "that Otter Tail's refusals to sell at wholesale or to wheel were solely to prevent municipal power systems from eroding its monopolistic position." 410 U.S. at 378, 93 S.Ct. at 1030.

In Aspen Skiing, the Court criticized a monopolist's unilateral refusal to deal in a very different situation, casting serious doubt on the proposition that the Court has adopted any single rule or formula for determining when a unilateral refusal to deal is unlawful. In that case, an "all-Aspen" ski ticket--valid at any mountain in Aspen--had been developed and jointly marketed when the three (later four) ski areas in Aspen were owned by independent entities. 472 U.S. at 589, 105 S.Ct. at 2850. Some time after Aspen Skiing Company ("Ski Co.") came into control of three of the four ski areas, Ski Co. refused to continue a joint agreement with Aspen Highlands Skiing Corp. ("Highlands"), the owner of the fourth area. Id. at 592-93, 105 S.Ct. at 2852-53. Although there was no "essential facility" involved, the Court found that it was exclusionary for Ski Co., as a monopolist, to refuse to continue a presumably efficient "pattern of distribution that had originated in a competitive market and had persisted for several years." Id. at 603, 105 S.Ct. at 2857.

It is not entirely clear whether the Court in Aspen Skiing merely intended to create a category of refusal-to-deal cases different from the essential facilities category or whether the Court was inviting the application of more general principles of antitrust analysis to unilateral refusals to deal. We follow the parties' lead in assuming that Grumman need not tailor its argument to a preexisting "category" of unilateral refusals to deal.

b. Unilateral Refusals to License

DG attempts to undermine Grumman's monopolization claim by proposing a powerful irrebuttable presumption: a unilateral refusal to license a copyright can never constitute exclusionary conduct. We agree that some type of presumption is in order, but reach that conclusion only after an exhaustive inquiry touching on the general character of presumptions, the role of market analysis in the copyright context, existing responses to the tension between the antitrust and patent laws, the nature of the rights extended by the copyright laws, and our duty to harmonize two conflicting statutes.

(1) The Propriety of a Presumption

We begin our analysis with two observations. First, DG's rule of law could be characterized as either an empirical assumption or a policy preference. For example, if we were convinced that refusals to license a copyright always have a net positive effect on the competitive process, we might adopt a presumption to this effect in order to preclude wasteful litigation about a known fact. On the other hand, if we were convinced that the rights enumerated in the Copyright Act should take precedence over the responsibilities set forth in the Sherman Act, regardless of the realities of the market, we might adopt a blanket rule of preference. DG's argument contains elements of both archetypal categories of presumptions.

Second, we note that the phrase "competitive process" may need some refinement in order to evaluate either an empirical assumption or a policy presumption concerning the desirability of unilateral refusals to license a copyright. Antitrust law generally seeks to punish and prevent harm to consumers in particular markets, with a focus on relatively specific time periods. See, e.g., Jefferson Parish, 466 U.S. at 18, 104 S.Ct. at 1561 (holding that "any inquiry into the validity of a tying arrangement must focus on the market or markets in which the two products are sold, for that is where the anticompetitive forcing has its impact"). Thus, in determining whether conduct is exclusionary in the context of a monopolization claim, we ordinarily focus on harm to the competitive process in the relevant market and time period. See generally 3 Areeda & Turner pp 517-28, at 346-88, pp 533-36, at 406-431. Confining the competitive process in this way assists courts in deciding particular disputes based primarily on case-specific adjudicative facts rather than generally-applicable "legislative" facts or assumptions. The use and protection of copyrights also affects the "competitive process," but it may not be appropriate to judge the effect of the use of a copyright by looking only at one market or one time period.

We now consider what appears to be an empirical proclamation from DG: "[T]he refusal to make one's innovation available to rivals ... is pro-competitive conduct."62 As support, DG cites Grinnell, 384 U.S. at 570-71, 86 S.Ct. at 1704, in which the Court held that willful maintenance of monopoly does not include "growth or development as a consequence of a superior product." It is not the superiority of a work that allows the author to exclude others, however, but rather the limited monopoly granted by copyright law. Moreover, one reason why the Copyright Act fosters investment and innovation is that it may allow the author to earn monopoly profits by licensing the copyright to others or reserving the copyright for the author's exclusive use. See Sony Corp. of Am. v. Universal City Studios, Inc., 464 U.S. 417, 429, 104 S. Ct. 774, 782, 78 L. Ed. 2d 574 (1984) (explaining that the limited copyright monopoly "is intended to motivate the creative activity of authors and inventors by the provision of a special reward"). Thus, at least in a particular market and for a particular period of time, the Copyright Act tolerates behavior that may harm both consumers and competitors. Cf. SCM Corp. v. Xerox Corp., 645 F.2d 1195, 1203 (2d Cir.1981) ("[T]he primary purpose of the antitrust laws--to preserve competition--can be frustrated, albeit temporarily, by a holder's exercise of the patent's inherent exclusionary power during its term."), cert. denied, 455 U.S. 1016, 102 S. Ct. 1708, 72 L. Ed. 2d 132 (1982).

DG does not in fact argue that consumers are better off in the short term because of the inability of TPMs to license ADEX. Instead, DG suggests that allowing copyright owners to exclude others from the use of their works creates incentives which ultimately work to the benefit of consumers in the DG service aftermarket as well as to the benefit of consumers generally. In other words, DG seeks to justify any immediate harm to consumers by pointing to countervailing long-term benefits. Certainly, a monopolist's refusal to license others to use a commercially successful patented idea is likely to have more profound anti-competitive consequences than a refusal to allow others to duplicate the copyrighted expression of an unpatented idea (although such differences may become less pronounced if copyright law becomes increasingly protective of intellectual property such as computer software). But by no means is a monopolist's refusal to license a copyright entirely "pro-competitive" within the ordinary economic framework of the Sherman Act. Accordingly, it may be inappropriate to adopt an empirical assumption that simply ignores harm to the competitive process caused by a monopolist's unilateral refusal to license a copyright. Even if it is clear that exclusive use of a copyright can have anticompetitive consequences, some type of presumption may nevertheless be appropriate as a matter of either antitrust law or copyright law.

(2) Antitrust Law and the Accommodation of Patent Rights

Antitrust law is somewhat instructive. Although creation and protection of original works of authorship may be a national pastime, the Sherman Act does not explicitly exempt such activity from antitrust scrutiny and courts should be wary of creating implied exemptions. See Square D Co. v. Niagara Frontier Tariff Bureau, Inc., 476 U.S. 409, 421, 106 S. Ct. 1922, 1929, 90 L. Ed. 2d 413 (1986) ("[E]xemptions from the antitrust laws are strictly construed and strongly disfavored."); cf. Flood v. Kuhn, 407 U.S. 258, 92 S. Ct. 2099, 32 L. Ed. 2d 728 (1972) (holding that the longstanding judicially created exemption of professional baseball from the Sherman Act is an established "aberration" in which Congress has acquiesced). The Supreme Court has suggested that an otherwise reasonable yet anti-competitive use of a copyright should not "be deemed a per se violation of the Sherman Act," Broadcast Music, Inc. v. CBS, Inc., 441 U.S. 1, 19, 99 S. Ct. 1551, 1562, 60 L. Ed. 2d 1 (1979), but a monopolistic refusal to license might still violate the rule of reason, see Rural Tel. Serv. Co. v. Feist Publications, Inc., 957 F.2d 765, 767-69 (10th Cir.) (analyzing reasonableness of monopolist's unilateral refusal to license copyrighted telephone listings to a competing distributor of telephone directories), cert. denied, --- U.S. ----, 113 S. Ct. 490, 121 L. Ed. 2d 429 (1992).63 Should an antitrust plaintiff be allowed to demonstrate the anti-competitive effects of a monopolist's unilateral refusal to grant a copyright license? Would the monopolist then have to justify its refusal to license by introducing evidence that the protection of the copyright laws enabled the author to create a work which advances consumer welfare?

The courts appear to have partly settled an analogous conflict between the patent laws and the antitrust laws, treating the former as creating an implied limited exception to the latter. In Simpson v. Union Oil Co., 377 U.S. 13, 24, 84 S. Ct. 1051, 1058, 12 L. Ed. 2d 98 (1964), the Supreme Court stated that "[t]he patent laws which give a 17-year monopoly on 'making, using, or selling the invention' are in pari materia with the antitrust laws and modify them pro tanto." Similarly, we have suggested that the exercise of patent rights is a "legitimate means" by which a firm may maintain its monopoly power. Barry Wright, 724 F.2d at 230. Other courts have specifically held that a monopolist's unilateral refusal to license a patent is ordinarily not properly viewed as exclusionary conduct. See Miller Insituform, 830 F.2d at 609 ("A patent holder who lawfully acquires a patent cannot be held liable under Section 2 of the Sherman Act for maintaining the monopoly power he lawfully acquired by refusing to license the patent to others."); Westinghouse, 648 F.2d at 647 (finding no antitrust violation because "Westinghouse has done no more than to license some of its patents and refuse to license others"); SCM Corp., 645 F.2d at 1206 (holding that "where a patent has been lawfully acquired, subsequent conduct permissible under the patent laws cannot trigger any liability under the antitrust laws"); see also 3 Areeda & Turner p 704, at 114 ("The patent is itself a government grant of monopoly and is therefore an exception to usual antitrust rules."). This exception is inoperable if the patent was unlawfully "acquired." SCM Corp., 645 F.2d at 1208-09 (analyzing legality of Xerox's acquisition of plain-paper copier patent); see generally 3 Areeda & Turner pp 705-707, at 117-45 (discussing effect of patent acquisition, internal development of patents, and improprieties in patent procurement on applicability of antitrust laws).

The "patent exception" is largely a means of resolving conflicting rights and responsibilities, i.e., a policy presumption. See, e.g., Miller Insituform, 830 F.2d at 609 (declaring summarily that "[t]here is no adverse effect on competition since, as a patent monopolist, [the patent holder] had [the] exclusive right to manufacture, use, and sell his invention.") (emphasis added). At the same time, the exception is grounded in an empirical assumption that exposing patent activity to wider antitrust scrutiny would weaken the incentives underlying the patent system, thereby depriving consumers of beneficial products. See, e.g., SCM Corp., 645 F.2d at 1209 (holding that imposition of antitrust liability for an arguably unreasonable refusal to license a lawfully acquired patent "would severely trample upon the incentives provided by our patent laws and thus undermine the entire patent system").

(3) Copyright Law

Copyright law provides further guidance. The Copyright Act expressly grants to a copyright owner the exclusive right to distribute the protected work by "transfer of ownership, or by rental, lease, or lending." 17 U.S.C. Sec. 106. Consequently, "[t]he owner of the copyright, if [it] pleases, may refrain from vending or licensing and content [itself] with simply exercising the right to exclude others from using [its] property." Fox Film Corp. v. Doyal, 286 U.S. 123, 127, 52 S. Ct. 546, 547, 76 L. Ed. 1010 (1932). See also Stewart v. Abend, 495 U.S. 207, 229, 110 S. Ct. 1750, 1764, 109 L. Ed. 2d 184 (1990). We may also venture to infer that, in passing the Copyright Act, Congress itself made an empirical assumption that allowing copyright holders to collect license fees and exclude others from using their works creates a system of incentives that promotes consumer welfare in the long term by encouraging investment in the creation of desirable artistic and functional works of expression. See Feist Publications, Inc. v. Rural Tel. Serv. Co., 499 U.S. 340, 349, 111 S. Ct. 1282, 1290, 113 L. Ed. 2d 358 (1991) ("The primary objective of a copyright is not to reward the labor of authors, but '[t]o promote the Progress of Science and useful Arts.' ") (brackets in original) (quoting U.S. Const. art. I. Sec. 8, cl. 8); Sony Corp., 464 U.S. at 429, 104 S.Ct. at 782 (discussing goals and incentives of copyright protection); Twentieth Century Music Corp. v. Aiken, 422 U.S. 151, 156, 95 S. Ct. 2040, 2044, 45 L. Ed. 2d 84 (1975) ("The immediate effect of our copyright law is to secure a fair return for an 'author's' creative labor. But the ultimate aim is, by this incentive, to stimulate artistic creativity for the general public good."). We cannot require antitrust defendants to prove and reprove the merits of this legislative assumption in every case where a refusal to license a copyrighted work comes under attack. Nevertheless, although "nothing in the copyright statutes would prevent an author from hoarding all of his works during the term of the copyright," Stewart, 495 U.S. at 228-29, 110 S.Ct. at 1764 (emphasis added), the Copyright Act does not explicitly purport to limit the scope of the Sherman Act. And, if the Copyright Act is silent on the subject generally, the silence is particularly acute in cases where a monopolist harms consumers in the monopolized market by refusing to license a copyrighted work to competitors.

We acknowledge that Congress has not been entirely silent on the relationship between antitrust and intellectual property laws. Congress amended the patent laws in 1988 to provide that "[n]o patent owner otherwise entitled to relief for infringement ... of a patent shall be denied relief or deemed guilty of misuse or illegal extension of the patent right by reason of [the patent owner's] refus[al] to license or use any rights to the patent." 35 U.S.C. Sec. 271(d) (1988). Section 271(d) clearly prevents an infringer from using a patent misuse defense when the patent owner has unilaterally refused a license, and may even herald the prohibition of all antitrust claims and counterclaims premised on a refusal to license a patent. See Richard Calkins, Patent Law: The Impact of the 1988 Patent Misuse Reform Act and Noerr-Pennington Doctrine on Misuse Defenses and Antitrust Counterclaims, 38 Drake L.Rev. 192-97 (1988-89). Nevertheless, while Section 271(d) is indicative of congressional "policy" on the need for antitrust law to accommodate intellectual property law, Congress did not similarly amend the Copyright Act.

(4) Harmonizing the Sherman Act and the Copyright Act

Since neither the Sherman Act nor the Copyright Act works a partial repeal of the other, and since implied repeals are disfavored, e.g., Watt v. Alaska, 451 U.S. 259, 267, 101 S. Ct. 1673, 1678, 68 L. Ed. 2d 80 (1981), we must harmonize the two as best we can, id., mindful of the legislative and judicial approaches to similar conflicts created by the patent laws. We must not lose sight of the need to preserve the economic incentives fueled by the Copyright Act, but neither may we ignore the tension between the two very different policies embodied in the Copyright Act and the Sherman Act, both designed ultimately to improve the welfare of consumers in our free market system. Drawing on our discussion above, we hold that while exclusionary conduct can include a monopolist's unilateral refusal to license a copyright, an author's desire to exclude others from use of its copyrighted work is a presumptively valid business justification for any immediate harm to consumers.64

c. DG's Refusal to License ADEX to non-CMOs

Having arrived at the applicable legal standards, we may resolve Grumman's principal allegation of exclusionary conduct. Although there may be a genuine factual dispute about the effect on DG equipment owners of DG's refusal to license ADEX to TPMs, DG's desire to exercise its rights under the Copyright Act is a presumptively valid business justification.

Apparently sensing the uphill nature of its allegation of an exclusionary refusal to license, Grumman seeks to overcome any obstacles primarily by characterizing DG's licensing policies as a monopolist's exclusionary withdrawal of assistance within the framework of Aspen Skiing. Citing Aspen Skiing, Grumman contends that DG's refusal to license ADEX to TPMs, in light of the fact that DG previously allowed TPMs to use DG diagnostics, is exclusionary conduct because "[a] monopolist that has helped a market develop may not withdraw its support without legitimate business justifications."

Assuming that such a claim can overcome the presumption that a refusal to license is not exclusionary, we nevertheless hold that Aspen Skiing cannot apply to the facts of this case. The reasoning of Aspen Skiing has little to do with the fact that defendant Ski Co. withdrew assistance upon which competitors may have relied when entering the market. Rather, the decision turns on a comparison of the behavior of firms in a competitive market (the Aspen ski market) with a monopolist's behavior once competition has been curtailed. The Court noted that the rich soil of competition had produced the all-mountain ticket in Aspen and other multimountain areas, justifying an "infer[ence] that such tickets satisfy consumer demand in free competitive markets." 472 U.S. at 603, 105 S.Ct. at 2858. See also Olympia Equip., 797 F.2d at 377 (suggesting that the facts in Aspen Skiing indicate that "competition required some cooperation among competitors" in the Aspen ski market). Ski Co.'s decision to eliminate the ticket in later years was a sign that the weeds of monopoly had begun to take hold, to the possible detriment of consumer welfare. Aspen Skiing, 472 U.S. at 604, 105 S.Ct. at 2858. Finally, after canvassing evidence of consumer preferences concerning skiing options at Aspen, the Court concluded that Ski Co.'s new policies did in fact harm consumers. Id. at 605-607, 105 S.Ct. at 2858-2860. In short, instead of prescribing a categorical approach, Aspen Skiing ultimately calls for an inquiry that is relatively routine in antitrust analysis: namely, whether the monopolist's actions unjustifiably harm the competitive process by frustrating consumer preferences and erecting barriers to competition. Cf. Olympia Equip., 797 F.2d at 379 ("If [Aspen Skiing ] stands for any principle that goes beyond its unusual facts, it is that a monopolist may be guilty of monopolization if it refuses to cooperate with a competitor in circumstances where some cooperation is indispensable to effective competition.").

Grumman attempts to analogize this case to Aspen Skiing by focusing on the fact that DG once encouraged firms to enter the DG service aftermarket by allowing liberal access to service tools, but no longer does so. The analytical framework of Aspen Skiing cannot function in these circumstances, however, because we are unable to view DG's market practices in both competitive and noncompetitive conditions. While TPMs have made inroads in the market for service of DG computers, DG has always been a monopolist in that market, and competitive conditions have never prevailed. Therefore, it would not be "appropriate to infer" from DG's change of heart that its former policies "satisfy consumer demand in free competitive markets." Aspen Skiing, 472 U.S. at 603, 105 S. Ct. at 2857.

Nor does it appear that Grumman would be able at trial to overcome the presumption on any other theory. There is no evidence that DG acquired its ADEX copyrights in any unlawful manner; indeed, the record suggests that DG developed all its software internally. Cf. 3 Areeda & Turner p 706, at 127-28 (arguing that although an internally developed patent may be as exclusionary as one acquired from outside a firm, labelling the former as exclusionary would "discourage progressiveness by monopolists"). And, while there is evidence that DG knew that developing a "proprietary position" in the area of diagnostic software would help to maintain its monopoly in the aftermarket for service of DG computers, there is also evidence that DG set out to create a state-of-the-art diagnostic that would help to improve the quality of DG service. Cf. id. p 706, at 128-29 (suggesting that "nearly all commercial research rests on a mixture of motivations" and that a search for an overriding "antisocial" motivation would be unilluminating). In fact, there is clearly some evidence that ADEX is a significant benefit to owners of DG's MV computers. ADEX is a better product than any other diagnostic for MV computers. The use of ADEX appears to have increased the efficiency and reduced the cost of service because technicians can locate problems more quickly and, through the use of the software's "remote assistance" capability, can arrive at customer sites having determined ahead of time what replacement parts are necessary. In addition to the possibility of lower prices occasioned by such gains in efficiency, ADEX also promises to lower prices through gains in effectiveness. For example, customers may save the cost of replacing expensive hardware components because the use of advanced diagnostics increases the possibility that technicians can locate a problem and repair the component.

d. DG's Other Restrictive Policies

Grumman's other allegations of exclusionary conduct are equally devoid of merit and require no extended analysis. It is essentially undisputed that DG will not provide spare parts, depot repair services, certain documentation, change order kits, or schematics to TPMs. But there is no evidence of any resulting harm to DG equipment owners. DG makes most of these items available directly to equipment owners. Equipment owners "may purchase ... depot repair services, rev-ups [change order kits], and spare parts directly from DG, regardless of whether their computers are serviced by DG, TPMs, or themselves." Grumman II, 761 F. Supp. at 189. We cannot presume that elimination of an intermediate seller of such items harms consumers; indeed, consumers are likely to benefit by not having to accept TPMs' mark-up of DG prices. Further, a direct sales policy does not act as a significant barrier to market entry by competitors offering lower prices for higher quality support services. TPM technicians may identify broken parts for the customer to send to DG's repair depot, use the change order kits to upgrade a customer's computer, and install spare parts the customer has ordered from DG.

Neither equipment owners nor TPMs may purchase schematics (blueprints of equipment that often contain manufacturing secrets), but Grumman has not introduced sufficient evidence that this policy constitutes exclusionary conduct. Grumman's theory below was that DG's refusal to sell schematics to TPMs prevented TPMs from acquiring the information necessary to develop fully competitive substitutes for ADEX. Even a monopolist, however, "may normally keep its innovations secret from its rivals as long as it wishes." Berkey Photo, Inc. v. Eastman Kodak Co., 603 F.2d 263, 281 (2d Cir.1979), cert. denied, 444 U.S. 1093, 100 S. Ct. 1061, 62 L. Ed. 2d 783 (1980). DG's policy might be exclusionary if DG had sought to alter its equipment (and therefore the schematics describing that equipment) in order to prevent technological advances by TPMs. But, as the district court noted, "Grumman ... makes no allegations that DG has in fact attempted to subvert competitors' efforts to develop and implement competing diagnostics." Grumman II, 761 F. Supp. at 191.

In conclusion, Grumman has not produced evidence from which a jury could find that DG engaged in exclusionary conduct by unilaterally refusing to license ADEX or sell schematics to TPMs, or by only selling other service tools directly to equipment owners. Therefore, there was no error in the district court's entry of summary judgment on Grumman's monopolization claim.

For the foregoing reasons, we affirm the district court in every respect save for its failure to instruct the jury on its duty to consider Grumman's plea for apportionment of Grumman's nonduplicative profits.65 We remand the case to the district court for the sole purpose of resolving that issue.

So ordered.