Today's case requires us to review Texas law on the misappropriation of trade secrets. Having done so, we conclude that the district court misconceived the nature and elements of this cause of action, a misconception that led it to direct a verdict erroneously in favor of appellee Bielefeldt. We also conclude that the court abused its discretion in excluding certain evidence. Accordingly, we affirm in part, reverse in part, and remand the case for a new trial.

We commence with a brief description of the scientific process concerned. Tungsten carbide is a metallic compound of great value in certain industrial processes. Combined with the metal cobalt, it forms an extremely hard alloy known as "cemented tungsten carbide"1 used in oil drills, tools for manufacturing metals, and wear-resistent coatings. Because of its great value, reclamation of carbide from scrap metals is feasible. For a long time, however, the alloy's extreme resistence to machining made reclamation difficult. In the late 1960's and early 1970's, a new solution--known as the zinc recovery process--was devised, a solution based on carbide's reaction with zinc at high temperatures. In the crucibles of a furnace, molten zinc will react with the cobalt in the carbide to cause swelling and cracking of the scrap metal. After this has occurred, the zinc is distilled from the crucible, leaving the scrap in a more brittle state. The carbide is then ground into a powder, usable in new products as an alternative to virgin carbide. This process is the generally recognized modern method of carbide reclamation.

Metallurgical Industries has been in the business of reclaiming carbide since 1967, using the more primitive "cold-stream process." In the mid-1970's, Metallurgical began to consider using the zinc recovery process. In that connection, it came to know appellee Irvin Bielefeldt, a representative of Therm-O-Vac Engineering & Manufacturing Company (Therm-O-Vac). Negotiations led to a contract authorizing Therm-O-Vac to design and construct two zinc recovery furnaces, the purchase order for the first being executed in July 1976.

The furnace arrived in April 1977. Dissatisfied with its performance, Metallurgical modified it extensively. First, it inserted chill plates in one part of the furnace to create a better temperature differential for distilling the zinc. Second, Metallurgical replaced the one large crucible then in place with several smaller crucibles to prevent the zinc from dispersing in the furnace. Third, it replaced segmented heating elements which had caused electrical arcing with unitary graphite heating elements. Last, it installed a filter in the furnace's vacuum-pumps, which zinc particles had continually clogged. These efforts proved successful and the modified furnace soon began commercial operation.

In the market for a second furnace in mid-1978, Metallurgical provided to Consarc, another furnace manufacturer, all its hard-won information about zinc-recovery furnace design. Apparently allowed to watch the first furnace operate, Consarc employees learned of its modifications. Because Consarc proved unwilling or unable to build what Metallurgical wanted, however, the agreement fell through, and Metallurgical returned to Therm-O-Vac for its second furnace. A purchase order was signed in January 1979, and the furnace arrived that July. Further modifications again had to be made, but commercial production was allegedly achieved in January 1980.

In 1980, after Therm-O-Vac went bankrupt, Bielefeldt and three other former Therm-O-Vac employees--Norman Montesino, Gary Boehm, and Michael Sarvadi--formed Fourtek, Incorporated. Soon thereafter, Fourtek agreed to build a zinc recovery furnace for appellee Smith International, Incorporated (Smith). The furnace Fourtek provided incorporated the modifications Metallurgical had made in its furnaces; chilling systems, pump filters, multiple crucibles, and unitary heating elements. Smith has been unable to use this furnace commercially, however, because a current shortage of carbide scrap prevents its economically feasible operation.

Metallurgical nevertheless brought a diversity action against Smith, Bielefeldt, Montesino, Boehm, and Sarvadi in November 1981.2 In its complaint, Metallurgical charged the defendants with misappropriating its trade secrets. Other causes of action, not brought forward on appeal, were breach of contract, interference with business relations, conversion, and unfair competition. Trial began on June 4, 1984, and Metallurgical spent the next ten days presenting its case in chief. During this time, testimony indicated Metallurgical's frequent notices to Bielefeldt that the process was a secret and that the disclosures to him were made in confidence. Another witness recounted meetings in which the modifications were agreed to; Bielefeldt was allegedly unconvinced about the efficacy of these changes and contributed little to the discussion. Metallurgical also presented evidence that it had expended considerable time, effort, and money to modify the furnaces.

Such evidence apparently did not impress the trial court; at the close of Metallurgical's case, it granted the defendants' motions for directed verdicts. Ruling from the bench, the court provided an array of reasons for its order. The principal reason advanced was the court's conclusion that no trade secret is involved. At trial, Metallurgical acknowledged that the individual changes, by themselves, are not secrets; chill plates and pump filters, for example, are well-known. Metallurgical's position instead was that the process, taken as a whole, is a trade secret in the carbide business. The court, however, refused to recognize any protection Texas law provides to a modification process. It also concluded that the information Bielefeldt obtained from working with Metallurgical is too general to be legally protected. Finally, it ruled that "negative know-how"--the knowledge of what not to do--is unprotected. All these findings seem to have coalesced into a general conclusion by the court that this case involves no trade secret. The trial court went on to provide further rationales for its decision; it ruled that there was evidence of neither Bielefeldt's improper use nor disclosure of any secret, nor of Metallurgical's having been damaged by any improper misappropriation. Metallurgical appeals all these conclusions against Bielefeldt and Smith only. It also contends that the court erred in excluding certain evidence.

We begin our discussion by noting the applicable standard of review. In Boeing Co. v. Shipman, 411 F.2d 365, 374 (5th Cir.1969) (en banc), we enounced the standard for deciding whether a directed verdict is proper:

[T]he [trial] court should consider all of the evidence--not just that evidence which supports the non-mover's case--but in the light and with all reasonable inferences most favorable to the party opposed to the motion. If the facts and inferences point so strongly and overwhelmingly in favor of one party that the Court believes that reasonable men could not arrive at a contrary verdict, granting of the motions is proper. On the other hand, if there is substantial evidence opposed to the motions, that is, evidence of such quality and weight that reasonable and fair-minded men in the exercise of impartial judgment might reach different conclusions, the motions should be denied, and the case submitted to the jury.

The standard of review on appeal is the same. See, e.g., Bobb v. Modern Products, Inc., 648 F.2d 1051, 1057 (5th Cir.1981). We must therefore consider all evidence and inferences most favorably to Metallurgical.

Because the district court provided so many reasons for its order, we feel compelled to discuss the law of trade secrets in detail. Our discussion concentrates on Texas law, despite a clause in both purchase order agreements that their interpretation is to be made under New Jersey law. This stems from the nature of this case; as will be explained below, we are dealing with a cause of action sounding in tort, not one based on contract.

Individual attention to the various elements of this tort is necessary to provide an easily-understood analysis, but we can here briefly summarize the discussion. A plaintiff must certainly show that a "trade secret" is involved; the definition of this term is therefore crucial and must be based on several factors. If the trial court concludes that a trade secret exists, it then must determine whether the defendant committed any wrongdoing. One who breaches the confidence reposed in him by the holder of a trade secret and one who obtains the secret can be held accountable. No defendant may be liable, however, unless he has "disclosed" or "used" the secret improperly; again, defining these terms is required. These considerations are the sum and substance of the cause of action involved. After expounding the substantive law, we then turn to issues regarding the district court's evidentiary rulings and the relief available to Metallurgical.

III. DEFINING A "TRADE SECRET"

We begin by reviewing the legal definition of a trade secret. Of course, to qualify as one, the subject matter involved must, in fact, be a secret; "[m]atters of general knowledge in an industry cannot be appropriated by one as his secret." Wissman v. Boucher, 150 Tex. 326, 240 S.W.2d 278, 280 (1951); see also Zoecon Industries v. American Stockman Tag Co., 713 F.2d 1174, 1179 (5th Cir.1983) ("a customer list of readily ascertainable names and addresses will not be protected as a trade secret"). Smith emphasizes the absence of any secret because the basic zinc recovery process has been publicized in the trade. Acknowledging the publicity of the zinc recovery process, however, we nevertheless conclude that Metallurgical's particular modification efforts can be as yet unknown to the industry. A general description of the zinc recovery process reveals nothing about the benefits unitary heating elements and vacuum pump filters can provide to that procedure. That the scientific principles involved are generally known does not necessarily refute Metallurgical's claim of trade secrets.

Metallurgical, furthermore, presented evidence to back up its claim. One of its main witnesses was Arnold Blum, a consultant very influential in the decisions to modify the furnaces. Blum testified as to his belief that Metallurgical's changes were unknown in the carbide reclamation industry. The evidence also shows Metallurgical's efforts to keep secret its modifications. Blum testified that he noted security measures taken to conceal the furnaces from all but authorized personnel. The furnaces were in areas hidden from public view, while signs warned all about restricted access. Company policy, moreover, required everyone authorized to see the furnace to sign a non-disclosure agreement. These measures constitute evidence probative of the existence of secrets. One's subjective belief of a secret's existence suggests that the secret exists. Security measures, after all, cost money; a manufacturer therefore presumably would not incur these costs if it believed its competitors already knew about the information involved. In University Computing Co. v. Lykes-Youngstown Corp., 504 F.2d 518, 535 (5th Cir.1974), we regarded subjective belief as a factor to consider in determining whether secrecy exists. Because evidence of security measures is relevant, that shown here helps us conclude that a reasonable jury could have found the existence of the requisite secrecy.

Smith argues, however, that Metallurgical's disclosure to other parties vitiated the secrecy required to obtain legal protection. As mentioned before, Metallurgical revealed its information to Consarc Corporation in 1978; it also disclosed information in 1980 to La Floridienne, its European licensee of carbide reclamation technology. Because both these disclosures occurred before Bielefeldt allegedly misappropriated the knowledge of modifications, others knew of the information when the Smith furnace was built. This being so, Smith argues, no trade secret in fact existed.

Although the law requires secrecy, it need not be absolute. Public revelation would, of course, dispel all secrecy, but the holder of a secret need not remain totally silent:

He may, without losing his protection, communicate it to employees involved in its use. He may likewise communicate it to others pledged to secrecy.... Nevertheless, a substantial element of secrecy must exist, so that except by the use of improper means, there would be difficulty in acquiring the information.

Restatement of Torts, Sec. 757 Comment b (1939). We conclude that a holder may divulge his information to a limited extent without destroying its status as a trade secret. To hold otherwise would greatly limit the holder's ability to profit from his secret. If disclosure to others is made to further the holder's economic interests, it should, in appropriate circumstances, be considered a limited disclosure that does not destroy the requisite secrecy. The only question is whether we are dealing with a limited disclosure here.

Prior caselaw provides no guidance on what constitutes limited disclosure. Metallurgical cites Hyde Corp. v. Huffines, 158 Tex. 566, 314 S.W.2d 763, cert. denied 358 U.S. 898, 79 S. Ct. 223, 3 L. Ed. 2d 148 (1958), and Sikes v. McGraw Edison Co., 665 F.2d 731 (5th Cir.), cert. denied 458 U.S. 1108, 102 S. Ct. 3488, 73 L. Ed. 2d 1369 (1982), in contending that subsequent disclosure of a trade secret does not free one from the constraint of a prior confidential disclosure. In both of these cases, however, publication of the trade secret by its holder followed an improper use by one in whom the holder had confided. This factual difference renders these cases inapposite.

Looking instead to the policy considerations involved, we glean two reasons why Metallurgical's disclosures to others are limited and therefore insufficient to extinguish the secrecy Metallurgical's other evidence has suggested. First, the disclosures were not public announcements; rather, Metallurgical divulged its information to only two businesses with whom it was dealing. This case thus differs from Luccous v. J.C. Kinley Co., 376 S.W.2d 336 (Tex.1964), in which the court concluded that the design of a device could not be a trade secret because it had been patented--and thus revealed to all the world--before any dealing between the parties. Second, the disclosures were made to further Metallurgical's economic interests. Disclosure to Consarc was made with the hope that Consarc could build the second furnace. A longstanding agreement gave La Floridienne the right, as a licensee, to the information in exchange for royalty payments. Metallurgical therefore revealed its discoveries as part of business transactions by which it expected to profit.

Metallurgical's case would have been stronger had it also presented evidence of confidential relationships with these two companies, but we are unwilling to regard this failure as conclusively disproving the limited nature of the disclosures. Smith correctly points out that Metallurgical bears the burden of showing the existence of confidential relationships. Contrary to Smith's assertion, however, confidentiality is not a requisite; it is only a factor to consider. Whether a disclosure is limited is an issue the resolution of which depends on weighing many facts. The inferences from those facts, construed favorably to Metallurgical, is that it wished only to profit from its secrets in its business dealings, not to reveal its secrets to the public. We therefore are unpersuaded by Smith's argument.

Existing law, however, emphasizes other requisites for legal recognition of a trade secret. In Huffines, 314 S.W.2d 763, a seminal case of trade secret law, Texas adopted the widely-recognized pronouncements of the American Law Institute's Restatement of the Law. The Texas Supreme Court quoted the Restatement's definition of a trade secret:

A trade secret may consist of any formula, pattern, device, or compilation of information which is used in one's business, and which gives him an opportunity to obtain an advantage over competitors who do not know it. It may be a chemical compound, a process of manufacturing, treating or preserving materials, a pattern for a machine or other device or a list of customers.

Id. at 776, quoting Restatement of Torts, Sec. 757 Comment b (1939). From this the criterion of value to the holder of the alleged secret arises, a criterion we have noted before. In Zoecon Industries, 713 F.2d at 1179, we concluded that a customer list was a trade secret in part because the list gave to its owner an advantage "over competitors who did not have this information." See also Lykes-Youngstown Corp., 504 F.2d at 535 ("[e]vidence was adduced at trial that [the product in question] had unique capabilities and features which made it a valuable competitive product").

Metallurgical met the burden of showing the value of its modifications. Lawrence Lorman, the company's vice president, testified that the zinc recovery process gave Metallurgical an advantage over its two competitors by aiding in the production of the highest quality reclaimed carbide powder. The quality of the powder, in fact, makes it an alternative to the more costly virgin carbide. Lorman testified that customers regarded Metallurgical's zinc reclaimed powder as a better product than that reclaimed by the coldstream process used by others. This evidence clearly indicates that the modifications that led to the commercial operation of the zinc recovery furnace provided a clear advantage over the competition.

Another requisite is the cost of developing the secret device or process. In Huffines' companion case, K & G Oil, Tool & Service Co. v. G & G Fishing Tool Service, 158 Tex. 594, 314 S.W.2d 782, 790 (1958), the court recognized the cost involved in developing the device in question; "[t]he record shows ... that much work and ingenuity have been applied to the development of a practical and successful device." See also Zoecon Industries, 713 F.2d at 1179 ("even if the names and addresses were readily ascertainable through trade journals as the defendants allege, the other information could be compiled only at considerable expense"). No question exists that Metallurgical expended much time, effort, and money to make the necessary changes. It clearly has met the burden of demonstrating the effort involved in making a complex manufacturing process work.

That the cost of devising the secret and the value the secret provides are criteria in the legal formulation of a trade secret shows the equitable underpinnings of this area of the law. It seems only fair that one should be able to keep and enjoy the fruits of his labor. If a businessman has worked hard, has used his imagination, and has taken bold steps to gain an advantage over his competitors, he should be able to profit from his efforts. Because a commercial advantage can vanish once the competition learns of it, the law should protect the businessman's efforts to keep his achievements secret. As is discussed below, this is an area of law in which simple fairness still plays a large role.

We do not say, however, that all these factors need exist in every case. Because each case must turn on its own facts, no standard formula for weighing the factors can be devised. Secrecy is always required, of course, but beyond that there are no universal requirements. In a future case, for example, should the defendant's breach of confidence be particularly egregious, the injured party might still seek redress in court despite the possibility that the subject matter was discovered at little or no cost or that the object of secrecy is not of great value to him. The definition of "trade secret" will therefore be determined by weighing all equitable considerations. It is easy to recognize the possibility of a trade secret here, however, because Metallurgical presented evidence of all three factors discussed above.

Appellees posit two other reasons why Metallurgical's modification process cannot be defined as a trade secret. The first is premised on characterizing the process in question as the installation of various devices all well known to modern manufacturing. This being so, the argument runs, the process itself can be no secret, either. The technologies of chill plates, multiple crucibles, pump filters, and unitary graphite heating elements are all said to be public knowledge. This may well be so, but it does not prevent Metallurgical from seeking legal protection. Ventura Manufacturing Co. v. Locke, 454 S.W.2d 431 (Tex.Civ.App.1970), involved a process of finding and cleaning small yet valuable titanium scraps lost during the manufacturing of airplanes. Appellee Locke refined a previously unsuccessful process by implementing new procedures, none of which was innovative. The court nevertheless found that his efforts could constitute a trade secret:

Obviously, the basic proposition involved in the cleaning of dirty parts by commercial detergents involves no trade secret. Also, the use of production line techniques as such must be considered as a matter of general knowledge. On the other hand, we cannot say that the details of the procedure followed in the cleaning of these fasteners are matters of general knowledge. To the contrary, the record demonstrates that such details as the types and amounts of chemicals to be used and the times and temperatures for cleaning the fasteners were developed by Ventura through extensive trial and error and at considerable expense.

Id. at 434. Similarly, in Sikes, 665 F.2d 731, 736, we quoted with approval the following passage from Water Services, Inc. v. Tesco Chemicals, 410 F.2d 163, 173 (5th Cir.1969).3

It is clear that the theory of the Treat-A-Matic was not new to the industry. The court also found that the design was apparent from inspection, and that the system had been sold and advertised without restriction. These findings are supported by the evidence. But these findings are not conclusive. The trade secret here was the application of known techniques and the assembly of available components to create the first successful system in the industry.... As the Second Circuit stated in Imperial Chemical, Ltd. v. National Distillers & Chemical Corp. (2 Cir.1965) 342 F.2d 737, 742, "a trade secret can exist in a combination of characteristics and components each of which, by itself, is in the public domain, but the unified process, design and operation of which in unique combination, affords a competitive advantage and is a protectible secret."

We believe these cases cogently refute appellees' assertions to the contrary.

The second defense advances a supposed concept of "negative know how" in reliance on one sentence from Hurst v. Hughes Tool Co., 634 F.2d 895 (5th Cir.) cert. denied 454 U.S. 829, 102 S. Ct. 123, 70 L. Ed. 2d 105 (1981). The court there ruled that Hughes Tool Company did not really use the information Hurst provided; "Hurst's information, while of some benefit, provided only negative, 'what not to do,' input to Hughes." Id. at 899. Both the district court and Smith interpret this to mean that negative know-how cannot constitute a trade secret. The sentence in Hurst, however, has nothing to do with whether the information was a trade secret. It is not found in the section of the opinion dealing with the existence of a trade secret; rather, we were determining whether Hughes Tool Company could be found to have used commercially any secrets Hurst conveyed.

Regardless of this misreading, the argument that negative know-how is involved here is unpersuasive. We do not understand how the changes that we have described can be seen as only showing what not to do. Metallurgical's evidence shows that it encountered many problems with the furnaces Therm-O-Vac delivered. Striving to solve these problems, Metallurgical took several steps to modify the furnaces. Installing unitary heating elements, chill plates, multiple crucibles, and pump filters are all "positive" steps resulting from conclusions on what to do. These changes allegedly turned unuseuable furnaces into ones fit for commercial operation. One may say, of course, that these actions result from realizations of what not to do; but so does every human process: the selection of one action at a given moment involves the rejection of every other conceivable one that might have been chosen. Using multiple crucibles, for example, stems from the conclusion that a single large crucible should not be used. This characterization, however, can always describe the invention or modification of a device. Knowing what not to do often leads automatically to knowing what to do. Although we decline to hold that this distinction will always be unavailing, in this case at least we regard the distinction between "positive" and "negative" knowledge to be unintelligible. Because this final claim of Smith is unpersuasive, we conclude that the evidence which was presented at trial could have led a reasonable jury to believe a trade secret existed.

Deciding whether a confidential relationship existed between Metallurgical and Bielefeldt must naturally precede an inquiry into his possible breach of Metallurgical's confidence. Once again, we look to the Restatement of Torts as our starting point:

One who discloses or uses another's trade secrets, without a privilege to do so, is liable to the other if ... (b) his disclosure or use constitutes a breach of confidence reposed in him by the other in disclosing the secret to him.

* * *

* * *

A breach of confidence under the rule stated in this Clause may also be a breach of contract which subjects the actor to liability.... But whether or not there is a breach of contract, the rule stated in this Section subjects the actor to liability if his disclosure or use of another's trade secret is a breach of the confidence reposed in him by the other in disclosing the secret to him.

Huffines, 314 S.W.2d at 769, quoting Restatement, Sec. 757 and comment j. In Huffines, the Texas Court cited several cases to emphasize this message. Quoted at great length was Adolph Gottscho v. American Marking Corp., 18 N.J. 467, 114 A.2d 438 (1955). The Gottscho court stressed the blameworthy conduct of the defendant as the basis of this tort:

[Defendants] now seek to appropriate these trade secrets to their own use and profit by a violation of their contractual agreements and a betrayal of the confidence reposed in them by plaintiff. This they may not do; such conduct is abhorrent to our conception of ordinary honesty.... Jackson learned the plaintiff's trade secrets in confidence and, in violation of his fiduciary obligations, he disclosed and used them for purposes other than his employer's benefit.... His conduct was grossly improper and gave rise to the plaintiff's cause of action, based on long-settled equitable principles and supported by the marked changes in the attitude of the law towards the need for commercial morality.

114 A.2d at 441-42, quoting Julius Hyman & Co. v. Velsicol Corp., 123 Colo. 563, 604-06, 233 P.2d 977, 999 (1952), cert. denied 342 U.S. 870, 72 S. Ct. 199, 96 L. Ed. 671 (1951). Our review of the evidence on the existence of a confidential relationship is hampered to some degree by the district court's exclusion of several items of evidence. As we discuss below, the exclusions were improper; but regardless of the evidence excluded, the record contains testimony of Metallurgical's president, Ira Friedman, that he informed Bielefeldt of the confidentiality Metallurgical expected. Although these references are few, they would have sufficed to allow a reasonable jury to have believed that a confidential relationship existed between Metallurgical and Bielefeldt.

At this point we must devote separate attention to Smith, which stands in a different light from Bielefeldt. It had no significant dealings with Metallurgical and apparently was not heavily involved in the design of the furnace it purchased. The question therefore becomes whether Smith as purchaser, and thus as beneficiary of Bielefeldt's alleged misappropriation, can also be held liable for it.

The law imposes liability not only on those who wrongfully misappropriate trade secrets by breach of confidence but also, in certain situations, on others who might benefit from the breach:

One who discloses or uses another's trade secret, without a privilege to do so, is liable to the other if ... (c) he learned the secret from a third person with notice of the facts that it was a secret and that the third person's disclosure of it was otherwise a breach of his duty to the other....

* * *

* * *

One has notice of facts under the rule stated in this Section when he knows of them or when he should know of them.... He should know of them if, from the information which he has, a reasonable man would infer the facts in question, or if, under the circumstances, a reasonable man would be put on inquiry and an inquiry pursued with reasonable intelligence and diligence would disclose the facts.

Restatement, Sec. 757 & comment 1. Under this standard, we believe a reasonable jury could find that Smith should have inquired into the relationship between Bielefeldt and Metallurgical. Testimony shows that, during negotiations for the purchase of a furnace, Bielefeldt told Smith of his current involvement in then-pending litigation with Metallurgical regarding trade secrets in New Jersey. Smith learned that Metallurgical claimed ownership of the design and manufacturing processes of the zinc recovery furnace, a furnace which Smith wished Bielefeldt to build. Apparently satisfied by Bielefeldt's assertion of the meritlessness of Metallurgical's claims, Smith eventually gave him the go-ahead for construction of the furnace. There is no indication that it ever investigated the danger that Bielefeldt was wrongfully misappropriating the ideas of others. The evidence as it stood at the end of Metallurgical's presentation thus suggests that Smith knew of possible problems and did nothing but rely on Bielefeldt's dismissals. We think that this inattention to possible wrongdoing, unless refuted, amounts to a failure to reasonably inquire into the facts involved. Under Sec. 757(c), Smith might therefore be held accountable, provided it used any trade secrets conveyed. This brings us to the next issue.

Wrongful misappropriation occurs if one "discloses or uses another's trade secret without a privilege to do so...." Restatement, Sec. 757. The district court directed verdict for appellees in part because it saw no evidence of Bielefeldt's actual use or disclosure of Metallurgical's secrets. In reviewing this conclusion, we keep in mind the rule of Boeing Co. v. Shipman by scouring the record for reasonable inferences favorable to Metallurgical. One fact jumps out from this review: in their original form, the furnaces delivered to Metallurgical differed from those that Smith purchased. The former furnaces lacked the key features needed to achieve commercial operation, while the latter possessed those features--features that Metallurgical had devised by extensive and expensive trial and error. Bielefeldt himself testified that he did not look to public sources of information in designing the Smith furnace; he instead claimed that he relied on his memory. That his earlier efforts lacked the features at issue suggests that his "memories" may well have been of working with Metallurgical. This issue is therefore an inappropriate ground for a directed verdict.

Smith's liability can arise, however, only if it in turn used the secrets gained from Bielefeldt. "Use," as it turns out, is not so easily defined. Smith claims that it never used any secrets gained because its inability to procure substantial quantities of scrap carbide prevented commercial operation of the furnace Fourtek provided. Lykes-Youngstown, 504 F.2d 518, guides us in determining commercial use. We must first recognize the unfortunate blurring of analyses in that case. The Lykes-Youngstown court's discussion of commercial use was in the context of inquiring whether damages might be available. It is preferable, of course, to divorce these concepts. Commercial use is an element of the tort as enounced in Sec. 757 of the Restatement; while the nature of the use may be relevant in determining the proper extent of damages, its existence must also be shown to establish wrongdoing in the first place. Despite this confusion, Lykes-Youngstown provides useful analysis.

Metallurgical looked to that case in arguing that the law provides a liberal definition of "commercial use." Lykes-Youngstown does indeed state a broad definition; "any misappropriation, followed by an exercise of control and dominion ... must constitute a commercial use...." 504 F.2d at 542. Lykes-Youngstown differs from our case, however, in one very important respect. It was a case in which "the trade secret itself was what was to be sold...." Id. at 540. The court there explicitly contrasted a case like ours, "where the trade secret is used to improve manufacturing, and subsequently manufactured items were sold at a profit...." Id. Although the court made this distinction in determining the proper method of computing damages, we think it also applies logically to developing a definition of "use." The discussion in Lykes-Youngstown following this distinction is therefore inapposite to our case, for which we instead employ the everyday meaning of the term. If Smith has not put the furnace into commercial operation to produce carbide powder it can then use, then no commercial use has occurred. Because Metallurgical failed to provide any evidence that Smith has so far benefitted from any misappropriation, directed verdict in Smith's favor was proper. Should it in future seek to profit from use or sale of the furnace, a new fact situation will be presented.

We now turn our attention to certain evidentiary rulings the district court made to exclude various documentary and testimonial items of evidence. At trial, Metallurgical offered Exhibit 246 in evidence; entitled "Non-Disclosure Agreement" and dated June 11, 1976, this document predates any purchase agreement between Metallurgical and Therm-O-Vac. In it, Bielefeldt agreed to hold all information gained from Metallurgical in confidence, using or disclosing this knowledge only with Metallurgical's approval.4 On July 30, 1976, Metallurgical and Therm-O-Vac signed the purchase order for the first furnace. Entered into evidence as Exhibit 826, the order form contained two clauses of significance here: an integration clause5 and one entitled "Property Interest."6 An analogous pair of documents also surrounds the purchase of the second furnace. Exhibits 247 and 247A are different copies of a Non-Disclosure Agreement dated March 1, 1978. Also signed by Bielefeldt, this agreement provided detailed recognition of the confidentiality involved and the expectation of nondisclosure absent Metallurgical's permission.7 On January 31, 1979, the parties signed the purchase order for the second furnace; this document, received in evidence as Exhibit 57, contained integration and nondisclosure provisions identical to those of Exhibit 826.

The district court eventually withdrew exhibits 246, 247, and 247A from evidence because of Texas' parol evidence rule, holding that the integration clauses of exhibits 826 and 57 "abrogated all prior agreements between the parties." The Texas parol evidence rule bars evidence of a prior or contemporaneous agreement that contradicts or changes the express terms of an unambiguous written agreement. See, e.g., Caviness Packing Co., Inc. v. Corbett, 587 S.W.2d 543, 545 (Tex.Civ.App.--Amarillo 1979, writ ref. n.r.e.). If Metallurgical were suing only on the contract, then the parol evidence rule unquestionably would bar evidence of prior agreements. This is not the case, however; the cause of action we are addressing on appeal is independent of any contract. See Huffines, 314 S.W.2d at 769. Even if the parties had not executed the purchase agreements, Metallurgical could have sued for misappropriation of its trade secrets. Exhibits 246, 247 and 247A are relevant in showing that this tort occurred. They were offered to show the parties' recognition of the unique nature of the zinc recovery process and the confidential relationship then existing; their purpose was not to change or undermine the agreements embodied in the purchase order. We regard their significance as stemming from legal concepts completely independent of any contractual relationship,8 their consequences, moreover, do not impinge upon the legal effect of that relationship. We thus conclude that the exclusion of these exhibits was an abuse of the district court's discretion.

The district court also excluded certain testimony of Friedman, Bielefeldt, and Paul Durkin, a former employee of Metallurgical. Because these exclusions were also premised on the applicability of the parol evidence rule, they too were improper. On appeal, appellees argue that the testimony was cumulative anyway, but we are unconvinced. On retrial, therefore, testimony about the parties' understanding of confidentiality existing before the purchase order agreements should not be excluded because of the parol evidence rule.

We may not disturb the district court's exclusion of the evidence, however, if that ruling can be upheld on other grounds, regardless of whether the court relied on those grounds. Fed.R.Evid. 103 requires a finding of error to be based on the exclusion of evidence only if "a substantial right of the party is affected." If other grounds exist supporting the evidence's exclusion, the district court's reliance on the parol evidence rule can constitute only harmless error. See United Communications, Inc. v. American Television & Communications Corp., 536 F.2d 1310, 1318 n. 6 (10th Cir.1976); 11 Wright & Miller, Federal Practice and Procedure: Civil Sec. 2885 (1973). It is thus necessary to address the best evidence rule and authentication challenges Bielefeldt raises regarding exhibits 247 and 247A. At trial, the court expressed concern about whether these exhibits satisfied the best evidence rule, but eventually did not base their exclusion on the rule. Bielefeldt now maintains, as he maintained at trial, that the exhibits' authenticity was insufficiently confirmed and that the exhibits were inaccurate copies of the original, which has apparently been lost.

Exhibit 247 is a poorly-xeroxed copy of the March 1, 1978 nondisclosure agreement. Because parts of the copy were difficult to read, Metallurgical introduced into evidence Exhibit 247A, a more legible reproduction. Both copies have unintelligible markings on the lower left hand corner. The markings do not cover the agreement's text or Bielefeldt's signature. Because they do not appear to alter or modify the typed text of the agreement, it is difficult, if not impossible, to see how their presence can affect the case in any way. While these markings might distinguish the copies from the original document, the alteration seems of a most trivial kind. The record therefore reveals no genuine issue as to the authenticity of the copies' contents; no suggestion of unfairness arises, moreover, by admitting the copies into evidence. The copies should not be excluded because of some small, unintelligible markings that apparently have nothing to do with the issues of this case. Rather, these markings should go to only the weight attached to the exhibits. See Greater Kansas City Laborers Pension Fund v. Thummel, 738 F.2d 926, 928 (8th Cir.1984).

Authentication of Exhibit 247 came from the testimony of Ira Friedman, who claimed to have been present when Bielefeldt signed the document. No testimony, however, authenticated Exhibit 247A. That exhibit appears to be a slightly enlarged and clearer copy of either the original or Exhibit 247. Authentication of Bielefeldt's signature can be made by comparing it to that of Exhibit 247, already authenticated. Had the case gone to the jury, it could have assumed this task. See Fed.R.Evid. 901(b)(3). We therefore refuse to affirm the ruling excluding these exhibits' exclusion on any other ground.

We now come to the issue of remedies available to Metallurgical. The district court apparently found crucial Smith's inability to operate its furnace profitably. Because there was no commercial use, it concluded that damages were unavailable. We have already concluded that Smith did not "use" the alleged secrets Bielefeldt provided; to say that this circumstance precludes all remedies goes too far, however. The court failed to distinguish consideration of the individual appellees; Smith is out of the picture, but Bielefeldt remains. Should he be found liable on retrial, the appropriate damages should be based on the tenets of Lykes-Youngstown. We there adopted the concept of the "reasonable royalty." This does not mean a simple percentage of actual profits; instead, the trier of fact, should it find Bielefeldt liable, must determine "the actual value of what has been appropriated." 504 F.2d at 537, quoting Vitro Corp. v. Hall Chemical Co., 292 F.2d 678, 683 (6th Cir.1961). We later expounded this concept:

[T]he proper measure is to calculate what the parties would have agreed to as a fair price for licensing the defendant to put the trade secret to the use the defendants intended at the time the misappropriation took place.

In calculating what a fair licensing price would have been had the parties agreed, the trier of fact should consider such factors as the resulting and foreseeable changes in the parties' competitive posture; the prices past purchasers or licensees may have paid; the total value of the secret to the plaintiff, including the plaintiff's development cost and the importance of the secret to the plaintiff's business; the nature and extent of the use the defendant intended for the secret, and finally whatever other unique factors in the particular case might have been affected by the parties' agreement, such as the ready availability of alternative process.

Id. at 540. Estimation of damages, however, should not be based on sheer speculation. If too few facts exist to permit the trier of fact to calculate proper damages, then a reasonable remedy in law is unavailable. In that instance, a permanent injunction is a proper remedy for the breach of a confidential relationship. Zoecon Industries, 713 F.2d at 1180, citing Huffines, 314 S.W.2d at 778. In this case, moreover, an injunction against Bielefeldt in no way depends on whether Smith achieved commercial use. Should Metallurgical prove its point on retrial, therefore, it has a remedy available to right the wrong done it.

We emphasize the limited consequences of today's decision. In no way do we pass judgment on whether Bielefeldt committed any wrong; that task is for the trier of fact. We hold only that the opportunity for this determination to be made must be given. It may be that on retrial Bielefeldt can successfully show that his knowledge came from somewhere else or that the process of modification was no secret.

We conclude by recognizing the other appellees in this case. Our decision has focused exclusively on Bielefeldt and Smith. Metallurgical, however, also named the other officers of Fourtek as appellees.9 It thereby forced Norman Montesino, Gary Boehm, and Michael Sarvadi to file an appellate brief. Only in their brief do we learn that Metallurgical itself moved to dismiss all counts against these three, a motion which the court granted. No evidence in the case was presented against them. Only in Metallurgical's reply brief, moreover, do we discover that "Metallurgical does not urge this appeal as to Montesino, Boehm, and Sarvadi...." We regard the naming of these three as appellees, thus forcing them to submit a brief, as constituting a frivolous appeal. We therefore award attorney's fees and single costs to Montesino, Boehm, and Sarvadi in accordance with Fed.R.App.P. 38. On remand, they are to submit their claims for fees to the district court.

The district court's order is AFFIRMED in part, REVERSED in part, and the cause is REMANDED.