When this case accusing Motorola of violating patent, contract, and trade secret rights became one of the oldest on its docket, the district court decided to expedite matters by splitting the case in two: the court required Ole Nilssen, the plaintiff, to file a new patent action (No. 96 C 5571) and retained the state- law theories under the original docket number (No. 93 C 6333), thinking that these could be handled quickly. Unfortunately this move backfired, extending rather than abbreviating the litigation. The patent action lingers on the district court's docket, though it looks to the statisticians at the Administrative Office of the United States Courts like a new and independent case. Meanwhile the contract and trade secret theories of relief, dismissed by the district court at the end of 1998, have wandered among appellate courts, looking for a home--a process complicated not only by the division between the regional circuits and the Federal Circuit, but also by the fact that the issues separated for appeal do not meet the criteria for dispatch under Fed. R. Civ. P. 54(b).

Final decisions in all cases originally based on federal patent law must be appealed to the Federal Circuit. 28 U.S.C. sec. 1295(a)(1). Jurisdiction of Nilssen's suit was based on 28 U.S.C. sec. 1338, so it looked like sec. 1295(a)(1) directed appeals to that court. Yet the district court had not resolved the patent issues, and the mitosis of this case produced one offspring relying entirely on state law. After concluding that jurisdiction in this judicially created case fragment did not arise under sec. 1338, a divided panel of the Federal Circuit concluded that the appeal should be transferred to the Seventh Circuit. Nilssen v. Motorola, Inc., 203 F.3d 782 (Fed. Cir. 2000). After filing lengthy briefs in this court, the parties discovered at oral argument that we too were restive about jurisdiction--both the district court's and ours. If one case really has become two, that ensures appellate jurisdiction, for the district judge entered a final decision in the state-law fragment. But if the proceedings now on appeal are a distinct case arising under state law, what is the basis of subject-matter jurisdiction in the district court? Motorola has its principal place of business in Illinois, and the complaint alleges that Nilssen is a "resident" of Illinois. He sought to pursue the state- law theories under the supplemental jurisdiction--all well and good, as long as the patent theory is in the case. But if the patent issues are gone--not just dismissed, after which the district court could have retained supplemental jurisdiction under 28 U.S.C. sec. 1367(c), but missing in the sense that sec. 1338 never supplied jurisdiction to begin with--then with no diversity of citizenship the case does not belong in federal court. And it looks like the Federal Circuit's decision establishes that sec. 1338 just is not in the picture, that the state-law fragment should be treated as a free-standing case that had never depended on sec. 1338 for federal jurisdiction. Whether or not that is right (Judge Rader's dissenting opinion in the Federal Circuit questioned it), it is the law of the case. See Christianson v. Colt Industries Operating Corp., 486 U.S. 800 (1988). If on the other hand Judge Rader's views were to prevail, then the appeal is in the wrong circuit--and probably should not be in any circuit, for "the case" (viewed as the original, pre-division unit invoking sec. 1338) would still be live in the district court.

Responding to our request at oral argument, the parties filed supplemental memoranda concerning jurisdiction. The first and most important revelation is that the allegations of the complaint are misleading: Nilssen may have been a "resident" of Illinois when the complaint was filed, but he was a domiciliary (and hence a "citizen") of Florida. It is cit izenship that counts for purposes of the diversity jurisdiction. Steigleder v. McQuesten, 198 U.S. 141 (1905); Denny v. Pironi, 141 U.S. 121 (1891); Robertson v. Cease, 97 U.S. 646 (1878). Nilssen has tendered papers under 28 U.S.C. sec. 1653 to show the correct jurisdictional facts. See Newman-Green, Inc. v. Alfonzo- Larrain, 490 U.S. 826 (1989). Motorola agrees that the new allegation of "citizenship" is correct. We have no reason to think that the parties are collusively asserting jurisdiction, so we grant Nilssen permission to amend his complaint, as he proposes. That resolves the question of subject-matter jurisdiction--and incidentally resolves appellate jurisdiction too, for even if the district court should not have divided the cases, the division is a fait accompli and Nilssen had to appeal immediately. It does not follow, however, that we should allow the division to continue. For reasons that will become apparent, this case must be re-integrated in the district court.

Nilssen is an inventor--a litigious inventor, as a check of Lexis or Westlaw will reveal. He has been unsuccessful in all published decisions and has been threatened with sanctions for making frivolous claims, see In re Nilssen, 851 F.2d 1401, 1402 n.1 (Fed. Cir. 1988); In re Nilssen, 1991 U.S. App. Lexis 9796 at *4 n.2 (Fed. Cir. May 8, 1991), but in this case he hopes for a brighter day. Formerly an employee of Motorola, Nilssen was fired in 1972. That, plus his penchant for litigation, might have led Motorola to send him packing when, in 1982, Nilssen wrote to Robert Galvin, Motorola's ceo, with a proposal. Actually Motorola did turn him away, several times, but in 1987 it decided to entertain Nilssen's ideas for electronic ballasts. (A ballast stabilizes electrical current. Ballasts are vital to the operation of fluorescent lamps, providing the high voltage necessary to ionize the gas and get the lamp going, then limiting the current's flow once the lamp is lit. Standard ballasts use electrical components such as transformers, resistors, and capacitors; an electronic ballast relies instead on transistors.)

Galvin's decision to open negotiations and receive information from Nilssen under a promise of confidentiality, coupled with Motorola's later decision to make and sell electronic ballasts without paying Nilssen, led to this lawsuit--now two lawsuits, plus at least one related suit against another firm. See Nilssen v. MagneTek, Inc., 130 F. Supp. 2d 985 (N.D. Ill. 2000). Nilssen contends that when Motorola entered the electronic-ballast business in 1989 it either stole his ideas (reflected in patents plus trade secrets revealed during the negotiations) or, by making ballasts without using his ideas, broke a promise to share with him a substantial portion of any profits. Nilssen contends that his technology was worth at least $50 million to Motorola, and he demands that sum in damages.

Since the 1970s Galvin had been evaluating electronic ballasts as a line of business that Motorola might want to enter. Until 1989, however, Motorola thought its chances of success dim. Nilssen believes that the technology he disclosed to Motorola during negotiations in the late 1980s made the difference; Motorola, by contrast, contends that Nilssen's ideas did not pan out and that it is now using technology developed by Carlile Stevens without any reliance on Nilssen's patents or trade secrets. After diverting the patents into a separate suit, the district judge turned to Nilssen's state-law theories. In April 1997 the court concluded that several of Nilssen's theories are preempted by the Illinois Trade Secrets Act, 765 ILCS 1065/1 to 1065/9, but that theories under this statute were viable. 963 F. Supp. 664 (N.D. Ill. 1997). In August 1997 the judge removed several varieties of trade secret theories from the case. 1997 U.S. Dist. Lexis 12899 (N.D. Ill. Aug. 22, 1997). A year later the court carved off still more elements of Nilssen's trade- secret theories. 1998 U.S. Dist. Lexis 12882 (N.D. Ill. Aug. 12, 1998). Next it entered an order barring testimony by Horace DePodwin, Nilssen's principal damages expert. 1998 U.S. Dist. Lexis 19161 (N.D. Ill. Nov. 30, 1998). DePodwin's testimony would be useless, the court concluded, because he did not try to identify those portions of Nilssen's loss that flowed from the theories of liability remaining in the case. Instead, the judge observed, DePodwin tried to estimate Motorola's gains (and thus areasonable royalty) from using all of Nilssen's intellectual property to which Motorola had access, including the patented inventions that had been carved off for the separate suit and the trade secrets that had been dealt with in the August 1998 opinion (which had been filed after DePodwin prepared his report). Because this ruling left Nilssen with no practical way to recover damages, the district court concluded that he was no longer entitled to a jury trial; the only remaining theories were equitable. At a status conference on December 29, 1998, Nilssen withdrew his request for injunctive relief. The next day, reading an opinion from the bench, the judge explained that this left Nilssen with no case and entered judgment for Motorola.

Two features of this history show why the separation of the patent from the state-law theories has led to trouble. One is the effect of the split on the damages calculation. The district judge demanded that DePodwin calculate trade- secret damages independently of patent damages. That task was necessitated by the segregation of legal theories into separate lawsuits, yet the district judge concluded that DePodwin had been unable to accomplish it (though DePodwin tried). The district judge never explained how it would have been possible (or practical) to calculate trade-secret damages on the assumption that Motorola did not infringe any of Nilssen's patents, or patent damages on the assumption that Motorola did not use any of Nilssen's trade secrets. In the parties' negotiations-- and, Nilssen insists, in Motorola's creation of its electronic ballasts--the trade secrets and patents were tied together, if only because the trade secrets concern the use of ideas reflected in the patents.

Second was the effect on the merits of the district court's rejection of DePodwin's damages report. This left Nilssen without a claim for legal relief and scuttled his request for a jury, which in turn sank the suit. But if the patent and trade secret theories were united in a single suit, damages claims would have remained, and Nilssen would have been entitled to a jury--at least on whatever elements survived summary judgment. When legal and equitable theories are joined in one suit, the jury's resolution of questions common to the two branches controls. Beacon Theatres, Inc. v. Westover, 359 U.S. 500 (1959); Dairy Queen, Inc. v. Wood, 369 U.S. 469 (1962). The district judge's order cleaving the case in two had the practical consequence of nullifying Nilssen's entitlement to a jury trial on issues common to its legal and the equitable aspects.

There is a third consideration--one apparent from the opening paragraphs of this opinion. By splitting the patent and trade-secret theories into two suits, the district court also directed the eventual appeals to two different courts. If things remain as they are, we will hear the state-law theories, and the Federal Circuit the patent theories. This would be bad enough if it required two appellate courts to master the intricacies of electronic ballasts, the Nilssen-Motorola negotiations, and the differences between the Nilssen and Stevens proposals for ballast design. The joint appendix in this appeal exceeds 4,000 pages, the record is very much longer, and we do not doubt that the patent appeal to the Federal Circuit will sacrifice additional reams of paper. Squandering judicial resources by requiring six appellate judges (at least two panels of two circuits) to master this material should be avoided. Moreover, as the district judge himself recognized when dismissing the suit (at least, the state-law fragment that the judge had created), Nilssen has only one "claim for relief." He does not have a patent claim based on one set of events and a trade secret claim based on different events. There is only one series of negotiations, disclosures, and product-design decisions. Nilssen contends that what Motorola did violates several statutes and common-law doctrines, but there is only one nucleus of operative facts, and thus only one "claim." See NAACP v. American Family Mutual Insurance Co., 978 F.2d 287, 291- 92 (7th Cir. 1992). That is why the state-law theories came within the supplemental jurisdiction, which depends on such a tight link among state and federal theories that they form part of a single controversy. 28 U.S.C. sec. 1367(a). It makes no sense to send patent-law theories to the Federal Circuit and state-law theories to the regional circuit, when only one "claim" is involved, and one circuit's decision may have preclusive effect on issues within the scope of the other's jurisdiction. Avoiding claim-splitting (with the waste of judicial resources and the potential for inconsistent appellate decisions) is why sec. 1295(a)(1) provides that, if jurisdiction is based on sec. 1338 even in part, then the Federal Circuit resolves all issues on appeal. The Federal Circuit concluded that the district court had all but defeated this aspect of sec. 1295 by turning one case into two; but it is within our power to pull victory from the jaws of defeat by turning the two cases back into one. See 28 U.S.C. sec. 2106.

Because sensible handling of this case includes only one appeal, to only one circuit--which means, to the Federal Circuit--we resist all temptation to comment on the issues covered in the current round of appellate briefs. Everything will be open to consideration on a new appeal after the patent issues have been wrapped up. But we do think that it would be prudent for the district court to take a fresh look at the admissibility of DePodwin's testimony once patent and trade-secret theories are reunited. In other words, Judge Darrah, who took over the patent-law fragment of the suit on his appointment to the bench last year, should not block DePodwin's testimony solely by invoking the law of the case. Moreover, it might make sense for the district court to allow DePodwin to submit a new damages report in light of Judge Shadur's rulings in August 1998 and any later developments with respect to the patent theories.

The judgment of the district court is vacated, and the case is remanded with instructions to consolidate this proceeding with the patent-law proceeding that had been spun off as No. 96 C 5571. The cases must be rejoined for all purposes, including any appeal from the final judgment. The district court should endeavor to consolidate all of Nilssen's actions concerning electronic ballasts before a single judge; Judge Darrah can and should initiate that process under local rules.