May a question “arise under” the patent laws, thus creating federal jurisdiction in the district court, but not “arise under” the patent laws for purposes of appellate jurisdiction? This is the question we must answer. No other court has faced it yet. Christianson v. Colt Industries Operating Corp., — U.S. —, 108 S.Ct. 2166, 100 L.Ed.2d 811 (1988), holds that jurisdiction under the patent laws in the district court is a necessary condition of the Federal Circuit’s appellate jurisdiction. Is it also a sufficient condition? Two Justices in Christianson said no, id. — U.S. at- -, 108 S.Ct. at 2178-82, (Stevens, J., joined by Blackmun, J.); the others did not directly address the question.
Samuel F. Kennedy received patents Nos. 3,583,112 and 4,073,110 for two inventions concerning the design of grain bins. The patents were practiced by S.F. Kennedy Industries and its subsidiary New Products, Inc., closely-held companies of which Kennedy and his family were the principal owners. The venture was successful for some time but in 1982 was liquidated under Chapter 7 of the Bankruptcy Code. The trustee in bankruptcy sold New Products’ name and assets, including “[a]ny and all patents or trademarks”. Robert Wright paid $1.25 million for these assets. On the morning of the closing, Kennedy purported to rescind his firm’s rights to practice the two patents, which he asserted depended on licenses terminable at will. Kennedy and Wright have been feuding ever since.
In 1984 Kennedy and Duane Young as trustee of a trust claiming title to the patents (collectively Kennedy), filed a complaint against Wright and Wright’s two firms (collectively Wright). This admirably short document asserted that the two patents were issued to Kennedy, that they are valid, and that Wright is infringing them. It sought a declaratory judgment that the patents are valid and infringed, an injunction against further infringement, and an accounting for damages and profits. The complaint invoked the district court’s jurisdiction under 28 U.S.C. § 1338(a). Wright’s answer copied out almost all of the statutory grounds of invalidity in 35 U.S.C. §§ 102, 103, and 112, asserting (for example) that each patent is invalid as obvious. Wright also filed a counterclaim, ask*965ing the court to declare that he owns the two patents, having purchased them from the trustee in bankruptcy. (Wright did not say what he planned to do with these patents if, as he insists, they are invalid.) This compulsory counterclaim invoked the district court’s ancillary jurisdiction. This may have been necessary, because all parties are citizens of Ilinois, although an aggrieved licensee sometimes may invoke the jurisdiction of § 1338 directly. Excelsior Wooden Pipe Co. v. Pacific Bridge Co., 185 U.S. 282, 22 S.Ct. 681, 46 L.Ed. 910 (1902). Either way, § 1338 supplied the whole of the district court’s authority.
The district court bifurcated the case into validity and ownership components. After a bench trial on the ownership dispute, the court entered judgment for Wright. 676 F.Supp. 888 (C.D.Ill.1988). The court found that S.F. Kennedy Industries and New Products, Inc., were the equitable owners of the two patents in 1982 even though they were issued to Kennedy personally and no change of ownership had been recorded. Kennedy did the work on company time, the court concluded, and the firms amortized the expenses of inventing, patenting, and developing the products. The district court found no evidence to support Kennedy’s contention that he held title to the patents and merely licensed the firms to use them. Because the firms equitably owned the patents, they passed to Wright by virtue of the trustee’s sale in bankruptcy. The court directed Kennedy to transfer legal title to Wright.
Kennedy filed a notice of appeal to this court under 28 U.S.C. § 1291. Wright has moved to dismiss it on the ground that the Federal Circuit has exclusive jurisdiction under 28 U.S.C. § 1295(a), which governs:
(1) ... an appeal from a final decision of a district court of the United States ... if the jurisdiction of that court was based, in whole or in part, on section 1338 of this title, except [for copyright and trademark eases].
As Wright points out, the jurisdiction of the district court was based “in whole” on § 1338. The plain language of § 1295(a)(1) puts this case into the Federal Circuit, according to Wright. Kennedy stoutly maintains that this court has jurisdiction because the district court resolved the case without a peep about patent law. We asked the parties to brief the question whether the regional circuits have jurisdiction of appeals in cases filed under the patent laws, when district courts dispose of them without construing those laws. After the parties filed these briefs, we deferred submission of the case pending the Supreme Court’s decision in Christianson.
The language of § 1295(a)(1) is indeed plain, its application here straightforward. This began as a real patent case properly before the district court under the “well-pleaded complaint” rule of Louisville & Nashville R.R. v. Mottley, 211 U.S. 149, 29 S.Ct. 42, 53 L.Ed. 126 (1908), anticipated for patent cases by Pratt v. Paris Gas Light & Coke Co., 168 U.S. 255, 259, 18 S.Ct. 62, 42 L.Ed. 458 (1897), and reiterated a few days ago in Christianson. Kennedy wanted relief that he could get only if he had valid patents, and he called on the district court to vindicate his entitlements. Although the district court resolved the case without disposing of any patent issue, the jurisdiction of that court was fixed by what the complaint contained, not by how it resolved the dispute. Section 1295(a)(1) says that if the district court’s jurisdiction rested on § 1338, then the Federal Circuit has exclusive jurisdiction of the appeal. That, it seems, is that.
Justice Stevens, concurring in Christian-son, contended that “Congress could not have intended precisely the same analysis” (— U.S. at-, 108 S.Ct. at 2180-81) of arising-under and appellate jurisdiction. He gave two examples of cases in which allocating an appeal among circuits according to the source of jurisdiction at the moment the complaint was filed would- not serve any of the purposes that led to the creation of the Federal Circuit’s exclusive patent jurisdiction. In one, the complaint made a claim under the patent laws, which was dismissed before trial; the case actually litigated would be a non-patent case suitable for review in the regional circuits. In the other example the complaint initially filed did not contain any patent claim, *966but one was added before trial and became the basis of decision, making it a patent case suitable for review in the Federal Circuit. In our case the patent claim has not been dismissed, but patent law had nothing to do with the district court’s decision. Perhaps Justice Stevens’ assessment of § 1295(a)(1) covers our case too.
The Federal Circuit’s exclusive jurisdiction under § 1295(a)(1) was created, after all, so that there could be a uniform jurisprudence of patent law. Whether Kennedy or Wright owns the two patents seems to depend on Illinois corporate and property law (conceivably on bankruptcy law, in light of the method of their transfer, if the Code has an applicable provision superseding state law) rather than on patent law. The regional circuits are better situated than the Federal Circuit to mull over questions of local law; “uniformity” interests cut in favor of distributing state law issues to courts with geographic jurisdiction, even as they support central handling of patent questions. Judge Friendly’s influential opinion in T.B. Harms Co. v. Eliscu, 339 F.2d 823 (2d Cir.1964), held that a dispute about the ownership of a copyright did not “arise under” the copyright laws because nothing unique to copyright was entailed. Ownership “will [not] be supported if the Constitution or laws of the United States are given one construction or effect, and defeated if they receive another”. Gulley v. First National Bank, 299 U.S. 109, 112, 57 S.Ct. 96, 97, 81 L.Ed. 70 (1936). Our case entails questions about validity as well as ownership, so Kennedy’s claim for relief arises under the patent laws, but perhaps the approach in Harms is well suited to determining appellate jurisdiction.
Such an approach would not be the first time an apparently “clear” statute has been modified to make appellate jurisdiction depend on the ground the district court adduced for its decision. Until 1974, 28 U.S.C. § 2281 (1970) required a three-judge district court to be convened whenever the complaint sought an injunction against the operation of a state statute on account of its unconstitutionality. Section 1253, which is still in force, provided for appellate jurisdiction:
Except as otherwise provided by law, any party may appeal to the Supreme Court from an order granting or denying ... an ... injunction in any civil action, suit or proceeding required by any Act of Congress to be heard and determined by a district court of three judges.
What to do when the three-judge district court issued an injunction against the statute on grounds other than its unconstitutionality — on the ground, for example, that it was preempted by a federal law? Such a case was “required ... to be heard and determined” by a three-judge court, because the complaint asked for an injunction on a constitutional ground, and the Supreme Court employed the well-pleaded complaint rule to determine whether a three-judge court had to be convened. Costello v. Wainwright, 430 U.S. 325, 97 S.Ct. 1191, 51 L.Ed.2d 372 (1977); Moody v. Flowers, 387 U.S. 97, 104, 87 S.Ct. 1544, 1549, 18 L.Ed.2d 643 (1967). After vacillating for decades, the Court decided that § 1253 did not establish jurisdiction of appeals unless “such order rests upon resolution of the merits of the constitutional claim presented below.” MTM, Inc. v. Baxley, 420 U.S. 799, 804, 95 S.Ct. 1278, 1281, 43 L.Ed.2d 636 (1975); see also Gonzalez v. Automatic Employees Credit Union, 419 U.S. 90, 95 S.Ct. 289, 42 L.Ed.2d 249 (1974). This was not the only creative device employed to limit the drain the compulsory jurisdiction placed on the Court’s resources. E.g., Swift & Co. v. Wickham, 382 U.S. 111, 86 S.Ct. 258, 15 L.Ed.2d 194 (1965) (complaints resting only on the Supremacy Clause do not even require the convening of a three-judge court); Goosby v. Osser, 409 U.S. 512, 518, 93 S.Ct. 854, 858-59, 35 L.Ed.2d 36 (1973) (no direct appeal if the constitutional claim is “obviously frivolous”); Gerstein v. Coe, 417 U.S. 279, 94 S.Ct. 2246, 41 L.Ed.2d 68 (1974) (no direct appeal if the district court issues a declaratory judgment and threatens an injunction only if the state does not capitulate). Perhaps we should give the same sort of reading to § 1295(a)(1) that MTM gave to § 1253, making appellate jurisdic*967tion turn on the ground of decision rather than the basis of the claim for relief.
We think not. The history of § 1253 tells more than one lesson. It reveals that a court can make appellate jurisdiction turn on the ground of decision. It reveals with equal clarity the costs of that choice. Section 1253 produced “a large body of confusing case law in which the Court ‘has not infrequently been induced to retrace its steps' ”, Robert L. Stern, Eugene Gress-man & Stephen M. Shapiro, Supreme Court Practice 72 (6th ed. 1986) (quoting from Gonzalez, 419 U.S. at 95, 95 S.Ct. at 292-93). The language of §§ 1253 and 2281 was “manhandled”, MTM, 420 U.S. at 805, 95 S.Ct. 1281-82 (White, J., concurring), as the Supreme Court tried to pare back its obligatory jurisdiction in order to make room for more-important cases. The ensuing lines — including the line, approved in MTM, based on the district court’s ground of decision — baffled lawyers and led to extended jurisdictional conundrums. The Freund Committee concluded:
When, where, and how to obtain appellate review of an order by or relating to a three-judge court is a hopelessly complicated and confused subject that in itself has produced much unnecessary litigation. Judicial and other literature on the subject is voluminous. There are rules and sub rules and exceptions to rules. ... [Rjeview of these matters has become so mysterious that even specialists in this area may be led astray.
Report of the Study Group on the Case Load of the Supreme Court 28-29 (1972). No one views the treatment of § 1253 as the Court’s finest hour — even as excusable, but for the need to make room on the docket for the most pressing eases of the day.
To construe § 1295(a)(1) as MTM construed § 1253 is to duplicate the confusion and uncertainty that gobbled up time for decades under the ill-starred three-judge court procedure. And why? To illustrate Santayana’s aphorism that “Those who cannot remember the past are condemned to fulfil it.”? I Life of Reason ch. 12 (1906). Surely not to save the time of the Federal Circuit, which unlike the Supreme Court can be expanded to handle additional business; surely not to direct cases to their “natural” tribunals, which for Kennedy’s appeal would be the Appellate Court of Illinois — to which we cannot send it no matter what we make of § 1295(a)(1). The Supreme Court reminded us recently that the rules governing appellate jurisdiction “should above all be clear ... Courts and litigants are best served by [a] bright-line rule”. Budinich v. Becton Dickinson & Co., — U.S. —, 108 S.Ct. 1717, 1722, 100 L.Ed.2d 178 (1988). We have said the same thing, frequently. E.g., Coté v. Wadel, 796 F.2d 981, 983 (7th Cir.1986) (“jurisdictional rules should be as simple as possible”); Exchange National Bank v. Daniels, 763 F.2d 286, 292 (7th Cir.1985) (“The first characteristic of a good jurisdictional rule is predictability and uniform application.”) (the Supreme Court approved the holding of Daniels in Budinich); In re Kilgus, 811 F.2d 1112, 1117 (7th Cir.1987) (“The more mechanical the application of a jurisdictional rule, the better. The chief and often the only virtue of a jurisdictional rule is clarity.”); Electrical Workers v. ICC, 832 F.2d 91, 93 (7th Cir.1987) (“a mechanical assignment [of eases among courts of appeals] allows everyone to address the merits more quickly”). Electrical Workers and Wronke v. Marsh, 767 F.2d 354 (7th Cir.1985), like other recent decisions dealing with the allocation of cases among circuits, stress the importance of simple rules that conserve the time that courts and counsel must devote to jurisdictional questions, so that more time is available to deal with the merits.
It is not hard to appreciate that trying to make appellate jurisdiction track the ground of the district court’s decision could complicate matters. What if the district court should give multiple or alternative bases of decision? (“I hold that Wright owns the patents; but if I did not believe this, I would hold that they are invalid.”) What if a contract respecting ownership cannot be construed without at least some reference to patent law principles? It would be very hard to discern the “true” ground of decision. What if there should *968be a dispute about whether a given ground of decision is a patent question? (“I hold that Wright owns the patents under Illinois law; and to the extent patent law influences the ownership dispute, I reach the same conclusion.”)
This is not a fanciful hypothetical. Despite our characterization of the ownership question as one of state law, at least some of the rules governing interests in and contracts about patents derive from federal law. Crown Co. v. Nye Tool & Machine Works, 261 U.S. 24, 43 S.Ct. 254, 67 L.Ed. 516 (1923), was a dispute turning on the validity of a purported assignment of a patent. The Court held that the dispute arose under federal law. Cf. Excelsior Wooden Pipe. The Patent Office maintains a title registration system. 37 C.F.R. § 1.331. Although this system is designed to protect creditors and subsequent purchasers, any federal filing system conceals the potential to produce federal questions out of otherwise-mundane ownership disputes. E.g., Philko Aviation, Inc. v. Shacket, 462 U.S. 406, 103 S.Ct. 2476, 76 L.Ed.2d 678 (1983), after remand, 841 F.2d 166 (7th Cir.1988). For other interactions between contract and patent law, see, e.g., Lear, Inc. v. Adkins, 395 U.S. 653, 89 S.Ct. 1902, 23 L.Ed.2d 610 (1969); Sola Electric Co. v. Jefferson Co., 317 U.S. 173, 63 S.Ct. 172, 87 L.Ed. 165 (1942); Unarco Industries, Inc. v. Kelley Co., 465 F.2d 1303, 1305-06 (7th Cir.1972); E.F. Drew & Co. v. Reinhard, 170 F.2d 679, 682 (2d Cir.1948) (L. Hand, J.); but cf. Aronson v. Quick Point Pencil Co., 440 U.S. 257, 99 S.Ct. 1096, 59 L.Ed.2d 296 (1979). The district court raised the possibility that federal law affects the outcome of the ownership squabble but wrote: “The dispute ... becomes academic for this Court can discern no difference in the law applied by the apposite decisions of the respective jurisdictions.” 676 F.Supp. at 891. Maybe; maybe not. ' The Federal Circuit has the last word, if ownership is a specialty of patent law, yet it has not said much about the subject. Beghin-Say International Inc. v. Rasmussen, 733 F.2d 1568 (Fed.Cir. 1983), holds that a “pure” ownership dispute does not arise under § 1338(a) but does not address the many possible variations. Suppose the regional circuit should conclude that ownership has patent-law attributes, making this a question suitable for the Federal Circuit — only to have that court conclude under Beghin-Say that ownership depends on state law? The number of other similarly unhappy interactions between jurisdiction and the merits is limited only by the imagination of judges and litigants.
Even if we were confident that a court quickly could separate patent from non-patent grounds of decision, we could not prevent a game of jurisdictional ping-pong. Imagine this sequence: the district court decides (as it has) that Wright owns the patents; we reverse and remand for a trial of the patent issues (presumably we would be forbidden to affirm on a patent ground not reached by the district court, even though such alternative-ground affirmanc-es are time-saving conveniences within a unitary system, see Massachusetts Mutual Life Insurance Co. v. Ludwig, 426 U.S. 479, 96 S.Ct. 2158, 48 L.Ed.2d 784 (1976)); on remand the district court holds the patents valid but not infringed; the Federal Circuit (which undoubtedly would have jurisdiction of an appeal from that decision) reverses on the ground that the patents are valid and infringed; on remand the district court concludes that bankruptcy law gives Wright some entitlement to use the valid patents without license; on appeal to this court from the disposition of the bankruptcy issue, we hold ... (and after we were done, there would be a final appeal to the Federal Circuit concerning royalties). Any considerations of convenience to the parties or to the court would be out the window. A court that becomes familiar with the case ought to hold on to it; if the ground of decision determines appellate jurisdiction, however, judicial time will be squandered.
Section 1295(a)(1) is too straightforward to permit (let alone require) a reading that would land the judiciary in such a bog. It gives the Federal Circuit exclusive jurisdiction of any case in. which the district court’s jurisdiction is based on the patent laws. Christianson holds that the principles of *969the well-pleaded complaint that determine the district court’s jurisdiction also determine appellate jurisdiction. — U.S. at -n. 2, 108 S.Ct at 2172-74 n. 2. Chris-tianson also reaffirms the time-honored rule that whether a complaint arises under federal law depends “on claims, not theories”. Christianson, — U.S. at -, 108 S.Ct. at 2174-76, citing Franchise Tax Board v. Construction Laborers Vacation Trust, 463 U.S. 1, 26 & n. 29, 103 S.Ct. 2841, 2855 & n. 29, 77 L.Ed.2d 420 (1983). Nothing in the legislative history of § 1295 suggests that the statute should receive other than a natural reading. Under Christianson a regional court could end up resolving patent issues in a case in which the complaint was based on other law; it follows that the Federal Circuit ends up with non-patent issues in a case based on the patent laws. We need not decide how Justice Stevens’ two hypotheticals should be treated; ours is easier.
The appeal is transferred to the United States Court of Appeals for the Federal Circuit under 28 U.S.C. § 1631.