POSNER, Circuit Judge.

Martha Olson had a commodity-trading account with Paine Webber, a brokerage house. She incurred a loss of some $18,000 in trading silver futures and, attributing the loss to a failure by her account executive at Paine Webber to follow her instructions, brought this suit for damages under the Commodity Exchange Act, 7 U.S.C. § 1 et seq. Paine Webber moved the district judge to stay the case pending arbitration of the dispute and to order arbitration, forms of relief expressly authorized by sections 3 and 4, respectively, of the United States Arbitration Act, 9 U.S.C. § 3, 4. The motion was based on a clause in the brokerage agreement between Olson and Paine Webber requiring arbitration of disputes. Olson opposed the motion on the ground that the clause was invalid because it violated regulations of the Commodity Futures Trading Commission. The district judge reformed the agreement to make it comply with the regulations and then enforced it by granting Paine Webber's motion to stay proceedings in the district court and by ordering arbitration. 627 F. Supp. 1317. The arbitration has not been conducted, however. Olson has appealed to this court, and Paine Webber has not pressed for arbitration before the appeal is decided.

The first question we take up is whether we have jurisdiction of this interlocutory appeal. The general rule in the federal as in virtually all state court systems is that only final judgments can be appealed. 28 U.S.C. § 1291. The reason is to conserve the time of litigants and appellate judges by limiting the parties, so far as possible, to one appeal per case and by ensuring the fullest possible ventilation of issues in the trial court before that appeal is taken, in the hope that maybe none will be taken, or, if one is, that the appellate court will only have to worry about dispositive issues. Neither of the orders that Olson is trying to appeal at this time (the order staying proceedings and the order to arbitrate) is a final, dispositive order. Her lawsuit remains pending in the district court. Should she lose the arbitration and fail to persuade the district court to set aside the arbitration award, and should her damage suit be dismissed as a result, there would then be a final judgment in the district court from which she unquestionably could appeal, raising among other issues the question whether the arbitration clause was valid. Of course, she might not appeal. The facts brought out at the arbitration hearing might convince her that Paine Webber had not violated the brokerage agreement after all. Or she might prevail in the arbitration, in which event her complaint that the dispute was not arbitrable would be moot. The fact that interlocutory orders in a litigation are frequently rendered moot by the final judgment in the trial court is powerful support for the final judgment rule. And however we decide this interlocutory appeal, we can't be sure it will be the last time we see this case.

The only basis for this appeal is the Enelow-Ettelson doctrine. Under this doctrine, named after Enelow v. New York Life Ins. Co., 293 U.S. 379, 55 S.Ct. 310, 79 L.Ed. 440 (1935), and Ettelson v. Metropolitan Life Ins. Co., 317 U.S. 188, 63 S.Ct. 163, 87 L.Ed. 176 (1942), an equitable stay (and a stay to permit arbitration to proceed is deemed equitable) of a suit at law (ordinarily, and here, a suit for damages) is deemed a preliminary injunction, and is therefore appealable under 28 U.S.C. § 1292(a)(1), an exception to the final judgment rule. Matterhorn, Inc. v. NCR Corp., 763 F.2d 866, 870 (7th Cir. 1985). It might seem that the order to arbitrate would be independently appealable as a mandatory injunction, but at least in this circuit it is not. See id. at 870-71, and cases cited there. It is not appealable as a final order, either, since Olson's lawsuit remains pending in the district court. The stay, however, is appealable, if the Enelow-Ettelson doctrine is still authoritative. We reluctantly conclude that it is.

Section 1292(a)(1), materially unchanged since its original enactment as part of the Evarts Act of 1891, 26 Stat. 828, which created the federal courts of appeals, creates a right of appeal from "interlocutory orders of the district courts . . . granting, continuing, modifying, refusing or dissolving injunctions, or refusing to dissolve or modify injunctions. . . ." The key word is "injunction," and no one just looking at it would suppose that when a judge stays further proceedings in a matter before him he is issuing an injunction. However, the Enelow-Ettelson doctrine deems such a stay an injunction, provided the ground for the stay is equitable and the suit in which it is issued is a suit at law.

Ordinarily a lower court has no authority to reject a doctrine developed by a higher one. See, e.g., Thurston Motor Lines, Inc. v. Jordan K. Rand, Ltd., 460 U.S. 533, 535, 103 S.Ct. 1343, 1344, 75 L.Ed.2d 260 (1983) (per curiam). If, however, events subsequent to the last decision by the higher court approving the doctrine — especially later decisions by that court, or statutory changes — make it almost certain that the higher court would repudiate the doctrine if given a chance to do so, the lower court is not required to adhere to the doctrine. See, e.g., United States v. Burke, 781 F.2d 1234, 1239 n. 2 (7th Cir. 1985); Norris v. United States, 687 F.2d 899, 902-04 (7th Cir. 1982); In re Korman, 449 F.2d 32, 39 (7th Cir. 1971) (per curiam) (dictum), rev'd, 406 U.S. 952, 92 S.Ct. 2055, 32 L.Ed.2d 340 (1972) (per curiam); Spector Motor Service, Inc. v. Walsh, 139 F.2d 809, 814 (2d Cir.), vacated 323 U.S. 101, 65 S.Ct. 152, 89 L.Ed. 101 (1944); Perkins v. Endicott Johnson Corp., 128 F.2d 208, 217-18 (2d Cir. 1942), aff'd, 317 U.S. 501, 63 S.Ct. 339, 87 L.Ed. 424 (1943); Healy v. Edwards, 363 F. Supp. 1110, 1117 (E.D.La. 1973) (three-judge district court), vacated, 421 U.S. 772, 95 S.Ct. 2410, 44 L.Ed.2d 571 (1975) (per curiam); Browder v. Gayle, 142 F. Supp. 707, 716-17 (M.D.Ala.) (three-judge district court), aff'd, 352 U.S. 903, 77 S.Ct. 145, 1 L.Ed.2d 114 (1956) (per curiam); United States v. Girouard, 149 F.2d 760, 765-67 (1st Cir. 1945) (dissenting opinion), rev'd, 328 U.S. 61 (1946). But we take seriously Judge Hand's warning against a lower court's "embrac[ing] the exhilarating opportunity of anticipating a doctrine which may be in the womb of time, but whose birth is distant." Spector Motor Service, Inc. v. Walsh, supra, 139 F.2d at 823 (dissenting opinion).

Experience since the cases establishing the Enelow-Ettelson doctrine were decided, coupled with later decisions by the Supreme Court concerning related issues of appealability, and a statutory change in the jurisdiction of the courts of appeals — all occurring against a background of extraordinarily rapid growth in the workload of those courts — have made clear that the doctrine is arbitrary, mischievous, and devoid of contemporary utility. We must consider therefore whether this is one of those rare cases where circumstances "have created a near certainty that only the occasion is needed for the pronouncement [by the Supreme Court] of the doom" of an obsolete doctrine. Salerno v. American League of Professional Baseball Clubs, 429 F.2d 1003, 1005 (2d Cir. 1970) (Friendly, J.); see also Buzynski v. Oliver, 538 F.2d 6, 7 (1st Cir. 1976).

In 1891 law and equity were still separate jurisprudential systems in the federal courts, even though administered by the same judges; a federal district judge was both a law judge and a chancellor. Historically, a chancellor could enjoin the plaintiff in a suit in a law court from proceeding further with the suit, to allow the chancellor to decide an equitable defense to it. Enelow v. New York Life Ins. Co., 293 U.S. 379, 55 S.Ct. 310, 79 L.Ed. 440 (1935), was a diversity suit at law for the proceeds of a life insurance contract. The insurance company pleaded the affirmative defense that the contract had been procured by fraud, and moved the district judge to try this defense, which the company contended was equitable in character, in advance of jury trial on the legal issues raised by the plaintiff's claim. The judge granted the motion, and the plaintiff appealed. The Supreme Court held that the order granting the motion was appealable because it was equivalent to the grant of an injunction "restraining proceedings at law precisely as if the court had acted upon a bill of complaint in a separate suit for the same purpose." Id. at 383, 55 S.Ct. at 311. Apart from the oddity of describing as an injunction an order regulating the sequence of trial before the judge issuing the order, the Court overlooked the fact that law judges were authorized to stay proceedings before them pending resort to equity. See, e.g., North Chicago Rolling Mill Co. v. St. Louis Ore Steel Co., 152 U.S. 596, 604, 14 S.Ct. 710, 713, 38 L.Ed. 565 (1894); Abraham v. North German Fire Ins. Co., 37 Fed. 731 (C.C.N.D.Ia. 1889); Hurst v. Hurst, 3 U.S. (3 Dall.) 512, 1 L.Ed. 700, 12 Fed.Cas. 1028 (C.C.D.Pa. 1799). There was no need to get an injunction. The order issued by the district judge in Enelow could just as accurately have been described as the stay by a law judge of the plaintiff's suit rather than as a chancellor's injunction against that suit; and a stay as such — a stay that is not also an injunction — is not appealable.

In Shanferoke Coal Supply Corp. v. Westchester Service Corp., 293 U.S. 449, 452, 55 S.Ct. 313, 314, 79 L.Ed. 583 (1935), decided the same day, the Court held that under the rule just laid down in Enelow the stay of an action at law pending arbitration in accordance with the contract between the parties (the kind of stay issued by the district judge in the present case) was appealable as an injunction. The Court offered no reason for this result beyond the authority of Enelow, so Shanferoke adds nothing to the strength of Enelow, the cases might just as well have been consolidated for decision. The facts of Shanferoke suggest, however, the pitfalls of a doctrine that requires characterizing stays as historically legal or historically equitable. At the time the Evarts Act was passed, executory contracts to arbitrate were not specifically enforceable; hence a chancellor would not have stayed a suit at law merely because the plaintiff had agreed to arbitrate his dispute with the defendant. See Red Cross Line v. Atlantic Fruit Co., 264 U.S. 109, 120-21, 44 S.Ct. 274, 275-76, 68 L.Ed. 582 (1924); Hayes v. Allstate Ins. Co., 722 F.2d 1332, 1339 (7th Cir. 1983) (dissenting opinion); Story on Equity § 804 (Gould 10th ed. 1892); Note, Appealability in the Federal Courts, 75 Harv.L.Rev. 351, 373 (1961). If the Court had gotten the history of the enforcement of contracts to arbitrate right it would have realized that Enelow could not be used to decide the issue of appealability in Shanferoke.

With the merger of law and equity by the Federal Rules of Civil Procedure in 1938, the artificiality of describing a stay of an action at law to permit consideration of an equitable defense as an injunction became transparent, as pointed out by Judge Charles Clark, the principal draftsman of the Rules:

[W]e lack any rationale to explain the concept of a judge enjoining himself when he merely decides upon the method he will follow in trying the case. The metamorphosis of a law judge into a hostile chancellor on the other "side" of the court could not have been overclear to the lay litigant under the divided procedure; but if now without even that fictitious sea change the judge in one form of action may split his judicial self at one instant into two mutually antagonistic parts, the litigant surely will think himself in Alice's Wonderland.

Beaunit Mills, Inc. v. Eday Fabric Sales Corp., 124 F.2d 563, 565 (2d Cir. 1942). Nevertheless, in Ettelson v. Metropolitan Life Ins. Co., 317 U.S. 188, 63 S.Ct. 163, 87 L.Ed. 176 (1942), the Supreme Court, in a brief opinion, reaffirmed Enelow. Virtually all the Court said to the point was that the plaintiffs were "in no different position than if a state equity court had restrained them from proceeding in the law action." Id. at 192, 63 S.Ct. at 164. The Court did not cite Beaunit, which had been decided a few months earlier.

Seven years later, in City of Morgantown v. Royal Ins. Co., 337 U.S. 254, 69 S.Ct. 1067, 93 L.Ed. 1347 (1949), the Court held that stays of equity actions, as distinct from stays of actions at law, are not appealable as injunctions. This completed (perhaps unwittingly) the edifice of Enelow-Ettelson; after Ettelson it had still been unclear whether the appealability of a stay depended on the underlying action's being legal. The plaintiff in Morgantown was the insurance company, seeking to reform the insurance contract in order to eliminate coverage. The defendant, the insured, counterclaimed at law for the proceeds of the contract. The judge struck the defendant's demand for a jury trial and set the case for trial before the court. The Supreme Court sensibly described this as "a case of a judge making a ruling as to the manner in which he will try one issue in a civil action pending before himself," rather than a case of a judge's issuing an injunction. Id. at 257, 69 S.Ct. at 1069. But Enelow and Ettelson — merely noting (in language that frequently presages an overruling) that "whatever the present validity of the analogy to commonlaw practice which supported those cases, it is of no help here." Id. The Court also remarked, as it had not done in Enelow or Ettelson, the general policy against piecemeal appeals; emphasized — contrary to Ettelson — the merger of law and equity brought about by the Federal Rules of Civil Procedure; and concluded that it "would ill serve the stated purposes of the Rules of Civil Procedure were we to perpetuate by analogy distinctions which the rules expressly disavow." 337 U.S. at 258, 69 S.Ct. at 1069. Justice Black, dissenting, stated that the Court had overruled Enelow and Ettelson, see id. at 261-63, 69 S.Ct. at 1071-72, because he did not believe either that those decisions had been limited to stays of legal proceedings or that Morgantown would be limited to equitable proceedings.

The Court followed Morgantown in Baltimore Contractors, Inc. v. Bodinger, 348 U.S. 176, 75 S.Ct. 249, 99 L.Ed. 233 (1955), holding nonappealable an order refusing to stay an equity suit pending arbitration. The opinion criticizes the Enelow-Ettelson doctrine as complicated, outmoded, and arbitrary, and suggests that it would be better for stays to be either all appealable or all nonappealable, regardless of whether the underlying suit is legal or equitable. But "the distinction has been applied for years, . . . and we conclude that it is better judicial practice to follow the precedents which limit appealability of interlocutory orders, leaving Congress to make such amendments as it may find proper." Id. at 184-85, 75 S.Ct. at 254. The Court did not, however, go so far as to suggest that Congress in revising the Judicial Code in 1948 had approved the Enelow-Ettelson doctrine — an obscure corner of federal jurisdiction of which the revisers, so far as we can tell, took not the slightest note. On the contrary, "If Congress were aware of the travail this obsolete and fictional distinction is causing the courts of appeals, relief would surely be forthcoming." Wright, The Law of Federal Courts 710 (4th ed. 1983). Almost certainly Congress is not aware; and if it is, it has more pressing matters on its agenda.

Lest anyone think the Enelow-Ettelson doctrine somehow connected with protecting the right to a jury trial, Morgantown and Baltimore Contractors show that it is not. True, the doctrine allows the plaintiff to appeal from a stay that has the effect of postponing his jury trial, as in Enelow and Ettelson; but it does not allow a defendant to appeal from a stay that postpones his jury trial in cases where the defendant has a legal counterclaim but the plaintiff's claim is equitable. Also, for Enelow-Ettelson purposes the courts treat mixed law-equity claims as equitable, see, e.g., Medtronic, Inc. v. Intermedics, Inc., 725 F.2d 440, 443-46 (7th Cir. 1984); Mellon Bank, N.A. v. Pritchard-Keang Nam Corp., 651 F.2d 1244, 1248-50 (8th Cir. 1981), so that a plaintiff who really wanted to use the doctrine to protect his right to appeal from a stay of a jury trial could do so only by giving up any claim he might have for equitable relief (other than incidental relief, see Medtronic, Inc. v. Intermedics, Inc., supra, 725 F.2d at 444-45). Finally, since by the language of section 1292(a)(1) an order denying a stay is appealable when the stay itself would be appealable under this section, a party who is trying to delay a jury trial can do so by asking the judge to stay it on some equitable ground. The denial of his motion will be appealable under Enelow-Ettelson, and often the practical effect of an interlocutory appeal is (as in this case) to stop further proceedings in the district court.

The Supreme Court's last significant mention of the Enelow-Ettelson doctrine came in Gardner v. Westinghouse Broadcasting Co., 437 U.S. 478, 98 S.Ct. 2451, 57 L.Ed.2d 364 (1978), where the issue was whether denial of class certification is appealable under section 1292(a)(1). In holding that it is not, the Court cited Baltimore Contractors for the proposition (inconsistent with Enelow and Ettelson) that "the statute creates an exception from the long-established policy against piecemeal appeals, which this Court is not authorized to enlarge or extend"; made clear that the exception was a "narrow one" geared to proof of irreparable harm; and held that an order denying class certification "did not have any such `irreparable' effect," since "it could be reviewed both prior to and after final judgment; it did not affect the merits of petitioner's own claim; and it did not pass on the legal sufficiency of any claims for injunctive relief." Id. at 480-81, 98 S.Ct. at 2453 (footnotes omitted). Exactly the same things could have been said about the stays held appealable in Enelow and Ettelson. The Court distinguished those decisions in a footnote — but merely by quoting the language of Morgantown that had distinguished without endorsing them. See 437 U.S. at 481 n. 8, 98 S.Ct. at 2453 n. 8. Gardner seems intended to confine appeals under section 1292(a)(1) to orders that are in a substantial sense injunctions, which the orders in Enelow and Ettelson are not.

The Supreme Court has read the statute narrowly in other post- Ettelson cases as well. For example, Switzerland Cheese Ass'n, Inc. v. E. Horne's Market, Inc., 385 U.S. 23, 87 S.Ct. 193, 17 L.Ed.2d 23 (1966), held that an order denying a motion for summary judgment is not appealable as the denial of an injunction, even though the motion sought injunctive relief. Carson v. American Brands, Inc., 450 U.S. 79, 101 S.Ct. 993, 67 L.Ed.2d 59 (1981), while holding that an order refusing to enter a consent decree providing for injunctive relief is indeed the denial of an injunction, seems to require that the appellant show that an immediate appeal is necessary to avert serious and perhaps irreparable consequences. See id. at 87-90, 101 S.Ct. at 998-99; Uehlein v. Jackson Nat'l Life Ins. Co., 794 F.2d 300, 302 (7th Cir. 1986); Donovan v. Robbins, 752 F.2d 1170, 1173-74 (7th Cir. 1985); Sentry Life Ins. Co. v. Borad, 759 F.2d 695, 697 (9th Cir. 1985). No such showing is possible in the usual case where one party asks that proceedings before the court be stayed, whether the court grants or denies the stay.

The cases since Ettelson, beginning with Morgantown, indicate that the Supreme Court is unwilling to interpret section 1292(a)(1) expansively and, in the case of unconventional injunctions — injunctions in other but the ordinary sense — will permit interlocutory appeals only if there is some emergency justifying departure from the final judgment rule, as will rarely be the case when a stay is granted or denied. The pattern created by these cases is inconsistent with the Enelow-Ettelson doctrine and may portend its demise. For a thoughtful exploration of the possible consequences of Carson alone for the continued vitality of the doctrine see Brandon v. Hines, 439 A.2d 496, 504-09 (D.C. 1981).

Even before Switzerland Cheese, Gardner, and Carson, it had been conjectured that the Supreme Court had implicitly rejected the Enelow-Ettelson doctrine in a series of cases that we have not yet mentioned. See Note, Appealability of Stay Orders in the Federal Courts, 47 Minn. L. Rev. 1099 (1963). For example, in Beacon Theatres, Inc. v. Westover, 359 U.S. 500, 79 S.Ct. 948, 3 L.Ed.2d 988 (1959), the Supreme Court had held that when the defendant in a suit in equity counterclaims in law, the defendant is entitled to a jury trial, even though before the merger of law and equity brought about by the Federal Rules of Civil Procedure the equitable claim would have been tried first to the judge, with collateral estoppel effect on any subsequent jury trial of the counterclaim. The premise of the Ettelson decision is that the Federal Rules did not affect the legal versus equitable character of the lawsuit.

Congress has not yet taken up the invitation tendered by Baltimore Contractors to amend section 1292(a)(1) to make stays either appealable or nonappealable regardless of the legal or equitable nature of the proceeding stayed. Congress may never have learned of the invitation, and if it did, still may have had too many other, more urgent problems on its platter to have time to take it up. But three years after Baltimore Contractors Congress did enact 28 U.S.C. § 1292(b), which allows interlocutory appeals from rulings on controlling questions of law if both the district court and the court of appeals agree to the appeal. It has been suggested that this is a sufficient response to the invitation tendered by Baltimore Contractors to warrant the overruling of Enelow and Ettelson. "The availability of permissive interlocutory appeal to secure review in cases of genuine need justifies reconsideration of rules that allow interlocutory appeals as a matter of right in many cases, in order to safeguard rights that may be important to only a few." 16 Wright, Miller, Cooper Gressman, Federal Practice and Procedure § 3923, at pp. 64-65 (1977) (footnote omitted). "The only justification for appeal from such an order [i.e., an order fixing the order of trial or requiring arbitration] is that unnecessary litigation may thereby be eliminated, the purpose for which 1292(b) was expressly designed." Note, Appealability in the Federal Courts, 75 Harv. L. Rev. 351, 380 (1961) (footnote omitted).

More important even than decisional and statutory changes since Ettelson was decided in 1942 is the evidence that has accumulated since then which shows that the doctrine serves no useful purpose, is difficult to apply, and places a burden on the federal courts of appeals that they are ill equipped to bear. This evidence led Professor Wright and his colleagues, almost a decade ago, to state that "experience with the established rule in the courts of appeals should provide sufficient justification for establishing a new rule in an area that is within their own special knowledge and responsibility." 16 Wright, Miller, Cooper Gressman, supra, § 3923, at p. 65.

Forty, even thirty, years ago the Enelow-Ettelson doctrine may have seemed a useful safety hatch in the final judgment rule. On reflection this seems unlikely, since unlike the collateral-order doctrine, Carson, mandamus, and section 1292(b), the Enelow-Ettelson doctrine neither requires a showing that an immediate appeal from a stay within its scope is necessary nor allows appeals from stays of equity suits no matter how urgent a case the appellant makes for an immediate appeal. In any event, experience with the doctrine has confirmed that it fills no gap in the federal appellate system. If the Enelow-Ettelson doctrine were necessary to mitigate the rigors of the final judgment rule, the Morgantown branch of the doctrine — the branch that denies appeal if the underlying suit is equitable rather than legal — would have provoked complaints that it was stifling a class of salutary interlocutory appeals. But, with one exception noted later (Judge Rubin's dissent in the MarLen case), no such complaints have been raised since Justice Black's dissent in Morgantown itself, which preceded the enactment of 1292(b). There is little or no felt need to make stays appealable as a matter of right. This is not surprising, when one considers the number of alternative grounds for seeking appellate review of an interlocutory order that now exist. These include the collateral-order doctrine (see Cohen v. Beneficial Industrial Loan Corp., 337 U.S. 541, 546-47, 69 S.Ct. 1221, 1225-26, 93 L.Ed. 1528 (1949)), section 1292(b), mandamus, the functional interpretation of "order . . . granting . . . [or] refusing . . . [an] injunction" adopted in Carson, and the cases that allow the immediate appeal of a stay if — but only if — it has the practical effect of a final judgment, see, e.g., Moses H. Cone Memorial Hospital v. Mercury Construction Corp., 460 U.S. 1, 103 S.Ct. 927, 74 L.Ed.2d 765 (1983); Mazanec v. North Judson-San Pierre School Corp., 750 F.2d 625 (7th Cir. 1984). All of these grounds, however, differ from section 1292(a)(1) in requiring a showing that immediate review really is necessary. See, e.g., First Nat'l Bank of Waukesha v. Warren, 796 F.2d 999, 1006 (7th Cir. 1986). The Moses H. Cone decision, in holding that an order which has the practical effect of a final judgment is appealable under 28 U.S.C. § 1291, is particularly significant, for it shows that any stay which has an irrevocable effect on the opposing party can be appealed. A stay (or denial of stay) that has no such effect should not be appealable.

The Supreme Court's decisions in the Enelow-Ettelson sequence treat as straightforward the questions whether the underlying suit is legal or equitable and whether the stay is the sort of thing that only a chancellor could (and would) have issued, although, as we have seen, the second question was not straightforward in the three cases in which an appeal was allowed ( Enelow, Shanferoke, and Ettelson). Experience since 1942 has revealed that both questions often are extremely doubtful and difficult — sometimes unanswerable — so that the Enelow-Ettelson doctrine has turned out to be much more difficult and uncertain of application than the Court expected. Among the perplexing questions with which our court has struggled in recent cases are whether the doctrine applies when the underlying suit seeks a mixture of legal and equitable relief (which is common), and if so, whether the equitable relief must be more than merely incidental to make the case a genuine mixed law-equity case; how to characterize remedies, such as an accounting, that can be legal or equitable; whether an order to arbitrate is the equivalent of a stay for arbitration (i.e., whether it too is in the nature of an injunction); whether a grant or denial of a motion to stay an arbitration is appealable if the motion is made in a suit at law; whether a stay to permit an appraisal (as distinct from an arbitration) is equitable in character; whether the denial of a motion to remand a proceeding to a state court is the denial of a stay ( = injunction) for Enelow-Ettelson purposes; and, finally, when is the ground for a stay legal and when is it equitable (since even if the suit sought to be stayed is legal, the doctrine does not apply unless the ground for the stay is equitable). See Medtronic, Inc. v. Intermedics, Inc., supra; Matterhorn, Inc. v. NCR Corp., supra, 763 F.2d at 870-71; Matterhorn, Inc. v. NCR Corp., 727 F.2d 629, 632 (7th Cir. 1984); Shearson Loeb Rhoades, Inc. v. Much, 754 F.2d 773, 778 (7th Cir. 1985); Pierson v. Dean, Witter, Reynolds, Inc., 742 F.2d 334, 336 n. 1 (7th Cir. 1984); Whyte v. THinc Consulting Group Int'l, 659 F.2d 817, 819-20 (7th Cir. 1981); Timberlake v. Oppenheimer Co., 729 F.2d 515, 518-19 (7th Cir. 1984); Hayes v. Allstate Ins. Co., supra, 722 F.2d at 1336-41 (dissenting opinion); Rohrer, Hibler Replogle, Inc. v. Perkins, 728 F.2d 860, 863 (7th Cir. 1984) (per curiam); Texaco, Inc. v. Cottage Hill Operating Co., 709 F.2d 452 (7th Cir. 1983).

Commenting on the problem of the suit that has both legal and equitable claims, the Ninth Circuit observed that " Enelow-Ettelson is virtually impossible to apply to a complaint like Danford's in which the averments and prayers are a puree of legal and equitable theories and of claims that had no antecedents in the old bifurcated system." Danford v. Schwabacher, 488 F.2d 454, 456 (9th Cir. 1973). Other difficulties in application have surfaced, but it would be tedious to recount them. For examples see Pepper v. Miani, 734 F.2d 1420 (10th Cir. 1984), and 16 Wright, Miller cooper, Federal Practice and Procedure § 3923 (1986 Supp.), and for an excellent overview see Federal Civil Appellate Jurisdiction: An Interlocutory Restatement, 47 Law Contemp. Prob., spring 1984, at 13, 126-34.

When Ettelson was decided in 1942, the annual number of cases appealed to the federal courts of appeals was only 3,228. In 1955, when Bodinger was decided, it was only a little higher — 3,695. Today it is almost ten times as great (34,292 in the year ending June 30, 1986), and growing (2.8 percent last year, 30 percent in the last five years). Although the number of court of appeals judges has increased since 1942, the increase has been much less than in the number of cases (less than three times compared to more than ten), resulting in severe pressure on the ability of these courts to provide conscientious consideration to all appeals. Arbitrary exceptions to the final judgment rule that were tolerable nuisances 44 years ago are intolerable today.

The tone of exasperation in which federal court of appeals judges complain about the Enelow-Ettelson doctrine should thus come as no surprise. See quotations collected in 16 Wright, Miller Cooper, Federal Practice and Procedure § 3923, at pp. 28-29 (1986 Supp.). It has been called "a remnant from the jurisprudential attic," Danford v. Schwabacher, supra, 488 F.2d at 455, and "an anachronism wrapped up in an atavism," Hartford Financial Systems, Inc. v. Florida Software Services, Inc., 712 F.2d 724, 727 (1st Cir. 1983). Every circuit (except the Federal Circuit, which is new and apparently has not yet had an Enelow-Ettelson appeal) is on record with criticism, often harsh, of the doctrine. See, e.g., Langley v. Colonial Leasing Co., 707 F.2d 1 (1st Cir. 1983); Standard Chlorine of Delaware, Inc. v. Leonard, 384 F.2d 304 (2d Cir. 1967); Nascone v. Spudnuts, Inc., 735 F.2d 763, 767-70 (3d Cir. 1984); Chapman v. International Ladies' Garment Workers' Union, 401 F.2d 626 (4th Cir. 1968); Wallace v. Norman Industries, Inc., 467 F.2d 824, 827 (5th Cir. 1972); Mansbach v. Prescott, Ball Turben, 598 F.2d 1017, 1022-23 (6th Cir. 1979); Matterhorn, Inc. v. NCR Corp., supra, 763 F.2d at 870-71; Mellon Bank, N.A. v. Pritchard-Keang Nam Corp., supra; Mediterranean Enterprises, Inc. v. Ssangyong Corp., 708 F.2d 1458, 1462 n. 3 (9th Cir. 1983); Pepper v. Miani, supra; Miller v. Drexel Burnham Lambert, Inc., 791 F.2d 850, 853 n. 3 (11th Cir. 1986) (per curiam); Lee v. Ply[*] Gem Industries, Inc., 593 F.2d 1266, 1268-69 and n. 18 (D.C. Cir. 1979). Judge Rubin has urged his court to reject the doctrine. See Mar-Len of Louisiana, Inc. v. Parsons-Gilbane, 732 F.2d 444, 445-47 (5th Cir. 1984) (dissenting opinion). Although a majority of the panel declined the invitation, they agreed that the doctrine was "artificial," "medieval," and "outmoded." Id. at 445 n. 1. One critic is so exercised that she has urged the courts of appeals to "practice a jurisprudential civil disobedience." Mathy, The Appealability of District Court Orders Staying Court Proceedings Pending Arbitration, 63 Marquette L.Rev. 31, 69 (1979).

Professor Moore and his collaborators have said, "It is to be hoped that the Supreme Court will accept the first opportunity offered to decide that the reason for the Enelow-Ettelson rule [i.e., the historical division between law and equity courts] having ceased, the rule is no more." 9 Moore's Federal Practice ¶ 110.20[3], at p. 245 (2d ed. 1985). Professor Wright and his collaborators have gone further, arguing that the extensive experience that the courts of appeals have had with the administration of the Enelow-Ettelson doctrine, coupled with the fact that the doctrine deals with a matter within the special knowledge and responsibility of those courts (namely, their jurisdiction), would justify them in rejecting it. See 16 Wright, Miller, Cooper Gressman, supra, § 3923, at p. 65.

The case against the doctrine seems to us conclusive. The Enelow, Shanferoke, and Ettelson decisions — the Supreme Court's only decisions actually allowing an appeal under the doctrine — rested not on the implications of a legislative text or evidence of legislative purpose, let alone on practical considerations, but on the historical mistake that a stay of proceedings at law was necessarily equitable in character; for in fact a law judge could stay a proceeding before him to enable a party to pursue his remedies in a court of equity. As early as 1949, in Morgantown, the Court was expressing profound skepticism about the correctness of Enelow and Ettelson; and the distinction it adopted in that decision, between stays of legal and equitable proceedings, seemed so threadbare as to move Justice Black to describe Enelow and Ettelson as having been overruled. What apparently was not foreseen in 1949, or six years later when Baltimore Contractors was decided, was that the attempt in Morgantown to limit Enelow and Ettelson to equitable stays of legal proceedings would produce a vast and sterile jurisprudence. Nor was the tremendous increase in the appellate workload of the federal courts foreseen; nor the enactment of section 1292(b). Finally, no justified reliance would be impaired by declaring the doctrine defunct — no reliance, period. No one plans his behavior, in court or out, in reliance on being able to take an interlocutory appeal under the Enelow-Ettelson doctrine.

Despite all this we think it would be improper for this court to reject the doctrine. The following considerations have persuaded us that discretion is the better part of valor in this instance:

1. The standard for declaring a decision or doctrine of a higher court defunct is, as suggested earlier, whether the lower court is certain or almost certain that the decision or doctrine would be rejected by the higher court if a case presenting the issue came before it. This is a high standard and will rarely be met. Any looser standard, however, would unsettle the law. The expansion of the federal courts makes it more rather than less important that lower courts respect the authority of higher ones in order to maintain a minimum of coherence in the law.

2. Although Morgantown gravely undermined the Enelow-Ettelson doctrine, Baltimore Contractors expressly if reluctantly reaffirmed it by saying that correction was for Congress. We do not think the enactment of section 1292(b) can be taken as a realistic sign that Congress wanted to overrule the doctrine, to which no reference is made in the legislative history. See S.Rep. No. 2434, 85th Cong., 2d Sess. (1958). Other decisions by the Supreme Court that undermine the doctrine, such as Carson and Beacon Theaters, do so only in the sense of adopting a view of other issues that, if carried to the limits of its logic, would be inconsistent with the Enelow-Ettelson doctrine. But the law tolerates a good deal of logical inconsistency. The fact that the Supreme Court would decide a similar question differently from Enelow-Ettelson is not a sure sign that it will discard the doctrine in order to produce logical order.

3. The Enelow-Ettelson doctrine has been reaffirmed in at least a hundred, and perhaps many more, court of appeals cases since the Supreme Court last considered it in Baltimore Contractors. If we declared it defunct in this circuit we would create a conflict among circuits and place great pressure on the Supreme Court to grant certiorari in order to clarify the status of the doctrine, since litigants in other circuits would wonder whether those circuits would reconsider their position in light of our decision and since a situation in which some courts deem a Supreme Court doctrine binding and others do not is an unseemly one that the Court is not likely to think tolerable. Thus we would be interfering with the Supreme Court's control of its discretionary docket.

4. The Enelow-Ettelson doctrine is an interpretation of a statute. Although it seems to us artificial in the extreme to suppose that the failure of Congress to amend section 1292(a)(1) to overrule this obscure doctrine constitutes legislative acquiescence in it, the Supreme Court has on occasion been hospitable to such arguments (see, e.g., Bob Jones University v. United States, 461 U.S. 574, 599-602, 103 S.Ct. 2017, 2032-34, 76 L.Ed.2d 157 (1983)) and might be in this instance. The argument from acquiescence is reinforced, a little bit anyway, by the 1948 revision of the Judicial Code, in which Congress reenacted section 1292(a)(1) without overruling Enelow and Ettelson. Again we think an inference of adoption from failure to overrule is artificial, both generally and in this instance. But the Supreme Court might disagree. See, e.g., Douglas v. Seacoast Products, Inc., 431 U.S. 265, 278-79, 97 S.Ct. 1740, 1748, 52 L.Ed.2d 304 (1977); Snyder v. Harris, 394 U.S. 332, 338-39, 89 S.Ct. 1053, 1057-58, 22 L.Ed.2d 319 (1969); cf. Phillips, Nizer, Benjamin, Krim Ballon v. Rosenstiel, 490 F.2d 509, 514 (2d Cir. 1973) (Friendly J.). We do not think it realistic "to assume that our elected representatives, like other citizens, know the law," Cannon v. University of Chicago, 441 U.S. 677, 696-97, 99 S.Ct. 1946, 1957-58, 60 L.Ed.2d 560 (1979), when "law" is defined to mean every judicial gloss, however recherche, on a technical statute. Hence we cannot think it realistic to assume that Congress knew about the Enelow-Ettelson doctrine and would have repealed it, either in the 1948 revision of the Judicial Code or later, if it hadn't liked it; moreover, even if it didn't like it, it might not have had the time, given the press of its other duties, to draft and enact a suitable repealer. But the Supreme Court's view is apparently otherwise. See, e.g., Merrill Lynch, Pierce, Fenner Smith, Inc., v. Curran, 456 U.S. 353, 379, 102 S.Ct. 1825, 1839, 72 L.Ed.2d 182 (1982). Still another point is that the Supreme Court is generally reluctant to overrule a statutory as distinct from a constitutional decision, see, e.g., Square D Co. v. Niagara Frontier Tariff Bureau, Inc., ___ U.S. ___, 106 S.Ct. 1922, 1930 and n. 34, 90 L.Ed.2d 413 (1986), although its practice in this regard has not been consistent. See, e.g., Monell v. Department of Social Services, 436 U.S. 658, 695-701, 98 S.Ct. 2018, 2038-41, 56 L.Ed.2d 611 (1978); Continental T.V., Inc. v. GTE Sylvania Inc., 433 U.S. 36, 47-49, 97 S.Ct. 2549, 2556-57, 53 L.Ed.2d 568 (1977); Boys Markets, Inc. v. Retail Clerks Union, Local 770, 398 U.S. 235, 240-42, 90 S.Ct. 1583, 1586-87, 26 L.Ed.2d 199 (1970).

5. Last and perhaps most important, merely to reject the Enelow-Ettelson doctrine would leave undecided the question whether all stays should be appealable as injunctions or no stays. Although obviously our own inclination would be toward the latter position, the question whether all stays or none should be appealable once the artificial distinction made by the doctrine is discarded is more difficult than whether the doctrine itself should be discarded, and we are not sure how the Supreme Court would answer it. Judge Rubin, in his dissent in Mar-Len advocating repeal of the doctrine, wanted to allow the appeal of a stay of an equity suit. See 732 F.2d at 445-47.

So though we hope the Supreme Court will someday repudiate the doctrine and hold (with exceptions unnecessary to dwell on here) that no stays are appealable under section 1292(a)(1), we do not think it would be proper for us to anticipate such a repudiation; we lack the very high degree of confidence necessary to justify disregarding decisions of a higher court.

We shall therefore proceed to decide the merits of the present appeal — an appeal properly within our jurisdiction if the Enelow-Ettelson doctrine remains good law.

The arbitration clause in Olson's contract with Paine Webber, as drafted, fails to comply with two regulations issued by the Commodity Futures Trading Commission in 1983. One requires that a brokerage agreement, if it contains an arbitration clause, advise the customer that he "will have the opportunity to elect a qualified forum for conducting" the arbitration. 17 C.F.R. § 180.3(b)(4). The other requires that the agreement (again only if it contains an arbitration clause) advise the customer that the broker will pay certain costs of arbitration that might otherwise be split between the parties. 17 C.F.R. § 180.3(b)(5). Neither required statement was in the agreement that Paine Webber prepared and Olson signed. Paine Webber has, however, conceded that the regulations are applicable to the agreement and has bound itself to comply with them in the arbitration.

This concession makes it impossible to understand how Olson has been hurt by Paine Webber's failure to advise her in the original agreement that she had these two rights (to choose the forum and to shift some of the costs of the arbitration to Paine Webber). The arbitration will be conducted exactly as it would have been had the advice been given as required. Olson does not argue that if she had been advised of these two rights she would not have agreed to the inclusion of an arbitration clause, for by conferring procedural advantages on customers the regulations in question make arbitration more rather than less attractive to them. She does argue that if Paine Webber is allowed to get away with what it has done here, a case may someday arise where a customer, not having been advised of these additional rights in arbitration, goes through an arbitration with Paine Webber without invoking them. But we shall decide that case when and if it arises. It is not this case; Olson has not been cheated out of any of her procedural rights in arbitration.

It is not yet the law that if a contractual provision is illegal (and as drafted the arbitration clause in this case was illegal — violated CFTC regulations whose validity is not challenged) the contract cannot be enforced, regardless of the nature and consequences of the illegality. As explained in Northern Indiana Public Service Co. v. Carbon County Coal Co., 799 F.2d 265, 273-74 (7th Cir. 1986), the contract defense of illegality is flexible rather than rigid, and will not be applied where it would produce a sanction disproportionate to the wrong, as it might well do here. Paine Webber's violations of CFTC regulations did not hurt Olson or help Paine Webber. On the contrary, the violations, by omitting to advise her of some of her rights in arbitration, made it less rather than more likely that she would agree — as Paine Webber very much wanted her to agree — to submit disputes under the brokerage agreement to arbitration. It seems that Paine Webber is likely to hurt only itself by such violations; and violations that hurt the violator even if he isn't punished are normally self-correcting, and certainly don't require extraordinary punishment.

Against this it can be argued that there is a danger — how great we do not know — that Paine Webber is obtaining an advantage from customers who (unlike Olson), not knowing that they have certain procedural rights by virtue of CFTC regulations, submit to arbitration without asserting those rights. Of course any customer advised by a competent lawyer would not be fooled; but not all customers are represented in arbitration proceedings, and of those who are, some unknown fraction are not competently represented. We are puzzled by the statement of Paine Webber's counsel at argument that Paine Webber has not yet reformed its standard brokerage agreement to comply with the 1983 regulations, though it anticipates being forced to do by the CFTC. Why wait?

Another argument for Olson is that when the Commission first adopted section 180.3 in 1976, it provided that all nonconforming agreements "will be null and void, including those heretofore signed by customers." 41 Fed.Reg. 42944 (1976). The courts have held that this language voids noncomplying agreements signed before the regulations were adopted. See, e.g., Ames v. Merrill Lynch, Pierce, Fenner Smith, 567 F.2d 1174, 1178-79 (2d Cir. 1977). The regulations at issue in this case, added in 1983 by amendment to section 180.3, presumably have the same effect. See Wotkyns v. D.E. Jones Commodities, Inc., 791 F.2d 749 (9th Cir. 1986), dealing with another of the 1983 amendments. We may take the "null and void" provision as a strong statement that non-complying agreements are unenforceable — provided there is some possibility that the violation harmed the plaintiff. In Ames, for example, the broker had violated the regulations by not telling the customer certain consequences of arbitration, such as that he would be giving up his right to sue, that might have dissuaded the customer from signing the arbitration clause. But in this case full compliance with the regulations would have made the customer more likely to sign.

That, we may assume, would make no difference if Paine Webber were trying to hold Olson to the letter of her agreement and thus deny her the procedural advantages in arbitration which the 1983 regulations give her. But it is not trying to do this. The agreement having been reformed, the only illegality is Paine Webber's failure to have advised Olson in advance of this dispute of the rights newly granted customers by the 1983 regulations. Since this illegality could not have harmed her — she does not, as we noted earlier, even argue that disclosure of her additional rights would have dissuaded her from signing the arbitration clause to which those rights pertained — we do not think it should provide a defense to Paine Webber's motion to enforce an arbitration clause to which she freely consented. We have no evidence that the Commission wanted to nullify agreements that, as enforced, would comply fully with the Commission's regulations.

In so concluding, we are not condoning Paine Webber's violations, for the CFTC remains free to impose whatever sanctions for them the law provides. Those sanctions are, in fact, extensive. See 7 U.S.C. § 9, 13a-1, 13b. But the arbitration clause having been reformed, Paine Webber is not trying to enforce any illegal provision in it against Olson. Since there is no possible way that she has been hurt by the illegality, it does not provide a ground on which she can avoid the clause.

AFFIRMED.