Appellant S. Robert Davis was convicted of violating the federal mail fraud statute, 18 U.S.C. § 1341. This court overturned the conviction on the ground that the indictment, which was based on the “intangible rights” theory disavowed by the Supreme Court in McNally v. United States, 483 U.S. 350, 107 S.Ct. 2875, 97 L.Ed.2d 292 (1987), was legally deficient. A superseding indictment was then returned, and Mr. Davis filed two motions to dismiss — one alleging a violation of the Double Jeopardy Clause and the other based on the statute of limitations. Both motions were denied by the district court, and Mr. Davis has taken an interlocutory appeal.

We conclude that we lack jurisdiction to entertain a pretrial appeal from the district court’s ruling on the statute of limitations question. With respect to the double jeopardy issue — as to which the government[*902] concedes we have jurisdiction under the collateral order exception to the final judgment rule, Abney v. United States, 431 U.S. 651, 97 S.Ct. 2034, 52 L.Ed.2d 651 (1977)-we shall affirm the order of the district court.

I

Appellant Davis owned a house and a tract of undeveloped land in the Squirrel Bend section of Upper Arlington, Ohio. He planned to subdivide his vacant land, and the city planned to construct a public waterline in the area at an estimated cost of $80,000. Mr. Davis agreed to provide water and sewer improvements to people who purchased lots from him, and he was expected to pay $60,000 of the city’s cost.

After discussions with Harold Hyrne, the city manager of Upper Arlington, Mr. Davis undertook to construct the Squirrel Bend waterline at his own expense. He had the line installed by Wright Brothers Excavating Company at a contract price of $36,296.54.

Under § 903.12 of the Upper Arlington City Code, abutting property owners may “tap in” to a privately built waterline without the builder’s consent if they pay the builder a charge set by the city manager on the basis of the builder’s certification of “the entire cost of such improvement to the builder.” One of Mr. Davis’s neighbors, Richard Schultz, paid Davis a tap-in fee set by City Manager Hyrne on the basis of a $71,472 cost figure allegedly certified by Davis. The United States asserts that the $71,472 figure (which Mr. Davis denies having “certified”) was substantially above the actual cost of the improvement to Davis.

A federal grand jury indicted Mr. Davis for mail fraud on June 12, 1986. The theory of the indictment was that Davis had devised “a scheme or artifice to defraud the citizens of the City of Upper Arlington Ohio, of their rights to honest and faithful services from their public officials,” in furtherance of which scheme he had used or caused the use of the United States mails on four occasions between June 18, 1981, and August 13, 1981. It was alleged that City Manager Hyrne — the public official of whose honest and faithful services the townspeople supposedly were deprived— had (1) agreed not to question any waterline construction cost certified by Davis as long as the amount certified was approximately the same as the cost estimated by the city, (2) approved a tap-in charge based on an inflated cost certification, and (3) failed to collect city inspection fees for the waterline. In return, Davis was alleged to have arranged for Hyrne to receive a discount on a purchase of certain common stock.

A pretrial motion to dismiss for failure to allege a legally cognizable scheme to defraud was denied, and the case went to trial before a jury. After seven days of trial and three days of deliberation, the jury returned a verdict of guilty on all four counts. Mr. Davis appealed. While his appeal was pending, the United States Supreme Court decided McNally v. United States, supra, 483 U.S. 350, 107 S.Ct. 2875, holding that the federal mail fraud statute is limited in scope to the protection of tangible property rights and does not make it a crime to deprive citizens of their intangible right to honest performance of duty by public officials. (Congress has now amended the mail fraud statute to incorporate the intangible rights theory, Anti-Drug Abuse Act of 1988, Pub.L. No. 100-690, § 7603, 102 Stat. 4181, 4508 (to be codified at 18 U.S.C. § 1346), but the amendment has no retroactive application, of course.) In an unpublished per curiam opinion filed on March 8, 1988, we reversed Mr. Davis's conviction on the ground that the indictment was defective under McNally:

“The indictment in the case at bar does not properly allege a violation of the mail fraud statute. In its initial paragraph the indictment defines the scheme to defraud as one directed at the rights [of the citizens of Upper Arlington] to honest and faithful services from their public officials_ Although in later paragraphs the indictment avers facts that may have involved a deprivation of property or money, nowhere is there an averment that the purpose or result of the[*903] scheme was to obtain property or money for the defendant.... We shall not strain to construe this defective indictment as implicitly charging that the purpose of the scheme was to deprive someone of money, as opposed to depriving the citizens of Upper Arlington of their ‘rights to honest and faithful services from their public officials.

United States v. Davis, No. 86-4076, slip op. at 3 (6th Cir.1988) [841 F.2d 1127 (Table)].

On remand, the district court dismissed the indictment. Ten days later a “superseding indictment” was returned against Davis, charging him with four counts of mail fraud based on the same jurisdictional mailings alleged in the original indictment.

The original indictment and its successor are much the same in wording, except for the introductory paragraph of Count I. Instead of alleging that Davis had devised a scheme to defraud his fellow citizens of their right to good government, the new introductory paragraph (which is incorporated by reference in Counts 2, 3, and 4) alleges that Davis devised “a scheme or artifice to defraud, and to obtain money by means of false and fraudulent pretenses and representations.... ” Like the original indictment, the superseding indictment charges that Davis prevailed upon Hyrne to accept a falsely inflated cost figure for the construction of the waterline, to approve an inflated tap-in charge based on that figure, and to refrain from collecting inspection fees, all in return for Davis’s arranging for Hyrne to buy stock at a discount. Minor changes tracking the language of the new introductory paragraph specify that the tap-in charges were in the form of “money,” and that the inspection fees Hyrne refrained from collecting were likewise monetary in nature.

After entering a plea of not guilty, Davis moved to dismiss the superseding indictment for violation of the Double Jeopardy Clause and for failure to bring the indictment within the five-year period prescribed by the applicable statute of limitations. The district court (Joseph P. Kinneary, J.) denied both motions. This interlocutory appeal followed.

II

The Double Jeopardy Clause of the Fifth Amendment provides that no person shall “be subject for the same offence to be twice put in jeopardy of life or limb.” U.S. Const. amend. V. In the leading case of United States v. Ball, 163 U.S. 662, 16 S.Ct. 1192, 41 L.Ed. 300 (1896), the Supreme Court held that a defendant who succeeded in having his murder conviction set aside because of a legal defect in the indictment was not “twice put in jeopardy,” in violation of the Constitution, when retried on a new and legally sufficient indictment. The Court provided no very extensive rationale for this holding, possibly considering it self-evident that the Double Jeopardy Clause was not designed to guarantee the defendant “the right of being hung, to protect him from the danger of a second trial.” United States v. Keen, 26 F.Cas. 686, 690 (C.C.D.Ind.1839) (No. 15,-510), as quoted in United States v. Scott, 437 U.S. 82, 92, 98 S.Ct. 2187, 2194, 57 L.Ed.2d 65 (1978). The modem interpretation of Ball is that “it effectively formulated a concept of continuing jeopardy that has application where criminal proceedings against an accused have not run their full course.” Price v. Georgia, 398 U.S. 323, 326, 90 S.Ct. 1757, 1759, 26 L.Ed.2d 300 (1970). Cf. Justices of Boston Municipal Court v. Lydon, 466 U.S. 294, 308, 104 S.Ct. 1805, 1813, 80 L.Ed.2d 311 (1984).

The second Mr. Justice Harlan offered these observations on the “Ball principle,” as he called it:

“While different theories have been advanced to support the permissibility of retrial, of greater importance than the conceptual abstractions employed to explain the Ball principle are the implications of that principle for the sound administration of justice. Corresponding to the right of an accused to be given a fair trial is the societal interest in punishing one whose guilt is clear after he has obtained such a trial. It would be a high price indeed for society to pay were ev[*904] ery accused granted immunity from punishment because of any defect sufficient to constitute reversible error in the proceedings leading to conviction. From the standpoint of a defendant, it is at least doubtful that appellate courts would be as zealous as they now are in protecting against the effects of impropreties at the trial or pretrial stage if they knew that reversal of a conviction would put the accused irrevocably beyond the reach of further prosecution. In reality, therefore, the practice of retrial serves defendants’ rights as well as society’s interest.”

United States v. Tateo, 377 U.S. 463, 466, 84 S.Ct. 1587, 1589, 12 L.Ed.2d 448 (1964).

The Court has endorsed Justice Harlan’s words in a number of subsequent opinions, the most recent of which is Lockhart v. Nelson, 488 U.S. -, 109 S.Ct. 285, 102 L.Ed.2d 265 (1988). See, 488 U.S. at -, 109 S.Ct. at 291, 102 L.Ed.2d at 272, where, before quoting extensively from Tateo, the Court cited both Ball and Tateo in support of the proposition that “[i]t has long been settled ... that the Double Jeopardy Clause’s general prohibition against successive prosecutions does not prevent the government from retrying a defendant who succeeds in getting his first conviction set aside ... because of some error in the proceedings leading to conviction.”

The error committed by the trial court in Ball — failure to dismiss a faulty indictment — was characterized in Burks v. United States, 437 U.S. 1, 14, 98 S.Ct. 2141, 2148, 57 L.Ed.2d 1 (1978), as “trial error,” the rectification of which has never been thought to be precluded by the Double Jeopardy Clause: “We have no doubt that Ball was correct in allowing a new trial to rectify trial error_” Id. (emphasis by the Court). In Scott, supra, 437 U.S. at 89, 98 S.Ct. at 2193, a case decided the same day as Burks, the Court observed that “Ball firmly established that a successful appeal of a conviction precludes a subsequent plea of double jeopardy ...”, and in Montana v. Hall, 481 U.S. 400, 404, 107 S.Ct. 1825, 1827, 95 L.Ed.2d 354 (1987), the Court cited Ball in declaring it to be “clear that the Constitution permits retrial after a conviction is reversed because of a defect in the charging instrument.”

Generalizing from Ball and subsequent decisions, the Scott Court went so far as to declare that “[t]he successful appeal of a judgment of conviction, on any ground other than the insufficiency of the evidence to support the verdict ... poses no bar to further prosecution on the same charge.” 437 U.S. at 90-91, 98 S.Ct. at 2193-94 (emphasis supplied). We did not reach the sufficiency of the evidence question in Mr. Davis’s earlier appeal, and given what the Supreme Court said in Scott, it may not be readily apparent how we could properly hold that Mr. Davis’s successful appeal on the very ground that was involved in Ball — a legal defect in the indictment — poses a bar to further prosecution. If, as Burks said, it was “trial error” not to dismiss the defective indictment in Ball, it would seem to have been “trial error” for the district court not to dismiss the initial indictment of Mr. Davis — and a reversal procured on the basis of “trial error,” to repeat, does not bar a new tria).

In Saylor v. Cornelius, 845 F.2d 1401, 1405 (6th Cir.1988), we recognized that “the only time a reversal of a conviction is the equivalent of an acquittal as far as the applicability of the Double Jeopardy Clause’s bar against retrial is concerned, is where the reversal is based on insufficiency of the evidence.” We nonetheless held, in Saylor, that where the defendant had been tried on an indictment that embodied two distinct legal theories (guilt by reason of acts performed as an accomplice and guilt by reason of acts performed as a conspirator), one of which theories (conspiracy) proved to be unsupported by the evidence and the other of which was inexplicably ignored in the instructions to the jury, the Double Jeopardy Clause barred a retrial on the uninstructed theory after the judgment entered against the defendant on the other theory had been reversed. The evidence introduced in support of the accomplice theory — the theory omitted in the trial court’s charge to the jury — may have been legally sufficient to support a convic[*905] tion. Because we were prepared to apply the Double Jeopardy Clause in Saylor notwithstanding the sufficiency of the evidence supporting the theory on which a retrial was sought, Mr. Davis argues (without conceding the sufficiency of the evidence against him in this case) that we ought to do the same thing here.

If Saylor was decided correctly (and Judge Kinneary doubted that it was), Mr. Davis’s argument has a certain superficial appeal despite the obvious distinction arising from the fact that our reversal of the Davis conviction turned on the inadequacy of the original indictment, while the reversal of the Saylor conviction had nothing to do with any defect in the indictment. We were concerned in Saylor about setting a precedent that would allow a prosecutor to “indict on several counts or theories, present evidence on each of them, and then go to the jury only on selected ones, in effect holding the others in reserve for a subsequent or improved effort” if the jury should fail to convict on the theory or theories actually submitted to it. 845 F.2d at 1408. Perhaps we ought to be equally concerned about setting a precedent that would allow a prosecutor to obtain an indictment on one theory (defrauding the electorate of an intangible right to honest government, e.g.) and let the case go to a jury on that theory, while holding in reserve a second theory (defrauding an identifiable individual of money or property) in order to get a subsequent bite at the apple if the jury failed to convict the first time.

In the factual situation presented here, such a concern would be unwarranted. If the jury had found Mr. Davis innocent of the offense with which he was charged in the original indictment, we do not believe for a moment that he could have been retried for the same offense on the theory embodied in the superseding indictment. The theories are different, to be sure, but there is a necessary logical nexus between them, and the conduct that would have to be proved in order to obtain a conviction would be the same under both indictments. That was not the situation in Saylor; in that case we thought there was no necessary logical nexus between the two theories,” and it seems clear from our opinion that if Mr. Saylor had been indicted and tried on one theory only, he might, as we saw it, have been subject to a later trial on the second theory regardless of the outcome of the first trial. 845 F.2d at 1408.

Judge Kinneary emphasized another distinction between this case and Saylor: the Saylor prosecutor was asleep at the switch (or so we assumed) when he failed to request that the jury be charged on the conspiracy theory, but no comparable fault could be attributed to the Davis prosecutor in deciding to base the indictment of Mr. Davis on an “intangible rights” theory alone. That decision was perfectly legitimate when made, the intangible rights theory having been endorsed by this court only weeks before in the very case that was ultimately to produce the McNally decision. United States v. Gray, 790 F.2d 1290 (6th Cir.1986), rev’d sub nom. McNally v. United States, 483 U.S. 350, 107 S.Ct. 2875, 97 L.Ed.2d 292 (1987). This court has been wrong before, of course, but the prosecutor is not to be faulted for assuming we were right.

The prosecutor gained no unfair advantage by limiting the indictment of Mr. Davis to an intangible rights theory. Had the prosecutor been given the prescience to realize that Gray would be reversed in McNally, the indictment of Mr. Davis would unquestionably have been drawn differently. As noted above, moreover, if the jury had acquitted Mr. Davis in the trial on the original indictment, there could have been no new trial even if the Supreme Court had affirmed Gray in McNally instead of reversing it. The potential for prosecutorial abuse that led us to find a termination of jeopardy in Saylor simply had no counterpart here.

The defect in the charging instrument at issue in United States v. Ball, supra, 163 U.S. 662, 16 S.Ct. 1192, 41 L.Ed. 300 (a failure to specify the time and place of a murder victim’s death), like the defect in the charging instrument in Montana v. Hall, supra (“the State simply relied on the wrong statute,” 481 U.S. at 404, 107[*906] S.Ct. at 1827), obviously reflected more poorly on the prosecutor than did the defect (as it proved to be) in the instrument with which Mr. Davis was charged. If, as Saylor seems to suggest, prosecutorial culpability may have some relevance in determining when jeopardy has been terminated, it would be more than a little anomalous to conclude that although a retrial was not barred in Ball or in Hall, Saylor requires us to block a retrial of Mr. Davis.

Mr. Davis argues that cases such as Ball and Hall involve the “acquittal equivalent” branch of double jeopardy law, under which the only question is whether the first trial ended in the functional equivalent of an acquittal. There is a second branch, he argues, that protects against reprosecution even where the first trial did not end in an acquittal or the equivalent of an acquittal. Thus in Oregon v. Kennedy, 456 U.S. 667, 102 S.Ct. 2083, 72 L.Ed.2d 416 (1982), where the prosecutor allegedly intended to cause a mistrial, jeopardy would have been terminated, and a retrial would have been barred by the Double Jeopardy Clause, had it been found as a fact that the prosecutor was guilty of misconduct that was intended to provoke the defendant into moving, successfully, for a mistrial. Oregon is a specimen of what we might call the “mistrial” branch of double jeopardy law, cf. United States v. Perez, 22 U.S. (9 Wheat.) 579, 580, 6 L.Ed. 165 (1824); but no mistrial was declared at Mr. Davis’s trial, and Oregon and its relatives do not apply here. Other cases cited by Mr. Davis are similarly inap-posite: United States v. Wilson, 420 U.S. 332, 95 S.Ct. 1013, 43 L.Ed.2d 232 (1975), for example, held that a government appeal from a ruling in favor of the defendant after a guilty verdict was not barred by the Double Jeopardy Clause.

It is true that our decision in Saylor did not involve an “acquittal equivalent,” but the significance of Saylor is limited by the unusual situation we were addressing in that case: because of prosecutorial absentmindedness, Mr. Saylor’s trial ended without an acquittal or a conviction on a charge that had been properly presented in an indictment and emphasized at trial. In the present situation, by contrast, the “money or property” theory of mail fraud was omitted from the indictment, and the theory was expressly disavowed at trial. The prosecutor was unequivocal on this, telling the jury

“This is is not a money case. When we talk about fraud, we are not talking about taking too much money from somebody or getting a little bit more money than you otherwise might have, we are talking about a fraud on a governmental process.”

In Saylor we read United States v. Scott, 437 U.S. 82, 98 S.Ct. 2187, 57 L.Ed.2d 65, as saying that trials which end inconclusively may sometimes be placed at one or another of two “poles” — one where trial proceedings are terminated inconclusively as a result of the defendant’s election, and the other where the proceedings end inconclusively as a result of the prosecutor’s action. At the first pole, reprosecution is allowed; at the second, it is barred by the Double Jeopardy Clause. The inconclusive outcome of the proceedings brought against Mr. Saylor on the accomplice theory fell between the two poles, we thought, but closer to the second than the first. But these poles have no relevance here, as we see it; proceedings against Mr. Davis on a “money or property” theory cannot fairly be said to have “terminated” because of the action or inaction of either party, no such proceedings having been initiated in the first place. This case does not fall between the metaphorical “poles” of Scott/Saylor; it is off the map entirely.

Our conclusion that the case at bar is not governed by Saylor is strengthened, finally, by what we said in United States v. Stack, 853 F.2d 436 (6th Cir.1988)-a decision handed down three months after Say-lor. Like the case at bar, Stack involved the appeal of a mail fraud conviction obtained on an indictment that proved to have been defective because it included an intangible rights theory. In reversing the conviction in Stack we noted, citing Montana v. Hall, supra, 481 U.S. 400, 107 S.Ct. 1825, that “since the conviction is not reversed due to insufficient evidence the government is not precluded from indicting[*907] defendant on a permissible theory of mail fraud.” 853 F.2d at 438. By the same token, we now hold, the government was not precluded from indicting defendant Davis on a permissible theory of mail fraud.

Ill

Purporting to forego, in his prior appeal, any claim of what he there characterized as “trial error,” ^[1] Mr. Davis advanced as his first argument in the earlier appeal a contention that “the evidence at trial was insufficient to prove beyond a reasonable doubt the existence of the alleged scheme to defraud.” We never reached that issue, and Mr. Davis now contends that our failure to make a specific finding on the sufficiency of the evidence ought to bar any retrial. We are not persuaded.

It is true that in Delk v. Atkinson, 665 F.2d 90 (6th Cir.1981), a habeas corpus case that did not involve a direct appeal from a federal conviction, we observed in a footnote — citing, inter al., United States v. Orrico, 599 F.2d 113, 116 (6th Cir.1979)-that

“[s]everal courts including this one have indicated that where it is claimed on appeal from a federal conviction that the evidence was insufficient, the reviewing court is required to decide the sufficiency question even though there might be other grounds for reversal which would not preclude retrial.” 665 F.2d at 93 n. 1.

In neither Delk nor Orrico, however, was this court dealing with a situation where the defendant had been tried on an indictment incorporating a legal theory that had turned out to be invalid. We faced just such a situation on the earlier appeal here — and it would have been pointless for us to sift through several days’ worth of trial evidence to determine whether the evidence would have been sufficient to support a conviction under the “intangible rights” theory if the Supreme Court had not already rejected that theory.

We saw no compelling need, similarly, to decide whether the evidence that was presented under the “intangible rights” theory on which Mr. Davis had in fact been indicted would have been sufficient to support a conviction under the “money or property” theory on which, as we held, he had not been indicted. We could not be certain that the government would ultimately decide to reindict, and considerations of judicial economy counseled in favor of the course we chose to follow even if the case or controversy requirement (U.S. Const., Article III, sec. 2) did not compel it.

But the new indictment is now a fact, and Mr. Davis invites us to decide now, before allowing the second prosecution to proceed, whether the evidence in the first trial would have been sufficient to support a conviction under the second indictment. We decline the invitation. Whether or not the evidence presented earlier would have warranted conviction under a valid indictment, the salient fact is that there was no such indictmefit earlier — and a valid indictment now having been returned, the government is both obliged and entitled to make its case anew. The government certainly would not be foreclosed, at a trial on the new indictment, from offering evidence that was not tendered at the first trial.

The new trial in this case, like the new trial allowed in a different context by Justices of Boston Municipal Court v. Lydon, supra, 466 U.S. 294, 104 S.Ct. 1805, 80 L.Ed.2d 311 (1984), can be regarded as “an enlarged, fact-sensitive part of a single, continuous course of judicial proceedings.” Id. at 309, 104 S.Ct. at 1813, quoting the court of appeals dissent of Campbell, J., 698 F.2d 1, 12 (1st Cir.1982). The proper time for determining the sufficiency of the evidence on the “money or property” theory embodied in the new indictment in the case at bar will come when the government has rested its case in the “enlarged, fact-sensitive” part of the proceedings against[*908] Mr. Davis. And the proper court to make the sufficiency determination, in the first instance, will be the court that hears the evidence.

If we are wrong in thinking it inappropriate for us to try to pass on the sufficiency of the evidence at this juncture, perhaps it is enough to say that we do not think the case presented by the government at the first trial would have been frivolous under a “money or property” indictment. Our initial impression (which is not to be deemed binding on the trial court) is that whether the $71,472 figure furnished to the city manager was intended to be a certification of Mr. Davis’s cost is a question of fact for the jury, and not a question for the court. What cost Mr. Davis actually incurred would likewise seem to be a question of fact — and it may prove to be a somewhat complex one, inviting expert testimony from both sides. Superficially, at least, the fact that Mr. Davis paid Wright Brothers Excavating Company only $36,297 for installing a line that may have been certified as costing Mr. Davis almost twice that amount suggests that the government was not way out in left field on this issue.

Citing a United States Attorneys’ manual, Mr. Davis tells us that “[tjhis case hardly rises to the usual level of federal concern embraced by the Department of Justice's published policies governing the exercise of discretion in the pursuit of mail fraud prosecutions....” That may be so, but there has been no showing here of any abuse of prosecutorial discretion of the sort claimed in United States v. Adams, 870 F.2d 1140 (6th Cir.1989). The government seems to have a colorable case, and we know of no reason why it may not bring it to trial, under a legally sufficient indictment, if it chooses to do so.

IV

The statutory limitations period for mail fraud prosecution is five years. 18 U.S.C. § 3282. Allowing for the tolling of the statute while the first indictment was in effect, Mr. Davis contends that the second indictment in this case was returned five years and ten days after the last of the four mailings on which the prosecution is based. The government maintains that the new indictment was timely under 18 U.S.C. § 3288, a savings statute that permits a new indictment to be returned within six months of the dismissal of an indictment which, after the statute of limitations has run, “is dismissed for any error, defect, or irregularity with respect to the grand jury_” Mr. Davis reads the savings statute as applying only where the indictment is dismissed for error of a sort that did not occur here, and he says that the new indictment broadened the charges im-permissibly in any event. The government reads the savings statute as applying where the indictment is found defective or insufficient for any cause, and the government denies that the new indictment is impermissibly broad. In denying Mr. Davis’s motion to dismiss the new indictment on statute of limitations grounds, the district court sided with the government. We shall not resolve the dispute here, because we are satisfied that we have no jurisdiction to do so in an interlocutory appeal.

Cohen v. Beneficial Loan Corp., 337 U.S. 541, 69 S.Ct. 1221, 93 L.Ed. 1528 (1949), and its progeny created a “collateral order” exception to the rule that only final judgments are appealable. To be appeal-able under the exception, the order must “conclusively determine the disputed question, resolve an important issue completely separate from the merits of the action, and be effectively unreviewable on appeal from a final judgment.” Coopers & Lybrand v. Livesay, 437 U.S. 463, 468, 98 S.Ct. 2454, 2457, 57 L.Ed.2d 351 (1978); Bauer v. Commerce Union Bank, 859 F.2d 438, 440 (6th Cir.1988), cert. denied, - U.S. -, 109 S.Ct. 1531, 103 L.Ed.2d 836 (1989).

Whether or not it “conclusively” determined the disputed question, the ruling that the statute of limitations did not bar the reprosecution was at least separate from the merits of the action. It did not, however, satisfy the third element of Cohen, which requires that the order be effectively unreviewable on appeal from a final judgment.

[*909] Mr. Davis suggests, in this connection, that he will irretrievably lose his right to be free from a costly, unnecessary, and inconvenient second trial if he has to wait until after a conviction to obtain an appellate determination as to whether the re-prosecution was untimely. But defendants in criminal proceedings often find themselves in comparable situations, and we are not convinced that Mr. Davis possesses any right arising from the Constitution or from a statute that will be irretrievably lost if we wait to decide the statute of limitations question until the conviction — if there is a conviction — is appealed.

Mr. Davis argues that his right to be free from an untimely prosecution is guaranteed by the words of the statute of limitations, 18 U.S.C. § 3282: “no person shall be prosecuted, tried, or punished for any offense, not capital, unless the indictment is found or the information is instituted within five years next after such offense shall have been committed.” Davis omits an introductory clause, however: “Except as otherwise expressly provided by law....” Read in its entirety, the statute of limitations seems to have been designed for a somewhat different purpose than the Double Jeopardy Clause. The Double Jeopardy Clause protects one against wrongly being tried or “placed in jeopardy,” but statutes of limitations, as the Third Circuit noted in an exhaustive discussion of this question, “acknowledge the state’s right to try certain persons, but then set boundaries on the exercise of that power.” United States v. Levine, 658 F.2d 113, 127 (3d Cir.1981).

“Limitations statutes ... are intended to foreclose the potential for inaccuracy and unfairness that stale evidence and dull memories may occasion in an unduly delayed trial. They provide relief from extended pretrial anxieties and, by encouraging investigation of recent crimes, contribute to a rational allocation of pros-ecutorial resources.... [I]t is primarily concerns of prosecutorial fairness and promptness, in contradistinction to insulation from repeated trials, that underlie limitations statutes. Consequently, pretrial review is neither critical nor always helpful in securing these interests.” (Emphasis in original.)

Id. As the Levine opinion says, the purpose that a statute of limitations is supposed to serve resembles the purpose of the Speedy Trial Clause of the Constitution. And in United States v. MacDonald, 435 U.S. 850, 98 S.Ct. 1547, 56 L.Ed.2d 18 (1978), the Supreme Court held that denial of a pretrial motion to dismiss a criminal prosecution on speedy trial grounds is not an appealable interlocutory order under Cohen. “Unlike the protection afforded by the Double Jeopardy Clause,” Justice Blackmun wrote, “the Speedy Trial Clause does not, either on its face or according to the decisions of this Court, encompass a 'right not to be tried’ which must be upheld prior to trial if it is to be enjoyed at all.” Id. at 861, 98 S.Ct. at 1553.

Courts have repeatedly held in other contexts that no generalized right exists to be free from prosecution — no such right, at least, as would create a corresponding right to an interlocutory appeal. See, for example, EEOC v. American Express Co., 558 F.2d 102 (2d Cir.1977), where the Court of Appeals for the Second Circuit rejected a claim of immediate appealability as to a district court ruling that the EEOC could maintain a discrimination suit notwithstanding that the same issue had been litigated, and lost, in a prior private action:

“Amex here does not claim that its right to proceed in another forum will be irretrievably lost, but rather that it should not have to litigate at all and thus save the expense and time involved in this suit in the district court. But that is true of all denials of motions to dismiss complaints. The defendant is compelled to try the cause on the merits, but by the same token the Court of Appeals is spared the problem of hearing piecemeal litigation.”

Id. at 104. See also Parr v. United States, 351 U.S. 513, 519-20, 76 S.Ct. 912, 916-17, 100 L.Ed. 1377 (1955) (“bearing the discomfiture and cost of a prosecution for crime even by an innocent person is one of the painful obligations of citizenship”).

[*910] In addition to filing his notice of appeal, Mr. Davis has petitioned this court for issuance of a writ of mandamus that would direct Judge Kinneary to dismiss the superseding indictment. The petition was filed as a “precaution,” to give us a jurisdictional basis for deciding the statute of limitations issue even if we should conclude that it does not fall within the collateral order exception of Cohen. Having so concluded, however, we do not agree that “precautionary” jurisdiction can be conferred. If it were that easy to get an appellate court to review a pretrial order, the final judgment rule would simply be eviscerated.

Mandamus is an “extraordinary” remedy, and may be granted only in “exceptional circumstances, amounting to a judicial usurpation of power.” Warden, Kentucky State Penitentiary v. Gall, 865 F.2d 786, 788 (6th Cir.1989), quoting Allied Chemical Corp. v. Daiflon, Inc., 449 U.S. 33, 35, 101 S.Ct. 188, 190, 66 L.Ed.2d 193 (1980). Mandamus ought not be “resorted to as a mode of review where a statutory method of appeal has been prescribed or to review an appealable decision of record.” Roche v. Evaporated Milk Association, 319 U.S. 21, 27-28, 63 S.Ct. 938, 942-43, 87 L.Ed. 1185 (1943). Cf. Vickers Motors, Inc. v. Wellford, 502 F.2d 967, 968 (6th Cir.1974) (Mandamus “should not issue as a substitute for appeal to review interlocutory orders that do not deprive ‘the parties of a trial before the court on the basic issues involved in the litigation.’ ”) Any other conclusion would “thwart the Congressional policy against piecemeal appeals in criminal cases.” Roche, 319 U.S. at 30, 63 S.Ct. at 943; Will v. United States, 389 U.S. 90, 96, 88 S.Ct. 269, 274, 19 L.Ed.2d 305 (1967).

The petition for a writ of mandamus is DENIED, and the order of the district court declining to dismiss the case on double jeopardy grounds is AFFIRMED.