with whom
Justice Brennan and Justice Marshall join, dissenting.A guilty plea, for all its practical importance in the day-today administration of justice, does not bestow on the Government any power to prosecute that it otherwise lacks. Here, after remand, the District Court found, and the Court of Appeals affirmed, that the two indictments brought against respondents charged two parts of the same conspiracy, and therefore sought to punish respondents twice for the same behavior, in violation of the Double Jeopardy Clause of the Fifth Amendment.
The Government, see ante, at 569, does not contest the finding that in fact there was only one conspiracy. It argues, however, that the defendants’ guilty pleas render this fact wholly irrelevant, and urges us to let stand convictions that otherwise are barred. Because I believe it inappropriate for a reviewing court to close its eyes to this constitutional violation, and because I find that the basis of respondents’ double jeopardy challenge is obvious from a reading of the two indictments and entitles respondents to a hearing, I dissent from the majority’s ruling that the guilty pleas are conclusive.
I
As noted in Brady1 and by the majority today, in most instances a guilty plea is conclusive and resolves all factual issues necessary to sustain a conviction. But in Blackledge v. Perry, 417 U. S. 21 (1974), and in Menna v. New York, 423 U. S. 61 (1975), this Court unequivocally held that a guilty *582plea does not waive a defendant’s right to contest the constitutionality of a conviction “[w]here the State is precluded by the United States Constitution from haling a defendant into court.” Id., at 62; see also Blackledge, 417 U. S., at 30. Although our recent decision in Ricketts v. Adamson, 483 U. S. 1 (1987), allows a defendant to waive a double jeopardy claim as part of a clearly worded plea agreement, none of our prior cases limited a defendant’s ability, under Menna and Blackledge, absent an express waiver, to challenge the Government’s authority to bring a second charge.
It is true, as the majority notes, that neither Blackledge nor Menna involved an independent evidentiary hearing to assess the defendants’ double jeopardy claims. But nothing in Blackledge or Menna indicates that the general constitutional rule announced in those cases was dependent on the fortuity that the defendants’ double jeopardy claims were apparent from the records below without resort to an evidentiary hearing. This is not surprising. There simply was no need for an evidentiary hearing in either Blackledge or Menna. Certainly, nothing in those cases suggests that a collateral proceeding would be inappropriate. Blackledge was a habeas proceeding in which the record was already fully developed, 417 U. S., at 23; and the remand in Menna from this Court to the New York Court of Appeals was not limited in any way, 423 U. S., at 63. To the extent that the majority reads the particular circumstances of those cases as compelling, or even implying, that the need for an evidentiary hearing alters the effect of a guilty plea, it infuses mere happenstance with constitutional meaning and draws distinctions where none belong.
The majority also justifies its outcome by looking to four words of dicta in a footnote in Menna, 423 U. S., at 62-63, n. 2. The relevant language in the Menna footnote is: “[A] plea of guilty to a charge does not waive a claim that — judged on its face — the charge is one which the State may not constitutionally prosecute” (emphasis added). The majority *583takes this language to mean that respondents can prevail only if they prove their claim by relying on nothing more than the indictment and the record.
A much better reading of the Menna footnote, however, is to place the emphasis on the word “claim.” Accordingly, if a claim that the Government was without power to prosecute is apparent on the face of the indictment, read in light of the existing record, a court should not consider the claim to have been waived, and must go on to consider its merits. This interpretation is true to the outcome in both Menna and Blackledge. It also gives appropriate force to the footnote’s language and its apparent purpose of placing some limit on the ability of a defendant who has pleaded guilty to make a later collateral attack without some foundation in the prior proceedings. Most important, it gives real content to the defendants’ constitutional rights.
II
This case provides a powerful example of why there is an especially great need to maintain the right collaterally to attack guilty pleas in the conspiracy context. Conspiracy, that “elastic, sprawling and pervasive offense,” Krulewitch v. United States, 336 U. S. 440, 445 (1949) (Jackson, J., concurring in judgment and opinion of Court), long has been recognized as difficult to define and even more difficult to limit. When charging a conspiracy, a prosecutor is given the opportunity to “cast his nets” in order to cover a broad timeframe and numerous acts and individuals, in part because conspiracies by their nature are clandestine and difficult to uncover. See, e. g., Blumenthal v. United States, 332 U. S. 539, 557 (1947). But this very permissible breadth of conspiracy indictments provides potential for abuse and confusion. Judge Parker said it meaningfully 50 years ago:
“Blanket charges of ‘continuing’ conspiracy with named defendants and with ‘other persons to the grand jurors unknown’ fulfil a useful purpose in the prosecution *584of crime, but they must not be used in such a way as to contravene constitutional guaranties. If the government sees fit to send an indictment in this general form charging a continuing conspiracy for a period of time, it must do so with the understanding that upon conviction or acquittal further prosecution of that conspiracy during the period charged is barred, and that this result cannot be avoided by charging the conspiracy to have been formed in another district where overt acts in furtherance of it were committed, or by charging different overt acts as having been committed in furtherance of it, or by charging additional objects or the violation of additional statutes as within its purview, if in fact the second indictment involves substantially the same conspiracy as the first. . . . The constitutional provision against double jeopardy is a matter of substance and may not be thus nullified by the mere forms of criminal pleading.” Short v. United States, 91 F. 2d 614, 624 (CA4 1937).
This Court noted in Sanabria v. United States, 437 U. S. 54, 65-66 (1978): “The precise manner in which an indictment is drawn cannot be ignored, because an important function of the indictment is to ensure that, ‘in case any other proceedings are taken against [the defendant] for a similar offense, . . . the record [will] sho[w] with accuracy to what extent he may plead a former acquittal or conviction,’” quoting Cochran v. United States, 157 U. S. 286, 290 (1895). See also Russell v. United States, 369 U. S. 749 (1962).
As the Court of Appeals recognized, the two indictments at issue here were broad and vague and substantially overlapped. Although the majority has included in the Appendix to its opinion, ante, p. 576, the few paragraphs in the two indictments which differ, it fails to acknowledge that the indictments otherwise are almost identical.
The indictments alleged acts occurring in the same place, having the same object of eliminating competition on a high*585way project, and having the same effect of restraining competition. More important, the first indictment is vague and open-ended in a number of material respects. While alleging a definite beginning date, the first indictment specified no termination date. As a consequence, the acts alleged in the second indictment were contained within the timeframe of the first. Moreover, the first indictment alleged that respondents conspired with “others known and unknown”; so, too, did the second indictment. Both indictments, therefore, may have involved the same participants. This vagueness, coupled with the express identical elements, provides a strong inference that the two agreements alleged were part of the same conspiracy.2 For this reason alone, there are sufficient grounds for raising a double jeopardy challenge under a proper reading of our decisions in Merma and Blackledge.
That the two indictments were duplicitous is further betrayed by the nature of the charged offense. The indictments state that the conspirators designated a low bidder on each project, submitted artificially high or complementary bids, and discussed paying consideration to other contractors to induce those contractors to submit noncompetitive rigged bids as well. Ante, at 577-578, 579. Although it is theoretically possible that such a conspiracy might involve only one project, it is highly unlikely. Rather, it seems reasonably clear to me, as it should have been to the Government, that in order to make any sense such an agreement must involve a number of projects, so that a conspirator who agreed to sub*586mit a sham bid on one project would be rewarded by being chosen for the successful bid on another project. In fact, a Justice Department release issued several weeks after the second indictment was filed described a Tennessee highway bid-rigging scheme as follows: “ ‘The prearranged low bidder would usually get the job as other contractors submitted intentionally high bids, knowing their turn as low bidder was coming.’” 42 BN A Antitrust & Trade Regulation Rep. 523 (1982), quoting unpublished release. See generally U. S. General Accounting Office, Report to the House Committee on Public Works and Transportation, Actions Being Taken to Deal with Bid Rigging in the Federal Highway Program (May 23, 1983). The very nature of the conspiracy alleged all but compels the conclusion that the initial indictment charged an ongoing agreement covering numerous projects.*
The Government argues that the respondents should have realized all this, and refused to plead to the second indictment. I agree. But it is no less true that the Government should have been aware that it could be charging duplicitous conspiracies, and, if so, not brought the second indictment. I fail to see why a reviewing court should punish the respondents’ oversight, but reward the Government’s.
“ ‘The Double Jeopardy Clause is not such a fragile guarantee that... its limitations [can be avoided] by the simple expedient of dividing a single crime into a series of temporal or spatial units.’” Sanabria v. United States, 437 U. S., at 72, quoting Brown v. Ohio, 432 U. S. 161, 169 (1977). As we pointed out in Braverman v. United States, 317 U. S. 49, 52 (1942), there may be a “single continuing agreement to com*587mit several offenses.” On the face of the two indictments, there was clear support for a claim that prosecuting the second indictment was barred by double jeopardy.
Ill
The question remains as to what procedures a reviewing court should follow when faced with such a double jeopardy claim.
As noted above, our prior cases and common sense require that the reviewing court consider the record in determining whether the claim of double jeopardy is sufficient to bar the second prosecution. It may be that in most cases the issue can be determined by reference to the record alone. Statements made at the plea hearing or other pretrial proceeding may be sufficient to clarify any ambiguity, or may constitute an express waiver of any double jeopardy challenge. But in the absence of a definitive record, an evidentiary hearing may be necessary in order to assure that the questioned indictment in fact alleges separate criminal conduct.
An evidentiary hearing on the double jeopardy issue would not be overly burdensome or replicate the trial that the guilty plea avoided. As noted in Abney v. United States, 431 U. S. 651, 659 (1977), in a claim of double jeopardy “the defendant makes no challenge whatsoever to the merits of the charge against him.” Although the nature of the evidentiary hearing obviously will depend on the facts of the particular case, for a challenge similar to the one here the hearing probably would involve only the Government’s explanation of how the conduct charged in the second indictment differs from the facts established by the guilty plea to the first indictment, and the defendants’ arguments to the contrary. The truth of many of the relevant facts will have been established by the guilty plea to the first indictment, and the legal sufficiency and independence of the second indictment should be determin*588able without substantial additional testimony.4 These challenges rarely should involve extensive proceedings.
The Government’s complaint that conducting an evidentiary hearing will present it with problems of proof, as well as administrative headaches, may have a modicum of force. Every prosecutor, however, has the power to avoid this by more carefully considering the actual scope of the alleged conspiracy, and by carefully drawing the indictment. The prosecutor also may ensure that any double jeopardy concerns are addressed at the plea hearing by describing with some particularity the scope of the agreement that is the basis of the conspiracy.5 While such steps are not absolutely required, each makes good sense, and would help to assure that every issue that should be raised at the plea hearing will be raised. Directly addressing double jeopardy questions at the plea hearing will prevent situations like the one at issue here. Once on notice, a defendant might expressly waive any double jeopardy challenge, see Ricketts v. Adamson, 483 U. S. 1 (1987), or might reconsider his inclination to plead guilty and, instead, litigate the issue.
This solution, it seems to me, properly balances the interests of the Government in finality of convictions pursuant to guilty pleas with those of criminal defendants who may have been unaware of their rights when pleading guilty. The Constitution’s prohibition against placing a defendant in jeopardy twice for the same conduct is fundamental, and no less applicable because a complicated question of conspiracy law *589may be presented. Because I believe that there is no legitimate interest in either punishing defendants twice for the same conduct or in allowing the Government to gain untoward benefits from the use of vague and imprecise indictments, and that an evidentiary hearing would not be a significant burden in the few cases where it would be necessary, I dissent.
Brady v. United States, 397 U. S. 742, 748 (1970); see also McMann v. Richardson, 397 U. S. 759, 766 (1970).
In determining how many conspiracies are involved in a particular case, courts have looked to a number of discrete factors. Some of these include the relevant (1) time, (2) participants, (3) statutory offenses charged, (4) overt acts charged, and (5) places where the alleged acts took place. See United States v. Ragins, 840 F. 2d 1184, 1188-1189 (CA4 1988); United States v. Atkins, 834 F. 2d 426, 432 (CA5 1987); see also United States v. Korfant, 771 F. 2d 660, 662 (CA2 1985) (considering eight factors).
The majority’s reading of the indictments appears to focus solely on the fact that each states a separate agreement, relating to a separate project. See ante, at 570-571. Had the majority reached the issue raised by Justice Stevens, in his separate concurring opinion, ante, p. 580, and decided that multiple conspiracies within an ongoing conspiracy could be prosecuted separately, then those allegations might be determinative. The majority, however, has not done this.
Indeed, we know already that this case did not require a long, complicated hearing. By the Government’s stipulation, the District Court considered the record in the Beachner case, see ante, at 566-567, as if that record had been a part of the plea proceedings.
It would also be worthwhile for the Government to provide a defendant with a copy of each indictment well in advance of the scheduled plea hearing. Here the defendants first received a copy of the second indictment on February 8, 1981, the same day on which the guilty pleas were entered. This may have contributed to respondents’ failure to raise the double jeopardy issue at that time.