Federal Rule of Criminal Procedure 6(d) states that only specified persons including "the witness under examination" may be present at a grand jury proceeding. In these cases, two Government witnesses testified in tandem before the grand jury, which indicted respondents and cross-petitioners (hereafter defendants) Mechanik and Lill for various drug-related offenses and conspiracy to commit such offenses. The Court of Appeals for the Fourth Circuit held that the simultaneous presence of these two witnesses violated Rule 6(d), and that even though the petit jury subsequently returned a verdict of guilty against defendants, the verdict must be set aside on any count that corresponds to a "tainted" portion of the indictment. We believe that the petit jury's verdict of guilty beyond a reasonable doubt demonstrates a fortiori that there was probable cause to charge the defendants with the offenses for which they were convicted. Therefore, the convictions must stand despite the rule violation.

A fairly detailed summary of the District Court proceedings will help to illustrate the nature and extent of our holding. A grand jury returned an indictment charging defendants with drug-related offenses and conspiracy. This indictment was concededly free from any claim of error. The grand jury then returned a superseding indictment in which the conspiracy charge was expanded. In support of this superseding indictment, the United States Attorney presented the testimony of two law enforcement agents who were sworn together and questioned in tandem before the grand jury.

The defendants did not learn about this joint testimony until after trial began. Before trial, they filed an omnibus motion requesting, inter alia, the names of all the people who appeared before the grand jury. The Government responded that there were no unauthorized persons appearing before the grand jury, and the District Court denied the motion. Trial began in February 1980, and concluded in early July of the same year. During the second week of trial, one Jerry Rinehart, an agent of the Drug Enforcement Administration, testified as a Government witness. At the time of his testimony, the Government furnished the defendants with a portion of the transcript of his grand jury testimony as required by the Jencks Act, 18 U.S.C. § 3500. The transcript disclosed that Rinehart and his fellow agent, Randolph James, had testified in tandem before the grand jury.

The defendants moved for dismissal of the indictment on the ground that the simultaneous presence of the two agents had violated Federal Rule of Criminal Procedure 6(d). Chief Judge Knapp, presiding over the trial, concluded that the presence and testimony of the two agents had not violated Rule 6(d), and he denied the motion. In May 1980, however, Chief Judge Knapp was unexpectedly hospitalized, and Judge Copenhaver took over as the trial judge. The defendants then moved for a rehearing of their motion to dismiss the indictment. Judge Copenhaver took the motion under advisement until the conclusion of trial.

In August 1980, after the jury had returned its guilty verdict, Judge Copenhaver ruled upon and denied the defendants' motion for dismissal of the indictment. 511 F.Supp. 50 (SD W.Va.1980). He first decided, contrary to Chief Judge Knapp's earlier ruling, that the joint testimony of Agents Rinehart and James did constitute a violation of Rule 6(d). Id., at 53-58. But he declined to set aside the defendants' indictment and convictions because, on the basis of a comparison between the two indictments and the evidence on which the indictments rested, the violation of Rule 6(d) had not harmed the defendants. Id., at 58-61. He justified this conclusion with respect to the substantive counts on the ground that they were materially unchanged from the valid initial indictment to the superseding indictment. Id., at 58-59. With respect to the conspiracy count, which had been expanded by the superseding indictment, he justified his conclusion on the ground that the grand jury "had before it ample independent evidence [apart from the joint testimony] to support a probable cause finding of the charges." Id., at 61. In light of these conclusions, Judge Copenhaver determined that a post-trial dismissal of the indictment would simply confer a windfall benefit on the defendants "who stand convicted after a three-month trial conducted at enormous expense to the United States and the defendants." Ibid. The judge nevertheless undertook to ensure future compliance with the one-witness rule by directing the Government to keep the court advised concerning compliance with Rule 6(d) in future criminal cases. Ibid.

A divided Court of Appeals reversed the conspiracy convictions, affirmed the others, and dismissed the conspiracy portion of the indictment. 735 F.2d 136 (1984). It reasoned that the language of Rule 6(d) is so "plain and unequivocal in limiting who may appear before a grand jury," id., at 139, that its transgression requires automatic reversal of any subsequent conviction regardless of the lack of prejudice. Id., at 139-140. But the court reversed only the conspiracy convictions because it found that the violation of Rule 6(d) tainted only the portion of the superseding indictment that related to them. Id., at 140. A divided en banc decision agreed. 756 F.2d 994 (1985) (per curiam ).

We assume for the sake of argument that the simultaneous presence and testimony of the two Government witnesses before the grand jury violated Rule 6(d), and that the District Court would have been justified in dismissing portions of the indictment on that basis had there been actual prejudice and had the matter been called to its attention before the commencement of the trial. But although the defendants appear to have been reasonably diligent in attempting to discover any error at the grand jury proceeding, they did not acquire the transcript showing that the two agents had appeared jointly in the grand jury proceeding until the second week of trial. Nor is there any suggestion that the Government designedly withheld the information. When the defendants made their motion to dismiss the indictment based on the joint testimony, Chief Judge Knapp denied the motion because of his view that there had been no violation of Rule 6(d). Judge Copenhaver eventually issued a contrary ruling on the resubmitted motion, but not until after the long and costly trial had been brought to its conclusion. Although we do not believe that the defendants can be faulted for any lack of diligence, we nonetheless hold that the supervening jury verdict made reversal of the conviction and dismissal of the indictment inappropriate.

Both the District Court and the Court of Appeals observed that Rule 6(d) was designed, in part, "to ensure that grand jurors, sitting without the direct supervision of a judge, are not subject to undue influence that may come with the presence of an unauthorized person." 735 F.2d, at 139. The Rule protects against the danger that a defendant will be required to defend against a charge for which there is no probable cause to believe him guilty. The error involving Rule 6(d) in these cases had the theoretical potential to affect the grand jury's determination whether to indict these particular defendants for the offenses with which they were charged. But the petit jury's subsequent guilty verdict means not only that there was probable cause to believe that the defendants were guilty as charged, but also that they are in fact guilty as charged beyond a reasonable doubt. Measured by the petit jury's verdict, then, any error in the grand jury proceeding connected with the charging decision was harmless beyond a reasonable doubt.1

It might be argued in some literal sense that because the Rule was designed to protect against an erroneous charging decision by the grand jury, the indictment should not be compared to the evidence produced by the Government at trial, but to the evidence produced before the grand jury. But even if this argument were accepted, there is no simple way after the verdict to restore the defendant to the position in which he would have been had the indictment been dismissed before trial. He will already have suffered whatever inconvenience, expense, and opprobrium that a proper indictment may have spared him. In courtroom proceedings as elsewhere, "the moving finger writes; and, having writ, moves on." Thus reversal of a conviction after a trial free from reversible error cannot restore to the defendant whatever benefit might have accrued to him from a trial on an indictment returned in conformity with Rule 6(d).

We cannot accept the Court of Appeals' view that a violation of Rule 6(d) requires automatic reversal of a subsequent conviction regardless of the lack of prejudice. Federal Rule of Criminal Procedure 52(a) provides that errors not affecting substantial rights shall be disregarded. We see no reason not to apply this provision to "errors, defects, irregularities, or variances" occurring before a grand jury just as we have applied it to such error occurring in the criminal trial itself. See United States v. Hasting, 461 U.S. 499, 509, 103 S.Ct. 1974, 1981, 76 L.Ed.2d 96 (1983); Chapman v. California, 386 U.S. 18, 21-24, 87 S.Ct. 824, 826-828, 17 L.Ed.2d 705 (1967); United States v. Lane, 474 U.S. 438, 106 S.Ct. 725, 88 L.Ed.2d 814 (1986).

The reversal of a conviction entails substantial social costs: it forces jurors, witnesses, courts, the prosecution, and the defendants to expend further time, energy, and other resources to repeat a trial that has already once taken place; victims may be asked to relive their disturbing experiences. See Morris v. Slappy, 461 U.S. 1, 14, 103 S.Ct. 1610, 1617, 75 L.Ed.2d 610 (1983). The "[p]assage of time, erosion of memory, and dispersion of witnesses may render retrial difficult, even impossible." Engle v. Isaac, 456 U.S. 107, 127-128, 102 S.Ct. 1558, 1571-1572, 71 L.Ed.2d 783 (1982). Thus, while reversal "may, in theory, entitle the defendant only to retrial, in practice it may reward the accused with complete freedom from prosecution," id., at 128, 102 S.Ct., at 1572, and thereby "cost society the right to punish admitted offenders." Id., at 127, 102 S.Ct., at 1572. Even if a defendant is convicted in a second trial, the intervening delay may compromise society's "interest in the prompt administration of justice," United States v. Hasting, supra, 461 U.S., at 509, 103 S.Ct., at 1981, and impede accomplishment of the objectives of deterrence and rehabilitation. These societal costs of reversal and retrial are an acceptable and often necessary consequence when an error in the first proceeding has deprived a defendant of a fair determination of the issue of guilt or innocence. But the balance of interest tips decidedly the other way when an error has had no effect on the outcome of the trial.

We express no opinion as to what remedy may be appropriate for a violation of Rule 6(d) that has affected the grand jury's charging decision and is brought to the attention of the trial court before the commencement of trial.2 We hold only that however diligent the defendants may have been in seeking to discover the basis for the claimed violation of Rule 6(d), the petit jury's verdict rendered harmless any conceivable error in the charging decision that might have flowed from the violation. In such a case, the societal costs of retrial after a jury verdict of guilty are far too substantial to justify setting aside the verdict simply because of an error in the earlier grand jury proceedings. The judgment of the Court of Appeals is therefore reversed to the extent it set aside the conspiracy convictions and dismissed the indictment, but is otherwise affirmed.

It is so ordered.

Chief Justice BURGER, concurring.

I concur in Justice REHNQUIST's opinion for the Court. I write separately only to state my view that this case is controlled by Justice Black's opinion for the Court in Costello v. United States, 350 U.S. 359, 76 S.Ct. 406, 100 L.Ed. 397 (1956).

Justice O'CONNOR, with whom Justice BRENNAN and Justice BLACKMUN join, concurring in the judgment.

I agree with the Court that the convictions obtained in the trial court against defendants Mechanik and Lill should not have been set aside. I write separately because I believe that the analysis adopted by the Court for determining the effect of a violation of the rules governing the conduct of grand juries effectively renders those rules a dead letter, thereby seriously undermining the grand jury's traditional function of protecting the innocent from unwarranted public accusation.

The grand jury has two principal functions. First, it bears the weighty responsibility of investigating crime and determining whether there is probable cause to believe that a crime has been committed. United States v. Calandra, 414 U.S. 338, 343, 94 S.Ct. 613, 617, 38 L.Ed.2d 561 (1974). The second, and no less important, task of the grand jury is to "serv[e] the invaluable function in our society of standing between the accuser and the accused, whether the latter be an individual, minority group, or other, to determine whether a charge is founded upon reason or dictated by an intimidating power or by malice and personal ill will." Wood v. Georgia, 370 U.S. 375, 390, 82 S.Ct. 1364, 1373, 8 L.Ed.2d 569 (1962). To further the grand jury's investigative function, the grand jury traditionally has been given "wide latitude" in its inquiries. Calandra, supra, 414 U.S., at 343, 94 S.Ct., at 617. See also United States v. Dionisio, 410 U.S. 1, 17-18, 93 S.Ct. 764, 773, 35 L.Ed.2d 67 (1973). Prosecutors have been accorded similar leeway in presenting their cases to the grand jury, see, e.g., United States v. Adamo, 742 F.2d 927, 936-938 (CA6 1984), cert. denied, 469 U.S. 1193, 105 S.Ct. 971, 83 L.Ed.2d 975 (1985), but they are bound by a few, clear rules which were carefully drafted and approved by this Court and by Congress to ensure the integrity of the grand jury's functions.

Federal Rule of Criminal Procedure 6(d) is one such rule; it is designed to guard the secrecy of the grand jury proceedings, prevent intimidation of jurors, and guarantee that the grand jury is given the opportunity to make an independent examination of the evidence and render its probable cause and charging determinations free of undue prosecutorial influence. See, e.g., United States v. Echols, 542 F.2d 948, 951 (CA5 1976), cert. denied, 431 U.S. 904, 97 S.Ct. 1695, 52 L.Ed.2d 387 (1977); United States v. Lill, 511 F.Supp. 50, 55-57 (SD W.Va.1980); 1 C. Wright, Federal Practice & Procedure § 105, p. 237, and n. 1 (2d ed. 1982). For example, if the Government, in violation of Rule 6(d), were to have all witnesses remain in the courtroom while the grand jury investigation was going on to ensure that all testified in a consistent manner, it cannot seriously be doubted that this practice would hinder the grand jury in its task of uncovering the truth. Similarly, if the prosecuting attorney were to remain in the jury room during the jury's deliberations in contravention of the Rule, a very real possibility would arise that the jury's deliberations or vote would be unduly influenced by the prosecutor's presence.

The Federal Rules clearly envision that dismissal of the indictment may be an appropriate remedy for such violations of Rule 6(d). See, e.g. Fed.Rule Crim.Proc. 12(b)(2). Indeed, courts have consistently employed the remedy of dismissal of the indictment for deviations from Rule 6(d) which may imperil the grand jury's independence. See, e.g., Lill, supra, at 58 (collecting federal cases). See also Nadel, Presence of Unauthorized Persons During State Grand Jury Proceedings as Affecting Indictment, 23 A.L.R. 4th 397 (1983) (hereinafter Nadel) (collecting state cases dealing with similar violations of state rules). To be sure, a violation must be clearly established before dismissal may be contemplated. The grand jury proceeding is accorded a presumption of regularity, which generally may be dispelled only upon particularized proof of irregularities in the grand jury process. See, e.g., United States v. Johnson, 319 U.S. 503, 512-513, 63 S.Ct. 1233, 1237-1238, 87 L.Ed. 1546 (1943). And not every violation of Rule 6(d) will create such a likelihood of prejudice as to warrant the drastic remedy of dismissal. The Federal Rules dictate that dismissal is appropriate only when a violation has impaired the substantial rights of the accused. See Fed.Rule Crim.Proc. 52.

Rule 52(a) provides that its harmless error inquiry extends to "[a]ny error, defect, irregularity or variance." The Advisory Committee's notes to Rule 52 state that "[t]his rule is a restatement of existing law," which specifically provided for a harmless error analysis of objections going to the validity of the indictment in 18 U.S.C. § 556 (1946 Ed.). See 18 U.S.C.App., p. 657. The language of Rule 6 does not exempt the Rule from a harmless error scrutiny. In fact, the commentary accompanying it states that the Rule "generally continues existing law" and expressly refers to the harmless error rule of 18 U.S.C. § 556 (1946 Ed.), thereby confirming the rulemakers' intent that violations of Rule 6(d) would be among those errors subject to harmless error review. See Advisory Committee's Notes on Fed.Rule Crim.Proc. 6(d), 18 U.S.C.App., p. 568.

A Rule 6(d) violation is one affecting the grand jury proceeding and is not in any sense a trial error. Accordingly, the logical focus of the harmless error inquiry is an examination of the influence of the error on the charging decision. Indeed, in most Rule 6(d) cases, a court conducting a harmless error inquiry will of necessity focus on the effect of the alleged error on the grand jury's charging decision, rather than the verdict, because the rules governing the disclosure of grand jury materials to defendants and the waiver provision of Rule 12(f) virtually ensure that all claims of violations of Rule 6(d) will be made before or during trial. See, e.g., 18 U.S.C. § 3500; Fed.Rules Crim.Proc. 6(e)(3)(C)(i), (ii); Dennis v. United States, 384 U.S. 855, 868-875, 86 S.Ct. 1840, 1848-1852, 16 L.Ed.2d 973 (1966); 8 J. Moore, Federal Practice ¶ 6.05[3] (2d ed. 1985) (discussing defendants' access to grand jury materials under the above cited authorities). See also Lill, supra, at 58, 61 (most reported Rule 6(d) claims raised and disposed of before trial).

In my view, when, as in these cases, a court decides to reserve the ruling on a timely raised and diligently pursued motion to dismiss based on an alleged violation of Rule 6(d) until after a verdict is returned, the focus of the court's inquiry should remain on the grand jury's charging decision. Yet the Court shifts the focus of the harmless error analysis in such circumstances from an examination of the violation's effect on the indictment to an assessment of the violation's effect on the trial verdict, without regard to the timing of the defendants' objection. See ante, at 70-71. The Court then concludes that a conviction automatically renders harmless any violation of Rule 6(d). This holding is not justified by the applicable rules, nor can it be reconciled with precedent.

A number of federal courts have employed a rule directly in conflict with that adopted by the Court: they presume the existence of prejudice from the presence of unauthorized persons in the grand jury room and apply a rule of automatic dismissal of the tainted indictment. See, e.g., United States v. Fulmer, 722 F.2d 1192, 1195, n. 5 (CA5 1983); Echols, 542 F.2d, at 951; Latham v. United States, 226 Fed. 420 (CA5 1915); Lill, 511 F.Supp., at 58 (collecting lower federal cases). See also Nadel § 4 (collecting state cases); 2 W. LaFave & J. Israel, Criminal Procedure § 15.6 (1984). Other courts have applied a harmless error test to Rule 6(d) violations, but have uniformly evaluated the prejudice to the defendant by looking to the violation's likely effect on the grand jury's deliberations, not merely to its significance in light of the trial verdict. See, e.g., United States v. Condo, 741 F.2d 238, 239 (CA9 1984) (per curiam ), cert. denied, 469 U.S. 1164, 105 S.Ct. 924, 83 L.Ed.2d 936 (1985); United States v. Computer Sciences Corp., 689 F.2d 1181, 1185-1186 (CA4 1982), cert. denied, 459 U.S. 1105, 103 S.Ct. 729, 74 L.Ed.2d 953 (1983); United States v. Kahan & Lessin Co., 695 F.2d 1122, 1124 (CA9 1982) (per curiam ); United States v. Rath, 406 F.2d 757 (CA6), cert. denied, 394 U.S. 920, 89 S.Ct. 1196, 22 L.Ed.2d 453 (1969). I have found no Rule 6(d) cases in which a per se rule based on the ultimate verdict at trial has been applied. Cf. also United States v. Lane, 474 U.S. 438, 447-448, 106 S.Ct. 725, 730-731, 88 L.Ed.2d 814 (1986) (disapproving use of per se rules in harmless error analysis).

The Court's focus on the effect of the verdict, in combination with its per se rule, gives judges and prosecutors a powerful incentive to delay consideration of motions to dismiss based on an alleged defect in the indictment until the jury has spoken. If the jury convicts, the motion is denied; if the jury acquits, the matter is mooted. The Court's approach thus undermines the authority of Rule 6(d), exposes to the ordeal of trial any defendants who would otherwise have a right to dismissal of the indictment, and undermines adherence to the very measures that this Court proposed and Congress implemented to guarantee that the grand jury is able to perform properly its screening function.

In my view, when a defendant makes a timely objection to the grand jury indictment based on a violation of Rule 6(d), the remedy of dismissal of the indictment is appropriate if it is established that the violation substantially influenced the grand jury's decision to indict, or if there is grave doubt as to whether it had such effect. See Lane, supra, at 449, 106 S.Ct., at ---- - ----; Kotteakos v. United States, 328 U.S. 750, 765, 66 S.Ct. 1239, 1248, 90 L.Ed. 1557 (1946). The focus of the prejudice inquiry should be on the effect of the alleged error on the grand jury's decision to indict even if the court postpones its decision until the conclusion of the trial.

This approach would not impose unwarranted costs on the criminal justice system. The prosecution has it within its power to avoid dismissals by adhering to the simple dictates of Rule 6(d), and to remedy any violation of the Rule by obtaining a superseding indictment so as to avoid the risks of subsequent dismissals. Even where an unremedied violation is proved, the trial judge must still be satisfied that the violation resulted in grand jury intimidation or improper influence on important witnesses' testimony and thus had a "substantial influence" on the indictment returned, or that there is grave doubt as to whether it had such effect. Kotteakos, supra, at 765, 66 S.Ct., at 1248. See also Lane, supra, at 449, 106 S.Ct., at ----.

In these cases, the District Court found, after a scrupulous examination of the record, that the violation was harmless, a finding which was not found wanting on appeal. The District Court preliminarily observed that:

"It is especially significant to note that the two indictments were returned by the same grand jury. The court's review of the attendance and voting records of that grand jury reveals that each of these indictments was returned by a unanimous vote. A nucleus of the same seventeen grand jurors voted for each indictment. In addition, one other grand juror voted for the first indictment but did not vote on the second, while two others voted for the second indictment but did not vote on the first." 511 F.Supp., at 58-59.

Both agents had testified separately before the same grand jury in support of the first, untainted indictment, giving the jury ample opportunity to weigh the credibility of each agent prior to their joint appearance. Moreover, both agents had access to all grand jury materials in the case pursuant to Rule 6(e)(3)(A)(ii), and thus the likelihood that their joint testimony created a potential for collusion not already available is minute.

Turning to an examination of the indictment itself, the District Court found that the substantive counts of which the defendants were convicted were "identical or virtually so" to the counts returned in the superseded indictment and that those counts had "a probable cause basis entirely independent of the testimony presented to the grand jury after the return of the [superseded] first indictment." Id., at 59. It concluded that as to the substantive counts, "there was neither prejudice nor potential for prejudice." Ibid. Thus, the only count upon which the defendants could have been prejudiced by the objectionable joint testimony of the Drug Enforcement Administration agents was the conspiracy count.

Although the District Court conducted its post-trial harmless error review in part with an eye to the effect of the error on the verdicts, its findings also make clear that the effect of the joint testimony on the grand jury's decision to indict on the conspiracy count was negligible. The District Court carefully isolated the alterations and additions to that count which were the subject of the joint testimony. After examining the testimony given by other grand jury witnesses, the trial judge concluded that "the grand jury would, in my view, undoubtedly have returned the very same second indictment even had [the] Agents . . . testified separately." Id., at 61. Accordingly, I would reverse the judgment of the Court of Appeals insofar as it set aside the defendants' conspiracy convictions and affirm the Court of Appeals' judgment regarding the defendants' cross-petitioners.

Justice MARSHALL, dissenting.

The Court concedes that federal prosecutors violated Rule 6(d) of the Federal Rules of Criminal Procedure in presenting their case against defendants Mechanik and Lill to the grand jury. The Court holds, however, that because defendants were ultimately convicted of some of the counts against them, "any error in the grand jury proceeding connected with the charging decision was harmless beyond a reasonable doubt." Ante, at 70. Because I believe that the majority's rule misconceives the role both of the grand jury and of the harmless-error doctrine, I dissent.

* The Court's decision today renders Rule 6(d) almost unenforceable. As the facts of this litigation demonstrate, Rule 6(d) violations are difficult for defendants to uncover. The grand jury conducts its investigation in secret, aided only by the prosecutors and witnesses. United States v. Calandra, 414 U.S. 338, 343, 94 S.Ct. 613, 617, 38 L.Ed.2d 561 (1974). Defendants are not entitled to grand jury transcripts before trial; due to the strictly enforced tradition of grand jury secrecy, defendants generally have access to no information whatsoever regarding the conduct of the grand jury proceedings. See M. Frankel & G. Naftalis, The Grand Jury 81-89 (1977). Requests by defendants pursuant to Rule 6(e)(3)(C)(ii) for disclosure of grand jury materials, "upon a showing that grounds may exist for a motion to dismiss the indictment because of matters occurring before the grand jury," are rarely granted; a defendant often can make the necessary showing only with the aid of the materials he seeks to discover. See 1 C. Wright, Federal Practice and Procedure § 108, pp. 263-265 (2d ed. 1982). Defendants' only access to grand jury materials is likely to be through the medium of the Jencks Act, 18 U.S.C. § 3500, which requires the prosecutor, after direct examination of a Government witness, to produce the witness' prior statements. That disclosure, however, does not take place until after trial has begun, and then only on a piecemeal and incomplete basis.

There is thus little likelihood that a defendant can raise a substantial claim under Rule 6(d) before his trial begins. After the start of trial, overwork and the press of events may prevent the district judge from disposing of a newly raised Rule 6(d) claim. The most attractive course for the district judge will be to defer a ruling until the close of trial, the course ultimately followed in this case. Indeed, the district judge may not have the opportunity to rule until that time. Under today's decision, however, deferring a meritorious Rule 6(d) claim until the close of trial disposes of it permanently. If the movant is acquitted, then his Rule 6(d) motion is moot; if the movant is convicted, under the majority's reasoning, then any error was harmless. The Court's decision thus offers busy district judges a new and unique way to reduce their workload; one need not believe in a judicial conspiracy against the assertion of Rule 6(d) rights to suspect that district judges, faced with Rule 6(d) motions necessarily raised in the middle of trial, will follow the Court's invitation.

Should a district judge decide a Rule 6(d) motion during trial, the majority's scheme insulates that ruling from appellate review. Appeal before judgment is unlikely; the Court has never allowed immediate appeal of an order issued after the start of a criminal trial. See Flanagan v. United States, 465 U.S. 259, 269, 104 S.Ct. 1051, 1057, 79 L.Ed.2d 288 (1984).1 And under the decision today, such rulings cannot be appealed after judgment. Enforcement of Rule 6(d) is thus left to the unreviewable largesse of the district court.

We have no reason to believe that Congress intended Rule 6(d) to have so little practical meaning. The legislative history of Rule 6, indeed, belies that approach. In 1933, Congress was faced with "a conflict of legal decision as to the right, under existing law, to permit stenographers in grand jury rooms without invalidating the subsequent conviction of defendant." S.Rep. No. 64, 73d Cong., 1st Sess., 1 (1933). It responded by adding a narrow clause to the harmless-error statute, Rev.Stat. § 1025 (later codified as 18 U.S.C. § 556 (1946 ed.)), providing that no indictment should be found insufficient, or conviction be reversed, because of the presence of stenographers in the grand jury room. But Congress nowhere expressed disagreement with the general proposition that the presence of an unauthorized person in the grand jury room invalidates a subsequent conviction. E.g., United States v. Fall, 56 App.D.C. 83, 84, 10 F.2d 648, 649 (1925); Latham v. United States, 226 Fed. 420, 424 (CA5 1915).

More recently, Congress amended Rule 6 in 1972 to incorporate by reference the provisions of the Jury Selection and Service Act of 1968; it provided that a defendant may move to dismiss the indictment based on the Government's failure to comply with that Act in the selection of the grand jury array or of individual grand jurors. Fed.Rule Crim.Proc. 6(b)(2). The advisory notes expressly state that the district judge may rule on such a challenge to the grand jury either before or after the verdict. Advisory Committee Notes on Fed.Rule Crim.Proc. 6, 18 U.S.C.App., p. 568 (1972 amendment). There is no hint that Congress, in providing for a ruling after the verdict, intended that ruling to be a mere intellectual exercise.

The majority's opinion misconceives the role of harmless-error analysis. We have recognized that harmless-error doctrine, denying any remedy in cases of clear prosecutorial misconduct, "can work very unfair and mischievous results." Chapman v. California, 386 U.S. 18, 22, 87 S.Ct. 824, 827, 17 L.Ed.2d 705 (1967). Denying defendants relief for clear violations of their procedural rights reduces the law to " 'pretend-rules,' " United States v. Borello, 766 F.2d 46, 58 (CA2 1985), quoting United States v. Antonelli Fireworks Co., 155 F.2d 631, 661 (CA2) (Frank, J., dissenting), cert. denied, 329 U.S. 742, 67 S.Ct. 49, 91 L.Ed. 640 (1946); it means that prosecutors are free to engage in prohibited conduct subject only to "purely ceremonial" words of appellate displeasure. 155 F.2d, at 661.

The Court's rule that all grand jury misconduct becomes harmless after conviction, however, is especially pernicious. Contrary to the majority's suggestion that reversal is too costly a remedy for grand jury misconduct, ante, at 72, it is the majority's refusal to reverse convictions for demonstrated grand jury misconduct that imposes unacceptable costs. There are few limitations on the conduct of the prosecutor before the grand jury. Those limitations are found only in Federal Rule of Criminal Procedure 6, the text of which takes up little more than a page in the official compilation of United States laws. Violations of even those isolated restrictions, in by far the majority of cases, will go undetected by defendants. The only way to allow even minimally effective enforcement of those rules is to reverse the convictions of defendants whose indictments were tainted by Rule 6 violations.

Such an approach would not hamper the enforcement of the criminal law. Violations of Rule 6(d) will be nonexistent if the prosecutor exercises proper control over access to the grand jury chambers. The substantive law is not onerous or ambiguous, and most violations are the product of the prosecutor's failure to adopt safeguards to ensure compliance. See 2 W. LaFave & J. Israel, Criminal Procedure § 15.6, p. 333 (1984). There is no danger of a prosecutor slipping into an inadvertent Rule 6(d) violation comparable to that, say, of making an ill-worded remark in the heat of trial. Courts would not often have cause to reverse convictions because of Rule 6(d) violations.

The majority's goal of upholding criminal convictions not marred by substantial defect does not justify reducing Congress' command regarding the proper conduct of grand jury proceedings to a mere form of words, without practical effect. Respect for the rule of law demands that improperly procured indictments be quashed even after conviction, because "only by upsetting convictions so obtained can the ardor of prosecuting officials be kept within legal bounds and justice be secured; for in modern times all prosecution is in the hands of officials." United States v. Remington, 208 F.2d 567, 574 (CA2 1953) (L. Hand, J., dissenting).2

The opinion concurring in the judgment suggests that the Rule 6(d) violation in this litigation should be viewed as harmless on the theory that the grand jury would have returned the same indictment regardless of the prosecutor's misconduct. Under that approach, a district court faced with a Rule 6(d) violation should examine the grand jury transcripts in an attempt to divine the effect of the violation on the jury's charging decision, and allow the indictment or conviction to stand only if it finds that there was no such effect. Such a rule would be contrary to the traditional black-letter law that "[a]ny violation of Rule 6(d) is per se prejudicial to the defendant and will result in dismissal of the indictment," 8 J. Moore, Federal Practice ¶ 6.04[7], p. 6-91 (2d ed. 1985).3 I believe that such an approach would be unworkable and would undermine the limits Congress imposed on the conduct of grand jury investigations.

Many of the reasons given above for rejecting the majority's view that grand jury impropriety is always harmless once a verdict is reached, apply in this context as well. Given defendants' difficulty in discovering Rule 6(d) violations, it is all the more important that dismissal of the indictment be certain when violations of the Rule are found. Only such a sanction can come close to providing prosecutors with an incentive to obey the Rule's commands. See United States v. Pignatiello, 582 F.Supp. 251, 255 (Colo.1984).

Such harmless-error analysis, moreover, overlooks the practical impossibility of determining the effect of a Rule 6(d) violation. The prejudicial impact of the unauthorized presence of persons in the grand jury room will often be impossible to quantify, and may not be apparent from the grand jury transcript. As one court wrote:

"A change in expression, a pressure on the hand or a warning glance would not be shown upon the minutes but might well influence, suppress or alter testimony to the prejudice of the defendant. There may have been prior expressions or conversations between the two witnesses which the one then giving testimony might well hesitate to repudiate or modify in the presence of the other. The District Attorney here contends . . . that defendant suffered no prejudice by the joint presence of the two sisters, but . . . '[t]he court cannot know that this suggestion represents the fact.' We think the practice offers too great a possibility for the exercise of undue influence to be condoned." State v. Revere, 232 La. 184, 207, 94 So.2d 25, 34 (1957) (emphasis omitted; citations omitted; internal quotations omitted).

Any case-by-case analysis to determine whether the defendant was actually prejudiced is simply too speculative to afford defendants meaningful protection, and imposes a difficult burden on the courts that outweighs the benefits to be derived. The distinction between the truly harmless error and the more dangerous one is not "such a pronounced one that the Court can cloak the one with the mantle of legality and yet recognize the dangers of the other and prohibit it." United States v. Carper, 116 F.Supp. 817, 821 (DC 1953).

That approach, finally, is likely to require a detailed inquiry that will frustrate and undermine the secrecy of grand jury inquiry. See United States v. Treadway, 445 F.Supp. 959 (ND Tex.1978). The district court may have to discuss the testimony of grand jury witnesses who did not appear at trial. The goals of grand jury secrecy, however, counsel that such analysis should not be spread across the public record. See United States v. Sells Engineering, Inc., 463 U.S. 418, 424-425, 103 S.Ct. 3133, 3138, 77 L.Ed.2d 743 (1983).4

This litigation illustrates the extent to which the Court is willing to reduce the substantive law to "pretend-rules," Borello, 766 F.2d, at 58, in order to affirm a criminal conviction. But by denigrating Congress' commands and eviscerating enforcement of Rule 6(d), the Court creates "a greater danger to a free people than the escape of some criminals from punishment." United States v. Di Re, 332 U.S. 581, 595, 68 S.Ct. 222, 229, 92 L.Ed. 210 (1948). I believe that the District Court in this case should have reversed defendants' conspiracy convictions without inquiry into the prejudice done to defendants by the Rule 6(d) violation. I therefore would affirm the judgment of the Court of Appeals.