No. 57

Argued: November 13, 1969Decided: April 6, 1970

[397 U.S. 436 , 437]   Clark M. Clifford, Washington, D.C., for petitioner.

Gene E. Voigts, Jefferson City, Mo., for respondent.

Mr. Justice STEWART delivered the opinion of the Court.

In Benton v. Maryland, 395 U.S. 784 , the Court held that the Fifth Amendment guarantee against double jeopardy is enforceable against the States through the Fourteenth Amendment. The question in this case is whether the State of Missouri violated that guarantee when it prosecuted the petitioner a second time for armed robbery in the circumstances here presented. 

Sometime in the early hours of the morning of January 10, 1960, six men were engaged in a poker game in the basement of the home of John Gladson at Lee’s Summit, Missouri. Suddenly three or four masked men, armed with a shotgun and pistols, broke into the basement and robbed each of the poker players of money and various articles of personal property. The robbers-and it has never been clear whether there were three or four of them-then fled in a car belonging to one of the victims of the robbery. Shortly thereafter the stolen car was discovered in a field, and later that morning three men were arrested by a state trooper while they were walking on a highway not far from where the abandoned car had been found. The petitioner was arrested by another officer some distance away. [397 U.S. 436 , 438]   The four were subsequently charged with seven separate offenses-the armed robbery of each of the six poker players and the theft of the car. In May 1960 the petitioner went to trial on the charge of robbing Donald Knight, one of the participants in the poker game. At the trial the State called Knight and three circumstances of the holdup and itemized witnesses. Each of them described the circumstances f the holdup and itemized his own individual losses. The proof that an armed robbery had occurred and that personal property had been taken from Knight as well as from each of the others was unassailable. The testimony of the four victims in this regard was consistent both internally and with that of the others. But the State’s evidence that the petitioner had been one of the robbers was weak. Two of the witnesses thought that there had been only three robbers altogether, and could not identify the petitioner as one of them. Another of the victims, who was the petitioner’s uncle by marriage, said that at the ‘patrol station’ he had positively identified each of the other three men accused of the holdup, but could say only that the petitioner’s voice ‘sounded very much like’ that of one of the robbers. The fourth participant in the poker game did identify the petitioner, but only by his ‘size and height, and his actions.’

The cross-examination of these witnesses was brief, and it was aimed primarily at exposing the weakness of their identification testimony. Defense counsel made no attempt to question their testimony regarding the holdup itself or their claims as to their losses. Knight testified without contradiction that the robbers had stolen from him his watch, $250 in cash, and about $500 in checks. His billfold, which had been found by the police in the possession of one of the three other men accused of the robbery, was admitted in evidence. The defense offered no testimony and waived final argument. [397 U.S. 436 , 439]   The trial judge instructed the jury that if it found that the petitioner was one of the participants in the armed robbery, the theft of ‘any money’ from Knight would sustain a conviction. He also instructed the jury that if the petitioner was one of the robbers, he was guilty under the law even if he had not personally robbed Knight. The jury- though not instructed to elaborate upon its verdict-found the petitioner ‘not guilty due to insufficient evidence.’

Six weeks later the petitioner was brought to trial again, this time for the robbery of another participant in the poker game, a man named Roberts. The petitioner filed a motion to dismiss, based on his previous acquittal. The motion was overruled, and the second trial began. The witnesses were for the most part the [397 U.S. 436 , 440]   same, though this time their testimony was substantially stronger on the issue of the petitioner’s identity. For example, two witnesses who at the first trial had been wholly unable to identify the petitioner as one of the robbers, now testified that his features, size, and mannerisms matched those of one of their assailants. Another witness who before had identified the petitioner only by his size and actions now also remembered him by the unusual sound of his voice. The State further refined its case at the second trial by declining to call one of the participants in the poker game whose identification testimony at the first trial had been conspicuously negative. The case went to the jury on instructions virtually identical to those given at the first trial. This time the jury found the petitioner guilty, and he was sentenced to a 35-year term in the state penitentiary.

The Supreme Court of Missouri affirmed the conviction, holding that the ‘plea of former jeopardy must be denied.’ State v. Ashe, 350 S.W.2d 768, 771. A collateral attack upon the conviction in the state courts five years later was also unsuccessful. State v. Ashe, 403 S.W.2d 589. The petitioner then brought the present habeas corpus proceeding in the United States District Court for the Western District of Missouri, claiming that the second prosecution had violated his right not to be twice put in jeopardy. Considering itself bound by this court’s decision in Hoag v. New Jersey, 356 U.S. 464 , the District Court denied the writ, although apparently finding merit in the petitioner’s claim. The Court [397 U.S. 436 , 441]   of Appeals for the Eighth Circuit affirmed, also upon the authority of Hoag v. New Jersey, supra. We granted certiorari to consider the important constitutional question this case presents. 393 U.S. 1115 , 89 S. Ct. 995.

As the District Court and the Court of Appeals correctly noted, the operative facts here are virtually identical to those of Hoag v. New Jersey, supra. In that case the defendant was tried for the armed robbery of three men who, along with others, had been held up in a tavern. The proof of the robbery was clear, but the evidence identifying the defendant as one of the robbers was weak, and the defendant interposed an alibi defense. The jury brought in a verdict of not guilty. The defendant was then brought to trial again, on an indictment charging the robbery of a fourth victim of the tavern holdup. This time the jury found him guilty. After appeals in the state courts proved unsuccessful, Hoag brought his case here.

Viewing the question presented solely in terms of Fourteenth Amendment due process-whether the course that New Jersey had pursued had ‘led to fundamental unfairness,’ 356 U.S., at 467 , 78 S.Ct. at 832-this Court declined to reverse the judgment of conviction, because ‘in the circumstances shown by this record, we cannot say that [397 U.S. 436 , 442]   petitioner’s later prosecution and conviction violated due process.’6 356 U.S., at 466 . The Court found it unnecessary to decide whether ‘collateral estoppel’-the principle that bars relitigation between the same parties of issues actually determined at a previous trial-is a due process requirement in a state criminal trial, since it accepted New Jersey’s determination that the petitioner’s previous acquittal did not in any event give rise to such an estoppel. 356 U.S., at 471 . And in the view the Court took of the issues presented, it did not, of course, even approach consideration of whether collateral estoppel is an ingredient of the Fifth Amendment guarantee against double jeopardy.

The doctrine of Benton v. Maryland, 395 U.S. 784 , puts the issues in the present case in a perspective quite different from that in which the issues were perceived in Hoag v. New Jersey, Supra. The question is no longer whether collateral estoppel is a requirement of due process, but whether it is a part of the Fifth Amendment’s guarantee against double jeopardy. And if collateral estoppel is embodied in that guarantee, then its applicability in a particular case is no longer a matter to be left for state court determination within the broad [397 U.S. 436 , 443]   bounds of ‘fundamental fairness,’ but a matter of constitutional fact we must decide through an examination of the entire record. Cf. New York Times Co. v. Sullivan, 376 U.S. 254, 285 , 728-729d 686; Niemotko v. Maryland, 340 U.S. 268, 271 , 327; Watts v. Indiana, 338 U.S. 49, 51 , 1348; Chambers v. Florida, 309 U.S. 227, 229 , 473-474; Norris v. Alabama, 294 U.S. 587, 590 .

    • ‘Collateral estoppel’ is an awkward phrase, but it stands for an extremely important principle in our adversary system of justice. It means simply that when an issue of ultimate fact has once been determined by a valid and final judgment, that issue cannot again be litigated between the same parties in any future lawsuit. Although first developed in civil litigation, collateral estoppel has been an established rule of federal criminal law at least since this Court’s decision more than 50 years ago in United States v. Oppenheimer, 242 U.S. 85 . As Mr. Justice Holmes put the matter in that case, ‘It cannot be that the safeguards of the person, so often and so rightly mentioned with solemn reverence, are less than those that protect from a liability in debt.’ 242 U.S., at 87 , 37 S.Ct. at 69.7 As a rule of federal law, therefore, ‘(i)t is much too late to suggest that this principle is not fully applicable to a former judgment in a criminal case, either because of lack of ‘mutuality’ or because the judgment may reflect only a belief that the Government had not met the higher burden of proof exacted in such cases for the Government’s evidence as a whole although not necessarily as to every link in the chain.’ United States v. Kramer, 289 F.2d 909, 913.

[397 U.S. 436 , 444]   

    • The federal decisions have made clear that the rule of collateral estoppel in criminal cases is not to be applied with the hypertechnical and archaic approach of a 19th century pleading book, but with realism and rationality. Where a previous judgment of acquittal was based upon a general verdict, as is usually the case, this approach requires a court to ‘examine that record of a prior proceeding, taking into account the pleadings, evidence, charge, and other relevant matter, and conclude whether a rational jury could have grounded its verdict upon an issue other than that which the defendant seeks to foreclose from consideration.’ 8 The inquiry ‘must be set in a practical frame and viewed with an eye to all the circumstances of the proceedings.’ Sealfon v. United States, 332 U.S. 575, 579 , 240. Any test more technically restrictive would, of course, simply amount to a rejection of the rule of collateral estoppel in criminal proceedings, at least in every case where the first judgment was based upon a general verdict of acquittal.

    [397 U.S. 436 , 445]   

      Straightforward application of the federal rule to the present case can lead to but one conclusion. For the record is utterly devoid of any indication that the first jury could rationally have found that an armed robbery had not occurred, or that Knight had not been a victim of that robbery. The single rationally conceivable issue in dispute before the jury was whether the petitioner had been one of the robbers. And the jury by its verdict found that he had not. The federal rule of law, therefore, would make a second prosecution for the robbery of Roberts wholly impermissible.

    The ultimate question to be determined, then, in the light of Benton v. Maryland, supra, is whether this established rule of federal law is embodied in the Fifth Amendment guarantee against double jeopardy. We do not hesitate to hold that it is. 10 For whatever else that [397 U.S. 436 , 446]   constitutional guarantee may embrace, North Carolina v. Pearce, 395 U.S. 711, 717 , 2076, it surely protects a man who has been acquitted from having to ‘run the gantlet’ a second time. Green v. United States, 355 U.S. 184, 190 , 225.

    The question is not whether Missouri could validly charge the petitioner with six separate offenses for the robbery of the six poker players. It is not whether he could have received a total of six punishments if he had been convicted in a single trial of robbing the six victims. It is simply whether, after a jury determined by its verdict that the petitioner was not one of the robbers, the State could constitutionally hale him before a new jury to litigate that issue again.

    After the first jury had acquitted the petitioner of robbing Knight, Missouri could certainly not have brought him to trial again upon that charge. Once a jury had determined upon conflicting testimony that there was at least a reasonable doubt that the petitioner was one of the robbers, the State could not present the same or different identification evidence in a second prosecution for the robbery of Knight in the hope that a different jury might find that evidence more convincing. The situation is constitutionally no different here, even though the second trial related to another victim of the same robbery. For the name of the victim, in the circumstances of this case, had no bearing whatever upon the issue of whether the petitioner was one of the robbers. [397 U.S. 436 , 447]   In this case the State in its brief has frankly conceded that following the petitioner’s acquittal, it treated the first trial as no more than a dry run for the second prosecution: ‘No doubt the prosecutor felt the state had a provable case on the first charge and, when he lost, he did what every good attorney would do-he refined his presentation in light of the turn of events at the first trial.’ But this is precisely what the constitutional guarantee forbids.

    The judgment is reversed, and the case is remanded to the Court of Appeals for the Eighth Circuit for further proceedings consistent with this opinion.

    It is so ordered.

    Reversed and remanded.

    Mr. Justice BLACK, concurring.

    I join in the opinion of the Court although I must reject any implication in that opinion that the so-called due process test of ‘fundamental fairness’ might have been appropriate as a constitutional standard at some point in the past or might have a continuing relevancy today in some areas of constitutional law. In my view it is a wholly fallacious idea that a judge’s sense of what is fundamentally ‘fair’ or ‘unfair’ should ever serve as a substitute for the explicit, written provisions of our Bill of Rights. One of these provisions is the Fifth Amendment’s prohibition against putting a man twice in jeopardy. On several occasions I have stated my view that the Double Jeopardy Clause bars a State or the Federal Government or the two together from subjecting a defendant to the hazards of trial and possible conviction more than once for the same alleged offense. Bartkus v. Illinois, 359 U.S. 121, 150 , 79 S. Ct. 676, 695 (1959) (dissenting opinion); Abbate v. United States, 359 U.S. 187, 201 , 674, (1959) ( dissenting opinion); Cuicci v. Illinois, 356 U.S. [397 U.S. 436 , 448]   571, 575, 841 (1958) (dissenting statement); Green v. United States, 355 U.S. 184 (1957). The opinion of the Court in the case today amply demonstrates that the doctrine of collateral estoppel is a basic and essential part of the Constitution’s prohibition against double jeopardy. Accordingly, for the reasons stated in the Court’s opinion I fully agree that petitioner’s conviction must be reversed.

    Mr. Justice HARLAN, concurring.

    If I were to judge this case under the traditional standards of Fourteenth Amendment due process, I would adhere to the decision in Hoag v. New Jersey, 356 U.S. 464 (1958), believing that regardless of the reach of the federal rule of collateral estoppel, it would have been open to a state court to treat the issue differently. However, having acceded in North Carolina v. Pearce, 395 U.S. 711, 744 , 2085 (1969), to the decision in Benton v. Maryland, 395 U.S. 784 (1969), which, over my dissent, held that the Fourteenth Amendment imposes on the States the standards of the Double Jeopardy Clause of the Fifth Amendment. I am satisfied that on this present record Ashe’s acquittal in the first trial brought double jeopardy standards into play. Hence, I join the Court’s opinion. In doing so I wish to make explicit my understanding that the Court’s opinion in no way intimates that the Double Jeopardy Clause embraces to any degree the ‘same transaction’ concept reflected in the concurring opinion of my Brother BRENNAN.

    Mr. Justice BRENNAN, whom Mr. Justice DOUGLAS and Mr. Justice MARSHALL join, concurring.

    I agree that the Double Jeopardy Clause incorporates collateral estoppel as a constitutional requirement and therefore join the Court’s opinion. However, even if [397 U.S. 436 , 449]   the rule of collateral estoppel had been inapplicable to the facts of this case, it is my view that the Double Jeopardy Clause nevertheless bars the prosecution of petitioner a second time for armed robbery. The two prosecutions, the first for the robbery of Knight and the second for the robbery of Roberts, grew out of one criminal episode, and threfore I think it clear on the facts of this case that the Double Jeopardy Clause prohibited Missouri from prosecuting petitioner for each robbery at a different trial. Abbate v. United States, 359 U.S. 187 , 196-201, 671-674 (1959) (separate opinion).

    My conclusion is not precluded by the Court’s decision in Hoag v. New Jersey, 356 U.S. 464 (1958), although the basic fact situation there was identical to that in this case. Three armed men entered a tavern and robbed five customers. Hoag was tried and acquitted under indictments for robbing three of the customers. He was then brought to trial under a fourth indictment, the same as the first three in all respects except that it named a fourth customer as the victim. This time Hoag was convicted. The New Jersey courts, in rejecting Hoag’s double-jeopardy claim, construed the applicable New Jersey statute as making each of the four robberies, although taking place on the same occasion, a separate offense. This construction was consistent with the state courts’ view that a claim of double jeopardy cannot be upheld unless the same evidence necessary to sustain a second indictment would have been sufficient to secure a conviction on the first. The issues differed only in the identifications of the victims and the property taken from each; otherwise the State’s evidence covered the same ground at both trials. This Court stated that it was unable to hold that the Due Process [397 U.S. 436 , 450]   Clause of the Fourteenth Amendment ‘always prevents a State from allowing different offenses arising out of the same act or transaction to be prosecuted separately, as New Jersey has done. For it has long been recognized as the very essence of our federalism that the States should have the widest latitude in the administration of their own systems of criminal justice.’ 356 U.S., at 468 , 78 S.Ct. at 833. But in the present case Missouri did not have ‘the widest latitude,’ because Benton v. Maryland, 395 U.S. 784 (1969), decided after Hoag, held that the Fifth Amendment guarantee that no person shall ‘be subject for the same offense to be twice put in jeopardy of life or limb’ is enforceable against the States, and North Carolina v. Pearce, 395 U.S. 711 (1969), accorded fully retroactive effect to that holding. This means, under our decisions, that federal standards as to what constitutes then ‘same offence’ apply alike to federal and state proceedings; it would be incongruous to have different standards determine the validity of a claim of double jeopardy depending on whether the claim was asserted in a state or federal court. Cf. Malloy v. Hogan, 378 U.S. 1, 11 , 1495 (1964).

    The Double Jeopardy Clause is a guarantee ‘that the State with all its resources and power (shall) not be allowed to make repeated attempts to convict an individual for an alleged offense, thereby subjecting him to embarrassment, expense and ordeal and compelling him to live in a continuing state of anxiety and insecurity ….’ Green v. United States, 355 U.S. 184, 187 , 223 (1957). This guarantee is expressed as a prohibition against multiple prosecutions for the ‘same offence.’ Although the phrase ‘same offence’ appeared in most of the early common-law articulations of the double-jeop- [397 U.S. 436 , 451]   ardy principle,1 questions of its precise meaning rarely arose prior to the 18th century, and by the time the Bill of Rights was adopted it had not been authoritatively defined. 

    When the common law did finally attempt a definition, in The King v. Vandercomb, 2 Leach 708, 720, 168 Eng.Rep. 455, 461 (Crown 1796), it adopted the ‘same evidence’ test, which provided little protection from multiple prosecution:

      ‘(U)nless the first indictment were such as the prisoner might have been convicted upon by rpoof of the facts contained in the second indictment, an acquittal on the first indictment can be no bar to the second.’

    The ‘same evidence’ test of ‘same offence’ was soon followed by a majority of American jurisdictions, but its deficiencies are obvious. It does not enforce but virtually annuls the constitutional guarantee. For example, where a single criminal episode involves several victims, under the ‘same evidence’ test a separate prosecution may be brought as to each. E.g., State v. Hoag, 21 N.J. 496, 122 A.2d 628, (1956), aff’d, 356 U.S. 464 (1958). The ‘same evidence’ test permits multiple prosecutions where a single transaction is divisible into chronologically discrete crimes. E.g., Johnson v. Commonwealth, 201 Ky. 314, 256 S.W. 388 (1923) (each of 75 poker hands a separate ‘offense’). Even a single criminal act may lead to multiple prosecutions if it is viewed from the perspectives of different statutes. E.g., [397 U.S. 436 , 452]   State v. Elder, 65 Ind. 282 (1879). Given the tendency of modern criminal legislation to divide the phrases of a criminal transaction into numerous separate crimes, the opportunities for multiple prosecutions for an esentially unitary criminal episode are frightening. And given our tradition of virtually unreviewable prosecutorial discretion concerning the initiation and scope of a criminal prosecution,3 the potentialities for abuse inherent in the ‘same evidence’ test are simply intolerable. 

    The ‘same evidence’ test is not constitutionally required. It was first expounded after the adoption of the Fifth Amendment, and, as shown in Abbate v. United States, supra, 359 U.S. at 197-198 and n. 2, 671-673, has never been squarely held by this Court to be the required construc- [397 U.S. 436 , 453]   tion of the constitutional phrase ‘same offence’ in a case involving multiple trials; indeed, in that context it has been rejected. See In re Nielsen, 131 U.S. 176 (1889), discussed in Abbate v. United States, supra, 359 U.S. at 201. The ‘same evidence’ test may once have been defensible at English common law, which, for reasons peculiar to English criminal procedure, severely restricted the power of prosecutors to combine several charges in a single trial. In vivid contrast, American criminal procedure generally allows a prosecutor freedom, subject to judicial control, to prosecute a person at one trial for all the crimes arising out of a single criminal transaction. 

    In my view, the Double Jeopardy Clause requires the prosecution, except in most limited circumstances,7 to join at one trial all the charges against a defendant that grow out of a single criminal act, occurrence, [397 U.S. 436 , 454]   episode, or transaction. This ‘same transaction’ test of ‘same offence’ not only enforces the ancient prohibition against vexatious multiple prosecutions embodied in the Double Jeopardy Clause, but responds as well to the increasingly widespread recognition that the consolidation in one lawsuit of all issues arising out of a single transaction or occurrence best promotes justice, economy, and convenience. Modern rules of criminal and civil procedure reflect this recognition. See United Mine Workers v. Gibbs, 383 U.S. 715 , 724-726, 1137-1139 ( 1966). Although in 1935 the American Law Institute adopted the ‘same evidence’ test, it has since replaced it with the ‘same transaction’ test. 9 England, too, has abandoned its surviving rules against joinder of charges and has adopted the ‘same transaction’ test. 10 The Federal [397 U.S. 436 , 455]   Rules of Criminal Procedure liberally encourage the joining of parties and charges in a single trial. Rule 8(a) provides for joinder of charges that are similar in character, or arise from the same transaction or from connected transactions or form part of a common scheme or plan. Rule 8(b) provides for joinder of defendants. Rule 13 provides for joinder of separate indictments or informations in a single trial where the offenses alleged could have been included in one indictment or information. 11 These rules represent considered modern thought concerning the proper structuring of criminal litigation.

    The same thought is reflected in the Federal Rules of Civil Procedure. A pervasive purpose of those Rules is to require or encourage the consolidation of related claims in a single lawsuit. Rule 13 makes compulsory (upon pain of a bar) all counterclaims arising out of the same transaction or occurrence from which the plaintiff’s claim arose. Rule 14 extends this compulsion to third-party defendants. Rule 18 permits very broad joinder of claims, counterclaims, cross-claims, and third-party claims. Rules 19, 20, and 24 provide for joinder of parties and intervention by parties having claims [397 U.S. 436 , 456]   related to the subject matter of the action. Rule 23 permits the consolidation of separate claims in a class action; see particularly Rule 23(b)(3).

    In addition, principles of res judicata and collateral estoppel caution the civil plaintiff against splitting his case. The doctrine of pendent jurisdiction has furthered single trials of related cases. See United Mine Workers v. Gibbs, supra. Moreover, we have recognized the jurisdiction of three-judge courts to hear statutory claims pendent to the constitutional claim that required their convening. See, e.g., United States v. Georgia Pub. Serv. Comm., 371 U.S. 285 , 287-288, 398-399 (1963); King v. Smith, 392 U.S. 309 (1968).

    It is true that these developments have not been of a constitutional dimension, and that many of them are permissive and discretionary rather than mandatory. Flexibility in the rules governing the structure of civil litigation is appropriate in order to give the parties the opportunity to shape their own private lawsuits, provided that injustice, harassment, or an undue burden on the courts does not result. Some flexibility in the structuring of criminal litigation is also desirable and consistent with our traditions. But the Double Jeopardy Clause stands as a constitutional barrier against possible tyranny by the overzealous prosecutor. The considerations of justice, economy, and convenience that have propelled the movement for consolidation of civil cases apply with even greater force in the criminal context because of the constitutional principle that no man shall be vexed more than once by trial for the same offense. 12 Yet, if the Double Jeopardy Clause were in- [397 U.S. 436 , 457]   terpreted by this Court to incorporate the ‘same evidence’ test, criminal defendants would have less protection from multiple trials than civil defendants. This anomaly would be intolerable. It was condemned by a New Jersey court nearly a century and a half ago in words even more applicable today:

      ‘If in civil cases, the law abhors a multiplicity of suits, it is yet more watchful in criminal cases, that the crown shall not oppress the subject, or the government the citizen, by unnecessary prosecutions . … (This) is a case where the state has thought proper to prosecute the offence in its mildest form, and it is better that the residue of the offence go unpunished, than by sustaining a second indictment to sanction a practice which might be rendered an instrument of oppression to the citizen.’ State v. Cooper, 13 N.J.L. 361, 375-376 (1833).

    The present case highlights the hazards of abuse of the criminal process inherent in the ‘same evidence’ test and demonstrates the necessity for the ‘same transaction’ test. The robbery of the poker game involved six players-Gladson, Knight, Freeman, Goodwin, McClendon, and Roberts. The robbers also stole a car. Seven separate informations were filed against the petitioner, one covering each of the robbery victims, and the seventh covering the theft of the car. Petitioner’s first trial was under the information charging the robbery of Knight. Since Missouri has offered no justification for not trying the other informations at that trial, it is reasonable to infer that the other informations were held in reserve to be tried if the State failed to obtain a conviction on the charge of robbing Knight. Indeed, the State virtually concedes as much since it argues that the ‘same evidence’ test [397 U.S. 436 , 458]   is consistent with such an exercise of prosecutorial discretion.

    Four of the robbery victims testified at the trial. Their testimony conflicted as to whether there were three or four robbers. Gladson testified that he saw four robbers, but could identify only one, a man named Brown. McClendon testified that he saw only three men at any one time during the course of the robbery, and he positively identified Brown, Larson, and Johnson; he also thought he heard petitioner’s voice during the robbery, but said he was not sure. Knight thought only three men participated in the robbery, and he could not identify anyone. Roberts said he saw four different men and he identified them as Brown, Larson, Johnson, and petitioner. Under cross-examination, he conceded that he did not recognize petitioner’s voice, and that he did not see his face or his hands. He maintained that he could identify him by his ‘size and height’ even though all the robbers had worn outsized clothing, and even though he could not connect petitioner with the actions of any of the robbers. On this evidence the jury acquitted petitioner.

    At the second trial, for the robbery of Roberts, McClendon was not called as a witness. Gladson, who previously had been able to identify only one man-Brown-now was able to identify three-Brown, Larson, and petitioner. On a number of details his memory was much more vivid than it had been at the first trial. Knight’s testimony was substantially the same as at the first trial-he still was unable to identify any of the robbers. Roberts, who previously had identified petitioner only by his size and height, now identified him by his size, actions, voice, and a peculiar movement of his mouth. As might be expected, this far stronger [397 U.S. 436 , 459]   identification evidence brought a virtually inevitable conviction.

    The prosecution plainly organized its case for the second trial to provide the links missing in the chain of identification evidence that was offered at the first trial. McClendon, who was an unhelpful witness at the first trial was not called at the second trial. The hesitant and uncertain evidence of Gladson and Roberts at the first trial became detailed, positive, and expansive at the second trial. One must experience a sense of uneasiness with any double-jeopardy standard that would allow the State this second chance to plug up the holes in its case. The constitutional protection against double jeopardy is empty of meaning if the State may make ‘repeated attempts’ to touch up its case by forcing the accused to ‘run the gantlet’ at many times as there are victims of a single episode.

    Fortunately for petitioner, the conviction at the second trial can be reversed under the doctrine or collateral estoppel, since the jury at the first trial clearly resolved in his favor the only contested issue at that trial, which was the identification of him as one of the robbers. There is at least doubt whether collateral estoppel would have aided him had the jury been required toresolve additional contested issues on conflicting evidence. 13 But correction of the abuse of criminal process should not in any event be made to depend on the availability of collateral estoppel. Abuse of the criminal process is foremost among the feared evils that led to the inclusion of the Double Jeopardy Clause in the Bill of Rights. That evil will be most effectively avoided, and the Clause can thus best serve its worthy ends, if ‘same [397 U.S. 436 , 460]   offence’ is construed to embody the ‘same transaction’ standard. Then both federal and state prosecutors will be prohibited from mounting successive prosecutions for offenses growing out of the same criminal episode, at least in the absence of a showing of unavoidable necessity for successive prosecutions in the particular case. 14 

    Mr. Chief Justice BURGER, dissenting.

    The Fifth Amendment to the Constitution of the United States provides in part: ‘nor shall any person be subject for the same offence to be twice put in jeopardy of life or limb ….’ Nothing in the language or gloss previously placed on this provision of the Fifth Amendment remotely justifies the treatment that the Court today accords to the collateral- estoppel doctrine. Nothing in the purpose of the authors of the Constitution commands or even justifies what the Court decides today; this is truly a case of expanding a sound basic [397 U.S. 436 , 461]   principle beyond the bounds-or needs-of of its rational and legitimate objectives to preclude harassment of an accused.


    Certain facts are not in dispute. The home of John Gladson was the scene of ‘a friendly game of poker’ in the early hours of the morning of January 10, 1960. Six men-Gladson Knight, Freeman, Goodwin, McClendon, and Roberts-were playing cards in the basement. While the game was in progress, three men, armed with a sawed-off shotgun and pistols, broke into the house and forced their way into the basement. They ordered the players to remove their trousers and tied them up, except for Gladson who had a heart condition of which the robbers seemed to be aware. Substantial amounts of currency and checks were taken from the poker table and items of personal property were taken from the persons of the players. During the same period in which the men were being robbed in the basement, one man entered Mrs. Gladson’s bedroom three floors above, ripped out the telephone there, tied her with the telephone cord, and removed the wedding ring from her finger. The robbers then fled in a car belonging to Roberts.

    Four men-Ashe, Johnson, Larson, and Brown-were arrested later in the morning of the robbery. Each was subsequently charged in a separate information with the robbery of each of the six victims. Ashe, Johnson, and Larson were also charged with the theft of the car belonging to Roberts.

    Ashe went to trial on May 2, 1960, on the charge of robbing Knight. No charge as to other victims was presented. Four of the six men-Knight, Gladson, McClendon, and Roberts-testified about the robbery and described their individual losses. Mrs. Gladson did not [397 U.S. 436 , 462]   testify because she was ill on the day of trial. As Mr. Justice BRENNAN has stated, the victims’ testimony conflicted as to whether there were three or four robbers:

      ‘Gladson testified that he saw four robbers, but could identify only one, a man named Brown. McClendon testified that he saw only three men at any one time during the course of the robbery, and he positively identified Brown, Larson, and Johnson; he also thought he heard petitioner’s voice during the robbery, but said he was not sure. Knight thought only three men participated in the robbery, and he could not identify any one. Roberts said he saw four different men and he identified them as Brown, Larson, Johnson, and petitioner.’ Ante, at 458.

    Ashe put in no evidence whatever, as was his right, and even waived closing arguments to the jury; nonetheless, the jury did not reach a verdict of guilty but returned a somewhat unorthodox verdict of ‘not guilty due to insufficient evidence.’

    Then, on June 20, 1960, Ashe was tried for the robbery of Roberts. Mrs. Gladson testified at this trial, relating that she was asleep in her bedroom when one the robbers entered, awoke her, tied her up with a telephone cord, and took cash and her wedding ring. The robber stayed in her room for about 15 or 20 minutes, during which time she could hear scuffling and talking in the basement. She said that she was able to identify the robber by his voice, and that he was Johnson, not Ashe.

    The Court’s opinion omits some relevant facts. The other victims’ testimony at the second trial corroborated that of Mrs. Gladson that four robbers were present during the time in which the robbery took place. [397 U.S. 436 , 463]   Gladson identified three robbers-Brown, Larson, and Ashe-as having been in the basement for the first minutes of the robbery; also he stated that one or more of the robbers had left the basement after 20 or 25 minutes. Roberts identified Brown, Larson, and Ashe as the men who formed the original group who entered the basement and testified that after the robbery, two of the three men, including Ashe, left the room. Two men returned in a short time with car keys, but Johnson had replaced Ashe as one of the two. There can be no doubt that the record shows four persons in the robbery band. The jury found Ashe guilty of robbing Roberts-the only charge before it.

    Thereafter, as described in the opinion of the majority, Ashe’s conviction was reviewed and upheld by the Supreme Court of Missouri, the United States District Court for the Western District of Missouri, and the Court of Appeals for the Eighth Circuit; in turn each rejected Ashe’s double-jeopardy claim.


    The concept of double jeopardy and our firm constitutional commitment is against repeated trials ‘for the same offence.’ This Court, like most American jurisdictions, has expanded that part of the Constitution into a ‘same evidence’ test. For example, in Blockburger v. United States, 284 U.S. 299, 304 , 182 (1932), it was stated, so far as here relevant, that

      • ‘the test to be applied to determine whether there are two offenses or only one is whether each provision (i.e., each charge) requires proof of a fact which the other does not.’ (Emphasis added).

    [397 U.S. 436 , 464]   

      Clearly and beyond dispute the charge against Ashe in the second trial required proof of a fact-robbery of Roberts-which the charge involving Knight did not. The Court, therefore, has had to reach out far beyond the accepted offense-defining rule to reach its decision in this case. What it has done is to superimpose on the same-evidence test a new and novel collateral-estoppel gloss.

    The majority rests its holding in part on a series of cases beginning with United States v. Oppenheimer, 242 U.S. 85 ( 1916), which did not involve constitutional double jeopardy but applied collateral estoppel as developed in civil litigation to federal criminal prosecutions as a matter of this Court’s supervisory power over the federal court system. The Court now finds the federal collateral estoppel rule to be an ‘ingredient’ of the Fifth Amendment guarantee against double jeopardy and applies it to the States through the Fourteenth Amendment. This is an ingredient that eluded judges and justices for nearly two centuries; the Court finds itself endowed with a perception all our predecessors lacked and in an area of the law which encounters no new facts.

    The collateral-estoppel concept-originally a product only of civil litigation-is a strange mutant as it is transformed to control this criminal case. In civil cases the doctrine was justified as conserving judicial resources as well as those of the parties to the actions and additionally as providing the finality needed to plan for the future. It ordinarily applies to parties on each side of the litigation who have the same interest as or who are identical with the parties in the initial litigation. Here the complainant in the second trial is not the same as in the first even though the State is a party in both cases. Very properly, in criminal cases, finality and conservation of private, public and judicial resources are lesser values than in civil litigation. Also, courts that have applied the collateral-estoppel concept to criminal actions would [397 U.S. 436 , 465]   certainly not apply it to both parties, as is true in civil cases, i.e., here, if Ashe had been convicted at the first trial, presumably no court would then hold that he was thereby by foreclosed from litigating the identification issue at the second trial. 

    Perhaps, then, it comes as no surprise to find that the only expressed rationale for the majority’s decision is that Ashe has ‘run the gantlet’ once before. This is not a doctrine of the law or legal reasoning but a colorful and graphic phrase, which, as used originally in an opinion of the Court written by Mr. Justice Black, was intended to mean something entirely different. The full phrase is ‘run the gantlet once on that charge ….’ (emphasis added); it is to be found in Green v. United States, 355 U.S. 184, 190 , 225 (1957), where no question of multiple crimes against multiple victims was involved. Green, having been found guilty of second degree murder on a charge of first degree, secured a new trial. This Court held nothing more than that Green, once put in jeopardy-once having ‘run the gantlet … on that charge’-of first degree murder, could not be compelled to defend against that charge again on retrial.

    Today’s step in this area of constitutional law ought not be taken on no more basis than casual reliance on the ‘gantlet’ phrase lifted out of the context in which it was originally used. This is decision by slogan.

    Some commentators have concluded that the harassment inherent in standing trial a second time is a sufficient reason for use of collateral estoppel in criminal [397 U.S. 436 , 466]   trials. If the Court is today relying on a harassment concept to superimpose a new brand of collateral-estoppel gloss on the ‘same evidence’ test, there is a short answer; this case does not remotely suggest harassment of an accused who robbed six victims and the harassment aspect does not rise to constitutional levels. 

    Finally, the majority’s opinion tells us ‘that the rule of collateral estoppel in criminal cases is not to be applied with the hypertechnical and archaic approach of a 19th century pleading book, but with realism and rationality.’ Ante, at 444.

    With deference I am bound to pose the question: what is reasonable and rational about holding that an acquittal of Ashe for robbing Knight bars a trial for robbing Roberts? To borrow a phrase from the Court’s opinion, what could conceivably be more ‘hypertechnical and archaic’ and more like the stilted formalisms of 17th and 18th century common-law England, than to stretch jeopardy for robbing Knight into jeopardy for robbing Roberts?

    After examining the facts of this case the Court concludes that the first jury must have concluded that Ashe was not one of the robbers-that he was not present at [397 U.S. 436 , 467]   the time. Also, since the second jury necessarily reached its decision by finding he was present, the collateral-estoppel doctrine applies. But the majority’s analysis of the facts completely disregards the confusion injected into the case by the robbery of Mrs. Gladson. To me, if we are to psychoanalyze the jury, the evidence adduced at the first trial could more reasonably be construed as indicating that Ashe had been at the Gladson home with the other three men but was not one of those involved in the basement robbery. Certainly, the evidence at the first trial was equivocal as to whether there were three or four robbers, whether the man who robbed Mrs. Gladson was one of the three who robbed the six male victims, and whether a man other than the three had robbed Mrs. Gladson. Then, since the jury could have thought that the ‘acting together’ instruction given by the trial court in both trials6 only applied to the actual taking from the six card players, and not to Mrs. Gladson, the jury could well have acquitted Ashe but yet believed that he was present in the Gladson home. On the other hand, the evidence adduced at the second trial resolved issues other than identity that may have troubled the first jury. If believed, that evidence indicated that a fourth robber, Johnson, not Ashe, was with Mrs. Gladson when Ashe, Larson, and Brown were robbing the male victims. Johnson did go to the basement where the male victims were located, but only after the other three had already taken the stolen items and when the robbers were preparing for their departure in a car to be stolen from Roberts. [397 U.S. 436 , 468]   Accordingly, even the facts in this case, which the Court’s opinion considers to ‘lead to but one conclusion,’ are susceptible of an interpretation that the first jury did not base its acquittal on the identity ground which the Court finds so compelling. The Court bases its holding on sheer ‘guesswork,’7 which should have no place particularly in our review of state convictions by way of habeas corpus. As Mr. Justice Holmes said in Guy v. Donald, 206 U.S. 399, 406 , 64 (1906):

      ‘As long as the matter to be considered is debated in artificial terms there is danger of being led by a technical definition to apply a certain name, and then to deduce consequences which have no relation to the grounds on which the name was applied. …’


    The essence of Mr. Justice BRENNAN’s concurrence is that this was all one transaction, one episode, or, if I may so characterize it, one frolic, and, hence, only one crime. His approach, like that taken by the Court, totally overlooks the significance of there being six entirely separate charges of robbery against six individuals.

    This ‘single frolic’ concept is not a novel notion; it has been urged in various courts including this Court. One of the theses underlying the ‘single frolic’ notion is that the criminal episode is ‘indivisible.’ The short answer to that is that to the victims, the criminal conduct is readily divisible and intensely personal; each offense is an offense against a person. For me it de- [397 U.S. 436 , 469]   means the dignity of the human personality and individuality to talk of ‘a single transaction’ in the context or six separate assaults on six individuals.

    No court that elevates the individual rights and human dignity of the accused to a high place-as we should-ought to be so casual as to treat the victims as a single homogenized lump of human clay. I would grant the dignity of individual status to the victims as much as to those accused, not more but surely no less.

    If it be suggested that multiple crimes can be separately punished but must be collectively tried, one can point to the firm trend in the law to allow severance of defendants and offenses into separate trials so as to avoid possible prjudice of one one criminal act or of the conduct of one defendant to ‘spill over’ on another.

    What the Court holds today must be related to its impact on crimes more serious than ordinary housebreaking, followed by physical assault on six men and robbery of all of them. To understand its full impact we must view the holding in the context of four men who break and enter, rob, and then kill six victims. The concurrence tells us that unless all the crimes are joined in one trial the alleged killers cannot be tried for more than one of the killings even if the evidence is that they personally killed two, three, or more of the victims. Or alter the crime to four men breaking into a college dormitory and assaulting six girls. What the Court is holding is, in effect, that the second and third and fourth criminal acts are ‘free,’ unless the accused is tried for the multiple crimes in a single trial-something defendants frantically use every legal device to avoid, and often succeed in avoiding. This is the reality of what the Court holds today; it does not make good sense and it cannot make good law. [397 U.S. 436 , 470]   I therefore join with the four courts that have found no double jeopardy in this case.

    To borrow some wise words from Mr. Justice Black in his separate opinion in Jackson v. Denno, 378 U.S. 368 , 401, 407-408, 1797 (1964), the conviction struck down in this case ‘is in full accord with all the guarantees of the Federal Constitution and … should not be held invalid by this Court because of a belief that the Court can improve on the Constitution.’