'When the trial starts, the accused feels the full impact of the conspiracy strategy. Strictly, the prosecution should first establish prima facie the conspiracy and identify the conspirators, after which evidence of acts and declarations of each in the course of its execution are admissible against all. But the order of proof of so sprawling a charge is difficult for a judge to control. As a practical matter, the accused often is confronted with a hodgepodge of acts and statements by others which he may never have authorized or intended or even known about, but which help to persuade the jury of existence of the conspiracy itself. In other words, a conspiracy often is proved by evidence that is admissible only upon assumption that conspiracy existed. The naive assumption that prejudicial effects can be overcome by instructions to the jury, cf. Blumenthal v. United States, 332 U.S. 539, 559, 68 S.Ct. 248, 257, 92 L.Ed. 154, all practicing lawyers know to be unmitigated fiction. See Skidmore v. Baltimore & Ohio R. Co., 2 Cir., 167 F.2d 54.' Id., at 453, 69 S.Ct., at 723.

There are other disabilities. Often testimony will be receivable only against a particular codefendant yet it may also inculpate another accused such as where (a) a codefendant 'opens the door' to prejudicial evidence by placing his reputation in issue,3 (b) a codefendant wants to place before the jury information which is helpful to him but is damaging to other defendants, or (c) the Government desires to offer evidence admissible against less than all of the codefendants. Cautionary instructions, of course, are routinely given where such circumstances arise but we have often recognized the inability of jurors to compartmentalize information according to defendants. Bruton v. United States, 391 U.S. 123, 88 S.Ct. 1620, 20 L.Ed.2d 476 (1968). See also Jackson v. Denno, 378 U.S. 368, 388, 84 S.Ct. 1774, 1786, 12 L.Ed.2d 908 (1964); Krulewitch v. United States, 336 U.S. 440, 453, 69 S.Ct. 716, 723, 93 L.Ed. 790 (1949) (quoted above). This short-coming of the jury is compounded when, as here, the jury is also asked to digest voluminous testimony.

A victim of the multi-defendant conspiracy trial has fewer options for trial strategy than the ordinary defendant tried alone. Counsel may reluctantly give up the option of pointing the accusing finger at his client's codefendants in order to obtain similar concessions from other trial counsel. Counsel must also divert his preparation in part toward generating possible responses to evidence which may be admissible only against other codefendants. As for the defendant, he may be put to the choice of hiring less experienced counsel or less actively pursuing discovery or investigation because of the higher legal expenses imposed by longer joint trials. Furthermore, although an accused normally has 'the right to present his own witnesses to establish a defense,' Washington v. Texas, 388 U.S. 14, 19, 87 S.Ct. 1920, 1923, 18 L.Ed.2d 1019 (1967), an accused in a mass conspiracy trial may not put on his codefendants without their prior waivers of their absolute rights not to testify.4

All of these oppressive features were present to various degrees in this trial. But, in particular, the most onerous burden cast upon these petitioners was their inability to cross-examine each other as to comments which Government witnesses said they had heard them utter. The Court of Appeals recognized that 'There was much testimony as to statements made by various co-conspirators during the course, and in furtherance, of the conspiracy.' 451 F.2d 71. For example, one important prosecution witness testified that he had been a contractor hired by the city administration and that one of the accused conspirators, 'Tony Boy' Boiardo, had told him 'You pay me the ten percent . . . I take care of the Mayor. I take care of the Council.' (A. 2611). The lawyer for the former mayor, however, was not permitted to put Boiardo on the stand and to ask him whether Addonizio had, in fact, entered into an agreement with him to coerce kickbacks. This handicap of an accused is at war with the holdings of this Court that a defendant should be permitted to confront his accusers especially where, as here, their declarations might have been purposefully misleading or self-serving. Pointer v. Texas, 380 U.S. 400, 407, 85 S.Ct. 1065, 1069, 13 L.Ed.2d 923 (1965); Douglas v. Alabama, 380 U.S. 415, 85 S.Ct. 1074, 13 L.Ed.2d 934 (1965); Brookhart v. Janis, 384 U.S. 1, 86 S.Ct. 1245, 16 L.Ed.2d 314 (1966); Bruton v. United States, supra; Barber v. Page, 390 U.S. 719, 88 S.Ct. 1318, 20 L.Ed.2d 255 (1968); Roberts v. Russell, 392 U.S. 293, 88 S.Ct. 1921, 20 L.Ed.2d 1100 (1968). Dutton v. Evans, 400 U.S. 74, 91 S.Ct. 210, 27 L.Ed.2d 213 (1970), is not inconsistent with this proposition. There the Court found that the hearsay was probably reliable. '[T]he circumstances under which [the declarant] made the statement were such as to give reason to suppose that [he] did not misrepresent [his coconspirator's] involvement in the crime.' Id., at 89, 91 S.Ct., at 219. On the other hadn, involved here were declarants, as mentioned earlier, who might have been motivated to misrepresent the roles of other parties in order to induce contractors, such as Rigo (the Government's key witness), to make kickbacks. Moreover, in Dutton the hearsay was 'of peripheral significance at most' whereas here much of the case against the petitioners, as the Court of Appeals pointed out, was admitted under the coconspirator exception to the hearsay rule.5

In addition, the petitioners were deprived of the right to cross-examine codefendant Gordon (who is not one of the petitioners). He had testified at the prior grand jury proceeding and that testimony was introduced at trial by the Government to corroborate the story of the Government's key witness, Rigo, as to various kickback transactions. The circumstances at trial were substantially similar to those involved in Bruton except that Gordon's grand jury remarks did not directly mention his codefendants. Normally, that difference would be sufficient to support the lower court's finding that Bruton was inapposite but for the fact that Government's case against all of the defendants turned upon Rigo's credibility. On cross-examination of Rigo, the codefendants had relentlessly attacked his credibility. But when the Government introduced the grand jury transcript in rebuttal, the defense challenge was completely terminated because Gordon, who was also on trial, could not be called to the stand. The judge, of course, gave instructions to the jury to consider the impact of the transcript upon Rigo's credibility only when assessing Gordon's guilt but it is doubtful that the jurors could faithfully adhere to the delicate logic that Rigo may have told the truth as to Gordon but