This is an appeal from criminal convictions following one of the lengthiest and costliest trials in this nation’s history. The trial lasted over 16 months, produced over 30,000 pages of transcripts, and involved over 250 witnesses and thousands of exhibits presenting evidence involving over 2,000 narcotics transactions spanning an 11-year period. Of the 24 defendants charged in the 44-count superseding indictment, 15 initially went to trial. Three defendants reached plea agreements during the trial. Eleven of the remaining 12 defendants join in this appeal. We are called upon to consider not only the approximately 50 individual issues raised on appeal, but the practical and human limitations of our jury system itself.
Richard Rupley, Sr., John Bonnenfant, Dwain Baker, Edward Baker, Daniel Rupley, Dominic Cavallaro, Katherine Rupley, Richard Rupley, Jr., Byron Wimberly, Robert Rowen, and Robert Cole (collectively, “Appellants”) were members of a large criminal organization known as “the Company.”1 The Company was headed by Rupley, Sr., who expanded his operations by recruiting many teenagers and young adults (including his 15-year-old son) into the organization. The central count of the superseding indictment charged all defendants with conspiracy to manufacture, distribute, and possess with intent to distribute methamphetamine in California and Nevada between December 1981 and September 1987. Seven defendants were charged in a separate marijuana conspiracy, and four defendants were charged with conducting a continuing criminal enterprise. The remainder of the counts involved various combinations of defendants and primarily charged specific narcotics violations and interstate transportation in aid of racketeering.
The Company’s drug-related activities, as the district court found, “involved an extraordinary level of violence.” The indictment charged, and the district court found by a preponderance of the evidence, that Company members furthered their criminal conspiracies by attempting to kill a United States Forest Service employee, shooting at a low-flying police helicopter, and planning the murders of state and federal narcotics agents and government witnesses. The district court further found that the Company dealt with perceived acts of disloyalty by murdering one member (Rosie Osick), attempting to murder another (Dale Richmond), and beating and forcibly raping a third (Crystal Channell).
Additional pertinent facts will be stated in the discussions of relevant issues.
I. Joinder
A
Perhaps the central issue of this case is whether the district court should have granted Appellants’ motions to sever this massive trial into several more manageable, less prejudicial proceedings. The district court’s denial of a motion to sever is reviewed for an abuse of discretion. United States v. Cuozzo, 962 F.2d 945, 949 (9th Cir.), *1387cert. denied, — U.S. -, 113 S.Ct. 475, 121 L.Ed.2d 381 (1992). “The test for abuse of discretion by the district court is ‘whether a joint trial was so manifestly prejudicial as to require the trial judge to exercise his discretion in but one way, by ordering a separate trial.’” United States v. Patterson, 819 F.2d 1495, 1501 (9th Cir.1987) (quoting United States v. Abushi, 682 F.2d 1289, 1296 (9th Cir.1982)). This scope of review is “extremely narrow.” United States v. Mariscal, 939 F.2d 884, 886 (9th Cir.1991); see United States v. Stirling, 571 F.2d 708, 733 (2d Cir.) (severance question is “virtually unreviewable”), cert. denied, 439 U.S. 824, 99 S.Ct. 93, 58 L.Ed.2d 116 (1978); United States v. Campanale, 518 F.2d 352, 359 (9th Cir.1975) (trial court’s severance rulings “will rarely be disturbed on review”), cert. denied, 423 U.S. 1050, 96 S.Ct. 777, 46 L.Ed.2d 638 (1976).
Fed.R.Crim.P. 8(b) provides:
Two or more defendants may be charged in the same indictment or information if they are alleged to have participated in the same act or transaction or in the same series of acts or transactions constituting an offense or offenses. Such defendants may be charged in one or more counts together or separately and all of the defendants need not be charged in each count.
Because “joint trials ‘conserve state funds, diminish inconvenience to witnesses and public authorities, and avoid delays in bringing those accused of crime to trial,’” United States v. Lane, 474 U.S. 438, 449, 106 S.Ct. 725, 732, 88 L.Ed.2d 814 (1986) (quoting Bruton v. United States, 391 U.S. 123, 134, 88 S.Ct. 1620, 1627, 20 L.Ed.2d 476 (1968)), Rule 8(b) is construed liberally in favor of joinder. United States v. Sanchez-Lopez, 879 F.2d 541, 551 (9th Cir.1989); United States v. Portac, Inc., 869 F.2d 1288, 1294 (9th Cir.1989), cert. denied, 498 U.S. 845, 111 S.Ct. 129, 112 L.Ed.2d 97 (1990). “ ‘Co-defendants jointly charged are, prima facie, to be jointly tried.’ ” Mariscal, 939 F.2d at 885 (quoting United States v. Doe, 655 F.2d 920, 926 (9th Cir.1980)); see also Zafiro v. United States, — U.S. -, -, 113 S.Ct. 933, 937, 122 L.Ed.2d 317 (1993) (noting preference in federal system for joint trials of defendants who are indicted together); United States v. Escalante, 637 F.2d 1197, 1201 (9th Cir.) (defendants jointly charged in conspiracy eases are presumptively to be jointly tried), cert. denied, 449 U.S. 856, 101 S.Ct. 154, 66 L.Ed.2d 71 (1980).
Fed.R.Crim.P. 14 limits the presumption of Rule 8(b) where otherwise proper joinder may prejudice a defendant:
If it appears that a defendant ... is prejudiced by a joinder of offenses or of defendants in an indictment or information or by such joinder for trial together, the court may order an election or separate trials of counts, grant a severance of defendants or provide whatever other relief justice requires.
Rules 8(b) and 14 “ ‘are designed to promote economy and efficiency and to avoid a multiplicity of trials, where these objectives can be achieved without substantial prejudice to the right of the defendants to a fair trial.’ ” Bruton, 391 U.S. at 131 n. 6, 88 S.Ct. at 1625 n. 6 (quoting Daley v. United States, 231 F.2d 123, 125 (1st Cir.), cert. denied 351 U.S. 964, 76 S.Ct. 1028, 100 L.Ed. 1484 (1956)).
Recognizing that “some prejudice is inherent in any joinder of defendants,” United States v. Vaccaro, 816 F.2d 443, 448-49 (9th Cir.), cert. denied 484 U.S. 914, 108 S.Ct. 262, 98 L.Ed.2d 220, 484 U.S. 928, 108 S.Ct. 295, 98 L.Ed.2d 255 (1987), we previously have focused our inquiry into the prejudicial effect of a joint trial on whether the jury may reasonably be expected to collate and appraise the independent evidence against each defendant. United States v. Sherlock, 962 F.2d 1349, 1360 (9th Cir.1989), cert. denied, — U.S. -, 113 S.Ct. 419, 121 L.Ed.2d 342 (1992). Because limiting instructions may suffice to cure a risk of prejudice, Zafiro, — U.S. at -, 113 S.Ct. at 938, the judge’s diligence in instructing the jury on the limited purposes for which various evidence may be used is a “critical factor” in assessing the jury’s ability to compartmentalize the evidence against each defendant. Cuozzo, 962 F.2d at 950. We have also recognized that “[t]he best evidence of the jury’s ability to compartmentalize the evidence is its failure to convict all defendants on all counts.” United States v. Unruh, 855 F.2d 1363, 1374 (9th Cir.1987), cert. *1388denied, 488 U.S. 974, 109 S.Ct. 513, 102 L.Ed.2d 548 (1988). Both of these factors support the district court’s denials of severance in this case.
Appellants do not contend that the district court was careless or lackadaisical in instructing the jury; they argue that there were too many limiting instructions, not too few. The judge gave nearly 200 limiting instructions over the course of the trial, many of which were repeated several times. For example, when a witness testified over several days, the court would routinely repeat any applicable limiting instructions at the start of each day’s testimony. The court also instructed the jury at the close of the case to give separate consideration to each charge and each defendant. “[Ojur court assumes that the jury listened to and followed the trial judge’s instructions.” Escalante, 637 F.2d at 1202 (citations omitted); see Richardson v. Marsh, 481 U.S. 200, 206, 107 S.Ct. 1702, 1706, 95 L.Ed.2d 176 (1987) (recognizing “the almost invariable assumption of the law that jurors follow their instructions”). The district court’s careful and frequent limiting instructions militate against finding an abuse of discretion. See United States v. Kennedy, 564 F.2d 1329, 1334 (9th Cir.1977), cert. denied, 435 U.S. 944, 98 S.Ct. 1526, 55 L.Ed.2d 541 (1978); Cuozzo, 962 F.2d at 950.
The jury’s selective verdicts also support the district court’s decision. The jury acquitted Robert Rowen on count 3, Dwain Baker on count 11, and Daniel Rupley on count 15. The court declared mistrials when the jury was unable to reach verdicts regarding Dwain and Edward Baker on count 3 and Daniel Rupley on count 40. Other incidents demonstrate that the jury gave conscientious attention to each count as it applied to each defendant. For example, the jury noticed that Daniel Rupley was named in the pleading portion of count 10 but not in the caption, and this flaw led to the dismissal of that count against Daniel Rupley. These factors “dramatically demonstrate[ ] that the jury was able, under the careful instructions of the Court, to understand and separate the evidence as to each defendant and to individually determine the issues presented.” Kennedy, 564 F.2d at 1334-35; see Cuozzo, 962 F.2d at 950.
Moreover, we note that, “[although the jury had to evaluate a tremendous amount of evidence, the nature of the evidence and the legal concepts involved in the case were not extraordinarily difficult to comprehend, as they might be, for example, in a complex anti-trust case involving abstruse economic theories or an employment discrimination case involving technical statistical evidence and formulae.” United States v. Casamento, 887 F.2d 1141, 1150 (2d Cir.1989), cert. denied, 493 U.S. 1081, 110 S.Ct. 1138, 107 L.Ed.2d 1043; 495 U.S. 933, 110 S.Ct. 2175, 109 L.Ed.2d 504; 495 U.S. 958, 110 S.Ct. 2564, 109 L.Ed.2d 746 (1990). Drug manufacturing and distribution, even on such a large scale as in this case, is not beyond the competence of the ordinary juror. See United States v. Moten, 564 F.2d 620, 627 (2d Cir.), cert. denied, 434 U.S. 942, 98 S.Ct. 438, 54 L.Ed.2d 304; 434 U.S. 959, 98 S.Ct. 489, 54 L.Ed.2d 318; 434 U.S. 974, 98 S.Ct. 531, 54 L.Ed.2d 466 (1977). “The crimes here may have been large in number and variety, but they were rather ordinary in nature, except in their viciousness.” United States v. DiNome, 954 F.2d 839, 842 (2d Cir.), cert. denied, — U.S. -, 113 S.Ct. 94, 121 L.Ed.2d 56 (1992).
Finally, Appellants’ allegations of prejudice are substantively insufficient to require reversal. It is not enough for Appellants to show that separate trials would have created a better chance for acquittal. Zafiro, — U.S. at -, 113 S.Ct. at 938; Escalante, 637 F.2d at 1201. Nor is the fact that a defendant is to be tried with a more culpable defendant enough to require severance. United States v. Van Cauwenberghe, 827 F.2d 424, 432 (9th Cir.1987), cert. denied, 484 U.S. 1042, 108 S.Ct. 773, 98 L.Ed.2d 859 (1988). But see Zafiro, — U.S. at -, 113 S.Ct. at 938 (“When many defendants are tried together in a complex case and they have markedly different degrees of culpability, [the] risk of prejudice is heightened.”). Rather, “when defendants properly have been joined under Rule 8(b), a district court should grant a severance under Rule 14 only if there is a serious risk that a joint trial *1389would compromise a specific trial right of one of the defendants, or prevent the jury from making a reliable judgment about guilt or innocence.” Zafiro, — U.S. at -, 113 S.Ct. at 938. Appellants do not demonstrate with sufficient particularity how they were injured by the joint trial. For example, they do not allege any inconsistent defenses, violation of confrontation rights, or unavailability of eodefendants’ exculpatory testimony. Broad and general allegations of prejudice from the length of the trial are not enough to require the district court to grant a severance.
The primary case on which Appellants rely, United States v. Donaway, 447 F.2d 940 (9th Cir.1971), is readily distinguishable. With less than 50 of the 2300 transcript pages in the government’s case relevant to Donaway, we held in that case that the district court abused its discretion in failing to sever Donaway from the eight other defendants in a 10-count interstate gambling conspiracy trial. Id. at 943. In Dona-way, however, the conspiracy count had been dismissed as to all but two defendants. Id. at 941. In this case, all Appellants were charged in the methamphetamine conspiracy. Although some Appellants may not have been mentioned frequently by name during the trial, the voluminous evidence of this conspiracy and of the overt acts in furtherance of it was relevant to all defendants, and thus would have been admissible even in separate trials.
In Richardson v. Marsh, the Supreme Court observed:
Many joint trials — for example, those involving large conspiracies to import and distribute illegal drugs — involve a dozen or more codefendants.... It would impair both the efficiency and the fairness of the criminal justice system to require ... that prosecutors bring separate proceedings, presenting the same evidence again and again, requiring victims and witnesses to repeat the inconvenience (and sometimes trauma) of testifying, and randomly favoring the last-tried defendants who have the advantage of knowing the prosecution’s case beforehand. Joint trials generally serve the interests of justice by avoiding inconsistent verdicts and enabling more accurate assessment of relative culpability— advantages which sometimes operate to the defendant’s benefit.
481 U.S. at 209-10, 107 S.Ct. at 1708-09. We hold that the district court did not abuse its broad discretion in denying Appellants’ motions for severance.
B
Our holding should not be interpreted as an endorsement of the government’s bringing such mega-trials in the future. We commend Chief Judge Reed for his remarkably careful rulings throughout the district court proceedings. The judge gave great weight to the defendants’ rights and made a determined effort to ensure that the trial was fair. In the hands of a less experienced judge, however, this trial could well have resulted in a complete reversal and a colossally expensive waste of time. We therefore find it appropriate to set forth our concerns regarding trials of this magnitude and some standards by which to measure the viability of such trials in the future.
1
As discussed above, there are several frequently cited advantages of proceeding against all defendants in a single trial. We believe, however, that most of these purported advantages are overstated when a trial of this nature is involved.
First, the claim that joint trials save time and serve judicial economy is ludicrous under the present facts. Where trials of this magnitude are involved, judicial economy will often be better served by severance. As the government proceeds through separate trials, it learns the strengths of its case and makes a sharper and more streamlined presentation so that “[ejach successive trial moves at a quicker and smoother pace than the last.” United States v. Gallo, 668 F.Supp. 736, 757 (E.D.N.Y.1987). With fewer defendants and defense counsel involved, there is less need for the sidebars and continuances that contribute to the length of a joint trial. Moreover, “[t]he court itself becomes much more familiar with the nature of the *1390case and the evidence, thus enabling more expeditious and more efficient rulings.” Id. We do not believe that this case would have required 16 months in the courtroom had the defendants been tried in manageable groups of three or four.
Second, the government’s asserted concern with disclosing and weakening its case against later-tried defendants is unpersuasive. We can see no problems beyond those inherent in retrying a case reversed on appeal, a situation in which the government has proved fully capable of securing convictions. See Edward Weinfeld, The Problems of Long Criminal Trials, 34 F.R.D. 158, 161 (1963). Disclosure of the government’s method and quality of proof may even benefit the prosecution by inducing additional guilty pleas from severed defendants. Gallo, 668 F.Supp. at 757.
Third, avoiding inconsistent verdicts is not a significant concern in a trial such as this. When several defendants are charged with jointly committing a single criminal act and tried on evidence that implicates them all equally, inconsistent verdicts may appear unfair and undermine public confidence in the judicial system. In a trial involving multiple defendants charged with separate substantive offenses, however, the acquittal of some defendants in separate trials is no more problematic than their acquittal in a joint trial, which occurred here as to several counts. Even where defendants are charged with the same offense, “inconsistent” verdicts are as possible in a joint trial as in separate trials. For example, of the seven defendants charged in the marijuana conspiracy in this case, four were convicted, one was acquitted, and the district court declared a mistrial as to the remaining two. We cannot see how the same result occurring after separate trials would result in any greater “scandal and inequity.” Richardson, 481 U.S. at 210, 107 S.Ct. at 1709.
Finally, the government contends that a joint trial avoids the possibility of witnesses who testify at the first severed trial being intimidated or otherwise prevented from testifying again. Although this is a genuine concern in a drug conspiracy case involving allegations of severe violence, we do not see how the risk is any greater in separate trials than in a joint trial where, as here, the defendants know the identity of most of the government’s witnesses far in advance. Even if, as the government alleges, some witnesses might refuse to testify more than once, their prior testimony would probably be admissible at subsequent trials under Fed.R.Evid. 804(b)(1). We recognize, however, that possible loss of testimony and, more importantly, risk to the lives of witnesses must be factored into the equation on a case-by-case basis.
2
Against the questionable benefits of a joint trial of this scope and duration, we must weigh the indisputably staggering hardships. These burdens fall not only on the defendants, but on defense counsel, prosecutors, the jury, the district court, the court of appeals, and the taxpayers.
The risk of prejudice to the defendants increases sharply with the number of defendants and the length of the trial. A trial’s length expands with the number of defendants not only because of the amount of evidence that must be presented, but also due to the scheduling conflicts that abound when dozens of jurors, defendants, and attorneys must be present in court at all times. This may often result in defendants having to endure months or even years of incarceration while they are presumed, and may in fact turn out to be, innocent. The Sixth Amendment speedy trial guarantee is rendered toothless when a verdict is not returned until years after an indictment.
There are a myriad of other potential sources of prejudice to an effective defense in trials of this scope. Defense counsel must call witnesses to attempt to impeach the credibility of prosecution witnesses who testified months earlier. Armies of defense counsel risk undermining each other with conflicting trial tactics and strategies. Defendants may have difficulty obtaining their counsel of choice, either because they cannot afford the staggering attorney fees of a year-long trial or because attorneys are unwilling to suspend the balance of their practice for such a protracted period. See Edward B. Williams, *1391The Problems of Long Criminal Trials, 34 F.R.D. 181, 182-83 (1963).
Most importantly, the human limitations of the jury system and the consequent risk of spillover prejudice cannot be ignored. This risk is particularly acute for comparatively peripheral defendants such as Robert Cole, who was charged only in the methamphetamine conspiracy count and whose separate trial could have been concluded in a matter of days or weeks, but who was required to sit in the courtroom during months of proof involving entirely unrelated conspiracies and substantive offenses. At oral argument in this case, the Assistant United States Attorney averred that his multiple violations of the district court’s limiting instructions during closing argument were the inadvertent result of confusion. When a seasoned prosecutor is unable to keep track of nearly 200 limiting instructions given over the course of a 16-month trial, our faith in a lay jury’s ability to do so is stretched to the limit. Our presumption that a jury is able to follow the trial court’s instructions is “rooted less in the absolute certitude that the presumption is true than in the belief that it represents a reasonable practical accommodation of the interests of the state and the defendant in the criminal justice process.” Richardson, 481 U.S. at 211, 107 S.Ct. at 1709. The presumption is not irrebuttable.
Beyond the risk of prejudice to the defendants, mega-trials such as this place a tremendous burden on the attorneys involved. Appointed defense counsel may sacrifice their time and other practice to earn less than half of what they normally charge, and the government must commit experienced prosecutors to a single trial indefinitely.2 The government also takes on a significant risk of reversal on appeal, not only as a result of the prejudice of such an epic trial, but because of any number of evidentiary or instructional errors that occur in the most basic proceedings. Even the most fair and attentive trial judge will err during the course of a 16-month trial, and some of those errors may require reversal.
This type of trial also imposes on citizens who fulfill their civic duty by serving on the jury. Jurors have their employment and home life disrupted, often at great financial, physical, and personal expense. They are required to “sit stoically and silently for hours every day, day after day,” Gallo, 668 F.Supp. at 754, and are prohibited from engaging in many ordinary pursuits of their daily lives, such as reading the newspaper.
The trial court is another of the mega-trial’s victims. Because the judge must adjourn the remainder of his or her calendar during the trial, “[t]he already overburdened docket of the court reaches a breaking point, and the administration of justice in all of the court’s cases is unconscionably delayed.” Id. at 755. The pressure to avoid a mistrial or reversal may also affect evidentiary rulings. “The option of a mistrial and a restarting of the case is almost closed when such a large expenditure of time and effort would be wasted.” Id.
The problems continue on appeal. Transcripts in this case were not filed until almost a year and a half after the end of trial, and oral argument was not heard until nearly four years after trial. The difficulties in coordinating briefing schedules and oral argument, the practical impossibility of a thorough review of the record, and the strain on judges and court clerks from reading the “briefs” (over 1200 pages in this appeal) make it more difficult fully to consider the issues raised and significantly burden our already congested calendar.3 We are un*1392aware of any ease that has imposed a comparable drain on this circuit’s resources.
Finally, we are abundantly aware that it is the taxpayers who frequently foot the bill for an extended criminal trial. The legal fees of defense attorneys appointed under the Criminal Justice Act, 18 U.S.C. § 3006A (1988), exceeded $2 million. Counsel for Robert Cole, whose separate trial we believe could have been concluded in a couple of weeks, filed over $250,000 in CJA vouchers during the joint trial. An additional $550,000 in appointed defense counsel fees have been paid on this appeal to the date of oral argument. When these millions of dollars in defense costs are combined with the millions in prosecution and court costs (including extensive reconstruction of the courtroom to accommodate the large number of defendants), the price tag of these 12 convictions is virtually indefensible.
In short, a trial of this scope and duration challenges the most fundamental goals of our federal criminal justice system: “simplicity in procedure, fairness in administration and the elimination of unjustifiable expense and delay.” Fed.R.Crim.P. 2.
3
The solution to the problems of the mega-trial “is largely in the hands of the United States Attorney, for he [or she] is in a position in the first instance to determine whether it would be more in the interests of criminal justice to restrict the number of defendants tried at any one time.” United States v. Agueci, 310 F.2d 817, 840 (2d Cir.1962), cert. denied, 372 U.S. 959, 83 S.Ct. 1013, 1016, 10 L.Ed.2d 11, 12 (1963). Prosecutors should consider the concerns we have discussed in deciding whether to bring a trial such as this in the future. If the government chooses to proceed with a single indictment, it should make a conscientious effort to “determine where [its] case is strongest and focus upon that area, reducing the number of counts and of defendants to manageable proportions.” Weinfeld, supra, 34 F.R.D. at 160.
The district court must also evaluate the burdens of a lengthy trial on the defendants’ due process rights in view of the factors discussed above and any others that may apply to the particular case. In Casamento, an appeal from a drug conspiracy trial that exceeded even the scope of this one, the Second Circuit provided some useful guideposts for the district court’s exercise of discretion in similar casés. 887 F.2d at 1151-52. The court recommended that the district court elicit the prosecution’s good-faith estimate of the time needed to present its case. While a conservative estimate should not become the subject of a contested hearing, the district court may independently assess the estimate in view of the number of defendants, the scope of the indictment, and the court’s own experience. When the estimate exceeds four months, the judge should require the prosecution to justify its conclusion that a joint trial serves the ends of justice. When more than ten defendants are involved, this justification should be especially compelling. Finally, the judge and the prosecutor should consider limiting the prosecution of peripheral defendants to easily provable charges that carry adequate penalties. We adopt the Second Circuit’s recommendations in their entirety.
We further note that when the district court relies on the government’s representations in assessing the likely length of a trial, the court may hold the prosecutor to those representations absent a showing of special circumstances. For example, when the district court requires the prosecution to provide a pretrial witness list, it need not allow the government, as it did here, to present nearly 100 witnesses not on that list (including the first two witnesses called).4 While these decisions are within the district court’s discretion, we believe that stricter adherence to pretrial commitments will guard against the prosecution’s underestimating the length of its case and will require more thorough and accurate trial preparation.
When the government and the district court fail adequately to guard against the *1393harms of a mega-trial, we will have no alternative but to reverse, whatever the cost of a retrial. Although we find it unnecessary to exercise that option today, we hope that trials such as this remain exotic blooms among legal flora and not rampant weeds threatening to strangle our most basic ideals of a fair and efficient justice system.
II. Pretrial Detention
In a vague and unsupported argument, Appellants contend that the prosecution manipulated the court to become its “unwitting partner” in abusing the Bail Reform Act. They claim that they were detained because they exercised their constitutional rights while others who waived those rights were granted conditional release, and that the Bail Reform Act was thus “unconstitutionally applied as punitive rather than preventative detention.” Appellants’ Joint Brief on Pretrial Issues at 90. This issue was not raised in the district court, and we therefore review for plain error. United States v. Dischner, 974 F.2d 1502, 1515 (9th Cir.1992), cert. denied, — U.S. -, 113 S.Ct. 1290, 122 L.Ed.2d 682 (1993). Appellants point to no evidence in the record indicating a miscarriage of justice or a failure of the integrity of the judicial process. See United States v. Smith, 962 F.2d 923, 935 (9th Cir.1992).
Appellants also argue that the government’s abuse of pretrial detention resulted in coerced pleas and cooperation with the prosecution. Of the nine people who were allegedly subjected to such coercion, only one is a party to this appeal; that defendant, Byron Wimberly, neither pled guilty nor cooperated with the government, and his inclusion in this list is a mystery. As for the other individuals named, Appellants lack standing to raise due process violations suffered by third parties. See United States v. Wingender, 790 F.2d 802, 803 (9th Cir.1986).
III. Juvenile Delinquency Act
Three of the four counts of which Richard Rupley, Jr., was convicted involved conduct occurring before his eighteenth birthday. Rupley, Jr., thus fell within the protections of the Juvenile Delinquency Act (JDA), 18 U.S.C. §§ 5031-42, with respect to those three counts.5 Rupley, Jr., argues that his convictions on the juvenile counts must be reversed based on improper transfer of those counts to adult status and numerous procedural violations of the JDA.
A. Transfer to Adult Status
The district court transferred Rupley, Jr., to adult status on counts 3, 10, and 15 pursuant to 18 U.S.C. § 5032. We hold that transfer was improper as to counts 3 and 10 and therefore reverse Rupley, Jr.’s convictions on those counts.
1. Count 3
Count 3 charged Rupley, Jr., with violating 21 U.S.C. § 846 by conspiring to manufacture, distribute, and possess with the intent to distribute marijuana between January 1983 and September 1985. At the time of Rupley, Jr.’s transfer to adult status (and at the time the charged offense was completed), § 5032 permitted transfer for any act committed after the juvenile’s fifteenth birthday “which if committed by an adult would be a felony that is a crime of violence or an offense described in [21 U.S.C. § 841].” 18 U.S.C. § 5032 (1988). Because count 3 alleged a violation of § 846 (albeit a conspiracy to violate § 841), the district court based its transfer order on a finding that the charged conspiracy was a “crime of violence.”
“Crime of violence” is defined in 18 U.S.C. § 16 (1988) as:
(a) an offense that has as an element the use, attempted use, or threatened use of physical force against the person or property of another, or
*1394(b) any other offense that is a felony and that, by its nature, involves a substantial risk that physical force against the person or property of another may be used in the course of committing the offense.
Count 3 does not fall within part (a) of this definition because “the use, attempted use, or threatened use of physical force” is not an element of a drug conspiracy under § 846. The count does not fall within part (b) of the definition because a conspiracy to manufacture, distribute, or possess a controlled substance does not “by its nature” involve a substantial risk of physical force. Although the government argues, and the district court found, that the violent overt acts alleged in count 3 made the conspiracy a crime of violence, the “by its nature” language of § 16(b) “implies that the generic, rather than the particular, nature of the predicate offense is determinative in defining a crime of violence.” See United States v. Cruz, 805 F.2d 1464, 1470 (11th Cir.1986), cert. denied, 481 U.S. 1006, 107 S.Ct. 1631, 95 L.Ed.2d 204, 482 U.S. 930, 107 S.Ct. 3215, 96 L.Ed.2d 702 (1987); see also United States v. Gonzalez-Lopez, 911 F.2d 542, 547 (11th Cir.1990), cert. denied, — U.S. -, 111 S.Ct. 2056, 114 L.Ed.2d 461 (1991); cf. United States v. Mendez, 992 F.2d 1488, 1490 (9th Cir.1993) (employing categorical approach to determine whether offense was “by its nature” a crime of violence under 18 U.S.C. § 924(c)). Had Congress intended a case-by-case inquiry into whether the felony as committed constituted a crime of violence, there would have been no need for the phrase “by its nature.” Cruz, 805 F.2d at 1469; see also United States v. Juvenile Male, 923 F.2d 614, 619 (8th Cir.1991). The cases cited by the government in support of a case-by-case approach interpret “crime of violence” as used in § 4B1.2 of the Sentencing Guidelines, which does not include the “by its nature” language that is dispositive here. See, e.g., United States v. Cornelius, 931 F.2d 490, 492-93 (8th Cir.1991). Because not every § 846 conspiracy involves a substantial risk of physical force, count 3 did not allege a crime of violence and thus was not transferable.
2. Count 10
Count 10 charged Rupley, Jr., with manufacture of methamphetamine in violation of 21 U.S.C. § 841(a). According to the superseding indictment, Rupley, Jr., committed the offense “on or about July 30, 1983,” when he was 15 years old. The transfer statute in effect on that date permitted transfer only for crimes “alleged to have been committed ... after [the juvenile’s] sixteenth birthday which if committed by an adult would be a felony that is punishable by a maximum penalty of ten years imprisonment or more, life imprisonment, or death.” 18 U.S.C. § 5032 (1982) (emphasis added). The statute was amended in 1984 to permit transfer for acts “alleged to have been committed ... after [the juvenile’s] fifteenth birthday which if committed by an adult would be a felony that is a crime of violence or an offense described in [21 U.S.C. § 841].” 18 U.S.C. § 5032 (1988) -(emphasis added); Pub.L. No. 98-473, § 1201(b)(2), 98 Stat. 1837, 2150 (1984). Because under the law existing at the time of the offense Rupley, Jr., could be adjudicated only under the JDA for count 10, he contends that application of the amended statute to that count violated the Ex Post Facto Clause. We review an alleged ex post facto violation de novo. United States v. Kohl, 972 F.2d 294, 297 (9th Cir.1992). Because Rupley, Jr., failed to raise this issue before the district court, however, we will reverse only if the transfer of count 10 constituted plain error. United States v. Calabrese, 825 F.2d 1342, 1346 (9th Cir.1987).
The Ex Post Facto Clause prohibits statutes “ ‘[1] which punish[ ] as a crime an act previously committed, which was innocent when done; [2] which make[ ] more burdensome the punishment for a crime, after its commission, or [3] which deprive[] one charged with crime of any defense available according to law at the time when the act was committed.’ ” Collins v. Youngblood, 497 U.S. 37, 42, 110 S.Ct. 2715, 2719, 111 L.Ed.2d 30 (1990) (quoting Beazell v. Ohio, 269 U.S. 167, 169-70, 46 S.Ct. 68, 69, 70 L.Ed. 216 (1925)). In United States v. Juvenile Male, 819 F.2d 468, 470 (4th Cir.1987), the Fourth Circuit held that the Ex *1395Post Facto Clause prohibited applying the 1984 amendment to § 5032 retroactively because it exposed the juvenile defendant to a much more severe sentence. We find the Fourth Circuit’s analysis persuasive and conclude that applying the amended statute to a previously committed offense violated the Ex Post Facto Clause.
Moreover, we hold that the transfer of count 10 was plain error under Fed. R.Crim.P. 52(b). The error is “clear” and unquestionably affected Rupley, Jr.’s “substantial rights” by subjecting him to a much harsher sentence than otherwise would have been authorized. See United States v. Olano, - U.S. -, - - -, 113 S.Ct. 1770, 1777-78, 123 L.Ed.2d 508 (1993). We therefore reverse Rupley, Jr.’s conviction on count 10.
3. Count 15
Count 15 also charged Rupley, Jr., with manufacture of methamphetamine in violation of 21 U.S.C. § 841(a). This count, however, alleged violations “on or about July 19, 1983 through on or about December 6, 1983.” Thus the conduct involved in count 15 occurred both before and after November 10, 1983, Rupley, Jr.’s sixteenth birthday. Although the government took the position at trial that count 15 alleged a single incident of manufacturing, various witnesses testified that the activity in question occurred anywhere between October and early December.
Were it indisputable that Rupley, Jr., committed the count 15 offense when he was 15, we would be required to reverse this conviction on ex post facto grounds despite the lack of an objection at trial. There was evidence, however, to support a finding that the crime occurred after Rupley, Jr.’s sixteenth birthday, thus making the 1984 amendment to § 5032 irrelevant. A jury’s verdict represents a finding that the crime was committed as alleged in the indictment. Calabrese, 825 F.2d at 1346; Leyvas v. United States, 371 F.2d 714, 717 (9th Cir.1967). In this ease, the jury’s verdict represents a finding, supported by the evidence, that Rupley, Jr., could have committed the offense after his sixteenth birthday. Although the district court, had Rupley, Jr., raised an ex post facto challenge to the transfer, might have been obligated to resolve the ambiguity in the evidence, any error is not obvious and does not clearly affect substantial rights. See Olano, - U.S. at - - -, 113 S.Ct. at 1777-78. We therefore hold that the transfer of count 15 did not constitute plain error. See Calabrese, 825 F.2d at 1346.
B. Alleged JDA Violations
With respect to count 15, the one juvenile count of conviction that we do not reverse based on improper transfer, we must consider the several alleged violations of the JDA that Rupley, Jr., claims merit reversal. We apply a three-step inquiry to alleged procedural violations of the JDA. We first ask whether the government violated the JDA’s requirements. If it did, we ask whether the government’s conduct was so egregious as to constitute a deprivation of due process. If it was not, we must determine whether the violation was harmless to the juvenile beyond a reasonable doubt. United States v. Doe, 862 F.2d 776, 779 (9th Cir.1988).
1. Certification
18 U.S.C. § 5032 (1988) provides in part:
A juvenile alleged to have committed an act of juvenile delinquency ... shall not be proceeded against in any court of the United States unless the Attorney General, after investigation, certifies to the appropriate district court of the United States that (1) the juvenile court or other appropriate court of a State does not have jurisdiction or refuses to assume jurisdiction over said juvenile with respect to such alleged act of juvenile delinquency, (2) the State does not have available programs and services adequate for the needs of juveniles, or (3) the offense charged is a crime of violence that is a felony or an offense described in [21 U.S.C. § 841] ... and that there is a substantial Federal interest in the case or the offense to warrant the exercise of Federal jurisdiction.
The government filed the certification on February 13, 1987, nearly a year before trial but approximately four months after Rupley, *1396Jr.’s arrest and arraignment. Rupley, Jr., argues that the government violated § 5032 by failing to file a proper certification before any proceedings in the district court had begun. We review this issue of statutory interpretation de novo. See Doe, 862 F.2d at 779.
Certification is a jurisdictional requirement. United States v. Juvenile Male, 864 F.2d 641, 643 (9th Cir.1988). But see United States v. Gonzalez-Cervantes, 668 F.2d 1073, 1077 (9th Cir.1981) (filing of accurate certification is not jurisdictional). The statute includes no deadline for filing the certification, however, and no court has imposed one. See United States v. Chambers, 944 F.2d 1253, 1260 (6th Cir.1991) (certification filed at close of prosecution’s case-in-chief was timely where defendants failed to invoke JDA earlier), cert. denied, - U.S. -, 112 S.Ct. 1217, 117 L.Ed.2d 455, - U.S. -, 112 S.Ct. 1680, 118 L.Ed.2d 397 (1992); United States v. Cuomo, 525 F.2d 1285, 1290 (5th Cir.1976) (without deciding the latest time for filing, holding that certification filed before arraignment was timely). Our own cases are unclear on the point. Compare Gonzalez-Cervantes, 668 F.2d at 1077 n. 6. (“LSection 5032] does not set forth when the certification need be filed. The statute provides: ‘shall not be proceeded against ... unless ... ’ a certification is filed. It does not state that proceedings cannot be undertaken until a certificate is filed.”) with id. at 1077 (“[Section 5032] quite clearly requires the filing of a certification before the district court can institute proceedings against the juvenile....”) (emphasis added) and Juvenile Male, 864 F.2d at 643 (“[Section 5032] require[s] the government to file a special certification regarding the juvenile before it can proceed against that juvenile.”) (emphasis added). See also Chambers, 944 F.2d at 1260 (“[Establishing the factors for federal jurisdiction is not always a prerequisite to initiating federal proceedings.”); Cuomo, 525 F.2d at 1289 (“[Section 5032] does not expressly apply to the commencement of an action against a juvenile ....”)• (emphasis added).
Congress’ purpose in enacting the certification requirement was “to help ensure that state and local authorities would deal with juvenile offenders wherever possible, keeping juveniles away from the less appropriate federal channels.” Juvenile Male, 864 F.2d at 644. We need not decide in this case whether a deadline for filing a certification is necessary to effectuate this purpose. The certification was furnished nearly a year before trial, and more than eight months before the filing of the superseding indictment on which Rupley, Jr., was tried and convicted. The certification was clearly timely.
2. Confinement in Maximum Security Institution
18 U.S.C. § 5035 (1988) provides in part:
A juvenile alleged to be delinquent may be detained only in a juvenile facility or such other suitable place as the Attorney General may designate.... The Attorney General shall not cause any juvenile alleged to be delinquent to be detained or confined in any institution in which the juvenile has regular contact with adult persons convicted of a crime or awaiting trial on criminal charges.
Because Rupley, Jr., was under 21 and stood accused of violating federal law prior to his eighteenth birthday, he was “[a] juvenile alleged to be delinquent” within the meaning of the statute. See 18 U.S.C. § 5031. During the 15 months between his arrest and indictment in October 1986 and his transfer to adult status, however, Rupley, Jr., was confined with his codefendants in a maximum security prison. Rupley, Jr., argues that the government’s violation of § 5035 requires reversal of his convictions. The government concedes that the statute was violated but argues that no prejudice resulted and that reversal is not warranted. Because Rupley, Jr., although represented by counsel at all times subsequent to his initial arrest and indictment, did not raise this argument in the district court, we review for plain error. Calabrese, 825 F.2d at 1346.
Section 5035 does not prescribe a remedy for its violation, and no court has addressed the specific issue. Under the circumstances of this case, we do not find the government’s lapses so egregious as to require reversal. See Doe, 862 F.2d at 780. *1397“There is no presumption or general rule that for every duty imposed upon the court or the Government and its prosecutors there must exist some corollary punitive sanction for departures or omissions, even if negligent.” United States v. Montalvo-Murillo, 495 U.S. 711, 717, 110 S.Ct. 2072, 2077, 109 L.Ed.2d 720 (1990). Had Rupley, Jr., objected to the conditions of his detention as violating § 5035, the proper remedy would likely have been dismissal of the juvenile charges without prejudice.6 He did not make such an objection. In September 1987, the government filed a superseding indictment charging Rupley, Jr., with two adult offenses as well as several committed before his eighteenth birthday. The superseding indictment mooted the earlier indictment and any procedural violations that may have tainted it. From that time on, Rupley, Jr., was properly detained in an adult facility, and any non-civil remedies that might previously have been available were waived. Moreover, Rupley, Jr., was eventually transferred to adult status and received credit at sentencing for the full period of his pretransfer confinement, thus minimizing any prejudice from the government’s error. See Doe, 862 F.2d at 780-81. We therefore need not invoke the “exceptional remedy” of reversal for plain error “to prevent a miscarriage of justice or to preserve the integrity and reputation of the judicial process.” United States v. Bryan, 868 F.2d 1032, 1039 (9th Cir.) (quoting United States v. Bustillo, 789 F.2d 1364, 1367 (9th Cir.1986)), cert. denied, 493 U.S. 858, 110 5.Ct. 167, 107 L.Ed.2d 124 (1989).
3. Speedy Trial
18 U.S.C. § 5036 (1988) provides in part:
If an alleged delinquent who is in detention pending trial is not brought to trial within thirty days from the date upon which such detention was begun, the information shall be dismissed on motion of the alleged delinquent dr at the direction of the court, unless the Attorney General shows that additional delay was caused by the juvenile or his counsel, or consented to by the juvenile and his counsel, or would be in the interest of justice in the particular case.
Rupley, Jr., argues that because the government failed to bring him to trial or move to transfer him to adult status within 30 days of his detention, the charges against him should have been dismissed.7
Rupley, Jr., was arrested on October 10, 1986, and ordered detained on October 14, 1986. On November 3, 1986, he moved for an extension of time to file pretrial motions and filed motions to suppress and for severance, among others. The time attributable to these motions is excludable from the 30-day limit under the “caused by the juvenile or his counsel” exception. Moreover, on January 16,1987, the magistrate found that “the ends of justice served by the granting of the government’s motion to continue the trial date ... outweigh the best interests of the public and the defendant in a speedy trial.” [ER F-5-6] This finding was not clearly erroneous. See United States v. Murray, 771 F.2d 1324, 1327 (9th Cir.1985). Thus any remaining time before filing of the government’s motion to transfer on February 13, 1987, was excludable under the “interest of justice” exception.
We therefore hold that none of the alleged procedural violations of the JDA require reversal of Rupley, Jr.’s count 15 conviction.
*1398IV. Discovery
Appellants allege at least 40 separate discovery violations, including the calling of witnesses not on the government’s pretrial witness list and violations of Fed.R.Crim.P. 16, the Jencks Act (18 U.S.C. § 3500), and Brady v. Maryland, 373 U.S. 83, 83 S.Ct. 1194, 10 L.Ed.2d 215 (1963). Fortunately, we need not address each alleged violation individually-
To reverse a conviction for a discovery violation, we must find not only that the district court abused its discretion, but that the error resulted in prejudice to substantial rights. United States v. Michaels, 796 F.2d 1112, 1115 (9th Cir.1986), cert. denied, 479 U.S. 1038, 107 S.Ct. 893, 93 L.Ed.2d 845 (1987); see also United States v. Span, 970 F.2d 573, 582 (9th Cir.1992) (violation of Jencks Act will not result in reversal if error is more likely than not harmless), cert. denied, — U.S. -, 113 S.Ct. 1283, 122 L.Ed.2d 676 (1993); United States v. Tham, 884 F.2d 1262, 1266 (9th Cir.1989) (reversal for Brady violation requires reasonable probability that result would have been different had the evidence been disclosed); United States v. Portillo, 633 F.2d 1313, 1324 (9th Cir.1980) (reversal based on testimony of undisclosed government witness requires defendant to show that he suffered “prejudicial surprise” from the testimony), cert. denied, 450 U.S. 1043, 101 S.Ct. 1764, 68 L.Ed.2d 241 (1981).
Assuming arguendo that the district court abused its discretion in each alleged instance by not finding a discovery violation, Appellants fail to demonstrate that prejudice resulted. The district court rou-finely granted the defense recesses and continuances to examine newly disclosed physical evidence or to prepare for cross-examination of unanticipated witnesses. In many cases, the district court based the length of a continuance or recess on the defense’s own estimate of how much preparation time was needed, and confirmed that the defense was ready to proceed when the trial reconvened. Appellants do not point to a single instance when the district court denied a request for additional time or when an alleged discovery violation otherwise caused them prejudice.8
Moreover, the district court did not abuse its discretion in selecting remedies for those discovery violations that it found to have occurred. “[A] district court has broad discretion to fashion remedies for the violation of its discovery orders,” United States v. Spillone, 879 F.2d 514, 522 (9th Cir.1989), cert. denied, 498 U.S. 864, 111 S.Ct. 173, 112 L.Ed.2d 137, 498 U.S. 878, 111 S.Ct. 210, 112 L.Ed.2d 170 (1990), and the sanction chosen should not be “harsher than necessary to accomplish the goals of Rule 16,” United States v. Gee, 695 F.2d 1165, 1169 (9th Cir.1983), or “disproportionate to the conduct of counsel,” United States v. Aceves-Rosales, 832 F.2d 1155, 1157 (9th Cir.1987), cert. denied, 484 U.S. 1077, 108 S.Ct. 1056, 98 L.Ed.2d 1018 (1988). Given the district court’s positive evaluation of the government’s cooperation9 and the court’s own accommodation of defense requests for additional time, the district court did not abuse its discretion in failing to impose severe sanctions for all discovery violations.10
V. Disqualification of Defense Counsel
On February 3, 1987, the district court granted the government’s motion to *1399disqualify Dominic Cavallaro’s counsel, Carl Martillaro, based on his prior representation of Crystal Channell. Channell, a key government witness, was Cavallaro’s former girlfriend and a previous wife of defendant John Bonnenfant. Martillaro had represented Channell on felony narcotics charges in Nevada state court in 1981-82. Cavallaro waived any potential conflict arising from the prior representation, but Channell refused to waive her attorney-client privilege. Although Martillaro stated in his affidavit that he did not recall any confidential information obtained from Channell that could be used to impeach her testimony and that he was willing to limit his cross-examination of her to public record material and questions concerning her grant of immunity for her testimony, he acknowledged the possibility that he might remember privileged information during trial. Cavallaro argues that the district court’s disqualification of Martillaro deprived him of his right to choice of counsel in violation of the Sixth Amendment. We review a district court’s pretrial disqualification of counsel for an abuse of discretion. United States v. Kenney, 911 F.2d 315, 320 (9th Cir.1990).
Although a presumption exists in favor of a defendant’s right to counsel of choice,
that presumption may be overcome not only by a demonstration of actual conflict but by a showing of a serious potential for conflict. The evaluation of the facts and circumstances of each case under this standard must be left primarily to the informed judgment of the trial court.
Wheat v. United States, 486 U.S. 153, 164, 108 S.Ct. 1692, 1700, 100 L.Ed.2d 140 (1988). Because “[t]he likelihood and dimensions of nascent conflicts of interest are notoriously hard to predict,” id. at 162, 108 S.Ct. at 1699,
the district court must be allowed substantial latitude in refusing waivers of conflicts of interest not only in those rare cases where an actual conflict may be demonstrated before trial, but in the more common cases where a potential for conflict exists which may or may not burgeon into an actual conflict as the trial progresses.
Id. at 163, 108 S.Ct. at 1699; see United States v. Rewald, 889 F.2d 836, 857-58 (9th Cir.1989) (defendant’s waiver of conflict is not dispositive; potential for conflict is determinative), amended, 902 F.2d 18 (9th Cir.), cert. denied, 498 U.S. 819, 111 S.Ct. 64, 112 L.Ed.2d 39 (1990).
We conclude that the district court did not abuse its discretion in disqualifying Martillaro based on the potential for a conflict of interest. It was reasonable to presume that effective cross-examination of Channell would include questioning her ability accurately to recall, observe, and testify about CaVallaro’s activities, and that her drug use would be a significant factor in this impeachment. That Channell’s drug use was a matter of public record does not eliminate the possibility of an unwitting' disclosure of confidential communications. Martillaro could thus have been faced with either exploiting his prior, privileged relationship with the witness or failing to defend his present client zealously for fear of misusing confidential information.
Cavallaro argues that no potential conflict existed because the facts underlying the charges against him were not “substantially related” to Martillaro’s prior representation of Channell. See Nev.S.Ct.R. 159. Substantial relationship to the prior representation, however, turns not only on factual identity, but may be presumed where there is “a reasonable probability that confidences were disclosed which could be used against the client in later, adverse representation.” Trone v. Smith, 621 F.2d 994, 998 (9th Cir.1980). The district court found such a probability here. The disqualification of Martillaro was within the court’s discretion and did not violate Cavallaro’s Sixth Amendment rights.
VI. Government’s Preparation of its Case
Appellants contend that they were prejudiced by the government’s failure to prepare its case adequately. Specifically, they argue that the district court relied on the government’s erroneous prediction of the length of its case-in-chief in ruling on motions regarding severance, pretrial detention, shackling, and juror notetaking, and that *1400they wei’e prejudiced by the government’s deviations from its initial witness list.
At the start of trial, the government estimated that its case-in-chief would be completed within six months. In fact, the government’s case lasted nearly a year. While we do not condone the government’s spectacularly inaccurate estimate and suspect that the prosecution could have been better prepared, we also recognize that the government could not predict the continuances attributable to absences of defense counsel and juror illnesses or foretell the length of cross-examination of its witnesses. We do not believe that the government’s failure adequately to predict the length of its case warrants reversal.
Appellants’ argument that they were prejudiced by the government’s calling of witnesses not on its pretrial witness list is similarly meritless. Appellants do not specify any instance in which they were misled or harmed by the calling of a witness not on the pretrial list. The district court required the government to give the defense at least 24 hours notice of any witness’ appearance and allowed the defense time to prepare for witnesses not on the original list when the defense so requested. On all but one of the few occasions when the defense objected to an additional government witness, the district court determined how much notice had been given, found that no prejudice existed, and overruled the objection. As to the one sustained objection, the court delayed the witness’ appearance until the next day, although it concluded that no prejudice had been demonstrated. Appellants do not explain how these findings were in error. Admitting the additional witnesses’ testimony was within the district court’s sound discretion. See United States v. Franklin, 704 F.2d 1183, 1190-92 (10th Cir.) (where government witness was undisclosed in violation of pretrial discovery order, exclusion of testimony was not required where defense had sufficient time to prepare and did not request an additional continuance), cert. denied, 464 U.S. 845, 104 S.Ct. 146, 78 L.Ed.2d 137 (1983).
VII. Speedy Trial
Trial in this case was initially set for November 3, 1987. On October 1, 1987, the government filed a superseding indictment naming six additional defendants and adding about 30 new counts. On October 9, 1987, the district court, on its own motion and over the objection of some defendants, continued the trial date until January 19, 1988. Appellants contend that this continuance violated the “heart and spirit” of the Speedy Trial Act as well as their Sixth Amendment right to a speedy trial.
A. Speedy Trial Act
We review a district court’s interpretation of the Speedy Trial Act de novo. United States v. Sears, Roebuck & Co., 877 F.2d 734, 737 (9th Cir.1989). Factual findings relating to excusable delay under the Act are reviewed for clear error. United States v. Aviles-Alvarez, 868 F.2d 1108, 1111 (9th Cir.1989).
The Speedy Trial Act is violated if a defendant is not tried within 70 days of the latest of either the filing of an indictment or the initial court appearance. See 18 U.S.C. § 3161(c)(1) (1988). When several defendants are joined for trial, the 70-day period begins to run on the date the last codefendant is indicted or arraigned. Henderson v. United States, 476 U.S. 321, 323 n. 2, 106 S.Ct. 1871, 1873 n. 2, 90 L.Ed.2d 299 (1986); United States v. Morales, 875 F.2d 775, 776 (9th Cir.1989). The issue is therefore whether the 81 days that elapsed between the last arraignment on October 30, 1987, and the first day of trial on January 19, 1988, included at least 11 days of excludable time.
18 U.S.C. § 3161(h)(1)(F) (1988) provides for exclusion of any “delay resulting from any pretrial motion, from the filing of the motion through the conclusion of the hearing on, or other prompt disposition of, such motion.” This exclusion is automatic and no specific findings are required. Aviles-Alvarez, 868 F.2d at 1112. Nor is the exclusion limited to delays that are “reasonably necessary.” Henderson, 476 U.S. at 329-30, 106 S.Ct. at 1876. In this case, numerous overlapping pretrial motions were *1401pending from October 30, 1986, the date of the last arraignment under the initial indictment, until the first day of trial. Once time periods attributable to pending pretrial motions are excluded, the Speedy Trial Act was not violated.
B. Sixth Amendment
A Sixth Amendment speedy trial claim is reviewed de novo. United States v. Wallace, 848 F.2d 1464, 1469 (9th Cir.1988). Factors to consider in determining whether the Sixth Amendment right to a speedy trial has been violated include: (1) the length of the delay; (2) the reason for the delay; (3) the defendant’s assertion of the right; and (4) the prejudice to the defendant. Barker v. Wingo, 407 U.S. 514, 530, 92 S.Ct. 2182, 2192, 33 L.Ed.2d 101 (1972).
Some defendants waited over 15 months between their initial arraignment and trial. The district court, however, agreed with the government that this lengthy delay was justified by the extreme complexity of the case, the prolonged discovery process, and the need to give counsel for newly joined codefendants time to prepare. Indeed, Appellants were responsible for much of the delay. The superseding indictment and resulting continuance stemmed in large part from continuing criminal activity by the defendants while in custody or on pretrial release. Moreover, the major part of the delay resulted from the distant scheduling of the initial trial date, a process in which Appellants participated with the government and the magistrate.
Finally, Appellants fail to allege any meaningful prejudice,, such as impairment of defenses, resulting from the delay. See Barker, 407 U.S. at 532, 92 S.Ct. at 2193. Nor were Appellants prejudiced by extended incarceration, as they all received credit at sentencing for time served. The failure to present evidence of actual prejudice weighs heavily in favor of the government. See United States v. Holm, 550 F.2d 568, 569 (9th Cir.) (per curiam), cert. denied, 434 U.S. 856, 98 S.Ct. 176, 54 L.Ed.2d 127 (1977). We find no Sixth Amendment violation.
VIII. Shackling
After hearing the recommendations of the United States Marshal’s Service and the objections of the defendants, the court agreed with the Marshal that all nine in-custody defendants should be shackled during trial. We review a trial court’s decision to shackle a defendant for an abuse of discretion, Jones v. Meyer, 899 F.2d 883, 884 (9th Cir.), cert. denied, 498 U.S. 832, 111 S.Ct. 95, 112 L.Ed.2d 67 (1990), and the underlying factual findings for clear error, Spain v. Rushen, 883 F.2d 712, 717 (9th Cir.1989), cert. denied, 495 U.S. 910, 110 S.Ct. 1937, 109 L.Ed.2d 300, 495 U.S. 948, 110 S.Ct. 2208, 109 L.Ed.2d 534 (1990).
“[Generally, a defendant has the right to appear before the jury free of shackles or other restraints.” Wilson v. McCarthy, 770 F.2d 1482, 1484 (9th Cir.1985). This right, however,
is not absolute. The trial court has discretion to use shackles or other security measures when circumstances dictate. The trial court- must balance the prejudicial effect of shackling with considerations of courtroom decorum and security.... [S]hackling is proper where there is a serious threat of escape or danger to those in and around the courtroom or where disruption in the courtroom is likely in the absence of shackles.
Id. at 1484-85 (citations omitted); see also Illinois v. Allen, 397 U.S. 337, 343-44, 90 S.Ct. 1057, 1061, 25 L.Ed.2d 353 (1970); Spain, 883 F.2d at 716. A trial judge must engage in a two-step process before deciding to impose physical restraints. First, compelling circumstances must persuade the court that some measure -is needed to maintain courtroom security. Second, the court must pursue less restrictive alternatives to shackling. Jones, 899 F.2d at 884-85. “If the alternatives [to shackling] are less onerous yet no less beneficial, due process demands that the trial judge opt for one of the alternatives.” Spain, 883 F.2d at 728.
Compelling circumstances existed in this case. The Marshal’s Service, which bears responsibility for the safety and welfare of the trial judge, jury, courtroom *1402personnel, and defendants, spent almost a year evaluating the security needs of this trial and rated the trial at level four, the highest level of threat. The security inspector for the Marshal’s Service testified that the trial presented unique security concerns and expressed his opinion that it would not be safe to conduct the trial without shackling the in-custody defendants.
Appellants argue that the court erred in failing to independently evaluate the risks presented by each defendant. While an individualized assessment of the defendants’ characteristics or circumstances might have been appropriate had they been tried separately, the U.S. Marshal believed that for security purposes the defendants should be treated as a single criminal organization. It is certainly reasonable to conclude that nine defendants acting in concert pose a risk far exceeding that which would be posed by any one individually. Each defendant was alleged to have committed acts of violence or intimidation. Government witnesses were threatened and intimidated during pretrial proceedings. While in custody, several of the defendants conspired to kill a government witness and discussed the murder of the FBI agent in charge of the case. The district court did not abuse its discretion in accepting the Marshal’s recommendation that the defendants be treated as a group for security purposes.11
The court also found that no less restrictive alternatives to shackling existed. Deputy marshals testified that the only alternative to shackling would be to station two marshals with each in-custody defendant and an additional six outside of the courtroom well, requiring 24 marshals in the courtroom (in addition to those required for in-custody witnesses). Although the trial was in the largest courtroom in Nevada, marshals would have been standing virtually shoulder-to-shoulder around the defendants. The district court reasonably found that flooding the courtroom with deputy marshals
would be a much more repressive atmosphere, too many law officers in one place and that there is serious danger to all concerned by that great concentration of people and I would have to conclude that it would be more prejudicial to the defendants to have all those law officers here than to have them shackled.
R.T. 1/15/88 at 61.
Nor was the shackling itself unduly oppressive or prejudicial. Each of the nine in-custody defendants had one leg shackled to the pedestal of his chair. No waist chains or handcuffs were used. Appellants do not contend that the shackles caused them pain, impaired their mental faculties, or impeded their ability to communicate with counsel. See Castillo v. Stainer, 983 F.2d 145, 149 (9th Cir.1992), amended, 997 F.2d 669 (9th Cir.1993); Jones, 899 F.2d at 885. The shackles were wrapped with tape and foam rubber to mute any sound, and panels were installed in front of the defendants’ legs to shield the shackles from the jury’s sight. All shackling and unshackling took place out of the presence of the jury and the public. The jury was instructed that no one would rise when the judge took the bench. Although Appellants assert that the jury must have realized that the nine defendants were shackled, they offer no evidence in support of this claim. Because the shackling employed was minimal and there is no evidence that the jury was ever aware of the restraints, there was no prejudice. See Castillo, 983 F.2d at 149; Spain, 883 F.2d at 722.
Finally, Appellants urge that the need for shackling could have been eliminated by severing the trials. We rejected the same argument in Loux v. United States, 389 F.2d 911, 920 (9th Cir.), cert. denied, 393 U.S. 867, 89 S.Ct. 151, 21 L.Ed.2d 135, 393 U.S. 869, 89 S.Ct. 156, 21 L.Ed.2d 138 (1968). As in Loux, it does not follow that the security risks resulting in the level four rating would have disappeared in separate trials. Although the shackling was an appropriate consideration in the district court’s rulings on *1403the motions to sever, the court did not abuse its discretion in denying those motions.
IX. Juror Notetaking
The . district court denied a defense motion to allow jurors to take notes during the trial. We review the denial of a motion to permit juror notetaking for an abuse of discretion. United States v. Vaccaro, 816 F.2d 443, 451 (9th Cir.), cert. denied, 484 U.S. 914, 108 S.Ct. 262, 98 L.Ed.2d 220, 484 U.S. 928, 108 S.Ct. 295, 98 L.Ed.2d 255 (1987).
“Whether it is advisable to permit a jury to take notes is a subject of some debate, and reasonable arguments are advanced for and against the practice. The decision of whether to allow the jury to take notes is left entirely to the discretion of the trial court.” Vaccaro, 816 F.2d at 451. The district court articulated a number of reasons for denying the motion, including concerns that jurors would be distracted from observing witnesses, would record the evidence selectively, or would rely on their own or other jurors’ inaccurate notes. The ongoing argument over the validity of these concerns is not for us to decide. The district court did not abuse its broad discretion in declining to allow juror notetaking.
X. Jury Selection
Appellants claim several errors in the district court’s voir dire of prospective jurors. We review a district court’s voir dire procedures for an abuse of discretion. United States v. Anzalone, 886 F.2d 229, 234 (9th Cir.1989). The court’s findings of juror impartiality may be overturned only for manifest error. Mu'Min v. Virginia, 500 U.S. 415, -, 111 S.Ct. 1899, 1907, 114 L.Ed.2d 493 (1991).
A.Voir Dire on Pretrial Publicity
Appellants claim that the district court failed adequately to voir dire prospective jurors regarding media exposure and awareness of the defendants’ shackling. A district court has wide discretion in conducting voir dire regarding pretrial publicity or other potential sources of juror bias. Mu'Min, 500 U.S. at -, 111 S.Ct. at 1906. “Several general questions [about pretrial publicity] addressed to the entire panel of jurors, followed by individual questioning of jurors who respond affirmatively to the initial inquiries, may be sufficient if it becomes clear that few jurors have any knowledge of the case.” United States v. Giese, 597 F.2d 1170, 1183 (9th Cir.), cert. denied, 444 U.S. 979, 100 S.Ct. 480, 62 L.Ed.2d 405 (1979); see also United States v. Polizzi, 500 F.2d 856, 879-80 (9th Cir.1974), cert. denied, 419 U.S. 1120, 95 S.Ct. 802, 42 L.Ed.2d 820 (1975). In this ease, group questioning revealed that only six of approximately 88 prospective jurors were aware of information beyond that contained in the indictment. The district court conducted individual voir dire of those six jurors. Three of the six were excused for cause; the other three were not challenged. The district court’s failure to engage in the more intensive probe of each juror requested by the defense did not “render the defendants’] trial fundamentally unfair.” Mu'Min, 500 U.S. at -, 111 S.Ct. at 1905.
B. Voir Dire on Relationship to Government Witnesses
Appellants claim that reversal is required because they could not voir dire jurors on their relationship with those government witnesses not included on the pretrial witness list. We rejected the same argument in United States v. Dischner, 974 F.2d 1502, 1522-23 (9th Cir.1992), cert. denied, — U.S. -, 113 S.Ct. 1290, 122 L.Ed.2d 682 (1993). Although a trial court abuses its discretion in failing to ask prospective jurors any questions concerning acquaintance with any government witnesses, United States v. Washington, 819 F.2d 221, 224 (9th Cir.1987); United States v. Baldwin, 607 F.2d 1295, 1297 (9th Cir.1979), “[njeither Baldwin nor Washington ... requires disclosure of all witnesses or directs the trial court to question veniremen about every possible government witness.” Dischner, 974 F.2d at 1523. Furthermore, Appellants do not contend that any juror in fact was acquainted with any government witness.
C. Impairment of Peremptory Challenges
Appellants contend that the district court’s refusal to ask proposed supple*1404mental questions of six prospective jurors denied the defense the opportunity to probe bias and qualifications and precluded meaningful exercise of peremptory challenges. “It is wholly within the judge’s discretion to reject supplemental questions proposed by counsel if the voir dire is otherwise reasonably sufficient to test the jury for bias or partiality.” United States v. Powell, 932 F.2d 1337, 1340 (9th Cir.), cert. denied, — U.S. -, 112 S.Ct. 256, 116 L.Ed.2d 210 (1991). After examining the record, we are satisfied that the district court’s voir dire was sufficient without the supplemental questions. Even if there was error, however, Appellants fail to demonstrate prejudice. None of the six people at issue was on the jury that decided the case. Only one of the 12 jurors who decided the case was challenged for cause (based on the fact that she was the mother of an infant), and Appellants do not suggest that any of the 12 were biased or incompetent. Furthermore, although the defense exercised peremptory challenges to three of the six jurors, the Supreme Court has explicitly rejected the notion that the loss of a peremptory challenge is a constitutional violation. Ross v. Oklahoma, 487 U.S. 81, 88, 108 S.Ct. 2273, 2278, 101 L.Ed.2d 80 (1988). “So long as the jury that sits is impartial, the fact that the defendant had to use a peremptory challenge to achieve that result does not mean the Sixth Amendment was violated.” Id.12
For the same reasons, we reject Appellants’ argument that the district court committed reversible error by not excusing two prospective jurors for cause. The two prospective jurors in question were not on the jury that decided the case. Even if Appellants were forced to use peremptory challenges to remove jurors who should have been removed for cause, this does not amount to a constitutional violation.
Appellants also argue that their Fifth Amendment due process right to peremptory challenges was impaired by the district court’s refusal to ask the proposed questions. They rely on Swain v. Alabama, 380 U.S. 202, 219, 85 S.Ct. 824, 835, 13 L.Ed.2d 759 (1965), overruled in part by Batson v. Kentucky, 476 U.S. 79, 106 S.Ct. 1712, 90 L.Ed.2d 69 (1986), which held that “[t]he denial or impairment of the right [to exercise peremptory challenges] is reversible error without a showing of prejudice.” In Ross, however, the Supreme Court limited Swain by holding that the due process “right” to peremptory challenges is “denied or impaired” only if the defendant does not receive the full complement of challenges to which he is entitled by law. Ross, 487 U.S. at 89, 108 S.Ct. at 2278. Fed.R.Crim.P. 24(b) allots 10 peremptory challenges to the defense and permits the district court to grant additional challenges in multidefendant cases. The district court in this case granted 25 peremptory challenges, to be exercised jointly. Accordingly, there was no due process violation.
XI. Restrictions on Recross-Examination
Appellants were permitted to cross-examine all witnesses at great length. They claim, however, that their right to recross-examination of several witnesses was unduly restricted in violation of their Sixth Amendment right to confrontation. Whether the court’s limitation of recross-examination constitutes a violation of the Confrontation Clause is reviewed de novo. United States v. Jones, 982 F.2d 380, 383 (9th Cir.1992). Within the bounds of constitutionality, we review a district court’s limitation of questioning for an abuse of discretion. Id.
Allowing recross is within the sound discretion of the trial court except where new matter is elicited on redirect examination, in which case denial of recross as to that new matter violates the Confrontation Clause. See Cossack v. United States, 63 F.2d 511, 517 (9th Cir.1933); United States v. Riggi, 951 F.2d 1368, 1375 (3d Cir.1991); United States v. Caudle, 606 F.2d 451, 458 (4th Cir.1979). Although the general rule is *1405clear, the authorities are devoid of any analysis of what constitutes "new matter." If "new matter" is defined broadly, then any question asked on redirect that had not already been asked and answered would conceivably introduce "new matter" requiring the opportunity for recross insofar as it expanded or elaborated on the witness' previous testimony. Such an approach would conflict with the trial court's discretion to impose reasonable limits on cross-examination. Delaware v. Van Arsdall, 475 U.S. 673, 679, 106 S.Ct. 1431, 1435, 89 L.Ed.2d 674 (1986); see United States v. Tarantino, 846 F.2d 1384, 1405 (D.C.Cir.), cert. denied, 488 U.S. 840, 109 S.Ct. 108, 102 L.Ed.2d 83, 488 U.S. 867, 109 S.Ct. 174, 102 L.Ed.2d 143 (1988).
We conclude, however, that the district court applied an overly narrow definition of "new matter." The district court interpreted the standard to preclude recross if the questions fell within an "area" or "subject matter" for which cross-examination had previously been available. As applied, this test improperly denied the defense recross as to material new matter brought out on redirect. Two examples demonstrate this error.
Sergeant David Frey testified for the government concerning seizures of drugs, and the defense cross-examined him extensively regarding his drug expertise. On redirect, the prosecutor brought out that other defense lawyers had stipulated to Frey's exper-' tise in over 25 court appearances. The defense sought a brief recross to establish that the prior stipulations had been in the context of state preliminary hearings, where guilt is not determined and where such stipulations are common. The district court denied recross on the grounds that "[Frey's] qualifications as an expert have been gone into completely." Although Frey's expertise was not a new subject, the government introduced the matter of defense stipulations to Frey's expertise for the first time on redirect. The Confrontation Clause required the opportunity for cross-examination as to this newly elicited and potentially damaging testimony.
DEA Agent Dale Kitts testified regarding the capacity of a 22-liter flask, commonly used as a reaction vessel in manufacturing methamphetamine. On direct examination, Kitts testified that the flask has a production potential of "approximately four pounds." The defense did not press the production yield issue on cross-examination, apparently because it was satisfied with the 4-pound testimony. On redirect, the prosecutor ir~correctly stated Kitts' previous testimony as a yield of "about five pounds," 13 and then proceeded to establish that the potential production of a 22-liter flask was as high as 10 to 15 pounds. The district court barred the defense from challenging this testimony on recross, stating: "I don't see how that's anything new. That was gone into, he did say four pounds, now he says more, that's rehabilitation, it's not a new area that hasn't been gone into before." We cannot agree that the redirect was mere rehabilitation because the 4-pound testimony had never been challenged on cross-examination. The general "subject matter" of the flask's production potential was not introduced on redirect, but Agent Kitts' testimony tripling the production potential presented new, material, and potentially damaging matter requiring an opportunity for recross.
Although we conclude that the court's undue restrictions of recross violated the Confrontation Clause, this violation is subject to harmless error analysis. Jones, 982 F.2d at 384. Reversal is not required if, assuming the damaging potential of recross-examination were fully realized, we can say that the error was harmless beyond a reasonable doubt. Van Arsdall, 475 U.S. at 684, 106 S.Ct. at 1438. Factors to consider in determining harmlessness include "the importance of the witness' testimony in the prosecution's case, whether the testimony was cumulative, the presence or absence of evidence corroborating or contradicting the testimony of the witness on material points, the extent of cross-examination otherwise *1406permitted, and ... the overall strength of the prosecution’s case.” Id.
We hold, based on our analysis of these factors, that the error was harmless beyond a reasonable doubt. Appellants make no showing that the opportunity for further recross-examination would have affected the outcome of their case. For example, although Agent Kitts’ testimony was used in calculating the total amount of methamphetamine manufactured for purposes of the prosecution’s summary charts, we cannot conclude that the jury would not have found the “substantial proceeds” required for the continuing criminal enterprise convictions had the 10- to 15-pound production potential been successfully impeached. Similarly, we do not believe that the verdict would have been different had Appellants been allowed to show that prior stipulations to Sergeant Frey’s drug expertise occurred in the context of state preliminary hearings. Nor do Appellants cite any other recross restrictions sufficiently prejudicial to east doubt on the verdict.
Appellants also argue that the district court permitted the prosecution to introduce new matters on redirect but denied them the same opportunity. The trial court has discretion to allow a new line of questioning on redirect examination. United States v. Lopez, 575 F.2d 681, 686 (9th Cir.1978). Appellants make no showing that the district court abused its discretion in this regard.
XII. Continuing Criminal Enterprise
Appellants Richard Rupley, Sr., Dwain Baker, Edward Baker, and John Bonnenfant (collectively, “the CCE defendants”) raise a number of challenges to their convictions for engaging in a continuing criminal enterprise (“CCE”) in violation of 21 U.S.C. § 848. The CCE statute, sometimes referred to as the “kingpin” statute, provides in relevant part:
[A] person is engaged in a continuing criminal enterprise if—
(1) he violates any provision of this sub-chapter or subchapter II of this chapter the punishment for which is a felony, and
(2) such violation is a part of a continuing series of violations of this subchapter or subchapter II of this chapter—
(A) which are undertaken by such person in concert with five or more other persons with respect to whom such person occupies a position of organizer, a supervisory position, or any other position of management, and
(B) from which such person obtains substantial income or resources.
21 U.S.C. § 848(c) (1988). The “continuing series” requirement has been interpreted as consisting of three or more federal narcotics violations. United States v. Hernandez-Escarsega, 886 F.2d 1560, 1570 (9th Cir.1989), cert. denied, 497 U.S. 1003, 110 S.Ct. 3237, 111 L.Ed.2d 748 (1990). The statute authorizes harsh criminal penalties and forfeiture of the enterprise’s assets.
A. Evidence of Multiple Conspiracies
Among the 35 charged predicate acts to the CCE, count 1 of the indictment listed three separate conspiracies: (1) a conspiracy between Rupley, Sr., and Bonnenfant to distribute cocaine from November 1975 to December 1981 (not charged as a separate count due to the statute of limitations); (2) a conspiracy between all four CCE defendants to manufacture and distribute methamphetamine from December 1981 to September 1987 (also charged as count 2); and (3) a conspiracy between Rupley, Sr., and the Bakers to cultivate, grow, and manufacture marijuana from late 1982 to late 1985 (also charged as count 3). Although the district court initially instructed the jury not to consider evidence of the cocaine conspiracy against the Bakers or evidence of the marijuana conspiracy against Bonnenfant, it later reversed itself, admitting cocaine evidence against the Bakers and marijuana evidence against Bonnenfant without limitation. Thus, for example, cocaine evidence was admitted against the Bakers under count 1 although they did not join the Company until after the alleged cocaine conspiracy ended, and marijuana found in Bonnenfant’s room upon his arrest was admitted against him under count 1 although he was not charged in the count 3 marijuana conspiracy. Bon-*1407nenfant and the Bakers argue that the district court erred in admitting against them evidence of conspiracies in which they did not participate.
The government contended at trial that because the CCE statute encompasses the elements of a conspiracy, see Jeffers v. United States, 432 U.S. 137, 149-50, 97 S.Ct. 2207, 2216, 53 L.Ed.2d 168 (1977), vicarious liability for coconspirators’ substantive offenses under Pinkerton v. United States, 328 U.S. 640, 647, 66 S.Ct. 1180, 1184, 90 L.Ed. 1489 (1946), should apply equally to a CCE charge. Although the prosecution later disavowed the Pinkerton analogy and agreed to limiting instructions on the cocaine and marijuana conspiracies, the district court apparently adopted the government’s earlier conspiracy analogy and admitted against all four CCE defendants evidence of all acts allegedly committed in furtherance of the criminal enterprise.
It is true that the CCE statute requires a defendant to act “in concert” with five or more people. This “in concert” element, however, refers to the relationship between the CCE defendant and those whom he organizes, supervises, or manages; that is, it contemplates a “vertical,” or hierarchical, nexus. Insofar as a CCE includes a conspiracy, it is a conspiracy between the CCE defendant and his underlings. Some courts have concluded that Pinkerton liability applies along that vertical nexus. See United States v. Graewe, 774 F.2d 106, 108 (6th Cir.1985), cert. denied, 474 U.S. 1068, 106 S.Ct. 826, 88 L.Ed.2d 798, 474 U.S. 1069, 106 S.Ct. 828, 88 L.Ed.2d 800 (1986); United States v. Michel, 588 F.2d 986, 999 (5th Cir.), cert. denied, 444 U.S. 825, 100 S.Ct. 47, 62 L.Ed.2d 32 (1979).14 The statute, however, does not require that jointly charged CCE defendants act “in concert” with one another. Thus there is no basis for imputing predicate offenses under Pinkerton along a “horizontal” nexus between a CCE defendant and others charged in the same count.15
To admit evidence of one CCE defendant’s predicate offenses against other CCE defendants who never participated in those offenses assumes the existence of a conspiracy that the government in this case neither alleged nor proved. Bonnenfant was found to have acted “in concert” with those whom he organized, supervised, or managed, not with his CCE codefendants. Similarly, the Bakers acted “in concert” not with Rup-ley, Sr., and Bonnenfant, but with the individuals whom they organized, supervised, or managed with respect to their series of Title 21 violations. If the government believed that Rupley,.Sr., Bonnenfant, and the Bakers were conspirators in a single “umbrella” narcotics conspiracy that encompassed the other conspiracies, then the indictment should have charged such a conspiracy and acts in furtherance thereof could have been imputed among the defendants under Pinkerton. Charging the four defendants with individually meeting the CCE statutory requirements is not a substitute for alleging an additional conspiracy among the CCE defendants. We therefore hold that the district court misapplied Pinkerton in admitting the marijuana evidence against Bonnenfant and the cocaine evidence against the Bakers.
We conclude, however, that this error was harmless. See United States v. Chu Kong Yin, 935 F.2d 990, 994 (9th Cir.1991) (noneonstitutional evidentiary errors will be reversed for abuse of discretion only if they more likely than not affected the verdict). Rupley, Sr., was unanimously convicted of 11 separate Title 21 offenses, Bonnenfant was convicted of seven, and the Bakers were each convicted of five. As to each CCE defendant, the jury properly could have used any three of his convictions as predicate acts for the GCE conviction. Significantly, only Dwain Baker challenges the sufficiency of the evidence for his CCE conviction, and then only with regard to the “five or more persons” requirement. The prosecutor did not argue Pinkerton liability to the jury as a *1408basis for the CCE convictions. The most reasonable conclusion is that the jury properly found the “continuing series” element of each CCE conviction satisfied by drug crimes for which that defendant was convicted, not offenses in which the defendant was uninvolved.
The CCE defendants also argue that joining three conspiracies within count 1 rendered the count duplicitous. We disagree. A count is duplicitous if it joins “two or more distinct and separate offenses.” United States v. UCO Oil Co., 546 F.2d 833, 835 (9th Cir.1976), cert. denied, 430 U.S. 966, 97 S.Ct. 1646, 52 L.Ed.2d 357 (1977). Count 1 charges only a single offense (engaging in a CCE) which is by definition based on a series of predicate acts. Title 21 conspiracy violations may serve as predicate offenses under the CCE statute. Hernandez-Escarsega, 886 F.2d at 1571.
Finally, the CCE defendants allege that including multiple conspiracies in the CCE count raises the possibility of a non-unanimous verdict. The district court, however, instructed the jury that it was required to agree unanimously on which three acts constituted the continuing series of violations committed by each CCE defendant and on the five or more persons with whom each defendant committed those violations. In view of the numerous Title 21 violations of which each CCE defendant was unanimously convicted, there is no reason to believe that the jury did not unanimously agree on the three predicate violations. See United States v. LeMaux, 994 F.2d 684, 689 (9th Cir.1993) (even where unanimity instruction on predicate acts was not given, unanimous substantive convictions and overwhelming evidence made it “inconceivable” that jury would not have convicted CCE defendant).
B. Aiding and Abetting a CCE
The CCE defendants allege that the government improperly argued and the district court erroneously instructed that they could be convicted for aiding and abetting the leader of a CCE without independently meeting all the statutory requirements. The record reveals that this argument is meritless.
Whether a defendant who aids and abets a CCE “kingpin” may be convicted under § 848 based on the federal aider arid abettor statute, 18 U.S.C. § 2(a), is the subject of an intercircuit split. Compare United States v. Amen, 831 F.2d 373, 381-82 (2d Cir.1987) (aiding and abetting liability conflicts with plain terms and clear intent of § 848), cert. denied, 485 U.S. 1021, 108 S.Ct. 1573, 99 L.Ed.2d 889 (1988), with United States v. Pino-Perez, 870 F.2d 1230, 1233-34 (7th Cir.) (en banc) (§ 2(a) attaches automatically to any federal criminal statute, including § 848), cert. denied, 493 U.S. 901, 110 S.Ct. 260, 107 L.Ed.2d 209 (1989). In United States v. Miskinis, 966 F.2d 1263, 1267-68 (9th Cir.1992), we declined to address the issue because the defendant had been charged and convicted as a principal, not as an aider and abettor.16 Similarly, we find it unnecessary to resolve the question here.
The government’s theory, supported by the evidence at trial, was that each CCE defendant violated § 848 in his own right. In his closing argument, the prosecutor told the jury that they would “have to make individual decisions as to whether each and every one of [the CCE defendants] satisfied the essential [CCE] elements beyond a reasonable doubt.” In order to convict a defendant under § 848, the government must show that (1) the defendant committed a felony violation of federal narcotics law which was (2) part of a continuing series of three or more violations (3) undertaken in concert with five or more persons (4) with respect to whom the defendant acted as an organizer, supervisor, or manager,17 and (5) from which *1409the defendant derived substantial income or resources. 21 U.S.C. § 848(c); see United States v. Sterling, 742 F.2d 521, 525 (9th Cir.1984), cert. denied, 471 U.S. 1099, 105 S.Ct. 2322, 85 L.Ed.2d 840 (1985). Although Rupley, Sr., clearly held the highest position in the Company, a defendant need not fit the label of a “kingpin,” United States v. Medina, 940 F.2d 1247, 1251 (9th Cir.1991), or “ringleader,” United States v. Zavala, 839 F.2d 523, 527 (9th Cir.), cert. denied, 488 U.S. 831, 109 S.Ct. 86, 102 L.Ed.2d 62 (1988), to be convicted under § 848. Nor does anything in the statute suggest that there may only be one head of a CCE. Pino-Perez, 870 F.2d at 1236. Rather, anyone who meets the statutory requirements may properly be convicted under § 848.
In support of their argument, the CCE defendants cite the district court’s general aiding and abetting instruction, given because counts 10, 12, and 15 of the indictment alleged aiding and abetting violations.18 The district court’s separate instructions on the CCE charge made no mention of aiding and abetting and stated that “[t]he government must prove beyond a reasonable doubt each of the essential elements of this offense in order to establish the guilt of each defendant.” Although the CCE defendants allege that the district court’s aiding and abetting instruction on the CCE predicate offenses gave the jury the impression that the defendants could be convicted on the CCE charge as aiders and abettors, they failed to object on this ground at trial. We are therefore limited to a plain error review. United States v. Armstrong, 909 F.2d 1238, 1243-44 (9th Cir.), cert. denied, 498 U.S. 870, 111 S.Ct. 191, 112 L.Ed.2d 153 (1990). Even if giving the aiding and abetting instruction without specific limitation could have been misleading, it is not “highly probable that the error materially affected the verdict.” United States v. Kessi, 868 F.2d 1097, 1102-03 (9th Cir.1989).
C. The “Organizer, Supervisor, or Manager” Element
The government argued at trial that the four CCE defendants could undertake different roles in different predicate offenses and thus “manage each other.” Although the CCE defendants did not object to this argument at trial, they argue on appeal that one cannot be convicted of engaging in a CCE if he is managed by others. We disagree. Because § 848 deals with individual liability, a CCE defendant may be an organizer, supervisor, or manager in the predicate offenses underlying his CCE conviction and a subordinate in a predicate offense underlying another’s CCE conviction. In other words, if one meets all the statutory requirements individually, he may properly be convicted under § 848 even though he himself was “managed.” See United States v. Sophie, 900 F.2d 1064, 1078 (7th Cir.) (one CCE defendant supervised another CCE defendant for purposes of § 848), cert. denied, 498 U.S. 843, 111 S.Ct. 124, 112 L.Ed.2d 92 (1990); United States v. Rosenthal, 793 F.2d 1214, 1226 (11th Cir.) (same), modified, 801 F.2d 378 (11th Cir.1986), cert. denied, 480 U.S. 919, 107 S.Ct. 1377, 94 L.Ed.2d 692 (1987). The Congressional intent of creating a separate offense to punish leaders of major drug organizations is not limited to the single person on the top rung of the ladder. See Zavala, 839 F.2d at 527 (defendant need not be “ringleader” to be among class of offenders sought to be penalized under § 848).
The CCE defendants also argue that the jury instructions on the “organizer” element constituted reversible error under United States v. Jerome, 942 F.2d 1328, 1331 (9th Cir.1991). Because the defendants failed to object to these instructions at trial, we again review for plain error. The district court instructed that “the word ‘organizer’ does not carry with it the implication that the organizer is necessarily able to control those whom he or she organizes.”19 In Jerome, *1410we interpreted the language “or any other position of management” in § 848 to mean that “an ‘organizer’ must exercise some sort of managerial responsibility; one does not qualify if one simply sets up a system of supply.” Id. The district court’s instruction in this case squarely accords with our holding in United States v. Ray, 731 F.2d 1361, 1367 (9th Cir.1984). Moreover, contrary to the CCE defendants’ assertion, Jerome does not render the instruction defective. The people organized by the CCE defendants in this case were not, as in Jerome, “the suppliers of [their] suppliers,” Jerome, 942 F.2d at 1330, but rather underlings in the Company’s drug manufacturing and distribution activities. In view of the defendants’ failure to object or to propose an alternate instruction, we find no plain error here.
Jerome also held that where some of the people named by the prosecution could not legally be counted as supervisees, a unanimity instruction was required. Id. at 1331; see LeMaux, 994 F.2d at 688. The CCE defendants suggest that their convictions may have been based on nonunanimous verdicts as to the identity of the five subordinates in each series of predicate offenses. There was substantial evidence that each of the CCE defendants managed, organized, or supervised five or more persons even without including his codefendants. Even if the prosecution named people who could not be considered subordinates, however, the court satisfied Jerome by instructing the jury that they “must unanimously agree on each of the five or more persons with whom each of the defendants committed the violations.”
D. Statute of Limitations
The CCE defendants contend that the government improperly introduced evidence of time-barred predicate acts. This argument misapprehends the purpose of the statute of limitations. In United States v. Musacchio, 968 F.2d 782 (9th Cir.1991), a defendant who had been convicted of misapplying bank funds argued that evidence of his misrepresentations to the bank’s Board of Directors was erroneously admitted because the misrepresentations occurred outside the limitations period for the misapplication of funds offense. We rejected this contention, stating:
Musacchio is attempting to convert the statute of limitations from a procedural rule that requires the bringing of a complaint within a certain time after the completion of a crime to a rule that restricts the introduction of evidence. We find no support for this use of the statute of limitations .... The statute of limitations did not bar the bringing of this action. The function of the statute of limitations ends with this determination. The statute of limitations does not bar the introduction of evidence of acts that occurred outside the limitations period.
Id. at 790; see also United States v. Drebin, 557 F.2d 1316, 1333 (9th Cir.1977) (in prosecution for interstate transportation of stolen property, only interstate transportation, and not thefts, need occur within limitations period), cert. denied, 436 U.S. 904, 98 S.Ct. 2232, 56 L.Ed.2d 401 (1978).
Section 848 creates a distinct offense separate from its predicate acts. Garrett v. United States, 471 U.S. 773, 781-86, 105 S.Ct. 2407, 2412-15, 85 L.Ed.2d 764 (1985). The statute of limitations does not begin to run until all elements of the crime occur. Toussie v. United States, 397 U.S. 112, 115, 90 S.Ct. 858, 860, 25 L.Ed.2d 156 (1970); Musacchio, 968 F.2d at 790. Therefore, just as time-barred offenses may serve as predicate acts to establish a pattern of racketeering activity under RICO, see United States v. Torres Lopez, 851 F.2d 520, 528 (1st Cir.1988), cert. denied, 489 U.S. 1021, 109 S.Ct. 1144, 103 L.Ed.2d 204 (1989); United States v. Pepe, 747 F.2d 632, 663 n. 55 (11th Cir.1984), such offenses may serve as predicate CCE crimes as long as an act in the series of offenses occurred within the CCE limitations period. See Hernandez-Escarsega, 886 F.2d at 1578 (defendant’s drug-related activities in 1970s were valid CCE predicate offenses in 1986 trial).
XIII. Summary Witness
The prosecution’s final witness was FBI Special Agent Lee Besse, who placed a total *1411value on the narcotics transactions and expenditures testified to during the government’s case. Agent Besse sat at the prosecution table and took notes throughout the trial. Her testimony was aimed at establishing the “substantial proceeds” element required for convicting the four CCE defendants. By projecting values onto transactions and events and totalling known expenditures, the prosecution worked backwards to show substantial proceeds. When no witness testified as to the value of a particular transaction, Agent Besse estimated a value based on price lists she had prepared from the testimony. A 134-page computer summary was admitted for illustrative purposes and distributed to the jury for reference during her testimony. Summary charts of the printout’s totals in various categories were admitted for illustrative and demonstrative purposes.20
The district court admitted Agent Besse’s evidence under Fed.R.Evid. 1006 and 702, stating that a summary witness was needed to “help the jurors to put together the evidence.” The court instructed the jury that the summary testimony and exhibits were not evidence, did not represent an opinion of the court or the prosecution on the credibility of witnesses, and were to be disregarded to the extent the jury found them conflicting with the testimony and evidence received at trial. Appellants contend that the district court improperly allowed the prosecution to present an early summation of evidence favorable to its case by a witness favorable to the government and that the evidence was incomplete, biased, cumulative, confusing, and misleading. We review the admission of summary exhibits and expert testimony for an abuse of discretion. United States v. Marchini, 797 F.2d 759, 766 (9th Cir.1986), cert. denied, 479 U.S. 1085, 107 S.Ct. 1288, 94 L.Ed.2d 145 (1987).
Although this circuit has often allowed the use of summary charts and summary witness testimony based on testimonial evidence (most commonly in tax cases), there is scant analysis on the authority for this practice. See, e.g., United States v. Poschwatta, 829 F.2d 1477, 1481 (9th Cir.1987), cert. denied, 484 U.S. 1064, 108 S.Ct. 1024, 98 L.Ed.2d 989 (1988); Marchini, 797 F.2d at 765-66. Neither Rule 1006 nor Rule 702 authorizes Agent Besse’s testimony in this case.
Rule 1006 states that “[t]he contents of voluminous writings, recordings, or photographs which cannot conveniently be examined in court may be presented in the form of a chart, summary, or calculation.” Although at least one circuit has read Rule 1006 broadly enough to encompass summaries of previously admitted oral testimony, see United States v. Winn, 948 F.2d 145, 158 (5th Cir.1991), cert. denied, — U.S. -, 112 S.Ct. 1599, 118 L.Ed.2d 313 (1992), we reject this interpretation as contrary to the plain language of the rule. See id. at 158 n. 32 (acknowledging that “[sjummary of purely testimonial evidence is strictly speaking not within the purview of Rule 1006”). Furthermore, Rule 1006 provides for the use of summaries or charts as evidence; in this case, the district court instructed the jury that the summary testimony and exhibits were not evidence. See 5 Jack B. Weinstein and Margaret A. Berger, Weinstein’s Evidence ¶ 1006[07] (1992) (distinguishing between use of summaries as evidence under Rule 1006 and as pedagogical devices that summarize or organize previously admitted evidence).
Nor does Rule 702, which provides for admission of the testimony of a qualified expert if “specialized knowledge will assist the trier of fact to understand the evidence or to determine a fact in issue,” justify admitting Agent Besse’s testimony. Although Agent Besse was a certified accountant, she had absolutely no experience or expertise regarding the street value of narcotics. Rather, she calculated the value of the narcotics involved using figures derived from the *1412evidence presented at trial. Presumably the only thing that rendered Agent Besse more capable than the jury to perform this function was that she was permitted to take notes. Agent Besse’s simple arithmetical calculations were not an appropriate subject for expert testimony under Rule 702. Additionally, Rule 702, like Rule 1006, contemplates the introduction of expert testimony as evidence.
We conclude, however, that admitting Agent Besse’s testimony was a valid exercise of the district court’s discretion under Fed.R.Evid. 611(a), which authorizes the court to “exercise reasonable control over the mode ... of ... presenting evidence so as to (1) make the ... presentation effective for the ascertainment of the truth, [and] (2) avoid needless consumption of time.” See United States v. Paulino, 935 F.2d 739, 752-54 (6th Cir.) (testimony of nonexpert summary witness regarding cash generated from cocaine sales in drug conspiracy case was admissible under Rule 611(a) where trial court gave limiting instruction and defense had full opportunity to cross-examine), cert. denied, — U.S. -, 112 S.Ct. 315, 116 L.Ed.2d 257, — U.S. -, 112 S.Ct. 323, 116 L.Ed.2d 264, — U.S. -, 112 S.Ct. 660, 116 L.Ed.2d 751 (1991), - U.S. -, 112 S.Ct. 883, 116 L.Ed.2d 787 (1992); United States v. Gardner, 611 F.2d 770, 776 (9th Cir.1980) (summary chart admissible in tax evasion case under Rule 611(a) because it “contributed to the clarity of the presentation to the jury, avoided needless consumption of time and was a reasonable method of presenting the evidence”); United States v. Scales, 594 F.2d 558, 563-64 (6th Cir.) (summaries of testimonial evidence designed “to aid the jury in its examination of the evidence already admitted” do not come within Rule 1006, but are authorized by Rule 611(a)), cert. denied, 441 U.S. 946, 99 S.Ct. 2168, 60 L.Ed.2d 1049 (1979); 5 Weinstein, supra, at ¶ 1006[03] (summary “prepared by a witness from his own knowledge to assist the jury in understanding or remembering a mass of details ... is admissible, not under Rule 1006, but under such general principles of good sense as are embodied in Rule 611(a)”).
We further hold that admission of the summary evidence was not unduly prejudicial under Fed.R.Evid. 403. We are not blind to the dangers of witnesses summarizing oral testimony, and we believe that such summaries should be admitted under Rule 611(a) only in exceptional cases. Permitting an “expert” witness to summarize testimonial evidence lends the witness’ credibility to that evidence and may obscure the jury’s original evaluation of the original witnesses’ reliability. In this case, although Agent Besse did not expressly testify to the veracity of the underlying data,21 the exhibits that she prepared did not cover all events presented in the government’s case, but only those that Agent Besse believed to be sufficiently specific or reliable. Thus Agent Besse’s selective summary itself constituted a subjective determination of reliability, and the original testimony, an attack on the credibility of which formed the core of the defense’s case, was assumed to be true. Furthermore, a summary of oral testimony is generally the purpose and province of closing argument, and we believe that it would have been more appropriate for the prosecutor to present Agent Besse’s summary exhibits and valuations in his closing remarks.
Any prejudice in this case, however, was not undue because of the precautions taken by the district court. The court required the government to lay a foundation for the summary evidence outside the presence of the jury and granted a continuance of over one week to allow the defense to examine the materials in detail. The defense had full opportunity to cross-examine Agent Besse about her methods of preparing the summaries, her alleged selectivity, and her partiality. The court gave a thorough limiting instruction three times during Agent Besse’s testimony, minimizing any risk that the sum*1413mary would be treated as substantive evidence. The district court invited the defense to present its own summary witnesses, but the defense declined to do so. Finally, any possible prejudice from Agent Besse’s testimony would be harmless in view of the overwhelming evidence of narcotics transactions involving substantial sums of money. See United States v. Cuevas, 847 F.2d 1417, 1428 (9th Cir.1988), cert. denied, 489 U.S. 1012, 109 S.Ct. 1122, 103 L.Ed.2d 185 (1989).
XIV. Systematic Pattern of Violence
Count 1 of the superseding indictment, the CCE count, alleged that the defendants
engaged in a systematic pattern of violence in support of its drug-based criminal enterprise, with its members using force, fear, violence and intimidation to guarantee discipline, loyalty and control among its members, to maintain, expand and protect its drug distribution territories, and to protect their various clandestine laboratories established for the illicit production of methamphetamine.
Counts 2 and 3, the conspiracy counts, used the identical language under the heading of “Overt Acts.” Appellants argue that the “systematic pattern of violence” alleged in the indictment was either an element of the offenses in question, requiring a jury instruction, or prejudicial surplusage that injected a false issue into the trial. This argument was not raised at trial. Appellants further contend that evidence of the Company’s violent acts should have been excluded under Fed. R.Evid. 403 and 404(b).
A. CCE Count
Because a CCE charge is based solely on Title 21 offenses, the “systematic pattern of violence” alleged in the CCE count was not an element of that offense. It was therefore surplusage and need not have been proved. Gawne v. United States, 409 F.2d 1399, 1403 (9th Cir.1969), cert. denied, 397 U.S. 943, 90 S.Ct. 956, 25 L.Ed.2d 123 (1970). There is no reason to believe, however, that this allegation interfered with the jury’s consideration of the essential elements of a CCE conviction. Id. at 1404. All of the 35 predicate acts alleged in the CCE count were Title 21 violations; not one was an act of violence. In any event, individual violent acts were relevant to the CCE count insofar as they constituted overt acts in furtherance of the two charged conspiracies, which in turn were alleged as predicate CCE offenses.
B. Conspiracy Counts
The “systematic pattern of violence” alleged in the two conspiracy counts appeared in the context of overt acts, and a significant number of the overt acts listed in the superseding indictment were acts of violence. An overt act in furtherance of a conspiracy is not a substantive charge requiring instruction on essential elements, but is itself an element of the charged offense.22 As for Appellants’ Rule 403 argument, much of the testimony about violent acts was admitted without objection from the defense. The record shows that the district court balanced relevance against prejudicial effect under Rule 403 when objections were raised. The district court’s Rule 403 decisions will not be disturbed absent an abuse of discretion. United States v. Kessi, 868 F.2d 1097, 1107 (9th Cir.1989). We hold that the district court did not abuse its discretion. See United States v. Patterson, 819 F.2d 1495, 1504-05 (9th Cir.1987) (evidence of shooting charged as overt act in drug conspiracy was not unfairly prejudicial and was relevant because shooting incident was an example of pattern of violence alleged in indictment).
Appellants’ argument that the acts of violence were inadmissible under Fed. R.Evid. 404(b) as “other crimes, wrongs, or acts” is misguided. Rule 404(b) precludes the use of such evidence “to prove the character of a person in order to show action in conformity therewith.” The overt acts of violence in this case were not used to prove the defendants’ character or violent propensity, but as substantive evidence of the *1414charged conspiracies. See Patterson, 819 F.2d at 1504 n. 3; United States v. Soliman, 813 F.2d 277, 278-79 (9th Cir.1987).
C. The Mary Bacon Shooting Incident
One of the overt acts alleged in furtherance of the methamphetamine and marijuana conspiracies was that on August 16, 1983, Dominic Cavallaro attempted to kill Mary Bacon, an employee of the United States Forest Service, by shooting at her five times with a large caliber firearm while Bacon was on official duty in a remote part of California’s Shasta National Forest. Witness Dale Richmond testified that Cavallaro had admitted to shooting at Bacon to scare her away from the Squaw Creek marijuana growing area. The district court admitted the evidence over defense objection and subsequently denied a defense motion to strike this overt act from counts 2 and 3 on the grounds that the incident had not been tied to the charged conspiracies. On appeal, Cavallaro vehemently argues that he did not commit the shooting, that the act was not connected to the conspiracies and was therefore irrelevant, and that, even if relevant, the district court should have excluded the evidence under Rule 403.
Prior to Richmond’s testimony, the district court determined that Cavallaro’s admission was made in furtherance of the conspiracy under Fed.R.Evid. 801(d)(2)(E) and conducted a Rule 403 analysis. In denying the motion to strike the overt act from the indictment, the court stated that it found sufficient evidence that the shooting was in furtherance of the two conspiracies. The admission of the evidence was not an abuse of the district court’s discretion. As to Cavallaro’s culpability, the defense thoroughly cross-examined Richmond and presented exculpatory witnesses. It was for the jury to decide whether Cavallaro was responsible for the shooting.
XV. Defense Witness Immunity
The indictment alleged that Rupley, Sr., was responsible for the murder of methamphetamine “cook” Edward Osick. The defense attempted to call Carol Calabrese, Richard Calabrese, and Terry Logan, three convicted methamphetamine manufacturers, to rebut this charge. According to Carol Calabrese’s grand jury testimony, the Cala-breses’ partner, Dominic Esposito, had made several statements suggesting that he had killed Osick. The three witnesses invoked the Fifth Amendment, and the government refused to grant them immunity. Appellants claim that this refusal constituted prosecuto-rial misconduct and that the district court should have required the government to grant immunity.
Immunity is an executive, not a judicial, function, and “[t]his court has emphatically rejected the argument that the sixth amendment provides a defendant with a right to demand use immunity for defense witnesses who invoke their privilege against self-incrimination.” United States v. Brutzman, 731 F.2d 1449, 1451-52 (9th Cir.1984). There is an exception, however, where the prosecutor intentionally distorts the fact-finding process and thereby denies the defendant a fair trial. United States v. Lord, 711 F.2d 887, 891-92 (9th Cir.1983). Although this exception generally requires the prosecutor affirmatively to induce the witness to invoke the Fifth Amendment privilege, see United States v. Montoya, 945 F.2d 1068, 1078 (9th Cir.1991); United States v. Paris, 827 F.2d 395, 401 (9th Cir.1987), we may also find intentional distortion of the fact-finding process where the government “grant[s] immunity to a witness in order to obtain his testimony, while denying immunity to a defense witness whose testimony would directly contradict that of the government witness.” United States v. Westerdahl, 945 F.2d 1083, 1087 (9th Cir.1991).
If the government had a legitimate reason for denying immunity, they have not provided it to this court. Carol and Richard Calabrese and Terry Logan were not charged with committing any crimes in the District of Nevada, and the Calabreses had been granted immunity by the United States Attorney in the Southern District of California for the events that would have been covered in their testimony. Appellants do not show, however, that the government coerced or prompted any witness to invoke the privilege. The evidence demonstrated that Carol and Richard Calabrese asserted *1415their Fifth Amendment rights based on advice from their attorneys and that Terry Logan had no contact with the prosecutor.
Nor is it clear that the immunized testimony of the Calabreses and Logan would have directly contradicted government evidence of the Osick murder. The district court admitted Carol Calabrese’s grand jury testimony pursuant to Fed.R.Evid. 804(b)(1), but only to the extent that her live testimony would have been admissible. The incriminating hearsay statements attributable to Esposito were thus properly excluded, and would have been even had the government granted the witnesses immunity.23
Lord and Westerdahl, on which Appellants rely, are readily distinguishable. We reversed in those cases because the trial court failed to hold an evidentiary hearing to determine whether the government intentionally distorted the fact-finding process. The district court conducted such a hearing in this case and determined that the defense failed to prove that the government denied immunity with the intent to distort the fact-finding-process. This finding was not clearly erroneous.
Finally, even were we to find a Sixth Amendment violation, it would be harmless. Neither Rupley, Sr., nor any other defendant was charged with the murder of Osick. Although Appellants maintain that the allegation that Rupley, Sr.’s alleged murder of Osick was “a cornerstone of the government’s case,” the Osick homicide was only one of over 70 overt acts charged in connection with the methamphetamine conspiracy, only one of which needed to be proved for a conspiracy conviction. Appellants fantasize when they claim “a substantial possibility that the testimony of the Calabreses and Terry Logan would have materially affected the outcome of the ease.”
XVI. Closing Argument
A. Jill Rttpley
1. Misrepresentations of the Evidence
Appellant Jill Rupley alleges that the prosecutor exaggerated, misrepresented, and told ‘ outright lies” about the evidence against her during his closing argument and at her sentencing. No objection was raised to the prosecutor’s statements on either occasion, and review is therefore for plain error. United States v. Lane, 708 F.2d 1394, 1399 (9th Cir.1983).
Our review of the transcripts reveals that any possible misrepresentations were not substantial enough to amount to plain error. Counsel are given latitude in the presentation of their closing arguments, and “courts must allow the prosecution to strike ‘hard blows’ based on the evidence presented and all reasonable inferences therefrom.” United States v. Gwaltney, 790 F.2d 1378, 1385 (9th Cir.1986) (quoting United States v. Prantil, 764 F.2d 548, 555 (9th Cir.1985)) cert. denied, 479 U.S. 1104, 107 S.Ct. 1337, 94 L.Ed.2d 187 (1987). The evidence presented and inferences therefrom supported the prosecutor’s contention that Jill Rupley was an integral part of the Company, serving as a courier of money and methamphetamine, selling methamphetamine to students at the University of Nevada-Reno, and making personal use of the Company’s substantial assets. Significantly, Rupley does not challenge the sufficiency of the evidence on any of her six counts of conviction.
Rupley argues that the prosecutor improperly relied on her presence at meetings and manufacturing operations to suggest that she was involved in the illegal activity. This is a valid argument of inferences and does not constitute plain error. See United States v. Penagos, 823 F.2d 346, 348 (9th Cir.1987) (“While mere proximity to the scene of illicit activity is not sufficient to establish involvement in a conspiracy, a defendant’s presence may support such an inference when viewed in context with other evidence.”). Moreover, Jill Rupley requested and received a “mere presence” instruction.
2. Violations of Limiting Instructions
During his closing argument, the prosecutor twice argued Jill Rupley’s involve*1416ment in marijuana cultivation and distribution based on evidence that the court had limited to defendants charged in the marijuana conspiracy. The prosecutor also argued Jill Rupley’s use of cocaine based on evidence that had been limited to the CCE defendants. The government concedes on appeal that this was error, but insists that it was not malicious and does not require reversal of Jill Rupley’s convictions.
In United States v. Sherlock, 962 F.2d 1349, 1360 (9th Cir.1989), cert. denied, — U.S. -, 113 S.Ct. 419, 121 L.Ed.2d 342 (1992); the prosecutor urged the jury in his closing argument to consider against Sherlock a hearsay statement of a nontestifying codefendant, which the trial court had ruled admissible only against the declarant. We concluded that the prosecutor’s error revealed either “that the trial was so complicated that not even he could apply the limiting instructions or that he understood, and intended, his misconduct.” Id. at 1362. In either event, we held that the prosecutor’s argument significantly prejudiced Sherlock and that the trial court therefore abused its discretion in denying Shertock’s motion for severance.
There is a key distinction, however, between Sherlock and this case. In Sherlock, the defense moved for a mistrial on the basis of the prosecutor’s argument, id. at 1360; here, there was not even an objection. We therefore review only for plain error and must determine whether the prosecutor’s blatant violation of the court’s limiting instructions was “a highly prejudicial error affecting substantial rights.” United States v. Giese, 597 F.2d 1170, 1199 (9th Cir.), cert. denied, 444 U.S. 979, 100 S.Ct. 480, 62 L.Ed.2d 405 (1979). We find no plain error here. Unlike the prejudicial statement in Sherlock, which directly implicated the appellant in the crime of which he was charged and convicted, the erroneous argument in this case was tangential at best and more likely than not harmless. It is unlikely that the prosecutor’s comments on Jill Rupley’s involvement with marijuana and cocaine significantly contributed to her multiple methamphetamine-related convictions. Thus, while we have no doubt that a properly raised objection to the prosecutor’s argument should have been sustained, we find no basis for reversing Jill Rupley’s convictions.
B. Richard Rupley, Jr.
Appellant Richard Rupley, Jr., claims that the prosecutor improperly engaged in personal attacks on his defense counsel during rebuttal argument. The prosecutor’s statements that defense counsel was trying to “con [the jury] into an acquittal” and was “adopting] the criminal mentally [sic]” of his client were met with sustained objections. Rupley, Jr., did not object to the prosecutor’s statements that his counsel was “desperate,” telling the jury “some things that did not occur in the record,” and engaging in “legal ostriehism ... a rare medical disorder that affects defense attorneys ... [and] is manifested by sticking your head in the sand when the important evidence comes in against your client.”
“[I]mproprieties in counsel’s arguments to the jury do not constitute reversible error unless they prejudice the defendant and that prejudice has not been remedied by the trial judge.” United States v. Lopez-Alvarez, 970 F.2d 583, 597 (9th Cir.), cert. denied, — U.S. -, 113 S.Ct. 504, 121 L.Ed.2d 440 (1992). We have previously recognized that improper attacks on the integrity of defense counsel can cause prejudice. Bruno v. Rushen, 721 F.2d 1193, 1195 (9th Cir.1983), cert. denied, 469 U.S. 920, 105 S.Ct. 302, 83 L.Ed.2d 236 (1984). In this case, however, some of the prosecutor’s improper comments were neutralized by a sustained objection, and all were subject to the court’s instructions that statements and arguments of counsel were not evidence and were not to be considered as such. The comments to which no objection was made at trial do not rise to the level of plain error.
XVII. Jury Instructions
In an argument devoid of legal authority, Appellants claim that jury comprehension of instructions is ordinarily “pitifully low” and in this case was “abysmal.” They base this argument on the number of jury instructions given (112) and the purportedly confusing wording of these instructions. Appellants *1417did not object at trial to the jury instructions as a whole, and review is therefore for plain error.
Our test for the adequacy of jury instructions is “ ‘whether or not the instructions taken as a whole were misleading or represented a statement inadequate to guide the jury’s deliberations.’ ” United States v. Shortt Accountancy Corp., 785 F.2d 1448, 1454 (9th Cir.) (quoting Stoker v. United States, 587 F.2d 438, 440 (9th Cir.1978)), cert. denied, 478 U.S. 1007, 106 S.Ct. 3301, 92 L.Ed.2d 715 (1986). In this case,
the charge to the jury was as clear as instructions to juries ordinarily are or reasonably can be_ Juries have for centuries made the basic decisions between guilt and innocence ... upon the basis of the facts, as revealed by all the evidence, and the law, as explained by instructions detailing the legal distinctions, the placement and weight of the burden of proof, the effect of presumptions, the meaning of intent, etc. We think that to condemn the operation of this system here would be to condemn the system generally. We are not prepared to do so.
Leland v. Oregon, 343 U.S. 790, 800, 72 S.Ct. 1002, 1008, 96 L.Ed. 1302 (1952); see also United States v. Sullivan, 595 F.2d 7, 8-9 (9th Cir.) (“[0]ur whole jury system is based upon the recognized ability of the jury to follow the court’s instructions.”), cert. denied, 442 U.S. 933, 99 S.Ct. 2867, 61 L.Ed.2d 302 (1979). The careful attention to each count and each codefendant evidenced by the verdicts indicates that the court’s instructions provided adequate guidance.
XVIII. Segregation of Evidence During Jury Deliberations
During its deliberations, the jury requested to view all evidence seized from searches of several locations in a display on separate tables according to location. The district court granted the request and, over defense objections, ordered the prosecution to segregate and display the evidence. Appellant Robert Rowen argues that because the segregation of evidence required reliance on trial testimony and conclusions as to which items were seized from which locations, the court improperly participated in the fact-finding process. This argument is specious. The location from which evidence was seized was never a disputed issue at trial. Furthermore, the ease on which Row-en relies, United States v. Walker, 575 F.2d 209, 214 (9th Cir.), cert. denied, 439 U.S. 931, 99 S.Ct. 320, 58 L.Ed.2d 325 (1978), is concerned with the trial court answering inquiries from the jury, not simply presenting evidence that has been requested.
Rowen also alleges that the prosecutor displayed as evidence seized from the Wells Avenue residence where Rowen was arrested items that were not actually found during the search. These items included (1) $6,136 in cash that was seized from Rowen at the Sheriffs Department following his arrest, and (2) one ounce of methamphetamine, purchased from Tom Baldwin by an undercover officer, that was traced to Rowen and provided probable cause for the Wells Avenue search warrant. Although the defense did not object to including the money and methamphetamine in the Wells Avenue display, Rowen argues that the evidence improperly contributed to his conviction on count 30 for possession with intent to distribute approximately one pound of methamphetamine.
We find no plain error in the inclusion of this evidence. Count 30 charged Rowen with possession of “approximately one pound of methamphetamine” in violation of 21 U.S.C. § 841(a). We have noted that “[s]ection 841(a) does not specify drug quantity as an element of the substantive offense of possession with intent to distribute; quantity is instead relevant to the penalty provisions of section 841(b), and is a matter for the district court at sentencing.” United States v. Sotelo-Rivera, 931 F.2d 1317, 1319 (9th Cir. 1991), cert. denied, — U.S. -, 112 S.Ct. 1186, 117 L.Ed.2d 428 (1992); see also 9th Cir.Crim. Jury Instr. 9.04P (1992) (“The government is not required to prove that the amount or quantity of [a controlled substance] was as charged in the indictment.”). The jury was properly instructed that proof of the precise quantity charged was not essential and that it need only find “some substantially similar quantity.” Thus Row-*1418en’s conviction did not depend on the additional ounce displayed.
XIX. Sufficiency of the Evidence
Appellants Dwain Baker, John Bonnenfant, Robert Rowen, Richard Rupley, Jr., and Byron Wimberly claim that the evidence against them was insufficient to support their convictions on various counts.24 Evidence is sufficient to support a conviction if, “ ‘reviewing the evidence in the light most favorable to the prosecution, any rational trier of fact could have found the essential elements of the crime beyond a reasonable doubt.’ ” United States v. Bishop, 959 F.2d 820, 829 (9th Cir.1992) (quoting Jackson v. Virginia, 443 U.S. 307, 319, 99 S.Ct. 2781, 2789, 61 L.Ed.2d 560 (1979)). Circumstantial evidence and reasonable inferences are sufficient to sustain a conviction. United States v. Reyes-Alvarado, 963 F.2d 1184, 1188 (9th Cir.), cert. denied, — U.S. -, 113 S.Ct. 258, 121 L.Ed.2d 189 (1992).
A.Dwain Baker
Dwain Baker argues that his CCE conviction must be reversed because there was insufficient evidence that he organized, managed, or supervised at least five people in connection with his series of drug offenses. The government, however, presented evidence from which a rational juror could have found this element satisfied. Viewing the evidence in the light most favorable to the prosecution, Dwain Baker recruited and used Kevin Aeheson as a purchaser of chemicals and a distributor of methamphetamine; directed Darrel Fitch to pick up three to four pounds of methamphetamine and place it in Baker’s ear, and instructed Fitch regarding packaging of newly manufactured methamphetamine; provided Ed George with methamphetamine to sell on behalf of the Company and instructed George to obtain chemicals and manufacturing equipment; ordered Jon Grant to transport five to ten pounds of methamphetamine to San Diego in the air cleaners of Grant’s truck and accompanied Grant on the trip; and instructed Mike Shoaf to transport nine pounds of methamphetamine in Baker’s car, deliver it to Baker in San Diego, and dig up and deliver buried methamphetamine in advance of FBI raids.
B. John Bonnenfant
Count 6 charged Rupley, Sr., and Bonnenfant with distribution of approximately four pounds of methamphetamine on or about September 15, 1982. The referenced transaction is apparently the same as predicate act (k) of the CCE count, which alleges that on September 15, 1982, Rupley, Sr., distributed approximately four pounds of methamphetamine to Bonnenfant, and that Bonnenfant was arrested the following day in possession of the four pounds of methamphetamine. There is no evidence that Bon-nenfant further distributed the methamphetamine after receiving it. Bonnenfant correctly argues that the recipient of a drug delivery cannot be convicted of distribution. “The term ‘distribute’ means to deliver (other than by administering or dispensing) a controlled substance or a listed chemical.” 21 U.S.C. § 802(11) (1988); see United States v. Harold, 531 F.2d 704, 705 (5th Cir.1976) (per curiam) (“To distribute means to deliver.... [I]t does not mean to receive.”).
Citing Bonnenfant’s and Rupley, Sr.’s convictions in the methamphetamine conspiracy, the government contends that Bonnenfant’s count 6 conviction should be upheld based on Pinkerton liability. This argument fails because the government made no showing that the distribution from Rupley, Sr., to Bonnen-fant, apparently for personal use, was in furtherance of the methamphetamine conspiracy. Nor did the government charge this incident in the indictment as an overt act of the conspiracy. We therefore reverse Bon-nenfant’s conviction on count 6.
C. Robert Rowen
1. Count 27
Count 27 charged Robert Rowen with distributing /¿-ounce of methamphet*1419amine on or about September 1985. The count was apparently based on grand jury testimony in October 1985 by Wendel Den Hartog, the alleged recipient of the methamphetamine. The only evidence in support of this count at trial, however, was Den Har-tog’s testimony that he received /4-ounce to 2 ounces of methamphetamine from Rowen “every week to two weeks” from late 1983 to the fall of 1986. Rowen argues that the lack of evidence regarding any specific transaction during September 1985 requires reversal of the count 27 conviction.
The jury was instructed that the government was not required to prove the precise dates or quantities charged in the indictment, but only “a date reasonably near the date alleged” or a “substantially similar quantity.” See United States v. Auerbach, 913 F.2d 407, 414 (7th Cir.1990) (evidence was sufficient for marijuana possession conviction despite variance between dates alleged in indictment and proved at trial). A rational juror who believed Den Hartog’s testimony reasonably could have inferred that the transaction took place as alleged. See United States v. Lai, 944 F.2d 1434, 1440 (9th Cir.1991) (uncorroborated testimony of accomplice is enough to sustain conviction), cert. denied, — U.S. -, 112 S.Ct. 947, 117 L.Ed.2d 116 (1992).
2. Count 30
Count 30 charged Rowen with possessing with the intent to distribute approximately one pound of methamphetamine on or about March 1, 1986. Rowen argues that there was insufficient evidence to tie him to the methamphetamine seized from the Wells Avenue search and to support his conviction for possession of a pound or “substantially similar quantity.” Our review of the record reveals sufficient evidence, even without relying on constructive possession or vicarious Pinkerton liability, to support the jury’s verdict. Lorna McClary, the owner of the Wells Avenue house, testified that she had 5 to 8 ounces of “Bobby [RowenJ’s methamphetamine” in her house when it was burglarized in January 1986, and that Rowen replaced the lost amount with 56-pound of methamphetamine 10 days to two weeks later. These amounts add up to Rowen’s possession of approximately one pound of methamphetamine on a date reasonably near the date alleged in the indictment. Moreover, the $6,136 in cash recovered from Rowen upon his arrest could reasonably suggest a recent sale of several additional ounces.
D. Rickard Rupley, Jr.
Rupley, Jr., argues that his conviction on the methamphetamine conspiracy count, for which he was charged and tried as an adult, must be reversed because there was insufficient evidence of his involvement in the conspiracy after age 18. See United States v. Cruz, 805 F.2d 1464, 1475-76 (11th Cir.1986) (Juvenile Delinquency Act does not apply to defendant who became involved in conspiracy while a minor but whose conspiracy activity continued after he turned 18), cert. denied, 481 U.S. 1006, 107 S.Ct. 1631, 95 L.Ed.2d 204, 482 U.S. 930, 107 S.Ct. 3215, 96 L.Ed.2d 702 (1987). After Rupley, Jr., turned 18, he relayed messages relating to negotiations for a methamphetamine ingredient between his imprisoned father and a coconspirator. There was also testimony that, while in prison, Rupley, Jr., participated in a meeting with his codefendants to discuss killing the FBI Agent in charge of the case. A rational juror could have concluded that Rupley, Jr., participated in the conspiracy after his eighteenth birthday. We therefore affirm Rupley, Jr.’s count 2 conviction.
E. Byron Wimberly
Byron Wimberly’s convictions on counts 23 (possession with the intent to distribute methamphetamine) and 24 (interstate transportation in aid of racketeering) both stem from Wimberly’s presence on a midnight charter flight from San Diego to Reno transporting substantial amounts of cash and methamphetamine. Wimberly was accompanied by Rupley, Sr., and government witness Ed George, and all three used false identities. Even apart from Pinkerton liability, which provides a separate and sufficient basis for upholding Wimberly’s convictions, a rational juror could have concluded from George’s testimony and the surreptitious nature of the flight that Wimberly had the *1420requisite knowledge of the contraband and was thus guilty of the charged offenses.
XX. Sentencing
A. Disproportionate Sentencing
Appellants allege that the district court abused its discretion in sentencing them to lengthy prison terms while rewarding codefendants who cooperated with the government with lenient sentences or dismissal of charges. The offenses in this case were committed prior to November 1, 1987, the effective date of the Sentencing Guidelines. Prior to the Guidelines’ enactment, the district court had “virtually unfettered discretion in imposing sentence.” United States v. Barker, 771 F.2d 1362, 1364 (9th Cir.1985). A pre-Guidelines sentence that falls within statutory limits is generally not reviewable unless there are constitutional concerns. United States v. Citro, 842 F.2d 1149, 1153 (9th Cir.), cert. denied, 488 U.S. 866, 109 S.Ct. 170, 102 L.Ed.2d 140 (1988).
Our review of the sentencing transcripts makes clear that the district court properly assessed each defendant’s culpability individually and attempted to fashion sentences that were proportionate to those of other defendants. The court repeatedly recognized that defendants should not be penalized for going to trial and stated its goal of proportionate sentences. Whatever disproportionality did exist between Appellants and their cooperating codefendants was within the court’s discretion. Most of the Appellants were convicted of multiple counts and more serious offenses than those who pled guilty. Furthermore, lenient sentencing in exchange for cooperation with the government is a valid exercise of discretion. United States v. Brown, 761 F.2d 1272, 1278 (9th Cir.1985). Finally, the mere fact of disparate sentences does not indicate an abuse of discretion because a sentencing court will generally have access to considerably more information after a trial than after a guilty plea, particularly in a trial of this magnitude. See Alabama v. Smith, 490 U.S. 794, 801, 109 S.Ct. 2201, 2205, 104 L.Ed.2d 865 (1989) (“[I]n the course of the proof at trial the judge may gather a fuller appreciation of the nature and extent of the crimes charged.”).
B. Enhancement for Prior Conviction
Appellant Cavallaro received a 25-year sentence for his count 2 (methamphetamine conspiracy) conviction. The applicable statute imposes a maximum sentence of 20 years but provides for an enhancement of up to 10 years if the offense is committed “after one or more prior convictions ... for a felony under any ... law of a State ... relating to narcotic drugs ... have become final.” 21 U.S.C. § 841(b)(1)(C) (1988). Cavallaro had been convicted in California state court for narcotics offenses arising out of the same facts as the count 3 marijuana conspiracy charge, of which he was also found guilty. Cavallaro argues that the sentence enhancement was improper because (1) the state convictions were related to the charges in the federal trial, and (2) his entire involvement in the methamphetamine conspiracy predated the finality of his state convictions. We reject these contentions.
First, that the federal and state charges derive in part from the same activity does not preclude using the state convictions to enhance the federal sentence. Section 841(b)(1)(C) imposes no requirement that the convictions be totally unrelated; as the district court observed, “the only limitations placed on the use of a prior conviction are that it be drug related and that it be final.” In any event, although the state and federal charges involved some overlapping evidence, the dates, events, and locations involved in the federal trial covered a much broader range of criminal conduct than the state convictions.
Second, Cavallaro’s federal offense was not completed until after his state felony convictions became final. Cavallaro’s state convictions became final on June 16, 1986. The jury’s verdict in this case represents a finding that the methamphetamine conspiracy continued from December 1981 to September 1987, as charged in the indictment. See United States v. Calabrese, 825 F.2d 1342, 1346 (9th Cir.1987). Testimony at trial indicated that, subsequent to June 16, 1986, Ca-vallaro participated in a jailhouse conversation with other Company members regarding *1421the murder of the FBI Agent in charge of the case and conspired with Rupley, Sr., and Edward Baker to murder government witness Darrel Fitch. Thus there was sufficient evidence that Cavallaro participated in the conspiracy during the entire time period alleged in the indictment, which extended beyond the date of the state convictions. The district court properly enhanced Cavallaro’s sentence.
CONCLUSION
We reverse the convictions of Richard Rupley, Jr., on counts 3 and 10, and of John Bonnefant on count 6. The sentences on these counts were to run concurrently with sentences of equal or greater length for the counts on which Rupley, Jr. and Bonnefant were properly convicted. Further, nothing in the record indicates that the counts whose convictions we reverse enhanced the sentences on the other counts. We therefore see no need to remand for resentencing. United States v. Ray, 731 F.2d 1361, 1368 (9th Cir.1984).
Insofar as their offenses were committed prior to November 1, 1987, Rupley, Jr. and Bonnefant retain the option of filing a motion for a reduction of sentence, pursuant to then Fed.R.Crim.P. 35(b). This Rule “permits the district judge to reduce a sentence sua sponte or on motion of the defendant for up to 120 days after the underlying conviction and sentence have been affirmed on appeal.” United States v. Soto, 793 F.2d 217 (9th Cir.1986) (amending 779 F.2d 558 (9th Cir.1986)), cert. denied, 484 U.S. 833, 108 S.Ct. 110, 98 L.Ed.2d 70 (1987).
We affirm the convictions and sentences of all other Appellants.
AFFIRMED IN PART, REVERSED IN PART.
. Although Appellants allege that the government invented the "Company” label to prejudice them, defense counsel repeatedly used the term in their questions at trial, and several of Appellants' counsel used the term at oral argument before this court.
. Astonishingly, this massive trial was prosecuted by a single Assistant United States Attorney with no assistant counsel. While we admire the prosecutor’s endurance and intestinal fortitude, we believe such conservative staffing to be an exception rather than the norm.
. There was an overabundance of briefing in this case. We remind counsel that, although they play a partisan role, they are officers of the court with an obligation to assist in the orderly and efficient search for the truth. Counsel serve neither their clients nor this court by burying the debatable issues in this appeal among numerous implausible and unsupported arguments, many based on grounds not objected to at trial and several asserting the rights of individuals who are not even parties to this appeal. This criticism is not addressed to all appellate counsel, as many performed their duties admirably.
. Although the government is not required to furnish the defendant with a list of witnesses in a non-capital case, United States v. Sukumolachan, 610 F.2d 685, 688 (9th Cir.1980), the discovery order in this case required the government to produce such a list before trial.
. The JDA applies to persons under 21 who have committed an act of juvenile delinquency. 18 U.S.C. § 5031 (1988). An act of juvenile delinquency is defined as “the violation of a law of the United States committed by a person prior to his eighteenth birthday which would have been a crime if committed by an adult.” Id. A successful prosecution under the JDA does not result in a criminal conviction but rather in an adjudication of status as a juvenile delinquent. United States v. Juvenile Male, 864 F.2d 641, 643 (9th Cir.1988).
. Dismissal with prejudice might be within the district court’s discretion for violations that are “aggravated or intentional,” Montalvo-Murillo, 495 U.S. at 721, 110 S.Ct. at 2079, or “egregious," Doe, 862 F.2d at 780. Civil remedies may also be available. Because the issue is not before us, we express no opinion on the appropriateness of any of these remedies.
. United States v. Andy, 549 F.2d 1281 (9th Cir.1977), holds that the 30-day period of § 5036 runs from the earlier of “(1) the date that the Attorney General certifies, or in the exercise of reasonable diligence, could have certified, to the conditions stated in Section 5032, or (2) the date upon which the Government formally assumes jurisdiction over the juvenile.” Id. at 1283. This holding has been criticized as inconsistent with the plain language of § 5036. United States v. Doe, 882 F.2d 926, 928-29 n. 3 (5th Cir.1989); United States v. Doe, 642 F.2d 1206, 1207-08 (10th Cir.), cert. denied, 454 U.S. 817, 102 S.Ct. 94, 70 L.Ed.2d 86 (1981); United States v. Sechrist, 640 F.2d 81, 83-85 (7th Cir.1981). Because the dates set forth in the Andy standard did not precede the date of Rupley, Jr.'s detention, Andy does not affect the speedy trial analysis.
.The gist of Appellants’ claim of prejudice appears to be that the result would have been different not had they had more time to prepare, but rather had the district court chosen to sanction all alleged discovery violations by suppressing the evidence. They argue that had severe sanctions been imposed, "a large percentage of the prosecution evidence would not have reached the jury” and "the result of the trial would probably have been substantially different." This argument is misguided. The prejudice that must be shown to justify reversal for a discovery violation is a likelihood that the verdict would have been different had the government complied with the discovery rules, not had the evidence been suppressed. See United States v. Walker, 538 F.2d 266, 268-69 (9th Cir.1976).
. Shortly after the trial began, the district court stated that "the Government has bent over backwards to make information available.” Approximately one year later, the court noted that "the Government has gone out of its way to accommodate the defense, has gone far beyond what I have seen in cases here.”
. On at least two occasions, the district court suppressed evidence favorable to the government because the government failed to produce the evidence in a timely manner.
. Appellants suggest that the alleged high security risk was belied by the government’s failure to show that the defendants had been disruptive in any previous court appearances. The decision to physically restrain a defendant, however, is not governed by conduct in previous court proceedings. Jones, 899 F.2d at 885 n. 2.
. The Ninth Circuit cases cited by Appellants are all pre-i?oss and therefore no longer good law. See, e.g., United States v. Claiborne, 765 F.2d 784, 799 (9th Cir.1985) (holding that the Sixth Amendment is violated by "forc[ing] defendants to exhaust their peremptory challenges on persons who should be excused for cause”), cert. denied, 475 U.S. 1120, 106 S.Ct. 1636, 90 L.Ed.2d 182 (1986).
. Appellants raise a veiled argument of prosecu-tonal misconduct based on the prosecutor's "intentional" misstatement of Kitts' prior testimony. There is no reason to believe that the prosecutor's question represented anything more than a mistake. See Doan v. United States, 202 F.2d 674, 680-81 (9th Cir.1953).
. We express no opinion on the holdings of these cases.
. Although there are some situations in which a CCE codefendant may be counted among the requisite "five or more persons” supervised, satisfaction of the “in concert” element in such a case is based on the defendant’s role as subordinate with respect to the particular predicate offense, not his status as a CCE codefendant.
. The question whether one may be convicted under § 848 for aiding and abetting a kingpin (without independently satisfying the statutory requirements) is separate from the question whether aiding and abetting Title 21 offenses may serve as series predicates for a CCE conviction. We have answered the second question affirmatively, and the Second Circuit shares our view. Miskinis, 966 F.2d at 1268; United States v. Aiello, 864 F.2d 257, 264 (2d Cir.1988).
. The requisite five or more persons need not act in concert with each other, nor need they be involved with the organizer at the same time. United States v. Jerome, 942 F.2d 1328, 1330 (9th Cir.1991).
. Rupley, Sr., and the Bakers were charged in and convicted of all three aiding and abetting counts; Bonnenfant was convicted of two. These convictions could serve as CCE predicate violations. Miskinis, 966 F.2d at 1267-68.
. The district court’s instructions defined an "organizer” as "a person who puts together a number of people engaged in separate activities and arranges them in their activities in one operation or enterprise.” The court defined a "su*1410pervisor" as "one who manages or directs or oversees the activities of others.”
. The scope of Agent Besse’s testimony extended beyond the four CCE defendants. For example, the marijuana summaries were admitted against all defendants charged in the marijuana conspiracy, and the methamphetamine summaries were admitted against all defendants. Furthermore, many entries included in the summary printout had no value assigned to them because no quantities were available, and were therefore irrelevant to proving substantial proceeds.
. The district court properly warned the defense that attacks on the credibility of the witnesses who testified to the underlying data would open the door for the prosecution to elicit Agent Besse’s opinions on credibility. See Fed.R.Evid. 608(a). The court, however, suggested that the defense could point out inconsistencies and contradictions in the underlying data as a means of impeaching the summary.
. The language of 21 U.S.C. § 846, unlike the general federal conspiracy statute, 18 U.S.C. § 371, docs not expressly require proof of an overt act. Although every other circuit has held that an overt act is not an element of a § 846 drug conspiracy, this court continues to differ. See United States v. Shabani, 993 F.2d 1419, 1420-21 & n. 1 (9th Cir.1993).
. The district court concluded that Esposito's statements would not be admissible as statements against interest because corroborating circumstances did not clearly indicate the trustworthiness of the statements. See Fed.R.Evid. 804(b)(3). This ruling was not an abuse of discretion. See United States v. Smith, 924 F.2d 889, 895 (9th Cir.1991).
. Bonnenfant does not frame his argument as a sufficiency of the evidence dispute, but as a challenge to the district court’s denial of his Rule 29 motion for acquittal on count 6 at the close of the prosccution's case. The standards of review are the same. See United States v. Shirley, 884 F.2d 1130, 1134 (9th Cir.1989).