WALD, Circuit Judge:

A jury convicted Dennis S. Davis and Bobby A. Holton of crimes involving the unlawful possession and distribution of crack cocaine, conspiracy and the unlawful use of a communications facility. On appeal, both appellants challenge their convictions on the basis of alleged errors made by the district court when it (1) permitted the jurors to consider government — prepared transcripts of drug transactions during deliberations; (2) replayed recorded drug transactions for the jury during deliberations without the defendants being present; (3) refused to voir dire the jury about possible prejudice stemming from a television news program; and (4) determined that handwritten notes made by a government witness were not producible Jencks Act material. In addition, appellants contend that the Fifth Amendment requires a remand for resentencing under the guidelines applicable to individuals who have committed crimes involving powder cocaine. Finding none of these claims meritorious, we affirm the convictions and sentences.

I. Background

In October 1991, a confidential informant advised the Metropolitan Police Department ("MPD") that individuals were selling drugs at the Barry Farms housing complex in southeast Washington, D.C. The MPD launched an undercover investigation in which MPD Detective Michael J. Quander posed as the brother of the confidential informant in order to gain an introduction to the drug dealers. Detective Quander made six trips to the Barry Farms complex during which he purchased approximately 11 ounces of crack cocaine. Following the last purchase, the police executed search warrants at two homes in the complex and seized additional drugs and drug paraphernalia. They arrested appellants Dennis S. Davis, Bobby A. Holton and several others.

The government's evidence against appellants included body — wire tape recordings of conversations between Detective Quander and the individuals with whom he engaged in drug transactions. A trial was held at which the body — wire tapes were admitted into evidence but, based on an agreement between the parties, were never played to the jury. Appellants were convicted of all the charges against them. On appeal, this court vacated the convictions against appellants and remanded the case for a new trial.

At the second trial, the government announced its intention to play portions of the body — wire recordings to the jury. Defense counsel unsuccessfully moved to have the tapes excluded because of their poor quality. Just prior to calling Detective Quander to the witness stand, the government informed the court that it had not yet prepared transcripts of the tape recordings. The trial was recessed so that the government and defense counsel could listen to the tape recordings and the government could prepare transcriptions of the recordings. When presented with the government's transcripts, defendants objected to their accuracy, arguing that the tapes were inaudible in many parts and that, because defendants denied involvement in the drug transactions, the attribution of statements to them would be prejudicial. Appellant Davis' counsel proffered an alternative version of the transcript that eliminated the attributions and substituted "inaudible" for some of the statements. At the government's request, the court listened to the tape recordings while reading along with the government — prepared transcripts. The court heard objections from the defendants and ruled that the government transcripts of the recordings would aid the jury in listening to the tapes during the trial, but that the transcripts themselves would not be given to the jury during its deliberations.

The tapes were played and authenticated during the examination of Detective Quander. The government — prepared transcriptions were distributed to the jury before the tapes were played and were collected directly afterwards. On cross-examination of the detective, it was revealed that Detective Quander had prepared handwritten notes or transcriptions of the tape recordings. Defense counsel made a Jencks Act request for the written notes, but the district court concluded that the documents did not constitute Jencks material and refused to hold a hearing. At defense counsel's request, however, the court agreed to examine the documents in camera and ordered that the government file the handwritten notes under seal so that they could be examined on appeal.

The evening after the defendants began presenting their cases, ABC News showed a Nightline program concerning the disparate sentences imposed for crimes involving powder cocaine and those involving cocaine base or crack cocaine. Among the individuals interviewed on the program were District of Columbia Superior Court Judge Reggie Walton and Michelle Roberts, a criminal defense attorney who practices in the District of Columbia. Judge Walton talked about patterns of violent behavior related to the crack cocaine trade and speculated that high mandatory sentences may have contributed to the lowering of the homicide rate in Washington, D.C. Ms. Roberts criticized the sentencing structure for crimes involving crack cocaine as lacking deterrent value. When trial resumed the following morning, counsel for appellant Holton requested that the district judge voir dire the jury in order to determine whether any of the jurors had viewed the television program and, if so, what impact it might have on their deliberations. The district court refused to question the jurors because they had been probed about their ability to be fair and impartial during jury selection and frequently had been admonished not to decide the case based on anything heard outside of the courtroom.

After deliberations began, the court received notes from the jury asking to hear all of the tape recordings. The district judge consulted with counsel and arranged a procedure whereby his law clerk would play the tapes for the jury in the courtroom — which the judge described as "an extension of the jury room" — and distribute the transcripts as listening aids, to be collected at the completion of the playing of the tapes. Defense counsel objected to the jurors receiving the transcripts during deliberations because the transcripts had not been admitted into evidence, were not accurate and because the court had stated that the jury would not be given the transcripts during deliberations.

Appellant Holton was convicted of one count of Conspiracy to Distribute Cocaine Base in violation of 21 U.S.C. Section(s) 846 and 841, four counts of Unlawful Distribution of Cocaine Base Within 1000 Feet of a Playground and School in violation of 21 U.S.C. Section(s) 860(a) and 18 U.S.C. Section(s) 2, and one count of Unlawful Use of a Communication Facility in violation of 21 U.S.C. Section(s) 843(b). Appellant Davis was convicted of one count of Conspiracy to Distribute Cocaine Base in violation of 21 U.S.C. Section(s) 846 and 841, six counts of Unlawful Distribution of Cocaine Base Within 1000 Feet of a Playground and School in violation of 21 U.S.C. Section(s) 860(a) and 18 U.S.C. Section(s) 2, and one count of Unlawful Use of a Communication Facility in violation of 21 U.S.C. Section(s) 843(b). The court sentenced appellant Holton to concurrent terms of 363 months imprisonment for the conspiracy and distribution offenses and 48 months imprisonment for the communication facility offense, and appellant Davis to concurrent terms of 370 months imprisonment for the conspiracy and distribution offenses and 48 months imprisonment for the communication facility offense. These consolidated appeals ensued.

II. Discussion [11] A. Use of Transcripts and Replaying of Tapes During Jury Deliberations [12] 1. Transcripts

A transcript repeating in written form a conversation recorded on tape may help a juror listening to the tape follow the conversation when the tape is of questionable clarity, see United States v. West, 948 F.2d 1042, 1044 n. 1 (6th Cir. 1991); United States v. Slade, 627 F.2d 293, 302 (D.C. Cir. 1980), or contains the voices of multiple speakers who talk over each other or speak in quick succession. See Slade, 627 F.2d at 302. Ironically, the same circumstances that make a transcript helpful to a juror may prejudice the defendant if it is presented without proper safeguards, for the only transcripts worth fighting about are those on which important words may be susceptible to different interpretations. After all, the jurors are likely to notice a clear discrepancy between a tape and a transcript.

The principal risk of indiscriminately permitting the use of transcripts by jurors is that in the case of a poor quality or unintelligible recording, the jurors may substitute the contents of the more accessible, printed dialogue for the sounds they cannot readily hear or distinguish on the tape and, in so doing, transform the transcript into independent evidence of the recorded statements. See United States v. Howard, 80 F.3d 1194, 1198 (7th Cir. 1996); United States v. Strothers, 77 F.3d 1389, 1392-93 (D.C. Cir.), cert. denied, 117 S. Ct. 374 (1996). A related risk arises when a transcript attributes incriminating statements to a defendant that the defendant does not admit making. See Howard, 80 F.3d at 1200 (possibility of prejudice when transcript attributes statements to defendants who do not testify and there is no evidentiary basis for the attribution); see United States v. Berry, 92 F.3d 597, 600-01 (7th Cir. 1996) (possibility of prejudice when identification is central issue in case and transcript identifies defendant as speaker). Placing a transcript in the jury room during deliberations — after the completion of the supervised, adversarial portion of the trial — opens up the possibility that jurors will see the transcript as a neutral exhibit placed before them by the court and increases the chance that the document will be read without the tape recording playing alongside for the purpose of comparison. See Strothers, 77 F.3d at 1392-93 (error to admit transcript of tape into evidence because of the risk that the jury might rely on the government's version of the conversations during deliberations without simultaneously listening to the authenticated tapes to verify the transcript's accuracy).

These risks can be markedly decreased when certain precautions are taken. Thus, we have upheld the practice of giving the jury a transcript to help follow a tape recording played during trial so long as procedures are followed to ensure that the jury does not rely on one party's version of the transcript instead of the tape recording. See Slade, 627 F.2d at 302 (citing United States v. McMillan, 508 F.2d 101, 105 (8th Cir. 1974)). We have approved of two alternative methods to ensure that a transcript is a sufficiently accurate rendition of the conversation recorded on tape to be used by the jury as a listening aid during trial and a third procedure to be used when the accuracy of the transcript has not been verified. Id. In a criminal case, the preferred approach is for the prosecution and defense to stipulate as to the accuracy of a transcript. If the parties cannot agree, however, the trial court can make a pretrial determination of accuracy after comparing the proposed transcript against the tapes or provide the jury with one transcript reflecting the prosecution's interpretation of the recording and one version reflecting the defendant's interpretation. Id. With respect to any one of the three procedures, we require the district court to "instruct the jurors that their personal understanding of the tape supersedes the text in a transcript." Id. The transcript then may be used in conjunction with the playing of the tape.

We have not previously adopted any rule regarding the use of the transcript by the jury during its deliberations or indeed for the admission of a transcript into evidence so that, like any other piece of evidence, it can be brought into the jury deliberations. See Dallago v. United States, 427 F.2d 546, 552-54 (D.C. Cir. 1969) (jurors ordinarily are entitled to see any exhibits entered into evidence). In Slade, we upheld the district court's discretion to allow the jury to use an accurate transcript to assist them in listening to a tape, but it is not clear from the opinion whether use of the transcript in that case was limited to trial; the tapes were played both during trial and deliberations and the jury's use of the transcripts as a guide was cited in the opinion as a basis for finding that the tapes were not so marred by background noise so as to render them unintelligible. Slade, 627 F.2d at 302. In Strothers, we found error when government-prepared transcripts were admitted into evidence and could be read by the jurors without listening to the tapes, but we did not rule out the admission of a transcript into evidence altogether. Certainly a transcript can be helpful when the jury listens to replays of tape recordings during deliberations as well as during trial. Indeed, it might well confuse the jurors to permit them to use transcripts as a guide during trial, but not during deliberations. Moreover, without a transcript to guide them, the jury could find itself involved in repetitious and time consuming replaying of unintelligible recordings. Thus, we agree with the other circuits that have held it is within the district court's discretion to permit the jury to use transcripts during deliberations. See, e.g., United States v. Young, 105 F.3d 1, 10-11 (1st Cir. 1997); United States v. Delpit, 94 F.3d 1134 (8th Cir. 1996); United States v. Elder, 90 F.3d 1110, 1129 (6th Cir.), cert. denied, 117 S. Ct. 529 (1996).

Our survey of the practices of other circuits indicates that permissible procedures regarding transcripts vary widely. Some circuits have permitted transcripts to be brought into deliberations without requiring that they be formally admitted into evidence. See, e.g., United States v. Capers, 61 F.3d 1100, 1107 (4th Cir. 1995), cert. denied, 116 S. Ct. 1830 (1996); United States v. Bryant, 480 F.2d 785, 791 (2d Cir. 1973). Other circuits specifically have held that a transcript, disputed or not, may be admitted into evidence. See United States v. Onori, 535 F.2d 938 (5th Cir. 1976) (transcript used as a guide is analogous to the use of expert testimony as a device aiding the jury in understanding other types of real evidence); United States v. Hall, 342 F.2d 849 (4th Cir. 1965) (transcript similar to photograph, drawing, or mechanical model used by a witness to amplify testimony). Every circuit agrees, however, that, while a jury may draw inferences of guilt from testimony or from a tape recording, it should not draw such inferences from a transcript. The transcript is viewed by our circuit and others as an illustration of a witness's interpretation of the recording. The Advisory Committee Notes to Rule 401 of the Federal Rules of Evidence explain that "relevant evidence" permits the use of evidence that is "admitted as an aid to understanding." Fed.R.Evid. 401 advisory committee's note. In addition, the court has the power to "exercise reasonable control over the mode and order of . . . presenting evidence so as to (1) make the interrogation and presentation effective for ascertainment of the truth, [and] (2) avoid needless consumption of time. . . ." Fed.R.Evid. 611(a). Transcripts admitted for the limited purpose of being used as a jury aid can help prevent jury confusion and wasted time as a tape is being played. Appellants argue that providing transcripts to the jury during deliberations always is prejudicial because it unduly emphasizes those pieces of evidence. Although "double exposure to transcripts of crucial evidence, i.e., during the trial and in the jury room, can be prejudicial to a defendant," United States v. Rengifo, 789 F.2d 975, 982 (1st Cir. 1986), the district court retains ample discretion to exclude transcripts in circumstances where the prejudice might outweigh their usefulness as an aid. Id.; see also Fountain v. United States, 384 F.2d 624 (5th Cir. 1967); see Fed.R.Evid. 403. Moreover, Federal Rule of Evidence 611 would permit a judge broad discretion to exclude a transcript, particularly if it will not add anything to a juror's understanding of a tape recording.

To protect jury deliberations from improper influence, we ordinarily restrict the jury's access only to exhibits that have been accepted into evidence and "consideration by the jury of documents not in evidence is error. . . ." United States v. Treadwell, 760 F.2d 327, 339 (D.C. Cir. 1985) (citing Dallago, 427 F.2d at 553). Although it makes little practical difference if a transcript is formally admitted into evidence, so long as its transmission to the jury is accompanied by appropriate instructions limiting its use to that of an aid, we believe that it is the better practice to require formal admission into evidence during the trial so that parties objecting to the transcript will be apprised of the possibility that the document could be given to the jury during deliberations and so that the judge will be sure to provide the jury with appropriate instructions as to its limited use on admission. Formal admission would reduce the risk of surprising counsel by a last minute decision to send the transcript into the jury room if the jury requests a replaying of the tapes.

Our precedents indicate that it is permissible to present the jury either with an "accurate" transcript or two alternative transcripts. We believe that the best practice remains for the parties to devise a stipulated transcript. See Onori, 535 F.2d at 948. On many occasions a defendant may object to the accuracy of a transcript and argue that the underlying tape recording is too unintelligible to be effectively challenged. If there is a general objection to the accuracy of the transcript, but no alternative transcript is offered, then the judge may review the transcript against the tape and whatever other evidence is presented and certify the transcript's accuracy, but the jury must be informed that the transcript is only one party's version. If a party makes specific objections to the transcript, or offers an alternative transcript, then "the jury is entitled to consider the divergence in two transcripts of the same conversation, with the recording of it, as a problem of fact to be resolved in the traditional manner." Howard, 80 F.3d at 1199. The jury may be given one transcript containing both versions of the disputed portions or two separate transcripts. The parties should each be given an opportunity to put on evidence supporting the accuracy of its version or challenging the accuracy of the other side's version. Id. No matter which of these procedures is utilized, the jury should be instructed that the tape recording constitutes evidence of the recorded conversations and the transcript is an interpretation of the tape. The jury must be instructed that they should disregard anything in the transcript that they do not hear on the recording itself. Moreover, the court must ensure that the transcript is used only in conjunction with the tape recording.

To sum up: The district court, in the exercise of its discretion, should decide whether transcripts should be admitted into evidence as an aid to the jury. It is within the district court's discretion to decide whether, upon request of the jury to hear tape replays, it will permit the jury to use the transcripts during deliberations. The court should reach a decision in this respect sufficiently early in the trial to allow defense counsel an opportunity to challenge the accuracy of any government-prepared transcript, to support the accuracy of any alternative transcript that the defendant wishes to introduce, to raise objections, and to request appropriate jury instructions. Upon an objection of one of the parties, and failing agreement between the parties on a stipulated transcript, each party should be permitted an opportunity to introduce its own transcript or the parties should create a single transcript that shows both versions at the points of disagreement. When the jury receives two transcripts of the same recording, it should be instructed that there is a difference of opinion as to the accuracy of the transcripts and that it is up to them to decide which, if any, version to accept. The jurors also should be instructed that they should disregard any portion of the transcript or transcripts that they think differs from what they hear on the tape recording. Further limiting instructions will depend on the circumstances of the case.

Although we recognize that the district court could not have anticipated the rule we lay down today, we find that the court failed to meet the standards previously set forth in Slade for the use of a transcript during trial. A careful reading of the transcript reveals that the trial judge acknowledged that the tapes were sufficiently audible and intelligible to be played to the jury and that the transcripts would aid the jury in listening to the tapes, but that he never explicitly found that the transcripts accurately reflected statements recorded on the tapes or that the attributions in the transcripts were accurate.

The trial judge did indeed listen to each tape recording while he read the government-prepared transcripts. After playing the October 4, 1991 tape for the judge, the government stated "I would suggest that the court can now find that not only are the excerpts that the government intends to play audible and intelligible, but that the material on the tape that the government intends to play is highly probative. . . ." After making two changes to the transcript of the October 4, 1991 tape, the government stated that it "stands by its transcript 100 percent." The court ruled "I am satisfied and make essentially the findings that are articulated by [the government] with respect to the October 4th transcript." Because the only proposed findings articulated by the government related to the quality of the tape, not the transcript, we are unable to conclude that the judge verified the accuracy of the transcript. After playing the October 16, 1991 tape, the government stated its belief that its transcript "is the appropriate transcript to aid the jury in playing these tapes." The court agreed but did not state that the transcript was accurate. The court appears to have made the same "finding" with respect to the transcript of the October 23, 1991 recording and the transcript of the October 30, 1991, recording. The court never ruled on any of the specific objections that were raised by defense counsel.

Despite our skepticism, it is possible that the court's "findings" could be interpreted as verifying the accuracy of the transcripts. There was no way, however, for the judge to know whether the attribution of certain voices to certain defendants was accurate; Detective Quander had not yet testified as to the identity of the speakers and the defendants had not testified. Nor did the district judge condition his accuracy "finding" on subsequent proof that the attributions were correct or ever revisit the issue. We must conclude, therefore, that the district judge did not follow any of the procedures described in Slade that must precede the jurors being given the transcripts.

The immediate question then is whether the court's error in placing the transcripts before the jury during deliberations requires reversal of the convictions. As we discussed above, we restrict the jury's access to information in order to protect the jury's deliberations from improper influence. If there was an intrusion in the jury room that did not result in harm, then reversal would be pointless. See United States v. Olano, 507 U.S. 725, 738 (1993). In examining the possibility of prejudice, we consider whether we can "reasonably conclude that the judgment was not substantially swayed by the error." Treadwell, 760 F.2d at 339.

The facts of this case do not suggest that the court should presume prejudice or that there is a reasonable probability that the jury was prejudiced by the transcripts. The record "provides substantial support for the relative accuracy of the transcripts." Slade, 627 F.2d at 303. During trial, Detective Quander, a party to the taped conversations, corroborated the identity of the parties who had been recorded. Officer Curtis, who observed the October 23rd and 30th transactions, corroborated Detective Quander's testimony as to the identity of the participants on those particular days. See Dallago, 427 F.2d at 559. In addition, the information on the tapes was only a portion of a larger set of facts that the prosecution put before the jury through proper means. Cf. Treadwell, 760 F.2d at 340. That evidence included testimony that appellant Davis attempted to flee from the bathroom of a house at 1361 Stevens Road when police executed a search warrant at that location; the recovery of ziploc bags containing nearly 32 grams of crack cocaine from the toilet in that bathroom; a plate and razor blade next to the sink in the bathroom; the recovery of narcotics packaging paraphernalia from the house; the recovery from Davis of $3000 in prerecorded funds that Detective Quander had earlier used to pay for a purchase of cocaine; the arrest of Holton beside the open driver's door of an Acura automobile in which a cellular telephone used in the drug transactions was found; and the recovery of $559 from Holton.

Appellants insist that the court's "about-face" with respect to the jury having the transcripts during deliberations compromised defense counsel's ability to highlight alleged inaccuracies or present alternative transcripts, but the facts simply do not bear out such a claim. Defense counsel seized the opportunity to challenge the preparation and accuracy of the government-prepared transcripts during their cross-examination of Detective Quander and during their closing arguments. Thus, "the jury was made aware that the transcripts offered only the government's interpretations" of the tape recordings, Slade, 627 F.2d at 303, and that it had to resolve an issue as to the identity of the speakers on the tape in light of the evidence introduced at trial. See Strothers, 77 F.3d at 1393.

The district court's frequent limiting instructions also served to avert the possibility of prejudice. The court instructed the jury prior to the distribution of each transcript and at the close of trial that the transcripts were not to be considered independent evidence and that only the tape recordings constituted evidence of the conversations recorded therein. The jury was also provided with a written instruction as to the limited use to be made of the transcripts and was referred to that instruction by the judge when he informed the jury that arrangements had been made for them to hear the tapes. See United States v. Crowder, 36 F.3d

Finally, it is clear that the jury did not use the transcripts as a substitute for listening to the tapes in deliberations. The record indicates that the transcripts were made available only in conjunction with the relevant tape recordings and the record suggests that the jury did listen to the tapes being replayed during deliberations. Cf. United States v. Collazo, 732 F.2d 1200, 1204 (4th Cir. 1984). For all these reasons, we conclude that, although the trial court erred when it permitted the jury to consider the government-prepared transcripts without verifying their accuracy or, alternatively, admitting a defense version, there was no reasonable probability that the jury substituted the transcripts for the evidence in the record or that prejudice resulted.

Appellants argue that by allowing the transcripts to be seen and read by the jury, the district court violated the best evidence rule, which requires the production of an original recording, when available, to prove its contents. Fed.R.Evid. 1002. We note that the best evidence rule would not be implicated in circumstances in which the district court follows the proper procedure. "The elementary wisdom of the best evidence rule rests on the fact that the [recording itself] is a more reliable, complete and accurate source of information as to its contents and meaning than anyone's description [of it]." Gordon v. United States, 344 U.S. 414, 420 (1953). Thus, the rule is a mechanism to prevent fraud or mistransmission of information, i.e., to ensure accuracy. When the original tape is available and presented to the jury and the accuracy of the transcript has been stipulated or is made an issue for the jury to decide, concerns addressed by the best evidence rule are not at issue.

2. Replaying of Tapes

Appellants also challenge the trial court's decision to replay the tapes for the jury during deliberations when neither they nor their attorneys were present. They argue that they were denied their Sixth Amendment right to confront the evidence against them and their Fifth Amendment right to due process. The government asserts that defendants' counsel agreed with the district court's proposal that only his law clerk be present, but the trial transcript is decidedly vague on that point, probably because there were prior discussions off the record between the lawyers and the judge which are not a matter of record.

First, appellants speculate about problems that might have occurred during the replaying, which was conducted in the courtroom by the judge's law clerk according to procedures of which both parties were informed beforehand. But, there is in fact no evidence suggesting that the law clerk either made independent decisions about whether or how to replay tapes or remained in the courtroom while the jury was deliberating, except for the actual playing of the tapes. Under these circumstances, there is no reason to presume harm. See United States v. Sobamowo, 892 F.2d 90, 96 (D.C. Cir. 1989) (citing United States v. Kupau, 781 F.2d 740, 743 (9th Cir. 1986) (even where tape-playing government agent and jury were alone in the courtroom, error found harmless beyond reasonable doubt given "the obvious efforts of the court, by clearing the courtroom, to prevent any outside influences, and the absence of any suggestion that extraneous matters came before the jury")); see United States v. Florea, 541 F.2d 568, 570-71 (6th Cir. 1976); cf. Riley v. Deeds, 56 F.3d 1117, 1119-20 (9th Cir. 1995) (complete abdication of judicial control where law clerk independently granted jury's readback request and control over what was read to jury was in hands of jurors and clerk).

Second, appellants assert that Sobamowo requires supervision by defense counsel during the replaying of tapes. In that case, the court reasoned that the district court's procedure requiring the presence of the prosecution and at least one defense attorney rendered any potential error harmless. In that same case, however, we held that "tape replaying [for the jury during deliberations is] not a stage of trial implicating the confrontation clause or Rule 43(a)." Sobamowo, 892 F.2d at 96 (citing Dallago, 427 F.2d at 552-53 (defendant's presence not required when exhibits are submitted to the jury during deliberations)); but see United States v. FelixRodriguez, 22 F.3d 964, 967 (9th Cir. 1994) (defendant has a right under Rule 43 to be present when a tape-recorded conversation is replayed to a jury during its deliberations). The Sobamowo court also held that a defendant's absence during replaying does not violate the due process clause because the absence has no "`relation . . . to the fulness of [the defendant's] opportunity to defend'" himself. Id. (quoting United States v. Gagnon, 470 U.S. 522, 526 (1985)). Accordingly, we reject appellants' constitutional arguments.

B. Jencks Act Claims

Appellants contend that the district court committed reversible error when it denied appellant Davis' request — under the Jencks Act, 18 U.S.C. Section(s) 3500 (1988) — for documents created by Detective Quander during the course of preparing the transcripts of the body-wire recordings. The Jencks Act provides, in pertinent part:

(b) After a witness called by the United States has testified on direct examination, the court shall, on motion of the defendant, order the United States to produce any statement (as hereinafter defined) of the witness . . . which relates to the subject matter as to which the witness has testified. . . .

* * * * * *

(e) The term "statement," as used in subsection (b) . . . of this section in relation to any witness called by the United States, means —

(1) a written statement made by said witness and signed or otherwise adopted or approved by him. . . .

Id. Section(s) 3500(b), 3500(e). The Act "was intended to enable the defense to impeach a government witness by bringing any variances [between Jencks material and the witness's testimony] to the attention of [the finder of fact]." Norinsberg Corp. v. United States Dep't of Agriculture, 47 F.3d 1224, 1229 (D.C. Cir. 1995) (quoting United States v. Prieto, 505 F.2d 8, 11 (5th Cir. 1974)); see also United States v. Harrison, 524 F.2d 421, 431 n. 29 (D.C. Cir. 1975).

The trial court has a duty under the Jencks Act to engage in an adequate inquiry into the nature of documents requested before ruling against Jencks Act production. United States v. North America Reporting, Inc., 740 F.2d 50, 55 (D.C. Cir. 1984). In this case, the district court concluded that the documents at issue were not "statements" under the Jencks Act because they represented early drafts of a work in progress that were not "adopted" by the detective witness until he verified the final version of the transcript. The court refused to hold a hearing but agreed to review the notes in camera and ordered the government to file the documents under seal for the purposes of appeal. The government did not provide the documents to the district court and did not file the notes until it filed the appendix to its brief in this appeal. Therefore, the notes were not available to appellant until after he filed his initial brief on appeal. We are displeased by the government's failure to meet its express undertakings in this case but, in the absence of prejudice caused by the omission, will assume this is an aberration rather than evidence of any pattern of failing to comply with court orders on potential Jencks material.

After careful review of the documents belatedly supplied by the government we find the attributions and the statements reflected in Detective Quander's handwritten notes to be entirely consistent with his testimony on the witness stand and we cannot conceive of any reasonable basis on which appellant was prejudiced by his inability to inspect the notes before Detective Quander testified. Thus, the government's failure to produce the documents had no actual consequence. When the information contained in undisclosed Jencks materials does not vary from that provided by the witness at trial, "it would offend common sense and the fair administration of justice to order a new trial." Rosenberg v. United States, 360 U.S. 367, 371 (1959). See also United States v. Rippy, 606 F.2d 1150, 1154 (D.C. Cir. 1979) (Jencks violation harmless where evidence of guilt was strong, undisclosed statement was relatively insignificant, and no evidence was adduced of bad faith by government). Therefore, we need not reach appellant's contention that the handwritten notes are "statements" in order to support our conclusion that any error by the trial court with respect to the disclosure of Detective Quander's handwritten notes was harmless.

C. Jury Contamination

Every criminal defendant is entitled to an impartial jury, as free as practicable from extraneous influences that could subvert the fact-finding process. Appellant Holton claims he was denied that right because, on the evening after the defendants began to present their cases, members of the jury may have seen a Nightline program dealing with sentencing standards for crack cocaine offenses. During the broadcast, the link between crack and violence was discussed extensively. Holton particularly targets a televised comment by District of Columbia Superior Court Judge Reggie Walton that credited high sentences for crack offenses with a reduction in the murder rate in the District of Columbia. The program did not refer specifically to anyone involved in this case.

There is no general presumption of prejudice where jurors are exposed to media coverage about the case on which they are sitting, United States v. Williams — Davis, 90 F.3d 490, 501 (D.C. Cir. 1996), and we find no special circumstances to invoke one on the basis of a program which did not specifically concern the trial or the parties in this case. "It is for the trial judge to decide at the threshold whether news accounts are actually prejudicial; whether the jurors were probably exposed to the publicity and whether jurors would be sufficiently influenced by bench instructions alone to disregard the publicity." Gordon v. United States, 438 F.2d 858, 873 (5th Cir. 1971); Waldorf v. Shuta, 3 F.3d 705, 709-10 (3d Cir. 1993) (cited in Williams-Davis, 90 F.3d at 501). "The trial judge has broad discretion in ruling on the issue of prejudice resulting from a jury's exposure to news articles concerning a trial." United States v. Aragon, 962 F.2d 439, 443 (5th Cir. 1992). Therefore, the court reviews the judge's decision not to conduct a mid-trial voir dire of the jury for an abuse of discretion.

First, courts must determine if the nature of the news material in question is innately prejudicial. Courts are more likely to find publicity prejudicial if it was broadcast close to the time of trial, discussed information that would have been inadmissible at trial, reported on the outcome of proceedings against co-defendants, or was inflammatory. See, e.g., Salemme v. Ristaino, 587 F.2d 81, 88 (1st Cir. 1978); United States v. Holman, 680 F.2d 1340, 1348 (11th Cir. 1982). The second inquiry involves the likelihood that the publicity has in fact reached the jury. The prominence of the coverage and the nature and number of judicial warnings against viewing outside coverage become relevant at this stage. United States v. Manzella, 782 F.2d 533, 542 (5th Cir. 1986) (citations omitted).

In this case, the Nightline broadcast was a half-hour national network program televised midway through appellants' trial at 11:30 p.m. (a promotional trailer may have appeared earlier in the evening). The focus on the Nightline show was the level of sentencing for crimes involving crack cocaine. Although it dealt with an issue that collaterally affected the defendants, it had nothing to do with their guilt (which they stoically denied) and prejudice to their individual defenses was extremely unlikely. We agree with the district court that its frequent reminders to the jury not to consider information gleaned outside of the courtroom also acted to stem any conceivable prejudicial inferences a viewer might take away from the program about sentencing in crack cases generally. In light of the absence of any obvious connection between the broadcast and the defendants, a mid-trial voir dire of the jury might well have prejudiced the defendants by drawing attention to the sentencing issues discussed in the broadcast before these defendants had even been found guilty. Accordingly, we conclude that the district court did not abuse its discretion when it denied appellant's request for a mid-trial voir dire.

D. Equal Protection

Appellant Holton contends that minorities are denied the equal protection of the law because they are disproportionately impacted by the higher mandatory minimum sentences applied for crimes involving crack cocaine than for crimes involving powder cocaine. He notes that the United States Sentencing Commission and an article published in the Journal of the American Medical Association support the conclusion that there is no rational basis for treating crack cocaine differently than powder cocaine. See Dorothy K. Hatsukami and Marian W. Fischman, Crack Cocaine and Cocaine Hydrochloride: Are the Differences Myth or Reality?, 276 J. Am. Med. Ass'n 1580-88 (1996). He argues that "where a law has a grossly disproportionate impact on one racial group, and where there are strong indications that the disparity is the result of an irrational statutory distinction and discriminatory enforcement," Holton Brief at 28, the Fifth Amendment right to equal protection of the law is violated.

In analyzing whether the sentencing disparity denies constitutional equal protection, the first inquiry is whether the mandatory crack minimums discriminate based on race. In order to prove that a facially neutral statute, such as the one involved here, violates equal protection guarantees, a challenger must demonstrate a racially discriminatory purpose behind the statute. Washington v. Davis, 426 U.S. 229, 239 (1976). Disparate racial impact can be probative of such purpose, but it is not dispositive without more. See id.

Here, appellant has not offered any specific evidence of disparate impact. Because he concentrates on the irrationality of the sentencing disparity, we surmise that his current argument is that since the Sentencing Commission and scientific analysis, such as that published in the Journal of the American Medical Association, indicate that crack cocaine and powder cocaine are not distinguishable, Congress should change the law and equalize the mandatory minimums. The opinions of the Sentencing Commission and scientific journals do not provide the requisite proof that Congress was motivated by any impermissible considerations. Moreover, this circuit recently reaffirmed its conclusion that there is a race-neutral explanation for the sentencing disparity. "As evident from our analysis in United States v. Cyrus and Thompson, Congress acted in light of the distributional efficiencies, heightened potency and acute violence associated with crack cocaine. The crack/powder distinction in the statute is thus readily explained on grounds other than race." United States v. Johnson, 40 F.3d 436, 440 (D.C. Cir. 1994), cert. denied, 115 S. Ct. 1412 (1995); see United States v. Thompson, 27 F.3d 671, 678-79 (D.C. Cir. 1994); United States v. Cyrus, 890 F.2d 1245, 1248 (D.C. Cir. 1989).

The convictions and sentences below are affirmed.

So ordered.