Appellant Roland Marshall Dumas was convicted in a jury trial of conspiracy to possess cocaine with intent to distribute, in violation of 21 U.S.C. § 846 (1988). The district court (Woodlock, J.) imposed a sentence of 262 months imprisonment, and denied Dumas's Motion for New Trial and Revised Motion for New Trial based on newly discovered evidence. Dumas appeals this ruling denying his motions for a new trial.

On appeal, Dumas presents two instances in which he claims the district court abused its discretion in denying his motions for a new trial, which we address seriatim. First, he claims that the district court should have granted his motions because the government failed to disclose allegedly exculpatory documents relating to the impeachment of a government witness. This witness's lack of credibility, he claims, was vital to his entrapment defense. Second, he claims that the district court should have granted the motions for a new trial in order to remedy a violation of due process springing from the government's failure to disclose, and subsequent destruction of, notes that Dumas claims constituted exculpatory material. For the reasons discussed below, we affirm the district court's denial of the motions, finding no abuse of discretion.

The facts of this case are byzantine in their complexity. We recount only those facts that are essential to our decision. In the fall of 1993, Dumas was serving a sentence in federal prison in Pennsylvania. At the request of the United States Attorney's Office, the Bureau of Prisons transferred Dumas to the Plymouth County House of Corrections ("Plymouth"), a Massachusetts state prison facility. According to Dumas, the United States Attorney's Office sought this transfer in order to obtain his testimony before a federal grand jury in a public corruption case. Dumas refused to testify, and was subsequently held in contempt and remained incarcerated at Plymouth. This refusal to testify, Dumas argued at trial, set in motion the events that led to his ultimate entrapment by the government. He claimed at trial, and continues to do so on appeal, that the government entrapped him in order to gain additional leverage with which to secure his grand jury testimony.

While at Plymouth, Dumas met Christopher Coyne, who invited Dumas to share a cell with him. Coyne, unbeknownst to Dumas, was an informant for the Drug Enforcement Agency ("DEA"). Coyne's initial investigative goal did not involve Dumas; Coyne sought, unsuccessfully, to set up drug transactions involving four others, and only turned his attention to Dumas after he failed in that endeavor.

Dumas claims that during the time in which the two shared a cell at Plymouth, Coyne put extraordinary pressure on him to engage in a drug transaction. The amount of time that the two spent in a cell together was a matter of dispute, and one which Dumas felt was vital to his entrapment defense. DEA Agent Steven Story testified that the two men shared a cell for "no more than a week." Coyne, for his part, recalled that they shared a cell for approximately two weeks. Dumas, on the other hand, testified that he and Coyne shared a cell for nearly two months.

Eventually, Coyne introduced Dumas to Agent Story, who was posing as a drug dealer affiliated with Coyne. Dumas participated in three telephone conversations with Agent Story, and planned a drug transaction. Though the transaction never took place, these conversations formed the basis for Dumas's conspiracy conviction.

None of the conversations between Coyne and Dumas, which Dumas claims constituted entrapment, were recorded. Because of this, Dumas claims that Coyne's credibility was crucial to his conviction. He points, inter alia, to the closing argument of the prosecutor, who stated: "If you believe Chris Coyne, the case is very simple."

Nearly two years after entry of judgment, Dumas moved for a new trial. He did so because he had learned from a corrections officer that, during the relevant time period, Coyne had been placed on a "Q-5 suicide watch" at Plymouth. This information had not been provided to the defense, although the defense had received general information on Coyne's overall psychiatric background. The government was aware of the suicide watch, because it was documented in the notes of DEA Agent Thaddeus Blazak. Apparently, Agent Blazak had spoken with corrections officials, and had taken notes at the meeting. Blazak had learned that Coyne had been placed on the Q-5 suicide watch because of "relationship problems." In the course of investigating this new revelation, Dumas also claims to have learned definitively that he and Coyne had shared a cell for nearly two months.

Based on this information, Dumas brought a motion for a new trial. The government, not surprisingly, opposed this motion. In doing so, the government contended that Coyne had requested the suicide watch in order to be segregated from the general population because he was afraid of Dumas. The government, in support of this position, offered the testimony of Agent Story, who had spoken with another informant who was incarcerated with Coyne. The informant stated to Agent Story:

[T]hat he was aware that there had been some problems surrounding Coyne backing out of the proposed undercover transaction and that he felt because of some heat, as he put it, that the defendant Dumas . . . had placed on him, that Coyne staged this incident or staged this ruse as an attempt to get him away from the defendant and to put him in a safe position without actually having to ask or request prison officials to be moved to isolation which invariably would mean that he had the status of being a Government informant.

A-182. After he was informed that he would be testifying in a hearing on Dumas's motion for a new trial, Agent Blazak shredded certain pages from his notebook which the defense alleges contained information relating to Dumas's case. Dumas claims that Blazak did this despite knowing that he would have to produce any relevant notes at this hearing.

Some pages pertaining to Dumas remained in Agent Blazak's notebook, including those pages that documented the Q-5 suicide watch. In addition, Blazak's notes contained information that supported Dumas's contention that Dumas and Coyne had shared a prison cell for nearly two months. These notes, Dumas now asserts, could have been used to impeach both Coyne and Agent Story. In addition, Dumas claims that the government discovered during trial that Story's testimony on the length of time Coyne and Dumas shared a cell was erroneous, and failed to correct it. Finally, Dumas asserts that the shredding of pages from Blazak's notebook constituted destruction of evidence in violation of his Due Process rights.

We review the district court's denial of a motion for a new trial only for manifest abuse of discretion. See United States v. Brimage, 115 F.3d 73, 79 (1st Cir.), cert. denied, 118 S. Ct. 321 (1997). In motions brought for a new trial based upon newly discovered evidence of prosecutorial nondisclosure, we accord deference to the district court's determination of materiality. See United States v. Hahn, 17 F.3d 502, 510 (1st Cir. 1994) ("[T]he district court's determination of the materiality of newly discovered evidence in prosecutorial nondisclosure cases is ordinarily accorded deference.") (quoting United States v. Sanchez, 917 F.2d 607, 618 (1st Cir. 1990)).

While Dumas states many facts, he alleges essentially two legal claims. First, he alleges that failure to disclose the contents of Agent Blazak's notes, which included information on the Q-5 suicide watch and the amount of time Coyne and Dumas shared a cell, which would constitute a Brady violation justifying a grant of a new trial. Second, he argues that the shredding of certain of Blazak's notes could be a violation of Arizona v. Youngblood, 488 U.S. 51, 58-58 (1988) (holding that bad faith destruction of evidence constitutes a due process violation) and Brady v. Maryland, 373 U.S. 83 (1963) (requiring disclosure of exculpatory evidence).

As the Supreme Court has stated: "We do not . . . automatically require a new trial whenever a combing of the prosecutors' files after the trial has disclosed evidence possibly useful to the defense but not likely to have changed the verdict . . . ." Giglio v. United States, 405 U.S. 150, 154 (1972) (internal quotation marks omitted). Instead, "[a] finding of materiality of the evidence is required under Brady." Id.

Evidence is material under Brady only when "there is a reasonable probability that, had the evidence been disclosed to the defense, the result of the proceeding would have been different." United States v. Bagley, 473 U.S. 667, 682 (1985) (opinion of Blackmun, J.); see also Kyles v. Whitney, 514 U.S. 419, 433-35 (1995) (endorsing Bagley test); United States v. Cunan, 152 F.3d 29, 34 (1st Cir. 1998) (same). This, as we stated in Cunan, does not mean that the reviewing court must be certain that a different result would obtain. See id. Rather, a defendant such as Dumas may prove a Brady violation by showing that "the favorable evidence could reasonably be taken to put the whole case in such a different light as to undermine confidence in the verdict." Kyles, 514 U.S. at 435; Cunan, 152 F.3d at 34.

Part of Dumas's prayer for a new trial is based on destruction of exculpatory evidence under Youngblood. We examine this claim under a different standard. As we noted in a previous case, "[T]he Supreme Court's jurisprudence divides cases involving nondisclosure of evidence into two distinct universes. Brady and its progeny address exculpatory evidence that is still in the government's possession. Youngblood and Trombetta govern cases in which the government no longer possesses the disputed evidence." United States v. Femia, 9 F.3d 990, 993 (1st Cir. 1993). This standard, culled from California v. Trombetta, 467 U.S. 479, 488-89 (1984), and Youngblood, 488 U.S. at 58, establishes three hurdles. The defendant must show that, in failing to preserve the evidence, the government, (1) acted in bad faith when it destroyed evidence, which (2) possessed an apparent exculpatory value and which (3) is to some extent irreplaceable. See Femia, 9 F.3d at 993-94.

1. Failure To Turn Over Evidence of the Q-5 Suicide Watch

According to Dumas, had he been able to introduce evidence of Chris Coyne's suicide watch, the jury would have assessed Coyne's credibility differently. Dumas suggests, probably rightly, that in order for the jury to reject his entrapment defense, it would have needed to find Coyne more credible than Dumas. Evidence of psychiatric problems, he asserts, would have called into question the reliability and accuracy of Coyne's testimony.

As a preliminary matter, it is clear that impeachment evidence falls within the Brady rule. See Giglio, 405 U.S. at 154. The simple fact that evidence could be used to impeach a government witness, however, does not end the materiality inquiry. "Impeachment evidence, even that which tends to further undermine the credibility of the key Government witness whose credibility has already been shaken due to extensive cross-examination, does not create a reasonable doubt that did not otherwise exist where that evidence is cumulative or collateral." United States v. Shelton, 588 F.2d 1242, 1248 (9th Cir. 1978), quoted in United States v. Sanchez, 917 F.2d 607, 618-19 (1st Cir. 1990).

The district court considered Dumas's contentions with respect to the suicide watch in some detail. It ultimately concluded that the evidence would not likely have aided Dumas's cause, in light of the explanation that would have been offered by the government for the suicide watch. The court stated:

With respect to the question of Mr. Coyne's placement on the Q-5 suicide watch, I start with what seems to me to be - in the light most favorable to the defendant - a very contested set of circumstances about what was really going on here. The evidence that I have before me of record suggests that Mr. Coyne undertook this as a self-help measure to protect himself from being identified as an informant and being exposed to some form of retaliation. It is, of course, also, consistent with a fragile psychological construct. And there have been some disclosures of psychological difficulties by Mr. Coyne. But the weight of the evidence simply does not support the proposition that there is a sufficient basis to have a jury find that he was put in administrative supervision on the basis of some sort of suicidal ideation.

. . . This much I know: That given the state of the record now, this would not have been exculpatory rather than inculpatory of Mr. Dumas.

This evidence, which arguably impeaches Coyne with respect to mental stability and his ability to recall and perceive events accurately, does not rise to the level of materiality required. While we recognize that impeachment evidence, if powerful enough, could constitute grounds for a new trial, we do not think that such is the case here. See Sanchez, 917 F.2d at 618-19.

In United States v. Slade, 980 F.2d 27, 29 (1st Cir. 1992) we noted, in conducting abuse of discretion review of a motion for a new trial, that "the district court is usually in a much better position to judge the credibility of the witnesses and to assess the highly nuanced relationship between the purported new evidence, and what previously transpired at trial . . . ." We think Slade's words instructive in this case. Therefore, mindful of the district court's superior position from which to assess the new evidence and the credibility of the witnesses, we are unwilling to disturb the court's careful conclusion on this point.

2. Failure To Turn Over Evidence of the Amount of Time Dumas and Coyne Shared a Cell

Dumas, as we have stated, argued that he was entrapped by Coyne, a government agent. He argues on appeal that his entrapment defense would have been successful had Dumas's testimony about the length of time he shared a cell with Coyne been corroborated by Brady material that should have been turned over. This evidence, too, was considered by the district court on the motion for a new trial. The court stated:

I turn to the question of the amount of time that he shared the cell with Mr. Coyne. That was the disputed issue at trial. But was it material? Would it have made a difference in . . . the outcome? I think not. [Defense counsel] makes a forceful point that it was a kind of reed from which he was attempting to construct a defense, but it was a thin reed. And the defense, given the state of the record, was grasping at such reeds. And this difference in the amount of time simply wouldn't have made a difference in the result of the case. That really was a dispute over subsidiary and, frankly, secondary kinds of issues that do not provide a touchstone for the kind of proof or evidentiary material that Bagley and all of its progeny have required for there to be a finding of materiality with respect to late disclosed exculpatory evidence . . . in the final analysis, it would not have made a difference.

Upon careful review, we find an adequate basis for the court's conclusion that there was not a reasonable probability that evidence would have made a difference in the result of the trial. We are particularly persuaded by the district court's view of the collateral nature of the disputed evidence. This view, to which we owe deference, see Slade, 980 F.2d at 29, provides sufficient basis for the court's decision.

We are impressed by the fact that Dumas himself admitted on cross-examination that Coyne did not mention a drug transaction for some time after they began sharing a cell. The government presses this point on appeal, noting the following colloquy from the trial record:

Q. Christopher Coyne didn't say anything to you about a drug deal the first three weeks you were there; is that your testimony?

A. The first three or four weeks, he said nothing until - I think it was right around the time I went to the Grand Jury.

Q. When was that?

A. I have no idea. It probably was - I think I went to the Grand Jury ten days after I was there. And about a week after I got back, I was complaining because I was going to get the contempt charge . . . So, I'd say it was probably the last week of October.

Dumas's argument for the materiality of the disputed evidence is that Coyne pressured him into the drug transaction over an extended period of time. The longer they shared a cell, Dumas argues, the more time Coyne had to exert pressure. But Dumas's own testimony undercuts this view. He admits that for a significant proportion of the time they were together, Coyne did not pressure him.

The district court's decision on this point is further bolstered by the fact that the evidence goes to only one prong of an entrapment defense. See United States v. Gifford, 17 F.3d 462, 468 (1st Cir. 1994) (affirmative defense of entrapment includes proof of both inducement and lack of predisposition). Given Dumas's criminal history,1 the latter element might likely have proven just as fatal as the former.2

In order to prevail on a motion for a new trial based on destruction of evidence, a defendant must show that the government: (1) acted in bad faith when it destroyed evidence, which (2) possessed an apparent exculpatory value and which (3) is to some extent irreplaceable. See Femia, 9 F.3d at 993-94. Dumas's argument that the government violated his due process rights by destroying Blazak's notes calls us to split this first prong of the analysis into two distinct inquiries. First, did the government destroy evidence at all? Second, was that evidence destroyed in bad faith?

With respect to the first question, the district court found as a matter of fact that Blazak did not destroy notes that were relevant to the instant case. The court stated:

[I]t is bad business for the Government to fool around with integrated documents that are going to provide the basis for examination and they do so at their peril. And I'm astonished that an agent with even a couple of years of experience would treat the notebook that he has like a game: "This is relevant, that part isn't" and throw that part out and keep this part.

Nevertheless, I've looked at the notebook fairly carefully. I've looked at the agent's testimony fairly carefully and thought about it. And I believe that what remains here is all that is relevant to the Dumas case.

We share the district court's incredulity at this unquestionably poor law enforcement practice. This does not, however, undercut our faith in the district court's finding of fact that the relevant materials were not excised, but were in fact produced.

With respect to the second question, the court determined that any destruction was not in bad faith. The court stated: "[I]f it is all that is relevant - and I find that it is - then it does nothing more than show certain sloppy practice." Again, we are loathe to disturb a finding based so heavily on the credibility of witnesses. See Slade, 980 F.2d at 29.

The district court's findings with respect to either the first or second question would each individually suffice to support its denial of the motion based on destruction of evidence. Accordingly, we refuse to disturb that decision.

Having found that the district court did not abuse its discretion with respect to each piece of new evidence in isolation, all that remains is to determine whether it abused its discretion in failing to order a new trial due to the cumulative effect of the errors Dumas claims. The district court considered the cumulative effect, succinctly stating: "I can't find that, individually or taken together, the evidentiary materials upon which Mr. Dumas now relies would have been sufficient to change the outcome. And for that reason, they are not material and ultimately do not provide a basis for a new trial." Put just as succinctly, we agree.

For the foregoing reasons, we affirm the district court's ruling in its entirety.