Michael Edward Kennedy, a federal prisoner, timely appeals from the denial of a 28 U.S.C. Sec. 2255 motion to vacate his convictions for felony murder and rape. Kennedy contends the district court erred when it held that the government's failure to disclose a letter containing allegedly exculpatory information did not violate Kennedy's due process rights. We affirm the district court order.

* Kennedy was indicted in February 1980 for murder, felony murder, and rape, arising from the 1979 rape and murder of Maria Lopez de Felix. The juries at Kennedy's first two trials were unable to reach verdicts. Midway through the second trial, the court granted Kennedy's motion for judgment of acquittal on the murder charge, but denied the motion for the felony murder and rape charges. On December 15, 1980, a third jury found Kennedy guilty of rape and felony murder. The district court sentenced Kennedy to two concurrent life sentences. This court affirmed the convictions. See United States v. Kennedy, 714 F.2d 968 (9th Cir.1983), cert. denied, 465 U.S. 1034, 104 S. Ct. 1305, 79 L. Ed. 2d 704 (1984).

On May 14, 1985, Kennedy filed a 28 U.S.C. Sec. 2255 motion to vacate his convictions on grounds that the government improperly suppressed a letter it had received from Dr. John Burton that contained allegedly exculpatory evidence.1 The district court held an evidentiary hearing on Kennedy's motion and denied it.

During each of Kennedy's three trials, the defense presented testimony by Dr. Thomas T. Noguchi, an expert in forensic pathology. Dr. Noguchi testified that the degree of rigor mortis in a victim's body was a reliable indicator of the time of death.2 He also stated, after reviewing police and autopsy reports, that he believed Lopez had died six to eighteen hours before her body was discovered--a conclusion that tended to exculpate Kennedy. During the cross-examination, the government attempted to impeach Dr. Noguchi's testimony with excerpts from articles written by Dr. Burton, stating that rigor mortis was a highly unreliable indicator of a victim's time of death.

After the completion of the first trial, the government sent a copy of Dr. Noguchi's testimony to Dr. Burton for his critique. Dr. Burton responded on August 2, 1980 with a letter expressing his opinion that Lopez' death occurred from "one to four hours before she was found dead." The government later determined that Dr. Burton's phrase "before she was found dead" referred to the time at which the coroner examined Lopez' body.3 The government sent a copy of Dr. Burton's letter to Kennedy in June 1982, two and one-half years after his convictions.

In May 1985, the government sent additional transcripts from Kennedy's first trial to Dr. Burton for his review. These included transcripts of (1) the testimony of Joseph Cogen, the deputy coroner who made the original examination of Lopez, (2) the testimony of Dr. Francis Luibel, the physician who conducted the autopsy of Lopez' body and who appeared as the government's expert witness, and (3) the testimony of Dr. Noguchi. Dr. Burton responded on May 15, 1985 with a letter stating that his original opinion was incorrect and that he now fully agreed with Dr. Luibel's opinion that Lopez died as much as thirty-six to forty hours prior to the finding of her body, a statement tending to inculpate Kennedy. Dr. Burton later incorporated his May 1985 letter into a sworn affidavit.

On March 1, 1986, Kennedy's attorneys presented Dr. Burton with transcripts of testimony given at Kennedy's third trial by Mr. Cogen, and Drs. Luibel and Noguchi. After reviewing these transcripts, Dr. Burton changed his opinion again, this time stating that Lopez had died zero to sixteen hours before her body was discovered.

We must decide whether the district court erred by holding that the government's failure to disclose Dr. Burton's letter did not violate Kennedy's due process rights.4 This court reviews de novo challenges to a conviction based on a Brady violation. United States v. Gordon, 844 F.2d 1397, 1402 (9th Cir.1988). The court also reviews de novo the denial of a Sec. 2255 motion. See United States v. Grewal, 825 F.2d 220, 222 (9th Cir.1987).

The government has a duty to disclose to a defendant, upon request, information that is favorable to an accused "where the evidence is material either to guilt or punishment." United States v. Bagley, 473 U.S. 667, 683, 105 S. Ct. 3375, 3384, 87 L. Ed. 2d 481 (1985); Brady v. Maryland, 373 U.S. 83, 87, 83 S. Ct. 1194, 1196-1197, 10 L. Ed. 2d 215 (1963). The Brady doctrine has been expanded to include cases where the defendant has not requested the relevant material. See United States v. Agurs, 427 U.S. 97, 107, 96 S. Ct. 2392, 2399, 49 L. Ed. 2d 342 (1976). Evidence is material under Brady "only if there is a reasonable probability that, had the evidence been disclosed to the defense, the result of the proceeding would have been different. A 'reasonable probability' is a probability sufficient to undermine confidence in the outcome." Bagley, 473 U.S. at 682, 105 S.Ct. at 3383.

Here, the government received a letter in 1980 from Dr. John Burton, a pathology expert. Dr. Burton's letter was in response to a government request for a critique of the testimony of Kennedy's expert witness. In the letter, Dr. Burton stated an opinion as to the time of death of Kennedy's alleged victim which tended to exculpate Kennedy.

It is not clear whether the Brady doctrine is limited to the actual materials in a prosecutor's possession, or whether it also includes information that the defendant reasonably could acquire after the prosecution's disclosure of the materials. However, in Lee v. United States, 388 F.2d 737, 739 (9th Cir.1968), this court stated: "Upon a motion for new trial based upon nondisclosure of evidence ... the precise issue is what the Government had learned from the witness in question ... rather than what this witness might now state under examination by the defendant."5

Lee suggests we should consider only the materiality of Dr. Burton's letter, in and of itself. See id. at 739; see also United States v. Chen, 754 F.2d 817, 824 (9th Cir.1985). As the Supreme Court has recently stated, however, in determining the materiality of undisclosed information, a reviewing court may consider "any adverse effect that the prosecutor's failure to respond might have had on the preparation or presentation of the defendant's case." Bagley, 473 U.S. at 683, 105 S.Ct. at 3384. Accordingly, we will discuss both the materiality of Dr. Burton's letter and the materiality of Dr. Burton's expected testimony had he been called as a witness at one of Kennedy's trials.

* Kennedy first argues that Dr. Burton's letter is itself Brady material. Because the letter is hearsay, however, it is not admissible as substantive evidence. See Fed.R.Evid. 801, 802. Furthermore, although impeachment evidence falls within the Brady doctrine if it is material to guilt or innocence, see Giglio v. United States, 405 U.S. 150, 154, 92 S. Ct. 763, 766, 31 L. Ed. 2d 104 (1972), and see Bagley, 473 U.S. at 676, 105 S.Ct. at 3380, Dr. Burton's letter has no such effect.

To be material under Brady, undisclosed information or evidence acquired through that information must be admissible. See Brady, 373 U.S. at 89-90, 83 S.Ct. at 1197-1198; United States v. Oxman, 740 F.2d 1298, 1311 (3d Cir.1984), vacated on other grounds, 473 U.S. 922, 105 S. Ct. 3550, 87 L. Ed. 2d 673 (1985); United States v. Ranney, 719 F.2d 1183, 1190 (1st Cir.1983) ("Inadmissible evidence is by definition not material [for Brady purposes], because it never would have reached the jury and therefore could not have affected the trial outcome."). Kennedy is barred from introducing Dr. Burton's letter as substantive evidence because the letter contains out-of-court statements and does not fall within any exception to the hearsay rule. See Fed.R.Evid. 801-804.

Kennedy argues, however, that the letter would have been admissible for impeachment purposes. In both Kennedy's first and third trials, the government used portions of articles written by Dr. Burton in an attempt to impeach the testimony of Dr. Noguchi, Kennedy's expert witness. Dr. Burton's articles tended to refute Dr. Noguchi's assertions that rigor mortis is a reliable indicator of the time of death. In his letter, however, Dr. Burton himself appears to have relied on a certain type of rigor mortis sign, that which appears immediately after a "fight or flight" reaction to a "death struggle," as the basis for his August 2, 1980 opinion on the time of Lopez' death. Apparently, Kennedy wished to use these statements about rigor mortis in Dr. Burton's letter to impeach or contradict statements of opinion in Dr. Burton's articles.

The record shows that Dr. Burton's articles had a fairly insignificant impeachment effect on Dr. Noguchi's testimony. At Kennedy's first trial, for example, Dr. Noguchi responded to Dr. Burton's statement that rigor mortis is the "most misleading of all post-mortem changes" by stating:

I recollect that Dr. Burton made a statement before the Academy of Forensic Sciences' meeting. It was a very interesting presentation. I do agree in general that dogmatic statements should not be made without considering various factors that might exist. I do respectfully disagree that such a general statement, to discard the information or experience we have, I think that it may be erroneous.

Other than acknowledging Dr. Burton's credentials, this was the only apparent exchange between the prosecutor and Dr. Noguchi concerning Dr. Burton's articles during Kennedy's first trial.

At Kennedy's third trial, the prosecutor again used Dr. Burton's articles about the unreliability of rigor mortis as an indicator of the time of death in an attempt to impeach Noguchi. Kennedy has not directed our attention to any meaningful differences between the examinations of Dr. Noguchi in the first and third trials.6 It does not appear that the prosecution's use of Dr. Burton's articles significantly affected Dr. Noguchi's credibility at either trial. Prior inconsistent statements made by Dr. Burton in his August 2, 1980 letter would have only slightly diminished the already insignificant impeachment effect of the Burton articles. Accordingly, Dr. Burton's letter would have had only a negligible impact on Dr. Noguchi's testimony and was, therefore, not material to guilt or innocence for purposes of the Brady doctrine. See Bagley v. Lumpkin, 798 F.2d 1297 at 1300 (9th Cir. 1986) (to determine whether impeachment evidence is Brady material, the court must determine whether the evidence would have made a difference between conviction and acquittal if disclosed and used effectively).

Kennedy also argues that the letter is Brady material because of the testimony Dr. Burton would have given had Kennedy called him as a witness. Between 1980 and 1986, however, Dr. Burton stated three different opinions concerning the estimated time of Lopez' death.

First, the opinion Dr. Burton expressed in his 1980 letter is inadmissible because it was not based on facts of a type reasonably relied upon by experts in the field in forming such opinions. See Fed.R.Evid. 703 (facts or data on which an expert bases an opinion need not be admissible if they are "of a type reasonably relied upon by experts in the particular field"). In his August 1980 letter, Dr. Burton stated that Lopez died from one to four hours before her body was examined. Because the victim's body was discovered five and one-half hours before the examination occurred, this opinion is inconsistent with the facts. Furthermore, Dr. Burton admitted at the Sec. 2255 hearing that the 1980 opinion was not based on the type of materials which an expert would normally rely upon in giving such an opinion. Accordingly, had Dr. Burton proposed to testify in accordance with his 1980 letter, it is unlikely that his testimony would have been admissible.

Second, Dr. Burton's 1985 opinion tends to inculpate Kennedy, and consequently, is not favorable evidence to the accused. In May and October 1985, Dr. Burton stated that, after reviewing the relevant transcripts from Kennedy's first trial, he fully agreed with Dr. Luibel's testimony that Lopez died as much as thirty-six to forty hours prior to the finding of her body. Because these statements tend to inculpate Kennedy, they do not support his claim that Dr. Burton's letter is Brady material. See Brady, 373 U.S. at 87, 83 S.Ct. at 1196-1197.

Finally, in August 1986, Dr. Burton stated that Lopez had died zero to sixteen hours before her body was found. Dr. Burton stated that he arrived at this last opinion after reviewing the transcripts from Kennedy's third trial. Dr. Burton, however, was unable to indicate any appreciable difference between the transcripts of the first and third trials, and would not state any rational basis for his changed opinion.7 Furthermore, had Dr. Burton testified in accordance with his 1986 opinion, although tending to exculpate Kennedy, it would have been cumulative to Dr. Noguchi's testimony, Kennedy's expert witness. See United States v. Polizzi, 801 F.2d 1543, 1553 (9th Cir.1986) (no Brady violation where the undisclosed evidence was cumulative to evidence presented at trial); see also United States v. Van Brandy, 726 F.2d 548 (9th Cir.1984) (defendant must not only make a showing that the non-disclosed evidence is material and favorable to defendant, but also that it has not been included in any report provided to defendant).

Between 1980 and 1986, Dr. Burton stated three substantially divergent opinions. It is not easy to determine what Dr. Burton's testimony would have been had he been called as a witness.8 Accordingly, the district court did not err by finding that Dr. Burton's expected testimony did not support Kennedy's Brady claim. See Bagley, 473 U.S. at 682, 105 S.Ct. at 3383-3384.

AFFIRMED.