AMENDED ORDER GRANTING MOTION FOR DISCOVERY

PREGERSON, District Judge.

In many criminal trials, the government relies on the testimony of people who were involved with the defendant in the commission of the crime charged. Such "accomplice witnesses" are often the best, if not the only, source of information about the alleged crime. To ensure that accomplice witnesses testify truthfully and completely, the government often reaches agreements with such witnesses that offer leniency or immunity in exchange for truthful testimony.

In this matter, the Court is called on to determine the boundaries of the government's discovery obligations relating to the agreements between the government and accomplice witnesses. The defendants have moved the Court to compel disclosure of documents and information relating to the period between an accomplice witness's initial contact with the government regarding possible cooperation and the point at which the witness and the government reached an agreement concerning the accomplice witness's testimony.

I. Background

In a multi-count indictment the government charged defendants Jeffery Sudikoff and Michael Cheramy with various securities and related violations. One of Sudikoff's associates, Phil McInnes, received immunity from the government and will testify for the prosecution at trial.

As part of the discovery process, the government has disclosed information that relates to McInnes's testimony. The information dates back to April 1995, which was shortly after McInnes and the government reached an agreement as to immunity. Sudikoff requests that the Court order the government to disclose material dating back to the Fall of 1994, when McInnes first began communicating with the government regarding possible testimony. Specifically, Sudikoff requests "all notes or other evidence of any communication between the government and Phil McInnes or his counsel — including materials relating to 'proffer sessions' that occurred well before the date of the first FBI '302' for Mr. McInnes." (Mot. at 1.)

II. Discussion

Sudikoff asserts three grounds for discoverability: the doctrine of Brady v. Maryland, 373 U.S. 83, 83 S.Ct. 1194, 10 L.Ed.2d 215 (1963); the doctrine of Giglio v. United States, 405 U.S. 150, 92 S.Ct. 763, 31 L.Ed.2d 104 (1972); and the requirements of the Jencks Act, 18 U.S.C. § 3500. Because Giglio is a subcategory of Brady, see Giglio, 405 U.S. at 153-54, 92 S.Ct. 763, the Court will consider these doctrines together and then consider the application of the Jencks Act.

A. Brady and Giglio

In Brady, the Supreme Court held that "the suppression by the prosecution of evidence favorable to an accused upon request violates due process where the evidence is material either to guilt or to punishment. . . ." 373 U.S. at 87, 83 S.Ct. 1194. Evidence that weakens the credibility of a prosecution witness has long been considered Brady material. See, e.g., Thomas v. United States, 343 F.2d 49 (9th Cir. 1965). Thus, evidence that would show bias, motive to lie or exaggerate, or dishonesty of the witness is within the scope of Brady.

In Giglio, the Supreme Court found a Brady-type due process violation by the government's suppression of evidence of a leniency agreement with an accomplice witness. 405 U.S. at 151, 92 S.Ct. 763. Specifically, the Supreme Court stated that the accomplice witness's "credibility as a witness was . . . an important issue in the case, and evidence of any understanding or agreement as to a future prosecution would be relevant to his credibility and the jury was entitled to know of it." Id., 405 U.S. at 154-55, 92 S.Ct. 763. Thus, the suppression of such evidence violated due process.

In the present case, Sudikoff asserts that McInnes proffered various versions of his testimony during the period leading up to his immunity agreement. Sudikoff argues that the proffers and any notes from proffer sessions are Brady and Giglio because they "will bear directly on Mr. McInnes' credibility, as well as the motives for Mr. McInnes' story incriminating Mr. Sudikoff." (Mot. at 6.) In addition, Sudikoff claims that the proffers will be admissible at trial to impeach as prior inconsistent statements of a witness. ( Id.; Reply at 4-6.)

The government disputes Sudikoff's assertion that the materials would be admissible as impeachment. (Opp. at 4-7.) The government discusses various rules of evidence, determines that the materials do not satisfy the requirements for admissibility and concludes that "[t]herefore, the notes defendant seeks are not Brady or Giglio." ( Id. at 7.)

To determine whether the proffers and related materials are discoverable, the Court must first consider the standards that evidence must meet to be discoverable under Brady. The Court will then apply this standard to the information requested in this case.

1. Standard for discoverability under Brady

Unfortunately, the standard that evidence must meet to fall within the scope of Brady and require pretrial discoverability has not been clearly stated. Therefore, before discussing the appropriate standard, the Court will address why the "materiality" standard, the usual standard associated with Brady, should not be applied in this context. The Court will then discuss the appropriate standard.

a. The materiality standard

Numerous cases define the Brady obligation in the context of appellate review considering the ramifications of a prosecutor's failure to disclose evidence. Using this post-trial perspective, Brady held that it would be a due process violation only if the suppressed evidence was "material." Courts have concluded that "[e]vidence is considered material '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.'" Ortiz v. Stewart, 149 F.3d 923, 935 (9th Cir. 1998) (quoting United States v. Bagley, 473 U.S. 667, 682, 105 S.Ct. 3375, 87 L.Ed.2d 481 (1985)).

Because a finding of a Brady violation requires the appellate court to conclude that disclosure might have affected the outcome of the trial, it is understandable that such a violation would only occur if the withheld evidence was admissible or would have led to admissible evidence. See Coleman v. Calderon, 150 F.3d 1105, 1117-18 (9th Cir.), rev'd on other grounds ___ U.S. ___, 119 S.Ct. 500, ___ L.Ed.2d ___, (1998). If the withheld evidence was not admissible in the proceeding and could not have led to evidence admissible in the proceeding, its nondisclosure could not affect the proceeding's outcome. Thus, from the post-trial perspective, the suppression of evidence that would not have changed the outcome, either because it was inadmissable or because it lacked sufficient probity, would not create a due process violation.

This standard is only appropriate, and thus applicable, in the context of appellate review. Whether disclosure would have influenced the outcome of a trial can only be determined after the trial is completed and the total effect of all the inculpatory evidence can be weighed against the presumed effect of the undisclosed Brady material. See, e.g., Giglio, 405 U.S. at 154, 92 S.Ct. 763 (finding violation because "the Government's case depended almost entirely" on the cooperating witness's testimony, making impeachment crucial); Carriger v. Stewart, 132 F.3d 463, 480 (9th Cir. 1997) (finding violation because lack of corroborating evidence at trial made impeachment evidence material); Willhoite v. Vasquez, 921 F.2d 247, 249 (9th Cir. 1990) (finding no violation because there was "sufficient evidence apart from" the withheld evidence); see also United States v. Valenzuela-Bernal, 458 U.S. 858, 874, 102 S.Ct. 3440, 73 L.Ed.2d 1193 (1982) (noting that "[b]ecause determinations of materiality are often best made in light of all of the evidence adduced at trial, judges may wish to defer" such determinations "until after the presentation of evidence"). This analysis obviously cannot be applied by a trial court facing a pretrial discovery request.

Additionally, the post-trial review determines only whether the improper suppression of evidence violated the defendant's due process rights. However, that the suppression may not have been sufficient to violate due process does not mean that it was proper. This conclusion is clear from consideration of an analogous standard of review, the standard for ineffective assistance of counsel under Strickland v. Washington, 466 U.S. 668, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984).

In Strickland, the Supreme Court set the standard for ineffective assistance of counsel as assistance that falls below reasonably objective standards and that prejudiced the defendant. As its standard for prejudice, the Supreme Court drew from Brady's materiality standard. See Strickland, 466 U.S. at 694, 104 S.Ct. 2052 ("Accordingly, the appropriate test for prejudice finds its roots in the test for materiality of exculpatory information not disclosed to the defense by the prosecution. . . ."); United States v. Spawr Optical Research, Inc., 864 F.2d 1467, 1472 n. 6 (9th Cir. 1988) ("The Strickland standard for prejudice has been considered to impose virtually the same burden on the defense as the standard for materiality in Brady claims."). Thus, the tests for Strickland and Brady are similar.

In this light, it is clear that Brady's materiality standard determines prejudice from admittedly improper conduct. It should not be considered as approving all conduct that does not fail its test. Just as unreasonably deficient assistance of counsel is improper even if it does not meet the prejudice prong of Strickland and result in a Sixth Amendment violation, so suppression of exculpatory evidence is improper even if it does not satisfy the materiality standard of Brady and result in a due process violation. Though an error may be harmless, it is still error.

Therefore, post-trial standards and cases applying them are not helpful for determining the government's disclosure obligations.

b. The proper standard

Because the definitions of materiality as applied to appellate review are not appropriate in the pretrial discovery context, the Court relies on the plain meaning of "evidence favorable to an accused" as discussed in Brady.

The meaning of "favorable" is not difficult to determine. In the Brady context, "favorable" evidence is that which relates to guilt or punishment, see Brady, 373 U.S. at 87, 83 S.Ct. 1194, and which tends to help the defense by either bolstering the defense's case or impeaching prosecution witnesses, see Giglio, 405 U.S. at 154-55, 92 S.Ct. 763. The Court notes again that in the pretrial context it would be inappropriate to suppress evidence because it seems insufficient to alter a jury's verdict. Further, "[t]he government, where doubt exists as to the usefulness of evidence, should resolve such doubts in favor of full disclosure. . . ." United States v. Van Brandy, 726 F.2d 548, 552 (9th Cir. 1984) (citing United States v. Goldberg, 582 F.2d 483, 489 (9th Cir. 1978). Thus, the government is obligated to disclose all evidence relating to guilt or punishment which might reasonably be considered favorable to the defendant's case.

Though the definition of "favorable" is straightforward, the definition of "evidence" is more complicated. It could be interpreted as including only evidence that would be admissible at trial; it could include evidence that is inadmissible but is likely to lead to other evidence that would be admissible at trial; or it could include evidence that is inadmissible and is unlikely to lead to admissible evidence but would assist the defense with trial preparation. For the reasons the Court will discuss, the Court holds that the second standard is correct: the government must disclose upon request all favorable evidence that is likely to lead to favorable evidence that would be admissible.

As to the first definition of "evidence" as applying only to admissible evidence, the Court holds that this cannot be correct. In United States v. Kennedy, the Ninth Circuit stated the following standard in the context of appellate review: "To be material under Brady, undisclosed information or evidence acquired through that information must be admissible." 890 F.2d 1056, 1059 (9th Cir. 1989). Thus, even in the context of appellate review, which imposes a strict standard of materiality, the Kennedy court held that suppression of inadmissible evidence could create a due process violation if the suppressed inadmissible evidence would have led to admissible evidence. If such inadmissible evidence can give rise to a due process violation even in the appellate review context, it must surely be disclosed under the more lenient pretrial standard. Thus, the Court holds that it would be incorrect to conclude that only admissible evidence is discoverable under Brady.

As to the third definition of "evidence" as applying even to inadmissible evidence that will not lead to admissible evidence so long as that evidence will assist in the defense's trial preparation, the Court finds this standard overly broad. Although Federal Rule of Criminal Procedure 16(a)(1)(C) requires disclosure of certain items if they are "material to the preparation of the defendant's defense," Brady and its progeny do not impose such a sweeping requirement.

This conclusion is clear from the Supreme Court's holding in Weatherford v. Bursey, 429 U.S. 545, 97 S.Ct. 837, 51 L.Ed.2d 30 (1977). In Weatherford, a government informant assured the defendant that he would not testify for the prosecution at trial but ultimately did so testify. Id. at 548-49, 97 S.Ct. 837. The Fourth Circuit held that the informant's false assurances violated Brady because they frustrated the defendant's ability to prepare for trial. Id. at 559, 97 S.Ct. 837. The Supreme Court reversed, holding that "Brady is not implicated here" because a right to discovery to assist trial preparation "does not follow from [Brady's] prohibition against concealing evidence favorable to the accused. . . ." Id. Thus, the Supreme Court held that Brady did not create a right to assistance in trial preparation.

Moreover, Brady's holding cannot be read to include a right to evidence that assists trial preparation. If Brady were concerned with the defense's ability to prepare for trial, it "would necessarily encompass incriminating evidence as well as exculpatory evidence, since knowledge of the prosecutor's entire case would always be useful in planning the defense." United States v. Agurs, 427 U.S. 97, 112 n. 20, 96 S.Ct. 2392, 49 L.Ed.2d 342 (1976). Brady, however, does not provide for disclosure of inculpatory evidence, see Bagley, 473 U.S. at 675, 105 S.Ct. 3375, and therefore should not be interpreted to concern the defense's trial preparation, see also, LeFave Israel, Criminal Procedure, § 19.5, at p. 543 (1984). Therefore, that government material could assist the defense's trial preparation is not relevant under Brady.

The Court has concluded that Brady is not narrowly limited to admissible evidence but neither is it so broad as to include evidence that would only assist in trial preparation. The Court therefore holds that Brady requires disclosure of exculpatory information that is either admissible or is reasonably likely to lead to admissible evidence.

This conclusion is consistent with the doctrinal underpinnings of Brady. In Brady, the Supreme Court was not concerned with the defense's ability to prepare for trial, it was concerned with the prosecutor's ability to corrupt the trial by allowing the introduction of false testimony. The Supreme Court stated that Brady was an extension of two prior cases: First, Mooney v. Holohan, 294 U.S. 103, 55 S.Ct. 340, 79 L.Ed. 791 (1935), in which the Supreme Court found a due process violation in a conviction that was based on perjury solicited by the government; and second, Napue v. Illinois, 360 U.S. 264, 79 S.Ct. 1173, 3 L.Ed.2d 1217 (1959), in which the Supreme Court found a similar violation when the government, although not soliciting false evidence, allows it to go uncorrected. See Brady, 373 U.S. at 86-87, 83 S.Ct. 1194.

In Brady, the prosecution presented the testimony of a witness who claimed that the defendant murdered a man. Despite the defendant's request, the prosecutor did not disclose that the witness had earlier admitted to committing the murder. Drawing on Mooney, Napue, and other cases, the Supreme Court held that such suppression could violate due process. When seen as an extension of Mooney and Napue, it becomes clear that Brady concerned the danger that a witness may testify at trial, with the jury accepting the testimony as true, when the government has in its possession evidence that is relevant to the credibility of the witness. In other words, Brady does not concern the rights of the defendant as much as it seeks to guarantee that the trial is a proceeding that "comport[s] with standards of justice. . . ." Id. at 88, 83 S.Ct. 1194.

Because Brady seeks to protect the quality and completeness of the evidence upon which the jury bases its verdict, it is understandable that information that is not likely to result in admissible evidence is irrelevant. Therefore, Brady does not require the disclosure of information that would only assist the defense in creating its trial strategy. It does, however, require the disclosure of information that is likely to result in admissible evidence that would give the jury a court a more complete basis for judging guilt or punishment.

Here, Sudikoff requests that the government disclose all of McInnes's proffers and any other material that would show how the immunity agreement was reached. The Court must decide whether this material "might reasonably be considered favorable to the defendant's case," see Bagley, 473 U.S. at 696, 105 S.Ct. 3375, and, as discussed, the Court must also consider whether it is likely to lead to admissible evidence.

2. Applying the pretrial Brady standard to Sudikoff's requests

Having concluded that the pretrial standard under Brady is evidence that may reasonably be considered favorable to the defendant's case and that would likely lead to admissible evidence, the Court will consider whether the material requested by Sudikoff falls within this standard.

a. Evidence favorable to the defendant's case

The Court holds that proffers of an accomplice witness that led to a leniency agreement and information that reveals the negotiation pursuant to which that agreement was reached might reasonably be considered favorable to the defendant's case. This is for two reasons. First, to the extent the proffers and other information reveal that the witness's proposed testimony may have varied over time, they may reveal inconsistencies relevant to the accomplice witness's credibility and within the scope of Brady. Second, to the extent the proffers and other information reveal the accomplice witness's motives and desire to seek an immunity agreement, they are relevant to the witness's credibility and within the scope of Giglio.

(1) Inconsistent stories

Because this process can be lengthy and because it often carries some of the typical negotiating give-and-take, it is possible, maybe even likely, that the witness's proposed testimony that was proffered at the beginning of the process differed in some respects from the testimony proffered at the end of the process. For example, it is likely that during initial contacts with the government the witness would proffer a less detailed version of his testimony than he would once it became more likely that an agreement would be reached. Though such variations could stem entirely from the nature of this process, a defendant implicated by the accomplice witness could reasonably argue that they stem from the accomplice witness's tendency to embroider on the truth. Thus, the existence of such variations might reasonably be held to be favorable to the defense.

Although the government might argue that all such variations are innocuous and lack probity, the Court cannot comfortably so hold. While many, if not most, of these differences probably do result only from the nature of the process, the Court cannot conclude that this is sufficient to prevent disclosure under Brady. Neither the government nor the Court is aware of the details of the defense strategy and therefore neither the government nor the Court can accurately determine which variations are important.

Moreover, while it may be reasonable to conclude that any differences are innocuous, because it might be reasonable to likewise conclude otherwise, the government should disclose the information. "[W]here doubt exists as to the usefulness of evidence, [the government] should resolve such doubts in favor of full disclosure." Van Brandy, 726 F.2d at 552. This is consistent with the Court's conclusion that the proper standard under Brady is evidence "that might reasonably be considered favorable to the defendant's case." Bagley, 473 U.S. at 696, 105 S.Ct. 3375. There is no requirement that the exculpatory nature of Brady material be indisputable.

Therefore, the Court finds that any variations in an accomplice witness's proposed testimony could be considered favorable to the defense and the existence of such differences should be disclosed under Brady.

(2) Revealing motive and desire to seek leniency

When judging the credibility of testimony, a jury may properly consider a "witness' interest in the outcome of the case and any bias or prejudice. . . ." Ninth Circuit Manual of Model Jury Instructions: Criminal 37 (1997). Because accomplice witnesses who testify pursuant to immunity agreements may be motivated by more than just the desire to tell the truth, courts require specific jury instructions regarding their credibility. See e.g., id. at 65-66, 92 S.Ct. 763 (instructing jury to consider "with greater caution" testimony given pursuant to immunity agreement or by informant); cf. Napue, 360 U.S. at 269, 79 S.Ct. 1173 ("The jury's estimate of the truthfulness and reliability of a given witness may well be determinative of guilt or innocence, and it is upon such subtle factors as the possible interest of the witness in testifying falsely that a defendant's life or liberty may depend.").

In Hoffa v. United States, 385 U.S. 293, 87 S.Ct. 408, 17 L.Ed.2d 374 (1966), the defendants had been convicted based in part on the testimony of an informant who received from the government incentives to testify. On appeal, these defendants argued that the use of such witnesses violated due process because it raised unacceptable risks of perjurious testimony. The Supreme Court rejected this argument and permitted the use of such testimony but only because "[t]he established safeguards of the Anglo-American legal system leave the veracity of a witness to be tested by cross-examination, and the credibility of his testimony to be determined by a properly instructed jury." Id. at 311, 87 S.Ct. 408. Because the informant in Hoffa "was subjected to rigorous cross-examination, and the extent and nature of his dealings with federal and state authorities were insistently explored" and because the trial judge instructed the jury to carefully consider the informant's credibility, the Supreme Court found the testimony consistent with due process. Id.; cf. Guam v. Dela Rosa, 644 F.2d 1257, 1259-60 (9th Cir. 1980) (requiring cautioning instruction when witness has received benefit for testimony).

The concerns raised by the defendants in Hoffa are widely recognized. While "[c]ourts have countenanced the use of informers from time immemorial," Hoffa, 385 U.S. at 311, 87 S.Ct. 408 (quoting United States v. Dennis, 183 F.2d 201, 224 (2d Cir. 1950) (Hand, J.)), "[b]y definition, criminal informants are cut from untrustworthy cloth and must be managed and carefully watched by the government and the courts to prevent them from falsely accusing the innocent, from manufacturing evidence against those under suspicion of crime, and from lying under oath in the courtroom," United States v. Bernal-Obeso, 989 F.2d 331, 333 (9th Cir. 1993).

Consistent with Hoffa's conclusion that the use of informants is constitutional if the existence of any incentives is disclosed and considered by the jury, courts "expect prosecutors and investigators to take all reasonable measures to safeguard the system against treachery. This responsibility includes the duty as required by Giglio to turn over to the defense in discovery all material information casting a shadow on a government witness's credibility." Bernal-Obeso, 989 F.2d at 334 (citing United States v. Shaffer, 789 F.2d 682, 689 (9th Cir. 1986)); see also Carriger v. Stewart, 132 F.3d 463, 479 (9th Cir. 1997).

Though Giglio concerned the suppression of the very existence of a leniency agreement, information that illuminates the process leading up to the agreement may "cast a shadow" on an accomplice witness's credibility in a manner that disclosure of only the agreement itself would not accomplish.

The motive behind an accomplice witness's agreement to testify may range from a simple quid pro quo to an earnest desire to disclose the truth. The defense cannot distinguish between such motives unless the government reveals information about the negotiation leading to the agreement.

This conclusion applies even if the negotiating process was short. Even if the witness made only one proffer of proposed testimony and the government immediately made an offer of leniency that was immediately accepted, such a proffer is the motivating force behind the leniency agreement and as such can reveal what the witness was willing to do in return for leniency. Again, this information is relevant to the witness's credibility.

Thus, the Court concludes that information that reveals the process by which an accomplice witness and the government reach a leniency agreement is relevant to the witness's credibility because it reveals the witness's motive to testify against the defendant. Therefore, such information is discoverable under Brady and Giglio.

3. Likely to lead to admissible evidence

As discussed earlier, evidence is discoverable under Brady only if it is itself admissible at trial or likely to lead to evidence admissible at trial. As noted, the government argues that the evidence that Sudikoff seeks would not itself be admissible. The Court need not resolve this issue, however, because it holds that the evidence is likely to lead to admissible evidence.

Even apart from the possibility that the defense might find some outside source of admissible evidence to corroborate the information it receives from the government, this information may result in admissible evidence. Defense counsel may elicit testimony from the cooperating witness based on the information; to the extent the information is inconsistent with the accomplice witness's trial testimony, the information may be admissible to impeach; and the information may be used to refresh a witness's recollection. These are just some examples of how the information would result in admissible evidence. Therefore, such information is discoverable under Brady. 4. Conclusion as to Brady and Giglio

The Court finds that any information that reveals any variations in the proffered testimony of an accomplice witness testifying pursuant to a leniency agreement is relevant to the witness's credibility and therefore must be disclosed under Brady. In addition, any information that reveals the nature of the negotiation process that led to the leniency agreement is relevant to the witness's motives to testify and must be disclosed under Giglio.

B. Jencks

Sudikoff argues that any proffers by witness McInnes are also discoverable under the Jencks Act, 18 U.S.C. § 3500. The Jencks Act provides that the government must disclose to the defense any statement of any witness in the possession of the government if the statement relates to the witness's testimony. 18 U.S.C. § 3500(b). The Act defines a "statement" as

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

(2) a stenographic, mechanical, electrical, or other recording, or a transcription thereof, which is a substantially verbatim recital of an oral statement made by said witness and recorded contemporaneously with the making of such oral statement; or

(3) a statement, however taken or recorded, or a transcription thereof, if any, made by said witness to a grand jury.

In the present case, the Court cannot rule on whether the material Sudikoff seeks would fall within the scope of the Jencks Act because the nature of the proffers or any notes from proffer sessions have not been sufficiently described by the parties. This determination may require in-camera review of these materials. The Court will merely highlight certain considerations that would apply to the decision of whether these materials might be discoverable under the Jencks Act.

First, Sudikoff seeks proffers that were presented to the government by McInnes's lawyer, not by McInnes. In oral argument, the government raised the issue of whether there is any privilege that would be improperly violated should the information be disclosed. The Court holds that the disclosure of witness proffers would not violate any privilege.

The Supreme Court defined the scope of lawyer-client privilege in Supreme Court standard 503. When restating the right of a client to prevent disclosure of "confidential communications," the Supreme Court defined "confidential communications" as those that are "not intended to be disclosed to third persons. . . ." See McLaughlin, et al., Weinstein's Federal Evidence § 503.01, at p. 503-9 (1998). Although this standard was never adopted by Congress, it restates the common law scope of privilege that was adopted by Federal Rule of Evidence 501 and therefore can be relied upon. See id. at § 503.2 n. 2 (citing cases relying on Supreme Court Standards).

Thus, if a client communicates with his lawyer for the purpose of having that lawyer relay that communication to a third party, the communication is not "confidential" and not protected by lawyer-client privilege. See e.g., United States v. Oloyede, 982 F.2d 133, 141 (4th Cir. 1992) (no privilege for communications intended to be relayed to INS); Esposito v. United States, 436 F.2d 603, 606 (9th Cir. 1970) (no privilege for communications intended to be relayed to court); McLaughlin, supra, at § 503.15. Therefore, the Court holds that a client's communications of proposed testimony made with the intent that the lawyer relay the communications to the government are not protected by the lawyer-client privilege.

Another issue that must be considered if the Court were to rule on the applicability of the Jencks Act to the materials sought in this case is whether the materials fall within the Act's definition of a witness "statement." As quoted above, the Jencks Act creates three general categories for statements: (1) written statements made, signed, adopted, or approved by the witness; (2) substantially verbatim recitals of oral statements made contemporaneously with the oral statement; and (3) recorded statements to grand juries. 18 U.S.C. § 3500(e).

The last category clearly does not apply to the material sought in this matter; but the first two categories may apply. First, if a witness submits a written proffer of proposed testimony, the written proffer should fall within the first category. This may be true even if the written proffer is proposed by the witness's attorney. It is reasonable to conclude that the attorney would only submit such material if it was approved by his client, the witness. If so, the proffer may fall within 18 U.S.C. § 3500(e)(1). Whether this is an accurate reflection of the material sought in any given case will depend on the district court's judgment of the nature of the proffers.

Second, if the witness actually made statements during proffer sessions and those statements are reflected in contemporaneous notes made by those who attended the session, these notes may fall within the second category of statements under § 3500(e). Again, in-camera review may be necessary to determine whether the notes are substantially verbatim recitals.

A final issue that a court should consider when determining the application of the Jencks Act to these materials is whether the information "relates to the subject matter as to which the witness has testified." 18 U.S.C. § 3500(b). The purpose of the Jencks Act is to allow the defense the opportunity to impeach prosecution witnesses if their prior statements were inconsistent with their testimony at trial. See United States v. Brumel-Alvarez, 991 F.2d 1452, 1458 (9th Cir. 1992); United States v. Dupuy, 760 F.2d 1492, 1496 (9th Cir. 1985). Consistent with this purpose, the Act restricts disclosure to those statements that "relate to the subject matter" of the witness's testimony. 18 U.S.C. § 3500(b). The reason for this is simple: if the statement does not relate to the witness's trial testimony, it cannot be used to impeach that testimony and is therefore beyond the scope of the Jencks Act.

The Ninth Circuit has defined the Jencks obligation as follows: "it is sufficient that '[i]n determining whether the statements in question "related to" the direct testimony of the witness, it must relate generally to the events and activities testified to.'" Brumel-Alvarez, 991 F.2d at 1464 (alterations and emphasis in original) (citations omitted) (quoting United States v. Derrick, 507 F.2d 868, 871 (4th Cir. 1974)); see also 125 A.L.R.Fed. 157 (1995).

In the present case, although statements reflecting proposed testimony would likely satisfy this standard, it is possible that some statements may not. As discussed earlier, a witness seeking an agreement with the government could engage in negotiation to attempt to receive a favorable agreement. If such a witness made statements only in an attempt to "feel out" the government's position, these statements might not be considered to relate to the proposed testimony; instead of reflecting proposed testimony, they may reflect the negotiation process. If so, they would not be probative as to the content of the witness's testimony and would therefore not be useful as impeachment. Therefore, they would not fall within the scope of the Jencks Act.

In conclusion, without reviewing the requested materials the Court will not decide whether they fall within the scope of the Jencks Act. Because the Court concluded earlier that much of the requested material is likely to fall within the scope of Brady and Giglio, the Court will not now order an in-camera review. The Court, however, has pointed out a number of considerations that would apply should a decision as to the Jencks Act become necessary.

III. Conclusion

While the use of witnesses testifying pursuant to leniency agreements may be necessary, the government must take extra precautions when using such witnesses.

The need for disclosure is particularly acute where the government presents witnesses who have been granted immunity from prosecution in exchange for their testimony. We have previously recognized that criminals who are rewarded by the government for their testimony are inherently untrustworthy, and their use triggers an obligation to disclose material information to protect the defendant from being the victim of a perfidious bargain between the state and its witness.

The government must disclose to the defendants all proffers by any witnesses receiving any benefit, whether immunity or leniency, in return for testimony. Included in this category are any proffers made by lawyers for such witnesses. By "proffers" the Court refers to statements that reflect an indication of possible testimony, whether or not it seems likely that the witness would actually so testify. In addition, the government must disclose any notes or documents created by the government that reflect this information. Further, the government must disclose any material that indicates any variations in the witness's proffered testimony.

The government must also disclose to the defendants any information in its possession that reveals the negotiation process by which the immunity agreement was reached. This includes materials authored by a witness, a witness's lawyer, or the government.

Though the Court orders the government to disclose this information, this order is not intended to indicate that any of this information will itself be admissible at trial. If any party wishes to introduce at trial any of the material disclosed pursuant to this order, it must file a motion in limine seeking a ruling on admissibility at least two weeks prior to trial.

IT IS SO ORDERED.