BUTZNER, Senior Circuit Judge:

Richard Craig Smith awaits trial on charges of violating the Espionage Act, 18 U.S.C. §§ 793(d), 794(a) (c). Because he gave notice that he expects to disclose classified information, the district court conducted a hearing prescribed by section 6(a) of the Classified Information Procedures Act, 18 U.S.C. app. 3. Invoking the Act's interlocutory review provision, section 7, the government appeals the district court's ruling that certain classified information will be admissible at Smith's trial. The government asserts that the primary error in the district court's analysis involves "a matter of law: the lower court's failure to give any consideration whatsoever to the harm resulting to the national security, and its failure to balance that harm against the relevancy of the proffered facts." It is the government's position that classified evidence that is relevant to Smith's defense must nevertheless be excluded if the court determines that Smith's need for the evidence is outweighed by the harm its disclosure will cause to the national security. The government also contends that the district court erred in determining that the facts in question met the relevancy test of Federal Rule of Evidence 401.

We conclude that the district court properly construed the Act and that its admission of classified evidence complied with Rules 401 and 403. Accordingly, we affirm the order the district court entered at the conclusion of the section 6(a) hearing.

I

The Senate Report explains the purpose of the Act:

[The bill] provides pretrial procedures that will permit the trial judge to rule on questions of admissibility involving classified information before introduction of the evidence in open court. This procedure will permit the government to ascertain the potential damage to national security of proceeding with a given prosecution before trial.

Section 5 of the Act requires the defendant to give notice of the classified information he expects to disclose. Section 6(a) directs the court to conduct a pretrial hearing to determine "the use, relevance, or admissibility of classified information . . ."

This appeal concerns only the order admitting classified evidence that the district court entered following the 6(a) hearing. Although the briefs and this opinion necessarily discuss sections 6(c) and (e), the procedures prescribed by these sections have not yet been invoked. Consequently, this opinion should not be read as prejudging any action the district court may take pursuant to these sections.

The district court ruled that the Act is procedural and that it does not change the standards for determining relevance and admissibility. Its interpretation of the Act is fully supported by the text, legislative history, and judicial precedent.

There is a practical reason why the district court refrained from balancing the harm to the national security against the relevancy of the classified evidence Smith proffered. Section 1(a) of the Act provides:

"Classified information," as used in this Act, means any information or material that has been determined by the United States Government pursuant to an Executive order, statute, or regulation, to require protection against unauthorized disclosure for reasons of national security . . . .

It is apparent, therefore, that the government pursuant to the authority mentioned in section 1 may determine what information is classified. A defendant cannot challenge this classification. A court cannot question it.

In conformity with the Act, the government informed the district court that certain evidence was classified. But the government did not tell the court why the evidence was classified or why its disclosure would harm the national security. It did not call the person responsible for the classification, or any other witness, to testify to the nature and extent of the harm the national security would suffer if the classified evidence were admitted. Without this information, the court was "ill-equipped," as it pointed out, to balance the harm to the national security against the relevance of the evidence.

We do not suggest that the government was remiss at this stage of the proceedings in omitting information about the identifiable damage to national security and the basis for classification. The Act does not contemplate that it should have presented this evidence at the 6(a) hearing. In subsequent proceedings governed by section 6(c), the Attorney General may submit this information in connection with the government's motion for an alternative procedure for disclosure of classified evidence. At the request of the government, the court must consider this information in camera and ex parte.

Significantly, section 6(c) does not require the government to furnish the court information about the harm disclosure threatens. The Attorney General may choose to withhold this information from the court. Moreover, under the procedures established by section 6(c), the Attorney General furnishes the information — if at all — only after the district court has decided that the classified information is relevant. It appears to us that if Congress had intended the district court to balance national security against relevancy in the 6(a) hearing, provision would have been made for transmission of information necessary for balancing during the 6(a) hearing and not after relevancy and admissibility have been determined.

Congress intended that the information from the Attorney General should be considered by the court in determining pursuant to section 6(c) whether the alternatives suggested by the government "will provide the defendant with substantially the same ability to make his defense as would disclosure of the specific classified information."

If the court grants the motion for substitution of alternative procedures, the government's concerns are put to rest subject to the defendant's statutory right to seek reconsideration or assign error on appeal. If the court denies the government's motion, the decision whether to permit disclosure of the classified evidence ultimately rests not with the court but with the Attorney General subject to the sanctions provided in section 6(e)(2). Even at this stage of the proceedings, the legislative history discloses that the court is not authorized to balance the harm to national security. After referring to the range of sanctions available to the court, the Senate Report states: "It should be emphasized, however, that the court should not balance the national security interests of the Government against the rights of the defendant to obtain the information. The sanctions against the Government are designed to make the defendant whole again."

Federal Rule of Evidence 501 provides that unless otherwise required the privilege asserted by the government "shall be governed by the principles of the common law as they may be interpreted by the courts of the United States in the light of reason and experience." The Act wrought no change in Rule 501. Consequently, federal judicial precedent dealing with the government's claim of privilege is pertinent. The government contends that precedent buttresses its argument that the district court must balance the harm to national security against the relevancy of the classified information in determining whether the proffered evidence is admissible. Of the several cases the government cites, none is precisely on point, and only two merit extended discussion.

The government relies primarily on Roviaro v. United States, 353 U.S. 53, 77 S.Ct. 623, 1 L.Ed.2d 639 (1957), which dealt with the government's privilege to withhold the name of an informant in a narcotics prosecution. Speaking of the admissibility of the informant's identity, the Court said: "The problem is one that calls for balancing the public interest in the flow of information against the individual's right to prepare his defense." 353 U.S. at 62, 77 S.Ct. at 628-29. It is this principle that the government emphasizes.

Roviaro is, however, distinguishable. That case did not involve disclosure of evidence that the Executive had determined affected the national security. Consequently, the trial court was furnished more information than the simple fact that the government deemed that the identity should be kept secret. To sustain its privilege, the government was required to disclose to the court why it wished to maintain secrecy and the nature and extent of the harm that it would suffer by disclosure. Only with this information at hand was the court in a position to balance the public interest and the defendant's need in determining whether to admit evidence of identity. This, as we have shown, is not the situation when the court rules on the relevancy of classified information in a 6(a) hearing.

Roviaro, moreover, places a significant limitation on the privilege of the government to withhold the informer's identity. It explicitly protects the right of the accused to obtain and introduce evidence of the informer's identity when this fact is relevant and helpful to his defense. The Court said, 353 U.S. at 60-61, 77 S.Ct. at 627-28:

Where the disclosure of an informer's identity, or of the contents of his communication, is relevant and helpful to the defense of an accused, or is essential to a fair determination of a cause, the privilege must give way. In these situations the trial court may require disclosure and, if the Government withholds the information, dismiss the action.

United States v. Reynolds, 345 U.S. 1, 73 S.Ct. 528, 97 L.Ed. 727 (1953), involved the government's privilege to resist discovery of military secrets in an action seeking recovery under the Federal Tort Claims Act. The Court held that a formal claim of privilege indicating a reasonable possibility that military secrets were involved was a sufficient showing to bar discovery. In view of the plaintiffs' dubious showing of necessity, the trial court was not required to probe further to determine whether the invocation of privilege was appropriate. Moreover, the invocation of the privilege did not justify imposition of a sanction sustaining the claim of negligence. Reynolds, the government urges, is authority for excluding Smith's proffer of relevant classified evidence after balancing it against the harm of disclosure.

The government's reliance on Reynolds is misplaced. Here there was no challenge to the appropriateness of the government's invocation of the privilege, nor could there have been since the classified material satisfied the legislative definition found in section 1. Furthermore, the Court explained why the assertion of privilege in a criminal case differs from a similar claim in a civil case. The Court said in Reynolds, 345 U.S. at 12, 73 S.Ct. at 534:

Respondents have cited us to those cases in the criminal field, where it has been held that the Government can invoke its evidentiary privileges only at the price of letting the defendant go free. The rationale of the criminal cases is that, since the Government which prosecutes an accused also has the duty to see that justice is done, it is unconscionable to allow it to undertake prosecution and then invoke its governmental privileges to deprive the accused of anything which might be material to his defense. Such rationale has no application in a civil forum where the Government is not the moving party, but is a defendant only on terms to which it has consented.

The distinction between civil and criminal cases previously was recognized in United States v. Andolschek, 142 F.2d 503, 506 (2d Cir. 1944):

While we must accept it as lawful for a department of the government to suppress documents, even when they will help determine controversies between third persons, we cannot agree that this should include their suppression in a criminal prosecution, founded upon those very dealings to which the documents relate, and whose criminality they will, or may, tend to exculpate. So far as they directly touch the criminal dealings, the prosecution necessarily ends any confidential character the documents may possess; it must be conducted in the open, and will lay bare their subject matter. The government must choose; either it must leave the transactions in the obscurity from which a trial will draw them, or it must expose them fully. Nor does it seem to us possible to draw any line between documents whose contents bears directly upon the criminal transactions, and those which may be only indirectly relevant. Not only would such a distinction be extremely difficult to apply in practice, but the same reasons which forbid suppression in one case forbid it in the other, though not, perhaps, quite so imperatively.

Recently, the court observed in a civil case where the government's privilege to withhold state secrets was upheld without dismissal of the government's civil complaint: "The civil/criminal distinction is important, and was emphasized by the courts in both Reynolds and Andolscheck." Attorney General v. Irish People, Inc., 684 F.2d 928, 953 (D.C. Cir. 1982).

The Act has not altered the distinction between the exclusion of classified evidence in civil cases and its admission in criminal cases. The Act reflects the distinction, for its procedures apply only to criminal prosecutions.

In sum, we conclude that a district court conducting a 6(a) hearing to determine "the use, relevance, or admissibility of classified information" is not empowered by either the Act or judicial precedent to exclude classified evidence that is relevant to the accused's defense on the ground that the prevention of harm to the national security outweighs the defendant's need. Subject to Rule 403, the Act protects the right of the accused to introduce relevant classified information in his defense when no effective alternative is available. It leaves to the Executive the ultimate decision whether to expose the classified material subject to the sanctions the Act mandates. Indeed, this is evident from the statement of the purpose of the Act set forth in the Senate Report. See n. 2 supra. Judicial precedent is not to the contrary.

II

Quite apart from its principal assignment of error, the government complains that the district court erred in its application of the traditional relevancy test of Federal Rule of Evidence 401. We have carefully considered the government's numerous objections.

During the nine-day 6(a) hearing, the court excluded a great deal of evidence by insisting that Smith identify precisely the information he intended to use from each classified file and document that he proffered. This procedure, of course, technically dealt with Smith's obligation under section 5(a) of the Act to give notice of his intention to disclose or cause the disclosure of classified information. See United States v. Collins, 720 F.2d 1195, 1197-1200 (11th Cir. 1983). The court's rigorous application of the requirement to give specific notice eliminated at the outset much classified evidence that Smith proffered. The court also excluded relevant classified documents and proffered testimony because its probative value was outweighed by one or more of the factors mentioned in Rule 403. Other evidence was excluded simply because it was not relevant.

The court restricted Smith's use of some admissible evidence by prohibiting its use when he presents his case in chief. Under the ruling of the court, the restricted evidence can be offered by Smith only if the government opens the door for its use by cross-examination or rebuttal. Upon consideration of the record, the briefs, and arguments of counsel, all of which are classified, we conclude, for reasons fully set forth in parts IV and V of the classified version of the district court's opinion (before its redaction for publication) that the admission of the evidence that is the subject of this appeal clearly met the standards for relevancy set forth in Rule 401 and that it should not be excluded under Rule 403.

AFFIRMED.