(dissenting):
I disagree with both aspects of the majority opinion. As to the instruction on uncharged misconduct there is a vast difference between the failure of defense counsel to request an appropriate instruction as constituting a waiver of the trial judge’s duty to instruct on issues for determination by the court members and an affirmative request by the defense that a particular instruction not be given. In my book, the latter instance represents defense-induced error which, except in the case of a manifest miscarriage of justice, will not be considered by an appellate court as a ground for reversal of an otherwise valid conviction. Even if United States v. Graves, 23 U.S.C.M.A. 434, 50 C.M.R. 393, 1 M.J. 50 *125(1975), is as expansive as the majority now perceive it to be, a construction with which I disagree,1 it does not overrule the self-induced error concept. In United States v. Morales, 23 U.S.C.M.A. 508, 508, 50 C.M.R. 647, 647, 1 M.J. 87, 88 (1975), the Court noted that Graves did not extend to “an affirmative waiver by defense counsel of appropriate judicial action required by the evidence”. In United States v. Harden, 1 M.J. 258 (1976), the court iterated the limitation of Graves with a reference to United States v. Brux, 15 U.S.C.M.A. 597, 602, 36 C.M.R. 95, 100 (1966), in which the Court held that, except to prevent a miscarriage of justice, “self-induced error . . . may not later be claimed as the basis for appellate reversal.”
As to the public trial issue, the principal opinion acknowledges that the right to a public trial is “not absolute.” The fact that the trial judge held a preliminary hearing on the matter demonstrates to me that he was mindful of his responsibility to effect a sensitive accommodation between the accused’s right to a public trial and the Government’s need to protect classified information affecting national security. His declaration that he would “bend over backwards” to preserve the accused’s right demonstrates to me that his criterion for exclusion of the public was at least as stringent as that contemplated by the majority. The question then is whether the judge disregarded his own declaration and, in fact, wielded “an ax in place of the . required scalpel,” as the majority conclude.
Contrary to the majority’s disdain of “mere percentages of the total pages of the record,” as indicative of the scope of exclusion, in my opinion, that circumstance is very important to resolution of the issue. The defense brief represents that the trial was conducted “almost entirely in secret.” However, Government counsel’s analysis of the transcript of the record, with which I agree, indicates that over 60 percent of the proceedings were “open to the public,” and that the “bulk of the closed session of the court-martial . . . contained numerous and repeated references to classified matters.” In my opinion, therefore, the record does not reflect “blanket exclusion of the public,” as the majority describe the trial judge’s ruling, but rather it convinces me the trial judge was firmly committed to, and properly applied, the “logic and rationale governing the exclusion” of the public, which the majority posit as an appropriate standard for measuring the validity of the trial judge’s ruling. I would affirm the determination by the Court of Military Review that the accused was not improperly denied the right to the presence of the public at portions of his trial.
Turning to the standards postulated by the majority by which to determine when to exclude the public, I am constrained to express some of my misgivings. The majority, I believe, have developed their standards from an amalgam of cases dealing not only with the right to public trial, but with the right, under the Freedom of Information Act,2 to information contained in public records. As matter classified to protect national security interests is a basis to exclude the public from a trial and is also exempt from disclosure under the Freedom of Information Act, the admixture has a surface appeal. But, in my opinion, it is wrong to measure the correctness of a ruling excluding the public by the requirements of the Information Act.
The majority say that a “witness whose testimony is only partially concerned with . [classified material] should testify in open court on all other matters.” The majority thus seem to equate exclusion of the public during testimony by a particular witness to the process of piecemeal excision that a trial judge may undertake in an action under the Freedom of Information *126Act to separate unclassified matter from the classified in a particular public record. I do not believe that sort of easy compartmentalization of the testimony of a witness, especially in a prosecution of this kind, is so commonplace and practicable as to be elevated to a standard for decision. But more importantly, the picture conjured up by this standard is that of a series of entrances and exits by the public, as the witness oscillates in his testimony between the classified and the unclassified.
In Stamicarbon, N.V. v. American Cyanamid Co., 506 F.2d 532, 537 (2d Cir. 1974), the Second Circuit Court of Appeals perceived “frequent shuttling between public and in camera proceedings” as presenting a serious risk of prejudice to the fairness of a trial. Inferentially, the majority concede their standard entails the risk that, in a trial with court members, the closed session testimony might be accorded greater weight, merely because of its apparent greater significance; they would control the risk by requiring the trial judge “to sua sponte instruct the court members both as an introductory matter and in greater detail during his final instructions as to the underlying basis for the use of this bifurcated process.” In my opinion, the better approach is to assess the probable extent to which a witness’ testimony will deal with classified material. If the major part of the testimony is concerned with such matter, or such matter is embraced in many different parts of the anticipated testimony, then the whole of the testimony should be given with the public excluded. I believe that approach is not only conducive to a more orderly trial than the multiple exclusions demanded by the majority’s standard, but it minimizes, if it does not entirely eliminate, the risk that the court members might, from the comings and goings of the public, give added weight to the testimony heard in the nonpublic sessions. Further, I believe the approach I suggest will tend to lessen, rather than encourage, later claims by an accused that the public should not have been excluded for a particular series of questions and answers because their subject matter did not deal with classified information.
A second disturbing aspect of the majority’s standards is their apparent disregard of what to me is a very important difference between a Freedom of Information case and exclusion of the public at a trial. In a Freedom of Information Act proceeding, if the trial judge mistakenly determines that classified documents are not entitled to exemption from disclosure because not properly classified, his decision is appealable; and his decision can be stayed pending appeal. See Department of the Air Force v. Rose, 425 U.S. 352, 96 S.Ct. 1592, 48 L.Ed.2d 11 (1976). Consequently, national defense interests sought to be safeguarded by the classification are protected until the trial ruling can be reviewed. No such protection is accorded the Government against a ruling by a trial judge refusing exclusion of the public from the whole or part of a trial because classified information is involved. True, the Government must disclose classified evidence to the accused and his counsel, as the accused has the right to know the evidence against him. Alderman v. United States, 394 U.S. 165, 181, 184, 89 S.Ct. 961, 22 L.Ed.2d 176 (1969). See also United States v. Nichols, 8 U.S.C.M.A. 119, 125, 23 C.M.R. 343, 349 (1957). However, because the accused has the right to know does not mean that disclosure must also be made to the public in general, especially if the trial is held in a foreign country and foreign nationals are in attendance. It seems to me, therefore, that as a trial ruling against the Government is not reviewable, the potentiality for irreparable harm to national security interests must be taken into account by the trial judge.
Aside from the Freedom of Information Act, the need to protect national security interests, and military secrets in particular, is so strong that, while recognizing a trial judge should not abdicate his responsibilities in the conduct of a trial to the Executive, the Supreme Court held that the judge should not jeopardize national security interests “by insisting upon an examination of the [classified] evidence, even . alone, in chambers.” United States v. *127Reynolds, 345 U.S. 1, 10, 73 S.Ct. 528, 533, 97 L.Ed. 727 (1953). The Supreme Court took the same approach to the Freedom of Information Act, as originally worded, and held that the exemption from disclosure of classified information, provided by the Act, did not “permit in camera inspection of such documents to sift out so-called ‘nonsecret components.’ ” Environmental Protection Agency v. Mink, 410 U.S. 73, 81, 93 S.Ct. 827, 833, 35 L.Ed.2d 119 (1973). In its wisdom, Congress later amended the Act to allow the trial court to examine classified records in camera and to “determine the matter de novo.” 5 U.S.C. § 552(a)(4)(A); Department of the Air Force v. Rose, supra.
It may be that the Freedom of Information Act’s empowerment of the court to determine the correctness of a security classification and other developments in the area of in camera proceedings have undermined the limitation on the judge’s authority propounded in Reynolds, supra. See Alderman v. United States, supra, 394 U.S. at 198-99, 89 S.Ct. 961 (Harlan, J., concurring in part and dissenting in part). It may also be that the basic policy of the act favoring disclosure over secrecy requires that, in a trial under the act, a judge resolve, in favor of disclosure, any doubt he may have as to the correctness of a classification of particular matter as information affecting national security interests. However, I believe that neither individually nor in combination do these circumstances support the procedure prescribed by the majority to determine whether the public should be excluded during the presentation of classified matter at a trial.
In footnote 14 of the principal opinion, the trial judge is portrayed as the bulwark against an “arbitrary and capricious” classification. A judge can also be arbitrary and capricious. An accused who is subject to an abuse of power in the denial of a public trial may obtain relief on appeal. But when the Government is subject to an arbitrary and capricious ruling, which requires disclosure of classified matter to the public, it has no means of relief, short of terminating the trial and dropping the charges. I do not believe, therefore, that the Government is obligated, as the majority indicate it is, to demonstrate “a compelling need” to exclude the public. It seems to me that a burden of that magnitude is not only inconsistent with the majority’s earlier suggestion that “[a]ll that must be determined” is that the classification was made by a proper authority in accordance with regulations, but is also wrong in principle. I believe that the correct test is whether the classified information is of a nature that presents a real, not merely colorable, basis for a conclusion that national security interests are implicated. If there is a rational doubt, that doubt must be resolved in favor of the Government. Balancing the respective interests of the accused and the Government in this way, strikes me as eminently preferable to that propounded by the majority. It takes account of, and tries to guard against, the risks of a wrong ruling by the trial judge. If the trial is closed to the public, the accused may still be acquitted; thus he has suffered no harm from the several evils deemed to exist in a closed trial and the Government has preserved its classified information from public disclosure. If, as here, the accused is convicted, and the exclusion of the public from trial was improper, and of a magnitude justifying reversal, the accused will get a new trial. And finally, if the accused is convicted and the exclusion of the public was proper, then neither he nor the Government has been deprived of a right. The intention of the majority to preserve the right to a public trial is laudable, but I just cannot agree with the course they have chosen to fulfill that purpose.
A third issue on which we granted review was whether the findings of guilty of the offense of willfully attempting to communicate information relating to national defense, in violation of 18 U.S.C. § 793(d), (Charge I and its specification), must be reversed because the statute “is unconstitutional as applied in this case.” The argument in support of the assignment of error is a mosaic of alleged instructional deficiencies. Some of the purported deficiencies relate to matters that are wholly evidentia*128ry in nature, with no discernible connection to a constitutionally impermissible application of the statute. For example, it is contended that the trial judge should have directed the court members to disregard at least part of the testimony of a Government witness because he was not properly qualified as an expert “on why . . . [the information which was the subject matter of the charge] was classified or how a person might have reason to believe that its release could injure the national defense” [emphasis in original], and that he should have rejected, or charged the court members to disregard, certain testimony as “irrelevant” to one of the essential elements of the offense. Other aspects of the argument, which also seem to me not to raise any question as to the constitutionality of § 793, attack the correctness of various parts of the instructions that were given. In this category is the contention that the trial judge erred in instructing the court members, if they determined the information in issue bore a security classification, they could consider that fact “with all the other evidence in determining whether” the information “relates to the national defense.”3 I am satisfied that none of the asserted instructional deficiencies, either alone or in combination, demonstrate a constitutional infirmity in the application of § 793 or justify reversal of the findings of guilty of Charge I and its specification.
I would affirm the decision of the Court of Military Review.
. See my dissent in United States v. McGee, 23 U.S.C.M.A. 591, 594, 50 C.M.R. 856, 859, 1 M.J. 193, 196 (1975), and United States v. Nelson, 1 M.J. 235 (1975).
. 5 U.S.C. § 552.
. As to the correctness of this instruction, I agree with appellate Government counsel that classification of information alleged to relate to national defense by the Government is a proper factor for the court members’ consideration in a case of this kind. Such classification is evidence that the Government has acted to protect the information from public domain. See United States v. Drummond, 354 F.2d 132, 152 (2d Cir. 1965), cert. denied, 384 U.S. 1013, 86 S.Ct. 1968, 16 L.Ed.2d 1031 (1966); United States v. Soblen, 301 F.2d 236 (2d Cir. 1962), cert. denied, 370 U.S. 944, 82 S.Ct. 1585, 8 L.Ed.2d 810 (1962); United States v. Heine, 151 F.2d 813 (2d Cir. 1945). Further, I discern no reasonable risk that the instruction could have misled the court members to conclude “that proof of classification was sufficient to prove relation to the national defense.”