dissenting.
I dissent. In my view, because the refusal of the court below to close the trial is not a final decision, we lack appellate jurisdiction over the case under 28 U.S.C. § 1291. However, the extraordinary nature of appellant’s claim, which involves both an allegedly life and death situation for him and a dramatically unsettled area of law, makes this one of those rare instances where we should treat the appeal as a petition for writ of mandamus. For the following reasons, I would grant the petition for writ of mandamus and order the trial court to close to the public the testimony about appellant’s past activities as a police informant.
I. Appellate Jurisdiction
Our jurisdiction to review the district court’s decision depends on 28 U.S.C. § 1291, which provides for review of “final decisions.” Appellate jurisdiction is not thereby limited only to review of final judgments. “While a final judgment always is a final decision, there are instances in which a final decision is not a final judgment.” Stack v. Boyle, 342 U.S. 1, 12, 72 S.Ct. 1, 7, 96 L.Ed. 3 (1951) (opinion of Jackson, J.). Where a “collateral order” disposes of some right not bound up with the final judgment in the case, the order may be a reviewable “final decision” under § 1291. See generally Cohen v. Beneficial Industrial Loan Corp., 337 U.S. 541, 69 S.Ct. 1221, 93 L.Ed. 1528 (1949). Examples of such final decisions in criminal cases include: denial of bail, Stack v. Boyle, supra, 342 U.S. 1, 72 S.Ct. 1, 96 L.Ed. 3; denial of a pretrial motion to dismiss on double jeopardy grounds, Abney v. United States, 431 U.S. 651, 97 S.Ct. 2034, 52 L.Ed.2d 651 (1977); and denial of a motion to dismiss an indictment because the allegedly criminal conduct was immunized by the Speech and Debate Clause, Helstoski v. Meanor, 442 U.S. 500, 99 S.Ct. 2445, 61 L.Ed.2d 30 (1979). The Supreme Court has, however, emphasized that appeals prior to final judgment are strongly disfavored in criminal cases where swift and economical dispensation of justice is especially critical. “The rule of finality has particular force in criminal prosecutions because ‘encouragement of delay is fatal to the vindication of the criminal law.’ ” United States v. MacDonald, 435 U.S. 850, 853-54, 98 S.Ct. 1547, 1549, 56 L.Ed.2d 18 (1978), citing Cobbledick v. United States, 309 U.S. 323, 325, 60 S.Ct. 540, 541, 84 L.Ed. 783 (1940). See also Abney v. United States, supra, 431 U.S. at 656, 97 S.Ct. at 2038; DiBella v. United States, 369 U.S. 121, 82 S.Ct. 654, 7 L.Ed.2d 614 (1962); Parr v. United States, 351 U.S. 513, 76 S.Ct. 912, 100 L.Ed. 1377 (1956); Cogen v. United States, 278 U.S. 221, 49 S.Ct. 118, 73 L.Ed. 275 (1929).
*326The requirements for an appealable collateral order have been set forth by this court:
(1) [T]he order must be a final determination of a claim of right “separable from, and collateral to,” rights asserted in the action;
(2) it must be “too important to be denied review,” in the sense that it “presents a serious and unsettled question”; and
(3) its review cannot, in the nature of the question that it presents, await final judgment because “when that time comes, it will be too late effectively to review the * * * order and rights conferred * * * will have been lost, probably irreparably.”
United States v. Barket, 530 F.2d 181, 185 (8th Cir. 1975), cert. denied, 429 U.S. 917, 97 S.Ct. 308, 50 L.Ed.2d 282 (1976), citing 9 J. Moore, Moore’s Federal Practice § 110.10 (2d ed. 1970).
First, the right claimed by appellant is inextricably bound up with the trial of his case. The essential question — whether appellant will receive a fair trial — involves disposition on the merits. Here, appellant claims essentially that, because of the dangers to himself and his family, he cannot effectively present in open court a defense based on his claim that he was an informant and therefore did not act with the necessary criminal intent. In general, such due process issues, like speedy trial issues will not be finally resolved until after the trial. “Before trial, of course, an estimate of the degree to which delay has impaired an adequate defense tends to be speculative.” United States v. MacDonald, supra, 435 U.S. at 858, 98 S.Ct. at 1551.
Similarly, appellant’s claim in the present case is that he will be denied a fair trial. The language and holding of MacDonald therefore forecloses appellate jurisdiction. Appealable collateral orders in criminal cases have involved only the denial of rights that do not depend on how the trial itself is conducted. See Stack v. Boyle, supra, 342 U.S. 1, 72 S.Ct. 1, 96 L.Ed. 3 (denial of bail held appealable). By contrast; this appeal hinges on the conduct of the trial. It is distinguishable from appeals of failure to dismiss an indictment that runs afoul of the speech and debate clause or double jeopardy clause, wherein a defendant “is contesting the very authority of the Government to hale him into court to face trial on the charge against him.” Helstoski v. Meanor, supra, 442 U.S. at 507, 99 S.Ct. at 2449 (emphasis in original), citing Abney v. United States, supra, 431 U.S. at 659, 97 S.Ct. at 2040. In the present case, appellant claims the refusal to close his trial “has frustrated his ability to establish his innocence of the crime charged. Normally, it is only after trial that that claim may fairly be assessed.” United States v. MacDonald, supra, 435 U.S. at 860, 98 S.Ct. at 1552. Thus, the third requirement of a collateral order is not met.
The majority opinion recognizes that in general this court lacks appellate jurisdiction until entry of the final judgment (i. e., a conviction and sentence) over a decision below not to close proceedings. See 319-320 n. 2. Nonetheless, the majority assumes jurisdiction “because of the fallout from the Gannett case and the proliferation of cases coming on line dealing with the issue of closure of open and public proceedings . . . .” Id. In short, the majority takes jurisdiction of this case because of its “extraordinary nature.” However, the Supreme Court has admonished that the facts of a particular case should not be a basis for appellate jurisdiction:
The Court of Appeals’ ... rationale — that it was the “extraordinary nature” of [appellant’s] claim that merited interlocutory appeal, even though not all [such] claims would be so meritorious — is also unpersuasive. “Appeal rights cannot depend on the facts of a particular case.” Carroll v. United States, 354 U.S. 394, 405 [, 77 S.Ct. 1332, 1339,1 L.Ed.2d 1442] (1957). The factual circumstances that underlie a [particular] claim, however “extraordinary,” cannot establish its independent appealability prior to trial.
*327United States v. MacDonald, supra, 435 U.S. at 857-58 n. 6, 98 S.Ct. at 1551 n. 6. In my opinion, appellant’s claim is essentially an interlocutory appeal over which we lack appellate jurisdiction.
II. Mandamus Jurisdiction
However, I think the extraordinary nature of this case does mandate treatment of the appeal as a petition for writ of mandamus. E. g., Wilkins v. Erickson, 484 F.2d 969 (8th Cir. 1973). See also Iowa Beef Processors, Inc. v. Bagley, 601 F.2d 949, 953 n. 3 (8th Cir. 1979); Hartland v. Alaska Airlines, 544 F.2d 992, 1001 (9th Cir. 1976); id. at 1003-04 & n. 2 (Wallace, J., concurring) (cases cited therein).
This case reaches us in an extraordinary posture factually as well as legally. The facts indicate the grave nature of appellant’s predicament. The prosecution does not dispute appellant’s position that he was a secret police informer who provided key evidence in several major criminal trials which resulted in convictions; the law enforcement authorities have in the past kept his informant status secret on some occasions even at the expense of dropping several prosecutions. Some of appellant’s former associates have committed violent acts against people. Legally, the issue of a defendant’s right to close his trial in these circumstances is one of first impression in this court. Moreover, the district court’s decision failed to reconcile the refusal to close at least part of appellant’s trial with an abundance of precedent allowing closure to protect prosecution witnesses, including rape victims and police informants, by closing their testimony to the public. (These closure cases are more fully elaborated below in the discussion of the merits.)
Mandamus is the appropriate vehicle for this court to review a decision where “there are present extraordinary and exceptional circumstances which require the grant of the extraordinary remedy to prevent a grave miscarriage of justice which might otherwise result . . . .” Hartley Pen Co. v. United States District Court, 287 F.2d 324, 327 (9th Cir. 1961). See also Iowa Beef Processors, Inc. v. Bagley, supra, 601 F.2d at 953 — 54; Gialde v. Time, Inc., 480 F.2d 1295, 1302 (8th Cir. 1973) (Heaney, J., concurring and dissenting). “It is, of course, well settled, that the writ [of mandamus] is not to be used as a substitute for appeal . . . even though hardship may result from delay and perhaps unnecessary trial . . . . The writ is appropriately issued, however, when there is . a clear abuse of discretion . . . .” Schlagenhauf v. Holder, 379 U.S. 104, 110, 85 S.Ct. 234, 238, 13 L.Ed.2d 152 (1964) (emphasis added) (citations omitted). See also Will v. United States, 389 U.S. 90, 107, 88 S.Ct. 269, 280, 19 L.Ed.2d 305 (1967) (Black, J., concurring); La Buy v. Howes Leather Co., 352 U.S. 249, 77 S.Ct. 309, 1 L.Ed.2d 290 (1957); 16 Wright, Miller, Cooper & Gressman, Federal Practice & Procedure, Jurisdiction §§ 3932-3936 (1977).
[A] number of courts, including this one, have found mandamus to be an appropriate vehicle to review orders compelling . testimony claimed to be privileged or covered by other more general interests in secrecy. See, e. g. Diversified Industries, Inc. v. Meredith, 572 F.2d 596 (8th Cir. 1978) (en banc) (attorney-client privilege); Usery v. Ritter, 547 F.2d 528 (10th Cir. 1977) (disclosure of identity of informer in FLSA case); Breed v. United States Dist. Ct., 542 F.2d 1114 (9th Cir. 1976) (disclosure of personnel and inmate files of state youth authority); Pfizer, Inc. v. Lord, 456 F.2d 545 (8th Cir. 1972) (attorney-client privilege); United States v. United States Dist. Ct., 444 F.2d 651 (6th Cir. 1971), aff’d, 407 U.S. 297, 92 S.Ct. 2125, 32 L.Ed.2d 752 (1972) (disclosure to conspiracy defendant of his monitored conversations); Harper & Row Publishers, Inc. v. Decker, 423 F.2d 487 (7th Cir. 1970), aff’d by an equally divided court, 400 U.S. 348, 91 S.Ct. 479, 27 L.Ed.2d 433 (1971) (attorney-client privilege); Hartley Pen Co. v. United States Dist. Ct., 287 F.2d 324 (9th Cir. 1961) (disclosure of secret formula acquired under license forbidding disclosure). .
Iowa Beef Processors, Inc. v. Bagley, supra, 601 F.2d at 953-54.
*328While the difference between the mandamus approach I suggest and the appellate jurisdiction asserted in Judge Gibson’s opinion may seem purely formal at first glance, I believe there are important reasons for making the distinction. First, the majority opinion broadly holds, “Closure would be improper under these circumstances.” At 325. I do not think we should pretermit due process questions which may be more apparent after trial than now. (I think due process, however, requires at this point that the proceedings be closed in part. See part III infra.)
Secondly, we should give a more clear indication as to how we will respond to this kind of matter in the future. I agree with Judge Gibson that ordinarily the question of closure is a matter of the trial court’s discretion and that it would ordinarily be inappropriate to halt the trial court’s proceedings to review the decision. At 319-320 n. 2, citing Gannett Co. v. DePasquale, 443 U.S. 368, 398, 99 S.Ct. 2898, 2915 (1979) (Powell, J., concurring). But the majority opinion then proceeds to make a special exception in this case and grant review. This treatment may suggest that we would decline to review all such orders in the future. I think that interlocutory review of a closure decision should be available only in extraordinary cases by writ of mandamus. I would apply this rule in this case instead of granting appellate review, since we all seem to agree appellate review should generally be foreclosed at this stage in the proceedings.
III. The Merits
I have reviewed the record, including the transcript of the in camera hearing held below, to ascertain the basis for appellant’s closure motion. Appellant testified concerning the dangers he perceived from specific persons if his previous informing activities would be revealed. Although the prosecution thoroughly cross-examined appellant, bringing out factors that might reduce the danger, it did not offer any affirmative evidence to refute appellant’s testimony or demonstrate that appellant’s apprehension was feigned or unreasonable.
The decision of the court below left Powers with a choice between omitting details of his informing activities from his defense or disclosing publicly information that might well tend to endanger himself and his family.
The district court held that
an order of closure would not be proper unless the defendant established:
(1) by clear and convincing evidence that there was a clear and present danger of actual harm to the life of the defendant or a member of his family;
(2) consent to the closure on the part of the prosecution, see Gannett v. DePasquale, supra, or a compelling reason for not requiring the prosecution’s consent:
(3) by clear and convincing evidence that the closure of the defendant’s criminal trial would effectively prevent the harm alleged and proven to exist; and
(4) by clear and convincing evidence that there are no effective alternatives to the complete closure of the criminal trial.
The Court concludes that the defendant failed to prove by clear and convincing evidence that the open court presentation of his defense would result in a clear and present danger of actual harm to the life of the defendant or a member of his family. The evidence at most demonstrates a fear of harm based on the allegedly violent nature of some of the persons involved. Admittedly, the defendant has shown that the prosecution would never consent to the closure because the issue before the Court focuses on a possible defense to the prosecution’s charges against the defendant. Powers, however, failed to prove that closure would effectively prevent the alleged harm, and provided the Court with no insight on how to prevent the jurors and witnesses from discussing the case subsequent to the trial, or on how to prevent the disclosure of information if the trial court’s decision is appealed. In addition, *329it was brought to the attention of the Court in the closed portion of the hearing that the defendant may generally be known to have been an informant for the Federal Bureau of Investigation and the Iowa Department of Criminal Investigation.
Finally, there is an alternative that the Court believes will enable the defendant to present his defense without closing the trial to the public. Such alternative procedure is [a generally-worded stipulation that appellant had in the past been an informant]. Therefore, the defendant’s Motion to Exclude the Public During Trial shall be denied because the defendant has failed to establish the necessary circumstances set forth above.
United States v. Powers, 477 F.Supp. 497, 498-99 (S.D.Iowa 1979). In my view, both the standard and the result embody an unduly restrictive approach to the court’s power to guarantee Powers a fair trial.
In my analysis, this case involves the most fundamental aspects of due process. “The right of an accused in a criminal trial to due process is, in essence, the right to a fair opportunity to defend against the State’s accusations.” Chambers v. Mississippi, 410 U.S. 284, 294, 93 S.Ct. 1038, 1045, 35 L.Ed.2d 297 (1973). In the landmark case of United States v. Nixon, 418 U.S. 683, 94 S.Ct. 3090, 41 L.Ed.2d 1039 (1974), the court held that due process rights of criminal defendants to obtain evidence needed for their defense outweighs other constitutionally based interests such as a presidential claim that certain communications were privileged. “The need to develop all relevant facts in the adversary system is both fundamental and comprehensive.” Id. at 709, 94 S.Ct. at 3108. “The right to the production of all evidence at a criminal trial . . . has constitutional dimensions.” Id. at 711, 94 S.Ct. at 3109. Although the Nixon case directly involved the ability of criminal defendants to obtain evidence for their defense, the rationale applies equally strongly to the ability of defendants to present evidence in their defense at trial. Thus, we have said, “[t]he exclusion of critical evidence under state evidentiary rules may deprive a defendant of a fair trial and amount to a denial of due process,” Johnson v. Brewer, 521 F.2d 556, 562 n. 16 (8th Cir. 1975).
Powers claims that publication of his testimony would endanger his life. Therefore, publicity would deter him from putting on evidence of past informing activities much of which is obviously relevant to his defense that he may have committed the acts charged as crimes in the capacity of an informer and lacked criminal intent. The trial court seems to have recognized that, indeed, public proceedings would tend to inhibit appellant from revealing his full activities as an informant, and suggested that appellant rely on a general stipulation as to his status as an informant without going into specifies.
Our tradition of public trials is supported by strong social interests, as Judge Gibson points out. At 323, citing Estes v. Texas, 381 U.S. 532, 583, 85 S.Ct. 1628, 14 L.Ed.2d 543 (1965) (Warren, C. J., concurring). See also In re Oliver, 333 U.S. 257, 68 S.Ct. 499, 92 L.Ed. 682 (1948); Radin, The Right to a Public Trial, 6 TEMP. L.Q. 381 (1932); Note, 36 ORE.L.REV. 345 (1957); 6 J. WIGMORE, EVIDENCE §§ 1834-36 (Chadbourn rev. ed. 1976). But in this case we measure that tradition and the reasons for the tradition against the constitutional guarantee of due process to the defendant. To do so requires, I believe, a look at the considerations which continue to breathe life into the requirement of public trials.
Commentators have pointed out three different kinds of interests protected by keeping trials open. Primary among these general factors supporting open trials is the aid given to the search for truth by the public presence. “This open examination of witnesses viva voce, in the presence of all mankind, is much more conducive to the clearing up of truth, than the private and secret examination . . . where a witness may frequently depose that in private, which he will be ashamed to testify in a public and solemn tribunal.” 3 Blackstone, *330Commentaries on the Laws of England 373 (1768) (footnote omitted).
The publicity of a judicial proceeding is a requirement of much broader bearing than its mere effect upon the quality of testimony; it would be essentially desirable and demandable on additional grounds. Nevertheless, it plays an important part as a security for testimonial trustworthiness and would exist as an independent requirement for that reason only, even where other grounds wanting
. . Its operation in tending to improve the quality of testimony is twofold. Subjectively, it produces in the witness’ mind a disinclination to falsify; first, by stimulating the instinctive responsibility to public opinion, symbolized in the audience, and ready to scorn a demonstrated liar; and next, by inducing the fear of exposure of subsequent falsities through disclosure by informed persons who may chance to be present or to hear of the testimony from others present. Objectively, it secures the presence of those who by possibility may be able to furnish testimony in chief or to contradict falsifiers and yet may not have been known beforehand to the parties to possess any information.
6 J. Wigmore, Evidence § 1834 (Chadbourn rev. ed. 1976) (emphasis in original). See also In re Oliver, supra, 333 U.S. at 270 n. 25, 68 S.Ct. at 506 n. 25.
In this case, the fact that a public trial would inhibit truthful testimony undercuts the rationale that public trials encourage truthful testimony. In this regard this case stands on an entirely different footing than the pretrial publicity cases, such as Gannett Co. v. DePasquale, supra, 443 U.S. 368, 99 S.Ct. 2898, where nothing suggests that publicity would inhibit truthful testimony. In pretrial publicity cases it is necessary to weigh bias resulting from the publicity against the encouragement of truthful testimony by open proceedings. See also Nebraska Press Ass’n v. Stuart, 427 U.S. 539, 96 S.Ct. 2791, 49 L.Ed.2d 683 (1976). By contrast, Powers’s claim makes it necessary to evaluate whether the rationale that public proceedings encourage truthful testimony applies in this case at all.
Likewise, Powers’s claim places in a different light the potential for members of the public to come forward either in aid of the defendant’s case or with information reflecting upon the testimony. The government concedes that Powers has been an informant against some offenders who may well have substantial propensity for violence. The potential for such witnesses to appear therefore may have a deterrent effect on truthful testimony. In pretrial publicity case no comparable deterrence would normally occur, no matter how prejudicial the publicity.
Pretrial publicity is indeed the norm. There is nothing so inherently dangerous about pretrial publicity that would support an inference of prejudice absent a showing of likelihood the publicity would have a prejudicial effect. But the disclosure of informing activities is not the norm. The danger of such disclosure is evident and may be inferred by the court from the fact of previous informing activities, especially involving serious crimes that may involve violence or subject participants to long prison sentences. The classic example is the case of undercover narcotics agent or informant. See, e. g., United States ex rel. Lloyd v. Vincent, 520 F.2d 1272 (2d Cir.), cert. denied, 423 U.S. 937, 96 S.Ct. 296, 46 L.Ed.2d 269 (1975).
On the basis of Powers’s in camera testimony, it seems inescapable to me that the man would be deterred from testifying about activities as an informant. The trial court made no finding that satisfies me the danger is not really great enough in this case to have a deterrent effect. The court noted evidence that Powers may be generally known as an informant. To me a general reputation as an informant is one thing; definite public knowledge that Powers informed on a particular person and contributed to a particular punishment is quite another thing. General reputation seems to me much less likely than a report of times and places of actual informing *331activities to induce violence in revenge. Moreover, Powers’s general reputation for informing may well make disclosure more, not less, dangerous to him, for this reputation would surely prompt Powers’s former associates to keep an eye on reports of his trial. Cf. Pechter v. Lyons, 441 F.Supp. 115 (S.D.N.Y.1977) (refusal to close deportation hearing of alleged war criminal).
I think that Powers has shown that requiring him to testify in public to his informing activities would inhibit rather than further the search for truth in this case. That basis for the open trial tradition therefore does not apply to the testimony concerning his informing activities. The court made no finding that would support a contrary conclusion, but instead concluded only that Powers had not adequately shown danger of actual harm. In doing so, the court applied a “clear and present danger” test and placed the burden on Powers to demonstrate the danger by clear and convincing evidence. (The trial court apparently adopted the “clear and present danger” test upon Powers’s urging; however, he did not ask for the burden of persuasion to be placed upon him.) The “clear and present danger” test was developed to determine when speech becomes so inseparably integrated with action it falls within the government’s power to regulate; thus, that test shows what speech is protected, and only indirectly what government action is prohibited. See, e. g., Gitlow v. New York, 268 U.S. 652, 672, 45 S.Ct. 625, 632, 69 L.Ed. 1138 (1925) (Holmes, J., dissenting). I think a much more apt analogy for the infringement of appellant’s due process rights is the chilling effect test which determines when government action deters the exercise of fundamental rights. E. g., National Ass’n for the Advancement of Colored People v. Alabama, 357 U.S. 449, 78 S.Ct. 1163, 2 L.Ed.2d 1488 (1958). Undisputed testimony that a defendant has been an informant and that some of his associates at the time were prone to violence is enough to demonstrate to me that the defendant would reasonably be deterred from putting on evidence about informing activities.
I do not suggest that the rationale for public trials disappears any time a party claims that public proceedings will inhibit testimony. As the majority points out, a defendant does not have a right to waive a public trial regardless of circumstances. See 321 & n. 4, citing Singer v. United States, 380 U.S. 24, 35, 85 S.Ct. 783, 790, 13 L.Ed.2d 630 (1965). Thus, a defendant must make an evidentiary showing of something which has a reasonable tendency to inhibit truthful testimony. There must also be special circumstances to suggest that closure would not increase the likelihood of untruthful testimony; in this case the prosecution concedes that Powers previously did act as an informant. Moreover, if the effort to close any part of a trial is opposed, I think the testimony in question must be so critical to the case of the party seeking closure that the inhibitive effect of non-closure would result in denial of some fundamental right, in this case Powers’s due process rights in presenting evidence critical to his defense.
The trial court may have considered the proposed general stipulation of informing status adequate to give Powers’s due process in this case, but I cannot agree. The particular details of Powers’s previous activities undoubtedly would color the jury’s impression of the activities for which he is now prosecuted. A stipulation lacks the pungency of live testimony. Therefore the general rule is that a party may refuse to stipulate and offer proof on a point admitted. “The reason for the rule is to permit a party ‘to present to the jury a picture of the events relied upon. To substitute for such a picture a naked admission might have the effect to rob the evidence of much of its fair and legitimate weight.’ ” United States v. Peltier, 585 F.2d 314, 324 (8th Cir. 1978), cert. denied, 440 U.S. 945, 99 S.Ct. 1422, 59 L.Ed.2d 634 (1979), citing Parr v. United States, 255 F.2d 86, 88 (5th Cir.), cert. denied, 358 U.S. 824, 79 S.Ct. 40, 3 L.Ed.2d 64 (1958). See also United States v. Brickey, 426 F.2d 680, 686 (8th Cir.), cert. denied, 400 U.S. 828, 91 S.Ct. 55, 27 L.Ed.2d 57 (1970); 9 J. WIGMORE, EVIDENCE § 2591 & n. 2 (3d ed. 1940). Where the *332proposed stipulation involves evidence critical to the defense of a criminal accused, this rule takes on significant due process implications.
A second general consideration supporting public trials is the effect on trial conduct. Public scrutiny tends to discourage arbitrariness and improprieties of court proceedings and to impress upon the judge, jury and other officers of the court the importance and seriousness of their task. “The right to a public trial is not only to protect the accused but to protect as much the public’s right to know what goes on when men’s lives and liberty are at stake, for a secret trial can result in favor to as well as unjust prosecution of a defendant.” Lewis v. Peyton, 352 F.2d 791, 792 (4th Cir. 1965). While the potential disclosure of Powers’s informant activities would tend to discourage truthful testimony by him about informing, disclosure would not tend to undermine this second rationale for open trials.
But this general consideration does not negate the due process concerns that favor giving Powers every fair opportunity to present testimony about informing activities in his defense, and it is evident that the two considerations can be accommodated. The closure of only those parts of the trial where Powers’s testimony is taken relieves him of whatever danger would tend to discourage him from testifying. The remainder of the trial — voir dire, witnesses not testifying on the informant issue, and other proceedings that would not directly endanger Powers because of his testimony— should remain fully open to the public. Although Powers may be concerned that, for example, jurors or others necessarily witnessing the proceedings might after trial reveal the specifics of his informing activities, the trial court could minimize the danger by admonishing them to secrecy on this matter. Cf. United States v. Sherman, 581 F.2d 1358 (9th Cir. 1978). The trial court’s converse concern that secrecy could not be preserved in an appeal presents no real problem, as the portions of the record relating to Powers’s informant activities could be treated exactly the way his in camera testimony has been treated on this appeal. Therefore, while the need for public scrutiny of court proceedings forecloses any attempt to close the entire trial in this case, closure of only the testimony concerning Powers’s informing activities allows public scrutiny while accommodating the due process considerations in allowing Powers a full opportunity to put on his defense. See United States ex rel. Latimore v. Sielaff, 561 F.2d 691 (7th Cir. 1977), cert. denied, 434 U.S. 1076, 98 S.Ct. 1266, 55 L.Ed.2d 782 (1978); Doe v. Risher, 2 MEDIA L.REP. (BNA) 1300 (D.C.Sup.Ct.1976).
Finally, the public has an interest not only in fair procedures but in knowing the substance of legal controversies. Persons who think themselves affected by the outcome or proceedings in a case have an interest in attending or reading reports of the trial. Criminal proceedings may have a deterrent effect on those contemplating crime or on abuses by police or other official powers. The public needs to know how its laws work or democracy might easily be reduced to a formality. All of these interests support keeping trials open generally. See Cox Broadcasting Corp. v. Cohn, 420 U.S. 469, 491-93, 95 S.Ct. 1029, 1044-45, 43 L.Ed.2d 328 (1975).
In this case there even may be a particularly strong public interest in observing the proceedings. (Indeed, at least one newspaper opposed any closure below, although it does not pursue its interest in this court.) The fact that the defendant has been an informer for law enforcement agencies suggest to me that the public might have especial concern in the degree of police cooperation with criminal elements. On this record there does not appear to be any impropriety, but closed trials of informants might generally suggest to the public or press a potential for a whitewash of some unsavory dealings or conversely for a vendetta by the authorities to dispose of some shady involvement on the fringe or outside of the law. The public certainly has an interest in knowing more about the whole business.
*333I do not minimize the public’s right to pursue such matters which I think have constitutional dimensions under the First Amendment guarantee of freedom of the press. See generally Smith v. Daily Mail Publishing Co., 443 U.S. 97, 99 S.Ct. 2667, 61 L.Ed.2d 399 (1979); Landmark Communications, Inc. v. Virginia, 435 U.S. 829, 98 S.Ct. 1535, 56 L.Ed.2d 1 (1978). But Powers is entitled to a fair trial , and the press must find its sources elsewhere if its presence at the trial would deny the defendant due process. Gannett Co. v. DePasquale, supra, 443 U.S. at 378, 99 S.Ct. at 2904; id. 443 U.S. at 398, 99 S.Ct. at 2915 (Powell, J., concurring); id. 443 U.S. at 439, 99 S.Ct. at 2936 (Blackmun, J., dissenting); Sheppard v. Maxwell, supra, 384 U.S. 333, 86 S.Ct. 1507, 16 L.Ed.2d 600; Estes v. Texas, supra, 381 U.S. 532, 85 S.Ct. 1628, 14 L.Ed.2d 543.
The result I would reach, that particular testimony by Powers or others concerning the specifics of his past informing activities may be closed to the public, is not at all unprecedented. Limited closure by trial courts of specific proceedings has been upheld where appropriate not only to protect the defendant’s due process rights but also on the motion of the prosecution over the defendant’s objection, where the court has found compelling reason to do so. In particular there is a line of cases in which trials have been closed to protect the identities of police informants while they testify at trial. E. g., United States ex rel. Lloyd v. Vincent, supra, 520 F.2d 1272; People v. Hinton, 31 N.Y.2d 71, 334 N.Y.S.2d 885, 286 N.E.2d 265 (1972), cert. denied, 410 U.S. 911, 93 S.Ct. 970, 35 L.Ed.2d 273 (1973). I do not say that these cases satisfy me that a trial can be closed over the defendant’s objection without violating the Sixth Amendment’s clear language affording the defendant a right to a public trial. But the cases closing informant’s testimony on the prosecution’s motion certainly indicate that the interests of a party or a witness in this kind of case may require closure in the absence of a specific statutory or constitutional provision prohibiting closure.
Notably, in at least one of the informant cases, the court closed the testimony on the mere showing that the witness had been an informant. United States ex rel. Lloyd v. Vincent, supra, 520 F.2d at 1275. Other cases of closure over the defendant’s objection include: Stamicarbon, N. V. v. American Cyanamid Co., 506 F.2d 532 (2d Cir. 1974) (closure of criminal contempt hearing to protect trade secrets); United States v. Bell, 464 F.2d 667 (2d Cir.), cert. denied, 409 U.S. 991, 93 S.Ct. 335, 34 L.Ed.2d 258 (1972) (court cleared during pretrial proceedings to protect the secrecy of the government’s “skyjacker” profile which was the subject of testimony); United States ex rel. Bruno v. Herold, 408 F.2d 125 (2d Cir. 1969), cert. denied, 397 U.S. 957, 90 S.Ct. 947, 25 L.Ed.2d 141 (1970) (court cleared for one day of all spectators to protect prosecution witness from intimidation); Harris v. Stephens, 361 F.2d 888 (8th Cir. 1966), cert. denied, 386 U.S. 964, 87 S.Ct. 1040, 18 L.Ed.2d 113 (1967) (closure to protect rape victim from public scrutiny during her testimony concerning the crime); Reagan v. United States, 202 F. 488 (9th Cir. 1913) (simil); United States ex rel. Smallwood v. LaValle, 377 F.Supp. 1148 (D.C.), aff’d, 508 F.2d 837 (2d Cir. 1974), cert. denied, 421 U.S. 920, 95 S.Ct. 1586, 43 L.Ed.2d 788 (1975) (closure during testimony of prosecution witness to avoid intimidation); Hogan v. State, 191 Ark. 437, 86 S.W.2d 931 (Ark. 1935) (closure during testimony of rape victim); Lowe v. State, 141 Ga.App. 433, 233 S.E.2d 807 (1977) (closure during testimony of rape victim); People v. Hagan, 24 N.Y.2d 395, 300 N.Y.S.2d 835, 248 N.E.2d 588, cert. denied, 396 U.S. 886, 90 S.Ct. 173, 24 L.Ed.2d 161 (1969) (closure to prevent intimidation of witness to assassination of Malcolm X). Although enumerating cases does not of course demonstrate their correctness, the above cases indicate that at least partial closure of a trial has been considered within the discretion of trial courts.
In at least two cases comparable to this one some closure has been permitted on the defendant’s motion. In Kirstowsky v. Superior Court, 143 Cal.App.2d 745, 300 P.2d 163 (1956), the defendant, accused of mur*334dering her husband, sought to defend by presenting evidence that her husband had forced her to commit repugnant sexual acts. The trial court found that she was psychically unable to testify in public about the sexual acts and ordered the trial closed. In State v. Poindexter, 231 La. 630, 92 So.2d 390 (1956), the Louisiana Supreme Court held it reversible error for the trial court to refuse to exclude spectators whose presence would intimidate a defense witness. Cf. United States v. Rios Ruiz, 579 F.2d 670, 674 — 75 (1st Cir. 1978). See also Commonwealth v. Principatti, 260 Pa. 587, 104 A. 53, 57-58 (1918). But see State v. Velasquez, 76 N.M. 49, 412 P.2d 4, cert. denied, 385 U.S. 867, 87 S.Ct. 131, 17 L.Ed.2d 95 (1966).
I do not mean to indicate that courts have been undivided on the difficult issues touched on by this case: under what circumstances and to what extent a trial may be closed in derogation of our tradition of public trials. On the contrary, a considerable diversity of opinion exists. See, e. g., United States v. Cianfrani, 573 F.2d 835 (3d Cir. 1978); United States ex rel. Laws v. Yeager, 448 F.2d 74 (3d Cir. 1971), cert. denied, 405 U.S. 976, 92 S.Ct. 1201, 31 L.Ed.2d 251 (1972); United States v. Kobli, 172 F.2d 919 (3d Cir. 1949); Tanksley v. United States, 10 Alaska 443, 145 F.2d 58 (1944); Davis v. United States, 247 F. 394 (8th Cir. 1917); Commercial Printing Co. v. Lee, 262 Ark. 87, 553 S.W.2d 270 (1977); State v. Schmit, 273 Minn. 78, 139 N.W.2d 800 (1966); People v. Jelke, 308 N.Y. 56, 123 N.E.2d 769 (1954); E. W. Scripps Co. v. Fulton, 100 Ohio App. 157, 125 N.E.2d 896, dismissed, 164 Ohio St. 261, 130 N.E.2d 701 (1955). See generally Annot., 39 A.L.R.3d 852 (1971); Annot., 4 L.Ed.2d 2128 (1960); Annot., 48 A.L.R.2d 1436 (1956); Annot., 156 A.L.R. 265 (1945); J. Cook, Constitutional Rights of the Accused, Trial Rights §§ 100-103 (1974); Note, 91 Harv.L.Rev. 1899 (1978); Note, 1966 Wash.U.L.Q. 458; Comment, 52 Mich.L.Rev. 128 (1953).
In this case, we are presented with, in my opinion, a relatively narrow issue of the need to close a trial to allow a defendant to put on the kind of informant defense which Powers seeks to present to the jury. In my view, this case presents a very different issue than would a case of closure to prevent pretrial publicity. In my opinion, different questions would be presented if the government had evidence to controvert Powers’s assertion that he was an informant, or if Powers’s claim were that he was afraid to testify that someone else had committed the crime with which he was charged. Such claims may well exemplify the kind of proceeding in which publicity serves an essential function of encouraging truthful testimony. In this case, however, the trial court expressly found that “the prosecution would never consent to the closure because the issue before the Court focuses on a possible defense to the prosecution’s charges against the defendant.” As discussed above, I think that Powers has made sufficient showing that closure of certain testimony is required if his trial is to be a “search for truth.” The stipulation proposed by the trial court is clearly inadequate to present his case effectively to the jury. Therefore, I conclude that Powers’s right to due process compels closure of testimony by Powers or by others about Powers’s informing activities. I would not, however, close proceedings other than testimony on these informing activities.
I would therefore treat this appeal as a petition for writ of mandamus and grant the petition.