On July 25, 1978, in what was to become one of the most controversial and well-pub*111licized events in recent Puerto Rico history, two suspected terrorists were killed in a shootout with police officers at a mountainous location known as Cerro Maravilla. Young members of a radical pro-independence group, the two were allegedly on their way to sabotage a nearby communications facility when slain. Coming as it did in the midst of a heated debate over the island’s political future — and on the eve of a closely contested electoral campaign — the Cerro Maravilla incident attracted widespread popular and political attention.
One of the several aftershocks of that incident was a federal civil rights action brought by relatives of the two deceased, in which they alleged that police officers, senior law enforcement officials and the Governor of Puerto Rico had conspired to arrange the killings. That suit, which among other things calls into question the validity of two earlier Commonwealth Justice Department reports exonerating the defendants, has itself stirred enormous interest and publicity in Puerto Rico. Media coverage of the developing litigation has 'been intense, with defendants’ deposition testimony in particular reported — together with explanations provided by both sides — virtually at the time it was given.
In response to this publicity, the district court issued several orders. The first, which was reaffirmed by that court several times, limited physical attendance at all subsequent depositions to attorneys of record and clerical assistants. That order was sustained by us on an earlier review, and is not presently before us in any fashion. The second, which the district court has also reaffirmed on reconsideration, prohibits attorneys from disclosing any evidence obtained through subsequent depositions to the press, to the litigants themselves, or to any third party. Emphasizing that all information would become public once trial began, the court described its order as a protective order governing the taking of depositions under Fed.R.Civ.P. 26(c), with the requisite “good cause” found in a “reasonable likelihood” that the wide dissemination of prejudicial publicity would otherwise deny the defendants a fair trial.
The third order issued by the district court arose in part from these developments and in part from an entirely independent sequel to the Cerro Maravilla incident. After the two Justice Department inquiries into the killings exonerated all officials of any wrongdoing, and after each investigation was attacked by opponents of the administration as biased and incomplete, the Commonwealth Senate authorized the issuance of two subpoenas for investigation-related documents in the Department’s possession. The fact that some of the documents subpoenaed were identical to those on which the district court had imposed the second protective order described above touched off what may be described as an internal combustion cycle of confrontation among three branches and two sovereigns. Invoking the district court’s protective order — which provided that “only plaintiffs, their attorneys, and paralegals of record” were to have access to the documents, and that they were “in no way to divulge their contents to any person or entity” — the Secretary of Justice refused to comply with the legislative subpoena and moved to quash it in the federal lawsuit. The Secretary asserted that complying with the subpoena would violate the protective order, affront the integrity of the court, deny him a federally-recognized privilege, and deny the defendants the fair trial the order sought to ensure. The police defendants joined the motion, and the President of the Senate intervened in the federal proceeding for the purpose of opposing the motion. The district court subpoenaed the President to testify regarding the legislature’s motives for issuing its subpoena; the President moved to quash the court’s subpoena on grounds of legislative immunity, the court held immunity inapposite, and defendants called the President to the stand. The President refused to answer questions regarding legislative motivations, the court ordered him to answer, the President refused, and the court ordered the legislative subpoenas quashed.
In the wake of these developments, three separate parties raise three distinct issues *112before us. First, intervenor The San Juan Star challenges that portion of the district court’s order prohibiting disclosure of deposition evidence to members of the press or public. Second, plaintiffs seek review of that portion of the same order barring their attorneys from disclosing the contents of depositions to them. Finally, intervenor Miguel Hernandez Agosto, President of the Puerto Rico Senate, asks us to reverse the court’s order quashing the Senate’s subpoenas. We consolidated these cases and heard them on an expedited basis.
I.
At the outset we face a question of appellate jurisdiction. Because the orders contested are interlocutory in character, we must determine whether each satisfies the criteria of Cohen v. Beneficial Industrial Loan Corp., 337 U.S. 541, 546, 69 S.Ct. 1221, 1225, 93 L.Ed. 1528 (1949), which we recently set forth as follows:
“Four requisites of appealability under this exception can be gleaned from the Cohen opinion and the cases applying it. The order must involve: (1) an issue essentially unrelated to the merits of the main dispute, capable of review without disrupting the main trial; (2) a complete resolution of the issue, not one that is ‘unfinished’ or ‘inconclusive’; (3) a right incapable of vindication on appeal from final judgment; and (4) an important and unsettled question of controlling law, not merely a question of the proper exercise of the trial court’s discretion.” In re Continental Investment Corp., 637 F.2d 1, 4 (1st Cir. 1980), quoting United States v. Sorren, 605 F.2d 1211, 1213 (1st Cir. 1979).
In addition, we observed that the third criterion, variously referred to as urgency or irreparable harm should be the “central focus” and perhaps even the “dispositive criterion” of appellate jurisdiction over such orders. Id. at 6-7. The first and second criteria — separability and finality — were described as facets of the analysis of urgency, while the last, importance, was said to be either another facet of that element or simply not relevant. Id. We turn to the orders before us in the light of these standards.
First, we consider briefly the separability and finality components. While many discovery orders are typically not sufficiently separable from the merits of the underlying dispute to meet this part of the Cohen test, see Grinnell Corp. v. Hackett, 519 F.2d 595, 597 (1st Cir.), cert. denied sub nom. Chamber of Commerce v. United Steelworkers, 423 U.S. 1033, 96 S.Ct. 566, 46 L.Ed.2d 407 (1975), we think the orders in this case do satisfy the requirement. All issues presently before us derive from the district court’s attempts to limit publicity concerning the civil rights action before it, and all are entirely independent of the substantive matters at issue in that action. Resolution of these collateral disputes at this time will not disrupt the trial below, and might well facilitate its conclusion by freeing the court from the recurrence of what has obviously been a constant irritant. Similarly, we think it apparent that the district court’s judgment on all matters before us has now been rendered in its final form. That court has reaffirmed its position with respect to both plaintiffs and The San Juan Star on numerous occasions without material modification, and has issued a decisive order following a climactic confrontation with intervenor Hernandez Agosta. Both orders are final and complete by their terms, and the district court has manifested no contrary view of them.
We thus come to the central, and most difficult, object of inquiry: inseparability of harm, or the possibility of vindicating the various rights asserted on appeal from a final judgment in the underlying litigation. We note initially that the question “turns on whether irreparable harm would result to appellants, not from the district court order itself, but from a delay in obtaining appellate review of that order”. In re Continental Investment Corp., supra, 637 F.2d at 5. In addition, we reiterate our oft-stated view that potential burdens of litigation or relitigation cannot alone constitute the requisite harm. Id. at 5-6, quoting Lamphere v. Brown Univ., 553 *113F.2d 714, 717 (1st1 Cir. 1977). Because the potential for irreparable harm differs significantly for each party seeking review, we address the applicability of this standard separately with respect to each.
Perhaps the clearest ease of urgency is presented by intervenor The San Juan Star. The interest asserted is that of covering effectively an ongoing judicial proceeding of significant hard news interest. Time is of the essence to such coverage in an almost singular fashion. In addition to the harm thus suffered by any delay in review, the paper, as an intervenor, faces the added consideration that if no party appeals the eventual final judgment, it may be precluded from gaining review at any time. We thus find the claim or urgency asserted by The San Juan Star to be sufficient to meet the Cohen standard.
Intervenor Hernandez Agosto faces the same risk of potential preclusion from later review, and a similar urgency of purpose. A legislative investigation of a matter of pressing importance to a government may reasonably require expedition both to secure the availability or freshness of evidence and to have a legitimately desired impact, and the legislature has asserted such a need here. We thus conclude that its interests will be irreparably impaired by a further delay in review.
Plaintiffs’ challenge to the portion of the order prohibiting their counsel from disclosing the contents of depositions to them presents somewhat different considerations. The central right invoked, the right to effective participation in one’s own litigation, is in part derivative of the right to win vindication on the merits of the underlying claim, and is to that extent capable of vindication through vindication of those rights on an ultimate appeal. We would therefore normally be inclined to find a lack of the urgency necessary to make this part of the order appealable under Cohen. At the same time, however, plaintiffs have two distinct interests in immediate review. First, a litigant has a significant interest in ensuring that his counsel, possessed of some factual information to which the litigant himself is denied access, not take some action he does not wish to take during the course of litigation. See generally Faretta v. California, 422 U.S. 806, 95 S.Ct. 2525, 45 L.Ed.2d 562 (1975). In addition, we think— as discussed more fully below — that the pri- or restraint on significant communications imposed by this order does in fact in this instance implicate the First Amendment, see Bernard v. Gulf Oil Co., 619 F.2d 459, 466-73 (5th Cir. 1980) (en banc), aff’d on other grounds, 452 U.S. 89, 101 S.Ct. 2193, 68 L.Ed.2d 693 (1981), and that such a restraint constitutes immediate injury by its very nature. See id. at 470, citing Zwickler v. Koota, 389 U.S. 241, 252, 88 S.Ct. 391, 397, 19 L.Ed.2d 444 (1976). While we find the combination of these interests sufficient to compel immediate review in the unusual circumstances of this case, we stress that an asserted harm to a constitutional right is not in itself sufficient to make appealable an interlocutory order, especially one relating to discovery, see Grinnell, supra.
Because each of the three appellants-petitioners before us satisfies the tests of separability, finality, and urgency, we must address briefly the final test of appellate jurisdiction, that of importance. Although we have said that this element is not analytically a proper component of the collateral order test, see In re Continental Investment Corp., supra, 627 F.2d at 6, and although the Supreme Court has apparently never, rejected an appeal for failure to satisfy it, see id., the Court has continued to include it in its statement of the doctrine. See Firestone Tire & Rubber Co. v. Risjord, 449 U.S. 368, 374-75, 101 S.Ct. 669, 673-74, 66 L.Ed.2d 571 (1981). We think it sufficient to note that, as our analysis below makes clear, each appeal presents issues that are of clear importance and that certainly involve unsettled questions of controlling law. We thus hold that we have jurisdiction over all three appeals, and proceed to consider the merits of each in turn.
II.
We turn first to the appeal brought by The San Juan Star, which presents sev*114eral difficult questions. At the outset we determine whether communications prohibited by a protective order restricting the dissemination of information obtained through court-ordered depositions qualify for protection under the First Amendment.1 Finding that a constitutional right warranting some protection is involved, we then assess its significance and determine the standard by which judicial restrictions on the protected communications are to be measured. Finally, we determine whether the district court’s actions in this case complied with the properly applicable requirements. See generally In re Halkin, 598 F.2d 176 (D.C.Cir.1979) (Bazelon, J.); id. at 200 (Wilkey, J., dissenting); Note, Rule 26(c) Protective Orders and the First Amendment, 80 Colum.L.Rev. 1643 (1980); Case Comment, The First Amendment Right to Disseminate Discovery Materials, 92 Harv.L.Rev. 1550 (1979).
We think the answer to the first question to be that significant but limited First Amendment concerns are implicated by such protective orders. We begin with the undisputed fact that a court under Rule 26(c)(1) can deny access to information altogether by preventing discovery upon a showing of any “good cause”, specifically including but not limited to such relatively innocuous possibilities as “annoyance, embarrassment, oppression or undue burden or expense”, and that a denial of access does not implicate the First Amendment. This fact has led some judges to conclude that a power to deny access completely necessarily includes the lesser power to condition access to information on an undertaking not to disseminate it. See In re Halkin, supra, 598 F.2d at 200, 209 (Wilkey, J., dissenting). ¿But the “lesser included” rationale does not always solve the problem. For example, the police power of a municipality justifies zoning but does not permit preferential zoning for whites only. The power to deny access altogether does not necessarily allow a court to impose any condition upon access that it wishes — particularly when the condition at issue has an impact on separate constitutional concerns.
Our next step is to recognize the relevant but not determinative fact that litigants, their attorneys, and the press would have a strong interest in being able to convey and receive truthful reports of public criminal and civil judicial proceedings, see Richmond Newspapers, Inc. v. Virginia, 448 U.S. 555, 595-96, 100 S.Ct. 2814, 2838-39, 65 L.Ed.2d 973 (1980) (Brennan, J., concurring in the judgment), and that restraints on the transfer of newsworthy information of this nature would be subject to the highest degree of protection. See Nebraska Press Ass’n v. Stuart, 427 U.S. 539, 559, 96 S.Ct. 2791, 2802, 49 L.Ed.2d 683 (1976). When information is produced during civil discovery, however, the First Amendment interests and consequently the degree of severity of our scrutiriy of the restraints are different.
During discovery, information is produced through compulsion in order “to narrow and clarify the basic issues . .. and ascertain the facts”, Hickman v. Taylor, 329 U.S. 496, 501, 67 S.Ct. 385, 388, 91 L.Ed. 451 (1947), and thereby to make the subsequent *115trial “a fair contest with the basic issues and facts disclosed to the fullest practicable extent.” United States v. Procter & Gamble Co., 356 U.S. 677, 682, 78 S.Ct. 983, 986, 2 L.Ed.2d 1077 (1958). In pursuit of these aims, discovery rules are liberal, giving parties broad powers to gather information. As a result, a deponent may be obliged to reveal information that will not be admissible at trial. Unlike evidence at trial, it has not passed the strict threshold tests of relevance and admissibility, yet it has been compelled by dint of legal process. The information revealed may be irrelevant, prejudicial, or pose an undue invasion of an individual’s privacy.
Such undigested matter, forced from the mouth of an unwilling deponent, is hardly material encompassed within a broad public “right to know”. Its disclosure would not advance the informed civic and political discussion that the First Amendment is intended to protect. We do not see present in the case of civil discovery those interests that make publicity in a criminal trial an important “safeguard against any attempt to employ our courts as instruments of persecution,” In re Oliver, 333 U.S. 257, 279, 68 S.Ct. 499, 510, 92 L.Ed. 682 (1948). Nor can the discovery processes lay claim to the long tradition of openness enjoyed by criminal or civil trials. Compare Richmond Newspapers, Inc. v. Virginia, supra, 448 U.S. at 564-69, 100 S.Ct. at 2821-23 (opinion of /Burger, C.J.). We conclude, therefore, that although there is a First Amendment interest in information produced at trial that warrants full protection, a judicially-powered process compelling information that has not yet passed through the adversary-judicial filter for testing admissibility does not create communications that deserve full protection.
Having first recognized the full scale First Amendment interest in reports of public judicial proceedings, and also then having taken note of the series of factors distinguishing information developed during pre-trial discovery, we conclude that notwithstanding such factors the communications produced during discovery are not totally undeserving of protection under the First Amendment. Individuals have a well-recognized interest in self-expression, First National Bank of Boston v. Bellotti, 435 U.S. 765, 777 n.12, 98 S.Ct. 1407, 1416 n.12, 55 L.Ed.2d 707 (1978); Time, Inc. v. Hill, 385 U.S. 374, 388, 87 S.Ct. 534, 542, 17 L.Ed.2d 456 (1967), or, to turn the coin over, there is a First Amendment concern that the government not lightly engage in any restraints on communication, particularly when the order is issued prior to the expression taking place. Nebraska Press Ass’n v. Stuart, supra, 427 U.S. at 556, 96 S.Ct. at 2801. Moreover, some of the information produced during discovery presumably will be relevant at trial, newsworthy, and potentially subject to greater protection under the First Amendment.
The products of discovery, therefore, embody significant but somewhat limited First Amendment interests. The Supreme Court has recognized that the special characteristics of the particular type of speech at issue may render inapplicable the prohibition against prior-restraints. Virginia Pharmacy Board v. Virginia Consumer Council, 425 U.S. 748, 771 & 772 n.24, 96 S.Ct. 1817, 1830 & 1831 n.24, 48 L.Ed.2d 346 (1976) (commercial speech may not be subject to prohibition against prior restraints). Accordingly, we scrutinize the restraints imposed by protective orders under a less severe standard than that ordinarily applied to prior restraints.
This result allows the discovery process to continue to serve its function while also allowing judges efficiently and effectively to fulfill their duty to protect the litigants’ right to a fair trial and right to privacy. If the trial judge were required to allow virtually full publicity of utterances forced from the mouth of an unwilling deponent, even if irrelevant or inadmissible, he might well refuse to allow the discovery to proceed at all when the interests of a fair trial or personal privacy are seriously threatened. The alternative of having to supervise personally each question and response when interests of privacy or fair trial are at stake ‘would prevent discovery from accomplishing its purpose of expediting the trial proc*116ess. While allowing discovery to proceed with greater speed and greater scope, protective orders can also give judges the discretion to exercise their constitutional duty to protect litigants’ rights to a fair trial and privacy.
The standard we apply to review the validity of a restraint on communications produced during discovery is similar to the standard applied to prior restraints but less stringent due to the more limited nature of the First Amendment interests at stake. We look to the magnitude and imminence of the threatened harm, the effectiveness of the protective order in preventing the harm, the availability of less restrictive means of doing so, and the narrowness of the order if it is deemed necessary. Our adaption of this standard in a civil discovery context, however, allows a court to be slightly less severe when considering the degree and imminence of the harm that might ensue if communications to the press are not prohibited. Thus, we consider the magnitude of the harm threatened by disclosure and the likelihood of that harm occurring, but we think these considerations are best applied on a sliding scale: as the potential harm grows more grave, the imminence necessary is reduced. For example, a threatened material harm to a defendant’s deep-seated interest in an impartial jury in a civil trial need only be reasonably likely. A claimed injury to less important interests, by contrast, would need to be serious and unmistakable. In general, then, we find the appropriate measure of such limitations in a standard of “good cause” that incorporates a “heightened sensitivity” to the First Amendment concerns at stake, see Bruno & Stillman, Inc. v. Globe Newpaper Co., 633 F.2d 583, 596 (1st Cir. 1980); Brink v. DaLesio, 82 F.R.D. 664, 677-78 (D.Md.1979), a position again midway between those taken in the majority and dissenting opinions in Halkin. Compare In re Halkin, supra, 598 F.2d at 191 with id. at 209.
With respect to the remaining criteria to be applied — the effectiveness cf a protective order in preventing the harm, the availability of less restrictive means of doing so, and the narrowness of the order if it is deemed necessary — no countervailing interests are juxtaposed, and we would accordingly maintain the strict standard. Because no competing interest can serve to justify any restraint on expression more than that minimally necessary to its own protection, any restraint must embrace Iwhat the court, after sensitively addressing | the issue, determines to be the least restrictive means possible. Similarly, because no purpose sufficient to outweigh First Amendment concerns can be served by an ineffective restraint, the court must be reasonably certain that its order will have the effect sought.
We turn in light of these standards to consider the validity of the orders prohibiting dissemination of deposition contents to the press or public in this case. In its first order on the matter, issued June 11, 1980, the district court began by specifically identifying the issue as “one which involves a direct confrontation of two .. . fundamental rights”, those of free speech and fair trial. It went on to describe at length the ways in which the community had been “fully saturated” by the reports of the proceedings, and noted specifically the press conferences held by counsel, whom it properly described as “officers of the court” held to a “higher duty” of ensuring a fair trial. See, e. g., Hirschkop v. Snead, 594 F.2d 356, 366 (4th Cir. 1979). The court determined that the amount of publicity being generated would make it “difficult if not impossible for the defendants to obtain an impartial jury.” The court went on to note that its order applied only to deposition evidence, only to future depositions, and only until the time of trial; it also asserted explicitly that it saw “no other method or alternative” to its action. Finally, the court briefly addressed the First Amendment issues implicated by its order, concluding that they were “necessarily qualified or conditioned by the potential restrictions that are part of the [discovery] system” and that they did not outweigh the countervailing fair trial interests at stake.
*117In its decision reaffirming that order, issued December 30, 1980, the court again reviewed the “intense pretrial publicity”, asserting that it went “beyond that which is tolerable” and “cast very serious doubts” on its ability to select an impartial jury. The court reiterated its view that protective orders restraining dissemination of information obtained through discovery were permissible despite the significant First Amendment concerns presented if they were based on a “reasonable likelihood” that defendants would otherwise be unable to obtain a fair trial. The court recognized that such an order must be “narrowly drawn,” and again emphasized the limitations on the scope and duration of its own order. Taking specific note of In re Halkin, the court concluded that “other measures would ... fail to correct the threat” and that “without this prior restraint a fair trial will be denied.”
We think the court’s judgment sustainable. With respect to the two threshold elements of the test discussed above, it is clear that the court found at least a reasonable likelihood of a material harm to defendants’ right to a fair trial. We think this conclusion well-supported by both the massive amount of publicity indisputably attending this case and by the emotionally-charged nature of the trial itself. As to the other elements of the test, we think the issues closer but the order ultimately supportable.
First, the order is narrowly drawn. It does not restrain the press from publishing any information to which it gains access; it does not restrain any party or counsel from discussing with the press anything other than the contents of future depositions; and it does not seal any information beyond the time of trial. It would in general be more desirable for a court to review the contents of each deposition individually, restricting dissemination of only such information as it found specifically likely to endanger defendants’ right to a fair trial. See United States v. International Business Machine Corp., 82 F.R.D. 183, 195 (S.D.N.Y. 1979). However, given the probability that any additional publicity of deposition contents would be damaging to the defendants’ fair trial rights, and the need for the court to construct a manageable order which would not impose upon it an intolerable burden of involvement in the day-to-day problems of discovery, we think this arguable overbreadth alone insufficient to invalidate the court’s order.
With respect to the final two elements to be considered we conclude that, while the court did not explicitly articulate the reasons for its conclusions, the record amply supports them. The conclusion that the alternatives that should be considered, see Nebraska Press Ass’n v. Stuart, supra, 427 U.S. at 563-64, 96 S.Ct. at 2804-05, would be unavailing finds support in several factors peculiar to this case. As appellees have argued, Puerto Rico is singularly unsuited to a change of venue, and postponement of a trial of such urgent proportions could seriously jeopardize important interests in its resolution. On the district court’s conclusions that such intermediate measures as voir dire, instructions, or sequestration would be unavailing in the face of the jury pool’s “saturation”, we must defer to that court’s discretion and its closer familiarity with the nature of the publicity involved.2
Finally, the record also provides reasonable support for the belief that the court’s order would be effective, since deposition testimony — some of which might be both inflammatory and inadmissible, and thus especially prejudicial — had provided the basis for a large part of prior publicity, and since even a late bar on discussion could make a fair trial more likely at some late *118date.3 We recognize that criminal trials might be thought to present even more compelling fair trial rights than do civil trials, but think the importance and potential effects of this case sufficient to justify scrupulous care for those rights here. See Chicago Council of Lawyers v. Bauer, 522 F.2d 242, 257-58 (7th Cir. 1975); CBS, Inc. v. Young, 522 F.2d 234, 241 (6th Cir. 1975) (per curiam). Accordingly, the district court’s order prohibiting disclosure of deposition contents to the press or public is affirmed. In future cases, however, we would expect courts to assess these factors more explicitly and deliberately, and we will scrutinize those assessments accordingly-
III.
We turn next to the plaintiffs’ appeal of that part of the same order that prohibited disclosure of deposition material by plaintiffs’ counsel to their clients, plaintiffs themselves. We think this portion clearly invalid. First, the significance of plaintiffs’ right to such information can hardly be overstated. The very right sought to be vindicated by the court’s order — the fundamental right to a fair trial— might be thought as a general rule to depend critically on one’s ability to participate effectively in one’s own litigation. See Faretta v. California, supra; Geders v. United States, 425 U.S. 80, 88-89, 96 S.Ct. 1330, 1335-1336, 47 L.Ed.2d 592 (1976). In this particular case, moreover, undisputed affidavits have averred that plaintiffs — two of whom are tenured university professors — have consistently taken an active and significant role in the investigation, planning, and supervision of the litigation. In addition, because it is far from clear that it is within the court’s power to deny litigants access to information from counsel altogether, First Amendment concerns are implicated here even more strongly than with respect to the preceding appeal. See Bernard v. Gulf Oil Co., supra; Rodgers v. U.S. Steel Corp., 508 F.2d 152, 162-63 (3d Cir.), cert. denied, 423 U.S. 832, 96 S.Ct. 54, 46 L.Ed.2d 50 (1975). Finally, while plaintiffs concededly made numerous public statements about deposition testimony when such comment was unrestrained, there is no suggestion whatsoever either that they would fail to comply fully with a court order barring such disclosure in the future or that the court could not deal effectively with any intransigence. Indeed, no possibility of harm resulting from disclosure of this information to plaintiffs has ever been advanced. Accordingly, we face a significant infringement of fundamental constitutional rights supported by no direct or material interest, and we reverse that portion of the district court’s order prohibiting the disclosure of deposition contents to plaintiffs.
IV.
We turn finally to intervenor Hernandez Agosto’s appeal of the district court’s order quashing the Puerto Rico Senate’s subpoenas of investigatory documents from the Commonwealth Secretary of Justice. This question we also think clear, for we think the district court has simply misperceived the issue presented. The key fact, in our view, is that the documents sought had come into the Secretary’s possession through means entirely unrelated to the federal court’s discovery process — indeed, had come through the Department’s own investigations. The fact that copies of the documents had independently been obtained through discovery by a party to the federal litigation bears no more than an accidental relation to the Senate’s subpoena. As a result, none of the powers vested in the court by Fed.R.Civ.P. 26 — powers *119that were central to our disposition of The San Juan Star’s appeal — are of any relevance here whatsoever.
To address the Secretary’s two principal arguments directly, we first think it clear that the protective order cannot possibly be said to compel the Secretary not to disclose the documents in question. The order was issued at the Secretary’s request and for his protection, and for him, neither a party nor privy to the civil rights action, now to assert that it bars disclosure on his part stands its legitimate purpose on its head. See Rodgers v. U.S. Steel Corp., supra, 536 F.2d at 1006-07; Davis v. Romney, 55 F.R.D. 337, 341 (E.D.Pa.1972). The Secretary’s second contention, that the district court’s protective order invested him with a privilege which he was allowed to assert as valid against the Senate subpoenas, reflects the same fundamental misconception. The Secretary’s argument is akin to that of an immunized criminal defendant in a tax fraud case who seeks to prevent the IRS from using independently obtained information in a separate civil case. Any “privilege” recognized as an incident of federal discovery proceedings — the Secretary’s inappropriate term for an interest found worthy of protection under Rule 26(c) — serves solely to condition the access obtained through those proceedings and cannot shut off unrelated means of access independently available. This limitation is particularly strong where the party enjoined from seeking access through such alternate means is not in any way party or privy to the judicial proceedings. See Martindell v. ITT, 594 F.2d 291, 295 (2d Cir. 1979); GAF Corp. v. Eastman Kodak Co., 415 F.Supp. 129, 132 (S.D.N.Y.1976). Finally, the constraints on the court’s action are perhaps at their most acute where the nonparty whose nonjudicial access is precluded is a legislative body of a separate sovereign. We thus conclude that any rights or privileges the Secretary may have against the Senate must be asserted before that body, rendering unnecessary any consideration of the propriety of the court’s questioning the Senate President regarding legislative motives.
Because Fed.R.Civ.P. 26 is not relevant to this situation, any power to quash the subpoenas must derive from the court’s power to take any action necessary to ensure a fair trial, a power entirely independent of the fact that the subpoenas by chance relate to documents subject to a protective order in a discovery proceeding. See Rodgers v. U.S. Steel Corp., supra, 536 F.2d at 1006; International Products Co. v. Koons, 325 F.2d 403, 409 (2d Cir. 1963); 8 Wright & Miller, Federal Practice and Procedure § 2043 at 308. A court might in a proper case be empowered to enjoin a legislature from publishing information to ensure a fair trial, but the circumstances for such action have by no means been established here. Proeedurally, the Senate was not even a party to an action that amounted to an injunction against it. Substantively, the court would at a minimum have to meet the standards applied to prior restraints generally — first and foremost a showing of a serious and imminent threat to a fair trial. Here, however, the court made no finding that any publicity resulting from the Senate’s access would pose a grave and immediate threat to a fair trial, a higher standard than was necessary to sustain its order with respect to the deposition testimony. Nor will the Senate’s enforcement of its subpoena undermine the court’s restriction on dissemination of deposition contents, since — even were the Senate to make public all documents subpoenaed— those documents will not reveal the contents of depositions subject to the court’s order, and their disclosure can involve no risk of chilling future depositions. In addition, the district court made no attempt whatsoever to justify its order as either the most minimally intrusive measure possible or as a precise and narrowly drawn restraint, and they clearly are not — the court could for example have sought simply to avoid publication of the subpoenaed documents by the Senate. These limiting requirements are particularly important where, as here, considerations of both comity and separation of powers buttress the prior restraint concerns militating against interfering with dissemination generally. *120Accordingly, the court’s order quashing the Senate subpoenas is reversed.
In the appeal by The San Juan Star, we uphold the district court’s protective orders to the extent that they prevent parties to the litigation and their counsel from disclosing to the press and public information obtained through deposition proceedings. In the appeal by plaintiffs, we reverse the protective orders to the extent that they bar plaintiffs’ attorneys from sharing with plaintiffs the content of the depositions. Lastly, we reverse the district court’s order quashing the Senate subpoenas.
So ordered.
. It is not possible to dispose of this appeal on nonconstitutional grounds. The district court’s order provides sufficient evidence to support a finding that, in imposing a Rule 26(c) protective order, the court did not exceed the bounds of its statutory discretion. Compare Gulf Oil Co. v. Bernard, 452 U.S. 89, 101 S.Ct. 2193, 68 L.Ed.2d 693 (1981) (no need to consider whether restraint on communications imposed under Rule 23(d) violated First Amendment; application was abuse of discretion because record did not indicate that court had properly considered conflicting interests protected by rule).
In delineating the First Amendment interests involved, we think it useful to note what rights are not implicated by the court’s order here at issue. The court did not limit the right of the litigants or their counsel to express their opinions about the case or to disclose information obtained independently of the discovery process. See generally Case Comment, infra, at 1554. Nor is there a blanket prohibition of all speech concerning a trial, such as was struck down in Nebraska Press Ass’n v. Stuart, 427 U.S. 539, 96 S.Ct. 2791, 49 L.Ed.2d 683 (1976). Finally, this part of the order does not infringe on the plaintiffs’ acknowledged First Amendment right to litigate freely, see In re Primus, 436 U.S. 412, 98 S.Ct. 1893, 56 L.Ed.2d 417 (1978); NAACP v. Button, 371 U.S. 415, 83 S.Ct. 328, 9 L.Ed.2d 405 (1963).
. In assessing the district court’s treatment of this factor, we must give due deference to the lower courts’ familiarity with and expertise in managing the discovery process. Given the broad discretion accorded the district courts in discovery matters generally, we will not lightly step in and assess for ourselves, in hindsight, whether minor changes in the protective order might not have rendered it marginally less restrictive of speech while still protecting the other interests at stake. Of course, we will scrutinize the order to the extent necessary to assure ourselves that the district court has been as accommodating as possible of the speech interests involved under the circumstances of the case.
. A further justification for the order exists in the risk that the press conferences held after each deposition created substantial incentives for plaintiffs to depose well-known individuals who would not otherwise be deposed, or ask them questions that might not otherwise be asked. The court’s interest in the efficiency and integrity of the trial process, as well as its responsibility to protect deponents against harassment, would entitle it to take reasonable measures designed to minimize any incentive to lengthen and abuse discovery for non-judicial ends.