At issue in this case is whether a private corporation, Dow Chemical Company, threatened with possible government cancellation of certain herbicides it manufactures, may compel through administrative subpoenas University of Wisconsin re*1266searchers to disclose all of the notes, reports, working papers, and raw data relating to on-going, incomplete animal toxicity studies so that it may evaluate that information with a view toward possible use at the cancellation hearings. The administrative law judge, at the request of Dow and over the objection of the Office of General Counsel of the Environmental Protection Agency, issued such subpoenas, but the district court refused to enforce them. We affirm the judgment of the district court and hold1 that the present facts do not warrant forced disclosure of the university research information.
I. Background
This case arises out of four research studies at the University of Wisconsin involving the dietary ingestion by rhesus monkeys of a chemical compound, 2,3,7,8 — tetrachlorodibenzo-p-dioxin (TCDD), at levels of 500 parts per trillion (ppt), 50 ppt, 25 ppt, and 5 ppt, respectively. Relying in part on evidence from the 500 ppt study, the Environmental Protection Agency (EPA) ordered emergency suspension of certain uses of two herbicides Dow manufactures and scheduled cancellation hearings relevant thereto. The subpoenas now at issue relate to information about the 25 ppt and 5 ppt studies.1 The procedures and objectives of the studies, the parties to the litigation, the proceedings before the EPA, and the ruling of the administrative law judge (ALJ) are set forth in the opinion of the district court, as are the court’s reasons for refusing to enforce the subpoenas. See United States v. Allen, 494 F.Supp. 107 (W.D.Wis.1980).2 We rely on that background and will not recount those matters except as is relevant to our discussion of particular issues.
II. Judicial Enforcement of Administrative Subpoenas
A. Basic Principles
Under section 6(d) of the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA), 7 U.S.C. § 136d(d), a district court has jurisdiction to enforce a subpoena issued by an administrative law judge.3 Leaving enforcement to the courts indicates that Congress intended that judges should not merely rubber-stamp the subpoenas that come before them. See generally F.T.C. v. Owens-Corning Fiberglas Corp., 626 F.2d 966, 974 (D.C.Cir.1980); S.E.C. v. Arthur Young & Co., 584 F.2d 1018, 1032-1033 (D.C.Cir.1978); N. L. R. B. v. Northern Trust Co., 148 F.2d 24, 29 (7th Cir. 1945). But, while a district court’s enforcement function is “neither minor nor ministerial,” Oklahoma Press Publishing Co. v. Walling, 327 U.S. 186, 217 n.57, 66 S.Ct. 494, 509-10 n.57, 90 L.Ed. 614 (1946), it is well established that it is a narrowly limited one, see United States v. Morton Salt Co., 338 U.S. 632, 652-653, 70 S.Ct. 357, 368-69, 94 L.Ed. *1267401 (1950); Oklahoma Press, supra, 327 U.S. at 216-218, 66 S.Ct. at 509-10; F.T.C. v. Texaco, Inc., 555 F.2d 862, 871-872 (D.C.Cir.1977). Under what Dow refers to as the Morton Salt test, it is frequently said that in deciding whether to enforce, “it is sufficient if the inquiry is within the authority of the agency, the demand is not too indefinite and the information sought is reasonably relevant.” Morton Salt, supra, 338 U.S. at 652, 70 S.Ct. at 369; see also, e.g., F.T.C. v. Anderson, 631 F.2d 741, 745 (D.C.Cir.1979) (quoting Morton Salt). However, this formulation of the rule, standing by itself, is something of an overstatement, for it is also clearly recognized that disclosure may be restricted where it would impose an unreasonable or undue burden on the party from whom production is sought. See, e.g., F.T.C. v. Shaffner, 626 F.2d 32, 38 (7th Cir. 1980); Texaco, supra, 555 F.2d at 881-882. The burdensomeness test finds its genesis in the Fourth Amendment, which prescribes that disclosure shall not be unreasonable. See Oklahoma Press, supra, 327 U.S. at 208, 66 S.Ct. at 505. “What is unduly burdensome depends on the particular facts of each case and no hard and fast rule can be applied to resolve the question.” Shaffner, supra, 626 F.2d at 38. “Some burden on subpoenaed parties is to be expected and is necessary in furtherance of the agency’s legitimate inquiry and the public interest.” Texaco, supra, 555 F.2d at 882. “The burden of showing that the request is unreasonable is on the subpoenaed party ... [and] is not easily met. where ... the agency inquiry is pursuant to a lawful purpose and the requested documents are relevant to that purpose.” Id. at 882; see also United States v. Powell, 379 U.S. 48, 58, 85 S.Ct. 248, 255, 13 L.Ed.2d 112 (1964).
A finding by the district court that documents are reasonably relevant to a legitimate agency purpose cannot be overturned absent a showing that the factual determinations on which it is based are clearly erroneous, see Anderson, supra, 631 F.2d at 746; F.T.C. v. Lonning, 539 F.2d 202, 210 n.14 (D.C.Cir.1976), or that the ruling itself constitutes an abuse of discretion, see Northern Trust, supra, 148 F.2d at 29. Similarly, court assessments of whether disclosure would be burdensome and of what restrictions might be appropriate are decisions within the sound discretion of the trial court and should only be reversed for abuse of discretion, see Lonning, supra, 539 F.2d at 111, save where they are intimately tied to a misunderstanding of law, in which case the ordinary standard of error applies, see Texaco, supra, 555 F.2d at 881, 882.
B. The District Court’s Analysis
1. Scope of Review
Dow takes exception to the district court’s characterization of the application for enforcement as a de novo proceeding.4 It is true, as documented above, that the questions which are presented to a district court by such an application are quite limited. In resolving them, however, the court is not confined to examining the record made before the agency, as is more often the case in judicial review of administrative decisions. The AU’s decision to issue the subpoenas and his denial of the motion to quash are not binding on the court as to the questions properly presented.5
Dow also contends that the district court erroneously failed to apply the Morton Salt standard. Indeed, courts have held the Morton Salt test applicable to adjudicatory subpoenas as well as to those issued for *1268purposes of investigation. See Anderson, supra, 631 F.2d at 745; F.T.C. v. Browning, 435 F.2d 96, 102 (D.C.Cir.1970). Nonetheless, we do not think it was inappropriate for the district court to have differentiated these two situations. Even if both instances are governed by the Morton Salt criterion of reasonable relevance, “[tjhere is, of course, a difference in that the relevancy of an investigatory subpoena is measured against the ‘general purposes of [the agency’s] investigation,’ ... while relevancy of an adjudicative subpoena is measured against the charges specified in the complaint.” Anderson, supra, 631 F.2d at 746 (brackets in original; citations omitted). The bounds of relevance therefore tend to be broader in the investigatory context.
Thus, the district court properly took cognizance of the fact that the subpoenas were requested in aid of adjudicative discovery. We do not view the Morton Salt doctrine as excluding from consideration the question whether the burden of compliance was unreasonable. We reject Dow’s argument that the court adhered to an erroneous view of the law.
2. Balancing
The basis for the district court’s decision was that the subpoenas were unduly burdensome on the respondents. It stated that from the reference to protective orders in the applicable statutory section6 it logically follows that a subpoena request can be denied where the burden of compliance is too great. The court wrote:
“Whether a subpoena duces tecum is unduly burdensome ‘is, of course, a matter to be decided in the light of all the circumstances of the case, with an eye to the need for the material on the one hand, and the- burden imposed and the possibility of lightening it through a protective order on the other. See In re Zuckert (D.D.C.1961) 28 F.R.D. 29 ....’ 5A Moore’s Federal Practice ¶ 45.05[q] note 44.” 494 F.Supp. at 112.7
Addressing Dow’s need for the subpoenaed documents, the court stated:
“There is little question that the information is in fact relevant but that its probative value at this stage of the study is quite limited. Dow does not dispute [Mr. John] Van Miller’s statement that it is not possible to conclude even tentatively that there is a cumulative no-effect level8 for maternal or reproductive toxicity at this point of the studies. Absent the ability to infer a no-effect level of TCDD, information that the test animals are showing no effects is not particularly probative.” 494 F.Supp. at 113 (emphasis and footnote added).
Noting that Dr. James R. Allen was no longer scheduled to be a witness and that the EPA apparently did not intend to introduce the documents Dow sought to discover, the court decided that Dow’s need for the information was not great.
As to the burden of production element, the court stated:
“I take judicial notice that it would be a substantial burden on respondents to force them to produce the information requested from the 5 ppt and 25 ppt studies which are nowhere near completion and which have not been subjected to peer review. In the early stages of any research project there are likely to be *1269false leads or problems which will be resolved in the course of the study with no ultimate adverse effect on the validity of the study. To force production of all information demanded by the subpoenas is likely to jeopardize the study by exposing it to the criticism of those whose interests it may ultimately adversely affect, before there has been an opportunity for the researchers themselves to make sure the study is the result of their best efforts. This is not the kind of burden which can be lightened by a protective order. Putting this study in jeopardy would be a heavy burden not only on those involved in the research, but also on the public which has helped to fund it through tax money and which ultimately stands to gain from knowledge of the final results.” Id.
Summarizing its reasons for denying enforcement, the court concluded:
“that the probative value of the information sought by Dow is minimal; that it does not outweigh the substantial burden enforcement of the subpoenas would place on respondents; and that it could not be relieved by a protective order.” Id.
Dow challenges the district court’s analysis primarily on two grounds. It first argues that the court’s reliance on the statute’s reference to protective orders as justification for inquiring into burdensomeness and denying enforcement is misplaced because the statute contemplates the application of protective measures only once the decision to issue or enforce has already been made. Second, Dow claims that the district court’s balancing of need and probative value against burden of compliance far exceeds the limited review permissible in a subpoena enforcement action.
As to the first argument, it may well be unnecessary to rely upon an extrapolation of the statutory authorization of a protective order to justify the court’s inquiry into the extent of burden imposed by the subpoenas, and the resulting denial of enforcement.
In any event, the degree of burden is an aspect of reasonableness. Even if consideration of burdensomeness be not required by the Fourth Amendment, courts have not, and should not, lightly construe a statute as requiring a court to enforce compliance without deciding whether the burden is reasonable.9
It is well settled that disclosure of subpoenaed information may be restricted where compliance would force an unreasonable burden on the party from whom production is sought.10 This circuit has recognized that there is no set rule for determining whether a particular burden is excessive and that reference must be made to the facts of the individual case. See Shaffner, supra, 626 F.2d at 38. The assessment of what is unreasonable or unnecessary, by definition, depends to a large extent on the reasons or needs underlying the request. *1270Whereas significant need might justify imposition of a very substantial burden,11 the same cannot easily be said where the information requested, although generally relevant to contested issues in the sense that it is somehow related, is wholly unlikely to resolve the question or dispute. In short, resolution of the issue 'of reasonableness requires juxtaposition of need and probative value against burden of compliance.12 See Cohn, supra, 154 F.Supp. at 908 (In ascertaining reasonableness “[t]here is a delicate balance between the necessity of obtaining information required in the public interest in furtherance of a lawful inquiry, and the onerous burdens which the furnishing of this information may place on [the] respondents.”)
We proceed now to consider the factors relevant to the reasonableness determination.
C. The Probative Value of the Information
When the district court says that there is little question that the data is relevant, but then distinguishes relevance from probativeness, the court is using a very broad meaning of the term “relevance,” something akin to “not clearly irrelevant.” The Federal Rules of Evidence would suggest a closer linkage between the concepts of relevance and probativeness since F.R.E. 401 defines relevant evidence in terms of its tendency to make the existence of a fact more or less probable, and thus comes pretty close to the idea of whether evidence has probative value. Similarly, as noted earlier, the fact the subpoenas were issued in aid of adjudication (as opposed to general investigation) would to some extent favor a more restrictive reading of relevance. However, regardless of the district court’s distinction between relevance and probativeness, it is clear there was no prejudice to the rights of Dow. The only real consequence is that the potential usefulness of the data was considered under the rubric of probative value instead of relevance. In this we find no reversible error.
In deciding that the information concerning the 25 ppt and 5 ppt studies was, at that point in time, not particularly probative,13 the district court relied primarily upon the second affidavit of Mr. Van Miller, *1271(see Joint Appendix, pp. 135-139,113, 4, 5, 6, 9,10,11,12,13,14), which was among the six affidavits submitted to the court. That affidavit is consistent with Mr. Van Miller’s first affidavit (see Joint Appendix, pp. 97-114, 1126, 42) and with the affidavit of Dr. Allen (see Joint Appendix, pp. 67-95,1129, 30, 34) in stating that the 25 ppt study would have to be conducted for a minimum of 60 months and the 5 ppt study substantially longer before any conclusion could be reached as to the existence of a no-effect level for TCDD. In neither the district court nor in this court has Dow challenged or contradicted the scientific analysis supporting that conclusion. Instead, Dow raises several matters which we think are not in point.
First, Dow repeatedly insists, that according to the protocols for the studies, breeding trials would have taken place and thus data should have been available concerning conception, pregnancy, offspring, and the like. Assuming that is so, the mere existence of such information does not mean it would be useful or probative in the sense that conclusions could be drawn concerning cumulative reproductive and maternal toxicity. If neither the 25 ppt or 5 ppt study had reached the point where effects of TCDD intoxication could be expected to be observed, the absence of such effects is hardly significant.14
Next, Dow claims that contrary to the statement in the affidavits that five years or more of TCDD ingestion would be necessary before conclusions could be drawn concerning a cumulative no-effect level, the grant application filed with the Government in connection with the studies indi*1272cates that the monkeys will be exposed to TCDD for only two years. The implication seems to be that the affidavit statement is somehow contradicted. We cannot agree. The grant application was filed several months prior to completion of the 500 ppt study and even further before the 50 ppt study confirmed the critical intake level15 established by the 500 ppt study. Thus, it was not possible at that point in time to accurately estimate how long each of the studies would have to run before a cumulative no-effect level might be ascertained. The two year figure stands as nothing more than a preliminary appraisal of duration which has since been proved impracticable. That the studies were originally scheduled to run for two years in no way contradicts the fact that, according to data gathered in the 500 ppt and 50 ppt tests, the 25 ppt and 5 ppt studies will have to continue five years or longer before the desired information becomes available.
Dow also asserts that the “FIFRA Scientific Advisory Panel [which was consulted by the EPA pursuant to statutory command] recognized that the no-effect level for monkeys is likely to be found in the 25-50 ppt range and specifically recommended that at least the 25 ppt study be obtained because of its pertinence to consideration of the no-effect level.” But this argument, too, misses the point because regardless of whether the no-effect level is likely to be found in that range the actual determination cannot be made until the studies have continued long enough for the total dosage of TCDD received to be at least equivalent to the total dosage at which toxic effects were first observed when TCDD was administered at the higher dosages of 50 ppt and 500 ppt. Nothing in the advisory panel’s findings disputes this fact.
Further, Dow contends that the district court should have deferred to the AU’s determination that “relevant and useful data regarding conception, pregnancy and possible births should be available.”16 Again, we are unable to agree. At the time that finding was made, respondents Dr. Allen and Mr. Van Miller, the parties to whom the subpoenas were issued, were only nominally represented, without their knowledge, by someone from the EPA. The AU was without the benefit of their affidavits carefully explaining the earliest points at which valid conclusions could be drawn from the 25 ppt and 5 ppt studies. In our opinion, that information is critical to a proper assessment of the probative value of the documents sought. In any event, the AU himself substantially — if not completely — undercut his assertion that useful data could be obtained when he went on to state that while he found the information relevant,
“This is not to say that there has been a determination at this time as to whether or not any such discovered data will be probative or substantive in nature.”17
We conclude that the district court was entitled to give very substantial weight to the uncontradicted analysis of respondents as to how long it would take for information as to the possible existence of a no-effect level to become available. The court’s finding of little probative value is not clearly erroneous.
D. Dow’s Need for the Information
In concluding that Dow’s need for the information was not great, the district court took note of three specific factors: first, Dr. Allen no longer was scheduled to appear at the cancellation hearing, and even when he was to be a witness he did not intend to discuss the 25 ppt and 5 ppt studies; second, the EPA apparently did not intend to introduce as exhibits any of the documents or information Dow sought to discover since it failed to exchange them pursuant to the discovery provisions of 40 *1273C.F.R. § 164.50;18 and third, the data from the 25 ppt and 5 ppt studies was not needed for the purpose of evaluating the 500 ppt and 50 ppt studies since Dr. Allen would not testify and in any event the protocol for all four studies was the same. Each of these factors was, we think, relevant to an assessment of Dow’s need for the information. The first two related to whether Dow would be confronted at the hearing with evidence from the 25 ppt and 5 ppt studies and would therefore need access to the data so that it could test the validity of the evidence. The district court properly concluded that the unlikelihood of such confrontation tended to establish lack of need for the documents on the part of Dow. The third factor concerned whether the 25 ppt and 5 ppt studies were needed to ensure proper evaluation of the earlier studies. The argument apparently had been made that if the protocols for the 25 ppt and 5 ppt studies were different from those followed earlier, the changes might indicate defects or flaws in the earlier studies. The district court correctly observed, however, that the procedures in each of the studies were identical, and therefore concluded that that consideration, too, weighed against a finding of need for disclosure. The court’s finding that there was no convincing showing of need on the part of Dow is not clearly erroneous.19
E. The Burden of Compliance
1. The Danger of Premature Disclosure
Dow argues that it was improper for the district court to take “judicial notice” of the fact that a substantial burden would be placed on Dr. Allen and Mr. Van Miller by forced compliance with the subpoenas. We are convinced that the court was not using that term in the classic sense, compare F.R.E. 201, but merely describing certain findings of fact it had made. There is a significant correlation between the court’s assessment of the burden on respondents and uncontroverted sworn statements submitted to the district court.
Relevant portions of the affidavits of Dr. Allen and Mr. Van Miller stated, without contradiction by Dow, that public access to the research data would make the studies an unacceptable basis for scientific papers or other research; that peer review and publication of the studies was crucial to the researchers’ credibility and careers and would be precluded by whole or partial public disclosure of the information; that loss of the opportunity to publish would severely decrease the researchers’ professional opportunities in the future; and that even inadvertent disclosure of the information would risk total destruction of months or years of research. These statements provide clear support for the district court’s findings that “the 5 ppt and 25 ppt studies ... [were] nowhere near completion and ... [had] not been subjected to peer review,” and that “force[d] production of all of the information demanded by the subpoenas ... [was] likely to jeopardize the study . . . [and] would be a heavy burden ... on those involved in the research.” 494 F.Supp. at 113 (brackets added).
*1274Dow argues that the court’s observation that, “[i]n the early stages of any research project there are likely to be false leads or problems which will be resolved in the course of the study with no ultimate adverse effect on the validity of the study” is an assertion unsupported by anything in the record. We do not think the point is well taken, if for no other reason than that we fail to see that this consideration significantly influenced the decision of the court. The expression, we think, was nothing more than casual recognition of a fact which is well known in common experience: namely, that there are hazards in relying on incomplete research. We find no error in this aspect of the court’s decision.
There is also considerable dispute as to extent to which compliance with the subpoenas would have disrupted or interfered with continuation of the studies. Dow, of course, minimizes the burden by arguing that production of the documents could have taken place in stages and that respondents’ past adherence to federal reporting requirements in connection with its funding grants shows that information can be gathered without undue difficulty. Respondents, on the other hand, maximize the potential disruption, arguing that the exceedingly broad scope of the subpoenas20 makes this a far different situation from the minimal disclosure mandated for continuation of federal funding. We find it unnecessary to resolve these contentions. While the district court might have based its decision on the fact that “compliance threaten[ed] to unduly disrupt or seriously hinder normal operations,” see Texaco, supra, 555 F.2d at 882, it apparently chose not to do so. Rather, the court found that the risk of even inadvertent premature disclosure so far outweighed the probative value of and need for the information as to itself constitute an unreasonable burden on respondents. We cannot say that this approach, on the instant facts, constituted an abuse of discretion.
2. Academic Freedom
A point not discussed by the district court, but presented by the State of Wisconsin as amicus on appeal, is that the instant dispute touches directly upon interests of academic freedom. Essentially, the State argues that scholarly research is an activity which lies at the heart of higher education, that it comes within the First Amendment’s protection of academic freedom, and therefore judicially authorized intrusion into that sphere of university life should be permitted only for compelling reasons, which do not exist here. In response, Dow asserts simply that “[t]he First Amendment interests at stake in this case are no greater than those involved in the ordinary case of enforcement of a subpoena duces tecum.”21 We think this issue sufficiently important to merit our discussion.
“Academic freedom, though not a specifically enumerated constitutional right, long has been viewed as a special concern of the First Amendment.” University of California Regents v. Bakke, 438 U.S. 265, 312, 98 S.Ct. 2733, 2759, 57 L.Ed.2d 750 (1978). Nearly a quarter-century ago, Chief Justice Earl Warren wrote:
“The essentiality of freedom in the community of American universities is almost self-evident. No one should underestimate the vital role in a democracy that is played by those who guide and train our youth.... Scholarship cannot flourish in an atmosphere of suspicion and distrust. Teachers and students must always remain free to inquire, to *1275study and to evaluate, to gain new maturity and understanding; otherwise our civilization will stagnate and die.”
Sweezy v. New Hampshire, 354 U.S. 234, 250, 77 S.Ct. 1203, 1211, 1 L.Ed.2d 1311 (1957), quoted with approval in Keyishian v. Board of Regents, 385 U.S. 589, 603, 87 S.Ct. 675, 683-84, 17 L.Ed.2d 629 (1967), and Whitehill v. Elkins, 389 U.S. 54, 60, 88 S.Ct. 184, 187, 19 L.Ed. 228 (1967). To be sure, “Our Nation is deeply committed to safeguarding academic freedom, which is of transcendent value to all of us not merely to the teachers concerned.” Keyishian, supra, 385 U.S. at 603, 87 S.Ct. at 683, quoted with approval in Bakke, supra, 438 U.S. at 312, 98 S.Ct. at 2760.
The precise contours of the concept of academic freedom are difficult to define. See Cooper v. Ross, 472 F.Supp. 802, 813 (E.D.Ark.1979). One First Amendment scholar has written “[t]he heart of the system consists in the right of the individual faculty member to teach, carry on research, and publish without interference from the government, the community, the university administration, or his fellow faculty members.” T. Emerson, The System of Freedom of Expression 594 (1970). We think it clear that whatever constitutional protection is afforded by the First Amendment extends as readily to the scholar in the laboratory as to the teacher in the classroom. See generally, id. at 619.
Of course academic freedom, like other constitutional rights, is not absolute, and must on occasion be balanced against important competing interests. See, e.g., Barenblatt v. United States, 360 U.S. 109, 112, 129, 130, 79 S.Ct. 1081, 1085, 1094, 1095, 3 L.Ed.2d 1115 (1959) (the Court recognized academic freedom element, but upheld contempt conviction of teaching fellow who refused to answer questions about Communist Party membership, since investigation was not directed at controlling what was taught at university but at overthrow of government). Case law considering the standard to be applied where the issue is academic freedom of the university to be free of governmental interference, as opposed to academic freedom of the individual teacher to be free of restraints from the university administration, is surprisingly sparse. But what precedent there is at the Supreme Court level suggests that to prevail over academic freedom the interests of government must be strong and the extent of intrusion carefully limited. In Sweezy, supra, the Supreme Court recognized that questions posed by a state attorney general to a university professor about the contents of a lecture invaded the professor’s liberty of academic freedom. The four-justice plurality stated that it could “not now conceive of any circumstance wherein a state interest would justify infringement of rights in .. . [this field],” but concluded it was unnecessary to reach that fundamental question of state power. 354 U.S. at 251, 77 S.Ct. at 1212. It based its reversal of the professor’s contempt conviction on the ground that there was no clear showing that the legislature had authorized the type of inquiry made by the attorney general. Justice Frankfurter, joined by Justice Harlan, concurred in the result on First Amendment grounds. Applying a balancing test, which gave predominant weight to the grave harm resulting from governmental intrusion into the intellectual life of a university, Justice Frankfurter wrote:
“[Academic] inquiries ... must be left as unfettered as possible. Political power must abstain from intrusion into this activity of freedom, pursued in the interest of wise government and the people’s well-being, except for reasons that are exigent and obviously compelling.” 354 U.S. at 262, 77 S.Ct. at 1217.22
In Keyishian, supra, the Court struck down key provisions of state loyalty requirements for teachers as vague and overbroad. After discussing the importance of academic freedom, the opinion of the Court, by Justice Brennan, stated:
*1276“We emphasize once again that ‘[precision of regulation must be the touchstone in an area so closely touching our most precious freedoms, .... Because First Amendment freedoms need breathing space to survive, government may regulate in the area only with narrow specificity’ .... The danger of ... chilling . .. the exercise of vital First Amendment rights must be guarded against by sensitive tools . . . . ” 385 U.S. at 604-605, 87 S.Ct. at 684-85 (citation omitted).
In the present case, the 25 ppt and 5 ppt administrative subpoenas by their terms would compel the researchers to turn over to Dow virtually every scrap of paper and every mechanical or electronic recording made during the extended period that those studies have been in progress at the university. The ALJ’s decision would have further obliged the researchers to continually update Dow on “additional useful data” which became available during the course of the proceedings.23 These requirements threaten substantial intrusion into the enterprise of university research, and there are several reasons to think they are capable of chilling the exercise of academic freedom. To begin with, the burden of compliance certainly would not be insubstantial. More important, enforcement of the subpoenas would leave the researchers with the knowledge throughout continuation of their studies that the fruits of their labors had been appropriated by and were being scrutinized by a not-unbiased third party whose interests were arguably antithetical to theirs. It is not difficult to imagine that that realization might well be both unnerving and discouraging. Indeed, it is probably fair to say that the character and extent of intervention would be such that, regardless of its purpose, it would “inevitably tend[] to check the ardor and fearlessness of scholars, qualities at once so fragile and so indispensable for fruitful academic labor.” Sweezy, supra, 354 U.S. at 262, 77 S.Ct. at 1217-18 (Frankfurter, J., concurring in result).24 In addition, the researchers could reasonably fear that additional demands for disclosure would be made in the future. If a private corporation can subpoena the entire work product of months of study, what is to say further down the line the company will not seek other subpoenas to determine how the research is coming along?25 To these factors must be added the knowledge of the researchers that even inadvertent disclosure of the subpoenaed data could jeopardize both the studies and their careers. Clearly, enforcement of the subpoenas carries the potential for chilling the exercise of First Amendment rights.
We do not suggest that facts could not arise sufficient to overcome respondents’ academic freedom interests in the 25 ppt and 5 ppt studies. Nor do we say that a waiver of the protection afforded by the First Amendment is impossible. If, for example, Dr. Allen, Mr. Van Miller, or other researchers were likely to testify about the 25 ppt and 5 ppt studies at the cancellation hearing, there might well be justification for granting at least partial or conditional enforcement of the subpoenas. Of course, we need not decide that question now. For present purposes, our point is simply that respondents’ interest in academic freedom may properly figure into the legal calculation of whether forced disclosure would be *1277reasonable. See Richards of Rockford, Inc. v. Pacific Gas and Electric Company, 71 F.R.D. 388 (N.D.Cal.1976).26 Based on the facts before us which, among other things, show that potentially probative evidence will not be available from the 25 ppt or 5 ppt studies for months or years to come, that Dow will not be confronted by information from the studies at the cancellation hearing, and that present 25 ppt and 5 ppt study data cannot be used to test the validity of the 500 ppt study, we conclude there is little to justify an intrusion into university life which would risk substantially chilling the exercise of academic freedom. The district court could correctly have taken this factor into consideration in deciding that enforcement of the subpoenas would be unreasonable. We regard it as additional reason for affirming the judgment appealed from.
3. Third Party Status
Section 6(d) of FIFRA permits subpoenas for documents to be issued against non-parties as well as against those directly involved in the proceeding out of which the subpoenas arose. Notwithstanding that fact, a court may properly give account to the third-party status of those from whom production is sought in determining whether compliance would constitute an undue burden. See Richards of Rockford, supra, 71 F.R.D. at 390 (“Neither ... are parties to the underlying action .... [T]his considerationf] weigh[s] the balance toward non-disclosure”); F.T.C. v. Bowman, 149 F.Supp. 624, 629-630 (N.D.Ill.), aff’d, 248 F.2d 456 (7th Cir. 1957) (“the imposition of a heavy burden upon a witness not a party to that proceeding should be avoided”); but see Dresser Industries, supra, at p. 71,491 (“[0]ne who opposes an agency’s subpoena must bear a heavy burden. That burden is essentially the same even if the subpoena is directed to a third party.”) While the factor is far from a decisive one, it is entitled to some measure of consideration. The fact that Dr. Allen and Mr. Van Miller are not parties to the cancellation hearing adds its weight, however small, to the side of the scale favoring non-disclosure.
F. Protective Order
In its brief Dow contends that it has long been prepared to agree to a reasonable protective order for the data to safeguard its confidentiality pending respondents’ submission of any report on the studies for publication. However the extent to which Dow has manifested this readiness is open to question since Dow fails to dispute the statements in Dr. Allen’s and Mr. Van Miller’s affidavits that they were never contacted by Dow (or by the EPA) concerning the possibility of such an arrangement. Also, while the AU indicated a willingness to “entertain any motion for a protective order upon good cause shown,” 27 apparently no motion was ever made.
The district court recognized that in a subpoena enforcement action it is appropriate to consider the extent to which the burden of production may be lightened through a protective order, 494 F.Supp. at 112, but its discussion of this subject is exceedingly limited. After reviewing the nature of the burden imposed on respondents, the court simply concluded: “This is not the kind of burden which can be lightened by a protective order.” 494 F.Supp. at 113.
Although it would have been helpful if the district court had indicated its reasons for reaching that conclusion, we are not persuaded that the district court abused its discretion.
*1278A protective order presumably would have reduced the likelihood of public disclosure of the research data, but it could not have eliminated that risk.' Given the fact there was already some showing of at least inadvertent disclosure of raw data from the 500 ppt study entrusted to Dow28 and the fact that such disclosure of the 25 ppt and 5 ppt data could severely jeopardize the careers and professional reputations of the researchers, the district court could well have concluded that a protective order might not have been fully effective. A protective order, moreover, would not have eliminated the chilling effect which invariably accompanies governmentally authorized intrusions into the intellectual life of the university.
G. Miscellanea
Aside from, the issues heretofore discussed respondents argue that the subpoenas should not be enforced for several reasons, including that: respondents do not possess or exercise control over the documents; no witness fees were tendered; modified subpoenas were never served subsequent to the partial grant of the motion to quash; the subpoenas are overly broad and vague; production was required at three different locations at the same time; and the materials are protected from disclosure by the state of Wisconsin trade secrets privilege. In view of our disposition of the case, we find it unnecessary to address these questions.
H. Conclusion
The district court held that production of raw research data, far in advance of the time that relevant and probative results could be expected, should not be compelled over the non-frivolous objections of the researchers. The findings upon which the court based this decision were not clearly erroneous, nor was the holding itself an abuse of discretion. Accordingly, the judgment appealed from is AFFIRMED.
. The ALJ excluded the 500 ppt and 50 ppt studies from the subpoenas based on his understanding of representations by EPA counsel that respondents would voluntarily produce the documents and materials related thereto. See Joint Appendix, p. 26. Subsequently, the information relating to the 500 ppt study was turned over, but the materials relating to the 50 ppt study were not disclosed and became the subject of a second enforcement action, United States v. Barsotti, Civ. No. 80-C-677 (W.D. Wis. filed Dec. 31, 1980). That action was dismissed on April 27, 1981, after the EPA moved to withdraw the subpoenas relating to the 50 ppt study.
. The United States, petitioner in the district court, originally filed an appeal from the order denying enforcement, but later withdrew the same. It has not submitted a brief in the present appeal which was perfected by Dow, intervening petitioner in the district court. The respondents-appellees, Dr. James R. Allen and Mr. John Van Miller, were at times participants in the University of Wisconsin studies and were the parties to whom the subpoenas were issued. James P. Wachtendonk, et al, intervening respondents-appellees, are Vietnam veterans and families of servicemen who were permitted by the district court to enter this litigation because of the relevance of the studies to suits arising out of servicemen exposure to certain tree defoliants used in Southeast Asia between 1962 and 1971. The State of Wisconsin appears before us as amicus curiae.
. We are unaware of any judicial decisions concerning enforcement of administrative subpoenas issued under FIFRA. Consequently, our discussion is guided by cases dealing with the enforcement of administrative subpoenas issued in other contexts. Case law and the parties before us endorse this approach.
. The case cited by the district court on this point, E. E. O. C. v. United States Fidelity & Guaranty Co., 414 F.Supp. 227, 232 (D.Md.1976), held that “[a] subpoena enforcement proceeding is a de novo proceeding before [the District] Court” and therefore respondent was not barred from raising issues which it failed to raise before the Commission. That decision was affirmed on appeal without published opinion, 538 F.2d 324 (4th Cir. 1976).
. The ALJ’s decision was made without the benefit of the extensive affidavits submitted to the district court. See the discussion in the text, infra. It appears, also, that although the EPA made the motion to quash before the ALJ in the names of Allen and Van Miller, they may not have had a full opportunity to participate.
. 7 U.S.C. § 136d(d) provides in part:
“Upon a showing of relevance and reasonable scope of evidence sought by any party to a public hearing, the [ALJ] shall issue a subpoena to compel testimony or production of documents from any person. The [ALJ] shall be guided by the principles of the Rules of Civil Procedure in making any order for the protection of the witness or the contents of documents produced.... On contest, the subpoena may be enforced by an appropriate United States district court in accordance with the principles stated herein.”
. In a footnote, the district court observed:
“This passage from Moore’s and the case cited therein refer to Rule 45, F.R.C.P. which is not specifically applicable to the issuance of the subpoenas at issue in this case. However, they are ‘in pari materia’ and are persuasive regarding the meaning of § 136d(d).”
. A cumulative no-effect level is a dosage at which a toxin produces no adverse effects on a species regardless of the length of time for which the dosage is continued.
. This is so at least where there are facts sufficient to raise a true question of burdensomeness. See note 12, infra.
. Ordinarily such restrictions take the form of limitations on the scope of discovery, see Adams v. F.T.C., 296 F.2d 861, 870 (8th Cir. 1961), special provisions concerning manner and time of production, see Shaffner, supra, 626 F.2d at 38; S. E. C. v. Savage, 513 F.2d 188, 189 (7th Cir. 1975), or protective orders forbidding outside disclosure of the information, see Lonning, supra, 539 F.2d at 211, and cases cited therein. But where such types of protection are impracticable or clearly inadequate in terms of the burden or risk that is threatened, we see no reason why disclosure may not fully be denied. Cf. United States v. Powell, 379 U.S. 48, 58, 85 S.Ct. 248, 255, 13 L.Ed.2d 112 (1964) (subpoena issued not in good faith, but to harass or pressure subject of investigation, need not be enforced). Otherwise enforcement of the subpoena would run afoul of the Fourth Amendment’s reasonableness requirement, discussed by the Supreme Court in Oklahoma Press, supra, 327 U.S. at 208, 66 S.Ct. at 505, and in other cases. See F.C.C. v. Cohn, 154 F.Supp. 899, 908 (S.D.N.Y.1957) (“courts will plainly refuse to enforce an administrative subpoena which is not within the bounds of reasonableness”); see also Chapman v. Maren Elwood College, 225 F.2d 230, 231, 232 (9th Cir. 1955) (enforcement denied, notwithstanding fact that “matters ... sought were subject to jurisdiction of the [Veteran’s] Administration,” where uncontested “affidavit ... tended to prove that demand was entirely unreasonable”).
. See, e.g., F.T.C. v. Dresser Industries, [1977-1] Trade Cases ¶ 61,400, at pp. 71,492-71,493, (D.D.C.1977):
“It may very well be Dresser’s burden is greater than that of other subpoenaed companies, but ... it is Dresser’s dominance in the industry which makes the subpoena served upon it critical to Kaiser’s defense .... [T]he court must find that the burden imposed is not an unreasonable one....”
. It is important to distinguish the question of reasonable burden from the question of reasonable relevance. The Morton Salt “standard of reasonable relevance does not ... require a showing of specific need for the information.” United States v. First City Nat. Bank of El Paso, 598 F.2d 594, 599 (Temp.Emer.Ct.App.1979); see also Dresser Industries, Inc., supra, at p. 71,492 (“strong showing of relevance or need ... is not the correct standard”). Thus we agree with Dow that in the usual case a court need not inquire into the need or reasons of the party seeking enforcement of the subpoena. Where, however, as here, factually supported affidavits clearly raise a colorable claim of unreasonable burden, it is incumbent upon the court to make such an inquiry. Cf. S.E.C. v. Brigadoon Scotch Distilling Co., 480 F.2d 1047, 1056 (2d Cir. 1973) (“in appropriate circumstances a court may inquire into the reasons for an investigation and into its effects”).
. The district court’s finding that the data was not yet probative in itself provides reason for affirming the judgment appealed from. Section 164.51, 40 C.F.R. ch. 1, the regulatory provision which governs pre-hearing discovery, see Allen, supra, 494 F.Supp. at 112, permits discovery “only upon determination by the Administrative Law Judge ... that such information has significant probative value.” The ALJ found it impossible to say whether the information would be probative in nature, see text accompanying n.17, infra, and the district court went even further by finding that probative value was “quite limited,” 494 F.Supp. at 113. The court discussed the applicability of § 164.51 but rested its decision on other grounds, specifically avoiding the question of whether the regulation was consistent with the statute. We see no conflict between the Act and § 164.51 of the regulations. Section 6(d) of FIFRA, 7 U.S.C. § 136d(d), see n.6 supra, does not expressly mandate pre-hearing discovery, but instead sets down certain minimum requirements for exercise of the ALJ’s compulsory process authority. Such expressions by themselves create no entitlement to pre-hearing discovery. Cf. NLRB v. Interboro Contractors *1271Inc., 432 F.2d 854, 858 (2d Cir. 1970) (similar language in APA held not to create right to pre-trial discovery in the administrative process); accord Silverman v. Commodity Futures Trading Commission, 549 F.2d 28, 33 (7th Cir. 1977); see also Miner v. Atlas, 363 U.S. 641, 80 S.Ct. 1300, 4 L.Ed.2d 1462 (1960) (district court sitting in admiralty could not grant discovery depositions because admiralty rules previously adopted by Supreme Court did not expressly provide for such discovery). Formulation of appropriate pre-hearing discovery procedures is committed to the agency’s discretion under its general rulemaking authority. FIFRA § 25(a)(1), 7 U.S.C. § 136w(a)(1). Cf. Interboro Contractors, supra, 432 F.2d at 858 (rule-making authority under APA). The EPA exercised its rulemaking powers by adopting § 164.51. In discussing the proposed rules the Administrator noted:
"Inasmuch as a cancelled pesticide remains on the market until completion of the hearing, it has been advisable in the public interest and the orderly administration of justice to limit discovery to witness lists and exhibits, unless discovery ... is determined by the Administrative Law Judge not to unreasonably delay the proceeding, that the information is not otherwise obtainable and that the information has significant probative value." 38 Fed.Reg. 19371 (1973).
As such, § 164.51 reflects a sensitive appreciation of the circumstances surrounding a cancellation hearing. The requirements of § 164.51 in no way conflict with the “relevance” and “reasonable scope” requirements of FIFRA § 6(d). Thus, the district court could properly have denied the motion to enforce on the ground that the ALJ was unable to state that production would yield significantly probative information.
We reject Dow’s contention that the “relevant and material” standard of 40 C.F.R. § 164.70(a), instead of the “significant probative value” test of 40 C.F.R. § 164.51, governs its right to pre-hearing discovery. Section 164.70(a) provides:
“Subpoenas may be issued by the [ALJ] sua sponte or upon a showing by an applicant that evidence sought for hearing is relevant and material to the issues involved in the hearing or that the sought discovery pursuant to § 164.51 meets the standards set forth therein.”
Clearly, § 164.70(a) contemplates that in the context of pre-hearing discovery the more specific provisions of § 164.51 shall govern over the more lenient general provisions which it sets forth. To read § 164.70(a) otherwise would be to render superfluous the terms of § 164.51 and § 164.70(a)’s reference to them.
. There is no suggestion anywhere in the papers before us that monkeys receiving 25 ppt and 5 ppt TCDD have manifested signs of TCDD intoxication. The only relevant statements — and the ones upon which Dow hopes to capitalize, see Appellant’s Brief at pp. 35-36 —indicate that no adverse effects have been observed. For example, a progress report filed by Dr. Allen states that monkeys receiving 25 ppt for six months or 5 ppt for three months were free of signs of TCDD intoxication. Joint Appendix, p. 177. This is of no significance since even in the 25 ppt study no such effects are expected until the studies have continued for at least 60 months.
. The “critical intake level” is the point at which the total amount of TCDD ingested by the study animals first produces observable adverse effects.
. Joint Appendix, p. 27 (emphasis added).
. Id.
. Section 164.50, C.F.R. ch. 1, provides for “primary discovery,” that is, the exchange by the parties of witness lists, documents, and exhibits expected to be used at the hearing.
. Certain developments have taken place subsequent to the district court’s ruling, but we do not think they mandate a contrary result. First, Dow’s brief states that, “respondent Allen and a co-researcher have now been subpoenaed to testify at the hearing.” Appellant’s Brief, p. 38 n.17. Assuming that is so, there is still no indication that Dr. Allen will discuss the 25 ppt and 5 ppt studies, nor that the co-researcher [who Appellant’s Reply Brief indicates is Dr. Deborah A. Barsotti] would testify in that area.
Second, certain subpoenas relating to the 50 ppt study have been withdrawn and references to that study have been excluded from the EPA hearing record. See letter from respondents’ counsel to the Clerk, Aug. 12, 1981, p. 1. Dow’s need for the 25 ppt and 5 ppt studies’ data is therefore, if anything, even less than before. See letter from Dow’s counsel to the Clerk, Mar. 19, 1981, p. 3 (“[I]f [AU] ... eventually withdraws the 50 ppt subpoenas with prejudice and strikes any reference to the 50 ppt study from the agency record, ... Dow’s need for the 25 ppt and 5 ppt material .. . would be substantially diminished.”)
. The subpoenas required production of all “documents and records” related to the 25 ppt and 5 ppt studies and defined those terms as:
“[A]U letters, memoranda, correspondence, reports, notes, drafts, working papers, protocols for scientific studies, laboratory notebooks, raw data, data compilations, graphs, charts or papers of any kind, whether handwritten, typed, printed, or reproduced photo-statically or photographically, all film, photographs, videotapes, drawings, or other visual representations, and all magnetic, mechanical, or electronic recordings or other forms of data compilation. The term ‘documents and records’ does not include articles published in recognized scientific journals of wide circulation.”
. Appellant’s Reply Brief, p. 23.
. See also the dissent of Justice Douglas in Board of Regents v. Roth, 408 U.S. 564, 581-582, 92 S.Ct. 2701, 2711-12, 33 L.Ed.2d 548 (1972), quoting Justice Frankfurter’s statement with approval.
. Joint Appendix, p. 28.
. Cf. Gavison, Privacy and the Limits of Law, 89 Yale L.J. 421, 448 (1980):
“[P]rivacy ... contributes to learning ... by insulating the individual against ridicule and censure at early stages of groping and experimentation. No one likes to fail, and learning requires trial and error .... In the absence of privacy we would dare less, because all our early failures would be on record. We would only do what we thought we could do well. Public failures make us unlikely to try again.” (Footnote omitted.)
. It is not unduly speculative to imagine that a large private corporation, through repeatedly securing broad-based subpoenas requiring total disclosure of all notes, reports, working papers, and raw data relating to on-going studies, could make research in a particular field so undesirable as to chill or inhibit whole areas of scientific inquiry. Of course in the instant case we are concerned only with a single, albeit broad, demand for disclosure.
. Richards of Rockford was an action for breach of contract and defamation. In the course of discovery, the plaintiff moved to compel a third party university professor to testify and produce documents concerning confidential interviews with defendant’s employees made as part of a research project. The district court, in denying the motion, recognized that “society has a profound interest in the research of scholars, work which has the unique potential to facilitate change through knowledge,” and found that that consideration as well as other factors weighed in favor of non-disclosure. 71 F.R.D. at 390.
. Joint Appendix, p. 27.
. Paragraph 25 of Dr. Allen’s affidavit and paragraph 14 of Mr. Van Miller’s second affidavit alleged that Dow breached an understanding of confidentiality relating to the 500 ppt study by releasing the tissue analysis from that study to Dr. Henry Spencer, an EPA staff member. Paragraphs 3, 5, and 6 of the second affidavit of Dr. R. J. Kociba, a scientist for Dow, denied either an agreement of confidentiality or an intentional breach and suggested that, “Assuming that Dr. Spencer received the tissue analyses from Dow, my best assumption is that he may have happened upon the correspondence and notes in the files during the course of an EPA audit of Dow’s TCDD studies and analyses.”