Once in a great while, a case comes before this court which makes one wonder whether the judicial system is still equipped to deal with a litigant determined to frustate the workings of justice. Unfortunately, this is such a case.

It has now been seven years since the United Automobile Workers charged the Gyrodyne Company with an unfair labor practice for discharging some 30 union members at the height of an organizing campaign. For those seven years, the company has persistently refused to release relevant documents within its control which have a vital bearing on the proceedings. These documents have been subpoenaed, and a motion to revoke the subpoena has been denied. The company's case has now been the subject of a lengthy hearing before a trial examiner, two Labor Board decisions, and a decision by this court. The time when an effective remedy for the discharged employees might have been afforded passed years ago. Yet as the case comes before this court for the second time, there has still been no sanction imposed on Gyrodyne for its naked, willful suppression of the documents which could conclusively prove its guilt. If one takes the maxims of equity seriously, then the judiciary should not permit a party to profit from his own wrongdoing. See, e. g., Reynolds v. United States, 98 U.S. (8 Otto) 145, 160, 25 L. Ed. 244 (1878). The time has come to stop Gyrodyne from accruing interest on its investment in intransigence.

Despite the protracted character of this litigation, the facts are relatively simple. The Gyrodyne Company of America is a defense contractor specializing in the manufacture of helicopters. It employs some 800 workers in its plant in St. James, New York.1 As of 1964, when the events giving rise to the union's complaint occurred, its only customer was the United States Navy.2

Although previous attempts had been made to unionize Gyrodyne,3 the company was without a union when the United Automobile Workers began their organizing campaign in January 1963. The UAW's efforts apparently met with considerable apathy, and they remained low key until January 1964 when an intensive leafleting campaign began.4 On June 3, 1964 the union held its first open meeting, and subsequent meetings were held on June 17 and June 24.5 On June 10 the union announced that it had received enough authorization cards to petition the Board for an election.6

As the union campaign gathered momentum, the company began a course of conduct which ultimately led to the charges of unfair labor practices that are the subject of this litigation. On March 2 and 3, 1964 three Gyrodyne employees who were UAW members were discharged without warning or explanation.7 Then on June 11, one day after the union had announced its plan to petition for an election, Peter Papadakos, president of Gyrodyne, called all employees together for a speech on the company's prospects. Although the exact content of that speech and of a subsequent one to the company's production and maintenance employees the next day is hotly disputed, it appears that Papadakos promised to bear a greater part of the employees' medical insurance costs8 and warned against the possibility of a "wall" coming between the employees and their prospects for promotion.9 One week after the second of these speeches, the large-scale discharges began. On June 19, eleven employees were fired, ten of whom were members of the UAW. On June 29, an additional 16 UAW members were laid off, and still another union adherent lost his job on July 1.10

With its organizing efforts frustrated by these discharges, the union abandoned its campaign and sought redress before the Board. The union's principal contention was that the men had been fired for their union activity in contravention of Sections 8(a) (1) and 8(a) (3) of the National Labor Relations Act, 29 U.S.C. Secs. 158(a) (1) and 158(a) (3) (1970). The union argued that the labor cutbacks were unprecedented and that they followed company surveillance of and warnings against union adherents.11 In addition, the union offered testimony by Lieutenant Commander Cletus W. Scheperle, who was naval resident in charge of the Gyrodyne plant, tending to show that Papadakos had a general anti-union bias12 and testimony by Papadakos' wife and father-in-law who claimed that Papadakos had admitted to them that he had fired the men in order to crush the union.13

In defense, the company attempted to discredit a number of the witnesses called by the General Counsel and to demonstrate that some of the individual union members who had been discharged were fired for cause.14 The company's principal contention, however, was that the discharged employees had been caught up in a general cost-cutting program which Gyrodyne had instituted at the behest of the Government.15 President Papadakos testified that he had received letters from President Johnson and Defense Secretary McNamara, similar to those sent to other defense contractors around the time President Johnson assumed office, urging him to cut costs, and that he had responded to these pleas by increasing the efficiency of his plant, thereby necessitating the loss of some jobs.16

In order to meet this "cost-cutting" defense, the General Counsel subpoenaed a number of company records, including the payroll and personnel records of all persons hired or rehired in the company's production departments in 1964. The hiring records were particularly vital since if they showed that the discharged union adherents had been merely replaced it would be obvious that the cost-cutting defense was no more than a sham. But despite the fact that the company's motion to revoke the subpoena was denied by the Board, the records were never produced. Instead, the company chose to parry the General Counsel's repeated requests for the documents at the hearing and to rely solely on President Papadakos' self-serving oral statement that the men had not in fact been replaced.17

When the trial examiner handed down his final decision, the company's strategy of noncompliance seemed fully vindicated. The trial examiner chose to credit Papadakos' testimony and accept the costcutting defense.18 He discounted Lieutenant Commander Scheperle's testimony,19 found that Papadakos' father-in-law was an "inept stooge"20 and totally disbelieved Papadakos' wife.21 Although the examiner had indicated during one stage of the proceedings that he might attach an adverse inference to Gyrodyne's defiance of the subpoena,22 he gave no weight to this defiance in his final decision. Instead, the examiner's opinion totally ignores the records in question and the company's failure to produce them. In a four-paragraph opinion, a panel of the Labor Board adopted the trial examiner's proposed opinion and dismissed the complaint in its entirety.23

Thereupon, the union appealed to this court, arguing, inter alia, that the Board had abused its discretion by refusing to draw the usual adverse inference from Gyrodyne's refusal to obey the subpoena. In our first opinion in this case, we held that "[t]he subpoenaed material appears clearly relevant"24 and that "[i]f the adverse inferences were not to be drawn, failure to do so should have been explained."25 We therefore remanded to the Board in order that it might adopt one of three options: (1) explain its failure to draw the requested inferences, (2) draw the inferences and explain the consequences, or (3) require production of the records.26

On remand, the Board issued a notice to show cause why an adverse inference should not be attached to Gyrodyne's continued refusal to produce the subpoenaed records and, if it should draw such an adverse inference, why it should not reverse its initial decision dismissing the complaint.27 After receiving submissions from all parties, the Board issued a supplemental decision "adher[ing] to its original Decision and Order."28 Purporting to adopt the first alternative proposed in our opinion, the Board once again declined to attach an adverse inference to Gyrodyne's conduct and affirmed its decision to dismiss the complaint.29

Although the Board's supplemental decision discusses at length its refusal to draw an adverse inference from Gyrodyne's nonproduction of other, less important, documents, it devotes no more than a few concluding sentences to the nonproduction of the vital hiring records. While the Board's discussion of this point is exceedingly terse and elliptical, we think we are able to discern five reasons offered for the failure to draw the requested inference with respect to these records: (1) The trial examiner credited the oral testimony tending to show that none of the laid-off employees was subsequently rehired or replaced and that further terminations had later taken place which were not even alleged to be discriminatory30; (2) the General Counsel and the union refused to use some of the documents which Gyrodyne did in fact produce31; (3) Gyrodyne did produce many of the documents requested32; (4) even if an adverse inference were drawn from Gyrodyne's refusal to produce the documents, there was still not a sufficient evidentiary basis for reversing the original decision33; and (5) the General Counsel had adequate opportunity to seek enforcement of his subpoena in the District Court.34

Perhaps sensing that these reasons are less than fully convincing, counsel for the Board in its brief before this court advanced five additional reasons why the inference was not drawn: (6) While the adverse inference rule can be utilized to provide cumulative evidence, it will not supply the missing elements of proof to a party who has failed to make out a prima facie case35; (7) neither counsel for the union nor the General Counsel objected to the introduction of secondary evidence to prove Gyrodyne's rehiring record at the hearing-the point is therefore not open on appeal36; (8) Papadakos testified as to his own personal knowledge of Gyrodyne's rehiring record-not as to the content of the documents not produced-therefore the adverse inference rule does not apply37; (9) neither the General Counsel nor the union requested introduction of the suppressed documents during the hearing38; and (10) even in a case where the adverse inference rule does apply, the rule is permissive only and never requires the body of first impression to draw the inference if it chooses not to do so.39

The Board's and counsel's thoroughness in uncovering every conceivable argument for the Board's result is commendable. Nonetheless, it seems to us that there is some danger of confusing reasons that are compelling with those that are merely compendious. After giving careful examination to each of the explanations offered by the Board and its counsel, we are forced to conclude that none of them justifies the action which the Board has taken. In the aggregate, they amount to no more than a continued insistence that the adverse inference rule has no relevance to the Board's proceedings in the face of a previous decision by this court to the contrary. It therefore becomes our duty to reverse the Board for a second time and to order it to draw the inference adverse to Gyrodyne.

The Board's complex gyrations might lead the unwary to conclude that the adverse inference rule is one of those intricate gems of the common law which is riddled with nonsensical exceptions, encrusted with gloss upon gloss, and surrounded by an arcane lore last fully explicated in a three-volume treatise published in the late 19th century. In fact, however, the rule is disappointingly free of mystery and mumbo-jumbo. Indeed, it is more a product of common sense than of the common law.

Simply stated, the rule provides that when a party has relevant evidence within his control which he fails to produce, that failure gives rise to an inference that the evidence is unfavorable to him. As Professor Wigmore has said:

"* * * The failure to bring before the tribunal some circumstance, document, or witness, when either the party himself or his opponent claims that the facts would thereby be elucidated, serves to indicate, as the most natural inference, that the party fears to do so, and this fear is some evidence that the circumstance or document or witness, if brought, would have exposed facts unfavorable to the party. These inferences, to be sure, cannot fairly be made except upon certain conditions; and they are also always open to explanation by circumstances which make some other hypothesis a more natural one than the party's fear of exposure. But the propriety of such inference in general is not doubted."40

Although this rule can be traced as far back as 1722 when it was applied in the famous case of the chimney sweep's jewel,41 it has been utilized in scores of modern cases as well. See, e. g., Interstate Circuit, Inc. v. United States, 306 U.S. 208, 226, 59 S. Ct. 467, 474, 83 L. Ed. 610 (1939) ("The production of weak evidence when strong is available can lead only to the conclusion that the strong would have been adverse."); United States v. Roberson, 5 Cir., 233 F.2d 517, 519 (1956) ("Unquestionably the failure of a defendant in a civil case to testify or offer other evidence within his ability to produce and which would explain or rebut a case made by the other side, may, in a proper case, be considered as a circumstance against him and may raise a presumption that the evidence would not be favorable to his position."); Tendler v. Jaffe, 92 U.S.App.D.C. 2, 7, 203 F.2d 14, 19 (1953) ("[T]he omission by a party to produce relevant and important evidence of which he has knowledge, and which is peculiarly within his control, raises the presumption that if produced the evidence would be unfavorable to his cause.").42

Given the widespread acceptance of the rule, it is hardly surprising that the Labor Board itself has used it on numerous occasions. As the trial examiner stated in Welcome-American Fertilizer Co., 169 N.L.R.B. 862, 870 (1968), in an opinion subsequently adopted by the Board and in a case which, incidentally, is virtually on all fours with this one: "Respondent's unexplained failure to support and substantiate its economic justification for the layoffs by the production of probative and material documentary records within the power of the Respondent to produce, renders the purported reasons dubious and also warrants drawing an inference that if such [records] had been produced, they could not have been favorable to the Respondent. This failure to produce such evidence 'not only strengthens the probative force' of its absence 'but of itself is clothed with a certain probative force."' (Quoting from Paudler v. Paudler, 5 Cir., 185 F.2d 901, 903 (1950), cert. denied, 341 U.S. 920, 71 S. Ct. 742, 95 L. Ed. 1354 (1951).) See also Monahan Ford Corp., 173 N.L.R.B. 204 (1969); Mid States Sportswear, Inc., 168 N.L.R.B. 559 (1967); Crow Gravel Co., 168 N.L.R.B. 1040, 1047 (1967). Moreover, the courts have consistently upheld the Board when it has drawn an adverse inference from nonproduction of relevant evidence, see P. R. Mallory & Co. v. NLRB, 7 Cir., 400 F.2d 956, 959 (1968); NLRB v. A. P. W. Products Co., 2 Cir., 316 F.2d 899, 903-904 (1963); NLRB v. Wallick, 3 Cir., 198 F.2d 477, 483 (1952); NLRB v. Remington Rand, Inc., 2 Cir., 94 F.2d 862, 868, cert. denied, 304 U.S. 576, 58 S. Ct. 1046, 82 L. Ed. 1540 (1938),43 and have suggested that in proper circumstances they might reverse the Board for an unexplained failure to draw the inference. See NLRB v. Selwyn Shoe Manufacturing Corp., 8 Cir., 428 F.2d 217, 225 (1970); NLRB v. Ford Radio & Mica Corp., 2 Cir., 258 F.2d 457, 463 (1958).44 Indeed, the trial examiner in this very case, while inexplicably refusing to apply the rule in the face of Gyrodyne's stubborn defiance of the subpoena, nonetheless apparently applied it without hesitation in order to derive an adverse inference against the union. See Gyrodyne Company of America, Inc., 170 N.L.R.B. 236, 250 (1968) ("Alfieri's statement that Roman did in fact go to Florida is not denied.").

*****

* * *

Before proceeding with a discussion of how the adverse inference rule applies to the facts of this case, we should note a few more of its general characteristics particularly relevant here. First, it is important to realize that the applicability of the rule in no way depends on the existence of a subpoena compelling production of the evidence in question. The theory behind the rule is that, all other things being equal, a party will of his own volition introduce the strongest evidence available to prove his case. If evidence within the party's control would in fact strengthen his case, he can be expected to introduce it even if it is not subpoenaed. Conversely, if such evidence is not introduced, it may be inferred that the evidence is unfavorable to the party suppressing it. Of course, if a party has good reason to believe his opponent has failed to meet his burden of proof, he may find no need to introduce his strong evidence. See NLRB v. A. P. W. Products Co., supra, 316 F.2d at 903. Similarly, if the other party or the judge plays a role in suppression of the evidence, the force of the inference is dissipated. See NLRB v. Drennon Food Products Co., 5 Cir., 272 F.2d 23, 27 (1959). These special exceptions should not, however, be allowed to detract from the more general, commonsense observation that in most cases a party will introduce his most favorable evidence without being compelled by legal process to do so.

But while the adverse inference rule in no way depends upon the existence of a subpoena, it is nonetheless true that the willingness of a party to defy a subpoena in order to suppress the evidence strengthens the force of the preexisting inference. Indeed, in some circumstances defiance of a subpoena may justify striking a defense, cf. Hammond Packing Co. v. Arkansas, 212 U.S. 322, 351, 29 S. Ct. 370, 53 L. Ed. 530 (1909), or completely barring introduction of evidence on the point in question. Cf. Rule 37(b) (2) (ii), Fed. R.Civ.P. The reason why existence of a subpoena strengthens the force of the inference should be obvious. If a party insists on withholding evidence even in the face of a subpoena requiring its production, it can hardly be doubted he has some good reason for his insistence on suppression. Human experience indicates that the most likely reason for this insistence is that the evidence will be unfavorable to the cause of the suppressing party.

Moreover, the adverse inference rule plays a vital role in protecting the integrity of the administrative process in cases where a subpoena is ignored. It is, of course, always possible for the opposing party to seek enforcement of the subpoena in court. But enforcement against a really intransigent party can be costly and time consuming, particularly in administrative proceedings where the enforcement process is of necessity collateral to the main case. See Federal Maritime Com'n v. New York Terminal Conference, 2 Cir., 373 F.2d 424, 426 (1967). Cf. Wilmot v. Doyle, 9 Cir., 403 F.2d 811, 815 (1968). The adverse inference rule allows a tribunal to attach weight to a party's intransigence without resorting to the awkward enforcing process. It permits vindication of the tribunal's authority in situations where vindication might, as a practical matter, be impossible otherwise. See NLRB v. Selwyn Shoe Manufacturing Co., supra, 428 F.2d at 225. At the same time, the rule does not deprive the suppressing party of his constitutional right to an enforcement hearing on the subpoena, since he is not required to produce the evidence in question. Rather the tribunal simply utilizes the commonsense inference that if the evidence would do the suppressing party any good, he would readily produce it.45

Finally, it should be noted that, contrary to the Board's assertions in its brief,46 the adverse inference rule has nothing whatever to do with the separate rules requiring, under some circumstances, the production of best evidence. The best evidence requirement is an exclusionary rule which suppresses weak evidence in situations where nonproduction of best evidence is unexplained. In contrast, the adverse inference rule excludes no evidence. Moreover, again unlike the best evidence rule, the adverse inference rule need not be triggered by an attempt to introduce weak evidence. Instead, the rule allows a party to introduce any evidence for what it is worth, but always subject to the independent inference that nonproduction of putatively strong evidence indicates that this evidence would be harmful to the suppressing party.

III. The Enforceability of the Adverse Inference Rule

The mere fact that the adverse inference rule is widely recognized and followed does not, by itself, demonstrate that the Labor Board commits reversible error when it declines to utilize it. Generally, as the dissent argues, whether to draw the inference is a matter of discretion for the fact finder. Administrative agencies are, of course, required to obey the minimal requirements of rationality, see, e. g., Burlington Truck Lines, Inc. v. United States, 371 U.S. 156, 168, 83 S. Ct. 239, 9 L. Ed. 2d 207 (1962), and due process, see, e.g., Ohio Bell Telephone Co. v. Public Utilities Comm'n of Ohio, 301 U.S. 292, 301-305, 57 S. Ct. 724, 81 L. Ed. 1093 (1937). But it does not therefore follow that Wigmore's treatise has been incorporated by reference into the Administrative Procedure Act. Indeed, the trend has been toward freeing administrative bodies from the shackles imposed by some of the more arbitrary rules of evidence, see, e. g., John W. McGrath Corp. v. Hughes, 2 Cir., 264 F.2d 314, cert. denied, 360 U.S. 931, 79 S. Ct. 1451, 3 L. Ed. 2d 1545 (1959); 2 K. Davis, Administrative Law Treatise Sec. 14.01 (1958), and toward allowing such bodies to use "the kind of evidence on which responsible persons are accustomed to rely in serious affairs."47

Nonetheless, it seems to us there are several special reasons why we should continue to insist that the Labor Board apply the adverse inference rule. First, it should be noted that the movement away from the strict rules of evidence in administrative proceedings has centered on the various exclusionary rules. It has been argued with some force that it is pointless to insist on such rules, which were designed for jury trials, in situations where the trier of fact will have to hear the evidence in any event in order to rule on its admissibility. See Davis, Hearsay in Nonjury Cases, 83 Harv.L.Rev. 1362 (1970). But whatever the merit of this argument, it clearly has no relevance to the adverse inference rule which, as demonstrated above,48 requires exclusion of no evidence. Instead, the rule requires that evidence which might otherwise be ignored be considered-viz., the evidence that one party has suppressed relevant data and therefore might have something to hide. This is precisely "the kind of evidence on which responsible persons are accustomed to rely in serious affairs," and therefore precisely the type of evidence which the trier of fact should consider, whether he be judge, trial examiner, or member of a jury. The argument for allowing all evidence to be admitted "for what it is worth" in administrative proceedings thus cuts in favor of the adverse inference rule rather than against it.

Second, even if it is conceded that judicial supervision of agency rules of evidence is declining, there are nonetheless some special statutory requirements surrounding the Labor Board which argue for continuing judicial vigilance in at least this one area. 29 U.S.C. Sec. 160 (b) (1970) specifically provides: "Any [Labor Board] proceeding shall, so far as practicable, be conducted in accordance with the rules of evidence applicable in the district courts of the United States under the rules of civil procedure for the district courts of the United States * * *." While the exact contours of this requirement have never been precisely delineated, it is at least clear that the statute will on occasion support reversal of a Board decision when the Board relies on incompetent evidence49 or refuses to hear evidence which is competent.50 As shown above, the adverse inference rule is all but universally recognized in federal courts throughout the country.51 Moreover, there is nothing about the rule which makes it particularly burdensome for the Board to apply or ill-suited for the Board's proceedings. Thus if the requirements of Section 160(b) will ever justify a reviewing court in upsetting a Board decision on evidentiary grounds, it would seem that refusal to apply the adverse inference rule would be such a case.

Third, Gyrodyne's defiance of the Board's compulsory process makes this an especially appropriate case for judicial review. No one doubts that the Board has been given wide discretion and that courts should not lightly upset its judgments. See, e. g., Brooks v. NLRB, 348 U.S. 96, 104, 75 S. Ct. 176, 99 L. Ed. 125 (1954). But "the weight ascribed by law to the [Board's] findings 'rests upon the assumption that the officer who makes the findings has addressed himself to the evidence, and upon that evidence has conscientiously reached the conclusions which he deems it to justify."' Cupples Company Manufacturers v. NLRB, 8 Cir., 103 F.2d 953, 956 (1939), quoting from Morgan v. United States, 298 U.S. 468, 481, 56 S. Ct. 906, 80 L. Ed. 1288 (1936). When a party is in effect denied compulsory process, the failure of the Board to utilize the evidentiary sanctions available to it presents an issue as to the fairness of the decision making process rather than as to the correctness of the decision itself. See NLRB v. Selwyn Shoe Manufacturing Corp., supra, 428 F.2d at 224. Even when courts have felt it necessary to defer to substantive agency determinations, they have continued to insist on their competence to oversee the process by which those determinations were made. As Mr. Justice Cardozo argued long ago:

"Regulatory commissions have been invested with broad powers within the sphere of duty assigned to them by law. * * * Indeed, much that they do within the realm of administrative discretion is exempt from supervision * * *. All the more insistent is the need, when power has been bestowed so freely, that the 'inexorable safeguard' * * * of a fair and open hearing be maintained in its integrity. * * *"

Ohio Bell Telephone Co. v. Public Utilities Comm'n of Ohio, supra, 301 U.S. at 304, 57 S. Ct. at 730, 81 L. Ed. 1093. Where, as here, the Board's refusal to consider the evidentiary inference flowing from the company's nonproduction of its hiring records has the effect of denying a fair hearing to one of the parties, the argument for judicial intervention becomes overpowering.

Finally, and perhaps most significantly, an argument can be made for requiring the Board to apply the adverse inference rule in this case even if the Board is under no general compulsion to follow the rule. This requirement stems from the Board's voluntary adherence to the rule in previous cases indistinguishable from this one in all material respects.52 To be sure, the Board has the power to overrule those decisions, and if it does so the courts will have to respect that determination so long as it meets the requirements of the relevant statutes and the Constitution. But the Board has not chosen to follow that path. Instead, it has gone out of its way to reaffirm the general applicability of the adverse inference rule and to cite with approval previous cases which have adopted it.53

It is an elementary tenet of administrative law that an agency must either conform to its own precedents or explain its departure from them. See, e.g., Secretary of Agriculture v. United States, 347 U.S. 645, 653, 74 S. Ct. 826, 98 L. Ed. 1015 (1954); Columbia Broadcasting System, Inc. v. FCC, 147 U.S. App.D.C. 175, 454 F.2d 1018 (1971). Here the agency obviously has not followed the numerous precedents which seem to require use of the adverse inference rule.54 When this case came before us for the first time, we therefore elected to remand it to the Board so that the Board could explain its failure to follow the rule. On remand, however, the Board neither overruled its prior decisions nor effectively distinguished them. As indicated below, the reasons advanced for ignoring the rule in this one case range from unconvincing to trivial. If we were forced to accept such reasons, the remand would be revealed as a meaningless formality, and the requirement of administrative rationality and consistency as no more than a hollow sham.

Judicial supervision to ensure that agencies are "faithful and not indifferent to the rule of law"55 is too vital for us to be satisfied with a few illogical or hastily concocted rationalizations. Nor does the adverse inference rule rest on so weak a foundation as to topple when confronted with the barely disguised assertion that the rule can be ignored or applied at will. It is sometimes forgotten that the National Labor Relations Act deals with people's livelihoods and that the stakes in an unfair labor practice hearing are consequently extremely high. Given these stakes, we are not about to permit "the proceeding [to degenerate] into a game rather than a hearing to ascertain facts." NLRB v. Selwyn Shoe Manufacturing Corp., supra, 428 F.2d at 225. Nor are we about to watch the adverse inference rule sacrificed on the altar of administrative expertise when the real victim will almost certainly be the integrity and fairness of the Labor Board proceedings.

On its face, this case appears to be a textbook example of a fact situation which fairly begs for application of the adverse inference rule. It can hardly be doubted that Gyrodyne's hiring records are vitally relevant to this litigation and, indeed, we do not understand the Board to contest this proposition. If those records showed that the company had not replaced the men fired in 1964, Gyrodyne's cost-cutting defense would be substantially strengthened. Thus when Gyrodyne chose to suppress the records in the face of the General Counsel's assertion that they would show that the men had been replaced, it is only natural to suppose that the records would have been harmful to Gyrodyne's case. Moreover, this inference is further strengthened by Gyrodyne's defiance of the subpoena even after the petition to revoke had been denied and by the company's blatantly evasive tactics at the hearing.56

Nonetheless, the Board purports to find a plethora of reasons for its failure to apply the adverse inference rule. None of these reasons will withstand even cursory analysis, however, and some of them seem to be based on a total misapprehension of what the adverse inference rule is all about. For example, the Board contends that the rule is inapplicable because the trial examiner believed Papadakos was telling the truth when he asserted that the discharged men had not been replaced. It is, of course, obvious that this "reason" begs the very question to be decided. The principal factual question before the trial examiner was whether the testimony of Papadakos was truthful. One of the tools available to the examiner in evaluating the truthfulness of that testimony was the adverse inference rule. Yet the Board persists in arguing that the rule is inapplicable because it has already decided that question without applying the very rule which was supposed to be an aid in decision. Stripped of its verbal accessories, this formulation amounts to the barefaced assertion that the rule is inapplicable because the Board chose not to apply it. Such tautologies may be quite effective when dealing with a stubborn two-year-old, but they are hardly the type of argument conducive to persuading a court of law.

Next, the Board argues that it should not be required to apply the rule because Gyrodyne did produce many of the documents which were requested of it and because the General Counsel did not use some of the documents which Gyrodyne did produce. But the fact that Gyrodyne graciously condescended to obey part of the Board's subpoena in no way changes the fact that it remains in flagrant violation of the rest of it. It has never been thought that a party is free to pick and choose which parts of a subpoena it will obey and which parts it can ignore. Indeed, if anything, Gyrodyne's policy of selective obedience reinforces the adverse inference since it demonstrates that the company was quite capable of coming forward with evidence when it was favorable to its cause. Nor do we know of any cases holding that the party who subpoenas documents must introduce them on pain of losing the right to subpoena anything else. If Gyrodyne was so anxious to get those documents which it produced into the record, it could have introduced them itself. Surely the General Counsel's failure to introduce them has no bearing on the question whether Gyrodyne should obey the rest of the subpoena. Indeed, the Board makes no attempt to show a logical connection between the two, and we are unable to see one.

The Board then changes tack and argues that even if the adverse inference were drawn there would still not be a sufficient evidentiary basis for reversing the original decision. It seems to us, however, that it is insufficient for the Board to speculate on the possible results if the inference were drawn. Rather, absent a valid reason for bypassing the rule, the inference should actually be drawn and its impact evaluated. While this requirement may seem like no more than a technicality, it in fact has some important consequences. If the Board had actually drawn the inference, it would have had to conclude that the company's records would have shown that the discharged men were in fact replaced. This conclusion leads, in turn, to the further conclusions that Papadakos had perjured himself on the stand and that the cost-cutting defense was a blatant sham. To be sure, these propositions are not, in themselves, sufficient to make out a Section 8(b) (3) violation. But when they are put together with the other facts appearing in the record, they point very strongly to the ultimate conclusion that such a violation occurred. Indeed, any other decision by the Board might well be irrational.

Finally, the Board concludes the explanation of its refusal to draw the inference with a crowning non sequitur. There is no need to draw the inference, the Board argues, since there was adequate opportunity to enforce the subpoena in the District Court. If we were to take this theory seriously, it would mean that the adverse inference rule could never apply in a case where the suppressed evidence had been subpoenaed. Since it is presumably always open to a party to enforce his subpoena, there would never be reason to attach an inference to nonproduction of evidence covered by the subpoena. Yet, as demonstrated above, to the extent the existence of a subpoena has been thought to have anything to do with the adverse inference rule in the past, it has strengthened rather than weakened the inference. Surely a party should not be in a stronger position because he has willfully defied a subpoena than he would be in if he merely failed to introduce evidence which had not been subpoenaed. Yet by carving out an exception to the adverse inference rule for cases where the suppressed evidence is covered by an outstanding subpoena, the Board produces precisely this result.

Moreover, the suppressed premise in the Board's reasoning-that the subpoena was in fact readily enforceable in the District Court-is hardly apodictic. In fact, there are numerous cases holding that the charging party does not have standing to enforce a Board subpoena in collateral proceedings. See, e. g., Wilmot v. Doyle, supra; NLRB v. Selwyn Shoe Manufacturing Corp., supra, 428 F.2d at 224-225. The union's only remedy in the face of the company's intransigence and the Board's refusal to draw an inference from that intransigence was to continue with the hearing and then to assign as error the Board's failure to enforce the subpoena in a subsequent petition to reverse the Board's decision. True, it was open to the General Counsel to petition for immediate enforcement of the subpoena. But we can see no reason why the union should be penalized because of the General Counsel's laxity.

Nor do we see a justification for requiring either the union or the General Counsel to utilize the cumbersome and time-consuming enforcement procedure when there was an alternative, well recognized means available for vindicating the Board's power to require production of relevant documents. Cf. NLRB ex rel. Kohler Co. v. Gunaca, E.D.Wis., 135 F. Supp. 790, 794-795 (1955), affirmed, 7 Cir., 230 F.2d 542 (1956), vacated on other grounds, 353 U.S. 902, 77 S. Ct. 666, 1 L. Ed. 2d 660 (1957). It is important that we not lose sight of the fact that time is often of the essence in Section 8(b) (3) proceedings. There is a constant danger that the aims of the Act will be frustrated by a party determined to delay the workings of justice. See Note, NLRB Power to Award Damages in Unfair Labor Practice Cases, 84 Harv.L.Rev. 1670, 1672-1674 (1971). Indeed, the courts have discouraged collateral proceedings to enforce subpoenas for the very reason that they add to the already considerable delays inherent in the Board's procedures. See Wilmot v. Doyle, supra, 403 F.2d at 815. The adverse inference rule provides a quick, fair method of encouraging parties to come forward with all material relevant to the controversy between them. It would be foolhardy to jettison the rule on the erroneous assumption that the Board can tolerate enless delay for legal maneuverings while at the same time not sacrificing any of the goals of the National Labor Relations Act. After all, this case itself is already over seven years old.

Since none of the reasons advanced by the Board justifies the action which it took on remand, the Board's supplemental decision must be reversed. We note, however, that in its brief before this court counsel for the Board has offered some additional justifications for the failure to draw the inference which do not appear in the Board's opinion. Under well established principles of administrative law, we would be justified in ignoring these new reasons altogether. See, e. g., Burlington Truck Lines, Inc. v. United States, supra, 371 U.S. at 168, 83 S. Ct. 239, 9 L. Ed. 2d 207; Securities & Exchange Com'n v. Chenery Corp., 332 U.S. 194, 196, 67 S. Ct. 1575, 91 L. Ed. 1995 (1947); Columbia Broadcasting System, Inc. v. FCC, supra, 147 U.S.App.D.C. at 190, 454 F.2d at 1033. Had we desired counsel's reasons why the Board might have decided in the way it did rather than the Board's reasons why it actually did so decide, a remand would hardly have been required. It is, however, unnecessary for us to rest on this salutary administrative law requirement, since in fact counsel's reasons for the failure to draw the inference are no more persuasive than the Board's.

Counsel first suggests that the adverse inference rule cannot be utilized by a party, such as the union, who has failed to produce a prima facie case without the rule. This suggestion is totally without merit. First, it should be noted that, although counsel confidently states this rule as if it were part of the organic law of the land, there is in fact substantial authority to the contrary. See, e. g., Tendler v. Jaffe, supra, 92 U.S.App.D.C. at 6-7, 203 F.2d at 18-19. Cf. Welcome-American Fertilizer Co., supra, 169 NLRB at 870. But see Transcontinental Gas Pipe Line Corp. v. Mobile Drilling Barge, 5 Cir., 424 F.2d 684 (1970).

Of course, in a situation where a party has good reason to believe he will prevail without introduction of all his evidence, it would be unreasonable to draw any inference from a failure to produce some of it. But where, as here, the ultimate decision of the trial examiner remained in doubt until the end, it might well be assumed that the parties will do everything in their power to ensure that that decision will be favorable to them. Cf. NLRB v. A. P. W. Products Co., supra, 316 F.2d at 903-904. In this situation, some courts have found it proper for a tribunal to place a production burden on the party controlling vitally relevant evidence. These courts have treated failure to produce as giving rise to an inference that the suppressed evidence would be unfavorable to the suppressing party-an inference which can aid the other party in making out a prima facie case.

But whatever the abstract merits of limiting the adverse inference rule to situations where the party benefiting from it has made out a prima facie case, such a limitation clearly has no relevance here. In fact, the General Counsel here made out a prima facie case, and the trial examiner so ruled at the close of the General Counsel's evidence.57 Indeed, given the strong proof of discriminatory discharges which the General Counsel produced, any other ruling might well have been irrational.

Next, counsel for the Board contends that the adverse inference rule is inapplicable because the union and the General Counsel neither objected to introduction of secondary evidence as to Gyrodyne's rehiring record nor requested introduction of the suppressed documents. We are forced to conclude from these arguments that the Board's counsel totally misconceives both the purpose and the content of the adverse inference rule. Of course, the General Counsel did not object to production of the secondary evidence, since no valid ground existed for such an objection. As explained above, the adverse inference rule does not exclude any evidence. Gyrodyne was perfectly free to prove its rehiring record by secondary evidence, and whether it chose to do so or not has no effect on the applicability of the adverse inference principle. That principle is triggered by the mere fact that strong evidence which Gyrodyne chose not to use was in existence. It has nothing whatever to do with the evidence which Gyrodyne did choose to use. Nor does the General Counsel's alleged failure to request production of the suppressed documents affect the adverse inference which naturally adheres to Gyrodyne's conduct. In fact, while the record is not perfectly clear, it appears that the General Counsel did request production of the documents.58 Although the General Counsel might have been more careful in following the legal niceties necessary to preserve the point on appeal, his persistent questioning about the documents was at least sufficient to put Gyrodyne on notice that he had an active interest in them. Surely by the time of the first remand, when the documents had been subpoenaed, requested at the hearing, and then been made the basis of a reversal in this court, Gyrodyne must have known that the General Counsel and the union wanted them made part of the record.

Moreover, even if we assume for the moment that the General Counsel failed to request production of the records, this assumption in no way excuses the Board's refusal to draw the inference. As explained above, the adverse inference rule is based on the belief that a party will introduce all relevant evidence which is favorable to him on his own initiative. If a piece of evidence appears clearly relevant but the party without explanation nonetheless fails to introduce it, it must be inferred that that evidence would be unfavorable to him. Thus the adverse inference rule has been applied in countless cases where the party benefiting from it never requested production of the suppressed evidence. See, e. g., Transcontinental Gas Pipe Line Corp. v. Mobile Drilling Barge, supra; Washington Gas Light Co. v. Biancaniello, 87 U.S.App.D.C. 164, 183 F.2d 982 (1950). Given the clearly relevant character of the rehiring records, we can see no reason why Gyrodyne would not have come forward with them of its own volition if they had supported its case.

The final two arguments offered by counsel for the Board may be summarily disposed of. Counsel suggests that the adverse inference rule is inapplicable because Papadakos testified as to his own knowledge of Gyrodyne's rehiring record rather than as to the content of the suppressed documents. But this argument once again evidences a fatal confusion between the adverse inference rule and the quite distinct best evidence rule. In terms of the adverse inference rule, it makes not a whit of difference what Papadakos testified to or, indeed, whether he testified at all. Regardless of what other testimony was in the record, the fact remains that Gyrodyne had within its possession important evidence which it failed to produce and that an inference that that evidence was unfavorable naturally attaches to this conduct.

Finally, counsel for the Board argues that the adverse inference rule is permissive only and need not be applied if the trier of fact chooses to ignore it. Counsel takes this position in the teeth of our first opinion, in which our dissenting brother concurred, which remanded the case to the Board on the premise that the Board had to show valid reasons for not applying the adverse inference rule. Surely the panel in remanding did not intend that the Board's reasons be accepted however irrational they might be. More fundamentally, counsel's argument ignores the underlying supposition upon which the vast structure of administrative law is built. It has always been assumed that administrative agencies are not authorized to act arbitrarily and capriciously and that they, unlike juries, are required to give rational reasons for their decisions. In this connection, it should be noted that the dissent's citation of cases holding that the rule is voluntary in jury cases is wholly inapposite. In fact, there are numerous cases which refer to the rule as involving a rebuttable presumption. See, e. g., United States v. Roberson, supra, 233 F.2d at 519 (failure to produce relevant document "may raise a presumption that the evidence would not be favorable"); Tendler v. Jaffe, supra, 92 U.S.App.D.C. at 7, 203 F.2d at 19 (failure to produce "raises the presumption that if produced the evidence would be unfavorable"). Cf. Interstate Circuit, Inc. v. United States, supra, 306 U.S. at 226, 59 S. Ct. at 474, 83 L. Ed. 610 ("The production of weak evidence when strong is available can lead only to the conclusion that the strong would have been adverse" (emphasis added).). A presumption, of course, is mandatory unless reasons affirmatively appear for not applying it. See H. C. Black, Law Dictionary 1349 (4th ed. 1957). But even if we assume that the rule is merely permissive for juries, that fact in no way excuses arbitrary conduct by an administrative agency. Juries are by their nature discrete entities whose decisions set no precedent and whose mode of operation is deliberately insulated from judicial review. In contrast, administrative agencies must act within the context of their own decisional law, must consider all the probative evidence put before them, and must give reasons for their decisions. See e. g., 5 U.S.C. Sec. 706 (1970); Burlington Truck Lines, Inc. v. United States, supra, 371 U.S. at 168, 83 S. Ct. 239, 9 L. Ed. 2d 207.

There may, in fact, be occasions when valid reasons exist for not applying the adverse inference rule. Cf. NLRB v. Drennon Food Products Co., supra. But saying that the rule can be ignored for no reason is tantamount to saying that it is not a rule at all. Having rejected the other nine reasons advanced by the Board and its counsel, we are not about to validate the Board's conduct for the "reason" that that conduct is supported by no reason.

We come, finally, to the question of remedy. From the discussion above, it should be clear that the Board's refusal to draw an adverse inference from Gyrodyne's suppression of the rehiring records cannot be allowed to stand. However, this conclusion alone does not dictate the choice of any particular procedure to follow our second remand. It would be possible, for example, for us to return the case to the Board for reconsideration in light of this opinion. Alternatively, we could order the Board to seek enforcement of the subpoena in the District Court, or order the Board to draw the adverse inference and evaluate the consequences.

After giving the matter careful consideration, however, we have decided to adopt none of these alternatives. When "[t]he administrative conduct reflected in [the] record is beyond repair," Office of Communication of United Church of Christ v. FCC, 138 U.S.App.D.C. 112, 119, 425 F.2d 543, 550 (1969), and "the proceedings on remand [have] been hopelessly bungled," 138 U.S.App.D.C. at 120, 425 F.2d at 551 (statement of Judges McGowan and Tamm), it is occasionally necessary for the reviewing court to take direct action. See ibid. We are convinced there is no longer anything to be gained by a further remand which would, in essence, offer the Board the same three alternatives it rejected last time. The Board has had its chance to seek enforcement of the subpoena or to ruminate about the adverse inference rule. This case has now reached a posture where there is only one rational course for the Board to follow. Rather than allowing the Board to choose still again whether or not to adopt that course and then reversing it for a third time if it elects to ignore it, we have decided to specify precisely the disposition which the Board must now make of this case. We are therefore ordering the Board to strike Gyrodyne's cost-cutting defense unless Gyrodyne takes advantage of a last chance to produce the documents in question. In doing so, we are motivated in part by the need to bring the seemingly endless years of litigation surrounding this case to a rapid conclusion. Nowhere does the old adage about justice delayed being justice denied have more validity than in the field of labor law. As one commentator has remarked:

"Delay is especially troubling because it allows the party charged with committing an unfair labor practice to harass or injure the charging party with relative impunity by prolonging adjudication. * * * [T]he wilful violator who acts in bad faith has the power to postpone any relief for four or five years. Even then the sanction will merely be enforcement of a Board order directing him to cease and desist from his unlawful practices. Thus, the present system invites a rational person to improve his own position by violating the law and absorbing the half-hearted reprimand."

Note, NLRB Power to Award Damages in Unfair Labor Practice Cases, supra, 84 Harv.L.Rev. at 1673.

To be sure, there is something to be said as well for the glacial grandeur of the law. And clearly a party should never be deprived of his substantive rights because a court or agency is anxious to clear its docket. But to say that Gyrodyne has already had its day in court would be something of an understatement. It has now had seven years in court. Through its deliberate footdragging and continued defiance of the Board's supposedly compulsory process, the company has deprived the union of its right to a fair and expeditious resolution of its complaint.

We do not mean to suggest, however, that we are ordering the cost-cutting defense stricken as punishment for Gyrodyne's prior conduct. Rather, this disposition flows naturally from the adverse inference which the Board must attach to the company's refusal to produce the documents in question. The only rational inference which the Board could draw from the unexplained nonproduction is that the documents would show that Gyrodyne had in fact hired men to replace those whom it discharged. Such a showing-contained in the company's own records-would make it irrational to believe Papadakos' self-serving oral declaration to the contrary. In light of this finding, it is no longer possible for the Board to accept Gyrodyne's cost-cutting defense. If the Board finds-as it now must-that the discharged men were replaced, then it cannot rationally believe the men were discharged in order to save money. It follows that Gyrodyne's cost-cutting defense must be stricken. Cf. Hammond Packing Co. v. Arkansas, supra.

While this result may seem harsh to Gyrodyne, the company must bear the responsibility for its own intransigence. Had Gyrodyne produced the records either when they were first requested or after our first remand, it would have been unnecessary for us to deal so firmly with it now. Moreover, in order to be absolutely certain that no miscarriage of justice occurs, we think the company should be given one last chance to come forward with the documents. Now that the consequences of suppression have been made abundantly clear, surely Gyrodyne will produce the rehiring records if they are in any way exculpatory. If the company still prefers suppression-even at the price of having its cost-cutting defense stricken-then the tenor of the documents will be obvious to all.

By permitting Gyrodyne a last chance to come forward with the documents, however, we do not mean to suggest that the proceedings may be delayed indefinitely while Gyrodyne ponders its decision. We have seen quite enough pondering-and not nearly enough deciding -already. Therefore, the Board should allow the company 30 days to produce the rehiring records. If, by the end of that time, the company has still not come forward with the evidence, the consequences outlined above should swiftly follow.59

None of this is to say that the Board must find that Gyrodyne has committed an unfair labor practice. It is still open to the Board to accept Gyrodyne's other defenses and to conclude that the statute has not been violated. In evaluating these other defenses, however, the Board should take into account the fact that Papadakos' testimony as to the company's rehiring record has now been thoroughly impeached and that he therefore may not be the most reliable of witnesses. The Board is further put on notice that any decision which it makes will be subject to review to assure that it is rational and that it is supported by substantial evidence in the record as a whole. Cf. Universal Camera Corp. v. NLRB, 340 U.S. 474, 71 S. Ct. 456, 95 L. Ed. 456 (1951).

The Board's ultimate authority to resolve this controversy thus remains intact. It retains the power to judge credibility, make findings of fact, and apply the law to those findings. It is only because the Board inadvertently became enmeshed in Gyrodyne's deliberate attempt to obstruct justice that its previous decisions in this case were overturned. The reversals of Board decisions in this context should not be seen as an invasion of its authority. On the contrary, only after parties have learned that they have nothing to gain from defiance and delay will the Board's ultimate authority to enforce the National Labor Relations Act be vindicated. After 35 years, there are apparently some parties who have still not learned that lesson. We remain hopeful that Gyrodyne will be the last pupil who is forced to learn it the hard way.

One last word. Perhaps the most remarkable thing about this case is that after seven years it is still in litigation over the question whether the Board was justified in crediting the testimony of the president of the company as to the company's hiring records while that president continues to defy a Board subpoena to produce those records. A layman, unfamiliar with the esoteric mystery of the law but well fortified with common sense, might well say: "No wonder the Board and this court are behind in their work."

Reversed and remanded with instructions.

TAMM, Circuit Judge (concurring in part, dissenting in part):

The National Labor Relations Board (hereinafter the "Board") is permitted to draw reasonable inferences from facts presented to it, Republic Aviation Corp. v. NLRB, 324 U.S. 793, 800, 65 S. Ct. 982, 89 L. Ed. 1372 (1945), including an adverse inference from the failure to produce relevant evidence. NLRB v. Wallick, 198 F.2d 477, 483 (3rd Cir. 1952). The majority concludes that in the case at bar the Board must draw an adverse inference from Gyrodyne's failure to produce the requested documents.1 While I concur in the thirty day reprieve granted Gyrodyne, I respectfully dissent from the majority's mandate to draw the requested inference.

The adverse inference rule speaks to a permissive inference which may be drawn by the trier of fact, not a mandatory inference which an appellate tribunal directs must be drawn. Although the rule has been applied in the areas of criminal,2 civil,3 and administrative law,4 it has been consistently treated as a permissive one invocable by the trier of fact.5 See, e. g., Interstate Circuit, Inc. v. United States, 306 U.S. 208, 226, 59 S. Ct. 467, 83 L. Ed. 610 (1939) (failure to call witness "is itself persuasive"); Tendler v. Jaffe, 92 U.S.App.D.C. 2, 7, 203 F.2d 14, 19 (1953) ("aids the case"); Washington Gas Light Co. v. Biancaniello, 87 U.S.App.D.C. 164, 167, 183 F.2d 982, 985 (1950) ("permits" the inference); NLRB v. Wallick, supra, 198 F.2d at 483 (Board "warranted" in drawing inference).

Professor Wigmore states the rule as follows: "The nonproduction of evidence that would naturally have been produced by an honest and therefore fearless claimant permits the inference that its [tenor is unfavorable to the party's cause]." 2 J. Wigmore, Evidence Sec. 285, p. 162 (3rd ed. 1940) (Emphasis supplied). Where there has been a failure to produce, Professor Jones states the "court may properly instruct the jury that they may infer that the documentary evidence would have operated unfavorably to the party refusing to produce it." 1 B. Jones, Evidence, Sec. 28, p. 62 (5th ed. 1958) (Emphasis supplied).

In speaking of inferences this court has stated that it is a "conclusion which the jury is permitted, but not compelled, to draw from the facts." Bray v. United States, 113 U.S.App.D.C. 136, 140, 306 F.2d 743, 747 (1962) (Emphasis in original). See United States v. Johnson, 140 U.S.App.D.C. 54, 63-64, 433 F.2d 1160, 1169-1170 (1970); Pendergrast v. United States, 135 U.S.App.D.C. 20, 30-33, 416 F.2d 776, 786-789 (1969); Black's Law Dictionary 917 (4th ed. 1957) ("An inference being a deduction which the trier may or may not make according to his own conclusions.").

The law is succinctly stated in Aetna Casualty & Surety Co. v. Smith, 127 A.2d 556 (D.C.Mun.App. 1956):

Assuming there was sufficient foundation for an unfavorable inference of this kind, the fact that the trial court failed to draw the inference constitutes no error reviewable on appeal. The inference arising from the refusal or unexplained failure to produce relevant documentary evidence is permissive in nature, and is merely another factor which may be given consideration by the trier of the facts when weighing the evidence and determining the credibility of witnesses.

*******

* * *

The trial court could have drawn an unfavorable inference, but it was not required to do so and apparently chose not to. Failure of the trier of the facts to draw an inference, permissible in nature, cannot be advanced to an appellate court as a ground justifying reversal of the judgment.

Id. at 559. See also Edwards v. Zahner, 395 S.W.2d 185, 191 (Mo.1965); Sorby v. Three Rivers Motors, 178 Pa.Super. 187, 114 A.2d 347 (1955).

An inference is within the discretion of the trier of fact. Congress has placed the authority to draw inferences in the Board not the courts. NLRB v. Link-Belt Co., 311 U.S. 584, 597, 61 S. Ct. 358, 85 L. Ed. 368 (1941); NLRB v. Falk Corp., 308 U.S. 453, 461, 60 S. Ct. 307, 84 L. Ed. 396 (1940). Decisions, "beyond [the] possibility of misunderstanding, have repeatedly reminded us that the power to draw inferences was entrusted to the Board and not to the courts." NLRB v. American Creosoting Co., 139 F.2d 193, 195 (6th Cir. 1943). See Standard Generator Serv. Co. v. NLRB, 186 F.2d 606, 607 (8th Cir. 1951); NL RB v. Mt. Clemens Pottery Co., 147 F.2d 262, 264 (6th Cir. 1945). Indeed, the right has been denominated "unquestionably exclusive." NLRB v. Austin Co., 165 F.2d 592, 596 (7th Cir. 1947). Notwithstanding the majority's assertion that several decisions "suggest" that a court might reverse the Board for failing to draw an adverse inference, not a single decision cited by the majority does so.6 Nor is there a decision cited where-in a court ordered the Board to draw an adverse inference.7

Moreover, the decisional law is supported by the facts in the instant case. The Board justifiably refused to draw the inference8 because inter alia credited testimony indicated that none of the discharged employees were subsequently rehired.9 It is established that the credibility of witnesses is a matter peculiarly within the province of the trier of fact. Indeed, it has been held that the evaluation of oral testimony may not be upset by a court unless it is "hopelessly incredible." NLRB v. Dinion Coil Co., 201 F.2d 484, 490 (2nd Cir. 1952). Had the oral testimony been unreliable or controverted the subpoenaed materials would, of course have assumed greater importance.10

The majority itself alludes to the type of evidence in dispute here as that "which the trier of fact should consider, whether he be judge, trial examiner, or member of a jury." It is submitted that this is precisely what transpired. The trier of fact after due consideration concluded that the adverse inference rule should not be applied to the particular facts at bar. We should not substitute our judgment for that of the Board's.

Moreover, the Board in its supplemental decision stated:

Even assuming, arguendo, adverse inferences were to be drawn from the Respondent's failure to produce, we nevertheless do not believe that such inferences as could be drawn would produce a sufficient evidentiary base for reversing our Decision herein, particularly in view of the Trial Examiner's credibility findings which were based on evidence subsequently adduced by the Employer. (Supp.App. 39.)

The majority chides the Board for "speculating" on the possible results of an adverse inference and yet with necromantic elan goes on to do precisely that which it had just condemned by itself engaging in speculation.

Judicial review of administrative decisions is narrowly circumscribed in order to encourage autonomy and allow greater expression to agency expertise. Unfortunately, the decision of the majority abnegates the traditional respect we have accorded agencies as "collaborative instrumentalities of justice."11 I respectfully dissent.