This case arises upon a petition to enforce an order of the Labor Board, whose only direction that we need consider was to reinstate with back pay a 'supervisory employee,' named Chairman, whom the respondent discharged on January 24, 1944, avowedly for insubordination. If the Board was right, the discharge was in fact for giving testimony hostile to the respondent at a hearing conducted by the Board to determine who should be the representative of the respondent's 'maintenance employees.' Chairman was an assistant engineer, whose duties were to supervise the 'maintenance employees,' and he testified at the hearing in favor of their being recognized as a separate bargaining unit. The respondent opposed the recognition of such a unit, and several of its officers testified to that effect, among whom were Shapiro, the vice-president, Kende, the chief engineer, and Politzer, the 'plant engineer.' The examiner, who heard the witnesses, was not satisfied that the respondent's motive in discharging Chairman was reprisal for his testimony; but on review of the record a majority of the Board found the opposite, and on August 31, 1948, ordered Chairman's reinstatement. The respondent argues (1) that the majority's findings are subject to a more searching review under the New Act than under the Old; (2) that in the case at bar the findings cannot be supported, because they are not supported by 'substantial evidence'; and (3) that its liability to Chairman, if any, ended with the passage of the New Act.

The substance of the evidence was as follows. On November 30, 1943, Chairman and Kende testified at the hearing upon representation, after which Kende told Chairman that he had 'perjured' himself; and on the stand in the proceeding at bar Kende testified that Chairman 'was either ignorant of the true facts regarding the organization within the company * * * or * * * he was deliberately lying, not in one instance, but in many instances, all afternoon'; and 'that there was definite doubt regarding his suitability for a supervisory position of that nature.' The examiner believed the testimony of Chairman that two other employees, Goldson and Politzer, had cautioned him that the respondent would take it against him, if he testified for the 'maintenance employees'; and Kende swore that he told another employee, Weintraub- the personnel manager- that he thought that Chairman was a Communist. After Politzer reported to him on December second or third that this was a mistake, Kende told him to keep an eye on Chairman. From all this it is apparent that at the beginning of December Kende was hostile to Chairman; but he took no steps at that time to discharge him.

Nothing material happened until the very end of that month, when Chairman and Weintraub got into a quarrel, about disciplining a workman, named Kollisch. Chairman swore that Weintraub demanded that he discharge Kollisch for loafing; and Weintraub swore that he only demanded that Chairman put Kollisch to work. In any event high words followed; Chairman told Weintraub that he was drunk; Weintraub brought up a plant guard to put Chairman out of the premises, and the quarrel remained hot, until one, Zicarelli, a union steward, succeeded in getting the two men to patch up an apparent truce. Two days later Weintraub saw Politzer and told him that he had heard that Politzer was looking into Chairman's statement that Weintraub was drunk, and on this account Weintraub asked Politzer to discharge Chairman. Politzer testified that he answered that Chairman was going to resign soon anyway, and this the examiner believed. He did not, however, believe Politzer's further testimony that Chairman had in fact told Politzer that he was going to resign; he thought that Politzer either was mistaken in so supposing, or that he had made up the story in order to quiet Weintraub. Probably his reason for not believing this part of Politzer's testimony was that he accepted Chairman's testimony that ten days later Politzer intimated to Chairman that it would be well for him to resign, and Chairman refused. Whatever the reason, Weintraub did not, after his talk with Politzer, press the matter until January 24, 1944, when, learning that Chairman was still in the factory, he went again to Politzer and asked why this was. When Politzer told him that Chairman had changed his mind, Weintraub insisted that he must resign anyway, and, upon Politzer's refusal to discharge him, they together went to Kende. Weintraub repeated his insistence that Chairman must go, giving as the reason that his accusation of drunkenness had undermined Weintraub's authority, Kende took Weintraub's view and Politzer wrote out an order of dismissal. No one testified that at this interview, or any time after December first, any of the three mentioned Chairman's testimony at the representation hearing.

As we have said, the examiner was not satisfied that the Board had proved that Chairman's testimony at the representation proceeding had been an actuating cause of his discharge; but, not only did the majority of the Board reverse his ruling as to that, but they also overruled his finding that Politzer had told Weintraub on January first that Chairman was going to resign. They then found that Kende and Weintraub had agreed to bring about Chairman's discharge, at some undefined time after December first, because of Chairman's testimony; and that Weintraub's complaint on January 24 was a cover for affecting that purpose. Whether these findings were justified is the first, and indeed the only important, question of fact; and as a preliminary point arises the extent of our review.

This has been the subject of so much uncertainty that we shall not try to clarify it; but we must decide what change, if any, the amendment of 19471 has made. Section 10(e) now reads that the findings 'shall be conclusive' 'if supported by substantial evidence on the record considered as a whole'; and the original was merely that they should be conclusive, 'if supported by evidence.' In National Labor Relations Board v. Pittsburgh Steamship Company2 the Supreme Court refused to say whether this had made any change, and remanded the case to the court of appeals to decide the point in the first instance. Of the four decisions which have discussed it, two have held that no change, or no material change, was made;3 one has held that the amendment was intended 'to give the courts more latitude on review,' but did not decide how much;4 and the fourth merely held that it did not make the review a 'hearing de novo.'5 (Since the opinion in the last was written by the same judge who wrote the first, it is to be read as deciding that there was no change.) It is true that there were efforts, especially in the House, to give to courts of appeal a wider review than before; but the Senate opposed these, and, so far as concerns the adjective, 'substantial,' it added nothing to the interpretation which the Supreme Court had already put upon the earlier language.6 The most probable intent in adding the phrase, 'on the record considered as a whole,' was to overrule what Congress apparently supposed- perhaps rightly- had been the understanding of some courts: i.e. that, if any passage could be found in the testimony to support a finding, the review was to stop, no matter how much other parts of the testimony contradicted, or outweighed, it. That the words throughout section ten were chosen with deliberation and care is evident from the changes in Sec. 10(c), apparently intended to confine the Board to the record before it, and in Sec. 10(b), restricting it in the admission of evidence to Rule 43(a) of the Federal Rules of Civil Procedure, 28 U.S.C.A. It appears to us that, had it been intended to set up a new measure of review by the courts, the matter would not have been left so at large. We cannot agree that our review has been 'broadened'; we hold that no more was done than to make definite what was already implied.

Just what that review was in another and much more difficult matter- particularly, when it comes to deciding how to treat a reversal by the Board of a finding of one of its own examiners. Obviously no printed record preserves all the evidence, on which any judicial officer bases his findings; and it is principally on that account that upon an appeal from the judgment of a district court, a court of appeals will hesitate to reverse. Its position must be: 'No matter what you saw of the witnesses and what else you heard than these written words, we are satisfied from them alone that you were clearly wrong. Nothing which could have happened that is not recorded, could have justified your conclusion in the face of what is before us.' That gives such findings great immunity, which the Rules7 extend even to the findings of masters, when reviewed by a district judge. The standing of an examiner's findings under the Labor Relations Act is not plain; but it appears to us at least clear that they were not intended to be as unassailable as a master's. The Old Act provided for 'examiners';8 but they did not have to make reports, and, although Sec. 10(c) of the New Act requires them to do that,9 it does not undertake to say how persuasive their findings are to be. On the other hand, Sec. 8(a) of the Administrative Procedure Act10 provides that 'on appeal from or review of' the decision of an 'officer' who has presided at a hearing, 'the agency shall * * * have all the powers which it would have in making the initial decision.' It is clear that these words apply to the decisions of the 'agency' upon the evidence; but nothing is said as to what effect the 'agency' must give to the 'officer's' findings; except that, if the text be read literally, it could be argued that the 'agency' was to disregard it. The reports in Congress do not help very much. The Senate Report11 merely said that the findings 'would be of consequence, for example, to the extent that material facts in any case depend on the determination of the credibility of witnesses as shown by their demeanor or conduct at the hearing.' The House Report12 was the same, in ipsissimis verbis, although it did add that 'in a broad sense the agencies reviewing powers are to be compared with that of courts under Sec. 10(a) of the bill.' That would have made them as conclusive upon an 'agency' as the 'agency's' findings are upon a court; and it is safe to say that the words will not bear so much. When the same question came up under the Old Act, the courts left the answer equally uncertain. The Seventh Circuit in A. E. Staley Manufacturing Company v. National Labor Relations Board13 said that, 'where it' (the Board) 'reaches a conclusion opposite to that of an Examiner, we think the report of the latter has a bearing on the question of substantial support and materially detracts therefrom'; and that has in substance received the approval of the Eighth Circuit,14 and of the Sixth,15 as well as a recent reaffirmation by the Seventh itself.16 All this leaves the question in confusion. On the one hand we are not to assume that the Board must accept the finding, unless what is preserved in the record makes it 'clearly erroneous.' That would assimilate examiners to masters, and, if that had been intended, we should expect a plainer statement. On the other hand, the decisions we have cited certainly do mean that, when the Board reverses a finding, it shall count in the court's review of the Board's substituted finding; the case does not then come up as it does, when the testimony has been taken by deposition.17 On the whole we find ourselves unable to apply so impalpable a standard without bringing greater perplexity into a subject already too perplexing. The weight to be given to another person's conclusion from evidence that has disappeared, depends altogether upon one's confidence in his judicial powers. The decision of a child of ten would count for nothing; that of an experienced master would count for much. Unless we are to set up some canon, universally applicable, like that of Rule 53(e)(2), each case in this statute will depend upon what competence the Board ascribes to the examiner in question. Section 4(a) provides that he shall be an employee of the Board,18 which will therefore have means of informing itself about his work. We hold that, although the Board would be wrong in totally disregarding his findings, it is practically impossible for a court, upon review of those findings which the Board itself substitutes, to consider the Board's reversal as a factor in the court's own decision. This we say, because we cannot find any middle ground between doing that and treating such a reversal as error, whenever it would be such, if done by a judge to a master in equity.

The foregoing discussion is relevant in the case at bar for the following reason. One ground why the evidence failed to convince the examiner of any agreement between Kende and Weintraub to discharge Chairman, was that he thought it quite as likely that the quarrel between Weintraub and Chairman at the end of December still rankled in Weintraub's mind, and induced him to insist upon Chairman's discharge on January 24, 1944. It became important in this view to explain why Weintraub waited for over three weeks; and this the examiner did explain because he believed that Politzer had told Weintraub that Chairman was going to resign. When the majority of the Board refused to accept this finding, they concluded that, since this left Weintraub's delay unexplained, his motive was to be related back to the quarrel of Kende and Chairman on November 30. We should feel obliged in our turn to reverse the reversal of this finding, if we were dealing with the finding of a judge who had reversed the finding of a master, because the reasons given do not seem to us enough to overbear the evidence which the record did not preserve and which may have convinced the examiner. These were (1) that the examiner did not believe all that Politzer had said; and (2) that the finding was 'irreconcilable with the other related facts and all the other evidence bearing on Politzer's behavior and attitude.' It is no reason for refusing to accept everything that a witness says, because you do not believe all of it; nothing is more common in all kinds of judicial decisions than to believe some and not all. Nor can we find 'other related facts' which were 'irreconcilable' with believing that Politzer told Weintraub that Chairman was going to resign. Indeed Chairman himself swore that on January 11, Politzer suggested to him that he resign, which affirmatively serves to confirm the examiner's finding that Politzer told Weintraub that Chairman would resign in order to placate him. However, as we have said, we think that we are altogether to disregard this as a factor in our review, which we should confine to the bare record; and on that we cannot say that Politzer's testimony had to be believed, in the face of Chairman's denial that he ever told him that he would resign.

There remains the question whether, with this explanation of Weintraub's delay missing, there was 'substantial evidence' that the cause of Chairman's discharge was his testimony; and on that the Board had the affirmative; so that it is not enough that Kende and Weintraub might have agreed to find a means of getting rid of Chairman, or that Kende unassisted might have been awaiting an opportunity. Once more, if this was the finding of a judge, we should be in doubt whether it was sufficiently supported. When Weintraub went to Politzer on January 24, 1944, with his complaint at Chairman's continued presence in the factory, and when the two went to Kende because Politzer would not discharge Chairman, if Weintraub was acting in accordance with an agreement between Kende and himself, he was concealing the facts from Politzer. So too was Kende at the ensuing interview; indeed, we must assume that the two had arranged beforehand to keep Politzer in the dark, else Weintraub could scarcely have relied upon Kende to play his part. This appears to us to be constructed substantially out of whole cloth, so improbable is it that they should have gone to such devious means to deceive Politzer. On the other hand, although it is possible that Kende had been waiting for a proper occasion, independently of Weintraub, and that he seized upon Weintraub's complaint, being secretly actuated by his old grievance, we do not read the majority's decision as distinctly indicating that they meant so to find. But, if they did, unless we assume that Weintraub's complaint was trumped up ad hoc, to deceive Politzer, it becomes the merest guess that Kende did not find it alone a sufficient reason for his action, and reverted to his concealed spite.

Nevertheless, in spite of all this we shall direct the Board's order to be enforced. If by special verdict a jury had made either the express finding of the majority that there was an agreement between Kende and Weintraub, or the alternate finding, if there be one, that Kende without Weintraub's concurrence used Weintraub's complaint as an excuse, we should not reverse the verdict; and we understand our function in cases of this kind to be the same. Such a verdict would be within the bounds of rational entertainment. When all is said, Kende had been greatly outraged at Chairman's testimony; he then did propose to get him out of the factory; he still thought at the hearings that he was unfit to remain; and he had told Weintraub to keep watch on him. We cannot say that, with all these circumstances before him, no reasonable person could have concluded that Chairman's testimony was one of the causes of his discharge, little as it would have convinced us, were we free to pass upon the evidence in the first instance.

The question of law involved does not appear to us difficult. When Chairman was discharged, he was within the protection of the statute. It is true that he ceased to be so protected when the Old Act was repealed in 1947; but the repeal did not extinguish the respondent's 'liability,' if there was any;19 so that the only question is what was the measure of that 'liability.' By that word, we understand the duty which the statute imposed upon the respondent to make restitution for the wrong- the discharge. That comprised the restoration of Chairman to his position as 'supervisory engineer,' and the payment of any loss he may have suffered meanwhile. To preserve the 'liability' is to preserve this duty; and it would be irrelevant, if similar conduct- a discharge for testifying in a proceeding before the Board- would not now be a wrong at all, a question we assume, but do not decide. It is true that Chairman's restoration will give him no immunity from discharge for any reason, or for no reason, except one- that he gave the testimony on November 30, 1943, for which he was discharged on January 24, 1944. Of course by the Board's order he gains no status of which the New Act would have stripped him, and he remained in the respondent's employ. But, since we cannot say that he would not have still been in that employ, save for the wrong, it is part of his remedy that he should have the chance to qualify once more. It is no doubt unfortunate that upon restoration, any later discharge for subsequent cause will be apt to be interpreted as a covert reassertion of the old wrong; but that cannot be helped; it is a consequence of making motive the test of legal wrong.

An enforcement order will issue.

SWAN, Circuit Judge (dissenting).

In National Labor Relations Board v. A. Sartorius & Co., 2 Cir., 140 F.2d 203, 205 we said that 'if an administrative agency ignores all the evidence given by one side in a controversy and with studied design gives credence to the testimony of the other side, the findings would be arbitrary and not in accord with the legal requirement.' I think that is what the majority of the board has done in the case at bar. I would reverse its finding of motive and deny enforcement of the order.