(dissenting).
The real controversy before us involves Shaw’s notes entitled “Re: Conversation with Nelson Cantellops on October 10, :[957” and “10/11/57.” It is upon the failure of the Government to produce these notes that defendants’ motion for a new trial must rest.
Judge Murphy held that these “notes most assuredly should have been produced pursuant to § 3500.” 200 F.Supp. 711, 715-716. Yet he denied the motion for a new trial on the ground that the failure to produce them was not prejudicial. Judge Murphy concluded that even though these notes were “statements” and hence producible under the Jencks Act, and even though they contained new impeaching information not otherwise available to defendants, “in the context of this long trial and in view of the fact that the witness’s credibility had been so exhaustively explored we hold that the failure to produce Shaw’s [October 10-11, 1957] notes was not prejudicial to defendants.” 200 F.Supp. at 717. An appeal was taken from Judge Murphy’s order denying the motion for a new trial and in our opinion of March 8, 1963' we affirmed. We never reached the question of whether the nonproduction was. prejudicial. Instead, we took the position that these notes were not statements and hence not producible within the meaning of the Jencks Act. 315 F.2d 186, 191-192. We prefaced our inquiry into the question whether these notes were producible with the announcement that we were going to give very little weight, if any, to the findings of the District Judge: “As we have available to us all of the materials upon which Judge Murphy based his rulings, our conclusion with respect to this factual issue is entitled to no less weight than that of the Court below.” Id. at 191.
On May 27, 1963, little more than two' months after we handed down our opinion in the appeal from Judge Murphy’s order, the Supreme Court decided Campbell v. United States [Campbell II], 373 U.S. 487, 83 S.Ct. 1356, which held that a dis*561trict judge’s findings of fact supporting a ruling that a writing is a “statement” within the meaning of the Jencks Act and hence producible should not be disturbed by a circuit court of appeals unless these findings are “clearly erroneous.” On October 21, 1963, the petitions for writs of certiorari in these cases were granted by the Supreme Court, our judgment “vacated,” and the cases remanded to this Court “for consideration in light of” Campbell II and “for such further consideration as may be appropriate.” Evola v. United States, 375 U.S. 32, 33, 84 S.Ct. 24, 11 L.Ed.2d 106. Although this Supreme Court mandate is not free of ambiguity, I feel that my Brothers have misinterpreted its import. I, therefore, must respectfully dissent.
The most natural interpretation of the Supreme Court’s mandate is that we were in error in making the de novo inquiry into the question of whether the notes were producible and that the clearly-erroneous standard of Campbell II would apply even if the district judge’s factual findings were made under § 3500 (e) (2) rather than § 3500(e) (1) 1 and the evidence before the district judge was written and documentary rather than live testimony.2 Our judgment was vacated because it was premised on a conception of the role of a reviewing court which Campbell II implied, if not actually held, was mistaken. Yet the Supreme Court’s remand mandate left enough room for us to decide whether the findings of Judge Murphy were “clearly erroneous,” and, if we decided that question in the negative, to decide whether Judge Murphy erred in holding that the unlawful nonproduction was non-prejudicial and not sufficient to warrant a new trial.3
There are, of course, other reasonable interpretations of the remand mandate; for example, since our decision predated Campbell II by two months, the Supreme Court may have wanted to give us an opportunity to decide whether the “clearly erroneous” standard applied when the evidence before the trial judge was entirely written and documentary and then to take other appropriate action. Or it is even conceivable that the Court wanted us to follow the procedure used by the First Circuit after the remand in Campbell I, 365 U.S. 85, 81 S.Ct. 421, 5 L.Ed.2d 428 (1961)—retain jurisdiction on the appeal, and return the papers to Judge Murphy with instructions to take the necessary testimony and evidence to determine whether the notes were “statements” within the meaning of the statute. 296 F.2d 527, 534 (1 Cir. 1961).
The majority here is not content with choosing among these interpretations of the Supreme Court’s mandate, but instead takes quite a different tack: “As Judge Murphy’s findings of fact were induced by a mistaken view of the law, the limited review of a district judge’s findings prescribed by Campbell II does not apply, and, in consequence, we were free to determine the facts for ourselves.” I take issue with this disposition, not simply because it is wholly foreign to what was envisioned in the Supreme Court’s remand mandate, but for more basic reasons.
The premise underlying the majority’s position is that Judge Murphy used the wrong standard of law in determining whether the notes are “statements.” I *562cannot agree. True, Judge Murphy asked whether the Shaw notes “set forth in substance what the witness said,” but there is not the slightest intimation that this formula was anything less than a paraphrase of the “substantially verbatim recital” phrase of the statute. In fact, Judge Murphy often reverts to the statutory language in his discussion of these particular notes.4 There is no indication that under his view of the law he would have accepted, as the “statements” required by the Jencks Act, mere “summaries” “which evidence substantial selection of material or which were prepared after-the interview without the aid of complete notes.” He focused on the contemporaneous aspect of Shaw’s transcription and on the resemblance between what Shaw attempted to do and what a stenographer would attempt to do. It would be unfortunate to abstract the formula Judge Murphy employed from his actual use of it and to build a case on the presence of the words “in substance.”
Furthermore, even if I were willing to concede, and I am not, that Judge Murphy used the wrong standard of law, it does not at all follow that we, as circuit judges, are “free to determine the facts for ourselves.” Campbell II’s clearly-erroneous standard would of course not apply if the district judge used the wrong standard of law, but the basic value judgment underlying Campbell II, namely, that district or trial judges and not circuit judges are the ones best suited to make the findings of fact required in applying the Jencks Act (see 373 U.S. at 493, 83 S.Ct. 1356), suggests to me that the proper course for us to take would be to vacate the judgment of the district court, state the correct standard of law that should be used by the district court and remand to the district court to make the appropriate findings of fact. Cf. Rogers v. Richmond, 365 U.S. 534, 81 S.Ct. 735, 5 L.Ed.2d 760 (1961); Townsend v. Sain, 372 U.S. 293, 83 S.Ct. 745, 9 L.Ed.2d 770 (1963). It seems that this course would be the one to be followed under the majority’s premise that Judge Murphy used the wrong standard even if most of the evidence thus far is written and documentary, especially in light of the fact that on the remand to the district court Shaw could be called on to testify. Testimonial evidence may always become relevant and available when the new and proper standard is being applied and I fear that if we decided the facts for ourselves de novo we would be depriving the parties of the rich and special experience which a trial judge could always bring to any fact-finding process.
. My Brothers acknowledge this, see supra p. 556.
. As my Brothers recognize, supra pp. 556, 557 the Supreme Court in Campbell II based the clearly-erroneous standard on something more than the trial judge’s unique position to evaluate the “demeanor of witnesses.” See 373 U.S. at 493, 83 S.Ct. at 1360. Of course, as was true in Campbell II, the clearly-erroneous standard applies even if the district judge is not the same one who conducted the original trial. S.Ct. 1356.
. This interpretation appears all tire more reasonable because the Solicitor General, in opposing defendants’ petition for certiorari, conceded that, under Campbell II, Judge Murphy’s findings that these notes are producible should not be disturbed. Brief in Opposition, p. 36.
. “And since that statute requires production of statements if only ‘substantially verbatim’ and recorded but ‘contemporaneously,’ then so long as such a statement is relevant to the direct testimony of the witness it would be producible under the statute notwithstanding that it were written or recorded by a government attorney, at least if taken during the investigative stage of the case as here, as distinguished from notes made in preparation for trial.
“These October, 1957, notes were therefore producible if they contained ‘in substance what the witness said,’ there being no question here but that they were in fact contemporaneously written by Shaw. From the beginning to the end of these particular notes they are written in the first person and they contain numerous quotes for responses or questions of third persons with whom Cantellops apparently narrates he was conversing. This, together with the very ‘flavor’ of the notes, and bearing in mind Shaw’s testimony at the post-trial hearing, inter alia, as to the difficulty of taking a statement from Cantellops, constrains us to conclude (and no voir dire seems necessary for us to so find) that these notes would be found by the trial court to be, and are in fact, substantially verbatim statements, and of course, relevant to Cantellops’ trial testimony on direct examination.” 200 F.Supp. at 715-716. (Emphasis added.) See also Campbell II, 373 U.S. at 495 n. 10, 83 S.Ct. 1356.