(dissenting in part).
For the following reasons, I dissent as to the conviction of the defendants other than Liss and Conte.
There are some persons (and I am one of them) who have come to doubt the virtues of jury trials in civil actions. However, even of those who entertain such doubts, most agree that in 'Criminal actions the right to a jury trial should be preserved —that is, the accused should be tried by jury if he so elects (although it is perhaps less certain than is generally supposed that defendants fare -better with the average jury than with the average trial judge sitting without a jury). But whatever may be the views of judges about the jury system, it is their duty to maintain the function of the jury in all jury trials, until the jury is abolished by legislation and constitutional amendment. And that means that, in a criminal action tried before a jury, a defendant should not be convicted unless the jury finds him guilty — after a fair trial. Perhaps -I am old-fashioned 'in saying that I believe that the doctrine of “harmless error” does not dispense with the necessity of a fair trial of a defendant whom the appellate judges believe to be guilty. A jury trial does not seem to me to be fair when it plainly appears that the jury has not been well informed as to the facts. As I understand the fundamental principle of the jury system, we appellate judges do *1002not sit as a jury. It surely follows that we ought not to sustain a verdict of guilty after an unfair trial merely because, upon reading the printed record, we believe that, had wo been the jury, we would have convicted. Any other rule has the result that the presence of the jury will often become a mere formality; for then, if only the jury finds guilt, no matter by what unfair tactics it was persuaded to do so, we judges, in actual fact, decide the case.
I cannot subscribe to a rule that what is substantial reversible error depends not on whether it probably affected the jury to the substantial prejudice .of the defendant but on whether we appellate judges think the defendant guilty or innocent. If that is to be the rule, I would urge that criminal jury trials be abolished as expensive and time-consuming shams. I have known lawyers who assert that only by such a rule can the jury system be made “workable.” But such a rule renders the jury system workable by not working it; if we happen to think that the jury system is impractical and unworkable, let us frankly say so; then our citizens, informed of the realities, can decide whether to dispense with it.
At a minimum, if we are to adopt the rule described above, we should make that fact unmistakably clear, so that the public will thoroughly understand what a jury trial now really is — and is not. The “principle of publicity,” which governs in the American judicial process,1 calls not only for trials held in public but also for the publication of judicial opinions that disclose the grounds of decisions; that publicity is illusory if the true grounds are not stated.2 To adopt a rule of decision concerning jury trials without publicizing it prevents the public from knowing how its government operates. That is unwise in a democracy where the workings of all branches of government should be plain to the citizens. There can be no true democracy when citizens are treated like infants who must not be told the facts of life; mysteries as to the workings of government impair one of the most important features of democratic participation — the capacity of citizens intelligently to criticize and seek changes in those aspects of government which they consider undesirable— and breed a sense of frustration and a cynicism which may pave the way to the destruction of democracy.3 If we are to apply a rule by which we, who do not see or hear the witnesses, are, in many cases, to decide the facts, we should so announce. We ought not to become fact-finders while appearing not to be. Moreover, if we are to find the facts, we ought at least to impose on ourselves the obligation we now impose on trial judges when they sit without a jury — that of publishing the findings.4
In the case at bar there was the greatest likelihood that the jury was confused. The indictment, as my colleagues concede, was misleading in that it charged twenty defendants with engaging in one conspiracy when, as the evidence disclosed, there were in fact three distinct alleged conspiracies. One of these conspiracies was not unlawful. Conte and Liss were parties to the other two conspiracies. But, as my colleagues say, there was no evidence whatever that Rudy was a party >to the third, and no evidence that the other defendants, except Liss, Conte and Rudy, were in any way involved in the second. It is difficult for me to believe that the jury, at the end of this long trial, was able to keep in mind the distinction between the second and third conspiracies and to comprehend, with any degree of clarity, that evidence relevant to the second should not be considered in determining whether there was guilty participation in the third, or vice versa.
The jury may well, then, have erroneously combined the evidence and found guilt when it would not have done so had it *1003observed the distinction. In those circumstances, the least that the trial judge should have done was to aid the jury in his charge by differentiating the conspiracies. Not only did he not do so but, as my colleagues point out, his charge was, not too enlightening. Moreover, he expressly refused, when requested by Rudy’s counsel, to instruct the jury that there was no testimony showing Rudy’s connection with the third conspiracy — and that, too, after a trial which consumed eight days in which testimony was taken that fills more than seven hundred and fifty printed pages.5
My colleagues concede that there may have been confusion, but they conclude that this jury, in fact, was not misled. That, I submit, is unverified and unverifiable guessing. It is well-settled that, generally, it is improper for the court to be affected by anything it may learn, even from the jurymen themselves, as to how the jury arrived at its verdict.6 Unable, as we are, therefore, to interrogate the jurors, any effort on our part to determine that they were or were not confused is an undertaking which most psychologists would regard as hopeless. Who are we judges that, in such circumstances, we should so confidently probe the mental interiors of the jurors? The best that can be said here is that the chances are as good that the jurors were bewildered as that they were not. Where such chances are anywhere near equal, there has been substantial error.7 We judges ought to take judicial notice of what every ordinary person knows about juries, and therefore to recognize that the twelve citizens, casually summoned to serve as jurors, are not trained fact-finders and can be easily bewildered. Our experience shows that even trial judges, experienced in finding facts, when sitting without a jury, sometimes do not avoid confusion as to the evidence when, as here, there are many defendants and the trial is not brief. The need for safeguarding defendants from misunderstanding by the jury is peculiarly acute in conspiracy trials which lend themselves to unfairness, since, at best, they often permit the jury to hear evidence as to some of the defendants which the jurymen may easily but mistakenly believe has a bearing on the guilt of others.
It is true that a variance between the indictment and the proof is not invariably fatal error. Whether or not it constitutes such error depends upon whether the record in the particular case does or does not show a likelihood of substantial prejudice. My colleagues suggest that the issue is similar to that of the propriety or impropriety of joining separate crimes in a single indictment or of consolidating separate indictments for trial. Yet in all the cases which they cite, and in which such joinders or consolidations have been held proper, the courts have pointed out either that the separate crimes were closely re*1004lated (in some of them, the proof of the one being necessary for the proof of the other) or that, as in our decision in United States v. Lotsch, 2 Cir., 102 F.2d 35, 36, “the evidence as to each was short and simple; there was no reasonable ground for thinking that the jury could not keep separate what was relevant to each.” That factor of shortness and simplicity — and of consequent ease of comprehension by the jury — is absent here. McElroy v. United States, 164 U.S. 76, 80, 17 S.Ct. 31, 32, 41 L.Ed. 355, condemned the consolidation of an indictment against one defendant with an indictment against another when the charges were “not provable by the same evidence, and in no sense resulting from the same series of acts.” It has been thought that the full rigor of the McElroy doctrine has been somewhat mitigated. But in.the cases where it has been so held, the several charges were very closely connected.8 My colleagues’ ruling here is the farthest north of the McElroy case to be found in the books.
Of course I agree that upper courts should not, by demanding perfection and by flyspecldng scrutiny of trial records,8a, obstruct “the criminal prosecution Of complicated crimes” unavoidably involving large number of defendants. But a trial even for a single conspiracy is complicated. The complexity of such a trial should not be increased by needlessly injecting into it the trial of another Conspiracy. More ought to be done, I think, to prevent prosecutors from employing the excuse of need for “expedition” to use, unnecessarily, conspiracy trials, in which large numbers of defendants are herded into one suit, instead of bringing several actions. The trial dockets are not so congested as to compel such omnibus trials. Any district judge can do much to meet this situation by exercising his discretion, on his own motion, to compel severances.8b At any rate, aware of the potential dangers of injustice involved in all conspiracy actions, we ought to be singularly insistent that they be conducted with exceptional fairness. “Expedition” and “efficiency” in the prosecution of crimes ought not to be purchased at the expense of justice. The recent expansion of the activities of so-called “administrative agencies” has centered attention upon their conduct with a result which, on net balance, is beneficent: the demand that dispatch in administration must not lead to infringements on the rights of citizens. Because courts and prosecuting attorneys have a far older tradition than the administrative agencies, it is often forgotten that they, too, are engaged in administration,9 a truth which, strangely enough, has been obscured by burying it in a revealing phrase, “the administration ■ of justice.” That phrase should bring to mind the fact that courts and district attorneys, no less than administrative agencies, are administrators; their task is to “administer” what we call “justice.” No more than administrative agencies should they emphasize dispatch in administration to the neglect of their primary function — justice. Here, as elsewhere, the major problem of our times is to reconcile the Expert State and the Free State.10
*1005The variance, coupled with the judge’s charge, was especially unfair to Rudy. It was only slightly less so as to all the other defendants except Liss and Conte. True, unlike Rudy, none of them requested a charge that the jury keep in mind the respective participations in the second and third conspiracies. But when, as here, irregularities make it highly probable that there was substantial error, we have one of those exceptional cases in which we have been told that we must not refuse to reverse because of the failure of lawyers for defendants to object and except.11 The devotion to the contentious method of court procedure, according to which each party to the courtroom duel must take the chance that the lawyer he selects as his duelist exercises appropriate skill, can be excessive;12 particularly in criminal trials should judges refuse to be relegated completely to the role of mere umpires;13 the accused should not suffer unduly because of the mistakes of his counsel. A “trial in court is never * * * ‘purely a private controversy * * * 0fn0 importance to the public.’ ”14 The Supreme Court recently admonished us that Congress when, in 28 U.S.C.A. § 391, it provided that appellate courts shall give judgment “without regard to technical errors, defects, or exceptions which do not affect the substantial rights of the parties”, did not intend that, where there have been substantial infringements of the right to a fair trial, we can ignore them.15 I must confess that I do not understand those persons — I do not include my colleagues among them — who, while purporting to worship the jury, applaud when the “harmless error” doctrine is pushed to such an extreme as it is in the majority opinion here. My colleagues, in stating that there is a “modern disposition to assume that an error has been harmless,” have failed to note what five circuit courts have observed :16 there has been a new trend in the Supreme Court concerning that doctrine since at least 1936 ;17 the rule is now said to be that, if error is shown, there must be reversal “unless it affirmatively appears from the whole record that it was not prejudicial.”
See Millar, in Engelman, History of Continental Civil Procedure (1927) 77f.
There Holmes is our guide; he taught that it is wholesome for courts to spell out their mental processes and frankly to acknowledge just what they are doing. As we recently said, “In all provinces of thought, it usually promotes intelligence to bring out into the light the concealed actual postulates of thinking.” Beidler & Bookmeyer, Inc. v. Universal Ins. Co., 2 Cir., Mar. 5, 1943, 134 F.2d 828, 831.
■ As to Holmes, see, e.g., Holmes, Book Notices (ed. by Shriver, 1936) 10-11; Holmes, Collected Legal Papers (1921) 181, 184; Vegelahn v. Guntner, 167 Mass. 92, 104, 105, 106, 44 N.E. 1077, 35 L.R.A. 722, 57 Am.St.Rep. 443.
Cf. Laswell & McDougal, Legal Education and Public Policy, 52 Yale L.J. (1943) 203, 207, 217, 218, 219, 220, 225, 230.
Matton Oil Transfer Corp. v. The Dynamic, 2 Cir., 123 F.2d 999; United States v. Forness, 2 Cir., 125 F.2d 928, 942; Kulukundis Shipping Co. v. Amtorg Trading Co., 2 Cir., 126 F.2d 978, 980.
The trial judge refused Rudy’s request to charge the jury that there was “no testimony in the case that the defendant Rudy had any part in the purchase, sale, distribution, extraction or diversion of lead and opium-wash.” The trial judge gave as his reason for this refusal that the request was made too late under the Rules. In this he was mistaken. At one time there had been a Rule in the Southern District of New York that, when requests were not submitted before summing up, the judge, in his discretion, could refuse to consider them. But this Rule was abolished in 1938. The taking of the testimony was concluded on April 20, 1940 at about 4.00 P. M. Counsel for Iiss then delivered his summation to the jury which he concluded at about 5.15 P. M. Counsel for Rudy was then instructed by the court to deliver his summation and did so only after the court had denied his application to adjourn until the following morning. The next day, at 2.30 P. M., Rudy’s counsel submitted to the court his written request to charge while counsel for other defendants were still in the process of addressing the jury and when the prosecutor had not yet delivered his summation. The court commenced its charge to the jury at 3.40 P. M. At the conclusion of this charge, Rudy’s counsel asked the court to charge on his written request. In the circumstances, the refusal to charge as requested was, I think, an abuse of discretion.
McDonald v. Pless, 238 U.S. 264, 35 S.Ct. 783, 59 L.Ed. 1300; Mattox v. United States, 146 U.S. 140, 13 S.Ct. 50, 36 L.Ed. 917; 8 Wigmore, Evidence (3d ed. 1940) 668, 677, 282ff; cf. Fayerweather v. Ritch, 195 U.S. 276, 307, 25 S.Ct. 58, 49 L.Ed. 193.
See, e.g., Wilmington Star Mining Co. v. Fulton, 205 U.S. 60, 79, 27 S.Ct. 412, 51 L.Ed. 708; Schilling v. Delaware & N. R. Corp., 2 Cir., 114 F.2d 69, 72; Christian v. B. & M. R. R., 2 Cir., 109 F.2d 103, 105.
I have expressed such views more at length in Keller v. Brooklyn Bus Corporation, 2 Cir., 128 F.2d 510, 513.
United States v. Smith, 2 Cir., 112 F.2d 83, 85; Caringella v. United States, 7 Cir., 78 F.2d 563, 567; Davis v. United States, 5 Cir., 12 F.2d 253, 256.
National Labor Relations Board v. Air Associates, 2 Cir., 121 F.2d 586, 590; cf. Dunlop v. United States, 165 U.S. 486, 498, 17 S.Ct. 375, 41 L.Ed. 799; Kentucky Tax Cases (Cincinnati, New Orleans & Texas Pacific R. Co. v. Com. of Kentucky), 115 U.S. 321, 335, 6 S.Ct. 57, 29 L.Ed. 414.
The proposed new Federal Rules of Criminal Procedure provide in Rule 14: “If it appears -that a defendant or the government may be prejudiced by a joinder of offenses or of defendants in an indictment or information or in a trial,, whether by a multiplicity of counts or of defendants or otherwise, the court at any time upon motion of the defendant, of the government, or of its own motion may order an election or separate trials of counts, grant a severance of defendants, and provide whatever other relief justice may require.” (Italics added). ■
■ The Committee which drafted these Rules says that Rule 14 “states the present federal law.” As to the common-law rule, see, e.g., Drake v. Commonwealth, 214 Ky. 147, 282 S.W. 1066, 1067, 1068; Hoffman v. Commonwealth, 134 Ky. 726, 121 S.W. 690; Powell v. State, 224 Ala. 540, 141 So. 201, 205.
Cf. Thayer, A Preliminary Treatise on Evidence (1898) 274-275; Hutcheson, Separation of Powers and Administrative Law (Address, August 1936); Radin says “that all law is in a sense, and certainly in its inception, ‘administrative’ law.” Radin, Anglo-American Legal History (1936) 76; see Chapters 5 and 6 for historical discussion. Cf. Radin, in My Philosophy of Law (1941) 287, 298.
Catlin, The Story of The Political Philosophers (1937) 361; cf. Laswell and McDougal, supra, 208.
See, e.g., Wiborg v. United States, 163 U.S. 632, 658, 16 S.Ct. 1127, 1197, 41 L.Ed. 289; United States v. Atkinson, 297 U.S. 157, 160, 56 S.Ct. 391, 80 L.Ed. 555; Brasfield v. United States, 272 U.S. 448, 450, 47 S.Ct. 135, 71 L.Ed. 345; New York C. R. Co. v. Johnson, 279 U.S. 310, 318, 49 S.Ct. 300, 73 L.Ed. 706; Crawford v. United States, 212 U.S. 183, 194, 29 S.Ct. 260, 53 L.Ed. 465, 15 Ann.Cas. 392; Van Gorder v. United States, 8 Cir., 21 F.2d 939, 942; Meadows v. United States, 65 App.D.C. 275, 82 F.2d 881, 884.
See also Rule 10 of this court which provides that “the court, at its option, may notice a plain error not assigned”; United States v. Trypuc, 2 Cir., June 11, 1943, 136 F.2d 900.
In re Barnett, 2 Cir., 124 F.2d 1005, 1010, 1011.
Cf. my dissenting opinion in United States v. St. Pierre, 2 Cir., 132 F.2d 837, 849.
New York Central R. Co. v. Johnson, 279 U.S. 310, 318, 49 S.Ct. 300, 303, 73 L.Ed. 706.
Bruno v. United States, 308 U.S. 287, 294, 60 S.Ct. 198, 84 L.Ed. 257.
See Fort Dodge Hotel Co. v. Bartelt, 8 Cir., 119 F.2d 253, 259; Lynch v. Oregon Lumber Co., 9 Cir., 108 F.2d 283, 285, 286; Farris v. Interstate Circuit, 5 Cir., 116 F.2d 409, 412; Little v. United States, 10 Cir., 73 F.2d 861, 866, 96 A.L. R. 889; Worcester v. Pure Torpedo Co., 7 Cir., 127 F.23 945, 947, 948.
See McCandless v. United States, 298 U.S. 342, 347, 348, 56 S.Ct. 764, 766, 80 L.Ed. 1205; Bruno v. United States, supra. Cf. discussion in my dissent in Keller v. Brooklyn Bus Corp., 2 Cir., 128 F.2d 510, 513, 514, 515. The harmless error doctrine (if kept within proper limits) is highly desirable; cf. In Matter of Barnett, 2 Cir., 124 F.2d 1005, 1011.
Federal Rules of Civil Procedure, rule 61, 28 U.S.C.A. following section 723c, promulgated since the enactment of 28 U.S.C.A. § 391, embodies that doctrine. That statute applies, literally, only to the trial courts, because the Act authorizing the Rules did not authorize Rules affecting appellate jurisdiction (3 Moore, Federal Practice 3155), but the Rules have been applied by analogy on appeal; In Matter of Barnett, supra; N.L.R.B. v. Remington-Rand, 2 Cir., 130 F.2d 919, 925. It is hardly to be presumed that the Supreme Court intended the “harmless error” Rule to be given a broader interpretation than it has given to the “harmless error” statute. Cf. University City v. Home, etc., Ins. Co., 8 Cir., 114 F.2d 288, 295; see also the cases cited in note 16, four of which were decided since the Rules went into effect; the same is true of the Bruno case.
In any event, while 28 U.S.C.A. § 391 covers both civil and criminal cases, those civil Rules do not apply to criminal cases.