(dissenting):
In this capital case, the state trial judge, in an ex parte discussion with defense counsel, spontaneously volunteered that if petitioner pleaded guilty he would receive only a life sentence. It is clear that, at the time of the judge’s vital disclosure, no plea bargaining agreement had been reached between the prosecutor and defense counsel. It is equally clear that this information caused petitioner, who had theretofore steadfastly protested his innocence despite his court-appointed counsel’s recommendation that he admit guilt, to tender his plea. Petitioner testified that this was his impetus. His court-appointed counsel understood that this was petitioner’s motivation. Petititioner’s rejection of his attorney’s advice up to the point of knowing the judge’s views, the rapidity with which he receded from his prior assertion of his rights and his claim of innocence, and the absence of any other explanation for his change of position are additional corroboration.
The majority, nevertheless, holds that petitioner’s plea of guilty was voluntary —the free and untrammeled choice of his low mentality, uninfluenced and unmotivated by extraneous threat or promise. The majority reads Brady v. United States, 397 U.S. 742, 90 S.Ct. 1463, 25 L.Ed.2d 747 (1970), to sanction judicial *1358participation in plea bargaining before the accused has decided to plead guilty by permitting the judge to indicate a defendant’s ultimate fate so as to assist him in his decision. As a result of this reading, the majority neatly avoids the carefully considered contrary views of eminent jurists, prosecutors, defense counsel, and teachers of the law as expressed in the A.B.A. Project on Minimum Standards for Criminal Justice, Standards Relating to Pleas of Guilty (Approved Draft, 1968) § 3.3, while paying lip service to their desirability. I disagree.
-I-
However useful and socially desirable plea bargaining may be, “an overriding constitutional limitation is that the plea must not have been induced by promises or threats which deprive it of the character of a voluntary act.” Lassiter v. Turner, 423 F.2d 897, 900 (4 Cir. 1970). To a person in petitioner’s position, the statement of the trial judge may well have had either or both of two meanings: first, in the affirmative sense, it was a commitment that a certain desirable result would obtain if the guilty plea was forthcoming, and, second, in the negative sense, it was an implication that, if tried and found guilty, the likelihood of the imposition of capital punishment would be increased. Because the statements emanated from the very person that petitioner contemplated would preside at his trial, the legal principle, known to judges and lawyers and perhaps known to petitioner, that in Virginia the jury would sentence him if it returned a guilty verdict is of little significance. Petitioner could well suppose that if he refused the judge’s suggestion, the jury’s determination of sentence would not proceed without hostile judicial influence.
I cannot conclude that, in this case, petitioner’s plea was voluntary. Factually, it was involuntary. The record will only support the conclusion that petitioner tendered his plea because it was generated by the trial judge. It is, therefore, not necessary here to decide that due process always forbids judicial intervention in plea bargaining before tentative agreement between prosecutor and accused, but I find persuasive the views of Judge Weinfeld expressed in United States v. Gilligan, 256 F.Supp. 244, 254 (S.D.N.Y.1966):
The unequal positions of the judge and the accused, one with the power to commit to prison and the other deeply concerned to avoid prison, at once raise a question of fundamental fairness. When a judge becomes a participant in plea bargaining he brings to bear the full force and majesty of his office. His awesome power to impose a substantially longer or even maximum sentence in excess of that proposed is present whether referred to or not. A defendant needs no reminder that if he rejects the proposal, stands upon his right to trial and is convicted, he faces a significantly longer sentence. One facing a prison term, whether of longer or shorter duration, is easily influenced to accept what appears the more preferable choice. Intentionally or otherwise, and no matter how well motivated the judge may be, the accused is subjected to a subtle but powerful influence. A guilty plea predicated upon a judge’s promise of a definite sentence by its very nature does not qualify as a free and voluntary act. The plea is so interlaced with the promise that the one cannot be separated from the other; remove the promise and the basis for the plea falls, (footnotes eliminated, emphasis supplied.)
See also, Judge Weinfeld’s earlier expression in United States v. Tateo, 214 F. Supp. 560 (S.D.N.Y.1963). The better reasoned cases and judicial expressions are in full accord: Scott v. United States, 136 U.S.App.D.C. 377, 419 F.2d 264 (1969) (dicta); Brown v. Beto, 377 F.2d 950 (5 Cir. 1967) (dictum); United States v. LaVallee, 319 F.2d 308 (2 Cir. 1963) (Marshall, J., dissenting); Euziere v. United States, 249 F.2d 293, (10 Cir. 1957). See also United States v. Schmidt, 376 F.2d 751 (4 Cir. 1967); *1359United States v. Lias, 173 F.2d 685 (4 Cir. 1949).
The result I would reach in the instant case has the imprimatur of the A.B.A. Standards for Criminal Justice, supra. Section 3.3(a) of the Standards Relating to Pleas of Guilty flatly states that “[t]he trial judge should not participate in the plea discussions”. Other provisions sanction judicial consideration and approval of the agreement once a tentative bargain has been made. The standard and the reasons for its recommendation are set forth in the commentary:
The standard takes the position that judicial participation in plea discussions is undesirable. Compare Informal Opinion No. 779, ABA Professional Ethics Committee: “A judge should not be a party to advance arrangements for the determination of sentence, whether as a result of a guilty plea or a finding of guilty based on procf.” 51 A.B.A.J. 444 (1965).
There are a number of valid reasons for keeping the trial judge out of the plea discussions, including the following: (1) judicial participation in the discussions can create the impression in the mind of the defendant that he would not receive a fair trial were he to go to trial before this judge; (2) judicial participation in the discussions makes it difficult for the judge objectively to determine the voluntariness of the plea when it is offered; (3) judicial participation to the extent of promising a certain sentence is inconsistent with the theory behind the use of the presentence investigation report; and (4) the risk of not going along with the disposition apparently desired by the judge may seem so great to the defendant that he will be induced to plead guilty even if innocent. Comment, 32 U.Chi.L.Rev. 167, 180-83 (1964); Note, 112 U.Pa.L.Rev. 865, 891-92 (1964); Note, 55 Colum.L.Rev. 366, 371 (1955) * * *
The commentary also considers the argument, oft advanced and accepted by the majority, that judicial nonparticipation has the effect of requiring a defendant “to plead in the dark.” See Comment, 32 U.Chi.L.Rev. 167, 183 (1964). The short answer provided is that “the benefits which would be derived from this practice [involving the trial judge in plea discussion before tentative or conditional agreement] do not outweigh the risks which are inherent in judicial participation in plea discussions and plea agreements.”
The majority advances other reasons to reject a rule of judicial nonparticipation and they deserve brief consideration. To me, they are unpersuasive.
First, it is said that if a trial judge refuses ratification of a plea bargain made without his intervention, he necessarily becomes involved in the negotiation of subsequent revised plea bargains. Because this “indirect manner of judicial participation in the plea bargaining process” would be “an exercise in judicial circuitousness” the majority concludes that his direct initial participation is somehow justified. There are two answers to this thesis: First, the number of judicially rejected plea bargains is quite small, so that the problem rarely arises; and, second, if a plea bargain is rejected, it should be a simple matter for the judge to disqualify himself and send the parties to another judge for consideration of their revised agreement.
Second, it is said that, if judicial participation in plea negotiations is outlawed, a trial judge cannot be certain that a comment of his would not be considered to implicate him improperly, all to the end that the socially desirable process of plea bargaining may be effectively ended. I suggest that the judge can be certain that he has committed no impropriety if his comments and discussion about the ultimate disposition of a case are confined to occasions when both parties were present, in person or by counsel, and to the time after the parties have reached tentative agreement. I can hardly suppose that discussion and exploration of possible areas of agreement between prosecutor and defense counsel will dry up in the absence of judicial stimulation.
*1360Thus, I would decide that, unless Brady v. United States, supra, requires a different result, petitioner’s plea of guilty was not voluntary.
-II-
The majority concludes that the conduct of the trial judge “was not only not prohibited by Brady but was of a type in fact contemplated by the opinion.” Brady is pertinent here because it reaffirms the due process requirement that a plea of guilty be a free and voluntary act before an accused may be held to its consequences, but I do not read Brady as a license for what transpired here.
Brady turned on whether a guilty plea, under an indictment subjecting the accused to capital punishment for kidnapping if the jury so recommended, was voluntary. The death penalty provision of the statute had been held unconstitutional in United States v. Jackson, 390 U.S. 570, 88 S.Ct. 1209, 20 L.Ed.2d 138 (1968). The petitioner asserted that he pleaded guilty because of his fear of possible imposition of the unconstitutional death penalty, and for that reason his plea was not voluntary. His argument was rejected, and his plea held voluntary. In reaching this conclusion, the Court noted, as a factual matter, that Brady, advised by competent counsel, tendered his plea after his codefendant determined to plead guilty and because available to testify against him. The threat of the death penalty was thus only one factor entering into Brady’s determination to plead, and the Court declined to hold “that a guilty plea is compelled and invalid under the Fifth Amendment whenever motivated by the defendant’s desire to accept the certainty or probability of a lesser penalty rather than face a wider range of possibilities extending from acquittal to conviction and a higher penalty authorized by law for the crime charged.” 397 U.S. at 751, 90 S.Ct. at 1470. In conjunction with that principle, the Court construed its decision in Jackson, prohibiting the imposition of the death penalty, as one which “neither fashioned a new standard for judging the validity of guilty pleas nor mandated a new application of the [voluntariness] test * 397 U.S. at 747, 90 S.Ct. at 1468.
In its opinion, the Court employed the language — a quotation from Shelton v. United States, 246 F.2d 571, 572, n. 2 (5 Cir. 1957), reversed on confession of error, 356 U.S. 26, 78 S.Ct. 563, 2 L.Ed.2d 579 (1958) — on which the majority relies. Superficially, the language may be read to contemplate plea bargaining in which the judge has participated. But analysis of the antecedents of the statement, and consideration of other portions of the Court’s opinion, satisfy me that the majority’s reading is incorrect.
Shelton as an examination of the panel opinion in 242 F.2d 101 (5 Cir. 1957), which was rejected by the en banc court, will demonstrate, was not a plea bargaining case in which there was any claim that there had been judicial participation in the negotiations preceding the accused’s decision to plead guilty. The inducement for defendant to plead was a representation by the prosecutor as to the sentence defendant would receive, and the representation was fulfilled. The statement of the en banc court was, therefore, dictum taken from Judge Tuttle’s dissent from the panel’s decision, where he analyzed previous decisions in which, inter alia, a defendant was misled into pleading guilty by a promise of leniency from the court. Judge Tuttle relied on two authorities: United States v. Lias, 173 F.2d 685 (4 Cir. 1949), and United States v. Parrino, 212 F.2d 919 (2 Cir. 1954) (dictum).
In United States v. Lias, which was an attempted appeal by the United States from the granting of a motion to strike out guilty plea, we indicated that it was proper for the district judge to strike out the plea of an income tax offender, sentenced to five years on his guilty plea, who swore that he had been led to believe, as a result of what was said by the judge to his attorney, that he would receive probation if he would plead guilty. In this case the judge did not controvert the allegations that he used the language attributed to him. The case is thus clear*1361ly one of a judicial promise not kept and our approval of the district judge’s action in striking out the plea was no more extensive.
In United States v. Parrino, there was only dictum that misadvice by the court with regard to the collateral consequences of a plea of guilty, such as possible deportation or eligibility to enlist in the armed services, was not a ground to strike out the plea when the collateral consequences proved more onerous than the trial judge contemplated. The precise point decided was that innocent misadvice by counsel that a plea of guilty would not result in the accused’s being deported was not a ground to strike out his plea.
Thus, the quotation in Brady, itself not a case in which there was any claim of judicial participation in any bargain to plead guilty, is not founded upon authority to support what the majority reads the language to say. If this were not enough, the majority in Brady carefully reserved the very question which the majority decides, so that it can hardly be said that Brady intended tp decide it. In speaking of the voluntariness of Brady’s plea, the Court likened Brady to
(1) the defendant, in a jurisdiction where the judge and jury have the same range of sentencing power, who pleads guilty because his lawyer advises him that the judge will very probably be more lenient than the jury;
(2) the defendant, in a jurisdiction where the judge alone has sentencing power, who is advised by counsel that the judge is normally more lenient with defendants who plead guilty than with those who go to trial; (3) the defendant who is permitted by prosecutor and judge to plead guilty to a lesser offense included in the offense charged; and (4) the defendant who pleads guilty to certain counts with the understanding that other charges will be dropped.
379 U.S. at 751, 90 S.Ct. at 1470. Significantly thereafter, it added a footnote to the statement “[i]n each of these situations,” reading as follows:
We here make no reference to the situation where the prosecutor or judge, or both, deliberately employ their charging and sentencing powers to induce a particular defendant to tender a plea of guilty. In Brady’s case there is no claim that the prosecutor threatened prosecution on a charge not justified by the evidence or that the trial judge threatened Brady with a harsher sentence if convicted after trial in order to induce him to plead guilty.
397 U.S. at 751, n. 8, 90 S.Ct. at 1470.
As I have attempted to demonstrate in Part I, the effect of the judge’s statement in the case at bar was not only to promise petitioner leniency if he pleaded guilty, but to threaten him if he persisted in his plea of not guilty. This is the point that Brady does not decide and which, for the reasons I have stated, lead me to conclude that the judgment should be reversed with directions to issue the writ unless the state affords petitioner a new trial upon a plea of not guilty within a reasonable period.