Appellant John D. Tarrago was convicted in the United States District Court for the Southern District of New York before Richard H. Levet, J., and a jury of filing a false federal income tax return for 1956 and failing to file timely returns for 1957 and 1958. 26 U.S.C. §§ 7206(1), 7203. Appellant was sentenced to a prison term of one year and eight months for the first offense, and one year each on the other two; the three sentences were concurrent.1 The two issues on appeal relate to the test of criminal responsibility used by the jury and the trial judge's refusal to grant appellant a short continuance to obtain other counsel.2 Our disposition of the former makes it unnecessary to rule on the latter. For reasons set forth below, we reverse and remand.

The trial was held in October 1965, at a time when the district courts of this circuit "in the absence of appellate guidance on the subject," see United States v. Freeman, 357 F.2d 606, 608 (2d Cir. 1966), were applying the M'Naghten test of criminal responsibility. The evidence on this issue was conflicting, but it should be noted that appellant was twice committed to mental institutions and his expert medical witness offered a diagnosis of "schizophrenic reaction, paranoid type," a psychotic condition "in which a person has lost touch with reality." After having been charged under the M'Naghten rule, the jury returned a verdict of guilty.

A few months later, while Tarrago's conviction was on appeal, the opinion of this court in United States v. Freeman, supra, was announced. In that case, we adopted as the standard of criminal responsibility for the courts of this circuit the criteria provided by section 4.01 of the Model Penal Code, drafted by the American Law Institute:

(1) A person is not responsible for criminal conduct if at the time of such conduct as a result of mental disease or defect he lacks substantial capacity either to appreciate the wrongfulness of his conduct or to conform his conduct to the requirements of law.

(2) The terms "mental disease or defect" do not include an abnormality manifested only by repeated criminal or otherwise anti-social conduct.

The Freeman rule is substantially broader than the M'Naghten test; it focuses on a defendant's ability not only to appreciate the wrongfulness of his conduct but also to conform it to the requirements of law, and it recognizes that both capacities involve matters of degree.

Within a short time after the Freeman opinion, the question explicitly arose whether the test there adopted should be applied to a case tried before Freeman and on appeal when Freeman was decided. In United States v. Sheller, 369 F.2d 293 (2d Cir. 1966), a panel of this court decided that the Freeman rule should be given that "limited retroactivity." The same issue is again before us in this case. Because two members of the panel that originally heard argument of this appeal3 believed that the decision in Sheller was wrong, we convened the court in banc to consider the question again.

There has been much discussion of late of the broad question of "prospective limitation" of judicial decisions.4 However, the issue before us is only whether the Sheller court was correct in applying the Freeman rule to a case still on direct appeal when Freeman was decided. The question is a narrow one with slight impact; we have been informed by the Government of only two cases that might be affected, of which this is one.5

The Government argues that "To reverse a conviction because direct appeal was still pending but not on collateral review creates a capricious * * * discrimination."6 This seems to us to misstate and confuse the issue. We are not faced here with an attempt at collateral review; moreover, we have only the question of the retroactivity of a change in a rule of substantive, but not constitutional, law. Cf. Sunal v. Large, 332 U.S. 174, 67 S. Ct. 1588, 91 L. Ed. 1982 (1947). In relevant context, the real issue is whether the Freeman rule should be given what Professor Mishkin has called "normal retroactivity" or "normal judicial operation." Mishkin, supra note 4, at 77. Moreover, even if only constitutional law decisions are examined, the Supreme Court has made exactly the distinction the Government labels as "capricious"; i. e., it has denied full retroactivity for new constitutional doctrines, while applying them to cases still on appeal. Thus, in Linkletter v. Walker, 381 U.S. 618, 622 & nn. 4 & 5, 85 S. Ct. 1731, 14 L. Ed. 2d 601 (1965), the Court adopted such limited retroactivity for the rule announced in Mapp v. Ohio, 367 U.S. 643, 81 S. Ct. 1684, 6 L. Ed. 2d 1081 (1961), that evidence obtained through an unreasonable search and seizure was to be excluded from state criminal proceedings. In Linkletter, 381 U.S. at 627, 85 S. Ct. at 1736, the Court also stated that:

Under our cases it appears (1) that a change in law will be given effect while a case is on direct review * *.

In Tehan v. United States ex rel. Shott, 382 U.S. 406, 409 n. 3, 86 S. Ct. 459, 15 L. Ed. 2d 453 (1966), the Court similarly treated the rule announced in Griffin v. State of California, 380 U.S. 609, 85 S. Ct. 1229, 14 L. Ed. 2d 106 (1965), which forbade adverse comment by prosecutors and judges on the failure of a defendant to testify in a state criminal trial. However, relying on the later-decided opinions in Johnson v. State of New Jersey, 384 U.S. 719, 86 S. Ct. 1772, 16 L. Ed. 2d 882 (1966), and Stovall v. Denno, 388 U.S. 293, 87 S. Ct. 1967, 18 L. Ed. 2d 1199 (1967), the Government counters that the Supreme Court did not realize what it was doing in the earlier decisions discussed above. The Government claims that "if the Court had ever considered the question squarely"7 in those earlier decisions, it would have ruled the other way. However, a careful consideration of even the cases relied on by the Government and of the underlying considerations leads us to reaffirm Sheller.

In Stovall v. Denno, supra, the Supreme Court considered the retroactivity of the rules announced in United States v. Wade, 388 U.S. 218, 87 S. Ct. 1926, 18 L. Ed. 2d 1149 (1967), and Gilbert v. State of California, 388 U.S. 263, 87 S. Ct. 1951, 18 L. Ed. 2d 1178 (1967), which required "the exclusion of identification evidence which is tainted by exhibiting the accused to identifying witnesses before trial in the absence of his counsel." 388 U.S. at 294, 87 S.Ct. at 1968. The Court pointed out, id. at 296-297, 87 S.Ct. at 1969-1970, that the Linkletter, Tehan, and Johnson cases

"establish the principle that in criminal litigation concerning constitutional claims, `the Court may in the interest of justice make the rule prospective * * * where the exigencies of the situation require such an application' * * *." [Citation omitted.] The criteria guiding resolution of the question implicate (a) the purpose to be served by the new standards, (b) the extent of the reliance by law enforcement authorities on the old standards, and (c) the effect on the administration of justice of a retroactive application of the new standards.

Applying these criteria in Stovall, the Court held that the new rules were prospective only and refused to distinguish between "convictions now final * * * and convictions at various stages of trial and direct review." It regarded "the factors of reliance and burden on the administration of justice as entitled to such overriding significance as to make that distinction unsupportable." Id. at 300-301, 87 S.Ct. at 1972. The analysis in Johnson v. State of New Jersey, 384 U.S. 719, 726-733, 86 S. Ct. 1772 (1966), also relied on by the Government, was similar.

On the other hand, applying this reasoning to the question in this case leads to a completely opposite result. Extending the Freeman rule to cases still on appeal when it was announced will place no significant "burden on the administration of justice." In Stovall and Johnson, the Court was obviously concerned that great numbers of cases might be reopened even by limited retroactivity.8 In contrast, we know of only one other case that will be affected by our ruling here. Moreover, no hearings will have to be held on "the excludability of evidence long since destroyed, misplaced or deteriorated," compare Linkletter v. Walker, 381 U.S. at 637, 85 S. Ct. 1742, and witnesses available at the trial two and one-half years ago are presumably still available. There is no problem here of "a legalized mass jail break by rapists, murderers, and other felons." See Schwartz, supra note 4, at 745.

Similarly, no overpowering argument of reliance by "law enforcement authorities" can be made. Putting to one side whether the rejection of M'Naghten elsewhere9 precluded a justified "reliance" upon it in the absence of controlling precedent in this court, the rules on criminal responsibility — old or new — have nothing to do with police conduct at all. It is true that the United States Attorney's office expended time and effort in indicting and prosecuting Tarrago, as did the trial judge in presiding at the trial. Perhaps it could be said that in doing so both "relied" on prior cases. But that type of reliance is less compelling than reliance "on the old standards" by law enforcement authorities in obtaining evidence by following a course of conduct which suddenly becomes an alleged — and perhaps ineradicable — bar to a conviction. We know of no reason why there cannot be a speedy retrial here at which the Government can produce again the evidence it already has.

Finally, and most important, the purpose of the Freeman rule was to deal more humanely and intelligently with the substantive question whether "certain classes of wrongdoers are * * * properly the subjects of criminal punishment," 357 F.2d at 625. The issue before us on this direct appeal is whether appellant should suffer the penalties of the criminal law, even though we are not sure that he should; the doubt exists because the jury in this case applied "medically discarded concepts" to appellant's conduct rather than the more enlightened test of Freeman. Such doubt can easily be dispelled if appellant is retried under a test that accords with modern penal and psychiatric thought.

The Government points out that when two other circuit courts changed the definition of insanity, the new rule was made purely prospective, see Durham v. United States, 94 U.S.App.D.C. 228, 214 F.2d 862, 874, 45 A.L.R. 2d 1430 (1954); United States v. Shapiro, 383 F.2d 680, 687 (7th Cir. 1967) (in banc). However, in neither case was there a discussion of the retroactivity problem in terms of the criteria suggested by the Supreme Court to guide resolution of that issue; i. e., the purpose of the new standard, the extent of reliance on the old standard by law enforcement authorities, and the effect of retroactivity on the administration of justice.10 Therefore, we have no way of knowing, for example, what would have been the effect in those circuits of the "limited retroactivity" under consideration here. In any event, as we have pointed out above, we see good reasons for applying the Freeman test to cases not yet final and no compelling reason for not doing so.

Judgment reversed and case remanded for a new trial.

LUMBARD, Chief Judge, with whom MOORE, Circuit Judge, concurs (concurring in the result):

I reluctantly concur in the reversal of Tarrago's conviction.

I voted in favor of en banc consideration of this case because it seemed to me that it is important in all situations where, after submission to all active judges, we announce a new rule for the district courts to apply in determining issues raised in criminal cases that all of the active judges of the court should pass upon the question of the extent to which the new rule is to be applied to cases tried prior to the time the court announces the new rule. Unfortunately we did not give such consideration to the application of the new test for determining criminal responsibility which was announced in United States v. Freeman, 357 F.2d 606 (2d Cir. 1966).

In Freeman, the court adopted the test of criminal responsibility in § 4.01 of the American Law Institute's Model Penal Code. United States v. Malafronte, 357 F.2d 629 (2d Cir. 1966), was argued and decided concurrently with Freeman. The court held, "because the trial judge, in assessing [the criminal incompetence] defense, applied a standard which we have today declared in Freeman to be outmoded and hence insufficient, we reverse and remand for a new trial," 357 F.2d at 629, applying the Freeman rule retroactively to a case on direct appeal without explicitly considering whether this new rule should only be applied prospectively. Since both Freeman and Malafronte were circulated to all of the active judges of the court and neither my brethren nor I objected or desired to give these cases en banc consideration, I feel compelled to accept the logical consequences of the remand in Malafronte. If the issue of retrospective application of Freeman were being posed for the first time in this case, I would at least vote to limit Freeman to cases tried after the decision was announced. United States v. Shapiro, 383 F.2d 680 (7th Cir. 1967); Durham v. United States, 94 U.S.App.D.C. 228, 214 F.2d 862, 874, 45 A.L.R. 2d 1430 (1954). See United States v. Youngblood, 379 F.2d 365, 370 (2d Cir. 1967).

The panel in United States v. Sheller, 369 F.2d 293 (2d Cir. 1966) announced explicitly what was implicit in Malafronte: that Freeman has at least limited retroactive effect. But in neither case was the issue of possible prospective application of the Freeman rule fully briefed for consideration by the entire court. There was no argument against retroactivity in Malafronte and the government's argument against retroactivity to the panel in Sheller, citing Johnson v. State of New Jersey, 384 U.S. 719, 731-733, was, as the panel itself said, merely pro forma. See 369 F.2d at 295 n. 3.

I am sure that had our attention been directed to the retroactivity issue in Freeman and Malafronte, or had the United States Attorney sought en banc consideration of it, we might well have decided the matter differently from its decision by a panel of the court in Sheller.

As in Linkletter v. Walker, 381 U.S. 618, 622 and n. 4, 88 S. Ct. 1731 (1965) and Tehan v. United States ex rel. Shott, 382 U.S. 406, 409 n. 3, 86 S. Ct. 459, 15 L. Ed. 2d 453 (1966), the cases upon which the Sheller panel relied, the court had already reversed a conviction on direct appeal before the retroactivity issue was even raised and it was compelled to decide between full retroactivity and limited retroactivity rather than between full retroactivity and prospective application. See Mishkin, The Supreme Court 1964 Term — Foreword: The High Court, The Great Writ, and the Due Process of Time and Law, 79 Harv.L.Rev. 56, 72-73, 76, 100 (1965); Schaefer, The Control of "Sunbursts": Techniques of Prospective Overruling, 42 N.Y.U.L.Rev. 631, 644-646 (1967).

We have been informed by the United States Attorney that of the four defendants in this circuit, other than Freeman, who might be affected by Freeman because they had been tried before Freeman was announced on February 28, 1966, but the proceedings on appeal had not yet been concluded, two defendants have already received the benefit of Freeman by reversal of their convictions, namely, Malafronte and Sheller. Even though I believe this action to have been erroneous, I reluctantly conclude that we cannot treat Tarrago and Caballaro differently.

But inasmuch as it seems likely, in these times of new concepts of justice and newly discovered constitutional rights, that we may announce new rules or doctrines for application in criminal cases, I take this occasion briefly to set forth the reasons why it seems to me to be of the utmost importance to the orderly, efficient and fair administration of the criminal laws that such new rules should be prospective only and that they should be clearly and emphatically stated to be such at the time that they are announced. Indeed, I would go further and advocate that in situations such as faced us in Freeman, it would have accomplished our purpose to announce that the rule would be applied to all cases tried after the date the rule was announced and not even apply it to Freeman.

While the authorities were widely divided upon what circumstances justify a court's refusal to apply newly declared principles to previous adjudication, those who have critically examined the decisions are unanimous in finding that making a distinction between cases on direct appeal and those that have gone to final judgment is arbitrary, as it subjects appellants to disposition of their appeals on the basis of fortuities unrelated to the merits of the points they would urge upon the court and seems to penalize the more diligent who have processed their appeals without delay.1

Since the Sheller decision, the Supreme Court itself has strongly condemned the finality distinction which it had used in Linkletter and Shott. DeStefano v. Woods, 392 U.S. 631, 88 S. Ct. 2093, 20 L. Ed. 2d 1308 (June 17, 1968); Stovall v. Denno, 388 U.S. 293, 296-301, 87 S. Ct. 1967, 18 L. Ed. 2d 1199 (1967). As Judge Schaefer put it, "But when a court is itself changing the law by an overruling decision, its determination of prospectivity or retroactivity should not depend upon the stage in the judicial process that a particular case has reached when the change is made. Too many irrelevant considerations, including the common cold, bear upon the rate of progress of a case through the judicial system." 42 N.Y.U.L.Rev. at 645. Bender, The Retroactive Effect of an Overruling Constitutional Decision: Mapp v. Ohio, 110 U.Pa.L.Rev. 650, 678 (1962); Mishkin, 79 Harv.L.Rev. at 66 n. 37, 74; Schwartz, Retroactivity, Reliability and Due Process: A Reply to Professor Mishkin, 33 U.Chi.L.Rev. 719, 731, 733-734, 764 (1966). I do not see how we can justify this arbitrary dividing line between those who are permitted to raise Freeman retroactively and those who are not. Cf. Witherspoon v. State of Illinois, 391 U.S. 510, 88 S. Ct. 1770, 20 L. Ed. 776. (June 3, 1968).

I would add another consideration which militates against drawing a distinction based upon whether or not appellant has completed his direct appeal: the court should not give appellants an added incentive to delay prosecution of their appeals by any means and thus postpone final determination of their appeals. It would be ironic if those who desire to protect final judgments from collateral attack make the incentives to postpone finality so great that appellants adopt a strategy that substantially delays operation of the principles of finality they wish to preserve. See, Mishkin, 79 Harv. L.Rev. at 77 n. 71, 80 n. 80; Schwartz, 33 U.Chi.L.Rev. 733, 742-743. I see nothing to be gained by applying an arbitrary cutoff point which rewards delay and penalizes those who promptly perfect their appeals in accordance with the rules of the court. In my view, this is seriously disruptive of the administration of justice and should not be permitted when it may fairly be avoided.

However, to reiterate, there are only five cases in the circuit in which limited retroactivity of the Freeman rule requires a new trial and three of those cases have already been reversed and remanded. Fairness dictates that the remaining two appellants be accorded similar treatment.

In the future I would hope that any decision which announces a new rule would not be applied retroactively in other appeals without thorough consideration of all arguments for and against retroactive application. See Griffin v. People of the State of Illinois, 351 U.S. 12, 25-26, 76 S. Ct. 585, 100 L. Ed. 891 (1956) (Mr. Justice Frankfurter, concurring); Mishkin, 79 Harv.L.Rev. at 64; Note, Prospective Overruling and Retroactive Application in Federal Courts, 71 Yale L.J. 907, 937 (1962). Cf. Chicot County Drainage District v. Baxter State Bank, 308 U.S. 371, 374, 60 S. Ct. 84, 84 L. Ed. 449 (1940). I would hope that any limitations upon retroactive effect of future decisions will not make arbitrary distinctions based upon the stage of the appellate process at which the case happens to be at the time of the decision announcing the new rule. I would also hope that any panel of this court which entertains for serious consideration a proposal to change any rule regarding criminal cases generally would ask the parties to brief and argue the question of when the proposed new rule should become effective and the nature and number of cases which would be affected by such proposed new rule. And lastly, all the active judges of the court should be alerted regarding such questions before a panel of the court files and publishes its opinion.