An indictment charging a criminal offense is subject to dismissal if it is not filed within the period of time established by the statute of limitations for that offense. Rule 6(e) of the Federal Rules of Criminal Procedure permits a district court to direct that a timely-filed indictment may be sealed until the defendant is in custody. A sealed indictment is timely even though the defendant is not apprehended and the indictment is not made public until after the end of the statutory limitations period. United States v. Michael, 180 F.2d 55, 56-57 (3d Cir. 1949), cert. denied, 339 U.S. 978, 70 S. Ct. 1023, 94 L. Ed. 1383 (1950); United States v. Niarchos, 125 F. Supp. 214, 231-32 (D.D.C. 1954). When this occurs, a defendant can be expected to claim that he has suffered prejudice because of the long delay between the date of the offense charged and the unsealing of the indictment. The issue in this case is to determine the relevant period of time in which such a claim of prejudice is to be assessed.

The panel that first considered this appeal held, by a divided vote, that an indictment timely filed under seal and not unsealed until after the limitations period must be dismissed whenever the defendant can show "substantial actual prejudice" occurring any time during the entire period between the date of the crime and the unsealing of the indictment. United States v. Watson, 599 F.2d 1149, 1155 (2d Cir. 1979), modified, Nos. 78-1296-8 (2d Cir. Oct. 9, 1979). Believing that prejudice had occurred because during that period the memory of appellant Muse had faded, the panel reversed his conviction. We granted a rehearing en banc because the issue presented, though arising infrequently, is important in the administration of criminal justice. We now hold that in determining prejudice to the defendant, the relevant time period is no longer than the time between the sealing of the indictment and its unsealing. Since in this case there is no basis for claiming that prejudice occurred during any of the time the indictment was sealed, we reinstate the judgment of conviction and leave for another day the issue of whether the relevant period for assessing prejudice should be only the post-limitations portion of the period during which an indictment is sealed.

The indictment in this case charged that Aaron Watson, John Muse, and two others, whose names were unknown, had participated in a narcotics conspiracy during the months of March to November, 1971. This indictment was filed and sealed on June 1, 1976, some five months before the end of the five-year limitations period. See 18 U.S.C. § 3282 (1976). Although the Government was aware of Muse's whereabouts during much of the post-indictment period, it did not arrest him for fear that his arrest would prompt his co-conspirators to flee. On September 29, 1977, Watson was finally located and arrested, and Muse was arrested the next day. A third member of the conspiracy, Robert Whitley, was arrested some two months later, while the fourth has yet to be found.

Muse, Watson, and Whitley were tried together in the United States District Court for the Eastern District of New York before Judge Charles P. Sifton and a jury. The jury returned verdicts of guilty against all three defendants, and they appealed, claiming, among other things, that the delay in unsealing the indictment denied them the protection of the statute of limitations.1 A panel of this Court rejected Watson's and Whitley's claims, concluding that these two defendants were not prejudiced by the delay, and that the delay was justified by the Government's need to locate them. 599 F.2d 1149 (2d Cir. 1979). But with respect to the third defendant, Muse, the panel found that he had been significantly prejudiced by the length of time that had passed since the commission of the crime, as he could not remember relevant dates when testifying on his own behalf. Given this prejudice to the defendant, the panel concluded, "the Government must show that the delay is justified by a strong prosecutorial interest, not simply by a legitimate interest." Id. at 1154. Judge Friendly, in dissent, argued that no significant prejudice had been shown, and that the Government had established as strong a prosecutorial interest as was needed to justify the delay. Id. at 1158-60.

The Government then petitioned the panel for rehearing, with a suggestion for rehearing en banc, as to Muse. Although the rehearing petition was initially denied, a majority of the panel subsequently vacated its order denying the petition and substituted a revised opinion. Nos. 78-1296-8 (2d Cir. Oct. 9, 1979). This opinion abandoned the distinction between "legitimate" and "strong" prosecutorial interests, and concluded instead that "when the defendant can show substantial actual prejudice, the indictment must be dismissed, for even a legitimate prosecutorial interest is then insufficient to effectuate statute of limitations policies." Id. at 5024. The revised opinion adhered to the view that the entire period from the offense to the unsealing of the indictment was relevant in assessing prejudice. Judge Friendly filed an additional dissent, arguing that the revised standard was even more harmful to legitimate prosecutorial interests and that there was no reason why delay prior to the sealing of the indictment should be considered in determining prejudice. Id. at 5029. The Government renewed its suggestion for rehearing en banc as to Muse; rehearing was ordered December 10, 1979, and the matter was submitted after receipt of additional briefs.

This case can be decided with adequate recognition of the legitimate interests of both the prosecution and the defendant. The Government's interests were recognized in the Michael decision, which first upheld the Government's power to maintain an indictment under seal beyond the limitations period. The Michael decision was based on Fed.R.Crim.P. 6(e)(4), which now states: "The federal magistrate to whom an indictment is returned may direct that the indictment be kept secret until the defendant is in custody or has been released pending trial." The obvious purpose of this provision is to prevent the requirement of an indictment from serving as a public notice that would enable the defendant to avoid arrest. See United States v. Lyles, 593 F.2d 182, 196 n.17 (2d Cir.), cert. denied, 440 U.S. 972, 99 S. Ct. 1537, 59 L. Ed. 2d 789 (1979). United States v. Sherwood, 38 F.R.D. 14 (D. Conn. 1964). Extrapolating from Rule 6, Michael held that a timely-filed, sealed indictment satisfies the five-year statute of limitations established by 18 U.S.C. § 3282 even though it is unsealed after the limitations period. This decision was designed to prevent the defendant from securing an unwarranted benefit from the statute of limitations by continuing to avoid arrest until the limitations period had run. The power granted to the Government in Michael, like that granted by Rule 6(e)(4), is based on the legitimate prosecutorial needs of the Government to capture those properly indicted for criminal activity.

The defendant's interest in avoiding a stale prosecution is safeguarded primarily by the federal statute of limitations, 18 U.S.C. § 3282, which states: "Except as otherwise expressly provided by law, no person shall be prosecuted, tried, or punished for any offense, not capital, unless the indictment is found or the information is instituted, within five years next after such offense shall have been committed." The purpose of this provision is to prevent the Government from instituting prosecutions after excessively long delays, and thus prejudicing the defendant by increasing his difficulty in marshaling evidence on his behalf. Toussie v. United States, 397 U.S. 112, 114-15, 90 S. Ct. 858, 859-60, 25 L. Ed. 2d 156 (1970); United States v. Laut, 17 F.R.D. 31, 36 (S.D.N.Y. 1955). "Such a time limit may also have the salutary effect of encouraging law enforcement officials promptly to investigate suspected criminal activity." Toussie v. United States, supra, 397 U.S. at 115, 90 S. Ct. at 860.

Section 3282 establishes a clear outer time limit beyond which indictments are invalid, but it does not provide guidance for determining how much protection should be afforded a defendant when an indictment is returned within the limitations period but maintained under seal until that period has run in order to serve the prosecutorial interest in arresting an indicted defendant. We think the defendant's interests are adequately protected by permitting him to secure dismissal only when he can show prejudice occurring during the period when the indictment was sealed, or perhaps only during the post-limitations portion of that period.2 ] Only prejudice occurring during either of these periods could be attributable to the prosecutor's decision to seal the indictment; any prejudice occurring prior to the sealing of the indictment is unrelated to this decision, and thus irrelevant to the issue in this case. There is a general presumption against regarding elapsed time during the limitations period as prejudicial to the defendant. See United States v. Marion, 404 U.S. 307, 322-23, 92 S. Ct. 455, 464-65, 30 L. Ed. 2d 468 (1971). The policy of § 3282 does not require, or even suggest, that this presumption is to be reversed when the Government justifiably delays the unsealing of a timely-filed indictment beyond the limitations period. Under the approach taken by the panel's opinion, the death of a defense witness one week after the offense would be grounds for dismissal of an indictment filed under seal years later just prior to the end of the limitations period and unsealed just after the end of that period, even though the death of the same witness on the same date would not justify dismissal of an indictment unsealed just prior to the end of the limitations period. As Judge Friendly observed, "There must be something wrong with a view that could yield such a result." United States v. Watson, supra, slip op. at 5030.

In this case, the Government initially sealed the indictment, and kept it sealed past the limitations period, because it feared that the indictment and arrest of Muse would cause his co-defendants to flee. Muse himself could have been captured at any time during the period in question and, had he been the sole defendant, the Government might be hard put to justify its delay. But his confederates were indicted defendants as well, and the need to capture them is as justifiable a basis for delay as the need to capture Muse himself. Judge Sifton found that the Government acted in good faith, with a reasonable basis for estimating the effect of Muse's arrest on the other defendants.

Muse's claim of prejudice is that, by the time of the trial, he had forgotten many of the relevant events, including the dates of various meetings with his co-conspirators. While memory loss may be a valid basis for a finding of prejudice, it is more difficult to establish that the passage of a particular time period created this type of prejudice than the prejudice resulting from the more usual grounds such as the death or disappearance of a witness or the loss of physical evidence. In any event, Muse's memory loss would be relevant to a statute of limitations defense only if it occurred during the sixteen months in which the indictment was sealed, or perhaps only during the ten months after the limitations period expired. Judge Sifton, having observed Muse testify, found that Muse had not been prejudiced by any memory loss, and, a fortiori, found no prejudice occurring during the ten-or sixteen-month period at issue here. We see no basis for disregarding this conclusion. Judge Oakes, writing for the panel majority, and Judge Friendly, in dissent, engaged in a well-reasoned and scholarly debate about the degree of incremental memory loss over time. We do not suggest that the law cannot benefit from the teachings of psychology, or that they would not be relevant to a fact-finder's consideration of this issue. But we are not prepared to hold that either position in this psychological debate is incorrect as a matter of law. Consequently we accept the findings of Judge Sifton that no significant memory loss occurred at least during the sixteen-month interval that we conclude is the outer limit of the relevant period.

Since the Government's prosecutorial interest in keeping the indictment sealed past the limitations period was substantial, and there was no resulting prejudice to the defendant, we conclude that Muse has no valid statute of limitations defense, and his conviction is therefore affirmed.

Judge Newman's lucid opinion for the majority en banc points up a flaw in the original panel opinion as modified, which is that

(u)nder the approach taken by the panel's opinion, the death of a defense witness one week after the offense would be grounds for dismissal of an indictment filed under seal years later just prior to the end of the limitations period and unsealed just after the end of that period, even though the death of the same witness on the same date would not justify dismissal of an indictment unsealed just prior to the end of the limitations period.

Majority op. at 1044. Our initial response to this point, when it was made by Judge Friendly in dissent, was that this result follows from the strict policy against prejudicial delay embodied in the statute of limitations. We noted that it is no more anomalous than the dismissal, without inquiry into actual prejudice, of an indictment filed one day after the statute of limitations has expired.

In hindsight, it appears that Judge Smith's and my response was insufficient to persuade my esteemed colleagues, and therefore must be deemed inadequate. Our answer should have been that when a sealed indictment tolls the statute of limitations, and is not unsealed until after the statutory period has passed, it is necessary to dismiss the indictment if a defendant proves actual and substantial prejudice attributable to rather than incidental to the delay beyond the limitations period. Thus, the death of a witness before the period expired would not constitute such prejudice, but the death of a witness afterwards could justify dismissing the indictment. Prejudicial loss of memory by a defendant or a witness poses a different problem because, as discussed later in this dissent, the loss of memory through time cannot be precisely charted by the science of psychology. It may be impossible to pinpoint whether a loss of memory that has prejudiced a defendant occurred before or after the statute has run. Therefore, in cases in which prejudice from loss of memory is shown, the policy underlying statutes of limitations, outlined in section V, requires the court to presume that the prejudice occurred as a result of the delay. The overriding purpose of the statute is to protect a defendant from long delay which adversely affects his ability to defend himself.

This notion, that the policy underlying statutes of limitations itself requires the court to presume that prejudice due to loss of memory resulted from delay beyond the limitations period, as applied to the instant case concededly involves several contested facts as well as legal principles. I will take up those controverted facts and principles in the following order:

1. Was there loss of memory?

2. Was it prejudicial?

3. Are we, as the majority opinion considers us to be, bound by the trial judge's findings on the first two points?

4. If not, what do psychological studies contribute to judicial resolution of the problem raised?

5. In light of the policies of statutes of limitations and the state of the art in the psychology of memory, what should that resolution be?

I. Was there loss of memory?

The testimony at trial demonstrated that both of the Government's chief witnesses as well as the defendant Muse suffered losses of memory. To prove the existence of a conspiracy during the limited period from March to November of 1971, the prosecution relied almost exclusively on the testimony of "Dickie" Diamond and MacArthur "Skip" Egister, both of whom were former drug dealers. Diamond had been convicted of narcotics violations, grand larceny, and murder, and he testified to other crimes such as pimping and fencing stolen property; and Egister was at least at one time one of Diamond's "lieutenants." After beginning to cooperate with the Government, both of them received payments under the witness protection program, Diamond $24,000 and Egister more than $20,000. At the 1978 trial, both of these witnesses were unsure of specific details and confused about the dates on which crucial events occurred.

Diamond implicated Muse by testifying about an argument involving himself, Egister, and Muse over some "synthetic" (very low quality) heroin which Diamond had sold to Egister. Diamond said that this incident took place following his release on bail in May 1971 in connection with the murder for which he was later convicted. But throughout his cross-examination he demonstrated an inability to remember specific events and dates. For example, when questioned about certain drug transactions in the early 1970's, Diamond professed a near total inability to recall events.

Q. ... How many keys (kilos of heroin) did you buy from Matthews during 1970?

A. I can't say, you know.

Q. Was it more than three kilos?

A. I can't say.

Q. Four kilos?

A. I have to go to the transcript to refresh my recollection. Whatever the transcript say, it say. You're going back years, seven, eight years on me.

Q. What about '71? Did you do business with Matthews in '71 or Beckwith?

A. I can't remember that year.

Q. All right. Now, was this (relation with a Cuban supplier) during the-what years was this?

A. I would say that's around '69 or '70.

Q. Did you ever do business out of the Lincoln Lounge?

A. What street is it? Sounds familiar to me, the name of the Lincoln Lounge.

Q. I don't know. You don't remember that place?

A. I remember the name. But it's been so long, you know. You're out there so many places. You be in and out of there over a period of time. I can't say where it was located. But the name Lincoln Lounge do ring a bell.

Diamond also demonstrated severe loss of memory with respect to other events in his life during the crucial year of 1971.

Q. Did you travel in 1971 with various groups (of musicians)?

A. 1971?

Q. Yes.

A. 1971 it seems I did some traveling in 1971, I couldn't say for sure.

Q. Mr. Diamond, during what year did this 30-night tour (with the music group "The Moments") take place?

A. I don't know if it was '70 or '71. I think it was '71.

Q. What part of '71?

A. I don't know.

Q. But you testified at the Frank Matthews trial, didn't you?

A. Yes.

Q. And the Frank Matthews trial was 1971?

A. Do the records say '71?

Q. Yes.

A. Yes.

It appeared from his testimony that Diamond dated all past events in a very rough manner. When asked whether he was guessing in which years certain events occurred, he responded, "No, based on the fact we can relate with cars, I can relate what type of car was that year, because you are asking me about things that happened many moons ago, so I am trying to relate myself to the cars I had into my moving pattern."

At one point, Diamond even testified that he had never transacted any business during 1971 with Muse or the other defendants.

Q. When you testified before the Grand Jury against the three defendants in this case, you were asked various questions, weren't you?

A. They asked me various questions.

Q. Do you recall if they only asked you questions referring to the dates in 1971?A. That's all they could ask me because I never done business with them in '71.

Q. Let me rephrase it. You testified in '76 about this case, didn't you, before the Grand Jury?

A. I don't know. Was it '76? I don't know.

(Emphasis added.)

The other principal Government witness, "Skip" Egister, who was serving time for first degree robbery when he first became a Government witness, also had a great deal of difficulty recalling dates. He testified to a 1971 incident in which he had attempted to obtain heroin from Diamond's supplier after Diamond had been arrested. The supplier sent him to Muse's fellow defendant Watson who, through the third defendant Whitley, gave Egister a package of heroin. Egister said that he had delivered the package to Muse's apartment in Brooklyn on Lafayette near Gates. In fact, Lafayette and Gates are parallel streets and Muse's apartment was on Lexington. Even in his direct testimony, Egister was unable to remember the precise date of this episode, testifying alternatively that it had occurred in March 1971, April 1971, May 1971, some time late in 1971, at the time of the Ali-Frazier fight (which was on March 8, 1971), at the time Diamond came out of jail (which was on May 16, 1971), or right after Diamond entered jail (which was on February 1, 1971) in "March or something like that."

Egister also testified that he had worked out an arrangement with Muse, to periodically deliver heroin to Muse from a connection (Watson), "at the time Dickie had come out (May 16, 1971). It was right after the time he went into jail (February 1, 1971) and it was March or something like that." In response to questioning about the "synthetic" heroin disagreement, which arose during the course of this arrangement, Egister could not recall a specific date on which he had received the poor quality heroin, but he said at different points in his testimony that it was after Diamond's release from jail (which would be after May 16, 1971); in June or July 1971; in August, September or October 1971; or in "late '71," "right before the winter started." He also remembered another conversation in the "latter part of 1971" in which he had complained to Watson about low quality heroin. Like Diamond, Egister was also unable to pinpoint an exact date for any transaction in the drug delivery arrangement.

Egister claimed on direct that he had worked for Diamond for three or three and one-half years, beginning in early 1969, but on cross he said that he had quit working for Diamond during the first two weeks of 1971, which was at most two years from the date on which he said their association had begun. And when confronted with his testimony at an earlier trial that he had never dealt in drugs with anyone but Diamond, he remarked that he had not been "working" at all in 1971.

Q. Were you asked the question: "Are you positive you never worked for anyone else? Is that correct?" And did you answer: "Yes."

A. Yes. Talking about that time, '68, '69, '70.

Q. Yesterday-Mr. Weisswasser asked you yesterday and you responded you were referring to '69 only; is that correct?

A. '69 and '70.

Q. What about '71?

A. '71-I wasn't working during '71.

If Egister "wasn't working during '71" his direct testimony was almost entirely false.

It is not surprising to me to also find a loss of memory on the part of the defendant Muse himself. Taking the stand, he testified that he had done business with Diamond and Egister up until late 1970 but that he had stopped dealing with them well before June 1, 1971, which is the cut-off date under the five-year statute of limitations (in light of the June 1, 1976, filing of the sealed indictment). According to Muse, he had stopped dealing with Egister and Diamond shortly before the time of Diamond's arrest, following an "argument" about "some bad merchandise," the synthetic heroin referred to by Diamond and Egister in their testimony. Muse said that the argument had taken place among the three of them at the Sugar Tree Lounge in "(l)ate 1970, almost around Christmas time," when it was "(w)inter" and "(v)ery cold".

Cross examination of defendant Muse focused on his memory for past events such as his change of addresses on Lexington Avenue.

Q. You testified you were living where?

A. 729 Lexington Avenue.

Q. And for how long a period of time did you live there?

A. I lived there, 729 Lexington, until I was about-I got married in '62. '62. Stayed there until about '69.

Q. Until when?

A. '69.

Q. And then where did you move?

A. 733.

Q. You testified on direct I believe that you moved in 1971. Were you in error when you said that?

A. '71.

Q. That's what you said on direct. Were you in error?

A. Probably was. I don't exact date (sic)....

The prosecution also attacked Muse's ability to remember dates and past events by questioning him about the cars that he had owned during different years in the early 1970's.

Q. What kind of car did you have in 1970?

A. 1970? I do not think I had a car in 1970.

Q. How did you use to get around?

A. My friends had cars. I do not think I had any. I might have had one.

Q. You don't remember?

A. I don't remember that.

Similarly, the Government questioned him about the date of his arrest on narcotics charges involving acts subsequent to the ones for which he was convicted in this case.

Q. When were you arrested on the Federal narcotics charge?

A. I think it was in late '72.

Q. Could it have been '73?

A. Might have been. Because I got out on bail. I don't know exact dates.

So far, I hope, I have demonstrated that there were losses of memory on the part of the two chief prosecution witnesses and the defendant himself. This is not surprising. That such memory losses can and do occur is recognized in the policy underlying statutes of limitations and in the case law.

II. Was the loss of memory prejudicial?

The critical factual question in this case was whether there were narcotics transactions among the parties during the period specified in the indictment. Each of the two principal prosecution witnesses was highly motivated to make a case out against the defendants because each of them had been released from prison, and had received payments from the witness protection program, after he had begun cooperating with the Government. Diamond began cooperating in 1972, Egister in 1974 or 1975. Egister by his own testimony had an enormous heroin habit during the relevant time periods in this case, snorting heroin 20 to 25 times a day at a daily cost of $350 to $400. I say enormous because Diamond testified that "(an) elephant couldn't snort 25 times a day." At trial both witnesses were vague and often confused as to dates, unspecific on details, and indeed, each one testified inadvertently that he was not "working" or had not "done business" in narcotics in 1971.

The precise date of the synthetic heroin delivery, and the argument about it at the Sugar Tree Lounge, were crucial to the case. Muse said that the meeting was in late 1970, when it was very cold, before Diamond was jailed on the murder charge. Diamond and Egister said that it was after Diamond was released on bail in May 1971. The fact that Muse could not on cross-examination remember whether he owned a car in 1970, when he was arrested in 1972 or 1973, or when he moved from one apartment on Lexington Avenue to another, surely must have cut against him in the eyes of the jury.

From section I above it should be obvious that the lapse of time from the original events to the time of the trial made it difficult to conduct a defense. As the original panel in this case noted, the fact that the witnesses were unsure of specifics, confused, and vague, does not mean that the evidence was insufficient to take the case to the jury. It does mean that there were losses of memory on the parts of the prosecution witnesses that made it difficult to cross-examine them and tie them down to specific times, dates and places. And it does mean that the defendant's own vagueness or inability to recollect could be capitalized upon.

To a certain extent, it is precisely because of the lapse of time that it is difficult for Muse, as it was for the defendant in Ross v. United States, 349 F.2d 210 (D.C. Cir. 1965), to show that his defense would have been more adequate had he learned of the charges against him sooner, i. e., to show prejudice. In the Ross case the court noted:

The trial court further concluded that there was no "indication that the defendant would have been able to present a more adequate defense if he had been arrested sooner than he was." But it was the very nature of appellant's disability that tended to make such a showing impossible. His failure of memory and his inability to reconstruct what he did not remember virtually precluded his showing in what respects his defense might have been more successful if the delay had been shorter .... In a very real sense, the extent to which he was prejudiced by the Government's delay is evidenced by the difficulty he encountered in establishing with particularity the elements of that prejudice.

Id. at 215. Similarly, in Muse's case, not only were the losses of memory prejudicial, but the length of time between the alleged acts and the trial made it difficult to demonstrate the elements of that prejudice.

III. Are we bound by the trial judge's findings as to loss of memory and resulting prejudice?

The trial judge here made his findings as to loss of memory and prejudice not at the time of the trial but at sentencing, some months later. It is on these findings that the majority en banc opinion rests, although Judge Friendly in his dissent from the original panel opinion must have felt them to be at least to a degree insufficient because he relied to some extent on his judicial notice of the psychology of memory and "common sense" about memory, of which more later. The lower court's findings on loss of memory, in haec verba and in toto, were as follows:

Now, analyzing the extent to which an offense-the defense may have been impaired in this case has to take into account, as I foresaw in deciding that this issue could be most appropriately determined after the trial, what kind of a case the defendants had. It impresses me and is determinative in this case of my disposition of the motion which is to deny it, that the case which the defendants had was a case involving large scale events occurring over a substantial period of time of a consistent pattern of testifying to by individuals who themselves suffered from difficulties of memory and who were able to pinpoint harms and occurrences, not in terms of minor private encounters, but rather in terms of gross public events, such as the Muhammad Ali fight or easily their (sic) verifiable events listed in the public records such as an arrest and releases on bail.

So that the factor of memory lapse which seems to me the important factor here-this is not a situation where a charge involving records which have been lost, but rather a case involving the possibility of memory lapse. It just seems to me that one either does or does not remember occurrences of this duration, of this extent for the rest of one's life. It seems to me unlikely that events of this sort-a charge of this sort would be forgotten.

With all due respect for the extremely able trial judge, I deem these findings insufficient. In the first paragraph he finds that Diamond and Egister "suffered from difficulties of memory," but he finds that "in terms of gross public events" and "verifiable events" (the Ali fights, Diamond's arrest and his release on bail), they could "pinpoint harms and occurrences." There were references in their testimony to these fights and to Diamond's arrest and release on bail. But there were, from Muse's point of view, two crucial and related factual questions at trial: first, whether he began dealing with Egister alone after Diamond's arrest, and second, whether the argument that concededly took place at the Sugar Tree Lounge occurred before Diamond's arrest in February 1971 or after his release on bail in May 1971. Resolution of the second issue by the jury in Muse's favor could very well have led to resolution of the first in his favor. In any event, Muse said that the argument was before Diamond's arrest, while Diamond and Egister said that it was afterwards, although they testified with much confusion, backtracking, and even contradictory statements. I do not see how this can be called "pinpointing"; no public or otherwise verifiable event establishes who was telling the truth.

Here the key question was not the one the trial judge discussed-whether there were "large scale events occurring over a substantial period of time of a consistent pattern"-for Muse conceded that he had regularly done business with Diamond and Egister in 1969 and 1970. Instead, the question was whether he had continued to do business in 1971 after Diamond was jailed on the murder charge. The trial judge disposed of the question of memory loss by saying that "one either does or does not remember occurrences of this duration, of this extent for the rest of one's life." But Muse did not forget-none of the witnesses forgot-the general nature of their drug dealing. The district court finding therefore does not address the specific issues whether losses of memory occurred about what transpired in 1971 during the period of the conspiracy charged in the indictment, and, if so, whether such losses of memory were prejudicial. Because the trial court's findings do not relate to the specific issues in question, Judge Smith and I did not view them as binding.

There is yet another reason why the "findings" of the trial judge are, in my opinion, insufficient. They go, as will be seen by the comments below on the psychological literature, to "inherently intractable" facts not susceptible of accurate resolution. Cf. J. Ely, Democracy and Distrust 53 (1980). Therefore, the question before us is, or should be, not what rule of law to apply given a certain state of "facts," but what rule of law is preferable when the "facts" cannot be fully known.

IV. What do psychological studies contribute to judicial resolution of the problem raised?

Judge Friendly initially argued that,

(e)ven if we make the charitable assumption that Muse in fact suffered a memory lapse on these items, there is nothing to show it had not occurred before June 1, 1976, when the indictment, conceded by the majority to have been timely, was returned, or November, 1976, when the statute of limitations ran. The notion that Muse's lapse of memory about where he lived between 1969 and 1970 occurred in the interval between June, 1976 or November, 1976 and September, 1977 defies the teachings of psychology, 1 as well as common sense.

United States v. Watson, 599 F.2d 1149, 1159 (2d Cir. 1979) (dissenting from panel opinion). His footnote 1 set out the following teachings of psychology:

The basic work on memory is Ebbinghaus, Uber das Gedachtniss (1885). He shows memory declining along an asymptotic curve, with most loss occurring within a few days and almost no further decline in the time span here at issue. The theory and data behind the Ebbinghaus curve have been refined but not significantly changed. See, e. g., Hutchins & Schlesinger, Some Observations on the Law of Evidence-Memory, 42 Harv.L.Rev. 860, 862 (1928); W. & H. Mischel, Essentials of Psychology, 68-70 (1977).

In the dissent filed to the modification of the original panel opinion on rehearing Judge Friendly also said, slip op. 5023, 5036 n.5 (2d Cir. Oct. 9, 1979):

The "retroactive inhibition" theory cited by the majority in footnote 4 to the opinion on rehearing does not at all undermine the long-accepted psychological learning supporting the common sense view that memory declines along an asymptotic curve, see 599 F.2d 1159 n.1; it purports to explain how memory loss occurs, but does not challenge the conventional wisdom as to when. The fact that Ebbinghaus' studies were performed with learned nonsense syllables for obvious laboratory control purposes does not detract from the validity of his widely accepted conclusions. I am at a loss to understand why we cannot take notice of a principle of psychology so long and widely accepted, clearly a "legislative fact", see Advisory Committee Note to Federal Rule of Evidence 201, still less how the majority can proceed on a self-generated theory of memory loss directly contrary to what has been the received learning for nearly a century.

Ebbinghaus's studies did show memory declining most rapidly initially, to be sure, but his studies related to learning and relearning series of nonsense syllables, not to remembering meaningful events; and they involved experimental situations in which most of the materials learned were forgotten in a matter of days, not years. Ebbinghaus himself said that he had confined his study to "a small and definitely limited sphere." H. Ebbinghaus, Memory 65 (H. Ruger & C. Bussenius trans. 1964). It is true, as Judge Friendly's original footnote suggests, that various studies of "learned" material have "refined" but still accepted the Ebbinghaus curve. E. g., Dallenbach, The Relation of Memory Error to Time Interval, 20 Psych.Rev. 323 (1913) (memory of a picture); Strong, The Effect of Time-Interval Upon Recognition Memory, 20 Psych.Rev. 339 (1913) (recognition of words after viewing list of words); Nickerson, A Note on Long-Term Recognition Memory for Pictorial Material, 11 Psychonomic Science 58 (1968) (recognition of photographs previously viewed).

On the other hand, with respect to "meaningful materials," much longer periods of retention and more gradual declines in recall have been found, as our own experience would suggest. For example, two researchers found that their subjects' recall of the names of television shows and race horses faded gradually during the course of several years. Squire & Slater, Forgetting in Very Long-Term Memory as Assessed by an Improved Questionnaire Technique, 104 J. Experimental Psych.: Human Learning and Memory 50 (1975). Retention of material acquired over a long period of time, such as names and faces of one's high school classmates, has been found to decline gradually for many years. Bahrick, Bahrick & Wittlinger, Fifty Years of Memory for Names and Faces: A Cross-Sectional Approach, 104 J. Experimental Psych.: General 54, 65 (1975). According to Ernest Hilgard, "the rate at which forgetting occurs varies enormously with the materials used and with the circumstances under which memorization occurs.... About as much forgetting of nonsense syllables occurs in a day as occurs in a year with the materials of an academic course." Introduction to Psychology 292 (3d ed. 1962). Erdelyi and Kleinbard have suggested that "(Ebbinghaus's) nonsense-syllable tactic may have been responsible for some major distortions in memory research." Has Ebbinghaus Decayed with Time?: The Growth of Recall (Hypermnesia) over Days, 4 J. Experimental Psych.: Human Learning and Memory 275, 286 (1978).

Differences in rates of retention and recall have been found to be dependent not only on the nature of the material but also on a variety of other factors. One such factor-intervening events which interfere with memory-is posited by the interference theory referred to in Redmount, The Psychological Basis of Evidence Practice: Memory, 50 J.Crim.L.C. & P.S. 244, 253-54 (1959), quoted in slip op. 5023, 5027 (order modifying panel opinion). See, e. g., J. Adams, Learning and Memory 286 (1976); R. Crowder, Principles of Learning and Memory 225, 260-61 (1976); L. Cermak, Human Memory 45 (1972) ("The extent of long-term memory has been shown to be primarily a function of the amount of interference that is present at the time of recall."). In general, as Professor Adams has written, "psychologists no longer think in terms of the curve of forgetting. Forgetting is a function of many factors and there are many curves of forgetting." J. Adams, supra, at 225. To the same effect, Professor Redmount has concluded that

(m)emory for a particular event may also be a function of the dimensions of time and meaning of the event itself, and of time and meaning as carried through experience. Elapsed time since the event, the event's meaningfulness to the individual in general and personal terms, the intensity and duration of its occurrence, and its uniqueness in general or personal experience, may be important determinants of memory sufficiency and accuracy....

Within this more complete conceptual framework memory must be conceived as a phenomenon of the individual to be studied in the context of his characteristics and experience.

Redmount, supra, at 262.

Thus, with all highest respect to Judge Friendly, neither the "teachings of psychology," "the received learning for nearly a century," nor, I submit, "common sense" dictates that we assume the memory losses at issue here occurred at any given point in time, either before or after June 1, 1976, or November 1976. At the very least the matter is one with respect to which scientific data and thinking are so varied that judicial notice may not be taken.

V. In light of the policies of statutes of limitations and the state of the art in the psychology of memory, what should judicial resolution be?

I have suggested above that, in light of both the policy underlying statutes of limitations and the fact that losses of memory cannot be pinpointed in time, when sealed indictments are opened after a statute has run and when prejudice due to loss of memory is demonstrated, it should not be necessary for the defendant to show the precise time at which that memory loss occurred. The law should presume that the prejudice occurred because of delay. A contrary presumption gives insufficient weight to the purpose of statutes of limitations, to the extent that their purpose is related to the forgetting of past events and the difficulty in getting at facts after years have passed-what might be called the staleness of a case after a certain period of time. In a given case one may find, of course, that there has been no memory loss, or that no prejudice has resulted from a memory loss, as, for example, when there is specific corroborative evidence. But here the case was essentially a swearing contest between the "paid" prosecution witnesses and the defendant.

The significance of statutes of limitations, in providing notice of charges, was emphasized in United States v. Marion, 404 U.S. 307, 92 S. Ct. 455, 30 L. Ed. 2d 468 (1971). While stating that the sixth amendment speedy trial provision is inapplicable to pre-indictment delay, the Court noted that

(t)he law has provided other mechanisms to guard against possible as distinguished from actual prejudice resulting from the passage of time between crime and arrest or charge.... (Statutes of limitations) provide predictability by specifying a limit beyond which there is an irrebuttable presumption that a defendant's right to a fair trial would be prejudiced.

Id. at 322, 92 S.Ct. at 464 (footnote omitted). The Court went on to quote with approval Toussie v. United States, 397 U.S. 112, 114-15, 90 S. Ct. 858, 859-60, 25 L. Ed. 2d 156 (1970):

The purpose of a statute of limitations is to limit exposure to criminal prosecution to a certain fixed period of time following the occurrence of those acts the legislature has decided to punish by criminal sanctions. Such a limitation is designed to protect individuals from having to defend themselves against charges when the basic facts may have become obscured by the passage of time and to minimize the danger of official punishment because of acts in the far-distant past....

Id. at 322-23, 92 S.Ct. at 464-65. The very same considerations that exist when an indictment is not timely filed also exist when, because of a sealed indictment, a defendant is not given notice of the charges against him until long after (here ten months after) the expiration of the statutory period.

Yet, under our case law, the defendant indicted under seal is not accorded protections against prejudice resulting from a delay beyond the limitations period in being advised of criminal charges. In such a case, the total delay from the criminal act until the time of trial consists of both pre-indictment delay and pre-unsealing delay. Under Marion, 404 U.S. at 324, 92 S.Ct. at 465, pre-indictment delay constitutes a due process violation only if there is substantial prejudice to the defendant and if the delay arises from the Government's intentional effort to improve its position. See United States v. Lovasco, 431 U.S. 783, 795-96, 97 S. Ct. 2044, 2051-52, 52 L. Ed. 2d 752 (1977). To be sure, a speedy trial claim may be advanced in connection with a delay after an indictment has been sealed; such a claim was made to and rejected by the original Muse panel. Although the pre-indictment period may be examined as a circumstance relevant to assessing post-indictment delay and speedy trial rights, see United States v. Vispi, 545 F.2d 328, 333 (2d Cir. 1976), only the post-indictment delay and any prejudice incident to it are strictly relevant, constitutionally speaking.

The statute of limitations is, or should be, a source of protection from delay beyond the limitations period, delay consisting of both the time from the crime to the sealed indictment and the time from that indictment to the date of unsealing and notification to the defendant. Marion indicates that the overriding policy of the statute of limitations is to protect a defendant from excessive delay which undermines his ability to defend himself. Such delay, of course, terminates only when the defendant is apprised of a criminal indictment. However, Federal Rule of Criminal Procedure 6(e), which allows an indictment to be kept secret until the defendant is in custody, has been interpreted to mean that the filing of a sealed indictment within the statutory period serves to toll the statute of limitations. United States v. Michael, 180 F.2d 55, 56-57 (3d Cir. 1949), cert. denied, 339 U.S. 978, 70 S. Ct. 1023, 94 L. Ed. 1383 (1950); see United States v. Watson, 599 F.2d 1149, 1154 (2d Cir. 1979) (original panel decision). An accused in appellant Muse's position therefore cannot claim a statute of limitations bar when an indictment is filed and sealed prior to expiration of the statutory period.

This leads to the ultimate question of this case: as a matter of supervisory power or otherwise, should this court formulate a judicial rule, grounded in the statute of limitations, placing some limit on the judicial extension of the statute of limitations in the case of a sealed indictment, thereby protecting a defendant from irreparable prejudice due to a delay in notification greater than the maximum delay allowed by the statute of limitations? Judge Smith and I thought it wise to define some boundary to the Government's privilege to toll the statute of limitations by filing a sealed indictment. We therefore held that when there has been a delay which extends beyond the limitations period, a defendant proves actual and substantial prejudice attributable to the delay, and the exact time the prejudice accrued is not ascertainable, then it is inappropriate to invoke the tolling provision of United States v. Michael and it is necessary to dismiss the indictment. Our decision was that in such a case an indictment should be dismissed, as it should be in a situation in which the prejudice demonstrated is clearly attributable to the delay beyond the limitations period. Although the statute of limitations irrebuttably presumes prejudice to the defendant, see Marion, 404 U.S. at 322, 92 S. Ct. at 464, our rule requires proof of actual and substantial harm.

Judge Friendly's dissent and the Government's rehearing petition raise several objections that, I believe, are not meritorious but warrant brief comment in light of the en banc opinion. One such objection is that the panel did not consider the Government's need for delay beyond the limitations period. But, the Government's need is respected by permitting the unsealing of an indictment after the limitations period; it is because the Government has a legitimate interest in sealing a particular indictment that the statute of limitations is tolled and that an indictment unsealed after the statute of limitations is considered timely.

It is also suggested that our rule erroneously looks to the entire delay period rather than to the post-sealing or to the post-limitation delay, and our choice of language in this regard was no doubt inept. To gain dismissal of an indictment for actual and substantial prejudice, a defendant should be required to demonstrate both that the prejudice has occurred and that it is attributable to the delay and not merely incidental to it-except where, as here, the time of occurrence is unascertainable, in which case the prejudice must be presumed to have occurred from the length of the delay. Thus, the death of a witness prior to expiration of the limitations period would not be prejudice attributable to the delay beyond the limitations period; but the death of a witness after the limitations date passed could be prejudice warranting dismissal. Faded memory of a defendant or witness, while posing a different problem because its date cannot be ascertained precisely, constitutes prejudice attributable to delay only if the defendant's inability to prove his defense results from the delay.

Here, it seemed to me that dismissal of the indictment was proper because Muse's inability to remember details of past events was matched by the Government witnesses' inability to recall. Muse was thus by reason of the delay unable to prove his defense or to impeach the Government witnesses who alleged that he was a participant in the conspiracy charged. At the same time, the Government's case was not strong enough to render Muse's defense implausible.

It is, of course, not for me to suggest that in the light of our usually strict approach toward rehearings en banc this case is not of earthshaking moment. I may note, however, that with the perspective of a few months' time, I see our initial panel decision as simply a rather minor limitation upon the rule of United States v. Michael permitting the filing of sealed indictments to toll the statute of limitations. I rather doubt that many cases will arise in which an indictment has been sealed, the statute of limitations has run, and the required prejudice is shown. I thought that such prejudice was shown here, and the fact immaterial under the circumstances that the precise time it arose was unascertainable. I still think so and therefore am required to dissent.