BAZELON, Circuit Judge.
On August 6, 1951, a first degree murder indictment was returned against appellant, charging him with the fatal shooting of his wife on June 20, 1951. The judgment of conviction of second degree murder from which he now appeals was entered on January 20, 1956. The time lapse between Wright’s indictment and his conviction is attributable to his mental condition.
At arraignment on August 9, 1951, the judge ordered a mental examination of Wright. Another order two weeks later designated Drs. Perretti and Gilbert of the District of Columbia General (then Gallinger Municipal) Hospital to make the examination. Both doctors made their examinations of Wright in September 1951 and both reported him incompetent to stand trial. On October 2,1951, the District Court, after a hearing, adjudicated Wright to be incompetent to stand trial and committed him to the custody of the Attorney General. Pursuant to that commitment, Wright was placed in the Springfield Medical Center for Federal Prisoners. In November of 1952, the psychiatric staff of that institution having concluded that his competency to be tried had been restored, he was returned to the District of Columbia Jail and in February 1953 was brought to trial and found guilty of first degree murder. But the trial judge, on the basis of testimony adduced during the trial,1 [7]*7ordered a further mental examination by the Mental Health Commission. The Commission examined Wright on June 4, 1953, and reported that he was then and had been during his trial of unsound mind. The judge, after holding a hearing, set aside the conviction and again committed Wright to the custody of the Attorney General. Pursuant to that commitment, after some delay, Wright was sent to St. Elizabeths Hospital where he remained until, on November 15, 1955, he was adjudged competent and was ordered to trial. In connection with the second trial, the court ordered Wright examined by Drs. Perretti, Rickman, Williams, Todd and Cavanaugh, all of whom had examined him before, and by Drs. Miller and Odenwald, who had not seen him before. All the doctors found him mentally competent at the time of that series of examinations, in December 1955.
At the trial, Wright’s principal defense was insanity and the issues raised on appeal relate largely to that defense. His court-appointed counsel argues that the Government failed to sustain its burden of proving beyond a reasonable doubt that Wright was sane when he shot his wife. He also argues error in the refusal of certain jury instructions relating to the insanity issue.
Evidence on the Issue of Sanity
It is not questioned that, when the defendant introduces some evidence to raise the issue of insanity, his sanity at the time of the offense becomes an element of the crime, which, like all other elements of the crime, must be proved by the Government beyond a reasonable doubt. Davis v. United States, 1895, 160 U.S. 469, 488, 16 S.Ct. 353, 40 L.Ed. 499; Tatum v. United States, 1951, 88 U.S.App.D.C. 386, 190 F.2d 612; Douglas v. United States, 1956, 99 U.S.App.D.C. 232, 239 F.2d 52. When the Government has not sustained its burden of proof, i. e., when it appears that reasonable jurymen could not conclude beyond a reasonable doubt that the act was not the product of defendant’s mental illness, “there is a duty to set aside a verdict of guilty and to direct a verdict of not guilty by reason of insanity” or to order a new trial. Douglas v. United States, 99 U.S.App.D.C. at page 237, 240, 239 F.2d at page 57, 60; Curley v. United States, 81 U.S.App.D.C. 389, 160 F.2d 229, certiorari denied, 1947, 331 U.S. 837, 67 S.Ct. 1511, 91 L.Ed. 1850.
The nature and quantum of evidence of sanity which the Government must produce to sustain its burden and take the issue to the jury will vary in different cases. Evidence of sanity which may suffice in a case where defendant has introduced merely “some evidence” of insanity may be altogether inadequate in a case where the evidence of insanity is substantial. Before considering the sufficiency of the Government’s proof in the instant case, therefore, we must assess the evidence of insanity introduced by the appellant.
Wright made his showing of insanity through the testimony of eleven of the Government or court-appointed psychiatrists who examined him or had him under treatment almost continuously during the four years between his arraignment and his trial. This testimony was supplemented by reading to the jury copious extracts from the records of the two mental institutions in which he had spent the major part of those four years.
Drs. Perretti and Gilbert, who had examined Wright in September 1951, testified that they had found him suffering from schizophrenia. In Dr. Perretti’s opinion, the condition dated back to the time of the crime (about three months before the examination). Dr. Gilbert, though not stating a specific opinion, said that the disease Wright had “doesn’t come or go overnight.” Drs. Rickman and Williams, who had examined Wright just before his 1953 trial and found him suffering from psychomotor seizures and schizophrenia, testified that the illness, in their opinion, dated back to the time of the crime. Drs. Todd and Cavanaugh [8]*8of the Mental Health Commission, who had examined Wright shortly after his 1953 conviction,'’testified that they had found him to have schizophrenia and to be mentally , deficient. Dr. Cavanaugh .could not say whether the condition could be dated back to the date of the shooting; Dr. Todd said that the condition observed in 1953 could be traced back several months, but that 1951 was too far back. Drs. Cushard, Epstein and Tartaglino, the St. Elizabeths staff doctors who examined Wright beginning early in 1954, testified that he was then suffering from schizophrenia. On the basis of the information they had, however, they could not say whether or not the condition had existed in June 1951. Dr. Cushard pointed out that they had never been asked to conduct an examination to determine Wright’s June 1951 condition. D.rs. Miller and Odenwald who examined Wright in December 1955 testified that he had been suffering from schizophrenia at the time of the shooting. Dr. Miller’s opinion was based on a history obtained from Wright and medical history derived from the hospital records. Dr. Odenwald based his opinion on what he could learn of the medical history from Wright him- ' self.
As to causal connection, four of the five witnesses who said that Wright was mentally ill at -the time of the shooting were asked whether the act was the product of the illness. Dr. Perretti said, “Yes”; Dr. Miller said, “Could very well be”; Dr. Williams said, “Likely”; and Dr. Odenwald said, “Surely possible.” Dr. Rickman, the fifth of the witnesses who testified that Wright had been mentally ill at the time of the shooting, and Dr. Gilbert, who indicated that he probably had been, were not asked whether they thought the act was a product of the illness. Nor was that question put to Drs. Todd, Cavanaugh and Tartaglino. Drs. Epstein and Cushard, when asked the causation question, replied that they ■had Insufficient data to support an'opinion, Dr. Cushard noting that the causal ■connection between an individual’s mental illness and his act “requires very intensive investigation and examination of the person.”
In summary of Wright’s showing of insanity, we observe that (1) five or six of the eleven psychiatrists who testified were of the opinion that he had been, mentally ill at the time of the shooting; and (2) several of them stated, with varying degrees of certainty, that the shooting was the product of the illness.
The Government challenges Wright’s showing of insanity on several grounds. It points out, first, that some of the doctors who examined Wright were unable to say that he was mentally ill at the time of the shooting. But this does not detract from the testimony of the doctors who were able to state an opinion that he was ill at that' time. None of the eleven witnesses said that he was not ill at the time of the shooting or that the shooting was not the product of the illness. Nor did any of the doctors who could • .not state an opinion challenge either the qualifications or the reasoning of those who did state an opinion. There was thus no conflict in the medical testimony.
Next the Government attacks as uncertain and inconclusive most of the testimony that the act was the product of the illness. Only Dr. Perretti, the first doctor to examine Wright after his arrest, gave an unequivocal opinion that the illness caused the act. Three other doctors expressed their opinions in degrees of probability. Obviously, unequivocal opinions, if obtainable, are more desirable than equivocal ones. But the opinion to which a psychiatrist testifies, need only be “the type of clinical opinion he is accustomed to form and to rely upon in the practice of his profession.” It need not consist of “mathematically demonstrable certainties.” Blunt v. United States, 1957, 100 U.S.App.D.C. 266, 275, 244 F.2d 355, 364. Moreover, Dr. Cushard, who was unable from 1954 examinations to say either whether Wright had been ill in 1951 'or [9]*9whether the illness caused the act, said he would need to make a “very intensive investigation and examination” to answer the causation question. That such an investigation and examination was not conducted is not Wright’s fault. Many deficiencies in the process by which we collect the evidence upon which cases like this turn, are within the prosecution’s capacity to remedy and should not be chargeable to the accused if he has no control over them. Blunt v. United States, 100 U.S.App.D.C. at page 275, 244 F.2d at page 364, note 23.
The same can be said of the Government’s third attack on the psychiatric evidence of insanity — that many or all of the psychiatric examinations were too remote in time from the date of the shooting to afford a basis for an expert opinion as to Wright’s mental condition on that date.2 The record shows that the question as to Wright’s sanity arose at his arraignment on August 9, 1951, and that the court ordered a mental examination on that date. It is not contended that Wright was responsible for the delay that ensued or for the fact that the examinations held in compliance with the August 1951 and the numerous subsequent court orders may not have been as thorough as psychiatrists would prefer. If the Government feels that psychiatric opinions which come into evidence ought to be based on examinations of greater scope and intensity than has been the practice heretofore, it can and should arrange to have such examinations made.
Wright’s showing of insanity, far from being merely “some evidence” of insanity, is about as strong as can ever be made, unless the accused happens to have had a psychiatric examination immediately prior to his act. To send the case to the jury in the face of so strong a showing of insanity requires more than minimal evidence of sanity.
The only evidence offered by the Government to sustain its burden of proving beyond a reasonable doubt that Wright was sane when the shooting occurred was the testimony of two policemen. The Government’s brief summarizes that testimony as follows:
“Officer Walker, who arrested the defendant within minutes after the shooting, stated that during the time he talked to the defendant he believed the defendant was rational and of sound mind although he was nervous (R. 135-136). Officer Hartnett, who was with the defendant both on the morning of the crime and the day after the defendant’s wife succumbed, was of the same opinion (R. 210-211). The officer stated that he had no trouble understanding the defendant, that the defendant gave responsive answers to his questions, and that as far as he could ascertain from looking and listening to the defendant, he was rational at the time (R. 210-211). The officer also said that the defendant was coherent during all the times that they talked, and he gave it as his opinion that the defendant was of sound mind during the critical period (R. 210-211).” [Emphasis supplied.]
Although the testimony of lay witnesses may be competent evidence on the issue of sanity, it does not follow that, in the face of a substantial showing of insanity, the Government may send the issue to the jury simply by having two policemen testify, “He looked all right [10]*10to me.” The probative value of any opinion on the issue of sanity depends on the facts upon which it is based. This is especially true of a lay opinion. Lay “witnesses may testify only upon the basis of facts known to them. They may testify as to their own observations and may then express an opinion based upon those observations. Of course the testimony of a lay witness with training in this or related fields may have more value than the testimony of a witness with no such training.3 Also obvious upon a moment’s reflection is the fact that, while a lay witness’s observation of abnormal acts by an accused may be of great value as evidence, a statement that the witness never observed an abnormal act on the part of the accused is of value if, but only if, the witness had prolonged and intimate contact with the accused.” Carter v. United States, 101 U.S.App.D.C. —, 252 F.2d 608.
Reasonable jurymen could not conclude beyond a reasonable doubt, from the testimony of these two policemen or from the rest of the record before us, that Wright was sane at the time of the shooting.
In addition to the evidence of sanity it offered itself, the Government relies on the fact that all the psychiatrists who testified at the trial for the defense said that Wright was at that time — in December 1955 — of sound mind. As was pointed out in Lyles v. United States, 101 U.S.App.D.C.-,-F.2d , “evidence as to insanity at any time — the present as well as any other— may be admitted * * * where it is used by a witness as part of the data upon which he bases a conclusion as to mental condition at the time of the offense.” Here, none of the psychiatrists gave an opinion that Wright was sane at the time of the crime. For subsequent mental condition to be a premise from which the jury may draw an inference as to mental condition at the time o.f the crime, there must be such a circumstantial relationship between the two periods as demonstrates the relevance of one to the other. Since four and one-half years elapsed between the time of the offense and the time of the trial and, during that interval, Wright went through long courses of treatment in two mental institutions, his 1955 soundness of mind cannot be said to be sufficiently probative of 1951 sanity to take the case to the jury. If proof of present soundness of mind necessarily takes the case to the jury, then, contrary to Douglas, the court could never direct a verdict of not guilty by reason of insanity, since all defendants, on trial, are at least of sound enough mind to be tried.
We hold that the Government failed to sustain its burden of proving sanity. The conviction is reversed and the case remanded to the District Court with instructions to grant a new trial if the Government shall request it;4 or, absent such request, to enter a judgment of acquittal by reason of insanity notwithstanding the verdict.
Since, under our mandate, Wright may be brought to trial again upon the present charge, we express our views on certain matters which would be pertinent upon retrial.
The Jury Instructions
1. The judge instructed the jury that it could return one of five possible verdicts: guilty of murder in the first degree, guilty of murder in the second degree, guilty of manslaughter, not guilty by reason of insanity, and not guilty. In connection with the fourth possible verdict, not guilty by reason of insanity, the charge included a recital, virtually in haec verba, of the sample instruction suggested in Durham v. United States, [11]*111954, 94 U.S.App.D.C. 228, 214 F.2d 862, 45 A.L.R.2d 1430, and a statement of the Davis rule that the prosecution must prove sanity beyond a reasonable doubt. It contained other matter, however, which could have negated the effect of the instruction as to the Davis rule, or, at least, raised confusion as to it.
The instruction on the fourth possible verdict begins with the following statement:
“It is claimed in behalf of the defendant * * * that he was insane at the time of the commission of the act. If you find this to be the case, then your verdict should be a verdict of ‘Not guilty because of insanity.’ ” [Emphasis supplied.]
Concluding the instruction on the fourth verdict and going on to the fifth, the verdict of “not guilty,” the judge stated:
“To repeat: You may return the fourth verdict, namely, not guilty because of insanity, if you find in the circumstances that the defendant at the time the crime was committed was either suffering from, a mental disease * * * or from a mental defect, * * * and that there was a causal connection between such mental defect or such mental disease and the commission of the crime with which the defendant is charged.
“There is a fifth possible verdict in this case. It is a verdict of not guilty. You may return that verdict if you find that the Government has failed to establish from the evidence beyond a reasonable doubt that the defendant committed the offense with which he is charged.” [Emphasis supplied.]
There is danger that the jury might infer from the contrasting treatment of the two possible verdicts that, in order to acquit by reason of insanity, it would be necessary for them to reach the affirmative conclusions that Wright was insane at the time of the crime and that the act was the product of the illness.
2. As we have noted, the judge gave the jury the sample Durham instruction on the meaning of insanity. After several hours of deliberation one juror asked for further instruction as to whether “in determining sanity * * * any other consideration * * * might be included * * * than dementia or schizophrenia.” The judge replied:
“The Court has read the instructions to you. The Court is not going to amplify the instructions. They are lenghty. They are inclusive. The Court feels they are adequate to enable you to return a verdict in this case. So that the answer to the question that you have asked is answered in the way the Court has just answered it.
“It will be necessary for the jury to reach its determination on the evidence and on the instructions which the Court has given you.”
A defense request embodying the juror’s request for explanation was also denied.
The refusal to answer the juror’s question and the denial of the requested instruction constitute reversible error. The sample Durham charge was not meant to be and is not an inflexible directive to be followed by rote. Douglas v. United States, 1956, 99 U.S.App.D.C. 232, 238, 239 F.2d 52, 58, note 10. It sketches the substance of what the judge “should in some way convey to the jury.” 94 U.S.App.D.C. at page 241, 214 F.2d at page 875. Where, as here, the need for more appears, it is the duty of the judge to fill in the sketch, as may be appropriate on the basis of the evidence, to provide the jury with light and guidance in the performance of its difficult task.
The terms “disease” and “defect” are not so self-explanatory and our definition of them in Durham is not so definitive as to make elucidation always superfluous. Whether elucidation is necessary and, if so, whether failure to give it is prejudicial error depend upon the evidence which the jury has heard. In this case there was testimony from nine psychiatrists that Wright, at various times, was afflicted by schizophrenia (or dementia [12]*12praeéox — an older term for the same disease). Some of this testimony dwelt 'at such length and with such detail upon the nature of that disease that it would not be surprising if a juror understood that schizophrenia was the only form of mental disease amounting to “insanity.” Where there is evidence to support a finding that the accused suffered from some other form of mental disease or from a mental defect the jury should be instructed that it may make that finding, especially when a juror has indicated by a question that the general instruction, as given, is ambiguous, confusing or inadequate. Here, in addition to the evidence of schizophrenia, there was testimony from two doctors that Wright was a victim of “psychomotor epilepsy” or “psychomotor seizures” at the time of the shooting. There was also evidence that tests had shown Wright was a moron. In view of this evidence and the juror’s question, merely reciting the sample Durham instruction did not accomplish what we said in that case was the function of the instruction — to “provide the jury with guides for determining whether the accused can be held criminally responsible.” 94 U.S.App.D.C. at page 241, 214 F.2d at page 875.
3. Two of the psychiatrists who testified that Wright had been mentally ill at the time of the shooting said further that, by reason of his illness, he had been unable to distinguish right from wrong. The court refused a defense request to instruct the jury to acquit Wright by reason of insanity if it found that he had been unable to distinguish between right and wrong.
The instruction should have been given. While capacity to distinguish right from wrong is no longer the earmark of legal sanity, the lack of that capacity is one of the earmarks of legal insanity. Durham v. United States, 94 U.S.App.D.C. at page 242, 214 F.2d at page 876. When evidence of such lack of capacity exists, “a trial court should permit the jury to consider [it] in resolving the ultimate issue ‘whether the accused acted because of a mental disorder,’ ” Douglas v. United States, 99 U.S.App.D.C. at page 238, 239 F.2d at page 58; “the jury should be instructed that it is relevant in determining whether the unlawful act was the product of mental disease or defect.” Stewart v. United States, 1947, 101 U.S.App.D.C. 51, 247 F.2d 42, 44.
4. The jury was instructed, in terms of the sample Durham instruction, that, even if Wright was suffering from a mental disease or defect at the time of the shooting, it could still convict him “if there was no causal connection between such mental abnormality and the act.” The court refused a defense request to explain “causal connection” as used in the charge.
When we say that one event; causes another, do we mean it is a cause of it, or the principal cause, or the exclusive cause? When a jury is told that it may convict the accused if it finds that, though he was mentally ill when he committed the act, the illness did not cause the act, may it convict if it finds that the illness was one of the causes of the act, but not the exclusive cause? The answers to these questions are to be found in Carter v. United States, 101 U.S.App.D.C. at page -, 252 F.2d at page 614; and Douglas v. United States, 1956, 99 U.S.App.D.C. 232, 239, 239 F.2d 52, 59. We said in Carter:
“When we say the defense of insanity requires that the act be a ‘product of’ a disease, we do not mean that it must be a direct emission, or a proximate creation, or an immediate issue pf the disease in the sense, for example, of Hadfield’s delusion that the Almighty had directed him to shoot George III. * * *
“ * * * There must be a relationship between the disease and the act, and that relationship, whatever it may be in degree, must be, as we have already said, critical in its effect in respect to the act. By ‘critical’ we mean decisive, determinative, [13]*13causal; we mean to convey the idea inherent in the phrases ‘because of’, ‘except for’, ‘without which’, ‘but for’, ‘effect of’, ‘result of’, ‘causative factor’; the disease made the effective or decisive difference between doing and not doing the act. The short phrases ‘product of’ and “causal connection’ are not intended to be precise, as though they were chemical formulas. They mean that the facts concerning the disease and the facts concerning the act are such as to justify reasonably the conclusion that ‘But for this disease the act would not have been committed.’ ” [101 U.S.App.D.C. at page-, 252 F.2d at page 616.]
Similarly, in Douglas, we said the accused should be discharged of responsibility if at the time of the crime he had a mental illness “without which [the crime] would not have occurred.”
On the evidence in the record the jury could have found two things to be simultaneously true: (1) that Wright would not have shot his wife but for his mental illness; and (2) that Wright might have had a “rational” motive for his act, i. e., the suspicions and passions engendered by the dissension between them which culminated in their separation. Without an explanation of “causal connection,” the jury may have erroneously concluded that, though the shooting would not have occurred but for Wright’s illness, the principal cause of the shooting was the rational one, i. e., his ill-feeling against his wife, and that, therefore, he ought to be held accountable for his act. While an instruction that conviction depends on “causal connection” may generally be sufficiently accurate, a defense request for an explanatory instruction should not be refused when there is latent ambiguity in the case.
The Admission of the Confession
The court refused Wright’s request for a preliminary hearing on the question of the voluntariness of his confession and submitted the question to the jury. This was error.
To substantiate a defendant’s contention that his confession was involuntary, it is generally necessary for him to take the stand. Since it is a. matter of course that the defendant should be allowed to testify to the involuntariness of his confession without waiving his privilege against self-incrimination, 3 Wigmore, Evidence 345 (3d Ed. 1940), it follows that, when he requests it, he should be given a hearing without the jury.
The majority rule is that the question of voluntariness is for the court alone to decide. Id., § 861. In this jurisdiction, however, by “law and practice,” the court first holds a preliminary hearing for the purpose of determining whether there is evidence from which the jury could properly conclude that the confession was voluntary. If the court concludes there is no such evidence, it must exclude the confession; but if it finds there is evidence on the basis of which it might be held to be voluntary, then the question of voluntariness is submitted to the jury. McAffee v. United States, 1939, 70 App.D.C. 142, 105 F.2d 21; Id., 72 App.D.C. 60, 111 F.2d 199, certiorari denied, 1940, 310 U.S. 643, 60 S.Ct. 1094, 84 L.Ed. 1410; Catoe v. United States, 1942, 76 U.S.App.D.C. 292, 131 F.2d 16; cf. Ziang Sun Wan v. United States, 1924, 266 U.S. 1, 45 S.Ct. 1, 69 L.Ed. 131. The trial judge, in refusing to hold a preliminary hearing, said a rule requiring it would make for cumbersome proceedings. Be that as it may,5 the rule does require it and the judge was wrong to refuse.
We held in Tyler v. United States, 1951, 90 U.S.App.D.C. 2, 6, 193 F.2d 24, 28, certiorari denied, 1952, 343 U.S. 908, [14]*1472 S.Ct. 639, 96 L.Ed. 1326, that, where a preliminary hearing was omitted because of a misunderstanding between court and counsel and the evidence showed there was ample basis for a conclusion of voluntariness, it was not prejudicial error to have submitted the question to the jury without preliminary hearing. We did not suggest, as the Government argues here, that the Mc-Affee rule was to be disregarded. That rule is to be followed when a contested confession is offered. The duty it places on the trial judge is mandatory.
Reversed and remanded.
BURGER, Circuit Judge, concurs in the result only.