Appellant was charged with second degree murder, but convicted of manslaughter for the fatal stabbing of her brother, Benjamin King, on November 22,1964. She was sentenced to imprisonment for two to ten years.
The salient physical facts with respect to the crime are briefly as follows: On the afternoon of November 22, 1964, appellant went out to get some wine and beer. She returned to the residence she shared with her mother, drank the wine and beer, and was more or less “high” at the time her brother returned to the house. Benjamin hit her, apparently out of anger because appellant was drinking again and talking loudly. Appellant and her mother presented differing testimony as to the nature of Benjamin’s assault — • whether a mere slap, or a reeling blow followed by his pursuit of appellant into the kitchen. In response to this assault by a large and powerful man, appellant, a tiny woman, plunged a small knife used for peeling potatoes into his heart. Testimony conflicts as to the timing of the fatal blow — whether in immediate reaction to the assault by the decedent, or only after having been backed into the kitchen by him.
I
Appellant’s main contention on appeal is that the District Judge should have granted her motion for a directed verdict of not guilty by reason of insanity. She urges that a reasonable doubt of sanity must be held to exist in this case, in view of the “uncontradicted and inherently credible expert testimony” of two St. Elizabeths psychiatrists who had examined appellant — Dr. Charles Ruch, then assistant clinical director of the Dix Pavilion, and Dr. Alkinoos Vourlekis.
The two psychiatrists, Dr. Ruch and Dr. Vourlekis, agreed that appellant suffered from a personality disorder — described as passive-aggressive personality, with organic and alcoholic features-— severe enough to be classified as a mental illness.1 Dr. Vourlekis testified that *386the crime was “probably” the product of appellant’s mental disease, while Dr. Rueh stated that the killing was “possibly” the product of her mental disease. Their testimony as to the dynamics of appellant’s mental illness and its relation to the crime was in substance the same as this excerpt from the official hospital report filed with the court before the trial began:
The possible relationship between the alleged offense and the patient’s personality disorder is best described in the following manner. The patient has extreme difficulty in getting her dependency needs met and tends to respond to aggression by passive obstructionism and by using alcohol to excess. However, when intoxicated her aggression is liberated and violent acts may occur such as the stabbing of her brother. Along with this, her lowered intellectual quotient causes her to display very poor judgment and not to consider alternative ways of dealing with situations.
Both doctors were asked about their observations of appellant in the hospital. Dr. Ruch noted that appellant was quiet and tended to stay to herself, showed an overly cooperative attitude, had “kind of a bland expression,” rarely showed emotion, tended to be circumstantial, and exhibited a “kind of farawayness to her appearance, as though she was not particularly concerned about her rather difficult spot.” Dr. Vourlekis testified that appellant was withdrawn and reclusive in the ward, that “she did not display the usual emotional range average people do,” that her judgment was poor, and that she exhibited complete displacement of responsibility.
Both doctors testified as to the existence of some organic brain damage. This testimony will be considered in greater detail below. Appellant’s history was also the subject of testimony by the psychiatrists. Dr. Ruch noted that appellant had a history of “pretty severe deprivation,” had suffered “quite a bit of physical abuse” from the deceased and another man, and had been drinking heavily since she was 25 (she was then 40). Dr. Vourlekis testified appellant “felt always inferior,” that she resented being treated as a baby and slapped around by her brother, that her children had been taken from her custody by court order in 1959.
The prosecutor elicited from Dr. Ruch that appellant was generally able to recognize the difference between right and wrong, although at the time of the stabbing she was “probably overwhelmed by emotion” and was “certainly emotional and pretty much out of control of herself.”
Both on direct, and in particular on probing cross-examination, the psychiatrists testified on the issue of productivity — whether there was such causal connection that the offense could be regarded as the product of the mental illness. Dr. Ruch testified as follows on direct:
[Where] we have a personality disturbance, there are so many things involved that it is hard to say what caused her to do what she did. I say possibly on the basis that her passive-aggressive personality — frequently someone with a relatively severe illness like this will turn to the use of alcohol.
******
Once someone is drinking- — -the drinking being related now to the personality disorder — once someone is drinking, there is a lessening of their inhibitions. Someone who is particularly angry for instance, when intoxicated may have a tendency to be more explosive, both verbally and possibly physically, and this is possibly what happened in this case.
On cross-examination he testified:
Once under the influence of alcohol a personality disturbance like this, where one tends to sit on one’s anger, the person is in trouble emotionally in controlling their anger, and hence an act like this can occur. Now, whether — I can’t say for certain it is a product. I can’t say for certain that it isn’t.
*387Dr. Vourlekis testified that in his opinion the crime was probably the result of appellant’s mental illness. Persons with a passive-aggressive personality condition manage to have their strong dependency needs taken care of. But he continued [Tr. 126-27]:
Such people also have difficulties in varying degrees with authority figures. They harbor quite often a great amount of resentment towards authoritative figures, which when their controls get lower, this resentment can be acted out, can be manifested in other words.
Now, in Miss King’s case, there seems to be an additional factor there, mainly alcohol which is well known for its effect in lowering controls. A person under the influence of alcohol does not control his actions as well as a person who is not under the influence.
The following colloquy on productivity between prosecutor and witness occurred after Dr. Vourlekis testified that appellant had told him that on November 22, 1964, she had consumed about a fifth and a pint of wine, plus six beers:
Q Was she capable of exercising the freedom of choice of having that first drink or not having that first drink that day?
A Well, I don’t know.
Q You don’t know. Why not?
A I don’t know whether she had the freedom of choice.
Q Was she suffering from any mental defect which would impair her behavior so that she couldn’t make that choice of taking that drink or not taking that drink?
A I am afraid I can’t relate the two.
******
Q It has been testified here that her brother reprimanded her for making too much noise, that he slapped her and she stabbed him with a knife. Was she capable of exercising the freedom of choice of stabbing or not stabbing?
A I would doubt it very much simply because she was under the influence of alcohol at that time.
On the insanity issue the Government also relied on the fact that appellant exhibited an excellent memory and the ability to respond logically to sustained questioning. As direct evidence, the Government produced only an officer who testified that on appellant’s arrest shortly after the homicide she gave clear answers to all questions asked her, and that he thought she was of sound mind and not intoxicated.
Although the testimony in support of the insanity defense was impressive, we conclude that it was not sufficient “to compel a reasonable juror to entertain a reasonable doubt concerning the accused’s responsibility,” and hence not sufficient to require the judge to direct a verdict of not guilty by reason of insanity. McDonald v. United States, 114 U.S.App.D.C. 120, 123, 312 F.2d 847, 850 (1962) (en banc). Since McDonald, where we examined at some length the problem of expert testimony and jury consideration of the insanity question, this court has been chary of directed verdicts of not guilty by reason of insanity, even where expert testimony is uncontradicted.2 The case at bar seems a particularly apt occasion to recall, by quotation rather than attempt at paraphrase, this court’s reasoning in McDonald, supra at 124, 312 F.2d at 851:
What psychiatrists may consider a “mental disease or defect” for clinical purposes, where their concern is treatment, may or may not be the same as mental disease or defect for the jury’s purpose in determining criminal responsibility. Consequently, for that purpose the jury should be told that a mental disease or defect includes any abnormal condition of the mind which substantially affects mental or emotional processes and substantially im*388pairs behavior controls. Thus the jury would consider testimony concerning the development, adaptation and functioning of these processes and controls.
In the case at bar a reasonable juror might have been convinced beyond a reasonable doubt that appellant was not suffering from a mental disease or defect, in the legal sense, or at least that the killing was not the product of such mental illness.
The jury may have felt that appellant’s passive-aggressive personality did not so substantially affect her mental or emotional processes and impair her behavioral controls, that she should be absolved of criminal responsibility for her actions. The question for the jury requires the application to medical knowledge, and the lay evidence as well, of the understanding and judgment of the community as reflected in the jury. This question is not answered by establishing that appellant’s condition resulted in some limitation of her freedom of choice. Many criminologists point out that even normal human behavior is influenced by such factors as training, environment, poverty and the like, which may limit the understanding and options of the individual. In appellant’s case, even the psychiatrist’s determination that she was suffering from a mental illness reflected a process of drawing a line between a classification of mere personality disorder and a personality disorder which amounts to mental illness.3 Among mental illnesses recognized as such for medical purposes a line must be drawn between those which do and those which do not result in denial of criminal responsibility. The ultimate issue posed to the jury is whether an individual’s mental condition “substantially” affects mental or emotional processes and impairs behavioral controls, and if so whether the action is a product of such mental illness.
These broad considerations are accentuated in appellant’s case by the fact that the issue of criminal responsibility is overlaid with the problem of intoxication, since the psychiatrists considered that her recourse to alcohol was the crucial connecting link between her mental illness and aggressive conduct. The jury may have concluded there was not a sufficiently substantial .distinction between appellant and the not unfamiliar “normal” angry person who has a resentment toward authority figures and who acts out his anger more readily when alcohol removes his inhibitions. Dr. Vourlekis could not describe appellant’s mental illness as negativing freedom of choice to begin drinking. But the rule that negatives voluntary intoxication as a defense to crimes of general intent like manslaughter in effect holds men responsible for their fateful drinking, without regard to the extent of control at the moment of homicide.4
As to the issue of non-productivity, we have already noted that although Dr. Vourlekis testified that the killing was “probably” the product of appellant’s mental illness, Dr. Ruch testified that it was “possibly” the product. Even if Dr. Ruch meant that he entertained a reasonable doubt as to non-productivity, the jurors were not bound by the conclusions of the psychiatrists. The facts and assumptions underlying the expert opinions were amply exposed during their testimony, and the jury could draw different ultimate inferences.
Alternatively the jury may have disbelieved appellant’s account as to the extent of her drinking. It will be recalled *389that the psychiatrists whose testimony is relied on by appellant in turn relied on her statements to them for information as to the extent of her alcoholic consumption earlier in the fateful day. The jury may have given weight to the officer’s testimony that appellant appeared sober and answered questions clearly at the time of the arrest. This again is a question of degree, since the mother testified appellant was “half high.” In short the jury may have taken issue not so much with the psychiatrists’ reasoning as with the factual premise, as to the extent of appellant’s drinking, underlying their testimony on productivity.
There may be cases where the facts adduced as to the existence and impact of an accused’s mental condition may be so overwhelming as to require a judge to conclude that no reasonable juror could entertain a reasonable doubt.5 But in view of the complicated nature of the decision to be made — intertwining moral, legal, and medical judgments — it will require an unusually strong showing to induce us to reverse a conviction because the judge left the critical issue of criminal responsibility with the jury. Although the case is not free from doubt, we conclude reversal on this ground is not warranted.
II
However the nature of the psychiatric testimony, and the fact that this is a close case on the issue of insanity, are significant in impelling us to the conclusion that the interests of justice require a reversal because of the prosecutor’s actions, which we find to constitute misconduct, in the course of cross-examination and summation on the significant subject of organic damage.
The presence of some organic damage was noted in the pre-trial report of the Superintendent of St. Elizabeths which set forth the opinion that appellant was then, and at the time of the act, “suffering from a mental illness, namely, 51.1 Passive Aggressive Personality (Organic and Alcoholic Features).” On direct, Dr. Ruch, the Assistant Clinical Director of the Dix Pavilion of St. Elizabeths, testified that organic brain injury or disease had been determined on the basis of psychological testing, even though skull X-ray, EEG, and neurological examinations had proven negative:
Q Will you tell us what the purpose of the skull X-rays and electroencephalogram were?
A These were to rule out any sort of organic findings.
Q Now what were the results of the skull X-rays and the electroencephalogram? Did you find any organic brain injury or disease?
A Not by skull X-rays or by electroencephalogram. We did, however, on psychological testing. [Tr. 72-73]
******
Also I should state that along this line our psychological tests pointed toward a moderately severe organic disturbance, probably secondary from the prolonged use of alcohol. This does not show up too much qn clinical examination, but did show up on psychological tests. That is to say her IQ is lowered. Her IQ, fullscale IQ is 69. She was unable to subtract, or refused to subtract seven from a hundred, was unable to give much information about current events, wouldn’t even estimate the distance between Los Angeles and Washington, for instance. Putting all this together we arrived at a diagnosis, that diagonsis being passive-aggressive personality. [Tr. 74-75]
*390Notwithstanding this testimony, the prosecutor persistently drummed into the jury — without evidentiary basis, and contrary to the uncontradicted testimony of the Government psychiatrists called by the defense — the assertion that organic damage was negatived by the failure to detect it by physical tests, and that psychological tests could not establish organic damage (“organieity”). Also improper and indeed glaring was his representation — or rather, misrepresentation —that Dr. Ruch agreed that organic damage had been negatived. We quote from the transcript at some and perhaps excessive length, though other references abound, with italics for passages that illuminate the nature and impact of the misconduct. We may assume that the prosecutor’s actions reflected his zeal and the excitement of the trial rather than any deliberate intent to impose upon the jury. Yet our reading of the transcript is such as to convince us that the prosecutor stepped out of bounds, that the impact of this plain error, in the context of a close case, was probably so prejudicial that our judicial conscience calls upon us to reverse and remand for a new trial that can be conducted free of similar error.
The prosecutor’s cross-examination of Dr. Ruch included these passages:
Q Now organically there was nothing wrong with this person; isn’t that correct, sir?
A Now, by organically, the tests for organieity are specifically, the tests that we do are, one, skull X-rays, which would show up something like a brain tumor, or something like that.
Q It didn’t show it in this case?
A Showed nothing.
Q The e.e.g. was negative?
A The electroencephalogram of the brain was normal.
Q Normal?
A Now the other tests we do for or-ganicity are the psychological tests.
Q Other than the psychological tests, the neurological test, the neurological examination was negative for organieity, wasn’t it?
A The neurological examination was, yes.
Q Now, are you saying that psychological tests are tests that are given by psychiatrists for the purpose of determining organieity ?
******
A I’m saying that psychological tests are tests given by psychologists for determination of organieity and extent of impairment.
Q But you as a psychiatrist, in order to be able to arrive at a determination whether a person is suffering from a mental illness or disease, you conduct or have conducted organic examinations to determine whether there are any structural defects of the brain, don’t you, as a psychiatrist?
A We do. The psychiatrist admitting a patient does a neurological examination, and also the neurologist at the hospital performs a neurological examination.
Q Let’s put it this way. There are only two ways, aren’t there, Doctor, that you as a psychiatrist can arrive at an opinion as to whether a person is suffering from a mental illness or disease or defect and that is to give organic examinations to determine whether a person is suffering from any structural defect, isn’t that correct, and if, having given those examinations and organically you find nothing wrong, then, you proceed to the functional or the behavior examination to arrive at an opinion as to whether a person is suffering from a mental illness or disease, don’t you ?
A I don’t quite understand your question. Let me elaborate on this, if you will.
Q Well, let me put the question a different way then if you don’t understand it.
*391As a psychiatrist you conduct or give or have examinations given organically to determine whether there is anything • organically wrong with the person, don’t you, to enable you to—
A You are speaking now physically wrong?
Q Neurologically, e.e.g.’s and skull X-rays — do you not determine from organic examinations wherein you find structural defects, can you not as a psychiatrist arrive at an opinion as to whether a person is suffering from a mental illness, disease or defect — can’t you ?
A Well, organic disease — as I indicated before, the only way that I can answer this question is this. The tests that are done at the hospital in reference to the presence or absence of organic brain damage are those that I have mentioned.
Q All right.
A Skull X-rays, neurological examination, electroencephalogram and psychological tests. [Tr. 86-88]
Q Now these organic examinations were given at the hospital and there was nothing structurally wrong with this person and you arrived, as you say, at an opinion that she was not suffering from any mental illness or disease by virtue of organicity? Isn’t that correct?
A No.
Q She wasn’t?
A I indicated that psychological tests—
Q I’m not talking about the psychological tests.
[Defense Counsel] : Your Honor, he’s arguing.
[Prosecuting Attorney] : No, I think I can cross-examine my witness, if your Honor please.
The Court: Go ahead.
By [Prosecuting Attorney] :
Q We will come back to the psychological tests later. Now on these tests, the skull X-rays, the e.e.g., and neurological indicate there was nothing organically wrong which would enable you as an expert to arrive at an opinion as to whether she was suffering from a mental illness or disease; isn’t that right.
A On the basis of the tests which I mentioned ?
Q Right. Well then when you find nothing organically wrong with an individual — there are no structural defects of the brain — then you proceed to a functional or a behavior examination, don’t you?
A Yes.
Q If you find nothing organically wrong, then proceeding to the functional or the behavior examination, from that functional or behavior examination you, as a psychiatrist, might be able to arrive, or are able to arrive at an opinion as to whether the person is suffering from a mental illness or disease or defect — right?
A Yes.
Q All right. So we eliminate organ-icity. Did you arrive now at this opinion that you arrived at that she was suffering from this, what, passive-aggressive personality, by virtue of the functional or behavior aspect of the psychiatric examination ?
A Now, on the basis of — let me state here that on the basis of mental status examination one also can determine, or get some idea as to the presence or absence of organicity. This is carried out by the psychiatrist. I imagine you are referring to this as being a behavioral or a functional examination. However, it can — it’s not something that we look at by X-ray, but may be organicity. [Tr. 89-91]
******
*392The following approach was employed during cross-examination of Dr. Vourle-kis:
Q Did I understand you correctly, sir, when you stated that she is suffering from a mental illness, this defendant is, of passive aggressive personality, organic? Is that what you said?
A Passive aggressive personality disorder. However, there are organic factors superimposed subsequent to alcoholic consumption.
Q Is this passive aggressive personality due to organicity?
A No, sir.
Q All right. You are not saying that there is anything organiely [sic] wrong with her, right?
A I am not saying that, you say?
Q Yes.
A I am.
Q What is organically wrong with her?
A Alcohol, after extensive use over a period of years, has an effect of destroying the brain.
Q All right. Now, were skull X-rays taken of this individual?
A Yes, sir.
Q And there was nothing wrong with the brain, was there?
A They were normal.
Q All right. Electroencephalograms were taken, right?
A Yes, sir.
Q Nothing wrong with the brain, right?
A The electroencephalogram was normal.
Q It was normal. And the neurologi-cals of the individual were normal, right?
A Yes, sir.
Q So there is nothing organically wrong with the person, is there ?
A May I point out one additional test, the psychological—
Q We will come to the psychological, but to arrive at an opinion as á psychiatrist, a person who is suffering from a mental illness, disease, or defect, you give these organic examinations to determine whether there is structural defects of the brain, don’t you ?
A Yes.
Q And all these tests were negative, right?
A Yes.
Q And so, therefore, do you not now on the basis of these tests which are negative, eliminate organicity?
A I would like to point out—
Q Do you not eliminate organicity on the basis that all these tests are negative, Doctor?
A I do not eliminate organicity on the basis of those tests.
Q Then do you disagree with Dr. Ruch when he says there is nothing organically wrong?
[Defense Counsel] : I object, Your Honor, that was not his testimony.
[Prosecuting Attorney] : It is up to the jury, Your Honor.
[Defense Counsel] : Well, I don’t think he can predicate a question to him based on what is allegedly the evidence if in fact that is not the evidence.
The Court: We will have to rely on the jury’s recollection of the testimony.
[Prosecuting Attorney] : That is my recollection, Your Honor.
The Court : Try to keep your questions within what has been testified to.
[Prosecuting Attorney] : Yes, Your Honor, I do nothing deliberately that is not in evidence, if Your Honor please. That is my recollection and I stand corrected if it is not the evidence. I try to give a fair interpretation of my recollection.
*393The Court: I want to warn the jury that the remarks of counsel are not evidence in the case. Take your evidence from the witnesses. By [Prosecuting Attorney] :
Q Do you disagree with Dr. Ruch? It is my recollection he testified that there was nothing — not taking the psychological into consideration 5a that there was nothing organically wrong with this individual.
A As far as the tests you mentioned are concerned, they did not show any organieity. [Tr. 128-131]
It will be noted how the prosecutor added to his initial proposition, that organic damage is negatived unless it appears on the physical clinical tests, the assertion that Dr. Ruch agreed there was no organic damage. Both points were driven home in the prosecutor’s summation, as appears from the following:
As has come out on cross-examination, any psychiatrist worth his salt— to use a colloquial expression, members of the jury — must necessarily admit that a psychiatrist conducts organic examinations to determine whether there are any structural defects of the brain and from which structural defects of the brain they, as psychiatrists, can come to a conclusion whether a person is suffering from any mental illness, disease or defect. In this case, they were conducted.
Skull X-rays were given which were negative. An electroencephalogram was given which was negative, and neurologicals were given which were negative. And he himself, as clinical director of the Dix Service, I believe he said, members of the fury, that there was nothing organically wrong with the person. [Tr. 233]
* * * * * *
Then as a psychiatrist, he had to admit when there is nothing wrong organically, members of the jury, then you must necessarily come to a conclusion if a person is suffering from a mental illness or disease by what is known as the functional behavior. And we had some time with that individual, as you recall it, who is an expert in the field of psychiatry. I am not an expert in psychiatry. That is not my field at all.
After some questioning, members of the jury, then he does recognize that the other aspect of a psychiatric examination is the functional behavior. When there is nothing organically wrong, members of the jury, when there is nothing wrong with the structural or working tools of the brain which would cause an irrational act on the part of an individual, then you have to turn to the behavior of the person to determine whether the acts of the person, members of the jury, are rational or not, whether they deviate from normal actions as you reasonable men and women can recognize whether the actions are normal or not. [Tr. 234-35]
*394We find error both in the prosecutor’s reiterated assertions — without eviden-tiary foundation — that nothing organically wrong had been found, and in his assertions that Dr. Ruch testified he had found no organic damage, plainly contrary to the actual testimony given.
It is elementary that a prosecutor may not import his own testimony into a criminal trial.6 The doctrine has full vitality not only where the prosecutor is asserting a fact within his individual cognizance, but also where, as here, the prosecutor is asserting a belief or opinion that is properly the subject of expert testimony. The prosecutor is not free to offer his opinion in lieu of calling an expert witness. It compounds the mischief that here his belief was contrary to the testimony of the duly qualified expert witnesses.
The prosecutor’s transgression would have been established beyond doubt if he had said in so many words: I believe, or in my opinion, or I can assure the jury — that there is nothing organically wrong with a person claiming mental illness unless that is revealed by a physical examination. In this case the prosecutor’s assertions were made more subtly, but with comparable effectiveness, when presented as comments made in passing, or preliminary to another question, as though they had been demonstrated by previous psychiatric testimony. It is misconduct for a prosecutor to make an assertion to the jury of a fact, put in the form of an assumption of a question, unless the prosecutor proffers direct testimony of that fact.7 The same principle applies to a matter of opinion, and is certainly applicable where as here, the prosecutor’s opinion contradicts the expert opinion of the witnesses called by the defense and he never offers opinion evidence to the contrary.
It may be that the prosecuting attorney sincerely believed in the proposition that when organic impairment of the brain is not shown by physical examinations a conclusion of organic damage may not be based on psychological tests. However, that is a proposition of medicine and cannot be asserted by the prosecutor to the jury unless it has some evi-dentiary foundation.
The prosecutor’s conduct cannot be justified on the ground that he was asking the jury to override the testimony of the defense experts as incredible. The fact is, as discussed below, that he expressly purported to be relying on those experts. And part of the vice of his unsupported general assertions was a tendency to lead a jury to suppose that these were in accordance with the premises of the experts. We see nothing incredible in the testimony of the St. Elizabeths psychiatrists. For all we know not a single responsible psychiatrist would contradict their assertion that psychological tests may indicate organic damage not revealed by physical examination — at least where, as here, the psychological test is assessed in the light of a case history showing regression from earlier test results as well as a history of drinking. The prosecutor did not even confront the witnesses with a divergent view expressed in a medical text.8
*395The mischief of the unsupported assertions was compounded by the prejudice from the prosecutor’s misbehavior in invoking and at the same time misrepresenting the testimony of Dr. Ruch — as appears both in his questioning of Dr. Vourlekis and in his summation to the jury. The two strains of misconduct are interrelated though analytically distinguishable.
Allowances must be made for an attorney, even Government counsel, who out of haste or confusion misunderstands the substance of the previous testimony. An honest misinterpretation of testimony without some indication of prejudice does not constitute reversible error.9
On a nerve-center issue we are more likely to hold the prosecutor to account. Cross v. United States, 122 U.S.App.D.C. 283, 353 F.2d 454 (1965). The prejudice digs in and holds on. As Judge Boreman pointed out in Wallace v. United States, 281 F.2d 656, 667-668 (4th Cir. 1960):
On at least six occasions during the trial and in closing argument, the United States Attorney misquoted the language [of a Government witness].We cannot agree with the Government that the prejudicial effect of such misquoting was removed by the United States Attorney and the trial judge telling the jurors that it was for them to recall the actual statement. Stewart v. United States, 1957, 101 U.S. App.D.C. 51, 247 F.2d 42, 47. These repeated and seemingly deliberate misquotes, * * * may well have become so firmly implanted on the jurors’ minds as to cloud the actual testimony.
We are not required nor do we undertake to determine whether the misconduct of the prosecutor was deliberate. We do note the care with which the prosecutor phrased his questions so as to defer, or set to one side, the subject of the conclusion drawn from psychological tests. Would that he had shown similar care in recalling the actual testimony of the psychiatrist — that psychological tests are given to determine organic damage, and that they showed organic damage in the case of appellant.
In any event, the probability of prejudice is clear. As we have already noted, the defense case on mental illness was strong. It would be difficult to conclude that the jury was not substantially, perhaps decisively, influenced by the prosecutor’s persistent but unwarranted assertions, and by his misstatements. Whether or not organic damage vel non would seem crucial to the initiate, the prosecutor obviously felt that his assertions would loom large in the jury’s deliberation. It seems likely that the verdict was stained by the misconduct.
Of course, corrective and clarifying measures were undertaken by defense counsel, and by the witnesses, but they could hardly catch up to, let alone overtake, the misstatements. Thus, defense counsel’s objection attempted to correct the prosecution’s distortion of Dr. Ruch’s testimony. This was remitted to the best recollection of the jury. On redirect defense counsel elicited testimony from both psychiatrists, reverting to the clarifying testimony they had managed to insert in the course of cross-examination. But it would have been well nigh impossible to dissipate the improper assertions —so prolonged and pervasive was the pattern. Sometimes, as the foregoing account reveals, the witness answered the prosecutor’s question without reiterating his issue with the prosecutor’s unwarranted preliminary assertions — perhaps because he assumed, with some reason, that his primary obligation was to focus on the ultimate question, without skirmishing on seemingly introductory matters, or perhaps because he was concentrating on the ultimate question and did not take particular note of the preamble. There were some timely correc*396tions by defense counsel. A prosecutor cannot fairly put defense counsel to the task of correcting a persistent course of assertion and misrepresentation, and thereby shoulder the risk that the jury might gain an impression of defense tactics as fly-specking and dilatory. Perhaps the situation could have been saved if the judge had stepped in, on his own initiative as well as on protest of defense counsel. A trial judge is understandably disinclined to depart from a more passive role as arbiter of an adversarial conflict, lest he perhaps unwittingly interfere with effective development of a case. But a cautionary warning in time to the prosecuting attorney, followed by control if necessary, would help both the defendant whose rights are protected directly and the Government whose case will otherwise run aground in the shoals of the prosecutor’s conduct.
In this case the District Judge did give the jury the customary caution not to rely on statements of counsel, but to rely solely on the testimony and their own recollections thereof. This instruction is, of course, appropriate, and in another case might help avoid reversal. But where the case is close, prejudice cannot be avoided by “mild judicial action,” and reversal is necessary to obviate the consequence of prosecutor’s persistent departure from his “duty to refrain from improper methods” to obtain conviction. Berger v. United States, 295 U.S. 78, 85, 88, 55 S.Ct. 629, 633, 79 L.Ed. 1314 (1935).
As the Supreme Court noted in Berger, 295 U.S. at 88, 55 S.Ct. 629, the particular vice of misconduct on the part of a prosecutor is the confidence of the average jury that the prosecuting attorney is faithfully observing his obligation as the representative, not of any ordinary party to a controversy, but of an impartial sovereign, whose interest “in a criminal prosecution is not that it shall win a case, but that justice shall be done.”10 The jury’s confidence may be at zenith on insanity issues for it knows that the prosecutor need not be concerned with precipitate release of a person dangerous to the community. A verdict of “not guilty by reason of insanity,” accompanied as it is by post-trial commitment, is a verdict of detention — in some cases longer than the maximum detention available on conviction. Compare Rouse v. Cameron,-U.S.App.D.C. -, 373 F.2d 451 (decided Oct. 10, 1966). Such a verdict is in some measure not unlike the Scottish verdict of guilty but insane, for it “presupposes a determination that the accused has committed the criminal act with which he is charged” and has no defense other than insanity. Lynch v. Overholser, 369 U.S. 705, 714, 82 S.Ct. 1063, 8 L.Ed.2d 211 (1962); Rucker ,v. United States, 108 U.S.App.D.C. 75, 280 F.2d 623 (1960).
The prosecutor may strike hard but not foul blows. On the subject of mental illness, we expect him to probe the conclusions even of Government psychiatrists on the merits, and to lay bare underlying assumptions. See Carter v. United States, 102 U.S.App.D.C. 227, 233-234, 252 F.2d 608, 614-615 (1957). In certain respects the prosecuting attorney’s cross-examination, though vigorous, was in accord with this approach.
However, the misconduct related above is so persistent and prejudicial as to compel us to take note of the misconduct as plain and prejudicial error. Fed.R.Crim.P. 52(b). We have already underscored the readiness of the courts to remit the issue of mental illness to the jury, save in exceptional cases. There is a correlative need for a trial which promotes the jury’s fair consideration of the merits of the issue without deflection by the prosecuting attorney on crucial aspects of the case. The issue is not detention or freedom but rather what kind *397of detention and treatment of the accused best furthers the public interest.
Ill
There are other aspects of the case which concern us. Without indicating whether they would independently warrant reversal, they reinforce our conclusion that the jury was impeded from providing the kind of consideration of the merits of the insanity issue to which appellant is entitled.
For example, the prosecutor inquired of Dr. Ruch whether appellant, with an I.Q. of 69, was an idiot, imbecile or moron. Answering in the negative, Dr. Ruch explained that such terms usually connote a mental defect persisting since childhood, whereas in appellant’s case the significant point is that there was a later decline from an earlier higher I.Q., probably on the basis of protracted use of alcohol and organic brain damage. On summation the prosecutor said that Dr. Ruch relied on the fact that appellant’s I. Q. was 69, and 70 was normal. This may have effectively misled the jury into disregarding the testimony of Dr. Ruch, whose conclusions were based in part on the I.Q. pattern. This came close to if it did not constitute distortion by selective omission, since Dr. Ruch focused not on the 69 level but on the significance, uncontradicted at trial, of the reduction from earlier higher levels. Since this was a basis for finding organic damage, the prosecutor’s tactic here interrelates with the misconduct already noted.
Less significant, perhaps, but not inconsequential is the prosecutor’s reiteration that neither of the Government psychiatrists was a diplómate of the American Board of Neurology and Psychiatry, not having completed the requisite five years of psychiatric training and experience. Assuming that the prosecutor acted properly in eliciting testimony defining the witnesses’ qualifications, we see no warrant in this case for his stressing to the jury, through summation, that the testifying psychiatrists were not diplomates where there was no contrary psychiatric testimony. The argument carried the implication that a more experienced expert might or would have reached a different conclusion. If there was any basis for such an implication it should have been adduced in the form of testimony presented by the Government, which had the burden of proof.11
In summation the prosecutor challenged the psychiatrists’ opinions as follows : “And when we asked the individuals was she able to distinguish between right and wrong, you know how hesitant they were to give an answer to that particular .question, particularly the last one.”
An expert may testify on the defendant’s capacity to distinguish between right and wrong, but this court has carefully noted that “[a]n expert may not be compelled to testify in these terms if he believes they are essentially moral or legal considerations beyond the scope of his special competence as a behavioral scientist.” McDonald v. United States, 114 U.S.App.D.C. 120, 124 n.9, 312 F.2d 847, 851 n.9 (1962) (en banc). The prosecutor’s action was improper. Indeed the defendant would have been entitled to an express instruction from the court advising that the expert did not have to answer that question, and that the jury could draw no negative inferences from his failure to do so, and secondly, that ability to distinguish right from wrong suggests only the absence of a symptom “which medical science has long recognized do[es] not necessarily, or even typically, accompany even the most serious mental disorder.” Durham v. United States, 94 U.S.App.D.C. 228, 242, 214 F.2d 862, 876, 45 A.L.R.2d 1430 (1954). As we have recently pointed out, in a case where the question of whether the defendant could distinguish right from wrong was introduced at trial, “the judge should also *398have instructed * * * [the jury] that, even if they believed that defendant did know right from wrong, they could still find, on the basis of other evidence, that defendant’s alleged act was a product of a mental disease or defect, which, of course, is the only ultimate test of criminal responsibility in this Circuit.” Blocker v. United States, 116 U.S.App.D.C. 78, 80, 320 F.2d 800, 802, cert. denied, 375 U.S. 923, 84 S.Ct. 269, 11 L.Ed.2d 167 (1963). (Footnote omitted.)
Finally we note that the court instructed the jury on drunkenness only as this was related to second-degree murder, and manslaughter. Intoxication as an element in the insanity defense was not mentioned, even though it is clear that the effect of alcohol on this defendant was crucial to her insanity defense. The thrust of the court’s instruction was aimed at suggesting that intoxication was not a defense to a general intent crime, and that the jury could not consider her intoxicated condition in deciding whether she acted in the heat of passion. On retrial it would be better, to obviate any possible misunderstanding, for the court to advise the jury explicitly that although voluntary intoxication is not a defense to the crime, the jury may give consideration to appellant’s claim of intoxication, together with the testimony of the psychiatrists concerning the interaction of appellant’s recourse to alcohol and her mental condition, in the course of determining the issue raised by the insanity defense, i. e., whether the Government had established beyond a reasonable doubt the elements of appellant’s criminal responsibility for the stabbing of her brother.12
Reversed and remanded.
. Dr. Vourlekis stated on cross-examination that at the diagnostic conference there had been some range in the opinions expressed as to the severity and nature of appellant’s disturbance — ranging from the ultimate and official hospital opinion to an opinion of “inadequate personality” (the record not showing who expressed that opinion).
. See, e. g., Hawkins v. United States, 114 U.S.App.D.C. 44, 310 F.2d 849 (1962). See also Carter v. United States, 102 U.S.App.D.C. 227, 252 F.2d 608 (1957).
. The jury may have given emphasis to the consideration that the symptoms presented by the psychiatrists as indicative of appellant’s mental illness tended to differ in degree rather than kind from the conduct of normal individuals. Symptoms of appellant’s condition as noted by the psychiatrists were passivity; reclusiveness; tendency to stay to herself; over-cooperativeness and tendency to be “too pleasant”; “circumstantial” speech, i. e., speech relating to events, not feelings; a “farawayness,” though without hallucination.
. As to the relation of intoxication to the defense of insanity, see the concluding paragraph of this opinion.
. See, e. g., Isaac v. United States, 109 U.S.App.D.C. 34, 284 F.2d 168 (1960); Hartford v. United States, 362 F.2d 63 (9th Cir. 1966) (where the evidence was held to require a verdict of not guilty by reason of insanity notwithstanding defendant’s testimony that he was sane, and the application, as the law of the case, of a more restrictive standard of insanity than that established by this court in McDonald).
. The United States Attorney suggests that this phrase sufficiently corrected any mis-statement by the prosecutor, and clarified the presentation to the jury sufficiently to avoid any inference of reversible error.
We take a diametric view of the significance of this phrase — which was inadvertently omitted from the opinion on first issue. The prosecutor’s insertion of the phrase modified the wholly inaccurate reference in his prior question objected to by defense counsel. By itself, it might well have been within the latitude of cross-examination. But it hardly served the function of providing a portrayal of Dr. Ruch’s testimony that was accurate over all, since it left the impression that somewhere along the way Dr. Ruch had testified there was nothing organically wrong with appellant.
The deeper significance of the inclusion of this phrase in the prosecutor’s questioning is its indication of the prosecutor’s awareness that Dr. Ruch had relied on the psychological — and found brain damage. This underscores the problem presented to the court by the prosecutor’s summation — when he told the jury, without any qualification, that Dr. Ruch said there was nothing organically wrong with appellant.
. Berger v. United States, 295 U.S. 78, 88, 55 S.Ct. 629, 79 L.Ed. 1314 (1935); Willie Lee) Stewart v. United States, 101 U.S.App.D.C. 51, 54-56, 247 F.2d 42, 45-47 (1957) (en banc); United States v. Spangelet, 258 F.2d 338, 342-343 (2d Cir. 1958).
The dissent in Stewart v. United States, supra, was based on the approach that in that case there was evidence, including inferences, to support the belief of the prosecutor and that the statements were not made repeatedly or dwelled on by the prosecutor. In the case at bar there is
no evidence in support of the repeated assertions of the prosecutor.
. See Callanan v. United States, 223 F.2d 171, 179 (8th Cir.), cert. denied, 350 U.S. 862, 76 S.Ct. 102, 100 L.Ed. 764 (1955) ; Richardson v. United States, 150 F.2d 58, 64 (6th Cir. 1945); 6 Wigmore, Evidence § 1808 (3d ed. 1940).
. Compare Reilly v. Pinkus, 338 U.S. 269, 275, 70 S.Ct. 110, 94 L.Ed. 63 (1949); Abrams v. Gordon, 107 U.S.App.D.C. 254, 276 F.2d 500 (1960). Such cross-examination would depend on a prelim*395inary determination by the court that the text was sufficiently authoritative to be used for this purpose.
. Nicholson v. United States, 221 F.2d 281 (8th Cir. 1955).
. In the Rotunda outside the Office of the Attorney General of the United States appears the formulation used in a brief filed in the Supreme Court by former Solicitor General Frederick William Leh-maim: “The United States wins its point whenever Justice is done its citizens in the courts.” See Frankfurter, The Government Lawyer, 18 Fed.B.J. 24, 27-29 (1958).
. We do not now consider the problem that may arise in a case where the Government impliedly challenges as inadequate the psychiatrists furnished to defendant.
. As already noted, these elements are sot forth in McDonald v. United States, supra, and need not be recapitulated here.