United States v. James E. McCracken

GOLDBERG, Circuit Judge: This is a direct appeal from a conviction for first degree murder. On August 18, 1971, James E. McCracken, then a fifty-six year old domiciliary patient at the Veterans Administration Hospital in Biloxi, Mississippi, shot and killed Dr. Hugh B. McGill, a member of the medical staff at the facility. Using the same .38 caliber pistol, he then shot and wounded himself. He made no attempt to escape or to conceal or deny his actions. Following his indictment for first degree murder, 1 McCracken was sent to the U.S. Medical Center for Federal Prisoners, Springfield, Missouri, pursuant to 18 U.S.C. § 4244 (1970), for examination into both his mental health at the time of the shooting and his present ability to aid in his own defense. Based on the report of the Springfield staff, the district court entered an order finding McCracken presently sane and mentally competent to stand trial for the charges against him. He was tried before a jury in the Southern District of Mississippi. His only defense was that at the time of the offense he was not legally sane. The jury rejected the insanity defense, and McCracken was convicted of first degree murder and sentenced to life imprisonment. On appeal McCracken specifies seven points of error, which raise three basic issues. First, was the Government’s evidence sufficient to make an issue for the jury on insanity and to sustain the jury’s conclusion .that McCracken was legally sane at the time of the offense [points 1, 2, and 7]; second, were the instructions relating to the definition of insanity so repetitious as to be prejudicial [point 4]; and third, did the trial court err in instructing the jury that if found “not guilty by reason of insanity,” the defendant would be “turned a loose” [points 3, 5, and 6]. We cannot agree with appellant that on the basis of the record as a whole the Government’s *409 evidence was insufficient as a matter of law to avoid a directed verdict on the issue of insanity or to sustain the jury’s verdict. The second issue is also without merit. As to the third issue, however, we have concluded that the trial judge committed harmful and substantial error, and we reverse. I. INSANITY DEFENSE Evaluating the sufficiency of the evidence of a defendant’s sanity is by no means an easy task, at either the trial or the appellate level. Nevertheless, the parameters by which the determination must be made are by now well-settled in this Circuit. 2 The defendant’s sanity is always an element of the offense charged. When no question of insanity is raised, the Government’s burden of proving sanity is satisfied by the so-called “presumption of sanity,” which stands in the place of evidence. When, however, “some” evidence is received, establishing the hypothesis of insanity, the burden is on the Government to prove beyond a reasonable doubt that the defendant was sane at the time of the offense. Blake v. United States, 5 Cir. 1969, 407 F.2d 908, 911, 912; Brock v. United States, 5 Cir. 1967, 387 F.2d 254, 257; Mims v. United States, 5 Cir. 1967, 375 F.2d 135, 140. Once raised, the question of the defendant’s sanity will normally be for the trier of fact to resolve, but at three different points the court may face the question of the sufficiency of the evidence as a matter of law. Initially, of course, the court must always determine whether the defendant has presented enough evidence to put his sanity in question. See United States v. Holt, 5 Cir. 1971, 450 F.2d 868. The court must also determine whether the Government’s evidence is sufficient to make an issue for the jury on the defense of insanity and thus avoid a directed verdict of acquittal. Finally, in the event of a guilty verdict, the court must decide whether the Government’s evidence is sufficient to support the conclusion of the trier of facts as to the defendant’s sanity at the time of the offense. United States v. Collier, 5 Cir. 1972, 453 F.2d 1173, 1176-1177; Gordon v. United States, 5 Cir. (1971) 438 F.2d 858, 885, cert. denied, 1971, 404 U.S. 828, 92 S.Ct. 63, 30 L.Ed.2d 56; Blake v. United States, supra, 407 F.2d at 911. This court has never precisely defined the quantum of evidence necessary to constitute sufficiency for each of these determinations, though the level obviously rises from the first to the third. Indeed exact quantification, even if possible, would probably be undesirable. Instead, through the painful process of case-by-case adjudication we have educed a set of general principles to guide both the trial court in exercising its role of weighing the evidence as a matter of law and the appellate court in reviewing that exercise. Foremost among these principles is the realization that “each [case] must be decided upon its own facts with careful attention to the weight of the evidence on each side.” Nagell v. United States, 5 Cir. 1968, 392 F.2d 934, 937. Applying these principles to the case sub judice, we have concluded that appellant’s challenges to the sufficiency of the evidence must fail. The defense theory was that Mc-Cracken suffered from psychomotor epilepsy, developed traumatically during World War II from a gun blast in the face, from malaria, or from both, which caused him to suffer frequent seizures, especially at times of stress, manifested by tearing, salivation, and jerking or jabbing movements of the right hand. *410 Acts of violence might be committed during the seizures, and McCracken would have no subsequent memory of such conduct. According to the theory, McCracken was bitter about his failure to receive a service-connected disability allowance. When he went to Dr. McGill, hoping a threat would produce action on his application, he suffered a seizure that caused the unintentional shooting. In support of this theory the defense presented the expert testimony of Dr. Morris H. Lampert, a Texas neurologist who had been treating McCracken for epileptic disorders for ten years. Dr. Lampert testified about the general characteristics and effects of psycho-motor epilepsy, including seizures, automatism, aggressive or hostile behavior, alteration in personality, and amnesia. He also testified concerning the defendant’s own condition, concluding that though treatable, McCracken was probably not curable. In addition, a number of lay witnesses — relatives, fellow patients and attendants at the V.A. hospital, and two jailers — described in detail McCracken’s actions during various observed seizures over the years. On a number of occasions this court has said that it takes only a slight quantum of evidence to raise the issue of insanity. E. g., United States v. Collier, supra, 453 F.2d at 1176; Gordon v. United States, supra, 438 F.2d at 885. The very strong case presented by the defense was certainly more than sufficient to this end, and placed the burden of proving sanity beyond a reasonable doubt on the Government. In determining whether the Government met its burden at least enough to warrant submission of the question of insanity to the jury, both the amount and kind of evidence presented by the Government are important. As we noted in Brock v. United States, supra, 387 F.2d at 258, however, “the nature and quantum of rebuttal evidence sufficient to present a jury question is to some degree determined by the strength of the case for insanity.” Thus it is possible for the Government to meet its burden without using expert testimony on the sanity question. In Mims v. United States, supra, 375 F.2d at 143-144, we outlined techniques by which the Government might meet its burden by discrediting the defense experts : [E] xpert opinion evidence may be rebutted by showing the incorrectness or inadequacy of the factual assumptions on which the opinion is based, “the reasoning by which he progresses from his material to his conclusion,” the interest or bias of the expert, inconsistencies or contradictions in his testimony as to material matters, material variations between the experts themselves, and defendant's lack of co-operation with the expert. Also, in cases involving opinions of medical experts, the probative force of that character of testimony is lessened where it is predicated on subjective symptoms, or where it is based on narrative statements to the expert as to past events not in evidence at the trial. In some cases, the cross examination of the expert may be such as to justify the trier of facts in not being convinced by him. No more than a brief review of the Government’s case is necessary for us to conclude that the trial court did not err in denying Mc-Cracken’s motion for directed verdict of acquittal and in submitting the question of insanity to the jury. The Government presented one expert witness, Dr. H. B. Fain, the Acting Chief of Psychiatry at the United States Medical Center for Federal Prisoners at Springfield, Missouri. Dr. Fain testified concerning the examination made about one month after the shooting to determine Mc-Cracken’s competency to stand trial. He included his own opinion about Mc-Cracken’s possible sanity at the time of the offense. 3 From the absence of re- *411 sidual evidence, Dr. Fain concluded that McCracken did not lack substantial capacity to appreciate the wrongfulness of his conduct or to conform his conduct to the requirements of law at the time of the offense. Dr. Fain did acknowledge that he would not know appellant’s mental condition on the day of the shooting if McCracken had in fact suffered a seizure at the time of the offense. The Government also presented lay witnesses to support its theory that McCracken had deliberately planned the confrontation with the goal of either achieving a satisfactory settlement of the disputed claim or killing Dr. McGill and himself. 4 Under the evidence presented by both sides, the crucial question for the jury was whether the shooting of Dr. McGill occurred as a result of an epileptic seizure by McCracken. In rejecting the insanity defense the jury evidently concluded that it had not. In reviewing the sufficiency of the evidence to support this result, we are conscious of the respected and revered admonition in Glasser v. United States, 1942, 315 U. *412 S. 60, 80, 62 S.Ct. 457, 469, 86 L.Ed. 680, 704, that “the verdict of a jury must be sustained if there is substantial evidence, taking the view most favorable to the Government, to support it” . . Lacaze v. United States, 5 Cir. 1968, 391 F.2d 516, 519. In this context “substantial evidence” means “evidence that a reasonably minded jury could accept as adequate and sufficient to support the conclusion of the defendant's guilt beyond a reasonable doubt.” United States v. Harper, 5 Cir. 1971, 450 F.2d 1032, 1040. Stated another way the test is whether “a reasonable mind must necessarily have had a reasonable doubt as to [the accused’s] legal sanity at the time of the offense.” United States v. Alvarez, 5 Cir. 1972, 458 F.2d 1343, 1344. When passing on the question of insanity, the jury should look at all the evidence, not just the opinion of the experts, though expert opinion evidence cannot be arbitrarily ignored. United States v. Harper, supra, 450 F.2d at 1042; Mims v. United States, supra, 375 F.2d at 143. The fact finder “need not be bound by expert testimony even if all the witnesses are presented by only one side,” United States v. Pitts, 5 Cir. 1970, 428 F.2d 534, 536, cert. denied, 1970, 400 U.S. 910, 91 S.Ct. 154, 27 L.Ed.2d 149, for as in other cases, questions of the credibility and weight of expert opinion testimony are for the trier of the facts, and . . . such testimony is ordinarily not conclusive even where it is uncontradicted. . . . [E]xpert opinion as to insanity rises no higher than the reasons upon which it is based. . United States v. Alvarez, supra, 458 F. 2d at 1344 (emphasis in the original). See also United States v. Collier, supra, 453 F.2d at 1176; Nagell v. United States, supra. Especially where there are “material variations between the experts themselves,” the fact finder need not be bound by defense expert testimony. United States v. O’Neal, 5 Cir. 1970, 431 F.2d 695, 696, cert. denied, 1971, 401 U.S. 917, 91 S.Ct. 898, 27 L. Ed.2d 818. This Court has not in the past been hesitant to reverse insanity cases because of insufficient evidence to support a verdict against the accused. We recognize that each case must be decided on its own “particular and peculiar facts,” United States v. Collier, supra, 453 F.2d at 1177, rather than on its supposed position along some imaginary spectrum of cases. Yet we feel constrained to mention that when in the past we found the evidence insufficient to sustain a guilty verdict, the Government had usually made a less complete effort on the insanity issue than in the case at bar. 5 Dr. Fain’s testimony certainly created a battle of experts for the jury to consider. ' Complementing this expert testimony was the testimony from various lay witnesses, apparently intended both to support the Government’s premeditation theory and to refute partially the defense expert’s testimony. Witnesses testified that McCracken had encoun *413 tered many difficulties with the Veteran’s Administration and believed he was being unfairly deprived of benefits. The day before the shooting, he traded a previously acquired pistol for the murder weapon. He announced his intention of settling his claim with “the clincher,” and according to a nurse, immediately after the shooting said, “I told you I would do it.” Other witnesses also saw McCracken shortly after the shooting and noticed none of the physical signs that usually accompanied the seizures. We do not mean to minimize or depreciate the evidence presented by the defense. It seems clear to us, as it no doubt did to the jury, that appellant is prone to epileptic seizures and in need of medical treatment. That the tragic shooting of Dr. McGill might have occurred as a result of a seizure is certainly not a frivolous hypothesis on these facts. The jury chose to reject that hypothesis. Had we been given the duty of finding the facts in this difficult case, we might or might not have reached a different result. Our duty here, however, is more limited, and on a review of the record as a whole we cannot say that the evidence viewed in the light most favorable to the Government is insufficient as a matter of law to sustain the jury’s conclusion that Mc-Cracken was legally sane at the time of the offense. II. REPETITIOUS INSTRUCTIONS Appellant next complains about the trial court’s instructions to the jury on the definition of insanity, the burden of proof, the definition and essential elements of the offense charged, and the meaning of “malice aforethought.” He makes no claim that the instructions misstated the law, but rather argues that they were prejudicially repetitious. This Court has often warned that in reviewing jury instructions we must evaluate “the charge as a whole, in its totality, without isolating statements which may appear prejudicial from the context in which they were made.” United States v. Williams, 5 Cir. 1973, 473 F.2d 507, 509; United States v. Jacquillon, 5 Cir. 1972, 469 F.2d 380, 386-387; United States v. Wilkinson, 5 Cir. 1972, 460 F.2d 725, 732. In this case appellant questions not the propriety of any particular statement or statements, but rather the effect of the mass taken together. Initially we must note that appellant voiced no objection at trial to the allegedly repetitious form of the instructions. Ordinarily this would preclude us from reviewing the jury instructions, Fed.R.Crim.Proe. 30, but we may notice “plain errors or defects” that affect substantial rights even though not brought to the attention of the trial court, Fed. R.Crim.Proc. 52(b). To bring himself within the grace of the “plain error” rule, appellant must show that “the charge deprived [him] of a constitutional right or was erroneous in matters which go to the very essence of the case ..” United States v. Fontenot, 5 Cir. 1973, 483 F.2d 315, 322. As Judge Wisdom wrote in United States v. Jacquillon, supra, 469 F.2d at 386, the application of the plain error rule is limited to exceptional situations involving serious deficiencies which affect the fairness, integrity, or public reputation of the judicial proceedings, United States v. Socony-Vacuum, 1940, 310 U.S. 150, 60 S.Ct. 811, 84 L.Ed. 1129, or which constitute obvious error, United States v. Atkinson, 1936, 297 U.S. 157, 56 S.Ct. 391, 80 L.Ed. 555. In Lacaze v. United States, 5 Cir. 1968, 391 F.2d 516, 521, we said that “ [r] epetitious and unnecessarily long charges are confusing to a jury and can be held prejudicial,” but we cautioned that “length and magnitude in a charge are not per se untenable.” We thus indicated our belief that challenges to jury instructions should be framed, not in terms of literary criticism— though literary craftsmanship is surely important in preparing good instructions — but rather in terms of specific el *414 ements of prejudice allegedly flowing from the questioned instructions. In theory the court’s charge to the jury is designed to inform the jury of its function of independently determining the facts and applying to those facts the law as charged by the court, United States v. Perisco, 2d Cir. 1965, 349 F.2d 6, 8. Instructions should be as succinct, clear and consistent as possible given their intended function and the nature of the particular case. See generally 1 Devitt and Blackmar, Federal Jury Practice and Instructions §§ 8.01, 8.02 (2d ed. 1970); Devitt, Ten Practical Suggestions About Federal Jury Instructions, 38 F.R.D. 75 (1965); Thomas, Improvement in Charges to Juries, 1 F.R.D. 141 (1940). Unnecessarily long and repetitive instructions increase the danger that the jury will perceive inconsistencies, become confused, and be unduly impeded in properly exercising its functions. More important, every criminal defendant has a right to trial by an impartial jury whose verdict is not commanded, coerced, or unduly influenced by the trial judge. Instructions must not, therefore, be argumentative or slanted in favor of either side, see Daniel v. United States, 5 Cir. 1959, 268 F. 2d 849, 854. Read as a whole they should neither “unduly emphasize the theory of the prosecution, thereby deem-phasizing proportionally the defendant’s theory,” Perez v. United States, 5 Cir. 1961, 297 F.2d 12, 16, nor overemphasize the importance of certain evidence or certain parts of the ease, see United States v. Curry, 5 Cir. 1973, 471 F.2d 419, 422. Under some conditions the repetition of particular instructions might have a tendency to force a verdict against an accused by unduly underscoring certain aspects of the case. Apparently, this latter type of prejudice is the gravamen of Mc-Cracken’s complaint on this point. We have studied the entire charge to the jury with care. While we agree with appellant that the instructions are unduly repetitive in some respects, we cannot agree that the repetition must have, by its very weight, influenced or induced the jury to find against appellant. 6 Nor, for that matter, can we detect any particular inconsistencies or other flaws that might have confused the jury or obstructed the performance of its functions. Boring the instructions might have been, but the repetition in these charges is a defect of the literary, not the “plain error” variety. 7 *415 III. FORM AND CONSEQUENCES OF THE VERDICT When' the jury departed to deliberate, they received a form of verdict containing four alternatives — not guilty, not guilty by reason of insanity, guilty of murder in the first degree, and guilty of murder in the second degree. At the close of his charge, the trial judge explained the various alternatives to the jury, saying of the second, [I]f your verdict is that the defendant is not guilty by reason of insanity the form of your verdict will be “we, the jury, find the defendant James E. McCracken, to be not guilty of the offense charged in the indictment by reason of insanity at the time of said act.” If that’s your decision you will put your X mark in the rectangle opposite that form of verdict and the defendant will be immediately released from custody. [TR. at 721, emphasis added.] The court also discussed the possible verdicts earlier in the charge, summarizing the insanity alternative in similar language: In the event you find that the defendant was not responsible for his conduct as a result of mental disease or defect upon all of the evidence which I have presented to you the form of your verdict will be “we the jury find the defendant, James E. McCracken, not guilty by reason of insanity,” in which event the court would turn the defendant a loose. That is the only alternative because there is no provision in the statute for doing anything [other] than completely liberating him in the event you find him not guilty by reason of insanity. [TR. at 704, emphasis added.] 8 Appellant perceives two prejudicial defects in these instructions. 9 Focusing on the form of the verdict alternatives, he argues that the “not guilty by reason of insanity” verdict [NGI] exists in federal practice only in the District of Columbia and is not otherwise authorized. By submitting this alternative, he continues, the court was in effect calling for a special verdict in a criminal case. Turning to the statements that the court would turn McCracken “a loose” in the event of an NGI verdict, appellant argues that the court thus allowed the jury to consider punishment in determining guilt or innocence. In fact, he concludes, the court invited, if not commanded, the jury to find him guilty in order to ensure his post-trial confinement. A. THE GAP IN FEDERAL LAW We begin our analysis of this part of McCracken’s appeal by recognizing that although McCracken has framed his complaint in terms of two particulars, a more fundamental problem over which we have absolutely no control underlies the challenged instructions. This prob *416 lem — somewhat euphemistically called a “gap” 10 but perhaps more accurately described as a chasm — is the absence of any federal legislation dealing with the disposition of a defendant tried in a federal court outside the District of Columbia and found not guilty by reason of insanity at the time of the offense. 11 The problem is as old as the acceptance in our jurisprudence of the insanity defense to criminal actions, but its solution is not yet in sight. At common law courts possessed the power to order the immediate confinement of defendants acquitted by reason of insanity at the time of the offense, usually in jail. Though the procedures vary, all fifty states today have statutes authorizing some type of civil commitment, either compulsory or discretionary with the trial judge, for criminal defendants found not guilty because of insanity. 12 In Mississippi, for example, the trial judge is authorized to order a defendant acquitted because of insanity committed to a State asylum for the insane if the jury has certified that the defendant is still insane and dangerous. Miss.Code Ann. §§ 41-21-21; 99-13-7 (1972). In the District of Columbia a statute prQvides for mandatory commitment to a mental hospital for those acquitted because of insanity at the time of the offense, D.C.Code § 24-301 (d). 13 In stark contrast, federal courts outside the District of Columbia have no statutory authority for special disposition of a defendant found to have been insane at the time of the offense. See generally 2 Wright, supra note 10, at 367-68. 14 Certainly the willingness of state authorities to institute state civil commit *417 ment proceedings against federal defendants acquitted because of insanity can mitigate to some extent the harsh effects of the federal statutory silence. See Wade v. United States, 9 Cir. 1970, 426 F.2d 64, 73; United States v. Freeman, 2 Cir. 1966, 357 F.2d 606, 625-626. Reliance on state cooperation, however, is at best a partial solution to a serious problem. See United States v. Chandler, 4 Cir. 1968, 393 F.2d 920, 927 (en banc); Tydings, A Federal Verdict of Not Guilty By Reason of Insanity and a Subsequent Commitment Procedure, 27 Md. L.Rev. 131, 133-34 (1967). Time and again federal courts have decried this gaping statutory hole, most often when adopting a broader definition of insanity for purposes of the insanity defense, 15 and have called upon Congress to take remedial action. E. g., Wade v. United States, 9 Cir. 1970, 426 F.2d 64, 73 (en bane); Blake v. United States, 5 Cir. 1969, 407 F.2d 908, 915 n. 2 (en banc); United States v. Chandler, 4 Cir. 1968, 393 F.2d 920, 927 (en banc); United States v. Freeman, 2 Cir. 1966, 357 F.2d 606, 625-626. Serious legislative proposals have been made, and legislation actually introduced in Congress, but the problem remains. See 2 Wright, supra note 10, at 368 n. 38; Tydings, supra. It is not our purpose here to repeat the call to Congress for help. For one thing, previous warnings have been clear and eloquent. More important, however, is our own inability to suggest an entirely satisfactory solution to the problem. We recognize, as have our judicial brothers many times in the past, the truth of Mr. Justice Clark’s words: [Jjustice does not permit punishing persons with certain mental disorders for commiting acts offending against public peace and order. But insane offenders are no less a menace to society for being held irresponsible . . The community has an interest in protecting the public from antisocial acts whether committed by sane or by insane persons. Lynch v. Overholser, 1962, 369 U.S. 705, 724, 82 S.Ct. 1063, 1074, 8 L.Ed.2d 211, 223 (Clark, J., dissenting). In the past courts have sought to frame the solution to the problem posed by Justice Clark in terms of choosing between two types of institutionalization — the prison and the mental hospital — rather than between prison and release. E. g., United States v. Chandler, 4 Cir. 1968, 393 F.2d 920, 927; United States v. Freeman, 2 Cir. 1966, 357 F.2d 606, 626. After all, Daniel M’Naghten himself was sentenced to life in an insane asylum and died after being incarcerated for the last twenty-two years of his life. Yet solving the problem has taken on new and perplexing dimensions with the rising disenchantment over indefinite confinement in a mental hospital as an alternative and the increasing disagreement over what type of disposition really serves the best interests of both the defendant and society. See, e. g., Szasz, The Insanity Plea and the Insanity Verdict, 40 Temp.L.Q. 271 (1967); Comment, supra note 12; Comment, supra note 15; Note, supra note 12. 16 Answers to these troubling *418 questions must come from elsewhere; in the meantime nagging difficulties such as the two posed by appellant in the case at bar will continue to beset courts trying to administer the insanity defense. B. NGI FORM OF VERDICT As appellant correctly states, usually the only proper verdicts in the federal criminal practice are “guilty” and “not guilty.” See 2 Wright, supra note 10. This might in part be explained in terms of need. In the District of Columbia, for example, the NGI verdict form is used to implement the special statutory provision on commitment. A simple verdict of “not guilty” in a case in which the defendant’s sanity is an issue might rest either on the jury’s belief that the accused was insane at the time of the offense or on its feeling that the Government had not proved the various elements of the offense charged beyond a reasonable doubt. To trigger the commitment provision the more explicit NGI verdict is necessary. For federal courts elsewhere, the absence of any similar statutory provision for defendants who have successfully asserted the insanity defense renders the NGI verdict unnecessary. A more important reason for the limited “guilty” — “not' guilty” alternatives in most federal cases is the strong distaste felt by the courts for so-called special verdicts in criminal cases. See generally 2 Wright, supra note 10, at 364-65. At the time of the adoption of our Constitution, English common law did not view special verdicts with such a jaundiced eye, see United States v. Noble, 3 Cir. 1946, 155 F.2d 315, 317 n. 4, but in our federal criminal practice, submission of a form calling for a special verdict or special interrogatories from the jury has been held to be error. E. g., United States v. James, 5 Cir. 1970, 432 F.2d 303, cert. denied, 1971, 403 U.S. 906, 91 S.Ct. 2214, 29 L.Ed.2d 682 (harmless error) ; United States v. Adcock, 2 Cir., 447 F.2d 1337, cert. denied, 1971, 404 U.S. 939, 92 S.Ct. 278, 30 L.Ed.2d 252 (1971) (reversible error); United States v. Spock, 1 Cir. 1969, 416 F.2d 165 (reversible error); Gray v. United States, 8 Cir., 174 F.2d 919, cert. denied, 1949, 338 U.S. 848, 70 S.Ct. 90, 94 L.Ed. 519 (reversible error). As then Chief Judge Aldrich explained in United States v. Spock, supra, 416 F.2d at 180-182, our dislike of special verdicts in criminal cases rests on reasons more basic than the suggestive but not determinative fact that the Federal Rules of Criminal Procedure do not provide for them. The principle that in a criminal case, no matter how overwhelming the evidence of guilt, the court may not direct the jury to return a verdict of guilty is a cornerstone of our jurisprudence. Unless the sacred right to trial by jury is to be a mere empty formalism, “[i]n the exercise of its functions not only must the jury be free from direct control in its verdict, but it must be free from judicial pressure, both contemporaneous and subsequent,” both of which “have been said to result from the submission of special questions.” Id. at 181. Thus the link between special verdicts or special interrogatories and directed verdicts is clear:. To ask the jury special questions might be said to infringe on its power to deliberate free from legal fetters; on its power to arrive at a general verdict without having to support it by reasons or by a report of its deliberations; and on its power to follow or not to follow the instructions of the court. Moreover, any abridgement or modification of this institution would partly restrict its historic function, that of tempering rules of law by common sense brought to bear upon the facts of a specific case. Id 17 *419 A general verdict insures the input of compassion into a jury’s decisional process. The rule against special verdicts and special questions in criminal cases is thus nothing more nor less than a recognition of the principle that “the jury, as conscience of the community, must be permitted to look at more than logic.” United States v. Spock, supra, 416 F.2d at 182. In the words of one thoughtful commentator, the prohibition of special verdicts affirms the notation that “[i]n criminal cases ... it has always been the function of the jury to apply the law, as given by the court in its charge, to the facts,” while preserving “the power of the jury to return a verdict in the teeth of the law and the facts.” 2 Wright, supra note 10, at 365. As strongly as we believe in the essential rightness of the above principles, we are unconvinced that they require reversal in this case. The NGI form of verdict may be an orphan in federal practice, for whom none will accept responsibility, but we do not agree that as used in the case sub judice it posed the dangers usually associated with special verdicts or special questions. In United States v. James, 5 Cir. 1970, 432 F.2d 303, cert. denied, 1971, 403 U.S. 906, 91 S.Ct. 2214, 29 L.Ed.2d 682, though we concluded that the error was harmless, we decried the action of the trial court in submitting in addition to “guilty” — “not guilty” verdicts six separate requests for “special findings,” inquiring which of the six offenses alleged to be objectives of a conspiracy the jury found the defendants to have committed. In United States v. Spock, supra, the trial court, sua sponte and over objection of defense counsel, put to the jury ten special questions to be answered “yes” or “no” that in effect asked for the jury’s conclusions on specific issues in the case. In Gray v. United States, 8 Cir., 174 F.2d 919, cert. denied, 338 U.S. 848, 70 S.Ct. 90, 94 L. Ed. 519 (1949), perhaps the clearest case of a true special verdict, the court submitted three forms of verdict, one for each count of the information against the defendant. The form for each count contained paraphrases of parts of the information describing particular aspects of the offenses alleged, and required the jury to make a determination of “guilty” or “not guilty” as to each particular. In the case at bar, by contrast, the trial court submitted only a third verdict alternative. To be sure, a finding of “not guilty by reason of insanity” would expose the reasons for the jury’s verdict more clearly than a simple finding of “not guilty,” although in this case the rationale for a not guilty verdict would have been clear in any event because of defense counsel’s virtual concession in open court that the insanity defense was the only issue. Nevertheless, we have difficulty finding in the NGI verdict form the step by step approach to a guilty verdict, the formal catechizing of a juror wishing to acquit, the “progression of questions each of which seems to require an answer unfavorable to the defendant” by which “a reluctant juror may be led to vote for a conviction which, in the large, he would have resisted.” United States v. Spock, supra, 416 F.2d at 182. Another, no less substantial reason for our conclusion that the use of the NGI verdict form was not erroneous lies in the case law. This Court has apparently not until now faced a direct challenge to the propriety of using the NGI form of verdict. Three prior decisions, however, compel us to answer as we do. In White v. United States, 5 Cir. 1967, 387 F.2d 367, we were asked to hold, based on two District of Columbia Circuit cases, 18 that it was error for the *420 trial court to fail to provide for a specific NGI verdict in its charge to the jury on the possible verdicts. We declined to do so, reasoning that the cited cases turned on the special D. C. statute, see supra, and noting that the “guilty”— “not guilty” alternatives used by the trial court “were consonant with pleas available to defendants under Rule 11 of the Federal Rules of Criminal Procedure.” 387 F.2d at 368. In United States v. Skinner, 5 Cir. 1971, 437 F.2d 164, the principal issue at trial was the sanity of the defendant at the time of the offense. The trial court instructed the jury that it could return verdicts of “guilty,” or “Not Guilty by Reason of Insanity.” On appeal this Court concluded that “[t]he effect of this instruction was to direct the jury to find Skinner guilty unless it appeared that he was lacking in mental capacity to commit the offense . . ..” 437 F.2d at 165. Thus in order to avoid the pitfall of partially directing a verdict against the defendant, we held that the trial judge should have included the “not guilty” alternative. United States v. Harper, 5 Cir. 1972, 460 F.2d 705, was a Skinner-based challenge to the trial court’s instruction that the only issue remaining for the jury was whether the defendant was guilty or not guilty by reason of insanity. Although the two cases were similar, we distinguished them because in Skinner the trial judge in instructing the jury commented on various admissions by the defendant, whereas in Harper the defense counsel had stipulated in open court that the allegations contained in the indictment were true, in effect pleading the defendant guilty but reserving the question of insanity. Thus, while in Skinner the jury should have been given the opportunity of finding that the elements of the offense had not been established beyond a reasonable doubt, in Harper the only issue put in dispute by the defendant was the defendant’s mental condition at the time of the offense. 460 F.2d at 707. Under the special circumstances of Harper, we concluded that instructing the jury in terms of “guilty” —“NGI” alternatives did not constitute direction of a verdict against the defendant. Appellant argues that because under White it is not error to omit the NGI alternative, it must therefore be error to include it. This step, though radiating a certain superficial logic, is neither as short nor as ineluctable as Appellant believes. The Government, on the other hand, concludes that Skinner and Harper require the trial court to use the NGI alternative in a case in which insanity is an issue. This step is as difficult for us to make as that suggested by Appellant, for at the very least the NGI verdict form cannot be required if, as White holds, omitting it is not necessarily error. 19 We think the relevant principles to be extracted from White, Skinner, and Harper are more limited. Quite simply, while the NGI alternative is neither required nor necessary, its use does not, standing alone, constitute error. The disposition of this case does not require us to express a final opinion regarding the wisdom as a policy matter *421 of using the NGI form in federal courts outside the District of Columbia. 20 Our previous cases have given at least tacit approval to that form of verdict, and we are mindful that both attorneys and trial judges might well have based their use of it on these decisions. We cannot now upset that reliance without more persuasive indications that significant prejudice results from the practice. One other consideration buttresses our conclusion that use of the NGI alternative by the trial court is insufficient, standing alone, to justify reversal in this case. On various occasions during the closing jury argument, defense counsel referred to a verdict of “not guilty by reason of insanity.” [e. g. TR. at 654, 664.] In addition, several of the instructions proposed by defense counsel and rejected by the court incorporated the NGI verdict language. [TR. at 727-28] Appellant cannot now object to the use by the court of specific wording his own counsel used and proposed for inclusion in the instructions. Cf. United States v. Easterly, 5 Cir. 1971, 444 F.2d 1236, 1240; Bianchi v. United States, 8 Cir., 219 F.2d 182, 194, cert. denied, 1955, 349 U.S. 915, 75 S.Ct. 604, 99 L.Ed. 1249. 21 C. TURNING McCRACKEN “A LOOSE” We have concluded that instructing the jury with the NGI alternative ver-diet is not, standing alone, reversible error, and have disclaimed any intention of expressing a final opinion about the wisdom of employing that verdict alternative. Nevertheless, the case sub jud-ice illustrates dramatically one problem that can arise when the NGI language is used to formulate a verdict — that it may not stand alone. The reason for the problem is simple enough: Jurors, in common with people in general, are aware of the meanings of verdicts of guilty and not guilty. It is common knowledge that a verdict of not guilty means that the prisoner goes free and that a verdict of guilty means that he is subject to such punishment as the court may impose. But a verdict of not guilty by reason of insanity has no such commonly understood meaning. Lyles v. United States, 1957, 103 U.S. App.D.C. 22, 254 F.2d 725, 728, cert. denied, 1958, 356 U.S. 961, 78 S.Ct. 997, 2 L.Ed.2d 1067. The lack of a common understanding of the meaning of the NGI verdict cannot be attributed solely to limited legal education of the public in general and jurors in particular, since the legal meaning of the NGI verdict in fact varies not only from state to state, but from state to federal practice. Given the manifest possibility of juror confusion, a trial judge who for whatever rea *422 son employs that language might understandably desire that the jury “know the meaning of this possible verdict as accurately as it knows by common knowledge the meaning of the other two possible verdicts.” Lyles v. United States, supra 254 F.2d at 728. The desire to prevent jury confusion and misunderstanding led to the D.C. Circuit to require, whenever the defense of insanity is fairly raised and unless the defendant affirmatively requests omission, carefully framed instructions explaining the NGI verdict alternative. McDonald v. United States, 1962, 114 U.S.App.D.C. 120, 312 F.2d 847, 851; Lyles v. United States, supra, 254 F.2d at 728. See generally Note, 16 Wayne State L.Rev. 1197, 1199-1201 (1970). 22 Because of the special D.C. statute, any assumption by the jury that if acquitted because of insanity the defendant will necessarily go free is simply erroneous. If allowed to exist in the jury’s mind, this erroneous assumption might well lead them to convict the defendant despite strong evidence of insanity at the time of the offense 23 — thus the need for instructions imparting “a sufficient intimation of continued incarceration to dull the danger of the jury’s having a vision of appellant walking out of the courtroom a free man upon its return of an insanity verdict.” United States v. Grimes, 1969, 137 U.S.App.D.C. 184, 421 F.2d 1119, 1125, cert. denied, 1970, 398 U.S. 932, 90 S.Ct. 1831, 26 L.Ed.2d 98. Given the lack of any comparable statute applicable to other federal courts, it is obvious that the instruction on the consequences of an NGI verdict used in the D.C. Circuit would be inapposite outside the District of Columbia. Accordingly, various courts, including this one, have held that it is not error for a trial judge to refuse to give an instruction regarding hospitalization following an NGI verdict. E. g., United States v. Borum, 10 Cir. 1972, 464 F.2d 896, 900-901; Apgar v. United States, 8 Cir. 1971, 440 F.2d 733, 737; White v. United States, 5 Cir. 1967, 387 F.2d 367; Pope v. United States, 8 Cir. 1967, 372 F.2d 710, 731, vacated, 392 U.S. 651, 88 S.Ct. 2145, 20 L.Ed.2d 1317 (1968); Powers v. United States, 10 Cir. 1962, 305 F.2d 157, 158, cert. denied, 1963, 375 U.S. 858, 84 S.Ct. 123, 11 L.Ed.2d 85; Pope v. United States, 5 Cir. 1962, 298 F.2d 507, 508-509, cert. denied, 1965, 381 U.S. 941, 85 S.Ct. 1776, 14 L.Ed.2d 704. See I Devitt & Blackmar, supra note 20. 24 The question in the case at *423 bar is whether it was error for the trial judge to instruct the jury as to the correct consequences under the law of a verdict of acquittal by reason of insanity. It might be argued that the reasoning of the D.C. Circuit is applicable by analogy to other courts without a comparable statutory scheme. Recognizing the likelihood that the jury will consider post-trial disposition of the defendant in arriving at its verdict, the D.C. Circuit simply decided to instruct on the consequences of an NGI verdict to insulate as much as possible the jury’s determination of guilt or innocence from erroneous factual assumptions. Jurors in other federal courts are no less likely to consider post-trial disposition of the defendant, and no less in need of instructions on the consequences of an acquittal based on insanity in order to avoid incorrect factual assumptions in arriving at a determination of guilt or innocence. Only in the latter case, they must be told, as the trial judge below quite correctly said, that in the event of an acquittal based on insanity, the defendant will be “turned a loose” and “released from custody” because there are no federal statutory provisions for doing other than “liberating him.” We do not believe that the conclusion follows from the premise. Except where a special statutory provision mandates a jury role in assessment or determination of penalty, the punishment provided by law for offenses charged is a matter exclusively for the court and should not be considered by the jury in arriving at a verdict as to guilt or innocence. 1 Devitt & Black-mar, supra note 20, at § 17.08; 2 Wright, supra note 10, at 365-66. In adopting the instruction on the effect of an NGI verdict, the D.C. Circuit recognized that it was, because of special circumstances, finding inapplicable the “doctrine, well established and sound, that the jury has no concern with the consequences of a verdict, either in the sentence, if any, or the nature or extent of it, or in probation.” Lyles v. United States, supra, 254 F.2d at 728. Yet a close reading of the cases from other circuits reveals that many courts have explicitly based their approval of not instructing on the consequences of an NGI verdict on the applicability of this principle as well as on the absence •of any statutory scheme for disposing of defendants acquitted by reason of insanity. E. g., United States v. Borum, 10 Cir. 1972, 464 F.2d 896, 901; White v. United States, 5 Cir. 1967, 387 F.2d 367; Pope v. United States, 8 Cir. 1967, 372 F.2d 710, 731, vacated, 392 U.S. 651, 88 S.Ct. 2145, 20 L.Ed.2d 1317 (1968). Writing in Pope v. United States, 5 Cir. 1962, 298 F.2d 507, 508, cert. denied, 1965, 381 U.S. 941, 85 S.Ct. 1776, 14 L. Ed.2d 704, Judge Gewin articulated the reasoning behind this result: The primary question raised here relates in large measure to the province of the court and the duty and function of a jury in a criminal case where the statute imposes the duty upon the court to determine the sentence to be given. Generally speaking, jurors decide the facts in accordance with the rules of law as stated in the instructions of the court. Unless otherwise provided by statute, it is the duty of the court to impose sentence, or make such other disposition of the case as required by law, after the facts have been decided by the jury. To inform the jury that the court may impose minumum or maximum sentence, will or will not grant probation, when a defendant will be eligible for a parole, or other matters relating to disposition of the defendant, tend to draw the attention of the jury away from their chief function as sole judges of the facts, open the door to compromise verdicts and to confuse the issue or issues to be decided. In a case of this nature [insanity defense to bank robbery charge] what they *424 were to decide was whether the defendant was guilty or not. [emphasis added] In Evalt v. United States, 9 Cir. 1966, 359 F.2d 534, the court reversed a Conviction because in the closing sentence of jury argument the prosecutor said, “If you find him not guilty, he walks out of this courtroom a free man, and I know, ladies and gentlemen, that you are not going to turn this man loose again on society.” When used in a case in which the only defense was insanity, the language was prejudicial because it was “in substance, an invitation to convict, for the protection of society, even though the jury might believe Evalt to have been insane.” In United States v. Birrell, 9 Cir. 1970, 421 F.2d 665, 666, another insanity defense case, the prosecutor in closing argument referred to the defendant’s record of car thefts and homosexual proclivities, and urged the jury not to “turn him loose on society.” Again the Ninth Circuit reversed because “the argument was fundamentally unfair in inviting the jury to convict even though it might believe that appellant was insane.” Cf. Bruce v. Estelle, 5 Cir. 1973, 483 F.2d 1031, 1039-1040. In Evalt and Birrell the prosecutor’s reference to releasing the accused was prejudicial notwithstanding the likelihood that jurors will perceive counsel for both sides as partisan advocates and regard statements made by counsel as to some degree the product of the heat of battle. When the same type of statements are made by the court, in theory the impartial arbiter, the one person in the courtroom who should not be committed to a particular side of the controversy, they are not likely to be received with the same degree of skepticism. The jury is far more likely to accept an invitation to consider certain factors if extended by the court than if made by counsel for one side or the other. This is the crux of the problem with instructing a jury on punishment or disposition in a ease such as the one sub judice— that it might under some circumstances induce a guilty verdict. When given to a jury that has manifested its inability to reach a verdict, such instructions on disposition of the defendant have been found erroneous because of their tendency to “unlock” the jury. Compare United States v. Davidson, 6 Cir. 1966, 367 F.2d 60, 62-63, and Demetree v. United States, 5 Cir. 1953, 207 F.2d 892, 895-896, with United States v. Jackson, 5 Cir. 1972, 470 F.2d 684, 688. In this ease the court’s statements that McCracken would be “turned a loose” if acquitted because of insanity came in the instructions given before the jury left to deliberate; however, it is not unreasonable to imagine that they could as easily have influenced the jury’s verdict by preempting the possibility of a deadlocked jury. The result — inducing a verdict — might well have been the same. We do not speculate on whether the references to releasing the defendant were calculated to coerce or induce a guilty verdict, but say only that, given the facts of the ease, those statements might reasonably have contributed to such a result. It might be argued that the prohibition on instructing the jury about possible punishment or disposition is inappo-site in a case in which the only references were to releasing the accused. While it is certainly true that “release” is not “punishment,” the prohibition is a general one, covering all consequences of a particular verdict. As Judge Gewin said in response to a similar argument in Pope v. United States, supra, 298 F. 2d at 509, “the problem presented relates to the fundamental division of duty between the court and the jury.” Nor does it matter that the instructions on turning McCracken “a loose” stated the law correctly, 25 though *425 it might seem ironic to reverse a conviction because of a jury charge embodying an accurate statement of the law. The simple answer is that application of the principle that it is error to instruct the jury on punishment or consequences of a certain verdict does not depend on the correctness of the information imparted to the jury. It is error to tell the jury about the consequences of a certain verdict even if they are mandatory. Chapman v. United States, 10 Cir. 1971, 443 F.2d 917, 920; United States v. Del Toro, 5 Cir., 426 F.2d 181, cert. denied, 1970, 400 U.S. 829, 91 S.Ct. 58, 27 L. Ed.2d 60. It was as erroneous for the trial judge to tell the jury that Mc-Cracken would be released if found insane as it would have been to tell them that the court intended to impose a life sentence if he were found guilty of murder in the first degree. The Government seeks to parry the thrust of Appellant’s argument on this point with the “harmless error rule,” see Chapman v. State of California, 1967, 386 U.S. 18, 87 S.Ct. 824, 17 L.Ed.2d 705; Fed.R.Crim.Pro. 52(a). Pointing to what it calls “overwhelming” and “uncontradicted” proof by the Government on the ultimate issue in the case — whether at the time of the shooting Appellant was suffering an epileptic seizure — the Government argues that the jury could have reached no other verdict than guilty. The Government also cites the trial court’s instruction that “punishment is a matter exclusively within the province of the Court and is not to be considered by the jury in arriving at an impartial verdict as to the guilt or innocence of the accused.” [TR. at 720] The Government apparently believes that this instruction neutralized whatever prejudice might have resulted from the language regarding McCracken’s release. See United States v. Smith, 3 Cir. 1971, 450 F.2d 312, cert. denied, 1972, 405 U.S. 932, 92 S.Ct. 989, 30 L.Ed.2d 807; Pittman v. United States, 9 Cir. 1966, 368 F.2d 560, cert. denied, 1967, 386 U.S. 995, 87 S.Ct. 1314, 18 L.Ed.2d 343. 26 *426 We cannot agree with either contention. Certainly the Government presented strong evidence supporting its theory of premeditation, strong enough for us to conclude that the evidence would support a jury verdict against Appellant. We categorically reject, however, the Government’s characterization of its own proof as “uncontradicted” or “overwhelming.” More important, we cannot ascribe to the cautionary instruction the beneficent power the Government proposes. There was clear evidence that McCracken is prone to fits and seizures, and strong evidence that he is sometimes violent. The court told the jury that he would be released in the event of a not guilty by reason of insanity verdict at two different points in the instructions, and the cautionary language was sandwiched between the two. The jury might have concluded after hearing the second mention of releasing McCracken that the cautioning words referred to something else, such as the number of years in prison possible under a guilty verdict. Or the jury might simply have been confused about the proper consideration to. be given the disposition of the accused. At the very least it seems likely, given the timing of the conflicting instructions, that the cautionary language did not divert the jury’s attention away from the issue of punishment and disposition back to the issue of guilt or innocence. We are mindful of the Supreme Court’s warning that “[w]e must guard against the magnification on appeal of instances which, were of little importance in their setting.” Glasser v. United States, 1942, 315 U.S. 60, 83, 62 S.Ct. 457, 471, 86 L.Ed. 680, 706. We recognize also that the fairness and accuracy of a jury charge is to be read and tested as a whole, and not by a single isolated sentence or remark, [citations omitted]. Only if the allegedly harmful instructions were either so overwhelmingly misleading as to be incurable, or were not effectively cured by statements elsewhere in the charge should a jury conviction be upset. United States v. Wilkinson, 5 Cir. 1972, 460 F.2d 725, 732. This is such a case. Perhaps it was a placebo, perhaps a palliative, but the cautionary instruction was not, in the circumstances of this case, a sufficient cure for the infection of prejudice caused by the court’s “turn him a loose” language. The Supreme Court in Kotteakos v. United States, 1946, 328 U.S. 750, 764-765, 66 S.Ct. 1239, 1247, 90 L.Ed. 1557, 1566, provided the benchmark for determining whether an error or defect affects substantial rights and thus may not be disregarded: The question is, not were they [the jury] right in their judgment, regardless of the error or its effect upon the verdict. It is rather what effect the error had or reasonably may be taken to have had upon the jury’s decision. The crucial thing is the impact of the thing done wrong on the minds of other men, not on one’s own, in the total setting, [citations omitted] This must take account of what the error meant to them, not singled out and standing alone, but in relation to all else that happened. And one must judge others’ reactions not by his own, but with allowance for how others might react and not be regarded generally as acting without reason. If, when all is said and done, the conviction is sure that the error did not influence the jury, or had but very slight effect, the verdict and the *427 judgment should stand, except perhaps where the departure is from a constitutional norm or a specific command of Congress, [citation omitted] But if one cannot say, with fair assurance, after pondering all that happened without stripping the erroneous action from the whole, that the judgment was not substantially swayed by the error, it is impossible to conclude that substantial rights were not affected. The inquiry cannot be merely whether there was enough to support the result, apart from the phase affected by the error. It is rather, even so, whether the error itself had substantial influence. If so, or if one is left in grave doubt, the conviction cannot stand. We find ourselves unable to say that the judgment was not substantially swayed by the error. In our bifurcated system of dividing the decision on guilt or innocence from the decision on disposition, we must be alert that each branch .operates independently. Yet because jury deliberations do not take place in a vacuum, the danger is always present that a jury facing A defendant prone to violence will base its verdict on a perceived need to protect society. When in addition the defendant relies on the insanity defense and the evidence of illness is uncontradicted, the danger is compounded by the fear with which many people still regard mental illness. The criminal law has not yet produced a final, satisfactory method for handling those who commit prohibited acts while suffering from a mental illness, but we decided long ago that such persons should not be held fully responsible. The wisdom, the justice, and the mercy reflected in that basic policy decision would be largely nullified if fears about mental illness are allowed to influence the decision on guilt or innocence. It is the trial judge, robed in the majesty, of the law and elevated above the heat of partisan advocacy, who must bear principal responsibility for insulating, to the extent it is possible, the ultimate decision on guilt or innocence from the force of these fears. The task is formidable, given the unresolved imperfections and holes in our current laws. We must be sensitive to loose language from the dais, lest the task be made impossible. Reversed. Footnotes [17] Judge Aldrich was quoting from United States v. Ogull, S.D.N.Y.1957, 149 F.Supp. 272, 276, aff’d sub nom. United States v. *419 Gernie, 2 Cir. 1958, 252 F.2d 664, cert. denied, 1958, 356 U.S. 968, 78 S.Ct. 1006, 2 L.Ed.2d 1073. [1] 18 U.S.C. § 1111(a) (1970). [2] Our en bane decision in Blake v. United States, 5 Cir. 1969, 407 F.2d 908, 916, established the definition of insanity for purposes of the insanity defense. A defendant must show that at the time of the offense by reason of a mental disease or defect, he/she lacked substantial capacity either to appreciate the wrongfulness of his/her conduct or to conform his/her conduct to the requirements of law. The term “mental disease or defect” does not include an abnormality manifested only by repeated criminal or otherwise antisocial conduct. [3] Appellant renews his complaint, raised first at trial, that Dr. Fain’s testimony relating to McCracken’s sanity at the time of the offense should be excluded from the evidence. *411 18 U.S.C. § 4244 proscribes the admission against the accused of any statements made by him in the course of a pre-trial examination, and the trial judge’s findings that the accused is competent to stand trial may not be introduced into evidence or brought to the jury’s attention. United States v. Moudy, 5 Cir. 1972, 462 F.2d 694, 696-697. Under the same statute a court granting an examination to determine competency to stand trial can provide in its order that inquiry also be made into competency at the time of the offense. Id. at 697. Writing for this court in Birdsell v. United States, 5 Cir., 346 F.2d 775, 780, cert. denied, 1965, 382 U.S. 963, 86 S.Ct. 449, 15 L.Ed.2d 366, Judge Friendly pointed out that although the issues of sanity at the time of the offense and competency to stand trial are indeed different, it does not follow “that a psychiatrist investigating the competency of an accused to stand trial can never be qualified to express an opinion as to his sanity a few months earlier.” The record may show the type of investigation necessary “to provide the trier of the facts with the information essential for a proper determination of criminal responsibility.” Id. In the case at bar the trial judge expressly directed an' inquiry into sanity at the time of the offense in his § 4244 order. The examination at Springfield occurred about one month after the shooting. At trial Dr. Fain was cross-examined extensively, and he made clear that he based his own conclusion that McCracken was sane at the time of the offense on both the tests and interviews done at Springfield and the records and testimony of the defense expert. Dr. Fain was honestly able to respond to all the pertinent legal questions relating to McCracken’s sanity. We conclude that the record amply demonstrates Dr. Fain’s qualifications for testifying about appellant’s sanity at the time of the offense, and we find the objection to that evidence to be without merit. [4] Appellant complains that the Government failed to use the various tactics enunciated in Mims v. United States, 5 Cir. 1967, 375 F.2d 135, 143-144. There are two answers to this complaint. In the first place, Mims presented the question whether the trier of fact could find defense expert opinion evidence inconclusive in the absence of countervailing Government expert evidence. The Mims court outlined how and when the Government might meet its burden of proving sanity beyond a reasonable doubt without presenting expert testimony of its own. That is not the problem in this case. To be sure, the methods outlined there for rebutting expert testimony in order to create a jury issue on the credibility and weight of the testimony may be useful and indeed even crucial in a case like the one at bar where the Government presents its own expert evidence. We find nothing in Mims to suggest, as the appellant seems to argue, that in a case revolving around a- battle between experts these several factors must be employed. The second answer to appellant’s Mims- based complaint is that the Government did utilize several of the rebuttal factors. (1) Dr. Lampert, the defense expert, was subjected to searching cross-examination that might justify the jury in not being convinced by him. (2) The Government showed the incorrectness of the factual assumptions on which Dr. Lampert’s opinion was based and the reasoning by which he progressed to his conclusion. Dr. Lampert, who had never witnessed the defendant during a seizure, explicitly rested his conclusion that McCracken had suffered a seizure at the time of the shooting on McCracken’s inability to remember what had happened; yet there was lay testimony that after the incident McCracken was able to remember what transpired and exhibited none of the usual seizure symptoms. We do not mean to express any opinion about the sufficiency of these tactics to raise a jury issue in the absence of expert testimony for the Government. [5] In Bishop v. United States, 5 Cir. 1968, 394 F.2d 500, for example, the Government offered no evidence on the insanity issue, and failed to utilize effectively any of the Mims techniques to rebut the defense expert evidence. In Brock v. United States, 5 Cir. 1967, 387 F.2d 254, the Government used three lay witnesses, two of whom had made observations far removed from the time of the offense, and again failed to use the Mims techniques to discredit the defense expert testimony. In United States v. Collier, 5 Cir. 1972, 453 F.2d 1173, the Government offered only lay witnesses and none of the Mims factors. In Nagell v. United States, 5 Cir. 1968, 392 F.2d 934, the Government presented two psychologists without special training in neurology. Because the defendant’s disorder was neurological in origin and because of inadequate factual assumptions for the Government’s expert opinions, we found insufficient evidence. Similar to Nagell was United States v. Parks, 5 Cir. 1972, 460 F.2d 736, in which we found that the psychiatrist offered by the Government had based his con-elusion of no mental illness on erroneous factual assumptions, [6] The bulk of the charge to the jury appears iu the trial transcript at pages 690-722. The greatest repetition occurs in the definition of insanity for purposes of the insanity defense, see, e. g., TR. at 695, 710-12, 716-17, and in the explanation of the Government’s burden of proof, see, e. g., TR. at 695, 696, 697-99 (“reasonable doubt”), 706, 709, 710, 711, 713, 714-17. The insanity instructions are couched in language taken from this Court’s case law, especially Blake v. United States, 5 Cir. 1969, 407 F.2d 908, and we. cannot understand how the repetition on either point operated to appellant’s disfavor. [7] We do not suggest that the prejudice resulting from repetitious instructions is to be determined by a numbers game, hut a comparison of this case with other repetitious or lengthy instructions cases reinforces our belief that the appellant here has not shown prejudice. Appellant cites a civil case, Howard v. Cincinnati Sheet Metal & Roofing Co., 7 Cir. 1956, 234 F.2d 233, in which a trial judge had given a total of eighty-nine instructions, more than 10,200 words, in a simple negligence case in which the only issue was proximate cause. The Seventh Circuit identified twice as many instructions favorable to the defendant as to the plaintiff, and agreed that the “voluminous and repetitious instructions” gave undue prominence and emphasis to the defendant’s theory of the case. The two cases we cited in Lacaze v. United States, 5 Cir. 1968, 391 F.2d 516, 521, are also instructive. United States v. Perisco, 2 Cir. 1965, 349 F.2d 6, involved the sixteen week trial of five defendants for robbery of merchandise in interstate commerce and conspiracy. There were 9,595 pages of transcript. The charge took thirteen hours over a two day period plus a four hour supplemental charge, and “[s]everal hundred pages of the charge were devoted to references to portions of the testimony, either by way of summary or direct quotation, for the purpose of marshaling the evidence or explaining points of law . . . .” Id. at 8. The *415 court reversed, in part because of the probability, given the length and repetition of the charge, that the jury must have become confused and unable to exercise its independent recollection of all the testimony. In United States v. Crosby, 2d Cir. 1961, 294 F.2d 928, on the other hand, the court rejected a challenge to the instructions based on length. Seven individual defendants and one corporate defendant were tried on a fifty-count indictment charging wire fraud, mail fraud, and the sale of unregistered securities. The trial took 3% months and produced a 10,000 page transcript. The charge took eight hours, but the Second Circuit found the extended instructions to have been necessary in such a lengthy, complicated case. The ease at bar, by contrast, involved a trial of one defendant on a one count murder indictment. The only issue in the ease was the legal sanity of the accused at the time of the offense. The trial took slightly more than three days, and produced a 734 page transcript. The charge to the jury took just over an hour, and occupies approximately thirty-five pages of transcript. [8] We note that in at least two other places in the charge, without listing the four alternatives, the court referred to finding the defendant “not guilty by reason of insanity.” TR. at 714, 717. [9] As required by Fed.R.Orim.Pro. 30, appellant objected to the challenged instructions at the appropriate time, thus preserving the error for review. [10] E. g., 2 Wright, Federal Practice and Procedure § 512, at 368 (1969). [11] Vai-ious other problems in the treatment of mentally defective persons who are taken into the federal criminal process are covered in 18 U.S.C. §§ 4241-4248 (1970), including detention for determination of present insanity and competency to stand trial (§ 4244), commitment pending competency to stand trial (§ 4246), transfer to hospital for treatment of inmates of federal prisons who are insane (§ 4241), and further detention or transfer to state authorities of federal prisoners who have served their sentences and are insane (§§ 4243, 4247-4248). [12] Lynch v. Overholser, 1962, 369 U.S. 705, 724-728, 82 S.Ct. 1063, 1074-1077, 8 L.Ed. 2d 211, 223-225 (Clark, X, dissenting). Since the date of Mr. Justice Clark’s dissent in Overholser, Tennessee has added such a provision, see Tenn.Code § 33-701 (a) (1972 Supp.). See generally Friginski, Commitment After Acquittal On Grounds of Insanity, 22 Md.L.Rev. 193 (1962); Comment, Disposition of the Insane Defendant After “Acquittal” — The Long Road From Commitment to Release, 59 J.Crim.Law 583 (1968); Note, Releasing Criminal Defendants Acquitted and Committed Because of Insanity: The Need for Balanced Administration, 68 Yale L.J. 293 (1958). [13] See Lynch v. Overholser, 1962, 369 U.S. 705, 82 S.Ct. 1063, 8 L.Ed.2d 211; United States v. Brawner, 1972, 153 U.S.App.D.C. 1, 471 F.2d 969, 997; Bolton v. Harris, 1968, 130 U.S.App.D.C. 1, 395 F.2d 642. [14] 24 U.S.C. § 211 (1970) provides that “[i]f any person, charged with crime, be found, in the court before which he is so charged, to be an insane person, such court shall certify the same to the Secretary of Health, Education and Welfare, who may order such person to be confined in Saint Elizabeths Hospital . . . .” Though the question has never been presented squarely in this Court, Judge Rives, dissenting in Howard v. United States, 5 Cir. 1956, 229 F.2d 602, 608, suggested that this provision might be available to federal district courts outside the District of Columbia. The subsequent decision by the en banc court vacating the panel decision and reversing the conviction neither turned on nor mentioned this point, Howard v. United States, 5 Cir. 1956, 232 F.2d 274. Judge Gewin has interpreted the en banc court’s silence as indicating doubts, at least in this circuit, that § 211 applies outside the District of Columbia. Pope v. United States, 5 Cir. 1962, 298 F.2d 507, 508. Although opinion elsewhere is split on the question, see authorities collected at Pope v. United States, 8 Cir. 1967, 372 F.2d 710, 731-732, vacated, 1968, 392 U.S. 651, 88 S.Ct. 2145, 20 L.Ed.2d 1317, on remand, 8 Cir. 1970, 434 F.2d 325, cert. denied, 1971, 401 U. S. 949, 91 S.Ct. 953, 28 L.Ed.2d 232 ; 2 Wright, supra note 10, at 367 n. 35, we share those doubts. We note also that the en banc court in Blake v. United States, 5 Cir. 1969, 407 F.2d 908, 915 n. 2, specifically referred to the absence of any federal statutory provision for commitment of defendants found not guilty because of insanity. [15] The passage of the provision in the D.C. Code resulted from the D.C. Circuit’s adoption of the broader Durham rule for escape from criminal responsibility. United States v. Brawner, 1972, 153 U.S.App.D.C. 1, 471 F.2d 969, 996; Comment, Commitment Following Acquittal By Reason of Insanity and the Equal Protection of the Laws, 116 U.Pa.L. Rev. 924, 927 n. 21 (1968). It has been suggested that the absence of any comparable federal statutory provision in fact impeded judicial acceptance of broadened definitions of insanity, see Tydings, A Federal Verdict of Not Guilty By Reason of Insanity and a Subsequent Commitment Procedure, 27 Md.L. Rev. 131, 138. But see Wade v. United States, 9 Cir. 1970, 426 F.2d 64, 73; Blake v. United States, 5 Cir. 1969, 407 F.2d 908, 915 n. 2. [16] Commentators have suggested a link between disenchantment with the mental hospital alternative and reluctance to use the insanity defense: where a successful insanity defense means-commitment, the well-informed defendant rarely feels that the insanity plea serves his best interests. He tends to avoid the plea, preferring punishment in jail to “treatment” in the mental hospital. *418 Szasz, The Insanity Plea and the Insanity Verdict, 40 Temp.L.Q. 271, 280 (1967). See also Comment, supra note 12. [18] McDonald v. United States, 1962, 114 U.S. App.D.C. 120, 312 F.2d 847, and Lyles v. United States, 1957, 103 U.S.App.D.C. 22, 254 F.2d 725, 728, cert. denied, 1958, 356 U.S. 961, 78 S.Ct. 997, 2 L.Ed.2d 1067. [19] The Government bases its argument principally on the following language from Skin ner: “[T]he district court should have given the jury three choices of verdict, guilty, not guilty or not guilty by reason of insanity.” 437 F.2d at 166. In concluding that Skinner boxes us in we think the Government has tried to prove too much. In the first place, that language, as well as the holdings of Skinner and Harper, was directed at the impropriety of omitting the “not guilty” alternative, not at the propriety of including the NGI alternative. We were there concerned with the problem of partially directing a verdict, which, as we have held today, is not normally a problem posed by the NGI form. While we clearly gave at least tacit approval to the NGI form of verdict, we would certainly not decide a question as important as whether a particular verdict form will be required without confronting it directly. Further indication that we were not concerned in Skinner with the propriety of the NGI alternative is that none of the cases there cited dealt with the problem of the triple alternative verdict in insanity cases. [20] Use of this verdict alternative is common in state courts, usually resting on statutory provisions. See, e. g., Sanford v. State, 217 Ga. 825, 125 S.E.2d 478 (1962), relying on Ga.Code Ann. § 27-1503 (1972); Tex.Code Crim.Pro. Vernon’s Ann. art. 46.02 Section 2(c)(2) (Supp.1972-73). Although we have no way of knowing the extent of its use in the federal district courts in this Circuit, it is obviously not unknown, notwithstanding the absence of any statutory foundation and the lack of any special need for it. It is, of course, entirely possible that a carefully framed NGI instruction might help; the jury better understand that there are several roads to acquittal. On the other hand, though we do not now intend to propose pattern instructions for the district courts in this Circuit, we note that the same kind of explanation can probably be made without using the NGI form. See, e. g., I Devitt & Blackmar, Eed-eral Jury Practice and Instructions, § 13.15 (2d ed. 1970) : If the evidence in the case leaves you with a reasonable doubt as to whether the defendant was sane at the time of the offense, you will find him not guilty, even though it may appear that he was sane at earlier and later times. [21] The principle that request for an instruction bars later objection is generally applied where one side complains on appeal of an instruction adopted literally or effectively from its own proposals. Appellant’s proposed instructions incorporating the challenged language were presented after the court had charged the jury rather than before, but this does not call for a different rule, especially in light of defense counsel’s use of the words at earlier moments in the trial. [22] Since the decision in Lyles, the D.C. Circuit lias had a number of occasions to review, refine, and ultimately reformulate the wording of the instructions to be given, which suggests that the mere passage of federal legislation providing authority for federal courts outside the District of Columbia to do other than release defendants found not guilty because of insanity would hardly solve all the problems. See, e. g., United States v. Brawner, 1972, 153 U.S.App.D.C. 1, 471 F.2d 969, 996-998, 1009-1010; United States v. Marcey, 1971, 142 U.S.App.D.C. 253, 440 F.2d 281, 286; United States v. Grimes, 1969, 137 U.S.App.D.C. 184, 421 F.2d 1119, 1125, cert. denied, 1970, 398 U.S. 932, 90 S.Ct. 1831, 26 L.Ed.2d 98; Bolton v. Harris, 1968, 130 U.S. App.D.C. 1, 395 F.2d 642, 651 n. 50; Starr v. United States, 1958, 105 U.S.App.D.C. 91, 264 F.2d 377, 382, cert. denied, 359 U.S. 936, 79 S.Ct. 652, 3 L.Ed.2d 639 (1959). [23] Lyles v. United States, 1957, 103 U.S.App. D.C. 22, 254 F.2d 725, 734, cert. denied, 1958, 356 U.S. 961, 78 S.Ct. 997, 2 L.Ed.2d 1067 (Bazelon, J., dissenting). [24] Thus the trial judge in the case sub judiee properly refused as incorrect statements of the law McCracken’s requested instructions that if you find the Defendant not guilty by reason of insanity, he, the Defendant, will be presumed to be insane and may be confined in a hospital for the insane as long as public safety and welfare require. [Sjliould you find the Defendant not guilty by reason of insanity this does not mean that Defendant will walk out of the Court a free man, but rather Defendant will be committed to a mental hospital or institution and would be confined there until such time as his sanity is restored, and he is certified to be sane and not a danger to society; otherwise, he shall be confined and remain at the said hospital or institution permanently. Tit. at 727-28. Of course the request for an *423 instruction incorporating an erroneous statement of the law in no way precludes or estops Appellant from complaining of the instruction actually given which correctly states the law. [25] It might be argued that this instruction was in fact inaccurate. After all, for all the judge knew, a psychiatrist and his custodial co-workers could have been waiting at the courthouse door pursuant to the state civil commitment statute in the event of an acquittal by reason of insanity. We do not know what percentage of federal defendants acquit *425 ted by reason of insanity in this Circuit are committed under state authority, and have already noted that reliance on state cooperation cannot be the total solution to the problem caused by the federal statutory silence. More important, however, is the fact that the danger in instructing the jury on disposition exists independently of the correctness of the instruction. [26] The Government cites three other statements as additional support for its argument that the error, if any, was harmless: a statement by defense counsel during voir dire that “the law provides no punishment for someone who is insane or incompetent” [TR. at 27]; a statement by defense counsel during closing argument that “if you entertain the least bit of doubt as to his guilty [sic] then you must let him go” [TR. at 652]; and the court’s instruction that McCracken would be released in case of a verdict of not guilts'. We find nothing in the quoted portions to cast doubt on our conclusion that the instructions challenged by Appellant in this portion of the appeal constitute reversible error. In the first place, it would be incorrect to say that because these two statements by defense counsel might have had the effect of drawing the jury’s attention to the consequences of a particular verdict, defendant is deemed to have invited error or in some way become estopped from complaining of the instructions. This is not a situation in which request for certain instructions estops the requester from later complaining if they are given. Moreover, the difference in the minds of the jurors between statements made by counsel and instructions given by the court is too clear to require elaboration. If the jury impermissibly considered disposition or punishment in reaching a verdict, as we believe they might have done, the instructions of the judge were far more likely to have been the proximate cause than these two statements by defense counsel made at widely separate points in time concerning the consequences of a not guilty verdict. Nor can we imagine how the court’s instruction on the consequences of a not guilty verdict had any effect on the jury one way or the other. As we have noted previously, it is commonly understood that a verdict of not guilty results in the release of the accused. Of course, it is better for the court to avoid whenever possible any mention of the consequences of a particular verdict; but a simple reference in passing to a matter of common understanding poses little danger. On the other hand, undue emphasis on even this commonly understood fact might in some *426 circumstances inflame or confuse the jury, diverting its attention from the real issue for decision, as by arousing passions against an unsympathetic defendant. We do not mean to imply that such a thing happened in the case at bar, but rather to suggest that if the court’s remark shows anything, it shows a trial judge overly anxious that the jury know the meaning of the verdicts about which he was instructing them, and insufficiently cognizant of the danger that the jury might give too much attention to the possible disposition of the defendant in deciding on his guilt or innocence.