(dissenting):
I respectfully dissent. While my Brothers and I disagree on numerous specific issues, my principal difficulty with the majority opinion concerns its approach which, as I see it, is insensitive to the delicate nature of a perjury prosecution and dangerously unrestrietive of the pervasive impact of Gebhard v. United States, 422 F.2d 281 (9th Cir. 1970) . By isolating each of Cook’s contentions without regard to its inherent relationship with the others, my Brothers issue an opinion that is fraught with distressing inconsistencies.
A perjury trial is not an ordinary criminal prosecution. The crime is unique in that perjury, unlike offenses such as robbery, arson, or murder, lacks an objectively demonstrable actus reus. Since the gravamen of the offense lies in the calculating utterance of testimony *763known by the actor to be false, the elements of actus reus and mens rea converge in a prosecution that focuses solely upon the person and character of the accused. This uniqueness has been recognized by the courts in fashioning rules of law peculiar to perjury prosecutions.1 The ’ time-honored application of the “two witness rule,” for example, has shrouded perjury trials with an aura of solemnity exceeded only by prosecutions for treason, the only other modern crime in which a quantitative rule of evidence is applied. See Weiler v. United States, 323 U.S. 606, 608-609, 65 S.Ct. 548, 89 L.Ed. 495 (1945); 7 Wigmore on Evidence §§ 2036-2041 (3d ed. 1940).
Cook was not accorded the full benefit of the protection generally afforded by the “two witness” rule, however, because this case was submitted to the jury in accordance with the recently announced holding of our court in Gebhard, supra, permitting proof of the subjective falsity of allegedly perjured statements by circumstantial evidence. While Gebhard adequately indicates when the traditional “two witness” rule will give way to allow proof of perjury by circumstantial evidence, that case does not explicitly explain the effect of so relaxing the proof requirement. The pervasive impact of Gebhard upon the whole trial cannot be overlooked. If the circumstantial evidence offered by the prosecution in this trial were excised from the record, the transcript, in all probability, would shrink to less than one-half of its present 2,000 pages. While I recognize the pragmatic reasons leading to our adoption of the circumstantial evidence rule in Gebhard, I cannot believe that the Gebhard court intended to impinge upon the traditionally exalted rights of one accused of perjury by sanctioning that alternate mode of proof. The clearest indication of the lack of any such intention appears in Gebhard itself, for the court there, upon its own initiative, enlarged upon the rights of one accused of perjury, extending protection to the interrogation stage of the proceedings and imposing restraint on the Government in framing an indictment. See 422 F.2d at 289-290.2
The teaching of Gebhard, therefore, is that the relaxation of the strict proof requirement is to be offset by increased vigilance on the part of the courts and the imposition of higher scruples on the prosecution. These measures are necessary to guard against the temptation to erode other safeguards erected for the protection of one accused of perjury, under the guise of merely facilitating proof of the offense. Unfortunately, I fear that my Brothers have fallen prey to that temptation, for my review of the record convinces me that, despite extremely conscientious efforts on the part of the trial judge, Cook did not receive a fair trial.3
*764I shall discuss my intense disagreement with the majority opinion under separate headings.
I. AMBIGUITY IN THE INDICTMENT.
The majority recognizes that the term “any knowledge” is vague, and the trial court’s attempted definition of the phrase indicates the variety of meanings to which the expression may be subjected.4 Moreover, as I discuss in greater detail below,5 the question set forth as item (a) of the indictment is so ambiguous that Cook could have been subjected to this prosecution regardless of his response to that query.
I do not think it proper to indict and prosecute an individual for perjury when the questions forming the basis of the charge are so vaguely and inarticulately phrased by the interrogator as to require the jury to probe the inner workings of the accused’s mind to seek to ascertain which of several plausible meanings he attributed to the ambiguous inquiries when he gave the allegedly perjurious responses. The patent vagueness of the present indictment, therefore, does present a defect of constitutional dimension.
“When elements in an indictment are so easily subject to divergent interpretation, the accused is not adequately apprised of the charges as to enable him to prepare a defense. It therefore fails to conform to the requirements of the Sixth Amendment and Federal Rules [Rule 7(c), Fed.R.Crim.P.].”
United States v. Lattimore, 127 F.Supp. 405, 410-411 (D.D.C.), aff’d, 98 U.S.App.D.C. 77, 232 F.2d 334 (1955) (footnotes omitted). See also Vitello v. United States, 425 F.2d 416, 425 (9th Cir.) (dissenting opinion), cert. denied, 400 U.S. 822, 91 S.Ct. 43, 27 L.Ed.2d 50 (1970).
II. LIFTING STATEMENTS OUT OF CONTEXT.
A careful review of the record and relevant decisions convinces me that either there is merit in Cook’s contention concerning the lifting of his statements out of context or the majority has implicitly overruled that portion of Gebhard v. United States, 422 F.2d 281, 289-290 (9th Cir. 1970), wherein our court condemned the practice of bludgeoning “a witness who is lying by repeating and rephrasing the same question, thus creating more possible perjury counts.”
The majority recognizes the well established rule that “a charge of perjury may not be sustained by the device of *765lifting a statement of the accused out of its immediate context and thus giving it a meaning wholly different than that which its context clearly shows.” Fotie v. United States, 137 F.2d 831, 842 (8th Cir. 1943); accord, Van Liew v. United States, 321 F.2d 674, 677-678 (5th Cir. 1963); Brown v. United States, 245 F.2d 549, 556 (8th Cir. 1957); United States v. Geller, 154 F.Supp. 727, 730 n. 3 (S.D.N.Y.1957); cf. United States v. Norris, 300 U.S. 564, 576, 57 S.Ct. 535, 81 L.Ed. 808 (1937).
I strongly disagree with the holding of the majority that this rule is inapplicable to the present case. While I recognize that the facts of the subject case are distinguishable from those of one of the cases in which the “out of context” rule was applied, Brown v. United States, supra, I do not believe that that circumstance suffices as a reason for refusing to apply the rule here. For when the application of the standard in other cases is examined, it becomes obvious that the purpose of the rule would be furthered by applying it to the facts before us. For example, in Van Liew v. United States, supra, the court concluded that when a question and answer in the indictment were read in context, the testimony of the accused would have to be construed as referring to a single phase of a manufacturing process, not to the entire process as the Government had there attempted to establish by lifting the question and answer out of context. Likewise, when Cook’s allegedly perjurious statements are read in context with the remainder of his testimony before the grand jury, it is evident that he did not deny “any knowledge” whatsoever, of the illicit practices. He forthrightly admitted in testimony omitted from the indictment, that he had acquired “some knowledge” and had heard “rumors” and “gossip” concerning police involvement in the payoff system.
Unlike my Brothers, therefore, I perceive a material difference between an apparently unqualified denial of “any knowledge,” and a denial coupled with an explanation of the extent of one’s “knowledge.” In the present case, the defectiveness in the indictment is all the more egregious in light of the misdirection of the jury as to the manner in which the term “knowledge” should be interpreted.6
What is even more disturbing to me, however, is that the majority, in reaching its conclusion, seems to ignore the effect of its decision on this court’s prior holding in Gebhard v. United States, 422 F.2d 281, 289-290 (9th Cir. 1970). During the grand jury proceedings, Cook was asked a sequence of three questions:
“Q. Do you have any knowledge of law enforcement officers being paid by operators of gambling establishments ?
A. No, I do not.
Q. Do you have any knowledge of any law enforcement officers taking shakedown money?
A. In my twenty-nine years’ experience I have had some knowledge that these things have existed and as a result the people that were involved were separated from the Police Department.
Q. You don’t have any knowledge of anybody currently on the force who participated in shakedowns?
A. I do not.”
As I read these queries they are redundant. The first and second questions are, for all practical purposes, identical. In addition, the third question merely asks Cook to repeat his previous response. In the indictment, however, Cook was charged with having committed perjury only with respect to the first and third of his responses.
While I realize that it is not our function to question the Government’s motives in framing an indictment, I can perceive of no valid reason why the see*766ond question and answer were deleted from the indictment here.
“[I]t is better practice, both from the standpoint of the Government and the defendant, to set forth enough of the testimony immediately before and after the alleged false statements to give some coherent context to the alleged false statements.”
Stassi v. United States, 401 F.2d 259, 262 (5th Cir. 1968).
It would be utterly unreasonable to assume that the Government felt that Cook had not perjured himself with respect to the omitted testimony, since it was entirely consistent with the other responses for which he was indicted. It is more likely that the Government was concerned that the inclusion of this question and answer in the indictment would weaken its case by giving a different connotation to the questions and answers that were included in the indictment. Yet if this were so, then my above conclusion as to the applicability of the “out of context” rule could not logically be challenged. Assuming, arguendo, that the majority has properly concluded that Cook’s statements were not improperly lifted out of context, then the only possible remaining explanation for the omission of the intervening exchange is that the Government framed the indictment so as to avoid the appearance of attempting to circumvent Gebhard v. United States, supra. If that consideration did motivate the Government, then, I submit, only the majority has been hoodwinked:
“[Tjhere is a defect . . . which appears on the face of the indictment. . [W]e do not think it proper that the government bludgeon a witness who is lying by repeating and rephrasing the same question, thus creating more possible perjury counts.”
Id., at 289-290.
Gebhard makes it clear that the mere asking of' .repetitive questions is the practice condemned. While a distinction could be urged that the Government in Gebhard was only precluded from premising additional counts of perjury upon reiterated or rephrased questions, I do not believe that the evident error in this case can be avoided by the Government’s action in omitting one offending question from the indictment and charging Cook in only one count of perjury. The Gebhard court emphasized that the damage this rule is designed to prevent occurs when an interrogator is allowed to hammer away at a witness, “thus creating more possible perjury counts,” necessarily enabling the prosecution to choose the most detrimental responses to identical questions as the bases for a charge of perjury.
The “bludgeoning” prohibition, therefore, is not unlike the “out of context” rule.7 Both proscriptions are directed at insuring that allegedly perjurious statements included within an indictment accurately reflect the testimony of the accused. This objective, in my opinion, has not been accomplished in the present case. The questions included in the indictment reflect a perversion of both Cook’s testimony and the interrogation process; hence, I would hold that the indictment is defective.
III. SUFFICIENCY OF THE EVIDENCE.
The first question set forth in the indictment made inquiry as to whether Cook possessed “any knowledge of law enforcement officers being paid by operators of gambling establishments.” Cook was asked this question on February 25, 1970. Yet there was no evidence whatsoever that the alleged payments were being made at or about that time. To the contrary, the record very clearly establishes, and the majority appears to *767recognize, that the payments had ceased approximately two years before the critical question was asked.
If as Cook contends, the participial phrase “being paid” must be construed as referring to an ongoing activity, then the evidence was insufficient to support Cook’s conviction on the basis of the question and answer first specified in the indictment. In light of the following authoritative definition and explanation of participial usage, I must agree with Cook:
“par'ti.ci.ple . . . Gram. A word that partakes of the nature of both verb and adjective; a verbal adjective, modifying a noun, but sharing the adjuncts and construction of the verb from which it is derived. The English verb has two participles: (1) the present, ending in -ing (hastily writing it down, he paused); (2) the perfect, ending for the most part in -ed, -d, -t, -en, or -n (I saw the note written and posted). These tense forms refer to the state of the action or occurrence as in progress or complete, rather than to its time of happening, which depends on the time expressed by the verb of the clause it occurs in (he climbed the stairs, smiling to himself, — in which sentence smiling refers to the same time as climbed). English participles and their phrases show the following characteristic distinctions of time reference and voice:
TIME DISTINCTIONS
ACTIVE VOICE
PASSIVE VOICE
Present
(progressive (perfect
writing arrived
being written written
Past
(progressive (perfect
having written
having been written
Perfect
having been
progressive
writing"
Webster’s New International Dictionary 1782 (2d ed. 1940).
Since it is virtually impossible to determine from the jury’s general verdict whether Cook’s conviction was based upon either item (a) or item (b) of the indictment, or both, Cook is entitled to a new trial. See Yates v. United States, 354 U.S. 298, 311-312, 77 S.Ct. 1064, 1 L.Ed.2d 1356 (1957); Vitello v. United States, 425 F.2d 416, 419 (9th Cir.), cert. denied, 400 U.S. 822, 91 S.Ct. 43, 27 L.Ed.2d 50 (1970).
As I read the majority opinion, three theories are relied upon to refute my conclusion. First, the majority makes the somewhat astounding assertions that, in light of the nature of the grand jury investigation, the disputed question should be viewed as an inquiry into past events, arid that because of his “superior intellect,” Cook “is in no position to say that he misunderstood the question and thought it was referring to the present” (emphasis added). If, as my Brothers contend, the question must be read in context to ascertain its true meaning, then it would seem obvious that the query was improperly lifted out of context and given a wholly different meaning when it was included in the indictment. This inconsistency is a minor consideration for, as even a standard dictionary demonstrates, the only proper interpretation of question (a) is that it constitutes an inquiry into events in progress at the time it was asked. The use of the participial phrase “being paid” in conjunction with the verb “have” reveals that the interrogator could only have been eliciting Cook’s knowledge concerning payments that were being made on or about February 25, 1970. The interrogator revealed in his other questioning that he was sufficiently familiar with English grammar to articulate inquiries into past transactions properly and unambiguously in the past tense.8 In addition, several of the questions posed to Cook did inquire as to his knowledge concerning present practices, and these queries were phrased correctly *768in the present tense.9 Hence, it is only if Cook “misunderstood” the question that he could have considered it an inquiry into past events.
Perhaps the most persuasive indication of the indefensibility of the majority’s position is the simple fact that had Cook responded to the question in the affirmative, there could have been no doubt, in light of the uncontroverted evidence that the payments had ceased, that such an answer would have been false. Cook was indeed “up the creek” without any semblance of a paddle when posed this question, for regardless of his response, he was subjecting himself to a perjury prosecution. I would not be surprised to find the majority’s opinion cited in the future as authority for the proposition that an unequivocal “yes” or “no” response to any question in the course of grand jury proceedings presents a jury question as to the perjurious nature of the reply.
I cannot join the majority’s distortion of the English language, nor can I conceal my concern over what I view as a lack of recognition of the unique nature of the offense of perjury. If there be any “misunderstanding” as to the meaning of the phrase “being paid,” it is one shared only by the interrogator, the Government, and the majority. I cannot uphold a conviction for perjury which was quite likely premised upon the Government agent’s lack of “knowledge” concerning proper English usage.10
I am equally troubled by the majority’s second basis for its inexplicable conclusion that no reversible error occurred, i.e., that the issue of Cook’s misunderstanding was, in any event, for the jury to resolve. My principal objection to that theory is that I can not believe that the disputed phrase is susceptible to being interpreted as merely an inarticulate inquiry into past events. However, even if I were willing to join in the speculation that “being paid” could be so construed, I could not alter my conclusion as to the sufficiency of the evidence. The burden was on the prosecution, and it did not, and apparently could not, produce adequate evidence that Cook interpreted the question in a manner inconsistent with proper English usage.
“In a ease where the question propounded admits of several plausible meanings, the defendant’s belief cannot be adequately tested and it is nec*769essary to determine what the question meant to him when he gave the disputed answer. United States v. Lattimore, 127 F.Supp. 405 (D.C.D.C.1955), aff’d, 98 U.S.App.D.C. 77, 232 F.2d 334 (1955).
“There was no evidence to show what the question meant to Mrs. Wall when she answered it. In the absence of such evidence, no determination could be made as to the falsity of her answer. The evidence was insufficient to support the conviction. The District Court erred in denying the motion for judgment of acquittal.”
United States v. Wall, 371 F.2d 398, 400 (6th Cir. 1967). Hence, a jury question had never properly matured.
The majority, therefore, cannot seriously contend that the prosecution’s evidence was sufficient to support Cook’s conviction with respect to item (a) of the indictment. Apparently recognizing the dearth of reason supporting the two theories discussed above, my Brothers advance a third ground for rejecting Cook’s contention — the applicability of Vitello v. United States, 425 F.2d 416 (9th Cir.), cert. denied, 400 U.S. 822, 91 S.Ct. 43, 27 L.Ed.2d 50 (1970).
The relationship between Vitello and Cook’s claim concerning the sufficiency of the prosecution’s evidence is not readily apparent. The nexus becomes evident, however, when the following language from Vitello is examined:
“In Yates, the court submitted to the jury two overt acts, including an overt act which had been barred by the California statute of limitations. Since there was no way of knowing whether the jury based its verdict upon the barred act, or the other alternative, the Supreme Court reversed.
“For the purposes of this decision, we accept appellant’s view of the effect of Yates, and that the teaching of that decision should be here applied if we find (1) that there was insufficient evidence to be submitted to the jury on any one or more of the specifications of falsity. . . . ”
425 F.2d at 419. The majority, therefore, seems compelled to hold the evidence sufficient with respect to item (a) so that it can invoke the holding in Vitello, concerning the jury’s supposed prowess at divining the correct rule of law from nonexistent instructions, to cure the obvious error specified by the majority as the trial court’s failure to “separate the items in telling the jury that its verdict must be unanimous.” I see the utilization of two unsound theories as a weak foundation for the majority’s application of the Vitello holding.
I harbor a more fundamental objection to the majority’s application of Vitello to the present ease. While I still entertain my doubts concerning the correctness of the reasoning employed by my Brothers in Vitello, see 425 F.2d at 425-426 (dissenting opinion), I recognize that I am bound by the decision of the majority in that case. Nonetheless, I am convinced that our present case is significantly distinguishable from Vitello.
My Brothers are most surely aware that we here have a somewhat unique situation for a perjury case. They recognize that the indictment is permeated with ambiguity and that the prosecution’s case was based primarily upon circumstantial evidence. The true significance of these factors has, however, been overlooked by the majority. Since these factors have hitherto been considered by our court only in isolation, compare Vitello v. United States, supra (ambiguity in indictment), with Gebhard v. United States, supra (circumstantial evidence rule), their confluence here obviously sets Cook’s case apart from other perjury trials that we have reviewed.
The importance of recognizing the impact of the Gebhard circumstantial evidence rule is easily demonstrated. When a unanimous panel of this court adopted the circumstantial evidence rule in Gebhard, it apparently recognized that substantial rights of the accused had been impaired thereby, and thus proceeded to fashion the aforementioned *770“bludgeoning” prohibition. I reemphasize the fact that when Congress recently abrogated the traditional “two witness” rule as previously applied in some circumstances, it too was careful to afford countervailing protection to a perjury suspect by prescribing the defense of recantation. See note 2, supra.
Since, as I demonstrate in Part VI, infra, my Brothers misconstrue the applicability of the Gebhard circumstantial evidence rule in the one instance in which they discuss it, it is understandable that they would also incorrectly characterize my concern over the impact of Gebhard on Cook’s trial as an “assault on the probity of circumstantial evidence.” I do not, and logically could not, object to Cook’s being deprived of the full benefit of the “two witness” rule. Nor do I question the reliability of the circumstantial evidence offered by the Government. I merely think it is important to recognize that Gebhard represents a departure from the traditional method of proof allowed in a perjury case, and that in cases in which Gebhard applies, we are under a duty, as exemplified by the prior actions of our court and the Congress, zealously to protect the accused’s remaining rights.
We need not go so far as the court in Gebhard or the Congress. We need only strictly enforce Cook’s right that his guilt should be determined by unanimous agreement of the jurors. On the record before us, there is no indication that Cook was so convicted, for, as the majority recognizes, “the judge did not separate the items in telling the jury that its verdict must be unanimous.” Neither the court’s instructions nor the general form of the jury’s verdict excludes the possibility that Cook’s conviction resulted from an impermissible “combination of jurors' opinions.” Vitello v. United States, supra at 426 (dissenting opinion). The majority’s only response to my conclusion is the speculation that the jurors must have been unanimous in light of the prior holding in Vitello. Yet Cook’s trial differs substantially from Vitello’s in that the degree of precision afforded by adherence to the “two witness” rule was here lacking. I believe that Gebhard commands us to insure that a similar degree of precision pervades all perjury trials, whether or not the “two witness” rule is applied. Cook’s conviction should be reversed, if for no other reason, because of the deficiency in the court’s instructions concerning required juror unanimity.
IV. HEARSAY TESTIMONY.
Cook has catalogued numerous examples of hearsay evidence which were admitted over objection, the most damaging of which, ultimately stricken, were admitted conditionally for the limited purpose of showing a conspiracy in the police department to maintain the payoff system. The trial court concluded, however, that no conspiracy had been proved, and thus instructed the jurors to disregard the inadmissible testimony.
I strongly disagree with the majority’s determination that Krulewitch v. United States, 336 U.S. 440, 69 S.Ct. 716, 93 L.Ed. 790 (1949), is not applicable. The observations of Mr. Justice Jackson concerning the “conspiracy” offense, id. at 445-448, 69 S.Ct. 716 (concurring opinion), are especially apposite in the context of the instant case — a perjury trial under Gebhard, supra, wherein the accused, lacking the mantle of protection generally afforded him by the “two witness” rule, is confronted with damaging hearsay evidence, admitted under an unsupportable theory of conspiracy. In my opinion, the injection of “conspiracy” into Cook’s trial is a factor which must be considered in assessing the prejudicial effect of this testimony.
I also deem it significant that portions of the inadmissible testimony are present in the testimony of Fuller, Jessup, and Corr, who were among the few prosecution witnesses to testify as to direct dealings with Cook that would tend to implicate him in the payoff system. The trial court did not read the stricken *771testimony to the jury because of his fear that such action would “aggravate” the situation with respect to that evidence;11 hence, I believe the majority’s assumption that the jury could sift through the mass of testimony and “forget” the hearsay portions is wholly unwarranted.
I am therefore unable to agree with my Brothers that the trial court’s instructions to disregard this virulent testimony were sufficient to accomplish that goal. The testimony was simply too pervasive, and the resultant prejudice to Cook too damaging for the harm to be obliterated by any number of instructions. See, e. g., Odom v. United States, 377 F.2d 853, 859-860 (5th Cir. 1967).
V. INSTRUCTIONS ON RIGHT OF APPEAL.
The majority has quoted portions of the statements by the trial judge wherein he delineated the parties’ respective rights to appeal, and intimated that any errors could be cured by an appellate court.12 Under normal circumstances, I would agree with the majority that these remarks, when read in context, do not rise to the level of “plain error.” Rule 52(b), Fed.R.Crim.Proc. As I have repeatedly emphasized, however, this is not a “normal” case.
In light of the serious nature of the offense of perjury and the peculiar eircumstances of the present prosecution, I believe we must recognize that these remarks do impinge upon “the integrity and reputation of the judicial process.” See People v. Morse, 60 Cal.2d 631, 36 Cal.Rptr. 201, 388 P.2d 33 (1964), cited with approval in United States v. Workman, 454 F.2d 1124, 1129 (9th Cir. 1972).
Hence, the effect of these admittedly improper remarks, when considered in cumulation with the other irregularities occurring throughout Cook’s trial, cannot be so lightly dismissed.
VI. EMPLOYMENT OF GEBHARD RULE.
Assuming, arguendo, that it was proper to submit this case to the jury, then I accept the majority’s determination that the trial court correctly instructed the jury, in accordance with Gebhard v. United States, 422 F.2d 281 (9th Cir. 1970), that the subjective falsity of Cook’s testimony could be proved by circumstantial evidence. I deem it necessary, however, to correct a statement made by the majority as to the applicability of that rule. That incorrect statement is that the trial judge instructed the jury “that the Gebhard circumstantial evidence rule applied only to the charge in Item (a), in which the issue of knowledge was presented.” My review of the instructions, however, reveals no such limitation on the rule’s *772applicability.13 The trial court did limit the circumstantial evidence rule to proof of the subjective falsity of Cook’s allegedly perjured statements. It did not, and properly could not, limit the rule to item (a) because both item (a) and item (b) inquired as to Cook’s “knowledge.” If Gebhard were not applicable to item (b), there would have been no evidence to support Cook’s conviction on that question and answer, and we would be required to reverse under Yates v. United States, supra.
VII. KNOWLEDGE INSTRUCTION.
Before delving into the myriad defects in that portion of the majority opinion pertaining to the “knowledge” instruction, I deem it necessary to set forth Cook’s contentions concerning the instruction and the proper disposition of those claims.
Cook quarrels with the trial court’s definition of the term “knowledge.” While I entertain grave doubts as to the propriety of basing a charge of perjury upon a word with so many varied meanings,14 I find no impropriety in the trial court’s defining the term to encompass both belief and direct, or personal, knowledge since the word obviously is susceptible to either interpretation.
Cook also challenges the test of guilt included within the “knowledge” instruction:
“It is your role as jurors in this case to read the Grand Jury testimony of the defendant, determine from the context of the questions and answers given there how the Government intended to use the term ‘knowledge’, determine whether the defendant understood what the Government was asking, and then determine whether defendant willfully perjured himself.” (Emphasis added.)
I am deeply troubled by the trial court’s framing the test of guilt with reference to the interrogator’s intended use of the term “knowledge” in the queries forming the basis for the charge, especially when no evidence was presented as to “how the Government intended to use the term. . . . ”
The perjury statute itself, clearly formulates a “defendant’s viewpoint” test of guilt: “Whoever, having taken an oath . . . that he will testify . truly, . . . willfully and contrary to such oath states any material matter which he does not believe to be true, is guilty of perjury. . . . ” 18 U.S.C. § 1621 (emphasis added). Furthermore, there is abundant judicial authority recognizing that the *773crucial inquiry in a perjury case is whether the accused believed his testimony to be true in light of the meaning that he, not his interrogator, attributed to the questions and answers. See, e. g., United States v. Wall, 371 F.2d 398, 400 (6th Cir. 1967); United States v. Winter, 348 F.2d 204, 210 (2d Cir.), cert. denied, 382 U.S. 955, 86 S.Ct. 429, 15 L.Ed.2d 360 (1965); Seymour v. United States, 77 F.2d 577, 582 (8th Cir. 1935); United States v. Lattimore, 127 F.Supp. 405, 408 (D.D.C.), aff'd, 98 U.S.App.D.C. 77, 232 F.2d 334 (1955); cf. United States v. Hagarty, 388 F.2d 713, 716 n. 2 (7th Cir. 1968); Van Liew v. United States, 321 F.2d 674, 679 (5th Cir. 1963); Fotie v. United States, 137 F.2d 831, 840, 842 (8th Cir. 1943).
This portion of the court’s charge to the jury, therefore, is legally incorrect because it formulates a test of guilt inconsistent with the legal standard prescribed by the statute and the cases decided thereunder. Moreover, there can be no doubt as to the misleading nature of this instruction. The expression “any knowledge” might well have conveyed one meaning to Cook and another to the interrogator. Yet Cook would not be guilty of perjury under Section 1621, even assuming he knew how the interrogator intended to use the term, if he answered the question truthfully according to his own interpretation of the inquiry.
At the trial, in fact, Cook’s principal ground of defense was that he construed the queries as eliciting his “personal knowledge” of the matters about which he was questioned, whereas the prosecution contended that the question could not be interpreted in that limited fashion. Under the instruction given by the trial court, however, the jury was not permitted to accept Cook’s theory of the case.15 While I recognize that the jurors would be free to disbelieve Cook’s assertions as to his interpretation of “any knowledge,” and thus find him guilty of perjury as charged in the indictment, I do not believe that it should be presumed that they would have taken that course of action had they been given a legally correct instruction as to the test of guilt under Section 1621.
Since this case turns wholly upon the subjective state of mind of Cook, and since we permit proof of this inherently vague and ambiguous charge by cireum*774stantial evidence, I firmly believe that the instructions to the jury should have made it perfectly clear that Cook’s guilt or innocence was to be determined with reference to Ms construction of the questions.
Prior to the trial court’s submission of the case to the jury, Cook requested that the trial court give an instruction which was in substantial conformity to the legally accepted test of guilt.16 In addition, Cook specifically objected to the defective portion of the court’s “knowledge” instruction, specifying the particular defect therein, in full compliance with Rule 30, Fed.R.Crim.Proc.17 Hence, this issue is properly before us, and, as I understand the appellate function, we are bound to give it full consideration.
As concisely stated by the Supreme Court, the guiding principle controlling the disposition of this contention is “A conviction ought not to rest on an equivocal direction to the jury on a basic issue.” Bollenbach v. United States, 326 U.S. 607, 613, 66 S.Ct. 402, 405, 90 L.Ed. 350 (1946).
Yet Bollenbach does not require the reversal of every conviction following a trial in which the jury has been given an incorrect or misleading instruction. Implicit within the above quotation are the two limitations attending application of the rule: (1) In determining whether a conviction could be said to “rest upon” an improper instruction, the instructions must be read as a whole. See, e. g., Howard v. United States, 128 U.S.App.D.C. 336, 389 F.2d 287, 291 (1967); Cohen v. United States, 378 F.2d 751, 755 (9th Cir.), cert. denied, 389 U.S. 897, 88 S.Ct. 217, 19 L.Ed.2d 215 (1967). (2) The defective instruction must relate to “a basic issue” in the case.
I have thoroughly reviewed the trial court’s entire charge to the jury and I can find no other instruction tending either to cure or to neutralize the highly prejudicial effect of the offending in*775struction. On the other hand, I reiterate that Cook did propose an instruction stating the correct rule of law and that he also drew the trial court’s attention to the defective instruction.
Furthermore, the record in this case reveals that Cook’s interpretation of the term “knowledge” was not merely “a basic issue,” but the paramount issue in this case. Cook’s defense rested almost entirely upon the ground that he interpreted the questions as inquiring into his “personal knowledge.” It is obvious, therefore, that the rule of Bollenbach applies with full force to the present action. Moreover, it should be noted that in the analogous situation presented in Weiler v. United States, 323 U.S. 606, 65 S.Ct. 548, 89 L.Ed. 495 (1945), the Supreme Court, even without the benefit of its subsequent decision in Bollenbach, did not hesitate to reverse a perjury conviction that may have been premised upon an incorrect instruction to the jury. In response to the Government’s contention in Weiler, that the trial court’s failure to instruct the jury in accordance with the “two witness” rule did not prejudice defendant because the record indicated that the rule had been satisfied, the Court stated,
“The jury convicted without being instructed that more than the testimony of a single witness was required to justify their verdict. This was no mere ‘technical’ error relating to the ‘formalities and minutiae’ of the trial. Bruno v. United States, supra [308 U.S. 287, 293, 294, 60 S.Ct. 198, 200, 84 L.Ed. 257]. We are not authorized to look at the printed record, resolve conflicting evidence, and reach the conclusion that the error was harmless because we think the defendant was guilty. That would be to substitute our judgment for that of the jury and, under our system of justice, juries alone have been entrusted with that responsibility.”
323 U.S. at 611, 65 S.Ct. at 551.
Both the Supreme Court18 and our own court19 have staunchly adhered to the principles announced in Weiler and Bollenbach. I can perceive of no valid reason to cast them aside here. I would therefore hold that because of the erroneous standard of guilt included within the “knowledge” instruction, Cook is entitled to a new trial.
The defects in the majority's analysis of this issue should be patent. I cannot understand how the instruction given by the trial judge “measures up to” the defendant’s viewpoint test of guilt followed in United States v. Lattimore, 127 F.Supp. 405, 408 (D.D.C.), aff’d, 98 U.S.App.D.C. 77, 232 F.2d 334 (1955). Furthermore, the record conclusively demonstrates that Cook’s requested instructions were, to use some of the majority’s language, far “more enlightening than that given by the court” because Cook’s proffered charge correctly stated the applicable law. Finally, I pass the majority’s suggestion that the length of an instruction may serve as a basis for ascertaining its validity. Surely, such a suggestion does not warrant an extended analytical reply.
VIII. CONCLUSION.
As the foregoing discussion indicates, many of the faults in the majority opinion are attributable to its analysis of each of Cook’s separate contentions, isolated from the others. Perhaps the clearest example of the dangers inherent in this mode of analysis is seen from the majority’s conflicting approaches in respect to the assumed operation of the jury.
*776In this case, the majority has elevated the jury tó.a.stature .unrivaled in my legal experience. In the eyes of my Brothers, the jury was able to adapt itself to a variety of positions with amazing ease,, and. an uncanny depth of legal perception. I fully agree with the majority that the law requires us to presume that the jurors followed the instructions given them by the trial court. I can therefore discern a tinge of reason in. my Brothers’ disposition of Cook’s contention concerning the hearsay evidence. Yet one must marvel at the jury’s supposed ability at the same time, wholly to ignore the trial court’s admittedly improper references concerning appellate review!
The jury’s most remarkable feats, however, derived from its facility to divine the correct rule of law from nonexistent or incorrect instructions in the charge, while adhering to its presumed role of following the instructions that were given. The majority concedes that no instruction was given the jury that a verdict of guilty had to be supported by the unanimous agreement of the jurors as to Cook’s guilt on at least one of the two allegedly perjurious statements. Nonetheless, my Brothers conclude that the jury was able to arrive at this conclusion on its own because of our court’s prior holding in Vitello, a case involving distinct facts and stating a limitation upon the holding, which, • as I have shown, make that decision wholly inapplicable to Cook’s situation. Also, while the majority apparently recognizes some defectiveness in the “knowledge” instruction, it implicitly assumes that the jury could not have been improperly influenced by the incorrect statement of law. I cannot believe that the jury was properly guided as to the applicable law when the instructions that the jury presumably followed foreclosed all consideration of the principal ground of Cook’s asserted defense.
I would reverse.
. Two of these rules that are directly applicable to the present case are the rule prohibiting lifting an allegedly perjurious statement of the accused out of its immediate context, and the prohibition against an interrogator’s “bludgeoning” a witness who is suspected of giving perjurious responses, both of which are discussed in Part II, infra.
. When Congress recently abolished the application of the “two witness” rule in certain situations (18 U.S.C. § 1623(e)), it provided another safeguard against unwarranted prosecutions for perjury, i. e., the privilege of recantation. 18 U.S.C. § 1623(d).
. The trial judge's attempt to conduct the trial in a dignified manner was thwarted by the prosecution’s utilization of Cook’s case as the device for exposing the payoff system. The sensationalism thus injected into this trial by the Government created an unwholesome atmosphere wholly inconsistent with the exacting standards attendant to the trial of a charge of perjury in the federal courts. While the prosecution may have proceeded in the utmost good faith, it is clear, as the district judge ultimately recognized, that its bounds were overstepped. This presented problems which, despite the court’s best efforts, were irremediable. Cook, like any other Accused, was entitled to the calm and dispassionate judgment of jurors uninfluenced by inadmissible testimony or other improper considerations, flee, e. p., Part VIII, infra.
The district judge, to some extent, at least, apparently shares some of my doubts about the validity of Cook’s convic*764tion. During the proceedings on Cook’s motion for a new trial, he remarked:
“Now these other alleged errors, and I think some of them may have been erroneous, some rulings of the Court, and there may have been some misconduct. I don’t think it was intentional on the part of counsel.” *****
“. . .1 certainly would think that this would be a case that I hope would be reviewed by the Court of Appeals because it does present some problems
. The court instructed the jury:
“As you know, this charge here involved alleged knowledge on' the part of the defendant, or the issue is whether or not he had knowledge.
“Now, the term ‘knowledge’ may have different meanings when it is used in different contexts. ‘Knowledge’ as it is generally used may mean acquaintance, understanding, or awareness of the facts or the truth. It includes information as to a fact, a clear perception of the truth, a firm belief, informing, an impression of the mind, the state of being aware.
“So, as you will note, it isn’t a fixed term. On the other hand, the term knowledge is sometimes used to mean ‘personal knowledge’. In that case it means knowledge of the truth in regard to a particular fact or allegation, which is original, and does not depend on information or hearsay.
“You know what you see; you know what you hear. That is direct knowledge. Whether what you hear is true or not is another matter, but you know what you hear and see, presumably, if you are in a normal condition.”
. See text following note 10, in Part II, infra.
. See Part VII, infra.
. In fact, tlie bludgeoning prohibition could be viewed as an extension of the “out of context” rule to the interrogation stage of the proceedings. In essence, this portion of Gebhard imposes the same type of restraint upon an interrogator in his questioning of a prospective perjury defendant as Fotie imposes upon the Government in framing a perjury indictment.
. “What was the policy with particular reference to pinballs and cardrooms, private clubs, punchboards, bingo raffles? Would you run down through the thing?” (Emphasis added.)
* * H* * S¡«
“Did you ever receive any specific instructions from anyone regarding the tolerance policy, namely, the mayor, the chief, council, in other words, directed to do a particular thing with regard to the tolerance policy?” (Emphasis added.)
. The following interchange best exemplifies the interrogator’s facility in this regard :
“Q. What role, if any, do you play in implementing the tolerance policy, making it function?
A. At the present time, none.
Q. What have you in the past?
A. I have tried to interpret it.”
. While the record in this case provides ample basis for a belief that Cook was somehow implicated in the payoff system, it is not our function as appellate judges to punish him for a crime with which he was not charged. Our consideration is to be confined to the crime charged to have been committed in the indictment. I do not believe the majority has properly exercised its function when it “convicts” Cook for a mistake made by the interrogator. The disposition of Cook’s appeal has led me to reflect upon the tenets so articulately expressed by Mr. Justice Frankfurter:
“[I]t may not be amiss to remind that the question is not whether guilt may be spelt out of a record, but whether guilt has been found by a jury according to the procedure and standards appropriate for criminal trials in the federal courts.”
Bollenbaeh v. United States, 326 U.S. 607, 614, 66 S.Ct. 402, 406, 90 L.Ed. 350 (1946).
“From presuming too often all errors to to be ‘prejudicial,’ the judicial pendulum need not swing to presuming all errors to be ‘harmless’ if only the appellate court is left without doubt that one who claims its corrective process is, after all, guilty. In view of the place of importance that trial by jury has in our Bill of Bights, it is not to be supposed that Congress intended to substitute the belief of appellate judges in the guilt of an accused, however justifiably engendered by the dead record, for ascertainment of guilt by a jury under appropriate judicial guidance, however cumbersome that process may be.”
326 U.S. at 615, 66 S.Ct. at 406.
. When discussing with counsel the instructions that would thereafter be given to the jury, the trial judge expressed his concern over the damage this testimony had wrought, and indicated that he would not reread the stricken testimony to the jury.
. When the trial judge was addressing the panel of prospective jurors lie went to great lengths to explain to them the functions of the court and the jury. While explaining the court’s duty to rule on objections to evidence, he stated, “It is always important, particularly to the Government, to protect the record from prejudicial error because if the defendant is found not guilty, there is no review; if the defendant is guilty, there can always be an appeal.”
At the close of the trial, in his final instructions to the jury, the trial judge again referred to the availability of appellate review:
“During the course of the trial it ultimately becomes the duty of counsel to make objections and for the Court to rule on them in accordance with the law, and the Court can make an error, those errors are usually corrected by an appellate Court if they have been damaging in any way.
“The jury should not be influenced in any way by the Court’s ruling as to any evidence, and you shouldn’t draw any inference from the fact that the question was put.”
. The trial judge clearly explained to the jury the distinctions between, and the applicability of, the two rules applicable to the Government’s burden of proof in this case:
“Now, the objective falsity of the statements must, as I say, must be established in conformity with the ‘two witnesses’ rule, which is peculiar to perjury prosecutions, and bear in mind that perjury requires a higher measure of proof than the usual criminal offense.
* * * Hi *
“To continue with the evidence that pertains particularly to knowledge, the alleged falsity of this case is in the knowledge of the accused, and knowledge can rarely be established by any means other than through circumstantial evidence. While witnesses may see and hear and so be able to give direct evidence of what a defendant does or fails to do, of course, there can be no eye witness account of the state of mind of the defendant as to what knowledge he possessed.
“Therefore, you are instructed that in cases such as this when it is alleged that the falsity is in the knowledge of the accused, the law is that such alleged falsity in knowledge, that is, the subjective falsity of the statements, may be proved by circumstantial evidence alone, that is, the two witnesses or one witness and corroborative evidence rule does not apply to the state of the defendants mind, provided such evidence is sufficient to produce conviction in your minds of such falsity beyond a reasonable doubt upon all the evidence in the case and in accordance, of course, with the other instructions I have given you.”
. See Part I, supra.
. Under the instruction given, the jury was foreclosed from determining whether Cook’s response was, in fact, premised upon his avowed interpretation of the term “any knowledge.” For example, the jury might have concluded that the interrogator intended to use the term “knowledge” to encompass “belief” as well as direct knowledge. If the jury also concluded that Cook understood that the term could have been so used, then, as I read the instruction, the jury would have been required to find Cook guilty.
It is not inconceivable that an individual possessing Cook’s superior intellectual capacity would have recognized the ambiguity in the term “any knowledge.” Hence, while Cook might well have understood how the interrogator intended to use these words, he may still have answered the question truthfully according to the interpretation he ascribed as better fitting the phrase. Yet the jurors were not allowed to determine whether Cook did interpret the question in such a manner, and, hence, they could not accept his primary ground of defense, even if they fully believed his testimony.
I do not mean, of course, to imply any belief that the jury accepted Cook’s testimony. I only wish to point out that under the imperfect instruction, the jury could have believed Cook’s version of the facts, and still would have been compelled to find him guilty. Even though the possibility of prejudice inuring to Cook from this misleading instruction is speculative, Cook was nonetheless entitled to have his guilt ascertained “by a jury under appropriate judicial guidance, however cumbersome that process may be.” Bollenbach v. United States, 326 U.S. 607, 615, 66 S.Ct. 402, 406, 90 L.Ed. 350 (1946) (emphasis added).
. Cook’s Supplementary Requested Instruction Number 2 read as follows:
“You are instructed that to sustain a conviction for perjury the burden is upon the government to establish by substantial evidence, excluding every other hypothesis than that of guilt, the essential elements of the crime charged. Perjury is the wilful, knowing and corrupt giving, under oath, of false testimony material to the issue or point of inquiry. An essential element is that the defendant must have acted with a criminal intent — he must have believed that what he swore to was false and he must have had the intent to deceive. If there was a lack of consciousness of the nature of the statement made or it was inadvertently made or there was a mistake of the import, there was not corrupt motive.”
This language is almost a direct quotation from United States v. Rose, 215 F.2d 617, 622-623 (3d Cir. 1954), which was cited as authority for the requested instruction, as was Van Liew v. United States, 321 F.2d 674, 679 (5th Cir. 1963).
. “With respect to the Court’s failure to give requested supplemental instruction Number 2 [supra note 17], we object on the grounds and for the reason that relative to the use of the word ‘knowledge’ as defined by the Court, this particular instruction would have clarified it to the jury and made it less confusing as to the import of the particular word used.
* * * * *
“And then in the Court’s instruction relative to the third paragraph in the definition of ‘knowledge’, your Honor, we would object to the Court giving that particular instruction based upon the cases set forth in the defendant’s trial brief which refer to the Fotie case, Fotie versus the United States in 137 Fed.2d, and also Brown versus the United States, 245 Fed.2d.
“That portion of the instruction allows the jury to determine from the context of the question and the answers given how the government intended to use the term ‘knowledge’ and determine whether the defendant understood the government was asking what was meant by the word ‘knowledge.’ ”
. E. g., Yates v. United States, 354 U.S. 298, 327, 77 S.Ct. 1064, 1 L.Ed.2d 1356 (1957).
. E. g., United States v. Bagby, 451 F.2d 920, 927 (9th Cir. 1971).