dissenting:
I respectfully dissent from that part of the majority opinion which holds that the defendant’s right to trial by a fair and unbiased jury was assuredly not interfered with by events that occurred before the trial began.
The majority’s position is problematical in a number of respects. First, the majority fabricates a test of its own design — requiring a “strong possibility” of prejudice — for evaluating claims of jury bias in the absence of voir dire. Such a test has no basis in the case law, and cannot be reconciled with the test adopted by this Circuit which requires that a “reasonable likelihood” of prejudice be shown. Wansley v. Slayton, 487 F.2d 90, 92 (4th Cir. 1973), cert. denied, 416 U.S. 994, 94 S.Ct. 2408, 40 L.Ed.2d 773 (1974).
Second, the case law from which the majority attempts to gain support for its conclusion is inapposite. The majority uniformly relies upon cases in which the defendant was permitted fully to question the jurors on voir dire and to develop a complete record from which the court could properly evaluate the existence or nonexistence of prejudice. The standard for proof of prejudice when a complete record is available will of course be stricter than the standard which applies where the defendant has been completely denied the opportunity to inquire into and establish prejudice. Here voir dire on the possibly prejudicial matter was summarily denied. Wells was denied any opportunity to prove prejudice on the part of the jurors earlier
chastised by the trial judge for acquitting a defendant in another criminal trial.
A third difficulty with the majority opinion is that although the case law requires a court to consider the “totality of facts,” Wansley v. Slayton, 487 F.2d at 92, in evaluating claims of jury prejudice, the majority seeks to characterize the present case as involving pretrial publicity only. Such a characterization blatantly ignores the “totality of facts” creating prejudice in the present circumstance, since the primary source of prejudice was the judge’s harsh criticism of jurors and the way their action might be influenced and not the publication of a newspaper article.
Finally, the majority attempts to create a distinction between the present case and other cases based on the fact that the prejudicial commentary did not focus specifically on the defendant. The distinction falls flat as a matter of common sense since what is at issue is not whether a comment was directed at a specific person, but whether the defendant was tried by an unbiased jury undeterred from applying the presumption of innocence, whatever the reason. The majority inadvertently acknowledges its error when it attempts to distinguish United States v. Bland, 697 F.2d 262 (8th Cir.1983) on exactly the opposite basis, namely that the judge’s commentary in that case was more general than the commentary at issue here.
Despite what amounts to an attempt by the majority to obscure the applicable standard, the rule is that where events have occurred outside the trial which, “by reason of [their] impact on the jury, raise the ‘reasonable likelihood ’ or probability that the accused has been prejudiced in his right to a fair trial, the trial court is obligated to take appropriate steps to determine whether in fact the accused can secure ... a fair trial.” Wansley v. Slayton, 487 F.2d 90, 92 (4th Cir.1973), cert. denied, 416 U.S. 994, 94 S.Ct. 2408, 40 L.Ed.2d 773 (1974) (emphasis added).1
*479The determination as to whether there is a “reasonable likelihood or probability” of prejudice requires an evaluation of the “totality of the surrounding facts in the matter.” Wansley v. Slayton, 487 F.2d at 92. The totality of facts in the present case without doubt demonstrate a reasonable likelihood of prejudice.
I will briefly list and then elaborate upon the totality of the facts suggesting prejudice in the present case. First, the harshness of the judge’s criticism is significant. Second, it is relevant that the critique came from a judge, a person who generally commands the respect of jurors as knowledgeable in legal matters, whether or not the critique was publicly reported. The publicity here only further enhanced an already disturbing situation. Third, the sensitivity of the determination which the jury was called on to make, assessing the mens rea of a defendant charged with first degree murder where the evidence as to intent was in conflict and setting a sentence of from one year to life imprisonment, made the situation ripe for bias to have an effect. Fourth, the trial concerned the shooting death of an eighteen-year-old honor student which made the case one of public concern. Jurors who had once been chastised would suspect that they might again be subject to public reprimand — or even reprimand by the relative few who, by reason of their job duties, or who from curiosity or interest, frequent the court while it is in session and spread the word even further via the grapevine — if they found the defendant not guilty or guilty of a lesser offense. Fifth, the recency of the criticism (six days) and the number of jurors who were exposed to it must not be overlooked.
At least five jurors who served on petitioner’s jury had been sharply criticized by another judge for returning a verdict of innocent in the last criminal case in which they had participated, the week before petitioner’s trial began. The judge in the first case stated:
I can tell you that in this case I would have found this defendant guilty and it would have taken about two minutes to do so. I appreciate the time that you all have taken ... in there, four hours and obviously there was area of disagreement. But I want to tell you what the effect of your verdict is. By this verdict you are telling the citizens and people of Culpeper County that it’s all right for an employee to take large sums of money from an employer without permission and convert it to his or her use so long as the employee leaves a note promising to pay for it over a long period of time____ This is the effect of your verdict. I think it’s a gross miscarriage of justice and I don’t think you have done the citizens of Culpeper County any favor.
The incident was the subject of comments from a presumably unbiased source in a newspaper article entitled “Innocent: Judge Raps Jury for Setting Richards Free.” Petitioner brought the incident to the trial judge’s attention and proposed several voir dire2 questions to no avail.3
The identity of the commentator must be considered in the evaluation of prejudice. *480The fact that the judge presides in an arena as powerful, austere and formal as the judicial courtroom magnifies the significance of his statements in the mind of the lay juror. In Quercia v. United States, 289 U.S. 466, 53 S.Ct. 698, 77 L.Ed. 1321 (1933), in deciding that a trial judge’s remarks during trial unduly prejudiced the jury, the Supreme Court articulated this point: “The influence of the trial judge on the jury is ‘necessarily and properly of great weight and his lightest word or intimation is received with deference, and may prove controlling.’ ” Quercia, 289 U.S. at 470, 53 S.Ct. at 699; Anderson v. Warden, 696 F.2d 296, 301 (4th Cir.1982) (en banc), cert. denied, 462 U.S. 1111, 103 S.Ct. 2463, 77 L.Ed.2d 1340 (1983).
With the judge’s reprimand still fresh in mind and without the benefit of voir dire, the jurors were led into the jury box where they were bound by their oaths to administer justice in the well publicized trial4 of a “long-haired, bearded” misfit who had caused the death of an eighteen-year-old honor student. Because the case involved the killing of a human being and the Commonwealth charged first degree murder, the jurors were responsible not only for deciding whether the accused was guilty or innocent, but for making sensitive determinations as to the degree of petitioner’s culpability. They were instructed to determine what petitioner’s mental state was when he caused the death so as to render an appropriate verdict of guilty — of first degree murder, second degree murder, voluntary manslaughter or involuntary manslaughter — or not guilty.3 In addition, by law, the jurors set petitioner’s sentence. Va.Code § 19.2-295 (1983).5 6 There can be little doubt that there was a not-negligible possibility for another judge’s criticism of only days before to have a substantial effect upon the verdict and a choice of sentences ranging from life imprisonment to one year.
Despite the facts suggesting prejudice and the absence of voir dire the majority concludes that Wells has not shown a “strong possibility” of jury bias and that, absent a showing of a “strong possibility” of prejudice, there is no error which warrants reversal. The majority takes some uncertain and fallacious steps to reach its conclusions. First, it is impossible to ascertain the basis for the majority’s “strong possibility” standard, since it is not the standard explicitly adopted in Wansley v. Slayton, supra, a Fourth Circuit opinion which states that defendant must demonstrate a “reasonable likelihood” of prejudice before becoming entitled to voir dire. Nor is such a “strong possibility” standard announced in any other case of which I am aware.
*481Second, the precedents which the majority cites in support of its stricter proof requirement are inapposite. In the cases on which the majority relies voir dire was permitted in every case to determine whether jurors had or had not been influenced by exposure to prejudicial pretrial events.7 Wells was flatly rebuffed in his attempts to conduct voir dire.
The principles which the majority extracts from cases in which voir dire was permitted cannot properly be applied in the present case where voir dire was denied.8 In the present case defendant was not permitted to question jurors and, consequently, need only show that there was a “reasonable likelihood or probability” that the jury was affected by the prejudicial events. Sheppard v. Maxwell, 384 U.S. at 363, 86 S.Ct. at 1522; Wansley v. Slayton, 487 F.2d at 92. The reasons for the rule are obvious: A thorough voir dire which unearths no bias creates a strong presumption that bias is absent. It is in such a context that the cases cited by the majority have established a strict standard for substantiating allegations of jury bias. On the other hand, once a reasonable likelihood of prejudice is shown, if voir dire is not permitted, there is nothing to rebut the presumption of prejudice established by the
defendant’s showing. More importantly, where the defendant’s hands are tied by the denial of the opportunity to inquire into bias or prejudice, he can not fairly be required to establish the existence of prejudice to the same degree of certainty as a defendant who was given the opportunity to probe the jurors’ consciences.
It is easy to perceive the unfairness of requiring the defendant to prove prejudice to the same degree of certainty whether or not he was permitted to voir dire the jury. Indeed portions of the majority opinion illustrate the Catch-22 the holding will create. The majority writes:
It is the defendant’s responsibility to demonstrate a strong possibility of jury bias. He must show, through adequate voir dire, that he was denied his right to a fair trial before a panel of unbiased jurors.
Slip op. at 472 (emphasis added). However, a defendant can hardly be expected to prove that he was prejudiced through the use of adequate voir dire when he has been denied all right to conduct voir dire.
The majority opinion rests on equally shaky ground when it ignores the admonition of the case law that prejudice must be evaluated by examining the totality of the circumstances and attempts to treat the *482present case as if the only claimed source of prejudice was the pretrial publication of a newspaper article detailing the criticism by the judge of the jury in another case. Such an argument is a straw man. Clearly, in and of itself the publication of the newspaper article with respect to another case was not so prejudicial as to require reversal. The publication is, however, only a factor in assessing prejudice since it may have compounded a predisposition by making the jurors aware that they performed their duties under public scrutiny and should give heed to an exhortation in another case to be tough on the defendant in Wells’ case. It should not be forgotten that the primary source of prejudice in the present case is the fact that five jurors who had been criticized for returning a verdict of not guilty were seated on petitioner’s jury.9 The criticism combines with other factors in Wells’ case to create a totality of circumstances supporting defendant’s claim of prejudice. By characterizing the case as involving pretrial publicity and ignoring the other facts, the majority simply does not fairly address the case .before it.
There is little more substance to the majority’s attempt to suggest a distinction based on whether the prejudicial events dealt specifically with the defendant. As even the majority concedes in its discussion of United States v. Bland, 697 F.2d 262 (8th Cir.1983), such a general principle is not absolute 10 and does not override the rule of our cases that the existence a reasonable likelihood of prejudice must be assessed after examination of all of the circumstances. The majority’s attempt to distinguish Bland reveals the majority’s error. In Bland the Eighth Circuit stated that certain remarks made by a trial judge would be “clearly prejudicial” to future defendants. The majority attempts to distinguish Bland on the ground that the remarks in that case were more general than the remarks made in the present case:
The observations made by the Bland judge about the differences between federal and state trials, and the culpability of most federal criminal defendants, had *483broad applicability. These statements could have influenced the jurors’ decisions in future criminal cases they might sit on.
Slip op. at 475. Certainly the majority cannot argue that the remarks at issue in the present case were not prejudicial because they were not specific and at the same time claim that the remarks were not sufficiently general to create prejudice. Such inconsistent reasoning reveals the fragility of the suggested distinction.
The Bland case illustrates the reality that a trial judge’s remarks to a jury may be prejudicial despite the fact that they do not concern a particular defendant. Just as in Bland, the jury in the present case may have interpreted the trial judge’s remarks as reaching broadly beyond their context. Although the trial judge’s statement did not address Wells directly, it may have worked subtly to undermine the confidence that the jurors had in their own judgment, so that they were compromised in their ability independently to consider the evidence and fully to participate in the deliberations. The remarks may have served to link, in the jurors’ minds, the judge and the prosecutor. Alternatively, some jurors may have felt, as the result of the criticism and the publicity surrounding the trial, that they would be subject to disapproval in the minds of others if they rendered a verdict of guilty of a lesser offense or of not guilty. More ominously, the comments could have been understood as a general exhortation to find guilt, resulting, in the minds of some jurors, in a lessening of the principle that the jury’s function is to judge the evidence according to the dictates of its conscience, resolving reasonable doubt in favor of the defendant. Although these effects are different from the effects of an article concerning a particular defendant, it is impossible to say categorically that they are not as gravely prejudicial, particularly in the absence of any inquiry as to the effects of the comments upon the jurors. A likelihood is all the law requires to necessitate a voir dire.
There are only a few cases close on their facts to the present one. Although the majority cites the cases as supporting the result it reaches, a more accurate assessment of the cases would conclude that none is sufficiently close to provide clear direction in the present matter. Despite the lack of definite signposts in the decided cases, the considerations outlined above compel the conclusion that there was a reasonable likelihood that petitioner was not tried by a fair and impartial jury. The contrary contention simply denies reality.
The position set forth here does not require that in every case where jury prejudice is alleged, appellate courts should freely order new trials. Rather, it reaffirms what is an eminently sound position, that where the defendant has shown facts which establish a reasonable likelihood that the jury may be prejudiced, the trial judge must permit the defendant to inquire as to the claimed prejudice on voir dire. Here, if voir dire had been allowed, the current problem would not have arisen. Quite possibly the jurors from the earlier case would have shown freedom from prejudice and the case would have proceeded free from a persistent, continuing, nagging doubt. Alternatively, one or more jurors would have displayed prejudice and been excused, to be replaced by an objective person or persons. Again the case could have proceeded without the unfortunate remark of a judge in the earlier case hanging over the proceeding. As the majority acknowledges “[t]he better practice would have been for the trial judge to prevent the Richards jurors from sitting in Wells’s trial.” Slip op. at 476.
By contrast, the standard adopted by the majority, relying entirely upon cases where a full record was developed through voir dire, creates an unfair dilemma for criminal defendants. The majority requires a defendant to prove that his jury was prejudiced and at the same time removes from his hands all opportunity to do so.
Also disturbing is the majority’s willingness to assume that the defendant was not prejudiced without the benefit of any evidence to that effect. The majority's willingness blindly to assume that jurors have not been biased is both a miscarriage of *484justice and a practice highly inappropriate for an appellate court, which of necessity must base its decisions on an adequate factual record.
I respectfully dissent.
. See Turner v. Murray, 476 U.S. 1, -, 106 S.Ct. 1683, 1694, 90 L.Ed.2d 27 (1986) (a duty to inquire is triggered by a "constitutionally significant likelihood that, absent questioning^] ... jurors would not be indifferent ..."); Sheppard v. Maxwell, 384 U.S. 333, 363, 86 S.Ct. 1507, 1522, 16 L.Ed.2d 600 (1966) ("reasonable likelihood”); Jordan v. Lippman, 763 F.2d 1265, 1266-67 (11th Cir.1985) ("significant possibility”); United States v. Sawyers, 423 F.2d 1335, 1344 (4th Cir. 1970) (in presence of prejudicial *479pretrial publicity, there is a duty to inquire regarding impact on veniremen).
. The threatened prejudice was not dissipated by the trial judge's general questioning of jurors as to whether they could render an impartial decision. At least in cases such as that here, where the potential for prejudice is very high, general questioning of the jurors will not suffice to unearth the sources of prejudice which can effect a verdict. In such circumstances the juror is poorly placed to make a determination as to his own impartiality. See Jordan v. Lippman, 763 F.2d 1265, 1281 n. 18 (11th Cir.1985); United States v. Davis, 583 F.2d 190, 197 (5th Cir. 1978). See also Murphy v. Florida, 421 U.S. 794, 803, 95 S.Ct. 2031, 2037, 44 L.Ed.2d 589 (1975) (where community has been permeated with prejudicial publicity, the reliability of jurors’ protestations of impartiality may be drawn into question); Irvin v. Dowd, 366 U.S. 717, 728, 81 S.Ct. 1639, 1645, 6 L.Ed.2d 751 (1961) ("No doubt each juror was sincere when he said that he would be fair and impartial____ [Sjuch a statement ... can be given little weight.”).
. Petitioner had no means to prevent these jurors all from serving because, under Virginia law, he was only permitted four peremptory challenges, Va.Code § 19.2-262 (1983), and there were five suspect jurors. Had petitioner exercised any of the four peremptory challenges to exclude a previously seated juror, he would have had to have foregone a peremptory challenge of some other potential juror he looked on with disfavor.
. Almost all of the jurors indicated on voir dire that they had heard or read about the case in local or national newspapers.
. The jury was instructed as to first degree murder, second degree murder, voluntary manslaughter, involuntary manslaughter and the sentences which it could set for each crime. The judge instructed the jury that it should find the defendant guilty of first degree murder if it found proof beyond a reasonable doubt that:
The defendant killed Joseph Maybury, that the killing was malicious and that the killing was wilful, deliberate and premeditated.
The jury was instructed to find petitioner guilty of second degree murder if the killing was malicious, but "you do not find ... that the killing was wilful, deliberate and premeditated." The jury was told that voluntary manslaughter was the appropriate verdict if
[T]he killing was a result of an intentional act and the killing was committed while in the sudden heat of passion upon reasonable provocation or in mutual combat.
Finally, involuntary manslaughter was defined as "the killing, although unintended, was the direct result of criminal negligence."
The facts would have permitted the jury to find petitioner innocent or to find that the defendant had any of the mental states associated with criminally causing the death of another human being. The testimony, though in conflict at points, could have been understood to portray accident, the heat of passion, a struggle, the consumption of substantial quantities of alcohol — which may have affected petitioner’s mental state — as well as premeditated killing. After the student was shot, petitioner helped transport him to the hospital. Petitioner voluntarily gave a statement to the police, explaining that "[tjhat kid was never supposed to get hit. Frightened, yes, paid back, yes, but never shot.”
. The jury was charged to sentence the defendant to terms ranging from life for first degree murder to one year for involuntary manslaughter.
. Sheppard v. Maxwell, 384 U.S. 333, 86 S.Ct. 1507, 16 L.Ed.2d 600 (1966) (publicity surrounding trial and circus atmosphere of court room prejudiced jurors despite voir dire); Rideau v. Louisiana, 373 U.S. 723, 83 S.Ct. 1417, 10 L.Ed.2d 663 (1963) (television broadcast of confession viewed by three jurors required reversal regardless of what jurors stated in voir dire); Irvin v. Dowd, 366 U.S. 717, 720, 81 S.Ct. 1639, 1641, 6 L.Ed.2d 751 (1961) (voir dire examination lasted four weeks, however, despite jurors’ statements that they could disregard any predisposition based on pretrial publicity, conviction was overturned); Wall v. Superintendent, 553 F.2d 359 (4th Cir. 1977) (jurors quizzed on voir dire stated without reservation that they would be fair, yet the Fourth Circuit reversed the conviction and remanded for a new trial); United States v. Haldeman, 559 F.2d 31, 65 (D.C.Cir. 1976) (extensive voir dire was conducted which corrected any prejudice that might have been suspected as the result of pretrial publicity), cert. denied, 431 U.S. 933, 97 S.Ct. 2641, 53 L.Ed.2d 250 (1977); Wansley v. Slayton, 487 F.2d 90 (4th Cir. 1973) (defendant is entitled to voir dire where a reasonable likelihood of prejudice is shown, however, where examination of the veniremen was extensive and none of the jurors empanelled indicated prejudice, the verdict could stand), cert. denied, 416 U.S. 994, 94 S.Ct. 2408, 40 L.Ed.2d 773 (1974); Ignacio v. Guam, 413 F.2d 513, 516-18 (9th Cir.1969) (attorneys for the defendants were given an opportunity freely to question jurors as to prejudice during voir dire, but declined to do so. Because the defendants had failed to show that adverse publicity had biased the jurors after being given the opportunity to do so, the court rejected the claim of error.), cert. denied, 397 U.S. 943, 90 S.Ct. 959, 25 L.Ed.2d 124 (1970).
. For example, in Irvin v. Dowd, 366 U.S. 717, 81 S.Ct. 1639, 6 L.Ed.2d 751 (1961), cited by the majority, extensive voir dire was undertaken. The Supreme Court stated that the test for jury bias on such a complete record is "whether the nature and strength of the opinions formed [by the jurors] are such as in law necessarily * * * raise the presumption of partiality.” 366 U.S. at 723, 81 S.Ct. at 1643. It is impossible even to apply the Supreme Court’s test if the defendant is not permitted to question the jury.
. Cases dealing with jury bias are not limited to pretrial publicity. In the area of racial prejudice, the Supreme Court has held that the mere fact that a crime was interracial does not raise the specter of bias or prejudice. Ristaino v. Ross, 424 U.S. 589, 96 S.Ct. 1017, 47 L.Ed.2d 258 (1976). However, the Court has reversed convictions in interracial crimes where voir dire as to racial bias was refused and race was “inextricably bound up with the conduct of the trial." Ristaino v. Ross, 424 U.S. at 597, 96 S.Ct. at 1021. See Ham v. South Carolina, 409 U.S. 524, 93 S.Ct. 848, 35 L.Ed.2d 46 (1973). The potential that racial bias will affect a juror’s judgment is greater where the jury is called upon to make the "highly subjective, ‘unique, individualized judgment regarding the punishment that a particular person deserves’” in a death penalty case. Turner v. Murray, 476 U.S. 1, 106 S.Ct. 1683, 1687, 90 L.Ed.2d 27 (1986).
Prior knowledge of a case is another cause which can defeat the presumption of juror impartiality although a juror is not required to be totally ignorant of the facts and issues pertinent to a case. Irvin v. Dowd, 366 U.S. 717, 81 S.Ct. 1639, 6 L.Ed.2d 751 (1961). Prejudice has been presumed and convictions reversed although voir dire was conducted and jurors denied that their judgment would be affected by extensive pretrial publicity to which they had been exposed. Irvin, supra; 28 A.L.R. Fed. 26 (1976).
Jurors have been held irrevocably tainted because of their participation in proceedings which concerned the defendant prior to sitting on a jury which convicted the defendant, despite statements on voir dire that they would be able to disregard the earlier information. Donovan v. Davis, 558 F.2d 201 (4th Cir.1977); Wall v. Superintendent, 553 F.2d 359 (4th Cir.1977).
Juries have also been held to have been prejudiced by events which occurred after they had been empanelled. Particularly relevant to the present case are cases where it was held that a jury was prejudiced by a trial judge's expression of his belief that the defendant was guilty. La Rosa v. United States, 15 F.2d 479 (4th Cir.1926); 7 A.L.R.Fed. 377 (1971). For additional examples of the type of circumstances which have been considered for prejudice see 28 A.L.R.Fed. 26 (1976); 9 Fed.Proc.L.Ed. § 22-788-793 (1982).
In any case it is not that Wells contends that he inevitably was tried by a jury shown of its impartiality. Wells objects that he was not allowed even to explore the real possibility of juror prejudice.
. For example, where race is the source of jury bias, the prejudice is not directed at an individual, but at a class of persons of which the individual is member. It is a generalized bias which has an effect upon a specific individual. So might a general statement by a trial judge about the error involved in holding a criminal defendant not guilty bias a jury toward a particular criminal defendant.