I dissent. I agree with Justice Mosk that error relating to the instructions and argument at the penalty trial requires reversal of that verdict. (See People v. Brown (1985) 40 Cal.3d 512, 538-544 [220 Cal.Rptr. 637, 709 P.2d 440].) Two other errors also affect the judgment. First, the use of voter registration lists as the sole source for the master jury list deprived defendant of his constitutional right to trial by a jury drawn from a representative cross-section of the community (U.S. Const., Amends. VI, XIV; Cal. Const., art. I, § 16). Second, the trial court erroneously excluded for cause a prospective juror because of his generalized opposition to the death penalty, despite the lack of any showing that he could not conscientiously perform his duties as a juror in a capital case. (See Witherspoon v. Illinois (1968) 391 U.S. 510, 519 [20 L.Ed.2d 776, 88 S.Ct. 1770]; Wainwright v. Witt (1985) 469 U.S. 412, 431 [83 L.Ed.2d 841, 856, 105 S.Ct. 844], People v. Stanworth (1969) 71 Cal.2d 820, 837 [80 Cal.Rptr. 49, 457 P.2d 889].)
*539I.
Defendant made a prima facie showing that the system by which jury venires1 were selected in Kern County at the time of his trial produced venires that were substantially underrepresentative of the Hispanic population of the community. Defendant argues that this system of jury selection violated his right to a jury drawn from a representative cross-section of the community under the Sixth and Fourteenth Amendments to the United States Constitution and article I, section 16, of the California Constitution. The majority opinion rejects this argument, citing People v. Bell (1989) 49 Cal.3d 502, 520-531 [262 Cal.Rptr. 1, 778 P.2d 129]. But this case is not the same as Bell. That decision required a defendant to prove not only that the system of selecting jurors underrepresented a cognizable group, but also to identify the probable source of the underrepresentation. (See p. 524.) Here defendant has done so, presenting proof that the county selected jurors exclusively from a source—the voter registration lists—known to under-represent Hispanics. When a defendant does identify the source of the underrepresentation, as has the defendant here, the burden should shift to the prosecution to demonstrate that the particular procedure at issue is justified.
In People v. Harris (1984) 36 Cal.3d 36 [201 Cal.Rptr. 782 [679 P.2d 433], we held that on the facts of that case, the exclusive use of the voter registration lists as a basis for selecting jurors deprived the defendant of his right to a representative jury. The reason, we explained, is “that there is a large and increasing proportion of the general population that fails to register to vote, and that the proportion of minorities failing to register is larger than that for the general population.” (P. 52.) If a minority group fails to register in proportion to its numbers, the county’s decision to use the voter registration list as the exclusive source of jurors is a decision that this group will be underrepresented on its juries. Indeed, the Legislature has now recognized that voter registration lists do not yield representative juries, and has mandated counties to supplement those lists in the selection of jury venires.2 (See Code Civ. Proc., § 197.)
*540The Sixth Amendment to the United States Constitution and article I, section 16 of the California Constitution guarantee a criminal defendant the right to a trial by an impartial jury drawn from a representative cross-section of the community. (Taylor v. Louisiana (1975) 419 U.S. 522, 530 [42 L.Ed.2d 690, 698, 95 S.Ct. 692]; People v. Wheeler (1978) 22 Cal.3d 258, 272 [148 Cal.Rptr. 890, 583 P.2d 748].) This guaranty means that “a party is constitutionally entitled to a petit jury that is as near an approximation of the ideal cross-section of the community as the process of random draw permits.” (22 Cal.3d at p.277.)
In Duren v. Missouri (1979) 439 U.S. 357 [58 L.Ed.2d 579, 99 S.Ct. 664], the United States Supreme Court described a defendant’s burden in asserting a fair cross-section challenge to a jury venire. The court held that in order to establish a prima facie violation of the fair cross-section requirement, a defendant must show (1) that the group alleged to be excluded is a “distinctive” group in the community; (2) that the representation of this group in venires from which juries are selected is not fair and reasonable in relation to the number such persons in the community; and, (3) that this underrepresentation is due to systematic exclusion of the group in the jury selection process. (P. 364 [58 L.Ed.2d at p. 579]; accord People v. Harris, supra, 36 Cal.3d 36, 50.)
Defendant in this case based his motion to quash on the testimony presented in three other jury challenges in Kern County Superior Court. (See People v. Cantu, Nos. 21891, 22229; People v. Robinson, No. 21518; People v. Streeter, Nos. 22346, 22056, 21910, 21368.) By stipulation, the evidence in these cases was incorporated into the record, and defendant’s motion was submitted on those transcripts.
The majority concede that defendant’s showing establishes that “the first prong of the Duren test is clearly satisfied.” (See maj. opn., ante, p. 491.) Defendant has also satisfied the second and third prongs of the Duren test. With regard to the second prong—that the number of members of the cognizable group in the jury pool is not fair and reasonable in relation to the number of its members in the relevant community—defendant showed a “significant disparity based on the use of total population figures.”3 (See Harris, supra, 36 Cal.3d at p. 54.) Under the calculation in People v. Cantu that Hispanics represented 17.76 percent of the county’s adult population (21.95 percent of total population) and 8.3 percent of the jury pool, there is an absolute disparity of 9.46 percent and a comparative disparity of minus 114 percent (or approximately 53 percent).
*541But while under Harris the showing based on total population figures is sufficient, defendant here went further and showed that the use of voter registration lists underrepresents the proportion of Hispanics eligible for jury service. The evidence presented in Cantu and Robinson, and incorporated into this case, showed that Hispanics comprised 16.3 percent of the jury eligible population (as compared to 17.76 percent of the total population), but still only 8.3 percent of the registered voters. Using those figures, there is an absolute disparity of 8 percent and a probability of less than one in one million that the disparity could have occurred by chance.4
Finally, defendant has also satisfied prong three of the Duren test, i.e., that the alleged underrepresentation must be due to systematic exclusion of a cognizable group in the jury selection process. (Duren v. Missouri, supra, 439 U.S. at p. 364 [58 L.Ed.2d at p. 587].) It is here that the majority rests its holding, concluding that defendant failed to make the requisite showing of systematic underrepresentation. The majority err. In my opinion, systematic underrepresentation means only that the disparity is inherent in the particular jury selection process utilized, and not the result of random variations in that system. (Bell, supra, 49 Cal.3d at p. 565 (dis. opn. of Broussard, J., citing Duren v. Missouri, supra, 439 U.S. at p. 367 [58 L.Ed.2d at pp. 588-589].) But even under the majority view in Bell a defendant satisfies this prong by showing a specific element of the selection system which causes that system to underrepresent a cognizable group. Defendant has done so here. Indeed the prosecutor here conceded that if there is underrepresentation of eligible Hispanics on the jury panel—and the evidence makes it clear that there is—“the only explanation must lie in the fact that this group is underrepresented among registered voters.”
*542The majority reply that the county’s system is racially neutral (maj. opn., ante, p. 496.) In fact, everyone—this court, the state Legislature, and the county itself—has by now recognized that a jury selection system which relies exclusively on voter registration lists is not racially neutral, and has to be changed.5 But the change to a multiple list system took time. The county chose to try defendant under the defective former system, even though that system resulted in racially unrepresentative juries, because it would have cost between $15,000 and $40,000 (less than the cost of this appeal) to put a multiple list system into effect earlier.
As we have explained (see, e.g., People v. Harris, supra, 36 Cal.3d at p. 58), to make a showing of violation of the fair cross-section requirement, defendant need not show that the jury commissioner intended to discriminate against Hispanics. All that need be shown is that the system of selection results in denial of a jury pool representing a fair cross-section of the community. The constitutionally mandated goal is a representative jury, not an unbiased commissioner or a facially neutral selection method. The decisions requiring the accused to show systematic, purposeful discrimination do not nullify others which condemn discrimination stemming from negligence or inertia. The latter recognize that official compilers of jury lists may drift into discrimination by not taking action to prevent it. In formulating a panel for a jury, officials must adhere to a standard more stringent than mere abstention from intentional discrimination; they have an affirmative duty to develop and pursue procedures aimed at achieving a fair cross-section of the community. (Avery v. Georgia (1953) 345 U.S. 559, 561 [97 L.Ed. 1244, 1247, 73 S.Ct. 891]; Akins v. Texas (1945) 325 U.S. 398, 403 [89 L.Ed. 1692, 1696, 65 S.Ct. 1276].)
Evidence of a regular and notable disparity led the United States Supreme Court in Duren, supra, 439 U.S. 357, and this court in Harris, supra, 36 Cal.3d 36, to conclude that the cause of the underrepresentation was inherent in the jury selection process. Here defendant showed more than notable disparity. Here defendant demonstrated that the underrepresentation of Hispanics is due to the selection of jurors solely from voter registration lists. I would hold that defendant has adequately met the third prong of the Duren test, showing systematic exclusion of Hispanics in the jury selection process.
*543The right to be vindicated here is the right to a trial by a jury drawn from a fair cross-section of the community. It goes to the very heart of the integrity of the fact-finding process—the impartiality of the jury. As we recognized in People v. Wheeler, supra, 22 Cal.3d 258, 276, “[T]he primary purpose of the representative cross-section requirement ... is to achieve an overall impartiality by allowing the interaction of the diverse beliefs and values the jurors bring from their group experiences.” When, as here, Hispanics are underrepresented on jury venires, just as when, in Wheeler, Blacks were peremptorily struck, “such interaction becomes impossible and the jury will be dominated by the conscious or unconscious prejudices of the majority.” (Ibid.)
Nowhere is a representative jury more critical than in a capital case. In such cases the jury not only resolves the question of guilt or innocence but also serves as the conscience of the community in its most compelling role: it decides whether a person shall live or die. In this life or death determination the interplay of diverse values and views is imperative if a fair and impartial judgment is to be rendered. A just decision in a capital case requires that people of all the different ethnic, racial and religious groups in our society have an equal chance at being selected as jurors. Defendant was denied that right in this case.
II.
In addition to its error in rejecting defendant’s prima facie showing of unconstitutionally unrepresentative jury, the trial court also erred in excluding one juror on the basis of his generalized opposition to the death penalty.
As required by Hovey v. Superior Court (1980) 28 Cal.3d 1 [168 Cal.Rptr. 128, 616 P.2d 1301], jurors were questioned separately to determine their views on the death penalty. The voir dire to determine whether Juror Giangregorio was eligible to sit on a capital jury was extremely brief:
“Q: How do you feel about the death penalty?
“A: I am against it, your Honor.
“Q: Are you against in it every case?
“A: Every instance. I do not believe in it.” Over defense objection and without further questioning, the court then granted the prosecutor’s motion to exclude the juror for cause.
*544Under Wainwright v. Witt, supra, 469 U.S. 412, a juror may be excluded because of his views on capital punishment only if “the juror’s views would ‘prevent or substantially impair the performance of his duties as a juror in accordance with his instructions and his oath.’ ” (P. 424 [83 L.Ed.2d at pp. 851-852], quoting Adams v. Texas (1980) 448 U.S. 38, 45 [65 L.Ed.2d 581, 100 S.Ct. 2521].) The burden of showing that the juror should be excluded is on the party seeking exclusion. (Witt, supra, at p. 423 [83 L.Ed.2d at p. 851].) Deference is due to the trial judge who sees and hears the juror (p. 426 [83 L.Ed.2d at p. 853]); thus the issue on review is whether the judge’s findings “are fairly supported by the record.” (P. 434 [83 L.Ed.2d at p. 858]; see Darden v. Wainwright (1986) 477 U.S. 168, 176 [91 L.Ed.2d 144, 154, 106 S.Ct. 2464].)
The brief, unsearching colloquy with Juror Giangregorio is wholly insufficient to show that “[his] views would prevent or substantially impair the performance of his duties in accord with his instructions and his oath.” (Wainwright v. Witt, supra, 469 U.S. 412.) Instead, it appears that the trial court simply excluded the juror based on the juror’s opposition in principle to the death penalty. Witherspoon v. Illinois, supra, 391 U.S. 510, expressly held that political, religious, or philosophical opposition to capital punishment is insufficient to disqualify a juror. While Wainwright v. Witt, supra, 469 U.S. 412, repudiated as dictum the rigid test set out in footnote 21 of Witherspoon, it expressly reaffirmed Witherspoon's holding that the exclusion of a juror on the bias of a generalized opposition to capital punishment was unconstitutional.
Witherspoon v. Illinois, supra, 391 U.S. 510, involved a statute which provided that in capital cases “it shall be a cause for challenge of any juror who shall, on being examined, state that he has conscientious scruples against capital punishment, or that he is opposed to the same.'' (P. 512 [20 L.Ed.2d at p. 779], italics added.) Acting pursuant to that statute, the trial court excused six jurors who “said that they did not ‘believe in the death penalty’ . . . without any attempt to determine whether they could nonetheless return a verdict of death.” (P. 514 [20 L.Ed.2d at pp. 780-781].) The United States Supreme Court found the exclusion of these jurors from the penalty jury unconstitutional, declaring that “[One] who opposes the death penalty, no less than one who favors it, can make the discretionary judgment entrusted to him by the State and can thus obey the oath he takes as a juror .... [W]hen [the State] swept from the jury all who expressed conscientious or religious scruples against capital punishment and all who opposed it in principle, the State crossed the line of neutrality. In its quest for a jury capable of imposing the death penalty, the State produced a jury uncommonly willing to condemn a man to die.” (Pp. 519-521 [20 L.Ed.2d at pp. 783-784].) “[W]e hold that a sentence of death cannot be carried out *545if the jury that imposed or recommended it was chosen by excluding veniremen for cause simply because they voiced general objections to the death penalty or expressed conscientious or religious scruples against its infliction. No defendant can constitutionally be put to death at the hands of a tribunal so selected.” (Pp. 522-523 [20 L.Ed.2d at pp. 784-785], fn. omitted.)
In footnote 21 of Witherspoon v. Illinois (supra, 391 U.S. 510), the court proposed a specific test: jurors could be excluded only if they “made unmistakably clear . . . that they would automatically vote against the imposition of capital punishment without regard to any evidence that might be developed at the trial of the case before them.” (P. 522, fn. 21 [20 L.Ed.2d at p. 785].) This language became the “Witherspoon test” as applied in the courts of this and other states. (See, e.g., People v. Velasquez (1980) 26 Cal.3d 425 [162 Cal.Rptr. 306, 606 P.2d 341].)
In Wainwright v. Witt, supra, 469 U.S. 412, the court rejected the Wither-spoon test in favor of the one originally set out in Adams v. Texas, supra, 448 U.S. at page 45 [65 L.Ed.2d at pages 589-590]. It acknowledged, however, that its standard “leaves trial courts with the difficult task of distinguishing between prospective jurors whose opposition to capital punishment will not allow them to apply the law or view the facts impartially and jurors who, though opposed to capital punishment, will nevertheless conscientiously apply the law to the facts adduced at trial.” (Witt, supra, 469 U.S. at p. 421 [83 L.Ed.2d at p. 850].) This language makes it clear that even after Witt, opposition to capital punishment by itself is still insufficient to justify removal of a juror. Thus, Witt reaffirmed that Witherspoon was correctly decided on the facts of that case.6 Confirming this conclusion, Witt described Witherspoon as representing a necessary balance between a defendant’s right to a jury drawn from a fair cross-section of the community and the prosecutor’s right to remove a biased juror, and stated “[W]e adhere to the essential balance struck by the Witherspoon decision in 1968 . . . ; we simply modify the test stated in Witherspoon’s footnote 21 to hold that the State may exclude from capital sentencing juries that ‘class’ of veniremen whose views would prevent or substantially impair the performance of the duties in accordance with their instructions or their oaths.” (P. 424, fn. 5 [83 L.Ed.2d at p. 852].)
The California cases following Wainwright v. Witt, supra, 469 U.S. 412, illustrate the circumstances under which a juror can be excluded because of *546his views on capital punishment. In People v. Miranda (1987) 44 Cal.3d 57, 95 [241 Cal.Rptr. 594, 744 P.2d 1127], the juror asserted that she would “ ‘never vote for a verdict of death.’ ” In People v. Guzman (1988) 45 Cal.3d 915 [248 Cal.Rptr. 467, 755 P.2d 917], one juror said his mind was a “closed book,” and that he was not sure he could conceive of a crime so heinous that he would vote for death. (P. 955.) Upon further questioning he affirmed that he would never vote for the death penalty. (Ibid.) Another juror said, “I guess conscientiously I would have to say that I would absolutely vote for the life without parole.” Asked if she would ever so vote “automatically”, she replied, “ T would have to say that I would.’ ” (P. 956.) In People v. Howard (1988) 44 Cal.3d 375, 418 [243 Cal.Rptr. 842, 749 P.2d 279], each excluded juror “unqualifiedly stated he or she would not vote to impose death.” In People v. Coleman (1989) 48 Cal.3d 112, 137 [255 Cal.Rptr. 813, 768 P.2d 32], the juror, asked if she would automatically vote against first degree murder, replied, “ ‘I’m afraid I would avoid the death penalty. I would vote so that I wouldn’t have the death penalty on my mind.’ ” In People v. Walker (1988) 47 Cal.3d 605, 624 [253 Cal.Rptr. 863, 765 P.2d 70], the juror asserted that she would vote against the death penalty “ ‘regardless of what the facts might be in this case.’ ” In People v. Boyde (1988) 46 Cal.3d 212 [250 Cal.Rptr. 83, 758 P.2d 25], a juror was asked whether he could under any circumstances return a death verdict; he responded, “I can’t do it, sir.” (P. 245.) Another juror said he would not impose the death penalty regardless of the severity of the aggravating evidence. (Ibid.) A third juror said the only case in which she might vote for death would be one involving the murder of her own child. (P. 246.) In sum, whenever we have upheld the exclusion of a juror, that juror has said something which indicates that he could not decide the case impartially on the evidence and relevant factors, but instead that his vote was predetermined by his views on the death penalty.
With these precedents in mind, it is instructive to look again at the voir dire of Juror Giangregorio. He was first asked, “How do you feel about the death penalty?” and replied, “I am against it, your Honor.” This is the sort of generalized opposition to capital punishment which is clearly insufficient to disqualify a juror under Witherspoon, supra, 391 U.S. 510, Wainwright v. Witt, supra, 469 U.S. 412, or any California precedent. The judge then asked, “Are you against it in every case.” The juror replied, “Every instance. I do not believe in it.” This second answer adds nothing, because a person who is against the death penalty is, by definition, opposed to it in every instance. One who favors the death penalty in some cases but not in others is considered a supporter of the death penalty. The court asked Giangregorio only whether he was “against” the death penalty “in every instance,” not whether he could apply the law if he concluded that under the law the death penalty was appropriate. The juror’s answer, “I do not *547believe in it,” is again just a general statement of belief, not an expression of whether his ability to apply the law would be impaired.
The majority opinion, however, asserts that “Giangregorio’s expressed antipathy to the death penalty in ‘every instance’ . . . would undoubtedly ‘substantially impair the performance of his duties as a juror.’ ” (Maj. opn., ante, at p. 503.) This assertion is squarely inconsistent with the language in Witherspoon where the court declared that “[One] who opposes the death penalty, no less than one who favors it, can make the discretionary judgment entrusted to him by the State and can thus obey the oath he takes as a juror.” (Witherspoon v. Illinois, supra, 391 U.S. 510, 519 [20 L.Ed.2d 776, 783].) As we have seen, this portion of Witherspoon was not overturned by Wainwright v. Witt, supra, 469 U.S. 412, and remains controlling law.
For the foregoing reasons, I would reverse the judgment and remand this case for a new trial.
Appellant’s petition for a rehearing was denied November 28, 1990.
As the majority notes (maj. opn., ante, p. 489, fn. 2), though defendant frames this issue in terms of a challenge to the assembly of the master jury list, defendant, trial court, and the People apparently understood the basis of the motion to be a challenge to the composition of Kern County jury venires.
At the time of defendant’s trial, the master jury list in Kern County was compiled through the random selection of names from the county’s voter registration list alone. Questionnaires were then sent to those chosen to determine eligibility to serve as jurors. Although there were plans to use the names of those holding driver’s licenses to assemble the master jury list in the future, such plans had not yet been implemented at the time of defendant’s trial.
Kairys et al., Jury Representativeness: A Mandate for Multiple Source Lists (1977) 65 Cal.L.Rev. 776, 788-792.
The trial court itself recognized a prima facie showing of underrepresentation of Hispanic persons on Kern County jury panels when, in denying defendant’s motion to quash, it stated, “I do believe that there has been a prima facie showing that there is a disproportionate number of people with Hispanic surnames called for jurors, but I am not sure that means a thing, and as everybody knows by reading the Los Angeles Times and any other publication that there are an enormous number of people in this country illegally or green card-wise [szc] who are not really citizens. So, I don’t think that really in and of itself means anything.” The court’s ruling was erroneous both in disregarding the proof of underrepresentation on the basis of allegedly common knowledge about the number of noncitizens in the county and in its finding that there was no showing of systematic exclusion.
First, as we held in Harris, supra, 36 Cal.3d 36, a defendant may rely upon total population figures, rather than pinpoint the exact number of people eligible for jury service, to make his prima facie case. The trial court here, like the court in Harris (p. 48), erroneously rejected defendant’s showing on the grounds that his percentage of Hispanics may have included a large number of people who were not eligible for jury service, and in so doing improperly relieved the state of its burden to refute or justify the underrepresentation.
Second, the trial court ignored the evidence which defendant presented relating to the jury eligible population. Rather the court casually referred to the common knowledge, as reported in the newspaper, that there is “an enormous number” of noncitizens in the United States without recognizing that the evidence that had been presented did take account of that factor.
The majority quote United States v. Cecil (4th Cir. 1988) 836 F.2d 1431, which said that “[I]t is sufficient that the system adopted provides a fair cross-section and we find both common sense and precedent establish that if the voter registration lists do this they are not tainted by some affirmative form of discrimination.” (Pp. 1448-1449, italics added.) Clearly this and other cases cited by the majority are not authority that voter registrations lists may be used as an exclusive source for jury selection if they do not provide a fair cross-section. We cannot expect perfection, as Cecil says (ibid.), but surely we can do better than the 53 percent comparative disparity shown here.
Likewise, the court apparently agreed that two post -Witherspoon cases (Maxwell v. Bishop (1970) 398 U.S. 262 [26 L.Ed.2d 221, 90 S.Ct. 1578] and Boulden v. Holman (1969) 394 U.S. 478 [22 L.Ed.2d 433, 89 S.Ct. 1138]), both of which applied the Witherspoon formula in overturning death penalties, were correctly decided, since its decision distinguishes them in part on the ground that “both involved jurors who were excused merely because they had ‘conscientious’ objections to, or did not ‘believe in,’ the death penalty.” (Witt, supra, 469 U.S. at p. 422, fn. 4 [83 L.Ed.2d at p. 850].)