People v. Myers

43 Cal.3d 250 (1987) 729 P.2d 698 233 Cal. Rptr. 264

THE PEOPLE, Plaintiff and Respondent,
v.
VENSON LANE MYERS, Defendant and Appellant.

Docket No. Crim. 21991.

Supreme Court of California.

January 2, 1987.

*255 COUNSEL

Frank O. Bell, Jr., State Public Defender, under appointment by the Supreme Court, Donald L.A. Kerson and Michael Tanaka, Deputy State Public Defenders, for Defendant and Appellant.

John K. Van de Kamp, Attorney General, Steve White, Chief Assistant Attorney General, William R. Weisman, Thomas L. Willhite, Jr., and John R. Gorey, Deputy Attorneys General, for Plaintiff and Respondent.

OPINION

GRODIN, J.

In January 1979, defendant Venson Lane Myers, along with several accomplices, committed two armed robberies of liquor stores, three weeks apart. During the second robbery, defendant shot both the store clerk and a patron, killing the patron. On the basis of these two incidents, a jury convicted defendant of one count of first degree murder (Pen. Code, §§ 187, 189),[1] one count of assault with intent to commit murder (§ 217), and three counts of robbery (§ 211); allegations of firearm use (§§ 12022.5, 1203.06) and great bodily injury (§ 12022.7) were also sustained. In addition, the jury found true one special circumstance allegation: that the murder was committed during the commission of a robbery. (§ 190.2, subd. (a)(17)(i).) At the initial penalty phase proceeding, the jury was unable to agree on penalty; a new jury was empaneled, and after a second penalty trial, the new jury returned a sentence of death. The matter is before us on automatic appeal. (§ 1239.)

Defendant raises only a single challenge to the guilt/special circumstance phase of the trial, contending that the trial court erred in failing to grant his motion for mistrial based on the procedure used in selecting the jury panel. Defendant's jury challenge was directed at the county's use of voter registration lists as the single source for jury panels, and was similar to a challenge that was recently before this court in People v. Harris (1984) 36 Cal.3d 36 [201 Cal. Rptr. 782, 679 P.2d 433]. Defendant contends that under Harris his conviction must be reversed. We conclude, however, by *256 analogy to a decision of the United States Supreme Court involving a comparable challenge to a well-established jury selection procedure, that Harris should not be applied retroactively to invalidate convictions returned by juries that were selected and empaneled before the Harris decision was rendered. Accordingly, we affirm the judgment of guilt and the special circumstance finding.

At the penalty phase of the proceedings, however, the jury was given several instructions which may have seriously misled it both (1) as to the matters it should properly consider in determining sentence, and (2) even more fundamentally, as to the very nature of its ultimate task at the penalty phase. Under existing authorities, these errors clearly compel the reversal of the penalty judgment and a remand for a new penalty phase proceeding.

I.

We briefly review the evidence presented at the guilt/special circumstance and penalty phases of the trial.

A. Guilt/Special Circumstance Phase Evidence.

1. The La Verne Incident.

Cynthia Lopez, employed as a clerk at the El Rancho Seco Liquor Store in La Verne, testified that on January 7, 1979, about 7:20 p.m., three men — later identified as defendant, Michael Myers (defendant's brother), and Ivan Hill — entered the store. Lopez looked up as defendant pointed a shotgun at her and ordered her to the ground.[2]

Michael Myers and Hill went behind the counter and attempted to open the cash register. When they were unable to do so, defendant ordered Lopez to get up and open it. She complied. Defendant then told the others to take the money out of the register. After they did so, the three ran from the store, defendant pointing the shotgun at Lopez as he left. Approximately $190 was taken during the incident.

Lopez testified that defendant did not appear to be intoxicated during the incident; he did not stagger and there was nothing unusual about his speech. No defense evidence was presented to contradict Lopez's testimony.

*257 2. The Glendora Incident.

According to the statement which defendant gave to the police after his arrest, defendant, Hill and Michael Benton drove by the B&J Liquor store in Glendora in the evening hours of January 23, 1979. They saw a long-haired clerk, whom they mistook for a woman, sitting in the store watching television, and decided to rob the store. After entering and discovering that the clerk was a man, however, they abandoned their robbery plan because he was tall and they did not have a gun. They bought a few items and left.

The next day, defendant borrowed a .22 caliber revolver from Angela Bryce, a friend, telling her that he needed it because somebody had "jumped" him. Later that evening, defendant, Hill, and Benton returned to the B&J Liquor store. They drove by four times before parking and leaving the car. As they approached the store, they decided not to enter because there was a customer inside and they did not want any witnesses. They reentered their car, drove around, came back and waited in the car. Finally, they decided to go into the store even though a customer was still inside.

When the three entered, Keith Hunt, the clerk, was sitting on a stepladder eating, while Thomas Leavall, the customer, was in another area of the store. Hunt testified to the events that occurred thereafter.

Immediately after entering the store, Benton and Hill jumped over the counter and went to the cash register while defendant drew the revolver and told Hunt to freeze. When Hill was unable to open the cash register, Hunt opened the register at Hill's direction. Defendant then approached Leavall, and ordered him to get down on the floor. After Leavall complied, defendant took his wallet.

Defendant then ordered Hunt to lie down on the floor and to throw his wallet on the ground. According to Hunt, after defendant picked up the wallet, defendant took aim at Hunt's head. As Hunt ducked, he heard a pop and then a click and then felt "a sensation in [his] head." He rolled over onto his side and played dead.

Twenty to thirty seconds later, Hunt heard another pop in the area where Leavall had been. He next heard the running of feet and the click of the buzzer near the door. He opened his eyes and saw that he was bleeding. Hunt went over to the cash register, tripped the silent alarm, examined Leavall briefly and went to the bar next door for help. Eventually the police were called.

Leavall died a few weeks later from a gunshot wound to the head. Hunt survived, but required surgery on his jaw.

*258 The police investigation produced abundant evidence tying defendant to the Glendora crimes. The police obtained the gun which defendant had returned to Angela Bryce after the shootings; bullet fragments removed from Leavall and Hunt were identified as having been fired by that gun. A search of defendant's bedroom, pursuant to a warrant, uncovered Hunt's and Leavall's wallets.

After his arrest, defendant admitted the shooting to the police but told them that the gun had gone off accidentally as he was reaching for Hunt's wallet. After that, he said, he had to shoot the other man. In addition, defendant's brother, Michael, told the police that shortly after the incident, he overheard defendant say that he had "just shot two white dudes in the head."

3. The Defense.

Defendant presented a diminished capacity defense with respect to the Glendora offenses. Benton, one of the accomplices in the Glendora incident, testified that he was with defendant from 9 or 10 a.m. on the day of the incident. He stated that during that day, defendant drank several cans of Old English Malt Liquor and smoked two or three PCP cigarettes, lighting the last one around 5 p.m. Benton testified that defendant was acting unusual around the time of the robbery and shootings, in that his head was bowed and he kept grabbing his head. Although defendant did not testify at trial, in his statement to the police, which had been introduced by the prosecution and thus was before the jury, defendant said that on the day in question, he drank four cans of malt liquor and had been "mentally drunk," not knowing half the time where he was.

Dr. Bernd Schulze, a clinical pharmacist at the Los Angeles County Drug Information Center, testified to the effects of PCP ingestion. Along with two colleagues, Schulze had conducted a study of approximately 1,000 PCP-related admissions to the Los Angeles County Medical Center Jail Ward. He found that there was no single typical effect produced by PCP ingestion. PCP commonly produced various types of bizarre behavior, including violent behavior.

4. Rebuttal.

The prosecution produced several witnesses to rebut the contention that defendant had been suffering from the effects of PCP and alcohol on the night of the incident. Hunt, the store clerk, testified that at no time in the liquor store did defendant slur his words or stagger. Jennifer Moller and Dody Allen, two bystanders who saw defendant and two others run out of *259 the liquor store, testified that defendant was running fast and not stumbling or weaving as the three ran to Hill's car. Finally, Wayne Reed, who was at Angela Bryce's house when defendant returned the gun on the night of the incident, testified that defendant appeared normal.

5. Verdict.

After considering the foregoing evidence, the jury convicted defendant of all the charged crimes and enhancements, and found true the robbery-murder special circumstance allegation.

B. Penalty Phase Evidence.

1. Prosecution's Evidence.

After the initial penalty trial, the jury deliberated for five days but was unable to agree on sentence. Pursuant to the provisions of the 1978 death penalty law, a new jury was empaneled to consider penalty.

At the second penalty trial, the prosecution presented the testimony of Cynthia Lopez and Keith Hunt, who described the two liquor store robberies and the shooting of Hunt and Leavall; this testimony in substance paralleled the evidence that had been presented at the initial guilt/special circumstance phase.

In addition, the prosecution introduced evidence of two earlier incidents of violent criminal activity committed by defendant. In October 1975, when he was 15 years old and already a ward of the juvenile court, defendant snatched the purse of an elderly woman, pulling her to the ground and bloodying her knee. The following year, after he had been committed to and released from juvenile camp following the purse snatching, defendant and a friend entered a van which was parked in a drive-in theater and raped the lone woman occupant at gunpoint. Defendant was committed to the Youth Authority as a result of the rape.

The prosecution also called a psychiatrist and a number of probation officers who had examined and supervised defendant at various times during his juvenile court wardship. These witnesses generally characterized defendant as a self-centered individual, who was lacking in moral judgment and felt no remorse for actions which hurt others.

2. Defense Evidence.

Benton and Dr. Schulze, who had testified in support of defendant's diminished capacity defense at the guilt/special circumstance phase, gave similar testimony at the second penalty trial.

*260 Defendant also called his mother, who testified to his family background. Defendant was born in March 1960, and was 18 years old at the time of the offenses at issue in this case. He is one of eight children; three of his brothers have also been involved in serious criminal activity. His mother testified that she and defendant's father separated when defendant was three months old, reconciled fourteen months later, but then permanently broke up several months thereafter. Defendant lived with his mother until he was 12 years old; she testified that while he was living with her, he did well in school, frequently earning A's. In 1972, however, he went to live with his father, and thereafter defendant lost interest in school and began to get into trouble.

Defendant also called two law enforcement officials — a probation officer and a parole agent — who had supervised and treated him while he was a ward. The probation officer testified that over a five-month period in 1975, while defendant was in a juvenile camp, defendant had had no disciplinary problems, had done well academically and, overall, had performed satisfactorily. The parole agent, who had counseled him for several months in the early part of 1978, testified that during that period of time he was at the top of the juvenile facility's "honor roll," which meant that he was performing well in school, in his trade, and in everything else that was expected of him. She reported that though he had frequent opportunities to escape from the facility he made no attempt to do so, and, in fact, had been instrumental in preventing two other YA wards from escaping. She stated that in her opinion defendant performs well in a structured setting, when he has rules to follow and people to enforce them. Finally, she noted that although she had recommended that defendant be kept at a re-entry program in San Diego for a substantial period of time and not be sent home, defendant had in fact been released and returned home more quickly than she had suggested.

After counsel gave their closing arguments and the trial court instructed the jury, the jury retired to consider the issue of penalty. After a day of deliberation, the jury returned a verdict of death. This appeal followed.

II.

As noted, defendant's sole guilt/special circumstance phase contention relates to the validity of the procedure that was used in the selection of the jury panel. We review the procedural background of his claim.

Jury selection in this case began on October 7, 1980, and was completed on November 7, 1980. On October 23, 1980, the 10th day of jury selection, defendant moved for a mistrial, contending that the procedure by which the jury panel had been selected was unconstitutional. Defendant based his *261 motion on a decision rendered three days earlier, on October 20, 1980, by Judge Fainer of the Los Angeles Superior Court in a separate case, In re Rhymes (Super. Ct. Los Angeles County, No. A.P.H.C. 000042). In Rhymes, the court, after comparing statistics showing the ethnic makeup of jury panels in the Pomona Judicial District with statistics relating to the ethnic makeup of the general population of Los Angeles County and the City of Pomona, concluded in a brief order that "a pool or panel of potential jurors taken only from the voter registration lists cannot meet the constitutional test, at least in the Pomona Judicial District, as to blacks," and vacated Rhymes's misdemeanor conviction, remanding the case to the municipal court for a new trial.

Relying on the trial court ruling in Rhymes, defendant claimed that the jury panel in his case had similarly been unconstitutionally selected. He moved for a mistrial and requested that the trial court continue the case until approximately February 1981, when, "at least from the information that I have gleaned from the L.A. Times, there will be a new system used ... that may rectify the error that the [Rhymes] court has found to presently exist." The prosecution opposed the motion, contending that the challenge was untimely and that an insufficient showing had been made as to the nonrepresentativeness of the pool.

To support his motion, defendant relied almost entirely on the statistical studies, legal briefs, referee's report, and order in the Rhymes case; copies of this material were submitted to the trial court. In addition, defendant called a deputy jury commissioner, who testified that in Pomona both the municipal and superior courts draw their jurors from the same jury pool. In her testimony, the commissioner confirmed that at that time Los Angeles County compiled its jury panels solely from voter registration lists, and further explained that in January 1981 the county was planning to begin using Department of Motor Vehicle records to supplement the voter registration lists.

In opposition to the motion, the prosecution relied on the People's trial brief in the Rhymes case, taking the position that the Rhymes ruling was in error and would be reversed on appeal. The prosecution also pointed out that, as a superior court decision, the Rhymes order was not binding on the trial court in this case. (See King v. Order of Travelers (1948) 333 U.S. 153 [92 L.Ed. 608, 68 S.Ct. 488].)

After reviewing the studies, briefs and Rhymes ruling, the trial court denied defendant's motion. At the same time, the court stated "for the record" that it did not believe the challenge was untimely.

*262 Defendant asserts that the trial court erred in denying his motion for mistrial.

A.

As a threshold matter, the Attorney General argues that defendant is precluded from challenging the composition of the jury panel on appeal because his challenge to the panel in the trial court came too late. Although the trial court found the claim timely, the Attorney General maintains that, on this point, its ruling was in error. In light of the governing statutes, we agree with the trial court.

(1a) Penal Code section 1060 provides in relevant part that "[a] challenge to the panel must be taken before a juror is sworn. ..." (Italics added.) The Attorney General suggests that the emphasized language refers to the swearing of a juror before voir dire, rather than to the swearing of a juror to try the case after voir dire. Although we have found no case that has addressed this precise question,[3] the statutory provisions accompanying section 1060 and the legislative history of the provisions belie the Attorney General's suggested interpretation. Section 1066, which concerns challenges to individual jurors, rather than a challenge to the jury panel, uses the same language as section 1060, directing the court to inform the defendant "that if he intends to challenge an individual juror he must do so when the juror appears, and before he is sworn." (Italics added.) Since a defendant obviously could not be required to challenge an individual juror before the voir dire examination, the Legislature apparently used the phrase "is sworn" in this statutory scheme to refer to the swearing of a juror to try the cause. This interpretation also finds support in the statutory predecessors of sections 1060 and 1066, for the earlier statutes also used parallel language to govern the timing of challenges to a jury panel and to individual jurors. (See Stats. 1851, ch. 29, §§ 331, 339, 341; see also Pen. Code of 1872, notes to §§ 1060, 1068.) Thus, while it might well be reasonable — as a matter of policy — to require a defendant to interpose any challenge to a jury panel before the voir dire examination of individual jurors begins,[4]*263 we conclude that under the governing statutes the trial court properly found the challenge in this case timely.

B.

Turning to the merits, defendant contends that our decision in People v. Harris, supra, 36 Cal.3d 36, establishes that his statistical showing was sufficient to constitute a prima facie case of an unconstitutional jury selection process; because the prosecution failed to come forward with evidence to attempt to justify the disparity, he asserts that the trial court erred in failing to grant a mistrial. In response, the Attorney General maintains (1) that the Harris decision should not be applied retroactively to overturn convictions by juries selected pursuant to well-established procedures before the Harris decision, and (2) that even if Harris is applicable, the trial court's ruling should in any event be upheld because defendant's statistics related to the wrong geographical area, i.e., to the entire county rather than to the community within a 20-mile radius of the Pomona courthouse. The question of the proper application of Harris in areas where the 20-mile rule is in effect is currently before our court in another case. (Williams v. Superior Court, L.A. 32206.) We need not reach that issue here, however, because, under the governing authorities and legal principles, we conclude that Harris should not be applied retroactively.

We begin with the Harris decision itself. In Harris, the defendant moved to quash the jury venire, claiming that Los Angeles County's use of voter registration lists as the single source for jury selection resulted in unconstitutionally unrepresentative jury panels in the Long Beach court in which he was to be tried. In support of his claim, Harris relied on statistics demonstrating a significant disparity between the percentage of Black and Hispanic persons on Long Beach jury panels and the percentage of Black and Hispanic persons in the population of Los Angeles County as a whole. The trial court denied the motion, concluding that defendant's showing was inadequate to establish a prima facie case because the general-population figures on which defendant relied were not necessarily a reliable indicator of the jury-eligible population pool, and thus defendant's statistics did not necessarily demonstrate that the makeup of the actual jury panels was unrepresentative of the jury-eligible pool.

On appeal, we noted that "[t]he principal question before us is whether evidence based on total countywide population figures, rather than jury-eligible population, is adequate to make out a prima facie case...." (36 Cal.3d at p. 48.) In addressing that question, we recognized that "[c]riticism has been appropriately leveled at the use of total population figures as the comparison base for showing underrepresentation [citations]" *264 (id., at p. 53), explaining that "[c]omparing the makeup of the actual jury pool with that of the entire population of persons presumptively eligible for jury service would be preferable to comparing with total population because if the cognizable class has a lower percentage of jury eligibles than the general population, a showing that that class' representation in the jury pool is less than the group's percentage of the general population does not necessarily show that the group is underrepresented." (Id., at p. 54.)

At the same time, however, we recognized that it is frequently very difficult to arrive at reliable figures of the jury-eligible population for any particular cognizable class. (Id., at pp. 53-54.) Observing that "[t]he question raised by the difficulty of producing statistics concerning the precise proportion of jury eligibles" (id., at p. 54) comes down to the issue of "who should bear the burden of proof on whether a disparity exists" (ibid.), we concluded that "although more refined statistics would be preferable if available, when they are not, it is sufficient for the defendant to show a significant disparity based on the use of total population figures. The burden then shifts to the state to either show that when the group total population is reduced to jury eligibles, no underrepresentation exists, or to justify the underrepresentation by showing `that a significant state interest [is] manifestly and primarily advanced by those aspects of the jury-selection process, such as exemption criteria, that result in the disproportionate exclusion of a distinctive group.' [Citation.]" (Id., at pp. 54-55.)

Applying this analysis to the Harris case itself, the plurality opinion in Harris concluded that because Harris's statistical evidence was sufficient to make out a prima facie case under the opinion's test, and because the prosecution had not attempted to rebut the defense claim "but [had] short-sightedly rested its entire argument on the mistaken claim that defendant failed to present a prima facie case" (id., at p. 59), the judgment should be reversed. A concurring opinion — which provided the crucial fourth vote in Harris — agreed that Harris's showing "should be regarded as sufficient to trigger further inquiry" (id., at p. 71 (Grodin, J., conc.)), but suggested that a remand to the trial court for such inquiry, rather than reversal of the conviction, was the proper disposition. (Id., at pp. 71-72.) In order to enable the court to enter a dispositive judgment in the case, however, the concurring justice "reluctantly join[ed] in the reversal of the judgment below." (Id., at p. 72.) Three justices dissented, concluding that Harris had not established a prima facie case. (Id., at p. 72 (Mosk, J., dis., joined by Richardson, J.); id., at p. 75 (Kaus, J., dis.).)

Harris expressly left open the question whether the decision would apply retroactively. The lead opinion stated in this regard: "A majority of the justices of this court ... do not agree on whether, and to what extent, the *265 rule announced in this case should be given retroactive effect. We therefore take no position as to the disposition of other cases presenting issues concerning the representative character of juries selected from voter registration lists alone." (Id., at p. 59.)

Although Harris left the retroactivity question unresolved, we do not write on a clean slate in addressing that issue here. Six months after Harris was filed, the Court of Appeal explicitly addressed the question of the retroactive application of Harris in People v. Cantu (1984) 161 Cal. App.3d 259 [207 Cal. Rptr. 460]. After a thorough analysis of the applicable retroactivity principles and precedents (161 Cal. App.3d at pp. 267-271), the Cantu court concluded that Harris should not apply retroactively. We denied a petition for hearing in Cantu in January 1985, and then subsequently retransferred two cases presenting Harris issues to the Court of Appeal for reconsideration in light of the Cantu decision. In both cases, the Courts of Appeal thereafter followed Cantu and held that Harris does not apply to cases in which the jury was selected before the Harris opinion was rendered. (2) (See fn. 5.) See People v. Pendleton (1985) 167 Cal. App.3d 413, 416-418 [212 Cal. Rptr. 524]; People v. Brown (1985) 169 Cal. App.3d 728, 734-735 [215 Cal. Rptr. 465].)[5]

Another Court of Appeal decision also sheds light on the retroactivity question presented here. At the time Harris was decided, we had already granted a hearing in the case of In re Rhymes, the decision which formed the basis for defendant's trial court motion in the present case. In Rhymes, the Court of Appeal — in a decision filed before Harris — had affirmed the trial court's ruling in that case, sustaining Rhymes's challenge to the Pomona jury venire. The Court of Appeal decision in Rhymes, however, had specifically directed that its ruling would not apply retroactively, stating: "Our holding shall apply to this case only, until this case becomes final, and thereafter shall apply prospectively. We so provide because of the extent of reliance on this single source for selection of jury venires and because a retroactive application would have a detrimental effect on the administration of justice. [Citation.]" In June 1985, after Harris was final, we acted on the pending Rhymes matter, retransferring the case to the Court of Appeal, *266 "with directions to refile its opinion." The Court of Appeal thereafter refiled its original opinion, with the prospectivity discussion intact. (See In re Rhymes (1985) 170 Cal. App.3d 1100, 1114 [217 Cal. Rptr. 439].)

Thus, while we have not yet had occasion to address the Harris retroactivity in one of our own decisions, our actions since Harris with regard to the relevant Court of Appeal decisions uniformly support the conclusion that Harris should not apply retroactively. Although defendant argues that these Court of Appeal decisions were wrongly decided and should be disapproved, we do not agree.

The United States Supreme Court decision in Daniel v. Louisiana (1975) 420 U.S. 31 (42 L.Ed.2d 790, 95 S.Ct. 704] provides perhaps the most direct support for the conclusion that Harris should be applied only prospectively. In Daniel, the Supreme Court was faced with the question whether its then-recent decision in Taylor v. Louisiana (1975) 419 U.S. 522 [42 L.Ed.2d 690, 95 S.Ct. 692] — invalidating a jury selection process which largely excluded women from jury panels — should be applied retroactively or prospectively. The Daniel court, noting that the seminal decision in Duncan v. Louisiana (1968) 391 U.S. 145 [20 L.Ed.2d 491, 88 S.Ct. 1444] — which first held that the Fourteenth Amendment's due process clause incorporated the Sixth Amendment's jury trial guaranty — had been given only prospective application in DeStefano v. Woods (1968) 392 U.S. 631 [20 L.Ed.2d 1308, 88 S.Ct. 2093], concluded that Taylor should similarly be applied only prospectively.

(3a) The Daniel court explained: "The three relevant factors, as identified in Stovall v. Denno, 388 U.S. 293, 297 [18 L.Ed.2d 1199, 1203, 87 S.Ct. 1967] (1967), are `(a) the purpose to be served by the new standards, (b) the extent of the reliance by law enforcement authorities on the old standards, and (c) the effect on the administration of justice of a retroactive application of the new standards.' In Taylor, as in Duncan, we were concerned generally with the function played by the jury in our system of criminal justice, more specifically the function of preventing arbitrariness and repression. In Taylor, as in Duncan, our decision did not rest on the premise that every criminal trial, or any particular trial, was necessarily unfair because it was not conducted in accordance with what we determined to be the requirements of the Sixth Amendment. In Taylor, as in Duncan, the reliance of law enforcement officials and state legislatures on prior decisions of this Court, ... in structuring their criminal justice systems is clear. Here, as in Duncan, the requirement of retrying a significant number of persons were Taylor to be held retroactive would do little, if anything, to vindicate the Sixth Amendment interest at stake and would have a substantial impact on the administration of criminal justice in Louisiana and in *267 other States whose past procedures have not produced jury venires that comport with the requirement enunciated in Taylor." (420 U.S. at pp. 32-33 [42 L.Ed.2d at pp. 792-793].) Accordingly, although the jury that had convicted Daniel had been drawn from a panel that did not comport with the Taylor holding and Daniel had timely raised the issue at his trial, the Supreme Court nonetheless affirmed Daniel's conviction.

(4a) The Daniel court's reasoning applies fully to the present context. First, like Taylor, our decision in Harris was aimed at furthering Sixth Amendment values by assuring that the procedures used by the state in the selection of juries generally produce jury venires that are fairly representative of the relevant community; Harris, like Taylor, "did not rest on the premise that every criminal trial, or any particular trial, was necessarily unfair" (Daniel, supra, 420 U.S. at p. 32 [42 L.Ed.2d at p. 793]) simply because the jury selection process had not been conducted or evaluated in accordance with the Harris decision.[6]

Second, again like Taylor, the reasonable reliance of local authorities on the pre-Harris precedents appears clear. In the decades before Harris, voter registration lists had been widely used as a source for jury panel selection both in California and throughout the country with repeated judicial approval.[7] Although pre-Harris decisions had clearly left open the possibility that the use of such lists could be subject to challenge if they resulted "`in *268 the systematic exclusion of a "cognizable group or class of qualified citizens"'" (see, e.g., People v. Sirhan, supra, 7 Cal.3d 710, 749-750), virtually all the California decisions which had addressed challenges to the use of voter registration lists before Harris had rejected evidentiary presentations which relied on general population, rather than jury-eligible, statistics.[8] Under the circumstances, we cannot properly fault local officials for failing to alter their jury selection procedures in anticipation of Harris.[9]

Third, and finally, because of the widespread use of voter registration lists as the source of jury panels,[10] a retroactive application of Harris would likely, as the Rhymes court observed, "have a detrimental effect on the administration of justice." (Rhymes, supra, 170 Cal. App.3d at p. 1114.) Cantu, supra, 161 Cal. App.3d 259, Pendleton, supra, 167 Cal. App.3d 413, and Brown, supra, 169 Cal. App.3d 728, demonstrate that similar challenges to the single-source procedure were not uncommon in the pre-Harris era.

*269 Indeed, in an important respect, the prospective application of Harris follows a fortiori from the United States Supreme Court's prospective application of Taylor. Under Taylor, it was, of course, clear that the juries that had tried the Daniel case and the numerous other pre-Taylor matters did not represent a cross-section of the community; by contrast, we do not know whether pre-Harris juries were, in fact, unconstitutionally unrepresentative, for Harris simply created a new standard for determining the existence of a prima facie case, and did not establish the unconstitutionality of prior jury panels. Inasmuch as Daniel held that prospective application of Taylor was warranted even though the prior jury panels were not representative, it would surely be anomalous to apply our Harris decision retroactively to overturn numerous past convictions, when it is not at all clear that the juries in those past cases were in fact unconstitutionally unrepresentative.[11]

(5a) Accordingly, we agree with the decisions in Cantu, supra, 161 Cal. App.3d 259, Brown, supra, 169 Cal. App.3d 728, and Pendleton, supra, 167 Cal. App.3d 413, and hold that Harris should not be applied retroactively to cases in which juries were selected before the Harris decision was rendered. Because the trial court's rejection of defendant's challenge to the jury panel in this case was correct under the "pre-Harris" authorities, what ruling provides no basis for reversing defendant's conviction.[12] Since *270 defendant raises no other challenge to the guilt/special circumstance phase, we affirm the judgment of guilt and the special circumstance finding.

III.

As noted above, at the initial penalty trial the jury was unable to agree on sentence. A new jury was empaneled, and after the second penalty trial, the jury returned a sentence of death.

As we explain, under recent decisions of this court it is clear that the instructions which the jury was given at the penalty phase were improper in a number of significant respects. Because of the nature and extent of the instructional errors, the penalty judgment must be reversed and the case remanded for a new penalty trial before a properly instructed jury. We review the most significant of the instructional errors.

A. Briggs Instruction.

(6a) At the second penalty trial, the court instructed the jury, pursuant to the terms of the 1978 death penalty law (Pen. Code, § 190.3, 5th par.), that "under the state Constitution a Governor is empowered to grant a pardon or commutation of a sentence following the conviction of the crime. [¶] Under this power a Governor may, in the future, commute or modify a life sentence without possibility of parole to a lesser sentence that would include the possibility of parole."

In People v. Ramos (1984) 37 Cal.3d 136 [207 Cal. Rptr. 800, 689 P.2d 430], we examined the validity of this instruction — the so-called "Briggs Instruction" — at some length, and concluded that "when viewed in relation to prior California decisions and to the overwhelming majority of related precedents in other states, the Briggs Instruction is incompatible with [the "due process"] guarantee of `fundamental fairness' both because it is seriously and prejudicially misleading and because it invites the jury to be *271 influenced by speculative and improper considerations." (37 Cal.3d at p. 153.)

On the first point — the misleading nature of the instruction — Ramos explained that while "[u]nder the California Constitution, the Governor's power of commutation or pardon extends equally to a sentence of death and to a sentence of life without possibility of parole ... [t]he Briggs Instruction ... informs the jury only that a sentence of life without possibility of parole may be commuted." (Ibid.) As such, Ramos found the instruction "a classic example of a misleading `half-truth' ... [which] would reasonably be understood by the average juror to mean, by negative implication, that while a sentence of life without possibility of parole may be commuted, a sentence of death may not." (Ibid.) Furthermore, Ramos emphasized that the potential consequence of misleading the jury in this fashion was extremely serious. "[T]he pernicious effect of the Briggs Instruction is that it may lead a jury that does not believe that the death penalty is necessary, but fears a future commutation, to return a death penalty in the mistaken belief that that sentence alone will preclude any possible release. Because an accurately informed jury would at least realize that the possibility of gubernatorial action cannot be avoided in any event, it is less likely to return a death sentence when it is not convinced that death is warranted. The prejudice to the defendant is manifest." (Id., at p. 154.)

(7a) Moreover, Ramos went on to hold that even if the Briggs Instruction were modified to make it totally accurate, "the instruction would still violate the state constitutional due process guarantee because its reference to the commutation power invites the jury to consider matters that are both totally speculative and that should not, in any event, influence the jury's determination." (Id., at p. 155.) As Ramos pointed out, by raising the commutation issue, the instruction inevitably leads the jury to attempt to predict not only "what a particular defendant is likely to be like some 10, 15, 20 or more years in the future when commutation may be considered" (id., at p. 156), "but also what some presently unknown person — a future Governor — will do in response to the defendant's then condition.... `When a jury's attention is ... thrust into speculation about the future action of as yet unknown actors, a serious possibility arises that each death sentence imposed ... is the result of an interjection of an unquantifiable factor into the deliberation process, thereby rendering the decision arbitrary....'" (Id., at p. 157.) Finally, Ramos makes clear that, beyond the problems of the speculative nature of the inquiry, such an instruction is improper because it is likely to divert the jury from its proper function, either by "diminishing the jury's appreciation of its personal responsibility for the sentencing decision" (id., at p. 158), or by "invit[ing] the jury to second-guess a future Governor's exercise of his constitutional authority and to impose a harsher sentence *272 than it might otherwise impose simply out of fear that the Governor and the parole authorities will make a mistake and will release the defendant while he is still dangerous." (Ibid.) As Ramos concluded, "[t]o permit a jury to act in this fashion is inconsistent with a defendant's right under the California Constitution to have the commutation decision made by the Governor and undermines the fairness of the jury's determination." (Ibid.)

(8a) The Attorney General concedes that, in light of Ramos, instructing the jury pursuant to the Briggs Instruction was error, but he contends that we should find the error harmless. The Attorney General, however, has cited no instance, and we are aware of none, in which this type of instructional error has been found nonprejudicial in a death penalty case, and in view of the very serious potential for prejudice emphasized in Ramos, we strongly doubt that we could ever confidently conclude that there was no reasonable possibility that this instruction improperly tainted the jury's decision-making process. (See, e.g., People v. Montiel (1985) 39 Cal.3d 910, 928 [218 Cal. Rptr. 572, 705 P.2d 1248].)

(9a) In any event, we certainly cannot say the error was harmless here. Because the trial court, over defendant's objection, indicated at the beginning of the second penalty trial that it intended to give the Briggs Instruction, defense counsel — in an attempt to counteract the instruction — introduced evidence at the penalty phase relating to the past practices of California governors with respect to commutation and argued to the jury that it was extremely unlikely that defendant would ever be released if he were sentenced to life without possibility of parole. In addition, the trial court — at defendant's request — supplemented the normal penalty phase instructions with a lengthy instruction explaining the commutation process.[13] Although *273 the Attorney General argues that the defense evidence and argument, and the court's supplemental instruction, "effectively neutralized" any prejudice caused by the Briggs Instruction, in reality the additional focus on commutation in this case had the inevitable and unfortunate effect of highlighting the ostensible importance of the commutation question. As we have seen, Ramos clearly points out the numerous untoward consequences that result when the jury's attention is improperly directed to the issue of commutation; although the defense evidence and the court's instruction may have given the jury a more complete and accurate picture of the commutation procedure than it otherwise would have had, the fact remains that the jury was never told "that it would be a violation of [its] duty to consider the possibility of such commutation in determining the appropriate sentence" (see Ramos, supra, 37 Cal.3d at pp. 159-160, fn. 12), but was instead left with the understanding that it would be appropriate to consider commutation. Under the circumstances, we think it clear that this error alone requires a new penalty trial. (See, e.g., People v. Haskett (1982) 30 Cal.3d 841, 861-863 [180 Cal. Rptr. 640, 640 P.2d 776].)

B. Unclarified "Mandatory" Sentencing Instruction.

(10) In advising the jury as to the appropriate "standard" or "test" it was to apply in determining which penalty — death or life imprisonment without possibility of parole — should be imposed on defendant, the court instructed the jury, in language which paralleled the 1978 statute (§ 190.3, 7th par.), that "[i]f you conclude that the aggravating circumstances outweigh the mitigating circumstances, you shall impose a sentence of death; however, if you determine that the mitigating circumstances outweigh the aggravating circumstances, you shall impose a sentence of confinement in the State Prison for life without possibility of parole."

In People v. Brown (1985) 40 Cal.3d 512, 538-545 [220 Cal. Rptr. 637, 709 P.2d 440], we rejected a claim by the defendant that the foregoing language demonstrated that the 1978 enactment had established an unconstitutional "mandatory" death penalty statute. At the same time, however, we acknowledged "that the language of the statute, and in particular the words `shall impose a sentence of death,' leave room for some confusion as to the jury's role." (Id., at p. 545, fn. 17.) Explaining that "such confusion [had] occasionally [been] reflected in records before this court" (ibid.), we directed trial courts, in future cases, "[to] instruct the jury as to the scope of its discretion and responsibility in accordance with the principles set forth in [Brown]" (ibid.), rather than simply in the terms of the 1978 statute. *274 With respect to cases — like this one — which were tried before Brown, we explained that "[e]ach such prior case must be examined on its own merits to determine whether, in context, the sentencer may have been misled to defendant's prejudice about the scope of its sentencing discretion under the 1978 law." (Ibid.)

As noted in our recent decision in People v. Allen (1986) 42 Cal.3d 1222, 1276-1277 [232 Cal. Rptr. 849, 729 P.2d 115], "we explained in People v. Brown ... that the statute's `weighing process' described in former CALJIC No. 8.84.2, is the method by which `the jury ... determines under the relevant evidence which penalty is appropriate in a particular case.' [Citation.] Our concern in Brown was that the unadorned statutory instruction might in two interrelated ways lead the jury to misapprehend its discretion and responsibility. [¶] First, we pointed out that the jury might be confused about the nature of the weighing process. As we observed: `[T]he word "weighing" is a metaphor for a process which by nature is incapable of precise description. The word connotes a mental balancing process, but certainly not one which calls for a mere mechanical counting of factors on each side of the imaginary "scale," or the arbitrary assignment of "weights" to any of them. Each juror is free to assign whatever moral or sympathetic value he deems appropriate to each and all of the factors he is permitted to consider.' [Citation.] [¶] Second, we were concerned in Brown that the unadorned instruction's phrase, `the trier of fact ... shall impose a sentence of death if [it] concludes that the aggravating circumstances outweigh the mitigating circumstances' (emphasis added), could mislead the jury as to the ultimate question it was called on to answer in determining which sentence to impose. Although the quoted phrase could be understood to require a juror (i) to determine whether `the aggravating circumstances outweigh the mitigating circumstances' without regard to the juror's personal views as to the appropriate sentence, and then (ii) to impose a sentence of death if aggravation outweighs mitigation even if the juror does not personally believe that death is the appropriate sentence under all the circumstances, we concluded in Brown that the statute was not intended to, and should not, be interpreted in that fashion. Instead we stated: `By directing that the jury "shall" impose the death penalty if it finds that aggravating factors "outweigh" mitigating, the statute should not be understood to require any juror to vote for the death penalty unless, upon completion of the "weighing" process, he decides that death is the appropriate penalty under all the circumstances. Thus the jury, by weighing the various factors, simply determines under the relevant evidence which penalty is appropriate in the particular case.' [Citation (italics added).]"

(11) In the present case, the trial court did not give any supplementary instructions to the jury which expanded, or clarified, the potentially ambiguous *275 language of the 1978 law. Nonetheless, with respect to the initial source of potential confusion — the nature of the "weighing" process — we think it is unlikely that the jury was misled into believing that the instruction called for a "mechanical" "counting" of the relevant aggravating and mitigating factors. In closing argument, both the prosecution and the defense indicated to the jury that it was free to "attach whatever weight is appropriate" to each of the relevant factors, and while the prosecutor did, at one point, use a "scale" motif in his argument, he did not suggest that the weighing process was a mechanical or arithmetic operation.

(12) With respect to the second area of potential confusion — the nature of the ultimate question which the jury must answer in determining which sentence to impose — however, there seems no question but that the jury may well have been misled to defendant's prejudice. Unlike the recent case of People v. Allen, supra, 42 Cal.3d at pages 1278-1280, in which we found that the prosecutor's argument would have led any reasonable juror to understand that his ultimate task was to determine whether death or life imprisonment without parole was the appropriate penalty for the defendant under all the circumstances, the prosecutor in the present case described the jurors' task at the penalty phase in the following terms: "Your job is a fact-finding process. The law dictates what the penalty shall be depending upon what weight you give, what the total weight is to each aggravating or mitigating circumstance; and you are instructed ... that you are to weigh the aggravating factors against the mitigating factors to determine which weighs the most and then the result is determined by that process. [¶] It would not be appropriate for you to determine what result you want to obtain and then seek to shade the factors or the weight to give to the various factors. That would not be fulfilling your duties as a juror, and I am sure each of you will recognize the importance in the jury system of fulfilling your obligation as a juror and doing that which the law instructs you to do."[14]

*276 As we have seen, Brown, supra, 40 Cal.3d 512, makes it clear that this is not a proper understanding of the jury's task under the 1978 law. The jury is not simply to determine whether aggravating factors outweigh mitigating factors and then impose the death penalty as a result of that determination, but rather it is to determine, after consideration of the relevant factors, whether under all the circumstances "death is the appropriate penalty" for the defendant before it. In light of the prosecutor's characterization of the jury's function — a characterization that was not refuted or corrected by the court's instructions — there is clearly a reasonable possibility that the jury was misled as to the nature of its ultimate duty at the penalty phase. This error in itself is sufficient to require a new penalty trial.

C. Additional Penalty Phase Contentions.

In addition to the Briggs Instruction and the "mandatory" sentencing question, defendant contends that the penalty judgment must be reversed because (1) the prosecutor introduced some evidence at the penalty phase that is not properly admissible under People v. Boyd (1985) 38 Cal.3d 762 [215 Cal. Rptr. 1, 700 P.2d 782], and (2) the court gave several additional instructions that our decisions establish should not be given at the penalty phase of a capital case. (See, e.g., People v. Lanphear (1984) 36 Cal.3d 163, 165 [203 Cal. Rptr. 122, 680 P.2d 1081] [no-sympathy instruction]; People v. Brown, supra, 40 Cal.3d 512, 538, fn. 7 ["regardless-of-what-the-consequences-may-be" instruction].) Because we have concluded that both the Briggs Instruction and the "mandatory" sentencing error independently require that the penalty judgment be reversed, we need not decide whether the additional matters raised by defendant would in themselves require reversal. Inasmuch as any new penalty trial will be conducted with full awareness of our recent rulings, there is no reason to anticipate that any error will recur.

IV.

CONCLUSION

The judgment as to guilt, and the finding of a special circumstance, are affirmed. With respect to penalty, the judgment is reversed and the case remanded for a new penalty trial before a properly instructed jury. If the prosecution, for any reason, chooses not to retry the issue of penalty, defendant shall be sentenced to life imprisonment without possibility of parole.

Mosk, J., concurred.

*277 LUCAS, J.

(1b-9b) I concur in the majority's affirmance of the guilt and special circumstance findings. I also concur in the reversal of the death penalty, but only under compulsion of Ramos. (People v. Ramos (1984) 37 Cal.3d 136 [207 Cal. Rptr. 800, 689 P.2d 430].) I prefer to withhold discussion of the Brown issue pending the United States Supreme Court's decision in that case. (People v. Brown (1985) 40 Cal.3d 512 [220 Cal. Rptr. 637, 709 P.2d 440], cert. granted, 476 U.S. 1157 [90 L.Ed.2d 717, 106 S.Ct. 2274].)

Panelli, J., concurred.

BIRD, C.J.

I respectfully dissent from the majority's affirmance of the convictions and special circumstance finding.

Appellant made a timely and adequate prima facie showing that his jury venire was improper in that Blacks were systematically excluded by the state's selection methods. The state failed to rebut that showing.

Further, given the similarities between the record in this case and the records in People v. (Lee) Harris (1984) 36 Cal.3d 36 (201 Cal. Rptr. 782, 679 P.2d 433] (hereafter Harris), certiorari denied (1984) 469 U.S. 965 [83 L.Ed.2d 301, 105 S.Ct. 365] and In re Rhymes (1985) 170 Cal. App.3d 1100 [217 Cal. Rptr. 439] (hereafter Rhymes), as well as the fact that this case arose within a few months of the lower court proceedings in Harris and Rhymes, I cannot concur in the majority's arbitrary decision to deny appellant the benefit of those decisions. There is no logic in refusing to apply those rulings to a similarly situated individual where (1) the records below are nearly identical and (2) but for the time it took the trial court to certify the record on appeal and this court to complete its task, this case, rather than Harris and Rhymes, might have made it out the appellate door faster.

As I shall demonstrate, this court should find appellant's underrepresentation claim meritorious for two interrelated reasons: (1) his prima facie showing of underrepresentation was adequate under Harris and Rhymes; and (2) the holdings in Harris and Rhymes are applicable in this case. A third point — that a reversal rather than a remand is appropriate — will also be discussed.[1] The factual and legal background necessary to analyze these issues is presented as each is considered.

*278 I.

The first question is whether appellant made a prima facie showing that his venire was underrepresentative of a cross-section of the community. This issue requires an examination of Harris and Rhymes.

"It is a fundamental tenet that a criminal defendant is entitled to trial by an impartial jury drawn from a representative cross-section of the community. This right is guaranteed by the Sixth Amendment to the federal Constitution (Taylor v. Louisiana (1975) 419 U.S. 522, 530 [42 L.Ed.2d 690, 698, 95 S.Ct. 692]) as well as by article I, section 16 of the California Constitution (People v. Wheeler (1978) 22 Cal.3d 258, 272 [148 Cal. Rptr. 890, 583 P.2d 748])." (Harris, supra, 36 Cal.3d at pp. 48-49.)

"In order to establish a prima facie violation of the fair-cross-section requirement, the 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 of such persons in the community; and (3) that this underrepresentation is due to systematic exclusion of the group in the jury-selection process." (Duren v. Missouri (1979) 439 U.S. 357, 364 [58 L.Ed.2d 579, 586-587, 99 S.Ct. 664].)

Only the second and third prongs of the Duren test were at issue in Harris, since it was clear — indeed, the state conceded — that Blacks and Hispanics, the groups alleged to be underrepresented, are distinctive or cognizable groups for cross-section purposes. (Harris, supra, 36 Cal.3d at pp. 50-51.) Those groups were the focus of appellant's motion here, and, as in Harris, the state has conceded their cognizability under the Duren test.[2]

Harris involved a challenge to the jury venire at the Long Beach courthouse in Los Angeles County. The accused contended that the use of voter registration lists as the sole source for potential jurors resulted in significant underrepresentation and systematic exclusion of Blacks and Hispanics.

*279 To buttress his claim, Harris relied on testimony and statistical data provided by Dr. Edgar Butler. Butler obtained and compared figures for the Black and Hispanic population of the county as a whole with the ethnic breakdown of potential jurors who appeared at the Long Beach courthouse for a three-month period in 1979. (36 Cal.3d at pp. 46-47.) Using 1980 census figures, as well as Butler's 1970 census figures and 1975 figures derived from the 1970 figures, a plurality of this court found the comparative disparities[3] for Blacks to be 49 percent (1970 figures), 53 percent (1975 figures), and 56.3 percent (1980 figures); and for Hispanics to be 81 percent (1970), 84 percent (1975) and 87.7 percent (1980). (Id., at pp. 47-48.)

At issue in Harris was the propriety of using total population figures rather than the more narrowly defined data reflecting only persons presumptively eligible for jury service such as those eligible to vote (hereafter "jury-eligibles"). (Id., at pp. 45, 51.) Harris held that "because of the difficulty in obtaining more accurate figures for jury eligibility, the defendant can present a prima facie case by showing through the use of total population figures a significant underrepresentation of a cognizable class. The burden then shifts to the state to demonstrate either that with more refined statistics, the underrepresentation would be reduced to a constitutionally insignificant disparity, or that there exists a compelling justification for the procedure which results in the underrepresentation." (Id., at pp. 45-46.) Since appellant made such a prima facie showing and the state failed to rebut that proof, "shortsightedly rest[ing] its entire argument on the mistaken claim that defendant failed to present a prima facie case" (id., at p. 59), reversal was required (ibid.).

Justice Grodin provided the fourth vote for reversal, but "[w]ith reluctance." (36 Cal.3d at p. 71.) He agreed that "defendant's showing [based on total population figures] should be regarded as sufficient to trigger further *280 inquiry." (Id., at p. 71, fn. omitted (conc. opn.).) He believed, however, that an accused "can at least be expected to refine those statistics on the basis of readily available census information reflecting the relative percentages of majority and minority populations over the age of 18." (Id., at p. 71, fn. 1.) As to remedy, he preferred that the matter be remanded to permit the state to rebut the prima facie showing. (Id., at pp. 71-72.) This point is discussed more fully, post, at pages 289-293.

The Attorney General does not take issue with the "total population figure" holding of Harris. Instead, he focuses on a question which he believes still lingers after Harris, and on which the members of the Harris court divided sharply. That question is whether an accused must make a prima facie showing of underrepresentation as to the area served by the particular courthouse in which the case is to be tried (e.g., a 20-mile radius from the courthouse, see Code Civ. Proc., § 203[4]), or whether the showing may be based on the county population as a whole.

In Harris, the plurality noted that both parties presented evidence and argued the case "on the assumption that all juries in Los Angeles County must be representative of the entire county." (36 Cal.3d at p. 48.) The plurality went on to note that "[t]he state has not attempted to rebut this prima facie showing by arguing that the Long Beach juries need only represent those persons living within 20 miles of the courthouse, and has not attempted to show that such juries were truly representative of that limited area." (Ibid., italics added.)

As noted, Justice Grodin, in concurrence, agreed that there had been a prima facie showing based on countywide figures. However, he noted that there "may be merit ... in the dissenters' view that the more appropriate focus for statistical analysis is the area within a 20-mile radius from the Long Beach courthouse, but the People did not challenge the statistics on that ground and this court has been presented with no basis for taking judicial notice of the geographical distribution of racial and ethnic populations within Los Angeles County. If my view as to the appropriate disposition were to prevail ... that would be a matter of inquiry upon remand." (Harris supra, 36 Cal.3d at p. 71, fn. 1.)

The three dissenters in Harris clearly did not share the view that the county is the appropriate focus for cross-sectional purposes. Noting that the relevant statute requires only a cross-section of the "area served by the *281 court" (Code Civ. Proc., § 197), Justice Mosk, with Justice Richardson concurring, suggested that "the community or area that is Long Beach" or "the supervisorial district in which the city is located (see Code Civ. Proc., § 206a)" should be the relevant focus. (36 Cal.3d at p. 73.) Since "defendant has produced no statistics relating to the ethnic composition" of these areas (ibid.), Justice Mosk presumably would have affirmed the trial court's finding that Harris failed to make a prima facie showing.

Justice Kaus took a similar stance. He argued that "there is no showing that population figures for the whole of Los Angeles County are relevant with respect to a 20-mile radius from Long Beach." (Id., at p. 75.) He also deemed irrelevant the state's failure to argue that point in the trial court. (Ibid.)

From this review of the opinions in Harris, it is clear that four members of this court were of the view that a prima facie showing of underrepresentation could be made with countywide figures,[5] but that such a showing might be rebutted with statistics indicating the lack of underrepresentation when the effects of the 20-mile rule (or some less-than-countywide rule) are taken into account. Therefore, the question becomes whether appellant made a sufficient showing.

As noted, the primary basis for appellant's motion consisted of several documents from the record before the appellate department in Rhymes. Specifically, appellant submitted (1) a report by Dr. Butler — similar to the Long Beach report — which studied 805 jurors who were summoned to the Pomona courthouse from May through August of 1979 and compared the racial makeup of Pomona Superior Court jury panels with countywide population; (2) the referee's findings from the evidentiary hearing; (3) three briefs filed in the appellate department; and (4) a copy of Judge Fainer's memorandum decision granting relief.

Basing his statistics, as in Harris, on total county population figures, Butler found absolute disparities of Blacks of 4.6 percent (using 1970 population figures) and of 5.4 percent (using 1975 figures); and comparative disparities of 43 percent (using 1970 figures) and 47 percent (using 1975 figures). For Hispanics, the absolute disparities were 11.2 percent (1970) and 14 percent (1975), while the comparative disparities were 61 percent (1970) and 67 percent (1975). Butler also reviewed in a much less detailed fashion other panels for August and September of 1979 and March and April of 1980.

*282 Butler concluded that (1) White male and female representation on the panels was balanced; (2) the panels were overrepresented by Caucasians, higher educated persons, persons of higher income, and persons of middle age; (3) there was an underrepresentation of Blacks; and (4) Blacks and Hispanics were less likely to register to vote.

The referee's report in Rhymes summarized the testimony of several witnesses who were called at the evidentiary hearing in that case. These witnesses established that there had been increases in the percentage of Blacks and Spanish-surnamed individuals in the Pomona district from 1975 to 1977, that this increase was expected to continue through 1980, and that 90 percent of the jurors called for service request that their service be limited to within 20 miles of their residence.

As noted, the appellate department decision in Rhymes was the first to find the makeup of the Pomona venires unconstitutionally underrepresented. The state appealed that decision, and in 1982, the Court of Appeal affirmed. (Rhymes, supra, 170 Cal. App.3d 1100.) That court, presaging our opinion in Harris, found that while several cases had previously concluded "that no reliable conclusion can be drawn when total population rather than voter population or eligible population is used [citations], uncertainty in the ratio between the eligible and the total population does not preclude a court from finding underrepresentation when total population is used as the base." (Rhymes, supra, 170 Cal. App.3d at pp. 1111-1112, fn. omitted.) The court concluded that the statistical disparities in the case were "considerable for both Blacks and Hispanics." (Id., at p. 1113.)

Moreover, "[n]o evidence was presented to indicate in any way that the data was not typical or representative of an ongoing condition." (Rhymes, supra, 170 Cal. App.3d at p. 1113.) It could, therefore, "be safely assumed that such underrepresentation existed long before the specific study was performed in 1979 and that the underrepresentation can be expected to continue as long as the voter lists are the sole source for jury lists." (Ibid.) The court applied its holding only to Ms. Rhymes "and thereafter ... prospectively." (Id., at p. 1114.)

This court granted hearing in Rhymes, held it for Harris, and retransferred it to the Court of Appeal approximately one year after Harris became final. The retransfer order directed the Court of Appeal to refile its opinion.[6]

*283 The only difference between Rhymes and this case is that Rhymes was tried by a municipal court jury in the Pomona Judicial district, while appellant was tried by a superior court jury in that same district. However, the deputy jury commissioner testified in this case that superior and municipal court juries in Pomona are chosen from the same pool. Therefore, appellant's prima facie showing, based on the record before the appellate department in Rhymes, was adequate. The only remaining question under Duren is whether the underrepresentation was due to "systematic exclusion of the group in the jury-selection process." (Duren, supra, 439 U.S. at p. 364 [58 L.Ed.2d at p. 587].)

Harris and Rhymes answer that question in the affirmative. In both cases, the "particular jury-selection process utilized" (Duren, supra, 439 U.S. at p. 366 [58 L.Ed.2d at p. 588]) — random selection solely from voter registration lists — was the cause of the disparity. That, of course, was precisely the method used to select the venire here.

Accordingly, appellant made a prima facie showing "of a gross disparity resulting in a violation of [his] right to an impartial jury drawn from a fair cross-section of the community. The burden then shift[ed] to the state to rebut the prima facie case." (Harris, supra, 36 Cal.3d at p. 59.) Since, as in Harris, the state made no effort to rebut appellant's prima facie showing, but argued only that the appellate department decision in Rhymes would be reversed on appeal, relief should be granted, assuming Harris and Rhymes apply in this case.

Before turning to that question and the related one of whether a reversal here is required, it is necessary to address the Attorney General's contention that the 20-mile radius area around the Pomona courthouse, rather than the county as a whole, constituted the appropriate geographical basis for a prima facie showing under Duren, supra, 439 U.S. 357. Although, as in Harris, respondent vigorously argues that point, the question is not presented on the record before this court.[7]

Respondent insists that the prosecutor's questions to the deputy jury commissioner and the portions of the Rhymes record submitted to the trial court reveal reliance on the "20-mile rule" area. The district attorney in this case elicited testimony from the commissioner that Pomona juries are selected from persons living within a 20-mile radius of the courthouse. Similarly, the jury commissioner in Rhymes testified that of the jurors who are *284 ultimately selected for jury service, 98 percent serve on juries within 20 miles of their residence. Finally, respondent emphasizes that in the brief to the appellate department in Rhymes, the district attorney argued "in part" that the 20-mile radius was the proper focus for evaluating underrepresentation.

The record does not support respondent's position that the point was preserved below. At no time during the abbreviated hearing in the trial court did the prosecutor argue that the relevant community was the "20-mile rule" area. The jury commissioner's testimony here and in Rhymes established only that a 20-mile rule existed, a fact obvious from a cursory reading of the Code of Civil Procedure. Yet at no time did the state rely on that testimony as a basis for claiming that the 20-mile rule governed resolution of this case or Rhymes.

Indeed, throughout both the appellate department and Court of Appeal proceedings in Rhymes, the state assumed that countywide statistics, rather than 20-mile area statistics, could be used in making a prima facie showing. Indeed, the district attorney's brief in the appellate department was devoted almost exclusively to the claim that jury-eligibles, rather than total population, should be used in assessing the second Duren criterion. The brief assumed throughout, as the state did in Harris, that countywide, rather than district-wide or 20-mile area-wide statistics were the appropriate focus. Thus, except for the fact that they occurred in different courthouses, this case is on all fours with Harris and Rhymes on the issue of whether the state disputed the propriety of using countywide statistics.

II.

The next question is whether the holdings in Harris and Rhymes should be applied here. I respectfully disagree with the majority's conclusion that Harris (and inferentially, Rhymes) should be held prospective only.[8]

The opinion in Harris expressly left the retroactivity question open and took "no position as to the disposition of other cases presenting issues concerning the representative character of juries selected from voter registration lists alone." (Harris, supra, 36 Cal.3d at p. 59, fn. omitted.) The Court of Appeal in Rhymes explicitly made its ruling applicable to Rhymes and thereafter prospectively only. (Rhymes, supra, 170 Cal. App.3d at p. 1114.)

Harris itself has been applied both retroactively and prospectively. The Fifth Appellate District, in People v. Cantu (1984) 161 Cal. App.3d 259 *285 [207 Cal. Rptr. 460], held it prospective. There, the court accepted the parties' concession that Harris established a "new rule" sanctioning the use of total population figures. (Id., at p. 268.) Thus, the court proceeded to apply the test set forth in Stovall v. Denno (1967) 388 U.S. 293, 297[9] [18 L.Ed.2d 1199, 1203, 87 S.Ct. 1967], and concluded that the Harris rule was (1) "collateral to the fair determination of guilt or innocence," (2) had been relied on extensively by law enforcement and (3) would result in substantial impact on the administration of justice if applied retroactively. However, the court noted that its answer to the first two Stovall factors did not actually require resolution of the third. (161 Cal. App.3d at pp. 269-271.)[10]

The same court considered a Harris-type claim five months later in People v. Alexander (1985) 163 Cal. App.3d 1189 [210 Cal. Rptr. 306]. There, the challenge was that the accused's venire had been underrepresented by Blacks and Hispanics due to the methods employed by the jury commissioner's office in excusing venirepersons for hardship. (Id., at p. 1198.) No claim was made that source lists — by then compiled from Department of Motor Vehicles and voter registration records (see Code Civ. Proc., § 204.7) — had resulted in underrepresentation. (163 Cal. App.3d at p. 1198.)

In reaching the conclusion that Alexander had made a prima facie showing in the trial court, the Court of Appeal noted that he had relied on total population statistics rather than "jury eligible" statistics. (Id., at p. 1202.) This was precisely the "new rule" the Cantu court said Harris had enunciated. Nevertheless, Alexander did apply that rule even though, as here, the trial proceedings occurred before Harris was decided. Alexander thus represents a case in which an accused was afforded the benefit of the Harris rule even though the showing was made before the Harris decision.[11]

Recently, in People v. Guerra (1984) 37 Cal.3d 385 [208 Cal. Rptr. 162, 690 P.2d 635], this court had occasion to reiterate the principles of retroactivity. *286 "To determine whether a decision should be given retroactive effect, the California courts first undertake a threshold inquiry: does the decision establish a next rule of law? If it does, the new rule may or may not be retroactive ... but if it does not `no question of retroactivity arises,' because there is no material change in the law." (Id., at p. 399, quoting Donaldson v. Superior Court (1983) 35 Cal.3d 24, 36 [196 Cal. Rptr. 704, 672 P.2d 110].) "`As a rule, judicial decisions apply "retroactively." [Citation.] Indeed, a legal system based on precedent has a built-in presumption of retroactivity.' [Citation.]" (Guerra, supra, 37 Cal.3d at p. 399, quoting Solem v. Stumes (1984) 465 U.S. 638, 642 [79 L.Ed.2d 579, 586, 104 S.Ct. 1338].)

Decisions which "explain or refine the holding of a prior case, those which apply an existing precedent to a different fact situation, even if the result may be said to `extend' the precedent, or those which draw a conclusion that was clearly implied in or anticipated by previous opinions" do not "establish a new rule of law" within the meaning of the retroactivity doctrine. (Guerra, supra, 37 Cal.3d at p. 399, citations omitted.)

Even if a decision establishes a "new rule," it must be determined whether there was a prior rule to the contrary on which there was "justifiable reliance...." (Ibid.) "And the emphasized word is crucial: `Unjustified "reliance" is no bar to retroactivity.' [Citation.]" (Ibid., quoting Solem v. Stumes, supra, 465 U.S. at p. 646 [79 L.Ed.2d at p. 589].) There is no issue of retroactivity when the appellate court resolves a conflict between lower court decisions, or addresses an issue not previously presented to the courts. "In each of these cases there was no clear rule on which anyone could have justifiably relied." (Guerra, supra, 37 Cal.3d at p. 400.)

Arguably, Harris established two "new" rules of law: first, that voter registration lists could no longer serve as the sole source of Los Angeles County venires; and second, that a prima facie showing of disparity could be based on total population, rather than jury-eligible figures.

As to the voter registration issue, Harris did not establish a new rule. The idea that voter registration lists could be held unconstitutionally invalid as the sole source of jury venires was noted in People v. Sirhan (1972) 7 Cal.3d 710 [102 Cal. Rptr. 385, 497 P.2d 1121], overruled on other grounds, Hawkins v. Superior Court (1978) 22 Cal.3d 584, 593, footnote 7 [150 Cal. Rptr. 435, 586 P.2d 916]. Although the Sirhan court held the use of such lists valid in that case, it clearly left open the possibility that their use might be improper in a case where the accused was able to show that they "resulted `in the systematic exclusion of a "cognizable group or class of qualified citizens"' ..., or that there was `discrimination in the compiling *287 of such voter registration lists.' ... `(Id., at pp. 749-750, citations and fn. omitted.)

Thus, Sirhan stands not as a blanket approval of the use of voter registration lists, but as a refusal to invalidate their use in the absence of sufficient evidence indicating that they have produced a constitutionally suspect venire. The accused in Harris presented that evidence, and this court found it sufficient to satisfy the standard in Sirhan. At most, Harris could be characterized as having drawn "a conclusion that was clearly implied in or anticipated by previous opinions" (Guerra, supra, 37 Cal.3d at p. 399) — not as having established any "new" rule.

The second aspect of Harris presents a more difficult problem. Several Court of Appeal cases before Harris had held the use of total population figures insufficient to establish a prima facie case of underrepresentation. (People v. Mooring (1982) 129 Cal. App.3d 453, 458-460 [181 Cal. Rptr. 71]; People v. Remiro (1979) 89 Cal. App.3d 809, 839-840 [153 Cal. Rptr. 89, 2 A.L.R.4th 1135]; People v. Spears (1975) 48 Cal. App.3d 397, 404 [122 Cal. Rptr. 93]; People v. Powell (1974) 40 Cal. App.3d 107, 129-130 [115 Cal. Rptr. 109]; see also People v. Lewis (1977) 74 Cal. App.3d 633, 646 [141 Cal. Rptr. 614].) Several others sanctioned the use of total population figures. (People v. Buford (1982) 132 Cal. App.3d 288, 291 [182 Cal. Rptr. 904]; People v. Newton (1970) 8 Cal. App.3d 359, 389-390 [87 Cal. Rptr. 394]; People v. McDowell (1972) 27 Cal. App.3d 864, 870 [104 Cal. Rptr. 181].)

It is questionable how much weight should be given to any of the foregoing cases which were decided before Taylor v. Louisiana, supra, 419 U.S. 522, since before that decision, a challenge to a jury venire based on underrepresentation required a showing of intentional discrimination on the part of the state. Thus, attention will focus on post-Taylor cases.

Of those rejecting the use of total population figures, only Mooring, supra, 129 Cal. App.3d 453 squarely stands for the sole proposition that an underrepresentation claim must be based on jury-eligible rather than total population figures.[12] As the court stated, "Eligibility to serve on petit juries and raw census figures, without breakdown, may not be used as a basis for a showing of systematic discrimination. Figures can be deceptive and hence *288 meaningless, and for that reason the comparison, to be meaningful, must be between those persons eligible as a cognizable class to sit as petit jurors and those in that class who are actually called." (129 Cal. App.3d at p. 459.)

On the other side stands Buford, supra, 132 Cal. App.3d 288, the only published post-Taylor decision which sanctioned the use of total population figures.[13]Buford involved the adequacy of the jury commissioner's efforts in ensuring venires representative of Blacks in Contra Costa County. The Court of Appeal found that the appellant had made out a prima facie showing of underrepresentation under Duren, supra, 439 U.S. 357, based on a "marked" disparity between the percentage of Blacks who appeared on panels and the percentage of Blacks "in the general population of the county." (Buford, supra, 132 Cal. App.3d at p. 296.) This disparity, Buford held, resulted from systematic exclusion. (Id., at p. 297.)

The Buford court did recognize that the use of total rather than eligible county figures made "no allowance for the impact of concededly permissible standards of eligibility." (Id., at p. 298.) However, it excused this aspect of the appellant's proof since "[i]t would be unrealistic and contrary to applicable principles ... to impose upon a defendant the burden of excluding all possible and permissible explanations for underrepresentation." (Ibid.) The court also noted that the appellant's case was not predicated on statistics alone, but also on "the informal procedure by which Contra Costa County goes about excusing prospective jurors from service." (Ibid.)

This nonunanimous body of state authority must also be understood in the context of parallel federal authority. The statistical showing in Taylor was based on women in the county who were eligible for jury service (419 U.S. at p. 524 [42 L.Ed.2d at pp. 694-695]); however, Taylor did not discuss whether the use of total population figures was inappropriate.

Castaneda v. Partida (1977) 430 U.S. 482 [51 L.Ed.2d 498, 97 S.Ct. 1272] involved a challenge to Hispanic representation on a county grand jury in Texas. The court held that a prima facie showing, which was based on total county statistics, had been made (id., at pp. 495-496 [51 L.Ed.2d at pp. 511-512]), and that the state had failed to rebut the showing with any evidence as to the effect of ineligibility criteria on the pool considered. The court also noted that only a small part of the Hispanics in the county were *289 noncitizens and therefore ineligible. (Id., at pp. 498-499 [51 L.Ed.2d at p. 513].)

Similarly, in Duren v. Missouri, supra, 439 U.S. 357, the court noted that the showing in Taylor had been made on jury-eligible statistics. (439 U.S. at p. 364 & fn. 21 [58 L.Ed.2d at p. 587 & fn. 21].) However, in concluding that Duren had made a prima facie showing, the court relied on census (i.e., total population) data. (Id., at p. 365 [58 L.Ed.2d at pp. 587-588].)

The court rejected the state's claim that more precise data should have been used. "Although the Missouri Supreme Court speculated that changing population patterns between 1970 and 1976 and unequal voter registration by men and women rendered the census figures a questionable frame of reference, there is no evidence whatsoever in the record to suggest that the 1970 census data significantly distorted the percentage of women in Jackson County at the time of trial." (Duren, supra, 439 U.S. at p. 365, fn. omitted [58 L.Ed.2d at p. 587].) Nor was the failure to use voter registration data (which would have excluded those then ineligible to vote (under 21 years)) fatal: "the fair-cross-section requirement involves a comparison of the makeup of jury venires or other sources from which jurors are drawn with the makeup of the community, not of voter registration lists." (439 U.S. at p. 365, fn. 23 [58 L.Ed.2d at p. 587, fn. 23], italics in original.)

What this review indicates is that while the Supreme Court had not decided — at least prior to this court's Harris opinion — whether a prima facie showing of underrepresentation had to be based on jury-eligible statistics, it had clearly permitted an accused to make such a showing based on total population figures. Indeed, even the Harris plurality read Duren as indicating that such a showing was permissible. (Harris, supra, 36 Cal.3d at pp. 55-56.)

Therefore, on the jury-eligible versus total population issue, it must be concluded either that Harris resolved a conflict between Mooring, supra, 129 Cal. App.3d 453, and Buford, supra, 132 Cal. App.3d 288 (without explicitly mentioning the conflict), or that there was no "prior rule" on which reliance was "justifiable" because of the federal authority sanctioning the use of total population figures. (Guerra, supra, 37 Cal.3d at pp. 399, 400.)

On both the voter registration and total population aspects of Harris, then, "the ordinary assumption of retrospective operation" (Donaldson v. Superior Court, supra, 35 Cal.3d at p. 37) must come into play. Harris, therefore, should be held applicable to all cases pending on direct review *290 when it was decided in which a prima facie showing of underrepresentation due to systematic exclusion of a cognizable group was made. People v. Cantu, supra, 161 Cal. App.3d 259, People v. Pendleton, supra, 167 Cal. App.3d 413, and People v. Brown, supra, 169 Cal. App.3d 728, if they were pending on direct review when Harris was decided, and to the extent they are inconsistent with this rule, should accordingly be disapproved.[14]

Even assuming, however, that Harris should be held prospective only, the question of whether this court should hold Harris inapplicable to the present case remains. Both this court and the United States Supreme Court have, on occasion, permitted individuals facing the death penalty to take advantage of rulings announced subsequent to trial when the ruling would not otherwise be given retroactive effect. For example, in People v. Wheeler, supra, 22 Cal.3d 258, the court held its rule applicable to Wheeler, an appellant in a companion case (see People v. Johnson (1978) 22 Cal.3d 296 [148 Cal. Rptr. 915]) "and to any defendant now or hereafter under sentence of death." (Wheeler, supra, 22 Cal.3d at p. 283, fn. 31; see People v. Allen (1979) 23 Cal.3d 286, 294, fn. 3 [152 Cal. Rptr. 454, 590 P.2d 30] [so applying Wheeler].)

Similarly, in In re Jackson (1964) 61 Cal.2d 500 [39 Cal. Rptr. 220, 393 P.2d 420], this court vacated a death sentence by applying People v. Morse (1964) 60 Cal.2d 631 [36 Cal. Rptr. 201, 388 P.2d 33, 12 A.L.R.3d 810] retroactively. The court reasoned that "[w]e cannot accept the paradox that would occur if defendant Morse, whose appeal overrules the precedent under which he was erroneously sentenced, obtains a new penalty trial, but petitioner Jackson, sentenced under the same precedent, must suffer death in a grim sequence of judicial error." (61 Cal.2d at p. 507, fn. omitted.)

Finally, the United States Supreme Court made its holding in Witherspoon v. Illinois (1968) 391 U.S. 510 [20 L.Ed.2d 776, 88 S.Ct. 1770] retroactive. (Id., at p. 523, fn. 22 [20 L.Ed.2d at p. 785, fn. 22].) In In re Eli *291 (1969) 71 Cal.2d 214, 215-217, 219 [77 Cal. Rptr. 665, 454 P.2d 337], this court applied that ruling to an individual whose death sentence the court had previously affirmed.

As appellant notes, "[i]t is the universal experience in the administration of criminal justice that those charged with capital offenses are granted special considerations." (Griffin v. Illinois (1956) 351 U.S. 12, 28 [100 L.Ed. 891, 904, 76 S.Ct. 585, 55 A.L.R.2d 1055] (dis. opn. of Burton & Minton, JJ.).) A "capital case exception" to a rule of prospectivity of Harris claims would thus appear fair and consistent with established principles affording capitally condemned individuals the retroactive benefit of favorable rulings.[15]

One final point should be made. Harris's judgment of death was filed on September 3, 1980. The judgment of death here was filed on April 21, 1981, eight months later. Because of the necessity to conduct a second penalty phase after the first penalty phase jury deadlocked, it is likely that Harris and appellant challenged their respective venires around the same time.

The only real difference in the cases is the speed with which the appellate record in Harris was certified and the decision of this court was filed. The record in Harris was certified within seven months of the judgment of death and the opinion was filed in April of 1984. The record here was not finally certified until April 26, 1985 — nearly four years after the judgment of death — in part because the trial judge, according to appellant, "continued the record correction proceedings which were needed to add the jury challenge evidence to the record on appeal." That appellant should be denied relief on the sole ground that Harris, rather than this case, fortuitously became the vehicle to resolve the underrepresentation issue, is an exceedingly arbitrary result.

In sum, this court should hold that Harris and Rhymes apply to appellant's case.

III.

The three members of the Harris plurality agreed that the trial court's failure to rule that Harris had made a prima facie showing of underrepresentation *292 was prejudicial per se. (See Harris, supra, 36 Cal.3d at p. 59; People v. Wheeler, supra, 22 Cal.3d at p. 283.) Since this case is on all fours with Harris, appellant's conviction should be reversed.

The question of remedy was raised in Harris and the Attorney General has argued that point here. Therefore, the issue of remand should be addressed.

In his concurring opinion in Harris, Justice Grodin advocated a remand under the authority of Penal Code section 1260.[16] That procedure would have permitted the state to rebut Harris's prima facie showing before the trial court. "Upon full consideration of relevant evidence [the trial court] might [conclude] that `no disparity of constitutional significance exists,' or that `even with the use of multiple sources and all other practical means, a certain level of disparity is unavoidable,' or that the underrepresentation which does exist is justified by a showing of overriding state interest." (36 Cal.3d at p. 72, quoting plur. opn., at p. 59.)

One Court of Appeal has already opted for the remand solution. People v. Alexander, supra, 163 Cal. App.3d 1189 was a pre-Harris Kern County case in which the Court of Appeal held the accused made a prima facie showing as to underrepresentation. That court remanded and ordered the trial court to take further evidence on both present and past jury selection practices. The Court of Appeal directed the trial court to decide "if the County of Kern was and is doing all that it reasonably may be expected to do to assure a trial jury representing a fair cross-section of the county...." (163 Cal. App.3d at p. 1206.) Such a disposition was deemed appropriate in view of the fact that in the trial court "practically no rebuttal evidence was presented by the People, and it [was] not entirely clear that more [could] be presented...." (Id., at p. 1203.) In addition, `[t]he People apparently relied upon [the] pre-Harris rule...." (Id., at p. 1204.)[17]

In Rhymes, this court's retransfer order directing the Court of Appeal to refile the opinion resulted in reversal of Rhymes's conviction. (See 170 Cal. App.3d at pp. 1103, 1114.) That result was mandated even though a remand for the purpose of permitting the state to rebut the prima facie showing of underrepresentation was certainly a well-known option. As a *293 matter of fairness and consistency, this court should not accord appellant on automatic appeal any less relief than was accorded the misdemeanant in Rhymes in collateral proceedings.

Further, as appellant urges, there may be several practical problems with a remand here. The first is the passage of time. A remand would require the prosecution to conduct an investigation in 1987 as to the adequacy of procedures Los Angeles County officials were using in 1979 and 1980 to obtain representative juries in Pomona. Any such inquiry would have to span at least as long a period as appellant's statistics did in order to be truly representative. To the extent that any of the information necessary to rebut appellant's showing would depend on documents once in the jury commissioner's possession, such documents may no longer be available. (See Gov. Code, § 26202 [requiring retention of data for two-year period only].)

A second problem is the enormity of such an undertaking. To the extent post-July 1981 statistics are relied upon on remand, the trial court would have to take into account the degree to which the expanded source pool (see Code Civ. Proc., § 204.7) has altered the disparities which appellant's evidence indicated for the relevant period. The rebuttal showing would also require measuring the effect of the 20-mile rule on appellant's prima facie showing (see Harris, supra, 36 Cal.3d at pp. 48 (plur. opn.), 71, fn. 1 (conc. opn.)), and the extent to which jury-eligible figures — admittedly difficult to come by (see id., at pp. 53-54 (plur. opn.)) — reduce the disparities which Dr. Butler found to exist.

Also, it is far from clear how much "judicial economy" (Harris, supra, 36 Cal.3d at p. 72 (conc. opn.)) would actually be fostered by a remand. A new penalty trial here is necessary in any event due to the trial court's instruction on the Governor's commutation power (maj. opn., ante, at pp. 270-273; People v. Montiel (1985) 39 Cal.3d 910, 928 [218 Cal. Rptr. 572, 705 P.2d 1248]; see People v. Ramos (1984) 37 Cal.3d 136, 155 [207 Cal. Rptr. 800, 689 P.2d 430]) and the instructions and argument regarding the scope of the jury's sentencing discretion (maj. opn., ante, at pp. 273-276; People v. Brown (1985) 40 Cal.3d 512 [220 Cal. Rptr. 637, 709 P.2d 440]). That trial will, of necessity, involve the presentation of some, if not all, of the guilt phase evidence, as well as the death qualification of the jury. These procedures would also be required if a new guilt phase trial were conducted. Therefore, it is questionable whether judicial resources would really be saved as a result of a remand.

Finally, appellant argues that a remand would not be "proper" or "just under the circumstances." (Pen. Code, § 1260.) He asserts that the state has consistently maintained that using voter registration lists as the sole *294 source of jury venires and population figures based on juror eligibility is proper. The prosecutor's only response to the merits of the motion here was that the appellate department decision in Rhymes would be reversed on appeal. Such categorical insistence on a position that was under question, he argues, makes a remand improper.

To support this claim, appellant points to Whiteley v. Warden (1971) 401 U.S. 560 [28 L.Ed.2d 306, 91 S.Ct. 1031], where the United States Supreme Court rejected a similar request to remand. In Whiteley, the accused prevailed on his claim — asserted at every stage of the proceedings — that incriminating evidence should have been suppressed as the result of an insufficient arrest warrant and the lack of probable cause. The state asserted before the high court — apparently for the first time in the case — that the magistrate who issued the warrant had additional information which would have supported the arrest.

Rejecting the request for remand, the high court stated: "Knowing the basis for petitioner's constitutional claim, the State chose to try those proceedings on the record it had developed in the state courts.... Its sole explanation for this state of affairs is that `the state has felt, based on precedent and logic, that no court would accept the legal reasoning of petitioner.' ... In the circumstances of this case, that justification, as we have shown, is untenable." (Id., at p. 569 [28 L.Ed.2d at p. 314]; accord Lorenzana v. Superior Court (1973) 9 Cal.3d 626, 640 [108 Cal. Rptr. 585, 511 P.2d 33] [state not entitled to reopening of the propriety of a search on the basis of "new legal theories"]); Panopulos v. Maderis (1956) 47 Cal.2d 337, 340-341 [303 P.2d 738] [party not permitted to advance new theories on appeal which contemplate factual situations "not put in issue or presented at the trial"].)

Given the foregoing considerations, a remand would be inappropriate here.

IV.

Appellant made a timely prima facie showing under Duren, supra, 439 U.S. 357 and Harris, supra, 36 Cal.3d 36, that his venire was underrepresented by Blacks because of the systematic exclusion of them by the state's methods of venire selection. Since the prosecution failed to rebut that showing, and since the majority's refusal to give appellant the benefit of the Harris rule is exceedingly arbitrary, I cannot join in the affirmance of the convictions and special circumstance finding.

Reynoso, J., concurred.

*295 BROUSSARD, J., Dissenting.

I agree with the dissent of the Chief Justice in all respects except one. She suggests that the majority of this court in People v. Harris (1984) 36 Cal.3d 26 [201 Cal. Rptr. 782, 679 P.2d 433], certiorari denied (1984) 469 U.S. 965 [83 L.Ed.2d 301, 105 S.Ct. 365], reached a consensus on the question whether a prima facie showing of underrepresentation could be rebutted with statistics indicating the lack of underrepresentation when the effects of the 20-mile rule are taken into account. The Chief Justice's dissent states: "From this review of the opinions in Harris, it is clear that four members of this court were of the view that a prima facie showing of underrepresentation could be made with countywide figures, but that such a showing might be rebutted with statistics indicating the lack of underrepresentation when the effects of the 20-mile rule (or some less-than-countywide rule) are taken into account." (Dis. opn. at p. 281, fn. omitted.)

I do not agree that a consensus was reached on the effect of the 20-mile rule. In Harris, we merely stated that the parties had assumed that it was the countywide figures which were significant, and that the state had not attempted to rebut the defendant's showing on the basis of statistics from the 20-mile area. Thus the issue was not presented in Harris. (See Harris, supra, 36 Cal.3d at p. 48.) Nor is it presented in this case since, as the Chief Justice's dissent amply demonstrates, the issue was not preserved below.

Appellant's petition for a rehearing was denied April 2, 1987.

NOTES

[1] All statutory references are to the Penal Code unless otherwise indicated.

[2] Lopez was unable to identify defendant as having been involved in the incident. Both defendant and his brother, however, admitted complicity in the incident in statements given to police following their arrests. Defendant admitted to the police that he had carried the gun. At trial, defendant did not deny his participation in the robbery.

[3] In all of the cases which have held challenges untimely under section 1060, the challenge was made after the voir dire examination had been completed and the jurors had been sworn to try the case. (See, e.g., People v. Sirhan (1972) 7 Cal.3d 710, 751 [102 Cal. Rptr. 385, 497 P.2d 1121]; People v. Oliveria (1899) 127 Cal. 376, 380 [59 P. 772]; People v. Ah Lee Doon (1893) 97 Cal. 171, 176 [31 P. 933]; People v. Flowers (1974) 38 Cal. App.3d 813, 818 [113 Cal. Rptr. 701].) This is apparently the first case in which the challenge was made during voir dire examination, and before any juror had been sworn to try the case.

[4] Federal law draws a general distinction between the time for challenging the jury panel and the time for challenging individual jurors, requiring that statutory challenges to the jury panel be made no later than "before the voir dire examination begins." (See 28 U.S.C. § 1867(a), (b); see United States v. Price (5th Cir.1978) 573 F.2d 356, 361.) It is not clear, however, that constitutional challenges to a jury panel are subject to the same time limitations. (See, e.g., United States v. De Alba-Conrado (5th Cir.1973) 481 F.2d 1266, 1269-1270 & fn. 5.)

[5] Although the dissent suggests that "Harris itself has been applied both retroactively and prospectively" (post, p. 284) the case which it points to as embodying a retroactive application of Harris — People v. Alexander (1985) 163 Cal. App.3d 1189 [210 Cal. Rptr. 306] — did not involve a challenge to the use of voter registration lists as a source for jury venires and contains no discussion of the retroactivity issue at all. From all appearances, the question of the retroactive or prospective application of Harris was never raised in Alexander. It is well established, of course, that a decision is not authority for propositions that were not considered in the court's opinion. (See, e.g., McDowell & Craig v. City of Santa Fe Springs (1960) 54 Cal.2d 33, 38 [4 Cal. Rptr. 176, 351 P.2d 344]; People v. Banks (1959) 53 Cal.2d 370, 389 [1 Cal. Rptr. 669, 348 P.2d 102].)

[6] Neither Harris nor Taylor involved a claim that the state had engaged in intentional invidious discrimination in the selection of juries in violation of equal protection principles. In cases in which a defendant does establish that the state has engaged in such invidious discrimination in the selection of grand juries, the United States Supreme Court has recently reaffirmed that any resulting conviction cannot stand. (See Vasquez v. Hillery (1986) 474 U.S. 254, 261-266 [88 L.Ed.2d 598, 607-610, 106 S.Ct. 617, 622-624].) Similarly, in other contexts, courts have distinguished the cross-sectional violation of Taylor from improper jury selection practices that threaten the impartiality of juries. (See, e.g., Woodard v. Sargent (8th Cir.1985) 753 F.2d 694, 697.)

[7] See, for example, People v. Sirhan (1972) 7 Cal.3d 710, 749-750 [102 Cal. Rptr. 385, 497 P.2d 1121]; People v. Remiro (1979) 89 Cal. App.3d 809, 839 [153 Cal. Rptr. 89, 2 A.L.R.4th 1135]; People v. Lewis (1977) 74 Cal. App.3d 633, 646 [141 Cal. Rptr. 614]; People v. Cabral (1975) 51 Cal. App.3d 707, 715 [124 Cal. Rptr. 418]; People v. Breckenridge (1975) 52 Cal. App.3d 913, 921 [125 Cal. Rptr. 425]; People v. Powell (1974) 40 Cal. App.3d 107, 126-127 [115 Cal. Rptr. 109]; People v. Newton (1970) 8 Cal. App.3d 359, 389-390 [87 Cal. Rptr. 394]; Foster v. Sparks (5th Cir.1975) 506 F.2d 805, 816-817, footnote 28 and cases cited.

When Congress established voter registration lists as the basic source of jury panels in the federal Jury Selection and Service Act of 1968, the enactment was widely heralded as a significant reform. As one federal court observed at the time: "`[T]he best thought of all three branches of government points toward voter registration lists as representing "the best cross-section of the community; indeed, they are probably the most broadly based lists available."' ... No significant improvement in our nation's federal judicial machinery has ever met such unanimity of opinion as to its importance and propriety as has adoption by the Congress of the system of jury selection at random from voter registration lists." (Simmons v. United States (5th Cir.1969) 406 F.2d 456, 463.)

[8] See, for example, People v. Powell, supra, 40 Cal. App.3d 107, 128-129; People v. Spears (1975) 48 Cal. App.3d 397, 404 [122 Cal. Rptr. 93]; People v. Lewis, supra, 74 Cal. App.3d 633, 646; People v. Remiro, supra, 89 Cal. App.3d 809, 839-840. Although two earlier decisions — People v. Newton, supra, 8 Cal. App.3d 359, 389-390, and People v. McDowell (1972) 27 Cal. App.3d 864 [104 Cal. Rptr. 181] — did suggest in dicta that a defendant could establish a prima facie case on the basis of general population figures, neither case recognized or discussed the potential problems — identified in the subsequent California cases — attributable to the differences between general-population, and jury-eligible, statistics.

As the dissent notes, there is one more recent, pre-Harris decision — People v. Buford (1982) 132 Cal. App.3d 288, 291 [182 Cal. Rptr. 904] — in which the Court of Appeal did entertain a jury challenge when the defendant relied on general-population, rather than jury-eligible, statistics. Buford, however, did not involve a challenge to a county's use of voter registration lists, but instead concerned the validity and impact of a county's informal procedure for excusing prospective jurors. In our view, Buford — which, in any event, was decided after the trial in this case — cannot reasonably be viewed as having put local authorities on notice that the prior decisions relating to voter registration lists were no longer authoritative.

[9] It is in this respect that the present case is distinguishable from Lee v. Missouri (1979) 439 U.S. 461 [58 L.Ed.2d 736, 99 S.Ct. 710], where the Supreme Court held that its decision in Duren v. Missouri (1979) 439 U.S. 357 [58 L.Ed.2d 579, 99 S.Ct. 664] should apply retroactively to cases in which juries were sworn after the 1975 decision in Taylor, supra, 419 U.S. 522. In Lee the court evidently concluded that in light of its holding in Taylor, state officials could not reasonably have believed that the practice at issue in Duren — which automatically permitted women, but not men, to be excused from jury service on request — was constitutionally permissible. (Lee, supra, 439 U.S. at p. 462 [58 L.Ed.2d at p. 739].)

As noted above, in our view the pre-Harris California precedents did not provide similarly clear guidance to governmental authorities that the random use of voter registration lists might be invalid if the makeup of actual jury venires departed from general-population figures. Unlike the jury selection practice at issue in Duren — which, on its face, accorded differential treatment to men and women and was clearly suspect under Taylor for that reason — the use of voter registration lists does not apply facially different juror selection standards on the basis of gender, race, ethnicity, or any other "suspect" classification.

[10] Harris indicates that at the time of the trial in that case, 17 of the 46 counties in California used voter registration lists as the sole source for the selection of jury venires. (Harris, supra, 36 Cal.3d at p. 46.)

[11] Even before Harris, of course, a defendant could have prevailed on a challenge to the county's use of voter registration lists if he could establish, through the use of the more accurate jury-eligible statistics, that the procedure had resulted in actual jury venires that were not fairly representative of the jury-eligible population in the relevant community. In declining to apply Harris retroactively, we simply hold that the more lenient prima facie standard adopted in Harris should not be applied to pre-Harris cases.

[12] Defendant additionally contends that even if we conclude, as we have, that Harris should generally apply only prospectively, he should nonetheless be accorded the benefit of that decision on three separate grounds. First, he claims that but for a delay in the preparation of the record on appeal in his case, the "Harris" issue might have been decided in his case, rather than in Harris, and that it is unfair to withhold the benefit of the ruling on such fortuitous grounds. This "fortuity," of course, is present any time a decision is applied prospectively; in Daniel, supra, 420 U.S. 31, the Supreme Court affirmed Daniel's conviction even though he had raised the same issue as Taylor and his case was pending before the court at the time Taylor, supra, 419 U.S. 522, was decided. (See Young v. Zant (11th Cir.1984) 727 F.2d 1489, 1491.) We note that the defendants in Cantu, Brown and Pendleton could equally complain about unfair treatment if defendant were to be accorded the benefit of Harris despite the fact that we have concluded that the applicable legal principles call for the prospective application of that decision.

Second, defendant alternatively suggests that we ought to adopt a "death penalty" exception to our prospectivity holding, by analogy to an exception we fashioned in People v. Wheeler (1978) 22 Cal.3d 258, 283, footnote 31 [148 Cal. Rptr. 890, 583 P.2d 748], when we adopted the rule outlawing the use of peremptory challenges on the basis of group bias. Unlike Wheeler — which established a means for rooting out the exercise of intentional invidious discrimination in the selection of a jury — Harris simply modified a defendant's burden in establishing a prima facie case of an unintentional violation of the fair-cross-section requirement. Because even before Harris a defendant was able to pursue a constitutional cross-sectional challenge through the use of more precise jury-eligible statistics, we perceive no comparable unfairness in applying the same rule of prospectivity to death penalty cases as applies to all other criminal cases.

Finally, defendant contends that he is entitled to relief under "collateral estoppel" principles. The claim is manifestly untenable. To begin with, the "issue actually litigated" in Harris — whether the statistical showing made in that case was adequate to demonstrate a prima facie case of underrepresentation in Long Beach jury venires — was plainly not identical to the issue here. More fundamentally, in light of our conclusion that the legal principle on which Harris rests should not be applied retroactively, it clearly follows that that decision's holding cannot be applied to earlier cases on a collateral estoppel theory. The retroactivity ruling means, of course, that earlier cases are governed by different legal principles, and it is settled that under such circumstances collateral estoppel does not apply. (See generally Rest.2d Judgments, § 28(2) & coms. b, c.)

[13] The supplementary commutation instruction read in full:

"Subject to the application procedures provided by statute, the Governor, on conditions the Governor deems proper, may grant a commutation of a sentence.

"The Governor shall report to the Legislature each commutation granted, stating the pertinent facts and the reasons for granting it.

"The Governor may not grant a commutation for a person twice convicted of a felony except on the recommendation of the Supreme Court, four justices concurring.

"In order to obtain a commutation of sentence, the following procedure must be complied with: Persons who have been sentenced to life in prison without possibility of parole and who have suffered more than one felony conviction must make application for commutation of sentence directly to the Governor of the state.

"Upon the request of the Governor, the Community Release Board shall investigate, report on any application for reprieves, commutation of sentence.

"The Board shall consider the application, the transcripts of the judicial proceedings and all documents submitted with the application.

"... [I]vestigators for the Board may take testimony, examine witnesses under oath and take all action necessary to conduct a full and complete investigation of the application.

"The Board may require the Court in which the conviction was had or the district attorney who prosecuted the action to furnish it immediately with a summary of facts proved at the trial, any other facts relevant to the issue or denying of the commutation of sentence and any recommendations, including the reasons concerning granting or denying of the commutation."

[14] The prosecutor reiterated this theme at a later point in his argument: "[Defense counsel] is seeking to inject an element into this matter by saying, remarking that before you kill Mr. Myers, think of this, think of that. [¶] Ladies and Gentlemen, the law has been written in a manner and you were voir dired very carefully about this. Your function is a fact-finding function solely, and that has been explained to you very carefully. [¶] Once you determine, once you make a determination as to whether or not the aggravating circumstances or factors outweigh those in mitigation or the mitigating circumstances or factors outweigh those in aggravation, then the law says what verdict you shall return. [¶] You don't make up your mind as to your preference for penalties. All you are doing is weighing, a weighing process; and the law is very explicit. [¶] It is your duty as a juror, if you find that mitigating factors outweigh those in aggravation, it shall be your duty then to return a verdict of life without possibility of parole. [¶] If you find that aggravating factors outweigh those in mitigation, it shall be your duty to affix the verdict of the death penalty. You don't determine that. The law has determined that for you. [¶] And his statement, `before you kill Mr. Myers,' is completely inappropriate. The law determines as to what the penalty is and it has always been that way. [¶] I asked each one of you in one form or another if you would have the courage to return such a verdict if you found this to be an appropriate case. [¶] In other words, the aggravating factors outweigh those in mitigation, would you return such a verdict. Each of you said yes. [¶] So I do not feel it appropriate to discuss that any further."

[1] A preliminary issue, that of the timeliness of appellant's challenge to the venire, has been resolved correctly by the majority. (See maj. opn., ante, at pp. 262-263.)

[2] Although appellant has asserted that both Blacks and Hispanics were unconstitutionally underrepresented in his venire, he relied primarily on a decision of the Appellate Department of the Superior Court of Los Angeles County (In re Rhymes, APHC 000042) which found that the venires in the Pomona Judicial District of the municipal court were unconstitutionally underrepresented by Blacks. That decision, rendered by then Superior Court Judge Robert Fainer, followed a hearing before a referee, granted Rhymes habeas corpus relief, and vacated her municipal court conviction. The court concluded that the use of voter registration lists as the single source of jury venires in the Pomona Judicial District "cannot meet the constitutional test... as to blacks[, and] will continue to result in the systematic exclusion of a fair and representative number of eligible blacks as potential jurors."

[3] "Comparative disparity" is a standard which "measures how the use of voter lists alters the probability that a member of a particular cognizable group will be summoned to serve on a jury." (Harris, supra, 36 Cal.3d at p. 47, fn. 2.) A comparative disparity of 43 percent for Blacks, for example, would mean that "a Black would have a 43 percent lesser chance of being on the jury panel for [a particular district] for the period [studied] than the percentage of Blacks in the population would suggest." (Rhymes, supra, 170 Cal. App.3d at p. 1105.)

Another standard is "absolute disparity." This standard "measures the difference between the proportion of the studied group in the overall population of presumptively eligible jurors and the proportion of the group appearing in the pool of jurors used by the state." (Harris, supra, 36 Cal.3d at p. 47, fn. 2.) An absolute disparity of 4.6 percent for Blacks, for example, would mean that "there would be 4.6 percent fewer Blacks on the jury panels than there are in the county population based on [a selected] census." (Rhymes, supra, 170 Cal. App.3d at p. 1104.)

A third test — the "statistical significance test" — describes "the probability disparity would appear by chance in a random draw from the population." (Harris, supra, 36 Cal.3d at p. 47, fn. 2.) This test was not utilized in the studies relied on in Harris.

[4] Code of Civil Procedure section 203 provides in part that "in the County of Los Angeles no juror shall be required to serve at a distance greater than 20 miles from his or her residence."

[5] Indeed, Justice Grodin's recent majority opinion in People v. Balderas (1985) 41 Cal.3d 144, 181, footnote 15 [222 Cal. Rptr. 184, 711 P.2d 480], read Harris as standing for this proposition.

[6] The Court of Appeal originally filed Rhymes on March 25, 1982. Hearing was granted on May 27, 1982. The Court of Appeal opinion was refiled at the direction of this court on August 8, 1985.

[7] As the majority note, the question of the proper application of Harris in other cases where the 20-mile rule applies is pending before this court in Williams v. Superior Court (L.A. 32206, review granted June 20, 1986).

[8] Since traditional principles of retroactivity compel the application of Harris here, appellant's alternative claim that the doctrine of collateral estoppel affords a basis for relief need not be addressed.

[9] The Stovall v. Denno test examines (1) the purpose to be served by the new standards, (2) the extent of the reliance by law enforcement authorities on the old standards, and (3) the effect of the administration of justice of a retroactive application for the new standards.

[10] Cantu was followed without discussion in People v. Pendleton (1985) 167 Cal. App.3d 413 [212 Cal. Rptr. 524] and People v. Brown (1985) 169 Cal. App.3d 728 [215 Cal. Rptr. 465].

[11] My colleagues dismiss Alexander on the grounds that it did not involve a challenge to the use of voter registration lists as a source for jury venires and did not discuss the retroactivity question "at all." (Maj. opn., ante, at p. 265, fn. 5.)

This distinction misses the point. What the majority fail to glean from Alexander is its application — retroactively — of the Harris rule that total population figures, as opposed to jury-eligible population figures, could be used in resolving an underrepresentation claim. The fact that the claim in Alexander involved the jury commissioner's practice of granting excuses for hardship rather than the compilation of jury lists from voter registration rolls is quite beside the point.

[12] Remiro rejected the challenge on the ground that the appellant had failed to prove that Blacks and Hispanics were underrepresented on voter lists or that those who failed to register to vote were a cognizable group. (89 Cal. App.3d at p. 840.) Lewis did reject the claim on the basis that the pool was overinclusive, but also rested its holding on the fact that the accused produced no evidence as to the number of Blacks who served on panels for any length of time. (74 Cal. App.3d at p. 646.)

[13] Of course, Rhymes, prior to being vacated by this court's grant of hearing, explicitly sanctioned the use of total population figures. However, reliance on Rhymes by courts and local officials while it was in its "vacated state" would not have been "justifiable" within the meaning of Guerra. (See Cal. Rules of Court, former rules 976(d) [prohibiting publication of Court of Appeal opinions superseded by grant of hearing], 977 [prohibiting reliance on opinions not published].)

[14] The majority's reliance on Daniel v. Louisiana (1975) 420 U.S. 31 [42 L.Ed.2d 790, 95 S.Ct. 704] for its "prospective-only" holding (maj. opn., ante, at pp. 287-290) is somewhat puzzling. While it is true that Daniel held Taylor v. Louisiana, supra, 419 U.S. 522, prospective only, it did so on the basis of the tripartite test of Stovall v. Denno, supra, 388 U.S. at page 297 [18 L.Ed.2d at page 1203] (see ante, fn. 9). However, as this court took pains to point out in People v. Guerra, supra, 37 Cal.3d at page 401, the Stovall inquiry takes place only after it has been established that "there can have been justifiable reliance on an old rule to the contrary.... "As the foregoing discussion indicates, there was no "old rule" upholding the use of voter registration lists for all purposes, nor was there an old rule sanctioning the use of jury-eligible pools which the state could justifiably have relied on. In addition, the resolution of a conflict between Court of Appeal decisions — if indeed Harris resolves them — does not in itself provide grounds for a prospective-only holding (Guerra, supra, 37 Cal.3d at pp. 399-400). Accordingly, the majority's reliance on Daniel's invocation of the Stovall v. Denno test is misplaced.

[15] Admittedly, the paradox is more hypothetical in this case than it was in Jackson, supra, 61 Cal.2d 500. Here, there was error at the penalty phase requiring reversal of the death sentence. Therefore, failure to give appellant the benefit of a "capital case exception" to Harris's prospectivity would not mean the difference between affirming and reversing his death sentence. On the other hand, that same fact was deemed irrelevant by the Wheeler court when it enunciated its "capital case exception" to prospectivity. Indeed, in Allen this court did not determine whether a new penalty phase would have been required in the absence of Wheeler error, but reversed the entire judgment on that ground alone.

[16] Section 1260 provides in part: "The court may reverse, affirm, or modify a judgment or order appealed from ... and may, if proper, remand the cause to the trial court for such further proceedings as may be just under the circumstances."

[17] As Alexander notes (163 Cal. App.3d at pp. 1199-1200), People v. Buford, supra, 132 Cal. App.3d at page 299, also held that the appropriate remedy would normally be a remand; however, in view of the passage of time and the fact that Buford had served his sentence, the Court of Appeal deemed reversal the appropriate remedy there.