[Opinion certified for partial publication. ]

Opinion

Statement of Facts

Early in the morning of December 2, 1980, appellant and a companion named Paul Soto broke into a camper located behind a house in Bakersfield. Once inside the camper, appellant and his companion threatened the two occupants of the camper with guns and demanded their money. A fight broke out between appellant and one of the occupants of the camper, Jose Martinez. Appellant hit Jose with the gun and knocked him out. At this point, Pedro Martinez, Jose’s brother who had been in the house adjacent to the camper, heard the commotion and came out to the camper to investigate. However, as Pedro entered the camper he was shot in the abdomen by appellant. Appellant and his companion then fled the scene of the crime. As a result of the gunshot wound, Pedro Martinez died.

Jerome Angelo Chavez testified at trial that he had known appellant for approximately 10 years. Between 7 and 8 o’clock on the morning of December 3, 1980, appellant visited Chavez at Chavez’ house in Bakersfield. Appellant told Chavez that the night before he, Paul Soto and Gregorio Franco had gone to a farm labor camp to rob wetbacks. Appellant told Chavez that Gregorio was driving the car and that he and Paul had gone to the rear of a house where two men were sleeping in a room. While they were attempting to force the men to turn over their money, another man came to the door and started a fight with them. To prevent the man from fighting with them, and because the man had hit appellant’s leg with a pipe, appellant turned around and shot him.

When appellant was arrested, he was given his Miranda rights by Sergeant Joe Orman of the Kern County Sheriff’s Department and asked to waive them. When appellant refused to waive his rights, Sergeant Orman took appellant to the booking station and began the booking process. At some point during the booking process, appellant indicated to Sergeant Or-man that he would like to speak to him alone. Sergeant Orman took appellant to a room down the hall where they could speak privately. Upon entering the room, appellant said, “We were over there, but I didn’t shoot him, Paul did.” Appellant also told Sergeant Orman that he would be willing to testify against Paul Soto in exchange for a guarantee of a one-year maximum sentence in the county jail.

While appellant was an inmate in the Kern County jail, he became acquainted with a fellow inmate named Wayne Blackwell. At a preliminary hearing, Blackwell testified that appellant told Blackwell about the crime that he had committed, including the fact that during the course of the robbery he had shot a man in the stomach. Appellant also told Blackwell that a person he had confided in about the crime had apparently decided to turn state’s evidence against him. Appellant asked Blackwell to kill the witness, Jerome “Bo-Bo” Chavez, and in return appellant promised to kill one of Blackwell’s enemies in the future. Blackwell testified that appellant drew a map detailing the way to Chavez’ house from the Kern County jail. Prior to and at the trial, however, Blackwell could not be located. After a hearing on the prosecution’s efforts to locate him, the trial court allowed Blackwell’s preliminary hearing testimony to be presented to the jury.

Daniel Thompson was also an inmate in the Kern County jail who became acquainted with appellant during his tenure there. Thompson testified that appellant told him that Paul Soto had gone with appellant to a trailer to rob some illegal aliens. Appellant also told Thompson that he had shot someone during the course of the robbery. In addition, Thompson also testified that appellant told him he had arranged for a fellow inmate to kill the witness in his case who was turning state’s evidence.

Appellant was charged by information on January 15, 1981, with murder in the first degree in violation of Penal Code section 187. The first count also alleged that the murder occurred while appellant was attempting to commit a robbery in violation of Penal Code section 211, within the meaning of Penal Code section 190.2, subdivision (a)(17)(i). A firearm-use allegation in violation of Penal Code section 12022.5 was also contained in the first count.

Counts II and III charged appellant with attempted robbery and included firearm-use allegations. Count IV charged appellant with burglary and a firearm-use allegation in connection with the burglary.

On April 17, 1981, another information was filed against appellant charging him with solicitation of murder in violation of Penal Code section 653f. The two cases were ordered consolidated.

On June 1, 1981, appellant filed a motion challenging the jury selection process. After receiving evidence on the motion to dismiss the jury panel, the court denied the motion.

Subsequently, the district attorney made a motion to have the court rule that the People had exercised due diligence in attempting to locate a witness named Wayne Blackwell. The court made the finding requested by the district attorney.

On September 14, 1981, jury selection began and, ultimately, appellant was convicted on all counts alleged. He was sentenced to life imprisonment without possibility of parole.

Discussion

I

Appellant contends that the trial court erred in refusing to find that he had presented a prima facie case in support of his allegation that Hispanics were underrepresented in Kern County jury pools in violation of his constitutional right to have a jury drawn from a fair cross-section of the community.

In support of his motion, appellant presented the testimony of Dr. Terry Newell, a psychologist who is an expert in statistics. Dr. Newell testified that 17.76 percent of the population of Kern County were Hispanics age 18 and older. This figure represented an estimate derived by Dr. Newell from statistics taken from the 1970 and 1980 United States census.

Dr. Newell also testified that approximately 8.3 percent of the jurors in Kern County jury pools were Hispanic. This figure was based upon surveys conducted by Dr. Newell from October 1980 through February 1981 and in May 1981. The surveys were conducted among both those who were merely summoned for jury duty and those who actually appeared for jury duty.

Based upon the statistics compiled by Dr. Newell, there is a 9.46 percent difference between the percentage of Hispanics in Kern County and the percentage of Hispanics on Kern County jury pools. According to Dr. New-ell, there is only an “infinitesimally small” probability that these statistics would occur by chance and, therefore, they represent a “consistent under-representation of Hispanics in the jury pool.”

At the hearing on the motion, testimony was also taken from Bereyl White, who is the jury coordinator for Kern County, and Larry Drake, who is the director of the Data Processing Department of Kern County. Both men testified that the jury pool is randomly selected from the roll of registered voters in Kern County. At the time the hearing was conducted, Department of Motor Vehicle lists were not in use.

Following the hearing, the court took the matter under submission and later denied the motion. The court said, “The motion is denied. The People are correct in pointing out that the statistics used are incorrect in that the 18 plus or minus percent of Spanish in Kern County includes both citizens, resident aliens and non-resident aliens.”

On appeal, appellant argues that the instant case is controlled by People v. Harris (1984) 36 Cal.3d 36 [ 201 Cal.Rptr. 782 , 679 P.2d 433 ]. Appellant contends that, pursuant to Harris , the evidence presented in the instant case constituted a prima facie case of underrepresentation in the venire by a cognizable class (Hispanics) and that the trial court, therefore, erred in refusing to excuse the jury panel. Appellant acknowledges that Harris was decided subsequent to the hearing and conviction in the instant case but argues that Harris should be given retroactive effect.

Respondent argues that Harris should not be given retroactive effect. Respondent concedes that a retroactive application of Harris to the instant case would require a reversal on this issue.

A brief discussion of the facts in Harris is necessary for a determination of the retroactivity issue. In Harris , the defendant moved to quash the jury venire which had been randomly selected from voter registration lists. The defendant argued that because a smaller percentage of blacks and Hispanics register to vote than whites, the use of voter registration lists as the sole source for the jury pool leads to the underrepresentation of blacks and Hispanics in the jury pool. Harris contended that this violated his constitutional right to a jury drawn from a fair cross-section of the community as guaranteed by the Sixth Amendment of the United States Constitution and article I, section 16 of the California Constitution.

In support of this argument, the defendant presented an expert witness who testified, based upon general population statistics taken primarily from the 1970 census, that in 1980, 33 percent of Los Angeles County was Hispanic and 15 to 17 percent were black. Following the trial in Harris , the 1980 census statistics were released. These statistics, which the Supreme Court relied upon, revealed that 27.3 percent of Los Angeles County was Hispanic and 12.6 percent were black. (People v. Harris, supra, 36 Cal.3d at pp. 47-48, 56.) In addition, based upon surveys of jury venires, the expert testified that 5.5 percent of potential jurors were black and 3.4 percent were Hispanic. Based upon the 1980 census figures used by the court, the statistics revealed that blacks were underrepresented by 7.1 percent and Hispanics by 24.2 percent. (Id., at p. 56.)

In response to the defendant’s contentions in Harris , the People pointed out that the evidence relied upon by Harris was based upon total population figures and was not based upon voter eligibility statistics. The People contended that the showing was, therefore, inadequate, unreliable and could not serve as a basis for finding ethnic underrepresentation. (Id., at pp. 45, 51-52.)

On appeal, the Supreme Court reversed the trial court and pointed out that the Sixth Amendment of the federal Constitution (Taylor v. Louisiana (1975) 419 U.S. 522, 530 [ 42 L.Ed.2d 690, 698 , 95 S.Ct. 692 ]) and 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 ]) guarantee criminal defendants the right to trial by an impartial jury drawn from a representative cross-section of the community. “ ‘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 ].)” (People v. Harris, supra, 36 Cal.3d at p. 50 .)

In the Harris case, it was undisputed that blacks and Hispanics constitute distinctive groups. The Supreme Court rejected the argument of the People that the defendant did not meet his burden of showing that these groups were not adequately represented on venires. The People relied upon prior case law that held that when a defendant uses statistics based upon total population figures, and not on the number of people within the group who are eligible to serve on a jury, the statistics cannot be used to establish a prima facie case of underrepresentation.

The Supreme Court noted that “[c]riticism has been appropriately leveled at the use of total population figures as the comparison base for showing underrepresentation. [Citations.]” (Id., at p. 53.) The Harris court then rejected the People’s argument and found that placing the burden upon the defendant to obtain more refined statistics was too burdensome. (Id., at pp. 53-58.) The Harris court held that a defendant can rely upon statistics based upon total population in order to establish a prima facie case where the figures show a gross disparity resulting in a violation of the fair cross-section rule. The Supreme Court further held that the burden then shifts to the People to rebut the prima facie case through the use of more refined statistics which show that the underrepresentation is at an acceptable level, or that even through the use of multiple sources for compiling the jury pool lists a certain disparity is unavoidable or, finally, that a significant state interest is advanced by use of the jury selection process under scrutiny. (Id., at p. 59.)

Referring to the third prong of the three-part test expressed in Duren v. Missouri (1979) 439 U.S. 357 [ 58 L.Ed.2d 579 , 99 S.Ct. 664 ], the court held that systematic exclusion is shown when the system of selection that is being employed results in a jury pool that does not represent a fair cross-section of the community. (People v. Harris, supra, 36 Cal.3d at p. 58 .)

In conclusion, the Harris court noted: “The error in concluding that defendant had failed to make a prima facie showing of a violation of his right to a jury drawn from a representative cross-section of the community is prejudicial per se. ‘ “The right to a fair and impartial jury is one of the most sacred and important of the guaranties of the constitution. Where it has been infringed, no inquiry as to the sufficiency of the evidence to show guilt is indulged and a conviction by a jury so selected must be set aside.” ’ [Citation.]” (Id., at p. 59.)

It is apparent that the three-prong test of Duren is satisfied in the instant case and the remaining question is whether Harris should have retroactive application.

In Donaldson v. Superior Court (1983) 35 Cal.3d 24 [ 196 Cal.Rptr. 704 , 672 P.2d 110 ], the Supreme Court set forth the test to be applied in determining whether or not a decision should be given retroactive effect. “In determining whether a decision should be given retroactive effect, the California courts undertake first a threshold inquiry, inquiring whether the decision established new standards or a new rule of law. If it does not establish a new rule or standards, but only elucidates and enforces prior law, no question of retroactivity arises. [Citations.] Neither is there any issue of retroactivity when we resolve a conflict between lower court decisions, or address an issue not previously presented to the courts. In all such cases the ordinary assumption of retrospective operation [citations] takes full effect. . . .

Both parties agree that Harris has established a “new rule.” The agreement of the parties finds support in the fact that a new rule of Harris “ ‘overturns a longstanding and widespread practice to which [the Supreme Court] has not spoken, but which a near-unanimous body of lower court authority has expressly approved.’ [Citation.]” (Donaldson, supra, 35 Cal.3d at p. 37 .)

In support of this observation, it is noted that the decisions of the Courts of Appeal have nearly unanimously held that the use of total population figures was insufficient to establish a prima facie case of underrepresenta tion. (People v. Mooring (1982) 129 Cal.App.3d 453, 459 [ 181 Cal.Rptr. 71 ]; People v. Remiro (1979) 89 Cal.App.3d 809, 840 [ 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. Spears (1975) 48 Cal.App.3d 397, 404 [ 122 Cal.Rptr. 93 ]; People v. Powell (1974) 40 Cal.App.3d 107, 128-129 [ 115 Cal.Rptr. 109 ]; contra, People v. McDowell (1972) 27 Cal.App.3d 864, 870 [ 104 Cal.Rptr. 181 ].)

Therefore, whether or not Harris is retroactive in its application must be determined by considering the three-prong test of Stovall v. Denno (1967) 388 U.S. 293 [ 18 L.Ed.2d 1199 , 87 S.Ct. 1967 ]. The first prong analyzes the purpose to be served by the new standards. Appellant contends that the primary purpose of the Harris decision is to insure that defendants receive a fair trial before an impartial jury. Respondent contends that the purpose of Harris is to lighten the defendant’s burden in showing a violation of the fair cross-section rule by allowing him to use total population figures. We concur with respondent’s contention.

The major thrust of the Harris opinion is the adoption of a rule allowing a defendant to establish a prima facie case through the use of general population figures. “The problem of narrowing readily available census figures to reflect a population defined to include only those presumptively eligible for jury service raises two fundamental questions presented by this case: Is use of total population figures sufficient to make a prima facie case of a violation of defendant’s right to a jury drawn from a representative cross-section of the community? Which party has the burden of showing that a more narrowly defined population would or would not result in a disparity of constitutional consequence? We conclude 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.” (People v. Harris, supra, 36 Cal.3d at pp. 45-46.)

In determining whether or not the Harris rule supports retroactive application, a determination must be made as to whether or not the rule of Harris is essential to a reliable determination of whether an accused should suffer a penal sanction or whether the rule is seen as vindicating interests which are collateral to or relatively far removed from the reliability of the fact-finding process at trial. (See In re Johnson (1970) 3 Cal.3d 404, 411 [ 90 Cal.Rptr. 569 , 475 P.2d 841 ].)

We find that the rule of Harris is not directed at vindicating a right which is essential to a reliable determination of whether an accused should suffer a penal sanction. This reasoning is strongly supported by Daniel v. Louisiana (1975) 420 U.S. 31 [ 42 L.Ed.2d 790 , 95 S.Ct. 704 ], wherein the court refused to apply the rule of Taylor v. Louisiana, supra, 419 U.S. 522 retroactively. Taylor held a system which automatically excluded women from jury panels to be unconstitutional in violation of the fair cross-section rule. The court stated that its decision in Taylor “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.” (Daniel v. Louisiana, supra, 420 U.S. at p. 32 [42 L.Ed.2d at p. 793]; see also People v. Navarette (1976) 54 Cal.App.3d 1064, 1074-1076 [ 127 Cal.Rptr. 55 ],) Utilizing the first criteria of Stovall, supra, we find that the Harris rule is not predicated on the integrity of the fact-finding process. Having determined that the Harris rule is collateral to the fair determination of guilt or innocence, we conclude that the first prong of the Stovall test leans toward nonretroactivity with respect to the Harris rule.

The second prong of the Stovall test determines the extent of reliance placed on the old standards by law enforcement authorities. Appellant contends that the only “law enforcement authority” to rely upon the old rule was the Kern County District Attorney’s office. As noted earlier, however, the Courts of Appeal of this state have been virtually unanimous in their determination that the use of total population figures was insufficient to establish a prima facie case of underrepresentation. Given the number of published opinions reexpounding the “old rule,” it is clear that many law enforcement authorities, other than the Kern County District Attorney’s office, have relied on the old rule. Therefore, we find that the second prong of the Stovall test favors a prospective application of the Harris rule.

The third prong of the Stovall test examines the effect on the administration of justice that the retroactive application of the new rule would have. The record does not disclose the number of cases that would have to be retried based upon a retroactive application of Harris . It is evident, however, that because the issue involves the composition of jury pools in all criminal cases in counties where voter registration lists have been the sole source for the selection of jury pools, we can infer that the impact would be substantial. In any event, it is “clear that the factors of reliance and burden on the administration of justice are of significant relevance only when the question of retroactivity is a close one after the purpose of the new rule is considered. (Desist v. United States [1969] 394 U.S. [244], 251 . . . .)” (In re Johnson, supra, 3 Cal.3d 404, 410 .) Therefore, our analysis of the first and second prongs of Stovall does not require that we determine the impact on the administration of justice of a retroactive application because we do not find that the first two prongs of Stovall necessitate retroactive application.

Having determined that Harris should not be given retroactive application, the only remaining issue pertaining to the motion concerning the jury venire was whether or not the trial court erred when it denied appellant’s motion to dismiss the jury panel prior to the Supreme Court’s pronouncement in the Harris case. The “pre-Harris” rule can be stated as follows: “Courts have repeatedly emphasized that no reliable conclusion can be drawn when total population rather than voter population is used as the base. (People v. Powell, supra, 40 Cal.App.3d 107, 128-129, 133 ; People v. Spears, supra, 48 Cal.App.3d 397, 404 ; Adams v. Superior Court, supra, 27 Cal.App.3d 719, 729 [ 104 Cal.Rptr. 144 ].)” (People v. Lewis, supra, 74 Cal.App.3d 633, 646 , fn. omitted.) We find that under the application of the “old rule” it is clear that appellant failed to establish a prima facie case and, therefore, we find that the lower court was correct in denying appellant’s motion under the then existing rule of law.

II

Appellant contends that, pursuant to Carlos v. Superior Court (1983) 35 Cal.3d 131 [ 197 Cal.Rptr. 79 , 672 P.2d 862 ], he is entitled to a new trial on the special-circumstances allegation. In Carlos , the California Supreme Court held that proof of intent to kill or to aid a killing is essential to sustain a felony murder special-circumstances allegation under the 1978 death penalty law. (Id., at pp. 153-154.) In People v. Garcia (1984) 36 Cal.3d 539 [ 205 Cal.Rptr. 265 , 684 P.2d 826 ], the Supreme Court held that Carlos is to be given full retroactive effect to all cases not yet final. {Garcia, supra, at p. 549 .) Garcia held that Carlos requires reversal if the instruction given the jury “completely eliminated the issue of intent to kill from the consideration of the jury.” (Garcia, supra, at p. 554 .)

The Garcia case recognizes four exceptions to its rule mandating automatic reversal. The first two exceptions are based on the United States Supreme Court case of Connecticut v. Johnson (1983) 460 U.S. 73 [ 74 L.Ed.2d 823 , 103 S.Ct. 969 ], A conviction will not be reversed “‘if the erroneous instruction was given in connection with an offense for which the defendant was acquitted and if the instruction had no bearing on the offense for which he was convicted,’ ” or “ ‘if the defendant conceded the issue of intent.’” (People v. Garcia, supra, 36 Cal.3d at p. 554, quoting Connecticut v. Johnson, supra, 460 U.S. at p. 87 [ 74 L.Ed.2d at p. 834 , 103 S.Ct. at pp. 977-978.)

The third exception expressed in Garcia is based upon People v. Sedeno (1974) 10 Cal.3d 703, 721 [ 112 Cal.Rptr. 1 , 518 P.2d 913 ], This exception allows a court to conclude that the omitted instruction was harmless error where “ ‘the factual question posed by the omitted instruction was necessarily resolved adversely to the defendant under other, properly given instructions.’ ” (People v. Garcia, supra, 36 Cal.3d at p. 555, quoting People v. Sedeno, supra, 10 Cal.3d at p. 721 .)

The fourth exception to automatic reversal expressed in Garcia is predicated on People v. Cantrell (1973) 8 Cal.3d 672, 685 [ 105 Cal.Rptr. 792 , 504 P.2d 1256 ] and People v. Thornton (1974) 11 Cal.3d 738, 768-769 , footnote 20 [ 114 Cal.Rptr. 467 , 523 P.2d 267 ]. This exception applies in “cases where the parties recognized that intent to kill was an issue, presented all evidence at their command on that issue, and in which the record not only establishes the necessary intent as a matter of law but shows the contrary evidence not worthy of consideration.” (People v. Garcia, supra, 36 Cal.3d at p. 556, fn. omitted.)

In the instant case, the jury was instructed that in order to find the accused guilty of first degree felony murder, they did not have to find that the accused acted with the intent to kill. In addition, the instruction given the jury regarding the special-circumstances allegation did not require the jury to find that appellant acted with the intent to kill before they could find the special-circumstances allegation to be true.

Applying the foregoing rules to the instant case, it is apparent that the first two exceptions pursuant to Connecticut v. Johnson and the fourth exception pursuant to Cantrell and Thornton are not applicable. Appellant was found guilty of each allegation he was charged with, including the special circumstances. Appellant did not concede the issue of intent at trial and there is nothing in the record to indicate that intent to kill was recognized by the parties as an issue that needed to be expressly resolved.

A review of the record in the light of the Sedeno exception demonstrates that while the jury was instructed that in order to find the defendant committed first degree murder they had to find that he acted with the specific intent to kill, the jury was also instructed that first degree felony murder only required them to find the defendant acted with a specific intent to commit the underlying robbery and not a specific intent to kill. Regardless of whether the premeditated murder instruction was superfluous, the record does not disclose, nor do we infer, that it was necessary for the jury to find that the defendant acted with the specific intent to kill in order for them to arrive at the decision concerning first degree murder and the special circumstance alleged.

We therefore find that because Garcia holds that Carlos is to be given full retroactive effect, the special-circumstances finding must be set aside and the case remanded to retry the special-circumstances allegation in compliance with the requirement of Carlos v. Superior Court.

III

Conclusion

Based upon our discussion herein, appellant’s conviction of the special circumstance is reversed pursuant to the mandates of Garcia and Carlos . In all other respects, his conviction is affirmed.

Appellant’s petition for a hearing by the Supreme Court was denied January 31, 1985.

Parts III, IV, V and VI are not published, as they do not meet the standards of publication contained in rule 976(b), California Rules of Court. Assigned by the Chairperson of the Judicial Council. The survey included a question about ethnic origin. Therefore, the 8.3 percent of Hispanics represent individuals who consider themselves Hispanics and is not based on the surname a person has. It should also be noted that People v. Wheeler, supra, 22 Cal.3d 258 , another landmark case from the California Supreme Court involving the fair cross-section rule, was not given full retroactive effect. (Id., at p. 283, fn. 31.) (People v. Mooring, supra, 129 Cal.App.3d 453, 459 ; People v. Remiro, supra, 89 Cal.App.3d 809, 840 ; People v. Lewis, supra, 74 Cal.App.3d 633, 646 ; People v. Spears, supra, 48 Cal.App.3d 397, 404 ; People v. Powell, supra, 40 Cal.App.3d 107, 128-129 ; contra, People v. McDowell, supra, 27 Cal.App.3d 864, 870 .) See footnote 1, ante.