*521Opinion
GRODIN, J.Defendant Albert Greenwood Brown, Jr., was convicted on one count of rape (Pen. Code, § 261, former subd. (3))1 with the infliction of great bodily injury (§ 12022.8) (count I) and one count of first degree murder (§§ 187, 189) (count II). The jury made a special finding that the murder was premeditated. A special circumstance that the murder was committed in the course of a rape (§ 190.2, subd. (a)(17)(iii)) was found true. Defendant admitted allegations that he had suffered a prior conviction and prison term for rape. (§§ 667.5, subd. (a), 667.6, subd. (a).)
Acting under the 1978 death penalty initiative law (§§ 190.1-190.7), the jury fixed the punishment on count II as death. The court denied the automatic application for modification of judgment (§ 190.4, subd. (e)) and imposed a further sentence of thirteen years on count I (the upper term for rape, plus a consecutive five years for the great bodily injury) with a five-year enhancement for the prior prison term. This appeal is automatic.
Defendant raises several claims of error at the guilt and special circumstance phase of his trial. We find merit in defendant’s contentions that the testimony of a hypnotized witness, evidence of forensic tests of crime-scene fluid stains, and statistics about defendant’s blood characteristics were improperly admitted at the guilt phase. However, we conclude that the errors were harmless in light of the extremely strong evidence against him. We will therefore affirm the guilt and special circumstance findings.
At the penalty phase, we agree with defendant’s objection to instructions that the jury must not be swayed by sympathy or consequences in choosing a sentence. Prior authority of this court flatly prohibits the giving of such antisympathy instructions at a capital penalty trial. (People v. Lanphear (1984) 36 Cal.3d 163, 166 [203 Cal.Rptr. 122, 680 P.2d 1081]; People v. Easley (1983) 34 Cal.3d 858, 876 [196 Cal.Rptr. 309, 671 P.2d 813].) We are persuaded that their inclusion in this case was prejudicial on the issue whether defendant should live or die. The penalty judgment must therefore be reversed.
Defendant also argues that the 1978 death penalty law is unconstitutional, and contends he cannot be resentenced under its provisions, on grounds among others that it impermissibly provides for a mandatory death penalty under certain circumstances. We will conclude that the 1978 statute, correctly construed, preserves the jury’s constitutional discretion to decide the appropriate penalty and is therefore valid. We will note, however, that *522trial courts in future cases should give instructions which clarify the sentenced s responsibility.
I. Guilt Trial
A. Prosecution case.
On October 28, 1980, about 7:30 a.m., 15-year-old Susan J. left her home on Victoria Avenue in Riverside to walk to school with her younger brother and sister. After the younger children left Susan to walk to their elementary school, she continued up Victoria Avenue toward Arlington High School. She never arrived, and efforts throughout the day to locate her were unsuccessful. Her parents telephoned the police.
Sometime between 7 and 7:30 that evening, the telephone at Susan’s home rang, and Susan’s mother answered it. The male-voiced caller said, “Hello, Mrs. J., Susie isn’t home from school yet, is she?” Mrs. J. replied that she was not. The voice then declared, “You will never see your daughter again. You can find her body on the corner of Victoria and Gibson.” At Mrs. J.’s request, the caller repeated the information, then hung up. Mrs. J. telephoned the police again.
Around 7:30, the Riverside Police Department received another call. A male voice said, “On the corner of Gibson and Victoria, fifth row, you will find a white Caucasian body of a young girl in the orange grove.”
Police officers were sent with a police dog to the orange grove at the corner of Gibson and Victoria. They found nothing. Officer Taulli, one of the policemen at the scene, went to the J. residence to get an article of Susan’s clothing to be used by the dog as a scent guide.
Officer Taulli arrived at the J. home about 8:30; Police Chaplain Phillip Morgan arrived at the same time. While Taulli was still there, the telephone rang, and he answered it. A male voice asked if “this [is] the [J.] house or the [J.] residence?” Taulli advised that it was. The caller said, “You can find Sue’s identification in a telephone booth at the Texaco station at Arlington and Indiana.” Taulli told Morgan and Mrs. J. to record any further calls verbatim, then returned to the grove with an article of Susan’s clothing.
After sniffing the item brought by Taulli, the dog shortly found a pair of torn panties in the grove. The dog then led police down the next six rows of trees. There Susan’s body was found lying face down, with dirt piled up on both sides of the head. The body was nude below the waist, except for *523socks, and Susan’s bra was partially pulled out from under her blouse. Several school notebooks and Susan’s tennis shoes were found near the body. Her jeans were located elsewhere in the grove. A shoelace, apparently from one of her shoes, was wrapped tightly around her neck. Susan was holding a spark plug wire cap in her hand, and a spark plug wire was discovered nearby.
Homicide investigators were called to the scene. They found signs of a struggle and indications that the body had been dragged for some distance. Shoeprints in a herringbone pattern were found around the body and photographed. Susan’s blouse was stained and swabs were taken from her vagina and abdomen.
The same night, officers were sent to the Texaco station at the corner of Arlington and Indiana. There, in a telephone booth, they discovered two Arlington High School identification cards belonging to Susan and a library pouch from a book. The pouch was stamped with the words “Arlington High School.”
Meanwhile, Chaplain Morgan had obtained a tape recorder and hooked it up to the J.s’ phone. About 9:30 p.m., the phone rang again. The same male voice said, “In the tenth row, you’ll find the body.”
Early next morning, the police set up roadblocks on the streets near the grove and questioned passersby. Several remembered seeing Susan the morning before, walking on a bike trail through the grove in the block of Victoria between Gibson and Van Burén Boulevard. Others had additionally seen a black man approaching Susan on the bike trail, standing in the grove as she walked by, or following her. The man was wearing jogging clothes; two witnesses more particularly described green running shorts and a green and white shirt.
A number of people also recalled seeing a late-model brown or copper-colored Pontiac Trans-Am parked nearby; it bore a distinctive paper license plate with the words “Made in USA” or “Made in America.” Peter Rodriguez saw a black man emerge from the grove and open the trunk of the Trans-Am; the man kept staring at Rodriguez. Margery Johnston also saw a black man in running clothes come out of the grove. He appeared startled and his legs were dusty or dirty.
Police began a surveillance of defendant’s Gertrude Street residence. On November 6, he drove up in a brown Trans-Am; he was arrested when he drove away again. A warrant to search the residence was obtained and executed that evening. Behind a water heater in the garage, police found a *524crumpled-up paper license plate which read “Made in America.” Inside the house, a telephone directory was turned back to the page containing the J.s’ listing. There were newspaper articles about Susan’s death under defendant’s bed, and two of her missing schoolbooks were found in the den. The library pouch found in the telephone booth had come from one of the books.
Defendant’s locker at work was also searched. Police seized jogging clothes, including green running shorts and a green and white shirt. Undershorts found in the locker had semen stains. The locker also contained running shoes; the pattern of their soles closely matched the shoeprints found at the crime scene.
Tests determined that Susan had died by strangulation sometime between 7 and 11 a.m. on October 28. Analyses of the stains and swabs obtained from the body revealed the presence of semen.
Three witnesses positively identified defendant at trial as the man they saw in the vicinity of the grove on the mornipg of October 28. Wiley Eng, a high school student, said he was riding his bicycle on the bike trail in the grove. He overtook Susan, who was travelling on foot in the same direction, in the block between Van Burén and Gibson. He then passed the man, who was walking toward him, at a distance of two or three feet. Eng had some three seconds to see the man’s face. He had picked defendant from a photo lineup, saying then he was 60 to 70 percent sure it was the same man. His identification at the preliminary hearing and at trial were unequivocal, though he admitted newspaper photos of defendant had helped him decide.
Julie Pim, another high school student, testified she was a passenger in her brother’s truck, which had stopped for a red light at Victoria and Van Burén on the way to school. From 40 feet, she saw Susan, whom she knew from elementary school, pass close to a man she identified as defendant while crossing the intersection. Ms. Pim had picked defendant from a photo lineup of eight black males, saying she was 60 percent certain and could tell better in person. She saw defendant’s newspaper photo before the preliminary hearing, but she denied it aided her positive identification at that proceeding. She admitted that she might have been influenced at the preliminary hearing by the fact that defendant was the only black person present.
Margery Johnston also positively identified defendant at trial. She had been unable to pick defendant from a photo lineup while under hypnosis. She testified that he appeared different in person than in the photos.
Henry Garcia and Peter Rodriguez testified that defendant looked like the man they had seen that morning, but neither could be certain. Several wit*525nesses confirmed that the photos of defendant’s car produced at trial matched the automobile parked near the grove on the morning of October 28.
Over defendant’s objection, Faye Springer and Rodney Andrus, two criminalists from the California Department of Justice (CDJ), testified on the results of their comparison of the victim’s and defendant’s blood, the blouse stains, the vaginal and abdominal swabs, and the semen stains on the undershorts taken from defendant’s locker. Tests were performed in four categories of inherited genetic characteristics (see discussion, post). Springer testified that the stains and swabs matched defendant’s genetic characteristics in several respects, and that defendant’s characteristics were shared by only a small percentage of the black population.
William Anderson and Norm Gibson, acquaintances of defendant, identified the voice on the taped call to the J. residence as that of defendant.
B. Defense case.
Defendant presented an alibi defense. His mother testified that defendant lived with her at the Gertrude Street house in October 1980. She arrived home from work at 7:40 a.m. on October 28. Defendant was at home. He left to buy milk and returned at 7:48. He ate breakfast and left for work at 8:14.
II. Penalty Trial
At the penalty phase, the prosecution presented evidence of defendant’s 1977 rape of 14-year-old Kelly P. Defendant pled guilty to the rape and was sentenced to state prison. He was released in June 1980 on one year’s parole.
The defense presented psychiatric and background evidence suggesting that defendant suffered severe emotional problems, including extreme sexual maladjustment and dysfunction. Numerous relatives testified to their affection for defendant. Defendant himself took the stand, expressed remorse for his rape of Kelly P., and asked the jury to show mercy.
III. Guilt Issues
A. Failure to permit expert testimony on eyewitness identification.
At trial, defendant offered Dr. Robert Shomer, a psychologist, as an expert witness on the factors which may cause mistaken observations by eyewitnesses. The People objected on grounds there was no showing that *526any witness had specific psychological problems which might cause abnormal perception. The trial court refused to permit Dr. Shomer’s testimony. Defendant argues he was thus denied his due process rights to present witnesses in his defense. (Chambers v. Mississippi (1973) 410 U.S. 284, 302 [35 L.Ed.2d 297, 312, 93 S.Ct. 1038].)
The trial court relied on the then-established rule (e.g., People v. Guzman (1975) 47 Cal.App.3d 380, 385-386 [121 Cal.Rptr. 69]; People v. Johnson (1974) 38 Cal.App.3d 1, 6-7 [112 Cal.Rptr. 834]) that expert testimony of this type may be excluded if it relates to “normal” eyewitnesses, since the variables aifecting normal perception are common knowledge and the expert testimony tends to invade the province of the jury. Our recent decision in People v. McDonald (1984) 37 Cal.3d 351 [208 Cal.Rptr. 236, 690 P.2d 708], however, makes clear that expert “eyewitness” testimony is not excludable on either of these traditional grounds.
As McDonald noted, many factors which compromise the accuracy of eyewitness observation extend beyond lay understanding so that expert information on the subject would “assist the trier of fact.” (P. 369; see Evid. Code, § 801, subd. (a).) Moreover, McDonald observed, such testimony does not invade the jury’s province, since it expresses no view on the credibility of a particular eyewitness; in any event, California has rejected the rule that expert testimony is inadmissible when it coincides with the “ultimate issue” in the case. (P. 371.)
Trial courts retain discretion under McDonald to exclude expert testimony of this kind on grounds that it is unnecessary in a particular case, but appellate deference is not absolute. “When an eyewitness identification of the defendant is a key element of the prosecution’s case but is not substantially corroborated by evidence giving it independent reliability, and the defendant offers qualified expert testimony on specific psychological factors shown by the record that could have affected the accuracy of the identification but are not likely to be fully known to or understood by the jury, it will ordinarily be error to exclude that testimony. [Fn. omitted.]” (P. 377.)
Applying the Watson standard of prejudice (People v. Watson (1956) 46 Cal.2d 818, 836 [299 P.2d 243]), we concluded that the trial court in McDonald had abused its discretion when it excluded the expert testimony proffered by defendant, and that the error warranted reversal. (P. 376.) McDonald was accused of shooting a robbery victim to death in the street, but no circumstantial evidence linked him to the crime. The People’s case hinged entirely on equivocal eyewitness identifications, and one prosecution eyewitness testified unequivocally that McDonald was not the killer. (Pp. 355-360.) Numerous defense witnesses insisted that McDonald was in Alabama on the day the robbery-murder occurred. (P. 360.) Thus, the trial court’s exclusionary ruling was “crucial”—it “undercut the *527evidentiary basis of defendant’s main line of defense”—and “impair[ed] the jury’s determination of an issue that [was] both critical and closely balanced . . . .” (P. 376.)
By contrast, ample circumstantial evidence connects defendant in this case to Susan’s death. That evidence includes the consistent description of defendant’s car, the distinctive license plate found concealed in his garage, the incriminating items retrieved from his residence, and the clothing and shoes recovered from his locker at work. Therefore, we cannot say the trial court abused its discretion by excluding expert testimony on eyewitness identification. Even if an abuse of discretion is assumed, the error was clearly harmless.
B. Hypnosis of prosecution witness.
Defendant urges that the testimony of Margery Johnston, including her in-court identification, must be excluded under People v. Shirley (1982) 31 Cal.3d 18 [181 Cal.Rptr. 243, 641 P.2d 775], because the witness had previously undergone hypnosis to enhance her recall. In Shirley this court ruled that “the testimony of a witness who has undergone hypnosis for the purpose of restoring his memory of the events in issue is inadmissible as to all matters relating to those events, from the time of the hypnotic session forward.” (Pp. 66-67.)
The People respond first that Shirley is inapplicable, because an extensive voir dire examination of Johnston revealed that her memory of events was not affected by the hypnotic session. Shirley itself concluded, however, that a mere comparison of pre- and posthypnosis statements does not dispel the adverse implications of hypnosis. Hypnosis, said Shirley, not only creates “pseudomemories,” it tends “to clothe the witness’ entire testimony in an artificial but impenetrable aura of certainty [fn. omitted], . . .” (P. 69.)
The People next suggest that the Shirley rule does not apply to hypnotic sessions conducted before Shirley was filed. This court’s recent opinion in People v. Guerra (1984) 37 Cal.3d 385 [208 Cal.Rptr. 162, 690 P.2d 635] has resolved that issue. Guerra concluded that the Shirley rule applies to all cases not yet final when it was announced, regardless of when the challenged hypnotic session occurred. Contrary authorities cited by the People, in particular People v. Williams (1982) 132 Cal.App.3d 920 [183 Cal.Rptr. 498], were expressly disapproved. (37 Cal.3d at p. 413, and fn. 24.)
However, any error in admitting Johnston’s testimony here was clearly harmless. Two other persons, Wiley Eng and Julie Pim, gave unequivocal in-court identifications of defendant as the man they saw near Susan in the orange grove, and his car, with its distinctive license plate, was *528clearly implicated. Numerous articles linking defendant to the victim were found in his residence, and further incriminating evidence was discovered in his locker at work. There is no reasonable probability that exclusion of Ms. Johnston’s testimony would have produced a different verdict on the issue of guilt.2
C. Admissibility of forensic analysis of blood and semen stains.
Over defendant’s repeated objections,3 two CDJ criminalists compared their genetic analyses of dried-fluid and stain samples obtained from the victim’s body and clothing with the results of blood-grouping tests they had run on samples of defendant’s and the victim’s blood. The import of their testimony was that certain of the crime-scene stains were semen which could have been deposited by defendant. One of these witnesses also stated that, according to theories of statistical probability, only 1.2 percent of the black population would match defendant’s genetic characteristics in all the categories tested.
Defendant and his amicus, Dr. Benjamin Grunbaum, first contend that admission of the stain-test evidence was error, since the prosecution failed to satisfy the Kelly/Frye rule (Frye v. United States (D.C.Cir. 1923) 293 Fed. 1013, 1014 [34 A.L.R. 145]; People v. Kelly, supra, 17 Cal.3d 24, 30) by demonstrating at trial the scientific acceptance of the tests performed. As we will explain, we agree.
*529The typing of blood under the inherited antigen groups A, B, and O has long been known, but science has more recently discovered inherited proteins and enzymes which can also be individually typed. Eight of ten persons are “secretors” who carry these substances not only in the blood, but in other body fluids such as semen, vaginal secretions, and saliva. The rough population distributions for the various types of each of these antigens, proteins, and enzymes have been established, and research suggests they are statistically independent. Thus, the greater the number of categories which can be reliably typed in a particular specimen, the smaller the number of potential donors.
Traditional ABO testing is based on the immunological principle that each antigen type will react characteristically in combination with the others. The typing of more recently discovered proteins and enzymes such as PGM, Peptidase-A, AK, and EAP is accomplished by a method known as electrophoresis. In this system, a test sample is placed on a gel medium in an ionized buffer solution. When an electric current is run through the solution, the sample separates and migrates on the medium into characteristic patterns. These are then fixed, dyed, and read visually by the analyst. (See generally Jonakait, Will Blood Tell? (1982) 31 Emory L.J. 833, 836-842.)
Using these tests, the criminalists in this case compared dried stains recovered from the victim’s clothing, abdomen, and vagina against samples of the victim’s and defendant’s blood. The crime-scene stains were collected some eight to twelve hours after the victim’s death. In one instance, testing occurred some two and one-half months later, after the test sample had been mailed from one GDI crime laboratory to another.
Three antigen and enzyme categories were tested; ABO, PGM, and Peptidase-A. The criminalists reported inconclusive results in certain instances. However, they determined that the crime-scene stains were semen, the typing of which matched defendant’s blood in certain of the categories tested. No test excluded defendant as a potential donor.
In Kelly, supra, 17 Cal.3d 24, this court affirmed California’s adherence to the rule first announced in Frye, supra, 293 Fed. 1013 for the admissibility of a new scientific technique: the technique must be “sufficiently established to have gained general acceptance in the particular field to which it belongs.” (Kelly, supra, at p. 30, italics added; Frye, supra, at p. 1014.) Under the Kelly/Frye rule, the proponent of the scientific evidence must establish “(1) the [generally accepted] reliability of the method . . ., usually by expert testimony, and (2) [that] the witness furnishing such testimony [is] properly qualified as an expert to give an opinion on the subject. [Citations.] Additionally, the proponent . . . must demonstrate that correct scientific procedures were used in the particular case. [Citations.]” (Kelly, supra, at p. 30, all italics in original.)
*530Kelly further defined who is a “qualified expert” on the issue of scientific acceptance. The witness must have academic and professional credentials which equip him to understand both the scientific principles involved and any differences of view on their reliability. He must also be “impartial,” that is, not so personally invested in establishing the technique’s acceptance that he might not be objective about disagreements within the relevant scientific community. (Pp. 37-40.)
The witness may cite and rely upon written studies and findings by scientists not actually before the court. (Shirley, supra, 31 Cal.3d at p. 56.) Moreover, because appellate endorsement of a technique ends the need for case-by-case adjudication (Kelly, supra, 17 Cal.3d at p. 32), this court has sometimes looked beyond the trial record, examining California precedent, cases from other jurisdictions, and the scientific literature itself, to ascertain whether a particular technique is generally accepted. (Shirley, supra, at pp. 33-34, 56; Kelly, supra, at pp. 32-35.)
Here, defendant does not seriously dispute the scientific validity of genetic typing tests in general. Rather, he and Dr. Grunbaum focus on the large body of literature which suggests that drying, aging, temperature, contamination (particularly with bacteria or other organic substances), and unknown composition of the test sample—conditions often encountered in forensic work—can affect test results in varying degrees. The defense suggests that no standard, proven, and accepted methodology exists to avoid these dangers.
The People concede the problem of sample deterioration. They urge, however, that the antigens, enzymes, and proteins accepted for forensic testing are those most resistant to adverse conditions. Moreover, they suggest, the issue is not the reliability of the tests in general but the testable quality of particular samples. They assert that the factors which may hamper or help preserve testability are generally accepted by scientists and well known to forensic analysts; since a forensic witness may be examined on these issues, the factfinder has an adequate basis to evaluate the accuracy of particular test conclusions.
No California appellate decision has ruled on the admissibility of aged-stain typing under Kelly/Frye. Several have confirmed that stain analysis is admissible as relevant evidence though it merely includes the defendant among the class of possible donors. (People v. Lindsey (1978) 84 Cal.App.3d 851, 866 [149 Cal.Rptr. 47, 2 A.L.R.4th 485]; People v. Vallez (1978) 80 Cal.App.3d 46, 56 [143 Cal.Rptr. 914].) This court, suggesting that such tests may be useful and “feasible” to the defense in a rape case, has concluded that vaginal swabs taken by the police are material evidence which must be preserved under People v. Hitch (1974) 12 Cal.3d 641 [117 Cal.Rptr. 9, 527 P.2d 361], (People v. Nation (1980) 26 Cal.3d 169, 175, *531177 [161 Cal.Rptr. 299, 604 P.2d 1051].) However, at least two Courts of Appeal, after a review of scientific literature, have expressed misgivings about the reliability of electrophoretic tests for ABO and PGM typing of postcoital vaginal swabs taken within hours after intercourse. (People v. Newsome (1982) 136 Cal.App.3d 992, 999, fn. 4 [186 Cal.Rptr. 676]; People v. Wilson (1982) 128 Cal.App.3d 132, 135-137, and fns. 2, 3 [179 Cal.Rptr. 898].)
Cases in other jurisdictions have considered the scientific-reliability issue directly, with mixed results. Admission of electrophoretic tests of crime-scene stains was upheld in Jenkins v. State (1980) 156 Ga.App. 387 [274 S.E.2d 618]. But Georgia appears to reject the Frye test, allowing the jury to assess the weight and relevance of all expert and scientific evidence. (P. 619.)
In Robinson v. State (1981) 289 Md. 500 [425 A.2d 211], Jean Hostetler, a forensic chemist employed by the Montgomery County Police Department, testified that electrophoretic techniques were “developed in the late ’60’s[, are] now an accepted practice in the field of forensic chemistry,” and are utilized by large numbers of law enforcement agencies including the Federal Bureau of Investigation. She conceded that electrophoresis was not widely used outside crime laboratories, since classification of the substances at issue, other than the A-B-O antigens, has no medical value. (P. 220.)
The Robinson court concluded that “general acceptance” within the field of forensic chemistry was sufficient; the proponent of electrophoretic evidence did not need to show approval by the larger scientific community. {Ibid.) It is questionable whether, under our Kelly criteria, the testimony of Hostetler alone would have been sufficient in California to establish acceptance by impartial scientists in the field of forensic chemistry.
Recently, the Michigan Supreme Court reversed an appellate court opinion which had upheld admission of electrophoretic stain tests. (People v. Young (1983) 418 Mich. 1 [340 N.W.2d 805, 815].) Adhering to the Frye test, the court found that nothing in the trial record sustained the prosecution’s burden of showing that the “novel” electrophoretic techniques employed in the case, though undoubtedly accepted as a “diagnostic and research tool,” also enjoyed the general support of “disinterested and impartial experts” in the forensic context. (Pp. 813-814.) The prosecution’s sole witness, Mark Stolorow, a Michigan State Police criminalist and codeveloper of the technique, did not demonstrate that the tests were either standardized or “sensitive and specific in measuring what [they purported] to measure.” (P. 814.) The Young court declined to undertake its own review of scientific acceptance, but remanded for further hearings in the trial court.
*532The most complete analysis of electrophoretic stain-test reliability appears in State v. Washington (1981) 229 Kan. 47 [622 P.2d 986]. At issue was the Multi-System, a simplified electrophoretic method developed under Law Enforcement Assistance Administration (LEAA) auspices, which allows up to three separate enzymes to be tested in one medium at the same time. The Washington court weighed testimony by Stolorow and Eileen Burnau, a forensic serologist for the Kansas Bureau of Investigation, supporting the method’s reliability and wide acceptance in crime laboratories, against contrary declarations by Dr. Grunbaum, amicus here.
Dr. Grunbaum, a criminalist and biochemist,4 had testified that the MultiSystem, and electrophoretic testing of dried stains in general, are unreliable. As a long-time forensic researcher at the University of California, he had participated in initial stages of the Multi-System project, but had later withdrawn. The court noted Stolorow’s testimony that Dr. Grunbaum’s line of research in the use of a particular gel medium had not proven fruitful and had been rejected. (P. 990.) In particular, it found meritless Dr. Grunbaum’s assertion that the enzyme EAP degrades rapidly in dried stains, accepting Stolorow’s contrary testimony in that regard. (Pp. 991-993.)
On the other hand, a panel of the appellate court of Illinois more recently ruled that the scientific acceptance of electrophoretic bloodstain analysis had not been proved to its satisfaction. In People v. Harbold (1984) 124 Ill.App.3d 363 [464 N.E.2d 734], the court concluded that the skepticism expressed in the prior decisions and legal commentaries, the apparent subjectivity of interpretation of test results, and inconsistencies in the testimony of the trial experts (including that of Stolorow) called for careful examination of the test reliability issue in any retrial. (Pp. 746-748.)
Our review thus makes clear that the acceptance of tests for typing stale body-fluid stains is a matter of substantial legal controversy, Where that issue remains open, the party offering the evidence has the burden of proving in the trial court that a consensus of scientific opinion has been achieved. (Shirley, supra, 31 Cal.3d at p. 54.)
The prosecution did present testimony by Springer and Andrus that the tests they had used were accepted and reliable. However, the People *533do not seriously contend that this showing was sufficient. Indeed, it fell short under several of the criteria discussed in Kelly and Shirley. Springer and Andrus were competent and well-credentialed forensic technicians, but their identification with law enforcement, their career interest in acceptance of the tests, and their lack of formal training and background in the applicable scientific disciplines made them unqualified to state the view of the relevant community of impartial scientists. (Kelly, supra, 17 Cal.3d at pp. 38-40.) Moreover, neither witness backed up his or her opinion with a discussion of the relevant scientific literature. (See Shirley, supra, 31 Cal.3d at pp. 55-56.)
Portions of their testimony actually supported defendant’s claim that procedures for testing forensic samples vary substantially from laboratory to laboratory (even from analyst to analyst), with little attention to scientific confirmation of reliability. Springer and Andrus also conceded that aged, dried, or contaminated samples could produce misleading results, but they did not explain how their test procedures overcame these dangers. The trial record was patently inadequate to establish scientific acceptance of the tests under Kelly/Frye.
In order to end case-by-case controversy over the acceptance of a particular technique, we have occasionally reviewed the scientific literature ourselves in an effort to determine whether a fair consensus on reliability exists. Shirley is the most notable example. (31 Cal.3d at p. 56 et seq.)
The People and, in a separate brief, the California District Attorneys’ Association, warn against such a course here. They urge that the subject matter is too technical and the relevant literature too vast to be assimilated by lay judges lacking the assistance of qualified expert witnesses. Hence, they contend, the result in this case should stand or fall on the trial record, leaving the issue for further development in the trial courts. Under the circumstances, we find wisdom in this view.
We recognize that Kelly/Frye does not demand judicial absorption of all the relevant literature, nor does it require a decision once and for all whether a particular kind of scientific evidence is reliable. The court need only conduct a “fair overview” of the subject, sufficient to disclose whether “scientists significant either in number or expertise publicly oppose [a technique] as unreliable.” (Shirley, supra, 31 Cal.3d at p. 56.)
Often, however, the technical complexity of the subject matter will prevent lay judges from determining the existence, degree, or nature of a scientific consensus or dispute without the interpretive assistance of qualified live witnesses subject to a focused examination in the courtroom. It is for this reason that Kelly/Frye properly emphasizes the record made in the trial court.
*534In Shirley, the scientific issue was relatively simple—the long-known tendency of hypnosis to create undetectable and unshakeable false memories, despite the best intentions of both hypnotist and subject. The experts had responded to that danger in straightforward terms, permitting this court to conclude that “major voices in the scientific community [absolutely] oppose the use of hypnosis to restore the memory of potential witnesses. . . .” {Ibid.)
Here,, both the technical problem and the state of current scientific opinion are more difficult to comprehend. Electrophoretic typing of human fluid stains is a relatively recent development. (See, e.g., Baird, The Individuality of Blood and Bloodstains (1978) 11 J. Canadian Forensic Sci. 83, 103; see also People v. Young, supra, 340 N.W.2d 805, 812.) The number of proteins and enzymes theoretically subject to classification is substantial. Each substance apparently has a somewhat different reaction to adverse environments and conditions, and the effects in each case are not yet fully known.
It does appear that aged or contaminated stains can undergo actual chemical conversions, resulting in spurious or “false positive” test results. (See, e.g., Sensabaugh et al., Genetic Markers in Semen III: Alteration of Phosphoglucomutase Isoenzyme Patterns in Semen Contaminated with Saliva (1980) 25 J.Forensic Sci. 470, 476-477; Kind et al., An Investigation into the Possible Sources of Adventitious ABH Substances in Bloodstain Grouping (1976) 16 J.Forensic Sci.Society 155, 160; Periera et al., Problems Involved in the Grouping of Saliva, Semen, and Other Body Fluids (1976) 16 J.Forensic Sci. Society 151, 152; Jenkins et al., The Problem of the Acquired B Antigen in Forensic Serology (1972) 12 J.Forensic Sci.Society 597, 600, 602.) Moreover, the electrophoretic method itself is apparently performed under substantial chemical and electrical variations, and considerable training and experience are necessary to interpret the visual results.
The People respond with opinions by certain scientists that proper methodology and experienced professional judgment can ensure that any typing results reported will be reliable. (E.g., Sensabaugh, Response to the Misapplication of Genetic Analysis in Forensic Science (letter to the edit.) (1984) 29 J.Forensic Sci.Society 12, 15; Culliford, The Examination and Typing of Bloodstains in the Crime Laboratory (1971) p. 75.) It is not clear from our unaided review of these authorities that impartial science has developed a consensus on the crucial issue: whether for the typing categories (ABO, PGM, Peptidase-A) and body fluids (semen, blood, saliva, vaginal secretion) at issue here, current methodology, employed by qualified technicians, can discriminate reliably between testable and untestable samples and between accurate and inaccurate results.
We do not suggest that such a consensus is lacking. We simply conclude that the answer must abide an adequate future trial record made *535with the help of live witnesses qualified in the applicable scientific disciplines. We therefore do not foreclose future attempts to admit stain-typing evidence based on a foundation such as we have described. (See Kelly, supra, 17 Cal.3d at pp. 40-41.)5 In this case, such a record not having been made, the evidence should not have been admitted.
We find, however, that the error was harmless in light of the overwhelming valid evidence against defendant. In a supplemental brief on the Kelly/Frye issue, defendant argues that the 10-hour jury deliberation suggests the jurors perceived a close case. The undue weight of invalid “scientific” evidence, he urges, was therefore prejudicial.
We are not persuaded by cases defendant cites for the proposition that long jury deliberations indicate prejudice. In People v. Rucker (1980) 26 Cal.3d 368 [162 Cal.Rptr. 13, 605 P.2d 843], defendant had presented an “excellent” diminished capacity defense, which his statements at an improper booking interview directly undermined. (P. 391.) In People v. Woodard (1979) 23 Cal.3d 329 [152 Cal.Rptr. 536, 590 P.2d 391], prior convictions were used improperly to impeach a defense eyewitness on identity; the only contrary evidence on identity was a second eyewitness called by the prosecution. (P. 341.)
In neither case was the length of jury deliberations the sole basis for finding prejudice. Here, despite the virtually unimpeachable prosecution case, the jury may simply have sifted the evidence with special care in light of the capital implications of a guilt verdict. It is not reasonably probable that the stain-test evidence affected the outcome. (Watson, supra, 46 Cal.2d at p. 836.)
D. Statistical use of stain-typing evidence.
Defendant suggests that, aside from the reliability of the stain-test results, it was error to permit testimony that the tests placed him within a *536relatively small percentage of the black population (1.2 percent according to Springer) who could have deposited the stains. He urges that admission of statistics suggesting a mathematical “probability of guilt” misleads the jury and deprives an accused of his right to be acquitted upon a reasonable doubt. (See People v. Collins (1968) 68 Cal.2d 319, 329-330 [66 Cal.Rptr. 497, 438 P.2d 33, 36 A.L.R.3d 1176].)
Since we have already concluded that the test results themselves were inadmissible on this record, we need not address the merits of defendant’s claim.6 We must simply determine whether the statistical use of the test results rendered prejudicial the otherwise harmless error in their admission. We conclude it did not. As we have indicated, the combination of valid eyewitness testimony and circumstantial evidence leaves little doubt that defendant is Susan J.’s killer. There is no substantial possibility that the statistical evidence altered the jury’s verdicts.
No other claims of error are raised at the guilt and special circumstance trial, and our investigation of the record discloses none. We therefore affirm the convictions and the special circumstance finding.
IV. Penalty Issues
A. Antisympathy instruction and argument.
At the penalty phase, the jury was instructed in the words of CALJIC No. 1.00 that it “must not be swayed by mere sentiment, conjecture, sympathy, passion, prejudice, public opinion or public feeling.” The prosecutor made similar arguments, both during the voir dire of jurors and at the close of the penalty case. Defendant contends that these admonishments not to consider sympathy were error which invalidate the penalty judgment.
Defendant is correct. Because of the individualized sentencing concerns inherent in the Eighth Amendment, “federal constitutional law forbids an instruction which denies a capital defendant the right to have the jury consider any ‘sympathy factor’ raised by the evidence when determining the appropriate penalty. ...” Hence, it is error to give an antisympathy instruction at the penalty phase of a capital trial. (Lanphear, supra, 36 Cal.3d at p. 165; Easley, supra, 34 Cal.3d at p. 876.)
Our prior cases also reject the People’s argument that such an instruction is vitiated by the standard admonition (CALJIC No. 8.84.1) to consider *537“any . . . circumstance which extenuates the gravity of the crime, even though it is not a legal excuse for the crime.” (Italics added; see also § 190.3, subd. (k).) As we suggested in Easley, supra, 34 Cal.3d at page 878, footnote 10, and Lanphear, supra, 36 Cal.3d at pages 165-168, this ambiguous standard instruction, at least when combined with an antisympathy warning, is calculated to divert the jury from its constitutional duty to consider “any [sympathetic] aspect of the defendant’s character or record,” whether or not related to the offense for which he is on trial, in deciding the appropriate penalty. (See Eddings v. Oklahoma (1982) 455 U.S. 104, 113-115 [71 L.Ed.2d 1, 10-11, 102 S.Ct. 869]; Lockett v. Ohio (1978) 438 U.S. 586, 604 [57 L.Ed.2d 973, 989, 98 S.Ct. 2954]; Woodson v. North Carolina (1976) 428 U.S. 280, 304 [49 L.Ed.2d 944, 961, 96 S.Ct. 2978].)
Nor was the error harmless here. Defendant presented testimony from numerous lay witnesses and a psychologist, suggesting that he possessed a gentle and nonviolent nature disturbed only by severe psychosexual problems resulting from a difficult childhood. Friends and relatives indicated their affection for him. Defendant himself took the stand to express remorse for the prior rape of Kelly P., revealed at the penalty phase. Counsel emphasized these considerations in closing argument.
Obviously, defendant intended that this character and background evidence, though unrelated to the offense charged, be considered sympathetically by the jury in fixing his sentence. Yet the jury had been told to consider only matters which extenuated the “crime” and to ignore sympathy. As we have previously held, the ambiguous tension between these instructions and defendant’s right to sympathetic consideration of all the character and background evidence he presented requires reversal of the penalty judgment. (Lanphear, supra, 36 Cal.3d at p. 169; Easley, supra, 34 Cal.3d at pp. 878-879; People v. Robertson (1982) 33 Cal.3d 21, 54, 57-59 [188 Cal.Rptr. 77, 655 P.2d 279]; see Eddings, supra, 455 U.S. at p. 119 [71 L.Ed.2d at p. 13] [cone. opn. of O’Connor, J.].)7
*538 B. Constitutional challenge to asserted “mandatory” aspect of 1978 death penalty law.
Under the 1978 death penalty statute, once the defendant stands convicted of a capital crime and the jury has found one or more charged “special circumstances” to be true, the case proceeds to a penalty trial in which the jury must decide between only two possible punishments, death or life imprisonment without possibility of parole. (§§ 190.2, 190.3.) In making that determination, the statute provides that the jury shall “consider and take into account and be guided by” evidence of enumerated “aggravating and mitigating circumstances” introduced at the penalty phase or gleaned from the earlier guilt trial. In these respects the 1978 law is similar to its 1977 predecessor.
In contrast with the 1977 law, however, the 1978 statute declares that if the jury finds that “the aggravating circumstances outweigh the mitigating circumstances” it “shall impose a sentence of death.” (Italics added.)8 Defendant argues that the statutory formula impermissibly restricts the jury’s constitutional sentencing discretion in two related ways. First, defendant notes, section 190.3, subdivision (k) directs the jury to consider, as a mitigating factor, “[a]ny . . . circumstances which extenuate the gravity of the crime even though it is not a legal excuse for the crime,” but it does not expressly state the jury’s further constitutional duty to consider in mitigation all other sympathetic evidence defendant may offer about his character and background, even if it is unconnected to the charged crimes. Second, defendant argues, by its use of the term “outweigh” and the mandatory “shall,” the statute impermissibly confines the jury to a mechanical balancing of aggravating and mitigating factors. Defendant urges that because the statute requires a death judgment if the former “outweigh” the latter under this mechanical formula, the statute strips the jury of its constitutional power to conclude that the totality of constitutionally relevant circumstances does not warrant the death penalty.9
*539If we were to accept defendant’s interpretation of the 1978 law, his constitutional argument would have considerable merit. Under the teachings of the United States Supreme Court, it appears that for a death penalty statute to be valid, the jury’s discretion “must be suitably directed and limited so as to minimize the risk of wholly arbitrary and capricious action” (Zant v. Stephens (1983) 462 U.S. 862, 874 [77 L.Ed.2d 235, 248, 103 S.Ct. 2733], quoting from Gregg v. Georgia (1976) 428 U.S. 153, 189 [49 L.Ed.2d 859, 883, 96 S.Ct. 2909] (plur. opn.)); at the same time, however, the statute must allow for the jury’s “consideration of the character and record of the individual offender and the circumstances of the particular offense.” (Woodson, supra, 428 U.S. at p. 304 [49 L.Ed.2d at p. 961] (plur. opn.); see also Eddings, supra, 455 U.S. at pp. 111-112 [71 L.Ed.2d at pp. 8-9]; Lockett, supra, 438 U.S. at p. 605 [57 L.Ed.2d at p. 990] (plur. opn.); Roberts v. Louisiana (1976) 428 U.S. 325, 333 [49 L.Ed.2d 974, 981, 96 S.Ct. 3001] (plur. opn.).) As the Lockett plurality concluded, a procedure which “prevents the sentencer . . . from giving independent mitigating weight” to all relevant evidence proffered by the defendant for that purpose “creates the risk that the death penalty will be imposed in spite of factors which may call for a less severe penalty. When the choice is between life and death, that risk is unacceptable and incompatible with the commands of the Eighth and Fourteenth Amendments.” (438 U.S. at p. 605 [57 L.Ed.2d at p. 990].)
A capital sentencing scheme can “guide” and “channel” the determination of penalty by strictly confining the class of offenders eligible for the *540death penalty, or by setting forth the factors society deems relevant to a penalty verdict, or both. (See, e.g., Pulley v. Harris (1984) 465 U.S. 37, 53 [79 L.Ed.2d 29, 42, 104 S.Ct. 871]; Zant, supra, 462 U.S. at pp. 875-880 [77 L.Ed.2d at pp. 248-251].) But with respect to the process of selecting from among that class those defendants who will .actually be sentenced to death, “[w]hat is important... is ah individualized determination on the basis of the character of the individual and the circumstances of the crime.” (Zant, supra, 462 U.S. at p. 879 [77 L.Ed.2d at p. 251].) (Italics in original.) It is not simply a finding of fabts which resolves the penalty decision, “ ‘but ... the jury’s moral assessment of those facts as they reflect on whether defendant should be put to death . . . .’” (Easley, supra, 34 Cal.3d 880, quoting People v. Haskett (1982) 30 Cal.3d 841, 863 [180 Cal.Rptr. 640, 640 P.2d 776].) The jury must be free to reject death if it decides on the basis of any constitutionally relevant evidence or observation that it is not the appropriate penalty.10 Moreover, the decision is the responsibility of the jury and no one else. “[I]t is constitutionally impermissible to rest a death sentence on a determination made by a sentencer who has been led to believe that the responsibility for determining the appropriateness of the defendant’s death rests elsewhere.” (Caldwell v. Mississippi (1985) 472 U.S. —, - [86 L.Ed.2d 231, 239, 105 S.Ct. 2633].)
We agree with defendant, therefore, that a statute would be invalid if interpreted to preclude juror consideration of any factors constitutionally relevant to imposition of the death penalty. Nor would a statute pass muster if it required jurors to render a death verdict on the basis of some arithmetical formula, or if it forced them to impose death on any basis other than their own judgment that such a verdict was appropriate under all the facts and circumstances of the individual case.11 We agree with the People, however, that the 1978 death penalty law need not, and should not, be so interpreted.
*541While subdivision (k) of section 190.3 speaks only of a “ ‘circumstance which extenuates the gravity of the crime,’ ” this court in Easley, supra, 34 Cal.3d at page 878, footnote 10, imposed a prospective requirement that “trial courts—in instructing on the factor embodied in section 190.3, subdivision (k)—should inform the jury that it may consider as a mitigating factor ‘any other circumstance which extenuates the gravity of the crime even though it is not a legal excuse for the crime’ and any other ‘aspect of [the] defendant’s character or record . . . that the defendant proffers as a basis for a sentence less than death.’” (Lockett, supra, 438 U.S. at p. 604 [57 L.Ed.2d at p. 990].) In doing so, we necessarily determined that the statutory language is susceptible to that clarification. Indeed, we noted that in Frierson, supra, 25 Cal.3d at p. 178, which dealt with identical language in the 1977 law, we had “apparently viewed [the phrase ‘any other circumstance which extenuates the gravity of the crime’] as an open-ended, catchall provision, allowing the jury’s consideration of any mitigating evidence.” (Easley, supra, 34 Cal.3d at p. 878.) We see no principled basis for holding, so late in the day, that the 1978 law invalidly excludes jury consideration of evidence of this kind.
Similarly, the reference to “weighing” and the use of the word, “shall” in the 1978 law need not be interpreted to limit impermissibly the scope of the jury’s ultimate discretion. In this context, the 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 various factors he is permitted to consider, including factor “k” as we have interpreted it.12 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.13
*542This view is reinforced by examination of general language in the 1978 law. For example, section 190.3 provides that, with narrow exceptions, “evidence may be presented by both the people and the defendant as to any matter relevant to aggravation, mitigation, and sentence including, but not limited to” the circumstances of the current offense, prior felony convictions or violent crimes, “and the defendant’s character, background, history, mental condition and physical condition.” (Italics added.) In deciding whether the aggravating circumstances outweigh the mitigating, the jury must consider, among other things, “all of the evidence” and “the arguments of counsel.”
As Justice Stevens noted in his concurring opinion in Barclay v. Florida (1983) 463 U.S. 939 [77 L.Ed.2d 1134, 103 S.Ct. 3418], rehearing denied, 464 U.S. 874 [78 L.Ed.2d 185, 104 S.Ct. 209], the Florida courts have imposed a similar construction on that state’s somewhat analogous “weighing” statute. Under the scheme applicable to defendant Barclay, once a defendant is convicted of capital murder, a sentencing hearing proceeds before judge and jury at which evidence bearing on statutory aggravating, and all mitigating, circumstances is adduced. The jury then renders an advisory verdict “[w]hether sufficient mitigating circumstances exist . . . which outweigh the aggravating circumstances found to exist; and . . . [b]ased on these considerations, whether the defendant should be sentenced to life [imprisonment] or death.” (Fla. Stat. § 921.141, subd. (2)(b), (c) (1976-1977 Supp.).) The trial judge decides the actual sentence. He may impose death if satisfied in writing “(a) [t]hat sufficient [statutory] aggravating circumstances exist. . . and (b) [t]hat there are insufficient mitigating circumstances ... to outweigh the aggravating circumstances.” (Id., subd. (3).) If the jury recommends life, the facts suggesting a death sentence “should be so clear and convincing that virtually no reasonable person could differ.” (Tedder v. State (Fla. 1975) 322 So.2d 908, 910.)
“Shortly after the enactment of the current statute, the Florida Supreme Court explained: ' “. . . [T]he procedure to be followed by the trial judges and juries is not a mere counting process of X number of aggravating circumstances and Y number of mitigating circumstances, but rather a reasoned judgment as to what factual situations require the imposition of death *543and which can be satisfied by life imprisonment in light of the totality of the circumstances present. . . .” ’ [Citations omitted.]” (Barclay, supra, 463 U.S. at p. 963 [77 L.Ed.2d at p. 1152] [cone. opn. of Stevens, J.], quoting State v. Dixon (Fla. 1973) 283 So.2d 1, 10, cert. den. sub nom. Hunter v. Florida (1974) 416 U.S. 943 [40 L.Ed.2d 295, 94 S.Ct. 1950]; see also Proffitt, supra, 428 U.S. at p. 251 [49 L.Ed.2d at p. 922]; but see Cooper v. State (Fla. 1976) 336 So.2d 1133, 1142, cert. den. (1977) 431 U.S. 925 [53 L.Ed.2d 239, 97 S.Ct. 2200].)14
*544The 1978 California initiative does, of course, represent a change from its 1977 predecessor; the 1978 law tells the jury to decide the appropriate punishment by weighing certain factors, while the 1977 version asked only that the sentencer “consider, take into account and be guided by” the factors listed. But this amendment does not rob the jury of its constitutional responsibility to decide what penalty is appropriate under all the relevant circumstances. It simply makes clear that, in resolving the ultimate issue of punishment under the 1978 law, the jurors are to limit their consideration to “the specific factors listed in the statute, ...” (.People v. Boyd (1985) 38 Cal.3d 762, 773 [215 Cal.Rptr. 1, 700 P.2d 782].)15 Nothing in the amended language limits the jury’s power to apply those factors as it chooses in deciding whether, under all the relevant circumstances, defendant deserves the punishment of death or life without parole.
This construction of the 1978 law honors the plain language of section 190.3. It also explains the most likely "constitutional" intent of the drafters and avoids the constitutional difficulties of a finding that the statute permits "mandatory" death penalties.16 We conclude that the 1978 law is not invalid on grounds that it withdraws constitutionally compelled sentencing discretion from the jury. 17
*545 C. Other penalty phase issues.
Defendant makes three remaining contentions. He urges that certain testimony by the victim’s father at the guilt phase was improperly calculated to arouse jury passions on the issue of penalty. He points to prosecutorial comments that he could take the stand to evoke jury sympathy without discussing his offense.1 ****************18 And he notes that the trial judge denied his automatic motion for modification of the death verdict without providing the required statement of reasons. (§ 190.4, subd. (e).) We deem it unnecessary to discuss these issues, since they are unlikely to recur in any penalty retrial.
V. Conclusion
The judgment as to guilt, and the finding of a special circumstance, are affirmed. The penalty judgment is reversed.19
Broussard, J., Reynoso, J., and Kaus, J.,* concurred.
All statutory references are to the Penal Code unless otherwise indicated.
Shirley expressly held that the Watson standard of prejudice would apply to erroneous admission of testimony by a previously hypnotized witness. (31 Cal.3d at p. 68, citing People v. Kelly (1976) 17 Cal.3d 24, 40 [130 Cal.Rptr. 144, 549 P.2d 1240] [Watson standard applies to admission of scientifically unreliable evidence].) Defendant argues that Shirley error should be evaluated under the more stringent reasonable-doubt standard of Chapman v. California (1967) 386 U.S. 18, 24 [17 L.Ed.2d 705, 710, 87 S.Ct. 824, 24 A.L.R.3d 1065], since it involves denial of the constitutional right to confront witnesses. (U.S. Const., Amend. VI; Cal. Const., art. I, § 15.) A number of recent decisions in other jurisdictions, both before and since Shirley, have suggested that because the false memories created by hypnosis are impervious to cross-examination, admission of hypnotically refreshed testimony may deny the right of confrontation. (E.g., United States v. Valdez (5th Cir. 1984) 722 F.2d 1196, 1202; Peterson v. State (Ind. 1983) 448 N.E.2d 673, 678-679; People v. Gonzales (1982) 415 Mich. 615 [329 N.W.2d 743, 748], mod. (1983) 417 Mich. 968 [336 N.W.2d 751]; State ex rel. Collins v. Superior Court, etc. (1982) 132 Ariz. 180 [644 P.2d 1266, 1273-1275]; State v. Mena (1981) 128 Ariz. 226 [624 P.2d 1274, 1280]; State v. Mack (Minn. 1980) 292 N.W.2d 764, 769.) At least one court has explicitly applied the reasonable-doubt standard of prejudice, apparently on the ground of constitutional error. (People v. Nixon (1983) 125 Mich.App. 807 [337 N.W.2d 33, 34-35].) Extended discussion of the issue is unnecessary, since we are persuaded, for reasons explained in the text, that any Shirley error here was harmless even under the Chapman test.
Prior to the criminalists’ testimony, defendant moved in limine to exclude the stain-test evidence; the court ruled the evidence could come in “if there was a proper foundation.” Defendant objected again after Andrus’ testimony, and the court reserved its ruling until after Springer’s testimony. The defendant objected again at the close of the People’s case; the objection was noted and overruled.
Dr. Grunbaum holds a Ph.D. degree in biochemistry and a master’s degree in criminology with a specialty in forensic identification. He has been employed by the University of California as a research biochemist for 30 years, specializing in analytical biochemistry and microanalysis, which includes the examination of body fluids. (Washington, supra, 622 P.2d at p. 989.) He is a leader in the development of electrophoresis to test body-fluid enzymes for purposes of forensic identification, and is an author of the most-cited studies on the distribution of enzyme and antigen phenotypes in the population. (See Grunbaum et'al., Distribution of Gene Frequencies and Discrimination Probabilities for 22 Human Blood Genetic Systems in Four Racial Groups (1980) 25 J.Forensic Sci. 428; Grunbaum et al., Frequency Distribution and Discrimination Probability of Twelve Protein Genetic Variants in Human Blood as Functions of Race, Sex, and Age (1978) 23 J.Forensic Sci. 577.)
We recognize that defendant’s amicus, Dr. Grunbaum, is himself a well-credentialed forensic scientist who has pioneered in the development of electrophoretic techniques. (See fn. 4, ante.) His record of opposition to their forensic use under current conditions is well-established. In addition to his brief in this case, we judicially notice his testimony in other trials that the visual subjectivity of the electrophoretic process, and the possibility that aging or contamination will cause type conversions and spurious results, render forensic tests for ABO and PGM types in crime-scene stains highly suspect. (E.g., State v. Washington, supra, 622 P.2d 986, 989-992; see also People v. Williams (S.F. Super. Ct., No. 110047) [proceedings of Dec. 8, 1983].) Arguably his public opposition alone is “significant [enough] in . . . expertise” to establish a legitimate scientific dispute. (See Shirley, supra, 31 Cal.3d at p. 56.) However, it is not clear that Dr. Grunbaum’s views are widely shared in degree, and we are reluctant to rely on a single opinion, however respectable, to foreclose all further evidence on the issue of scientific acceptance. (See Kelly, supra, 17 Cal.3d at p. 37.)
We note, however, that both California and the majority of other jurisdictions have traditionally admitted statistical blood-group evidence of this kind in criminal cases, even where it simply includes the accused within the class of possible donors. (See Lindsey, supra, 84 Cal.App.3d at pp. 863-866, and cases cited; Vallez, supra, 80 Cal.App.3d at p. 56; see generally Annot. (1980) 2 A.L.R.4th 500, 511 et seq.)
Defendant also challenges a second feature of CALJIC No. 1.00 as given at his penalty trial—its admonition that the jury must render a “just verdict regardless of what the consequences of such verdict may be." (Italics added.) As defendant notes, this phraseology, like all of CALJIC No. 1.00 (see Easley, supra, at pp. 875-877, and fn. 6), was designed for guilt trials, at which “defendant’s possible punishment is not ... a proper matter for juror consideration . . . .” (People v. Honeycutt (1977) 20 Cal.3d 150, 157, fn. 4 [141 Cal.Rptr. 698, 570 P.2d 1050]; see also People v. Moore (1968) 257 Cal.App.2d 740, 750 [65 Cal.Rptr. 450].) However, at the penalty phase of a capital trial, the “consequences”— the choice between the two most extreme punishments the law exacts—are precisely the issue the jury must decide. In this context, an instruction to ignore “consequences” can be understood by the jury in the same light as an admonition to disregard sympathy. We agree that this portion of CALJIC No. 1.00 should never be given in a capital penalty trial.
The relevant portion of section 190.3 provides: “After having heard and received all of the evidence, and after having heard and considered the arguments of counsel, the trier of fact shall consider, take into account and be guided by the aggravating and mitigating circumstances referred to in this section, and shall impose a sentence of death if the trier of fact concludes that the aggravating circumstances outweigh the mitigating circumstances. If the trier of fact determines that the mitigating circumstances outweigh the aggravating circumstances the trier of fact shall impose a sentence of confinement in state prison for a term of life without the possibility of parole.”
We are mindful of the principles of judicial restraint which caution against premature consideration of constitutional issues, especially in the uncertain constitutional waters which surround the death penalty. (See People v. Frierson (1979) 25 Cal.3d 142, at pp. 185-195 [158 Cal.Rptr. 281, 599 P.2d 587] (cone, and dis. opn. by Mosk, J.); id., at pp. 196-199 (cone, and dis. opn. by Bird, C. J.); id., at p. 199 (cone, and dis. opn. by Tobriner, J.); see also People v. Green (1980) 27 Cal.3d 1, 49-50 [164 Cal.Rptr. 1, 609 P.2d 468]). When this court decided Frierson, however, the 1977 death penalty law had been in effect for only *539two years and had already been superseded by the 1978 law. Except for retrials, most defendants who could be tried under the 1977 law had already been tried. By contrast, the 1978 law has been on the books for seven years, and it remains the law of this state. Automatic appeals of death judgments rendered under its provisions are reaching us at the rate of approximately 30 per year.
If this court were of the view that the law is unconstitutional on its face, as defendant contends, it would certainly be our obligation to say so. While we are not of that view, we do find a potential for confusion in the law which calls for certain prophylactic instructions in future death penalty trials, including any retrial which defendant may confront. {Infra, fn. 17.) Under the circumstances, it is not appropriate for this court to withhold guidance simply because defendant’s death penalty judgment is being reversed on other grounds. (See Code Civ. Proc., § 43; People v. Ramos (1984) 37 Cal.3d 136, 150 [207 Cal.Rptr. 800, 689 P.2d 430].) Defendant asserts other constitutional challenges which we do not find it necessary to confront at this time.
With all due respect to the Chief Justice, we do not understand how our decision to confront certain constitutional issues which defendant has raised “insulates” our decision on those issues from United States Supreme Court review. (See cone. & dis. opn. by Bird, C. J., post, at p. 547.) Our decision that the 1978 law is constitutionally valid, and that defendant may therefore be retried under its provisions, is subject to immediate review by the high court. More fundamentally, even if defendant chooses not to seek review at this time, or if the high court denies review-, defendant is no worse off than if we refrained from deciding those issues. If he is sentenced to death upon retrial, his judgment of conviction will come to this court automatically for review; and if we affirm his conviction, notwithstanding his constitutional objections, he will be free to seek review from the United States Supreme Court in precisely the same manner as if we had deferred consideration of the constitutional challenge until that time.
““Relevant” circumstances are those which the Constitution requires to be cqnsidered in reaching a penalty decision, or which are constitutionally permissible and must or may be considered under the terms of the death penalty statute. Sentencers’ racial, religious, and political prejudices are examples of constitutionally impermissible considerations.
In Jurek v. Texas (1976) 428 U.S. 262 [49 L.Ed.2d 929, 96 S.Ct. 2950], the court upheld a statute which imposed the death penalty for capital murder once the jury found beyond a reasonable doubt that the answers to three statutory questions were “yes.” The questions were (1) whether the victim died as a result of deliberate conduct committed with the reasonable expectation that it would cause death; (2) whether there was a probability of future criminal violence by the defendant, and (3) if relevant, whether the killing was an unreasonable response to any provocation by the victim. The court’s decision relied solely on Texas appellate decisions which interpreted question (2) to encompass consideration by the jury of all relevant mitigating evidence, including general character and background. (Pp. 272-274 [49 L.Ed.2d 938-939].) Implicit in Jurek is the notion that the jury could act on such evidence, answer question (2) “no,” and thus reject the death penalty if it believed under all the circumstances that death was not the appropriate punishment for the particular offense and offender. (See discussion, post.)
The mere fact that a statute requires the sentencer to “weigh” aggravating against mitigating circumstances, or to determine which “outweigh” the others, does not render the law invalid. (See Proffitt v. Florida (1976) 428 U.S. 242, 251-253 [49 L.Ed.2d 913, 922-923, 96 S.Ct. 2960], rehg. den. sub nom. Gregg v. Georgia, 429 U.S. 875 [50 L.Ed.2d 158, 97 S.Ct. 197, 97 S.Ct. 198].)
“Aggravating” and “mitigating” are not defined by the statute. However, we see no statutory intent to require death if the jury merely finds more bad than good about the defendant and to permit life without parole only if it finds more good than bad. At a capital penalty trial, defendant has already been convicted of committing, without legal excuse, an intentional first degree murder with at least one “special circumstance” necessary to make *542him eligible for the death penalty. Often a person in this situation will have a substantial history of criminal and antisocial behavior. It would be rare indeed to find mitigating evidence which could redeem such an offender or excuse his conduct in the abstract. Recognizing this, the statute requires at a minimum that he suffer the penalty of life imprisonment without parole. It permits the jury to decide only whether he should instead incur the law’s single more severe penalty—extinction of life itself. (§ 190.3.) It follows that the weighing of aggravating and mitigating circumstances must occur within the context of those two punishments; the balance is not between good and bad but between life and death. Therefore, to return a death judgment, the jury must be persuaded that the “bad” evidence is so substantial in comparison with the “good” that it warrants death instead of life without parole.
At least one other state supreme court has interpreted a somewhat similar statute to clarify the jury’s fundamental sentencing discretion. North Carolina’s death penalty law requires the sentencer to find (1) whether any statutory aggravating circumstances exist, (2) whether they are “sufficiently substantial” to call for the death penalty, and (3) whether any one or more mitigating circumstances “outweigh” the aggravating circumstances. “Based on these considerations,” the jury must then recommend whether the defendant should receive death or life imprisonment. (N.C. Gen.Stat. (Cum.Supp. 1981) § 15A-2000, subds. (b), (c).) In State v. McDougall (1983) 308 N.C. 1 [301 S.E.2d 308], the North Carolina Supreme Court affirmed that, if questions (1) and (2) were answered “yes” and question (3) was answered “no,” the jury had a duty to recommend death. (Pp. 323-324; see also State v. Pinch (1982) 306 N.C. 1 [292 S.E.2d 203, 226-227], cert, den., 459 U.S. 1056 [74 L.Ed.2d 622, 103 S.Ct. 474].) However, the court indicated, in answering the statutory questions, the jury “must be satisfied that the sentence is justified and appropriate upon considering the totality of the aggravating circumstances with the totality of the mitigating circumstances . . . . [f] The jury is not required to assign a value to the aggravating circumstances, subtract from it the value of the mitigating circumstances, and look to the remainder to determine if that value is sufficiently substantial to deserve the death penalty. We reject and disapprove such a mechanical mathematical approach to the decision of life or death. ” (301 S.E.2d at p. 326.)
McDougall found “instructive” the Utah Supreme Court’s analysis in State v. Wood (Utah 1982) 648 P.2d 71. Utah’s statute provided only that the sentencer should “consider” the proper penalty in light of specific aggravating and mitigating circumstances, but Wood deemed it “implicit in the statutory scheme that a comparison of aggravating and mitigating factors must be made and a decision reached on the result of the comparison. ...” (P. 79.) In conducting the weighing process, the Wood court said, “the sentencing body [must] compare the totality of the mitigating against the totality of the aggravating factors, not in terms of the relative numbers of the aggravating and the mitigating factors, but in terms of their respective substantiality and persuasiveness. Basically, what the sentencing authority must decide is how compelling or persuasive the totality of the mitigating factors are when compared against the totality of the aggravating factors. ...” (P. 83, quoted in McDougall, supra, 301 S.E.2d at p. 327; see also Pinch v. North Carolina (1982) 459 U.S. 1056 [74 L.Ed.2d 622, 103 S.Ct. 474] [opn. of Stevens, J. on denial of cert.].)
McDougall upheld the death judgment there at issue against claims that the jury had not been adequately instructed about the scope of its sentencing discretion. As the court noted, the jury was told, among other things, that “you are the sole judges of the weight to be given to any individual circumstance which you find, whether aggravating or mitigating. Your weighing should not consist of merely adding up the number of aggravating circumstances and mitigating circumstances. Rather, you must decide from all the evidence what value to give to each circumstance and then weigh the aggravating circumstances, so valued, against the mitigating circumstances, so valued, and finally determine whether the aggravating circumstances outweigh the mitigating circumstances.” Jurors were also instructed that “you may consider any circumstance from the evidence which you are satisfied lessens the seriousness of the murder or suggests a lesser penalty than otherwise may be required, such as the defendant’s character, education, environment, habits, mentality, propensities and record, and any other circumstances arising from the evidence which you deem to have mitigating value. ...” (301 S.E.2d at pp. 324-325.) These instructions were neither directly derived from, nor expressly required by, the statutory language.
As we explained in Boyd, supra, the drafters of the 1978 initiative may have believed the 1977 law was unconstitutional if, by requiring the jury only to “consider” the factors listed in the statute, it implied that the sentencer actually was free to determine the penalty on any basis it chose. Such fears, plausible in 1978, were later laid to rest by the United States Supreme Court. (Boyd, supra, at pp. 773-774, fn. 5; see Zant, supra, 462 U.S. at p. 875 [77 L.Ed.2d at p. 248].)
In Easley, this court ruled that it was prejudicial error to give the 1978, or “mandatory death penalty,” version of CALJIC No. 8.84.2 in a case properly tried under the 1977 law, since a defendant “is . . . generally worse off under [the mandatory feature of] the 1978 law.” (34 Cal.3d at pp. 883-884, italics added.) Our discussion assumed that the 1978 version of section 190.3 intended to require the death penalty in certain cases. No extensive analysis was provided, however, and the statutory interpretation was not necessary to our decision. Certainly the 1978 instruction given in Easley was prejudicial when compared to its 1977 counterpart, since the latter, unlike the former, contained no unexplained use of mandatory language. (See discussion, post.) Easley itself recognized that, even if statutory language was susceptible to a liberal saving construction, instructions given in the literal statutory language might nonetheless be deficient. (P. 878, and fns. 8, 10; see discussion, ante.) Nothing in Easley precludes us from holding that the 1978 statute permits the jury to reject death if persuaded by any evidence that it is an inappropriate penalty.
We acknowledge 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. Indeed, such confusion is occasionally reflected in records before this court. For that reason, trial courts in future death penalty trials&emdash;in addition to the instruction called for by Easley, supra, 34 Cal.3d at page 878, footnote 10&emdash;should instruct the jury as to the scope of its discretion and responsibility in accordance with the principles set forth in this opinion. We pass no judgment here upon the validity of death penalty verdicts previously rendered without benefit of the Easley instruction or the instruction we now require. Each 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.
Defendant testified at the penalty phase, expressing remorse for the prior rape and asking the jury for mercy, under a prior trial court ruling that he could do so without exposing himself to examination on the circumstances of the instant crimes.
The parties and other persons have asked us to adopt a specific jury instruction to guide the jury’s determination of the penalty in future capital cases. We have been advised that the Committee on Standard Jury Instructions, Criminal has drafted a proposed intruction to be inserted into CALJIC No. 8.84.2. It directs the jurors that:
“The weighing of aggravating and mitigating circumstances does not mean a mere mechanical weighing of factors on each side of an imaginary scale, or the arbitrary assignment of weights to any of them. You are free to assign whatever moral or sympathetic value you deem appropriate to each and all of the various factors you are permitted to consider. In weighing the various circumstances you simply determine under the relevant evidence which penalty is justified and appropriate by considering the totality of the aggravating circumstances with the totality of the mitigating circumstances. To return a judgment of death, each of you must be persuaded that the aggravating evidence [circumstances] is so substantial in comparison with the mitigating circumstances that it warrants death instead of life without parole.”
We do not adopt the exact language of this instruction, which, in any case, is subject to revision by CALJIC before it is finally adopted. By way of guidance to the trial courts, however, we believe it appropriate to. state that this language, if inserted to replace the language in the current third paragraph of CALJIC No. 8.84.2 (which says that if aggravating circumstances outweigh mitigating circumstances the jurors “shall” return a verdict of death), would conform to our opinion.
Retired Associate Justice of the Supreme Court sitting under assignment by the Chairperson of the Judicial Council.