The PEOPLE, Plaintiff and Respondent, v. Frank SMITH, Defendant and Appellant.
Frank Smith appeals his conviction for multiple offenses against two women on two separate occasions 1 which resulted in sending him to prison for 52 years. He contends the court committed prejudicial error when it admitted the deoxyribonucleic acid (DNA) analysis of his blood and its comparison to the semen swabs obtained from the rape victims. Smith also argues the court erred when it denied his motion to sever the two sets of offenses, and when it denied his motion for mistrial for prosecutorial misconduct. Neither of these arguments has merit. We affirm.2
In May 1989, Robbin E. drove home after socializing with some friends. She parked in front of her house and began reading her mail in the car by the light of nearby street lamps. Suddenly, she saw a man approach her, brandishing a knife. He forced his way into her car and drove her to an uninhabited area. He then committed a series of sex offenses against her, threatening her with the knife. After raping her, he drove her back to her home and ran off. Robbin was taken to a hospital where she underwent a rape examination which included seminal swabs. Robbin had perfect vision, was completely sober and able to give a very clear description of her attacker. She identified Smith as her attacker from a photographic lineup and expressed no doubt that Smith was her attacker at trial.
About five months after the attack on Robbin, Orba S. was curled up in her robe, watching a movie on television alone in her living room. Her children were asleep in their bedrooms when suddenly, Orba saw a figure run down the side of her house. She immediately ran to the phone to call 911, but the electric power had been cut and she could not see the buttons on the phone to call for help. A second later the rear sliding glass door opened, and a man rushed towards her, waving a knife at her. The man met the general description that Robbin had given earlier, except that Robbin's suspect had a beard while Orba's attacker sported only a mustache.
After grabbing Orba, the man held a knife to her and committed a series of sex offenses against her. After subjecting her to his cruelties, he turned the power back on and left. Only then was Orba able to call for help.
When the police arrived she described the man who assaulted her. Orba told them that because the house was dark and the man had warned her not to look at him, she could not give a more detailed description. She gave the police her robe and was then taken to the hospital where she underwent a rape examination from which the doctor obtained vaginal swabs.
Sixteen days after the assault, the police showed Orba over 100 photographs, none of which she identified as her assailant. However, two months later, Orba chose one photograph out of a lineup, stating that the man in the photo looked like the man who raped her. At Smith's trial, Orba testified that although she was not “a hundred percent positive” that Smith was her assailant, he looked very similar and had all the same physical features.
Robert Keister, senior criminalist at the Orange County (OC) Sheriff's DNA crime laboratory (lab), compared the vaginal swab and aspirate taken from Orba's rape kit with the same specimens from Robbin's rape examination using the restriction fragment length polymorphism (RFLP) technique and concluded the bands were numerically indistinguishable: Both women were raped by the same man. These results were also compared with samples taken from Smith. Keister found visual and numerical matches. Moreover, Keister calculated the probability of a random match to be one out of 1.3 million unrelated individuals, even though he used only three sites for his comparison and frequency calculation.3 He used the so-called product rule in this probability calculation, but by eliminating the one probe, it resulted in a frequency calculation more advantageous to Smith than if the fourth point was included. The product rule was described as a formula used to determine the statistical probability of an event by multiplying together the frequency with which each of its component events occur.
The prosecution also called a world-renowned expert in human population genetics, Dr. Bruce Kovacs, who testified as to the concept and technique of DNA RFLP comparison,4 and then testified as to the specific procedure and analysis in this case. He concluded the elimination of the D2S44 probe was unnecessary—as the particular binning method used by the lab obviated any undue affect caused by the excess—and resulted in an analysis unduly weighted in favor of the suspect: He opined that the actual chance of a random match at the four loci was much, much greater than the one in 1.3 million Keister calculated. He also testified at length concerning the discussion about the application of the product rule in probability calculations, noting there was no empirical evidence that population substructuring has any genetic affect on the VNTR 5 markers; any theoretical impact was well covered by the binning process by which the OC databank of DNA samples was categorized. He testified that he had read and considered the National Research Council's (NRC) report and the article in Science by Hartl and Lewontin which proposed that intermarriage might alter gene frequencies in subpopulations, affecting proper application of the product rule. However, he disagreed with their conclusions because they were premised on 20–year–old protein data from esterase genes. By basing their conclusions on blood groups and protein markers rather than VNTR markers, extension of their conclusions to DNA forensic analysis was mere theoretical speculation in his view. His considered opinion was that population substructuring was “irrelevant․ it's not a very pragmatic issue. It's more of a theoretical issue․ [W]e may want to study the frequency of genes in fruit flies, ․ or in yeasts. [¶] Now we all know that yeasts don't make their offspring the same way that people make their offspring, so there are certain other principles that are applied in determining what has significance in a yeast population, that is to say the number of yeasts that would occur in their little world in this cup versus human․ [¶] Human population genetics would have to do specifically with [:] we don't care what the yeasts are doing today, we only care what people are doing․ [¶] As I said, flies don't care. They will have offspring with their sisters or their mothers. They don't care. People don't do that.”
He then criticized the formula proposed by the NRC report. Actually, there were two formulas it recommended: One called the ceiling principle (to be used only after a series of studies had been completed) and another called the modified ceiling principle.6 Dr. Kovacs labeled it as nonscience: “[T]he ceiling principle says, well, I don't care, we're still going to stay with one in ten. You can't say here is science and now we're going to throw that in the trash can and use witchcraft. [¶] So the ceiling principle as enunciated in the NRC Report is something which is probably the most absolutely conservative that could be short of saying that none of it works at all and it's all just baloney [sic ]. But, in fact, it goes in the face of evidence which we have to suggest that use of the ceiling principle will push numbers down into ranges which are less than what they really should be. So we will actually be lying.” (Italics added.) His expert opinion was that the product rule was the only appropriate formula to use in human genetic probability calculations, and the only one which had ever been used by human population geneticists.
Mary Hong, senior criminalist in the serology section of the OC lab, analyzed the blood samples from Smith and the crime scenes for their electrophoretic blood similarities. The results of the ABO bloodtyping and PGM tests could not eliminate Smith as a suspect in either of the rapes. Robbin was an ABO type A secreter with a PGM type of two-plus, one-plus. Her attacker was either a B secreter or an AB secreter. Smith was an ABO type B secreter also with a PGM type of two-plus, one-plus. Orba S. was an ABO type O secreter with a PGM type of one-plus, and her assailant was either an ABO type B or AB with a PGM two-plus or two-plus, one-plus. Again, because Smith was a B secreter with a PGM of two-plus, one-plus he remained a viable suspect in this case as well.
The defense presented Dr. Laurence Mueller to respond to the prosecution's DNA evidence. His testimony concentrated on the use of the product rule as improper in the probability calculations for DNA RFLP. He testified that his “feeling is that the prevailing view among the scientific community who have seriously considered this issue now is, in fact, not to use the product rule and I take as primary evidence of that view point the NRC report which is itself an attempt to get the pulse of the field, of the scientific community and reflect that in their recommendations.” (Emphasis added.) He then testified the appropriate method with which to calculate frequencies is a combination of the “counting” method and the modified ceiling principle. He also rejected the ceiling principle proposed by that same NRC report because, in his opinion, it was not conservative enough until 15 to 20 populations are sampled for substructuring, which at that time, had not occurred. However, he calculated the frequency of Smith's four loci having a random match to be one in 49,200 if the modified ceiling principle was applied; he preferred his counting method which looked only at the samples actually observed and concluded the “probability” was only one in 560.7
IIIAdmission of the DNA Evidence
At the time of Smith's trial, the rule enumerated in People v. Axell, supra, 235 Cal.App.3d 836, 1 Cal.Rptr.2d 411 was binding on the court. It held that DNA RFLP evidence had developed sufficient acceptance within the relevant scientific community for admission, and this included the probability calculation which applied the product rule. (Id., at pp. 860, 868, 1 Cal.Rptr.2d 411.) Thus, the trial court limited the foundational hearing to evidence that would show the experts failed to follow the accepted procedures laid out in Axell; it refused to hold a Kelly–Frye 8 hearing to determine if something had occurred within the scientific community tending to show DNA RFLP evidence had lost acceptance. (Cf. People v. Leahy (1994) 8 Cal.4th 587, 595, 34 Cal.Rptr.2d 663, 882 P.2d 321 [“․ once a trial court has admitted evidence derived from a new technique and the decision is affirmed on appeal in a published opinion, it will become precedent controlling subsequent trials.”].) Specifically, the court relied on Axell (see People v. Axell, supra, 235 Cal.App.3d at pp. 866–868, 1 Cal.Rptr.2d 411) and the transcripts of the expert testimony in the earlier People v. Soto 9 case (see 39 Cal.App.4th 757, 35 Cal.Rptr.2d 846, [rev. gr. 3/16/95] ) to conclude that any variation in the probability calculations proffered by the different experts merely went to the weight of their respective opinions, not to the admissibility of the evidence altogether. (Cf. People v. Cooper (1991) 53 Cal.3d 771, 814, 281 Cal.Rptr. 90, 809 P.2d 865 [lack of accuracy or professionalism in testing or technique is matter for jury which assesses weight; Kelly–Frye is not involved].)
The foundational hearing also included testimony regarding what, if anything, had occurred within the scientific community regarding the application of the product rule. Smith contended a controversy had developed over such frequency calculations, rendering the DNA evidence admitted against him unduly prejudicial. Following the hearing, the trial court decided the OC lab followed all the requisite procedures and qualifications in its analyses and comparisons in this case, and there was no evidence of possible bias to discredit its conclusion. As to the statistical calculations, it ruled that any discussion on the issue was properly a matter of weight for the jury and was irrelevant for the admissibility issue.
Smith now contends that the relevant science had changed considerably, rendering the Axell opinion obsolete. Specifically, he points to the much-discussed article in Science magazine by Drs. Hartl and Lewontin in which they proposed that intermarriage might have a genetic impact on VNTR markers, thereby rendering supposedly random VNTRs not random at all. Without true independence of loci based on truly random selection of samples, application of the product rule in the probability calculation would be wrong. The database must be in—or “approaching”—Hardy–Weinberg equilibrium 10 and linkage equilibrium 11 for a classic probability calculation to be possible.
Concomitant with the publication of Hartl and Lewontin's paper was a retort by Drs. Chakraborty and Kidd in which they attacked the critical ideas espoused in the article as without empirical basis; the material which Hartl and Lewontin used in drawing their initial idea came from 20–year–old protein studies, and the only VNTR material was highly questioned by the responding authors due to its age and contents. Soon after, the NRC published their report concerning the dispute and recommending—for forensic use only—a different formula for probability calculations which would be geared so completely in favor of the suspect, that even if there were a genetic impact from intermarriage (called population substructuring by Hartl and Lewontin), it would not affect the frequency ratio presented in the courtroom.12 Based on the evidence, the court ruled that the relevant scientific community—i.e., human population geneticists—still used the product rule in probability calculations for DNA RFLP work by a clear majority of its constituents.
Smith contends the court failed to exclude the entire DNA evidence (or, in the alternative, at least the prosecution experts' testimony applying the product rule in forming their probability calculation) as without meeting the foundational requirements under the Kelly test.13 Relying on such cases as People v. Axell, supra, 235 Cal.App.3d 836, 1 Cal.Rptr.2d 411, People v. Barney, supra, 8 Cal.App.4th 798, 10 Cal.Rptr.2d 731, People v. Pizarro, supra, 10 Cal.App.4th 57, 12 Cal.Rptr.2d 436, and People v. Wallace (1993) 14 Cal.App.4th 651, 17 Cal.Rptr.2d 721, he argues the lower court was without any legal basis for its conclusion. We disagree, particularly in light of the dynamic changes which have rocked both the scientific world as well as the legal world in the last two to three years.14 Even the NRC report and Dr. Lander in the recent Nature magazine article 15 recommended that the jury receive both types of probability calculations: The calculation using the product rule as the scientists use in their laboratory work and the alternative calculation using the ceiling or modified ceiling method. However, we emphasize the authors of those alternative methods actually concede that the nature of their proposal is rejected by the scientific community as nonscience; the ceiling and modified ceiling method are intended solely for courtroom use. Essentially, they offered a “compromise” for the information to be forensically acceptable, but their compromise was based on what they concluded the judicial world would find acceptable.16 As Kovacs testified: “You can't say here is science and now we're going to throw that in the trash can and use witchcraft․ [U]se of the ceiling principle will push numbers down into ranges which are less than what they really should be. So we will actually be lying.”
More importantly, the publication of the Federal Bureau of Investigation's five-volume study of worldwide VNTR data in late 1993 eliminated the need for the proposed compromise formulas: the Hartl–Lewontin concerns rested on the assumption that intermarriage had a genetic effect on VNTR markers. The worldwide study proved the opposite. (U.S. Dept. Justice, FBI Rep., VNTR Population Data: A Worldwide Study (1993), pp. 1–5.) Thus, the rationale for the ceiling principle and its corollary was undermined.
It is undisputed that the scientific community of human population geneticists all agree that in their own work, they use the product rule in extrapolating the probability of a random match. Because a few scientists, based on nonVNTR studies, feel such calculations may in the future be shown to be inaccurate,17 they advocate use of a separate calculation only for courtroom testimony, one different than that used in their own scientific work.
What we have then is a special rule devised by a committee of scientists—whose selection prerequisites are unknown—for use in the courtroom but not in the laboratory. For such a method to pass muster under Leahy,18 we would have to say that the standard governing admissibility is not that which is generally accepted within the scientific community but that which is accepted by a scientific committee for use in the courtroom, regardless of whether it is generally accepted as “science.” And that is not the law. The evidence was properly admitted.
The judgment is affirmed.
1. Smith's first set of crimes included the kidnapping (Pen.Code, § 207, subd. (a)), sexual battery by restraint (Pen.Code, § 243.4, subd. (a)), forcible rape (Pen.Code, § 261, subd. (2)), and two forcible penetrations with a foreign object (Pen.Code, § 289, subd. (a)) of Robbin E. The second set of offenses consisted of the residential burglary (Pen.Code, §§ 459, 460.1), sexual battery by restraint (Pen.Code, § 243.4, subd. (a)), forcible rape (Pen.Code, § 261, subd. (2)), forcible penetration with a foreign object (Pen.Code, § 289, subd. (a)), and the forcible oral copulation (Pen.Code, § 288a, subd. (c)) of Orba S. The court also found Smith used a knife during the commission of all counts excepting the burglary.
2. Smith also raised an issue which we can summarily reject. He contends the standard instruction defining reasonable doubt contained in CALJIC No. 2.90 was prejudicially ambiguous. That argument was recently rejected in Victor v. Nebraska (1994) 511 U.S. 1, 114 S.Ct. 1239, 127 L.Ed.2d 583.
3. Four sites were actually preselected and compared, but one site—named the D2S44 locus—was found to have an excess of homozygote (single band) patterns at that site in the Orange County black database, so it was eliminated as a point of comparison.
4. This relatively new procedure has been repeatedly described in previous decisions (see People v. Axell (1991) 235 Cal.App.3d 836, 846–847, 1 Cal.Rptr.2d 411; People v. Barney (1992) 8 Cal.App.4th 798, 805–809, 10 Cal.Rptr.2d 731; People v. Pizarro (1992) 10 Cal.App.4th 57, 69–70, 12 Cal.Rptr.2d 436), and we need not repeat it. Essentially, four sites along four separate chromosomes are selected from both the sample found at the crime scene and the suspect's. Because human chromosomes occur in pairs, the four sites actually represent eight different genetic markers. Each of the eight markers analyzed from the crime scene can be compared to the eight markers isolated from the suspect's chromosomes at the same genetic positions, known as loci. The loci are preselected to exclude any DNA segment that controls development of aspects of the body commonly shared by every human, such as arms, legs, head, brain, etc. But there are parts of the DNA strand which are truly unique to each individual. These are sometimes referred to as “junk DNA,” as their purpose, if they have a purpose, is unknown; they appear to be “nonfunctional genetic nonsense.” (See Sagan & Druyan, Shadows of Forgotten Ancestors (1992) pp. 82–83, 125.)
5. VNTR is the abbreviation for “variable number of tandem repeats,” a description of those sites selected for comparison because of the ease with which they are located; when the same pattern of two or three of the four possible bases is repeated a multitude of times forming a long chain, it is easily detected.
6. The ceiling principle essentially compares the frequency at which a VNTR occurs in the various racial subgroups and artificially selects the highest frequency, or five percent, whichever was higher. Thus, if the frequency was less than five percent of the VNTRs in any subgroup, the frequency factor automatically became five percent, irrespective of the actual frequency. The modified ceiling principle, recommended for use until these racial subgroups were studied, used the same approach but used ten percent as the standard, instead of five percent.
7. Dr. Kovacs placed little weight on Mueller's use of the “counting method”: “The counting rule, if you want to call it that, or the counting method is simply an enumeration of how many times something occurred in a given set of observations. They don't—I mean, you can't put them on a scale and say, well, let's see. I mean, it's comparing two very, very different things. One is simply enumeration. [A probability calculation] is what you do with an enumerated value.”
FOOTNOTE. See footnote *, ante.
8. People v. Kelly (1976) 17 Cal.3d 24, 130 Cal.Rptr. 144, 549 P.2d 1240; Frye v. United States (D.C.Cir.1923) 293 F. 1013.
9. The trial in People v. Soto was conducted before the same judge as heard Smith's case. She permitted a Kelly–Frye hearing on the limited issue of whether something had occurred in the scientific community since the publication of the Axell opinion which would render the precedent of acceptance to be currently untenable. At trial, the court concluded there was nothing to undermine the opinion that the relevant scientific community accepted the product rule as an appropriate formula in the probability calculation for DNA RFLP work, notwithstanding the latter comments by Hartl and Lewontin and the NRC. Our opinion in Soto—which is now presently pending before the California Supreme Court (rev. gr. 3/16/95)—affirmed the conviction by agreeing with that conclusion.
10. The actual theory espoused by Professors Hardy and Weinberg almost a century ago predicts the probability of any given occurrence. Here, it predicts a heterozygote pattern at a given locus as “2pq” where p and q are the respective probabilities of each of the 2 bands; it likewise predicts the probability of a true homozygote pattern at a given locus as the square of that probability: “p-squared” or “q-squared.” The mathematical characterization of the rule is p-squared plus 2pq plus q-squared equals 1. (See People v. Castro (1989) 144 Misc.2d 956, 545 N.Y.S.2d 985, 992–993.)
11. Linkage equilibrium reflects a similar proportionality as Hardy–Weinberg, but multiple factors at different loci are involved rather than just two at a single loci.
12. At the conclusion of Smith's trial, People v. Barney, supra, 8 Cal.App.4th 798, 10 Cal.Rptr.2d 731 was published, finding this controversy to have “change[d] the scientific landscape considerably, and demonstrates indisputably that there is no general acceptance of the current process․” (Id. at p. 820, 10 Cal.Rptr.2d 731.) Obviously, the Barney opinion had no effect on the trial court, and whatever effect it carries at this juncture will be discussed later.
13. As recently reiterated by our Supreme Court in People v. Leahy, supra, 8 Cal.4th 587, 34 Cal.Rptr.2d 663, 882 P.2d 321, there are essentially three components which must be met under Kelly: (1) The scientific technique in question must be shown to be reliable; (2) the witness furnishing such testimony must be properly qualified as an expert to render an opinion on the subject; and (3) the proponent of the evidence must demonstrate that correct scientific procedures were used in the particular case. However, to show the technique is reliable, the proponent must have expert testimony that it has gained the general acceptance within the relevant scientific community. (Id. at p. 594, 34 Cal.Rptr.2d 663, 882 P.2d 321.)
14. At present, one appellate opinion in California affirms the admission of DNA evidence, but the probability calculation presented in that case employed the modified ceiling method, and not the product rule. (People v. Taylor (1995) 33 Cal.App.4th 262, 40 Cal.Rptr.2d 132.) As of the beginning of this calendar year, 32 state supreme courts admitted DNA RFLP comparison testimony but with differing records and rationales. (See Perry v. Alabama (1991) 586 So.2d 242; Prater v. Arkansas (1991) 307 Ark. 180, 820 S.W.2d 429; Fishback v. Colorado (1993) 851 P.2d 884; Andrews v. Florida (App.1988) 533 So.2d 841; Johnson v. Georgia (1995) 265 Ga. 668, 461 S.E.2d 209; Hawaii v. Montalbo (1992) 73 Haw. 130, 828 P.2d 1274; Idaho v. Faught (1995) 127 Idaho 873, 908 P.2d 566; Davidson v. Indiana (1991) 580 N.E.2d 238; Iowa v. Brown (1991) 470 N.W.2d 30; Kansas v. Smith (1991) 248 Kan. 217, 807 P.2d 144; Montana v. Weeks (1995) 270 Mont. 63, 891 P.2d 477; Mitchell v. Kentucky (1995) 908 S.W.2d 100; Louisiana v. Charles (1993) 617 So.2d 895; Cobey v. Maryland (1989) 80 Md.App. 31, 559 A.2d 391; Massachusetts v. Lanigan (1994) 419 Mass. 15, 641 N.E.2d 1342; Minnesota v. Bloom (1994) 516 N.W.2d 159; Polk v. Mississippi (1992) 612 So.2d 381; Missouri v. Davis (1991) 814 S.W.2d 593; New Mexico v. Duran (1994) 118 N.M. 303, 881 P.2d 48; New York v. Wesley (1994), 83 N.Y.2d 417, 611 N.Y.S.2d 97, 633 N.E.2d 451; North Carolina v. Pennington (1990), 327 N.C. 89, 393 S.E.2d 847; Ohio v. Pierce (1992) 64 Ohio St.3d 490, 597 N.E.2d 107; Taylor v. Oklahoma (Crim.App.1995) 889 P.2d 319; Oregon v. Herzog (1993) 125 Or.App. 10, 864 P.2d 1362; Pennsylvania v. Crews (1994) 536 Pa. 508, 640 A.2d 395; South Carolina v. Ford (1990) 301 S.C. 485, 392 S.E.2d 781; South Dakota v. Wimberly (1991) 467 N.W.2d 499; Bethune v. Texas (Cr.App.1992) 828 S.W.2d 14; Vermont v. Passino (1994) 161 Vt. 515, 640 A.2d 547; Spencer v. Virginia (1989) 238 Va. 295, 384 S.E.2d 785; West Virginia v. Woodall (1989) 182 W.Va. 15, 385 S.E.2d 253; Springfield v. Wyoming (1993) 860 P.2d 435.) Only four state supreme courts have rendered opinions excluding it. (Washington v. Cauthron (1993) 120 Wash.2d 879, 846 P.2d 502; Arizona v. Bible (1993) 175 Ariz. 549, 858 P.2d 1152; Nebraska v. Houser (1992) 241 Neb. 525, 490 N.W.2d 168; New Hampshire v. Vandebogart (1992) 136 N.H. 365, 616 A.2d 483.) There are also several cases which have been granted review by the California Supreme Court dealing with the propriety of admitting such evidence when the probability calculation was based on the product rule or a combination of both the product rule and the ceiling method (e.g., People v. Burks (1995) 37 Cal.App.4th 652, 43 Cal.Rptr.2d 791 [rev. gr. 11–16–95]; People v. Amundson (1995) 39 Cal.App.4th 468, 41 Cal.Rptr.2d 127 [rev. gr. 8–10–95]; People v. Marlow (1995) 39 Cal.App.4th 343, 41 Cal.Rptr.2d 5 [rev. gr. 7–20–95]; People v. Wilds (1995) 40 Cal.App.4th 166, 37 Cal.Rptr.2d 351 [rev. gr. 3–16–95]; People v. Soto (1994) 39 Cal.App.4th 757, 35 Cal.Rptr.2d 846 [rev. gr. 3–16–95] ), while only one case, which was also granted review, held the evidence's admission was error. (People v. Venegas (1994) 40 Cal.App.4th 128, 36 Cal.Rptr.2d 856 [rev. gr. 3–16–95].)
15. As we discussed in Soto, an article appeared in the science magazine, Nature, in October 1994 which was coauthored by two former adversaries on the subject: Eric Lander and Edward Budowle. As its title reflects, the “DNA Fingerprinting Dispute [Is] Laid to Rest.” Lander, a former opponent to the forensic use of the product rule, discussed how neither his papers nor the NRC report ever intended DNA RFLP evidence to be excluded because of the issues raised by Hartl and Lewontin. He intended to have a jury fully informed on the nature of frequency calculations and the continuing progress in the field of DNA research, but that would require presentation of both the product rule and the ceiling method in any probability calculation. (Lander & Budowle, DNA Fingerprinting Dispute Laid to Rest, Nature (Oct. 27, 1994) p. 736.)
16. Although we do not address the public policy considerations of admitting into evidence formula or datum an ad hoc committee of scientists concludes is appropriate for the courtroom but not the laboratory, they are enormous. The subjective judgments of a scientific committee which come into play in changing scientific data can affect the burden of proof in any number of areas of both the civil and criminal law.
17. Such qualifying language brings to mind the warning issued by Peter Huber in his book, Galileo's Revenge: “The language of ․ possibly, may, might, and maybe that so often litters fringe testimony in court is not the language of science.” (Huber, Galileo's Revenge, p. 210, citing Karl Popper.)
18. We do not attempt to determine whether the ceiling principle would be admissible under the new federal evidentiary rule of Daubert v. Merrell Dow Pharmaceuticals, Inc. (1993) 509 U.S. 579, ––––, 113 S.Ct. 2786, 2794, 125 L.Ed.2d 469.
FOOTNOTE. See footnote *, ante.
SILLS, Presiding Justice.
SONENSHINE and CROSBY, JJ., concur.