AFTER REMAND

The question presented, in the words of the opinion of the Court on an earlier submission of this appeal, is whether "the results of serological electrophoresis [of dried evidentiary bloodstains] have achieved general scientific acceptance for reliability among impartial and disinterested experts” in the scientific community.1 We conclude that the people have not established such general scientific acceptance.

i

Defendant Jeffrey Allen Young was convicted of first-degree murder on evidence that the homicide was committed during the perpetration of a felony. The Court of Appeals affirmed.

*473The felony described in the information was burglary. On the earlier submission of Young’s appeal to this Court, we held that at the time of the homicide2 the term burglary in the section of the Penal Code defining the offense of first-degree murder meant "breaking and entering of a dwelling house in the nighttime with an intent to commit a felony.”3 The people’s evidence tended to show that the breaking and entering occurred in the daytime.4

We further held that "the results of the blood analyses were [not] admissible at trial without a prior showing that the technique of serological electrophoresis enjoys general scientific acceptance among impartial and disinterested experts . . .5

We declined to respond to the question "whether the results of blood analyses are admissible to include an accused within the class of possible perpetrators” (emphasis in original) until "development of a record by the trial court at the hearing which we order to determine if serological electro-phoretic analysis has achieved general scientific acceptance for reliability, Frye v United States, 54 US App DC 46; 293 F 1013 (1923); People v Davis, 343 Mich 348; 72 NW2d 269 (1955), by disinterested and impartial experts, People v Barbara, 400 Mich 352; 255 NW2d 171 (1977); People v Tobey, 401 Mich 141; 257 NW2d 537 (1977).”6

We declined to reduce the conviction to second-degree murder "pending resolution of the remain*474ing issues in this ease following remand,”7 and deferred as "premature” in "light of the hearing which we order” consideration of the prosecutor’s argument that "a Davis-Frye error here, if any, was harmless beyond a reasonable doubt because of the other evidence of defendant’s guilt adduced at trial.”8

We retained jurisdiction.9 The hearing on the admissibility of the bloodstain evidence was held in the circuit court, and the record was transmitted to us. Our order did not require the circuit judge to file findings of fact, and he did not do so. Supplemental briefs were filed, and the cause was reargued.

ii

Evaluating the scientific community’s acceptance of the reliability of electrophoresis of dried evidentiary bloodstains presents some unusual problems. The number of scientists not working for a police agency who are familiar with electrophoresis of evidentiary bloodstains is small. If these scientists alone were considered, the community would be too small for a fair sampling of scientific opinion. There is, however, a larger number of nonforensic scientists using electrophoresis who are capable of evaluating the reliability of electrophoresis of evidentiary bloodstains if presented with the information they need to fill the gaps in their own knowledge and experience. The two groups combined constitute a group of scientists large enough to make a fair determination of whether electrophoresis of evidentiary bloodstains *475is generally accepted by experts in the scientific community.

The prosecution has the burden of establishing this community’s general acceptance of the reliability of electrophoresis of evidentiary bloodstains. In the instant case, there is disagreement within the community on three separate issues: the length of time that genetic markers, particularly erythrocyte acid phosphatase (eap), can be accurately read in dried blood, the reliability of the thin-gel multisystem analysis, and the effects of crime-scene contaminants. The prosecution did not fulfill its burden respecting the last two issues raised by the defense.

The only prosecution witness having substantial experience with electrophoresis of evidentiary bloodstains relied on his own unpublished observations and an unpublished reliability study by the developer of the multisystem to conclude that the thin-gel multisystem analysis was reliable. A defense witness questioned both the reliability of the technique and the study. The other prosecution witnesses were unfamiliar with the thin-gel multi-system, and their conclusion about the reliability of the method was based on the absence of any study showing that it did not work. No independently conducted reliability study supported that conclusion. Another defense witness said that the scientific community would not agree on the reliability of that conclusion without better supporting evidence.

Nor have comprehensive control tests been run with respect to the effects of crime scene contaminants. Prosecution witnesses testified, on the basis of their experience with bloodstains drawn under laboratory conditions, that they can identify bacterial contamination, at least if it is of the type normally encountered. They also claimed that bac*476terial contamination has not affected the reliability of the electrophoretic tests they have conducted. This is, however, the type of self-verification considered inconclusive in the scientific community. The record does not indicate that any work has been done on the effects of soil and chemical contamination on the reliability of electrophoresis.

We conclude that the scientific community’s general acceptance of the reliability of electrophoresis of evidentiary bloodstains has not been established in the instant case. Reliability remains in dispute and unresolved because of the questions unanswered. The questions are not likely to be answered and the reliability of electrophoresis of evidentiary bloodstains established until independently conducted validation studies on the thin-gel multisystem analysis are undertaken and comprehensive control tests evaluating the effects of different contaminants are run, and the results have been subjected to the scrutiny of the scientific community. The evidence produced by electrophoresis should, therefore, not have been admitted.

If it were clear that the erroneous admission of the electrophoresis evidence had not prejudiced the defendant, a new trial would not be required. In the instant case, however, it is not clear the error was harmless. The only witness testifying about electrophoresis at the trial said that this new technique was reliable and that the test results showed that only a small percentage of the population could have had the protein subtype and abo combination found in the blood at the crime scene. It does not clearly appear that the other evidence offered by the prosecution would have dispelled all reasonable doubts in the jurors’ minds.

*477III

Mitchell Lechtanski was last seen alive at 8:00 a.m. on May 16, 1978. At 11:00 A.M., Lechtanski’s housekeeper entered his house and found its contents disturbed. She left immediately and called the police. At 11:30 a.m., police officers entered the house and found Lechtanski’s body. There was a great amount of blood. The scene was secured and assistance summoned.

At 1:30 p.m., a detective found a trail of blood leading from Lechtanski’s driveway to an intersection near a parking lot. Bloodstains were also found on the floor of the rear porch and on a stairway near a window that had apparently been broken to gain entry to the house.

At 3:55 p.m., State Police Crime Laboratory personnel arrived. A laboratory scientist used a damp cloth to obtain samples of dried blood from the sidewalk, stairway, and porch. Detective Stolo-row received the stains at the end of May. The stains were accompanied by a brief description of "the location from which that blood was collected

99

The first electrophoretic testing was conducted on June 5.

Electrophoresis is a physical method for the separation of biologically important proteins through the use of electric current. Proteins are very complex molecules which assume positive, negative, or neutral charges, depending on the solution in which they are placed. When these charged molecules are placed on an appropriate medium and subjected to an electrical field, they will migrate toward the pole of the opposite charge. Blood proteins vary in size, shape, density, and charge; consequently they vary in electropho-retic mobility. Therefore, after electrophoresis, *478they are separated into distinct bands on the supporting medium.[10]

The distinct bands form characteristic patterns that reveal the protein subtypes.

The samples were tested for five different proteins: phosphoglucomutase (pgm), esterase d (esd), glyoxylase 1 (glo), erythrocyte acid phosphatase (eap), and haptoglobin (np). Three of the genetic markers, esd, pgm, and glo, were tested simultaneously using a thin-layer starch gel multisystem analysis. The other two genetic markers were tested individually.

A test can yield no interpretable result, or it can reveal the specific subtype of a protein. Protein testing, like abo blood-type testing, does not give a "yes-or-no” answer, rather, it indicates insufficient information, or it provides the subtype (i.e., for abo: a, b, ab or o; for pgm: pgm, pgm l, pgm 2, or PGM 2-1).

The electrophoresis of the dried bloodstain on the stairway yielded no interpretable results for all five of the genetic markers in question. The results of the test of the sidewalk stain were as follows: no detectable activity for Hp and glo, type b for eap (the first run was reported as inconclusive), type 2-1 for esd, type 2-1 for pgm (the first run was reported as inconclusive). The results of the porch stain were as follows: no detectable activity for glo, type B for eap, type 2 for Hp, type 2-1 for esd (the first run was reported as a questionable type 1 for esd), and type 2-1 for pgm. The tests were not repeated even if there had been inconsistent test readings._

*479The protein subtypes from the sidewalk and porch stains matched those of Young, but not those of Lechtanski.

At trial, Detective Mark Stolorow testified that less than 1.3 percent of the population has the protein subtype and abo combination found in the sidewalk stain and one-half of one percent has the subtype and abo combination found in the porch stain.11

IV

On the first submission of this appeal, this Court said that as "unanimous precedent unequivocally demonstrates, the party offering novel scientific evidence has the burden of demonstrating general scientific acceptance for reliability among impartial and disinterested experts before the evidence may be admitted.”12

The general acceptance rule is designed to "prevent the jury from relying on unproven and ultimately unsound scientific methods.”13 It recognizes that jurors are "scarcely” prepared to evaluate "complicated, scientific testimony concerning the theory and operation of the devices in the face of a difference of scientific opinion as to their accuracy.”14 Leaving the decision to disinterested and *480impartial experts in the scientific community assures that those best qualified to assess the reliability of a scientific method do so.15 Although requiring the technique to "attain sufficient currency and status to gain the general acceptance of the relevant scientific community” inevitably creates some delay in the admission of the type of proof supplied by a new technique, the delay is deemed necessary to assure that the technique is trustworthy.16

v

At the evidentiary hearing to determine the *481reliability of electrophoresis of evidentiary bloodstains, the prosecution presented seven witnesses, and the defense presented two witnesses. The prosecution and defense each presented one forensic scientist having substantial experience with electrophoresis of evidentiary bloodstains. Three of the prosecution’s witnesses and the other defense witness were geneticists, familiar with electrophoresis, but unfamiliar with electrophoresis of eviden-tiary bloodstains. The other three prosecution witnesses were technicians, two of whom were full-time employees of law enforcement agencies. Before analyzing their conclusions it is first necessary to determine whether some or all of them are "disinterested and impartial experts in the particular field.”17

Because a theoretical understanding is essential, the relevant scientific community is scientists not technicians.18 Practical experience with the process, however, is also necessary.19 Ideally the community would be scientists with direct empirical experience with the procedure in question.

Two of the witnesses fit this description. Dr. George Sensabaugh is an associate professor of public health at the University of California at Berkeley and a specialist in forensic science. He has also conducted electrophoretic studies of dried bloodstains. Dr. Benjamin Grunbaum is a retired biochemist from the University of California with a specialty in criminalistics, the science of identification of physical evidence in criminal cases. He *482has been recognized to be "a leader in the development of electrophoresis to test body-fluid enzymes for purposes of forensic identification.”20

Grunbaum and Sensabaugh appear to be a part of a small community of scientists doing work on electrophoresis of evidentiary bloodstains.21 The number of scientists within this community willing to testify seems even smaller. Grunbaum, Sen-sabaugh, and Mark Stolorow, the police detective who did the electrophoresis in the instant case, figure prominently in the few reported cases involving electrophoresis of evidentiary bloodstains.22 Those cases might be described as reflecting and reporting a debate between Stolorow and Grun-baum.

An argument could be made that neither Grun-baum nor Sensabaugh are disinterested and impartial, and should therefore be excluded despite their expertise. Grunbaum was the leader of . the team of scientists that sought to develop a bloodstain analysis system for use in crime laboratories. He brought in Brian Wraxall and Stolorow to work on the project. After expressing dissatisfac*483tion with the multisystem being developed, he withdrew from the project and suggested that it be discontinued. The project continued and when the results were published, he claimed they included misrepresentations. An independent review group found no grounds for Grunbaum’s charges, but the sponsors of the project decided not to publish its results. Arguably, Grunbaum is still seeking to vindicate his original position. Sensabaugh also is not clearly disinterested. He has been a collaborator with Brian Wraxall and a paid consultant with the Oakland Crime Laboratory.23 He has also contributed to a prosecution response to an amicus curiae brief in a case pending before the California Supreme Court.

Nevertheless, a certain degree of "interest” must be tolerated if scientists familiar with the theory and practice of a new technique are to testify at all. The standard developed by this Court is whether the expert’s "livelihood was not intimately connected with the new technique.”24 The *484livelihood of Stolorow and James Kearney, the prosecution witness who directs the fbi serology laboratory, is intimately connected with the new *485technique. The livelihood of Grunbaum and Sensa-baugh is not so intimately connected.25

The community of scientists having direct empirical experience with electrophoresis of eviden-tiary bloodstains does not seem "sufficiently large so that the Frye objective of receiving a consensus judgment of the scientific community can be met.”26 The community of nonforensic scientists using electrophoresis is, however, large enough to obtain an adequate sampling of scientific opinion. These scientists have sufficient theoretical understanding and practical experience to be able to evaluate the evidence. The geneticists testifying for the prosecution and the defense are therefore in the relevant community of scientists having experience with electrophoresis.27 They need only explain the gaps in their own knowledge and experience, and reach general agreement about the reliability of the information they are using to fill these gaps, for their judgment to be respected.

vi

The precise issue in the instant case is whether electrophoresis of evidentiary bloodstains passes *486the general acceptance test. General acceptance of electrophoresis in other areas is not necessarily relevant. The defendant concedes that serological electrophoresis of fresh blood in paternity testing and genetics research is considered generally reliable.28 Electrophoresis of evidentiary bloodstains presents, however, a number of complications, particularly the electrophoresis conducted in the instant case. The complications are the bloodstain is not fresh, it is tested by thin-gel multisystem analysis, and, most importantly, it has possibly been exposed to unknown contaminants.29

A

Electrophoresis for paternity testing and genetics research is generally done on fresh blood. Electrophoresis of evidentiary stains is for the most part done on dried blood. "There is no difference in methodology from fresh blood to dried blood except in preparation of the samples and the extra care that must be exercised in interpretation of results when degraded material is studied.”30 The important difference is that the blood is not fresh, and blood begins to degrade as soon as it *487leaves the body.31 The "crucial question is whether the marker detected in aged blood is a reliable indication of that found in fresh blood from the same person.”32 The dispute centers on the results of the eap test.

The defense raised the possibility of misreadings caused by the deterioration of aged blood. According to Dr. Grunbaum blood degrades very rapidly. In the eap system, the problem is said to be particularly acute, because the banding patterns for the b, c, and cb types are similar. In the instant case, the sidewalk stain was typed b although originally the reading was inconclusive.

Grunbaum suggested that the a band of the ba type could be lost through decay resulting in a ba that looked like a b. Testifying for the prosecution Sensabaugh agreed that a degraded ba might have been typed a b, but he said the change would signal itself. Grunbaum’s response was "if the A degrades faster there can be a . . . time where you will see only the b isoenzyme, and you will not see the A isoenzyme, and you will miscall this as a type b. The explanation given here by Sensabaugh that there is [sic] some guidelines to identify this still as a ba, really do not hold in degraded material, namely the seeing of the secondary B, because in degraded material the clarity of these bands and the resolution is very obscure.”

The question of eap deterioration was addressed in State v Washington, 229 Kan 47, 55; 622 P2d 986 (1981). The witnesses were Grunbaum and Stolorow, testifying for the defense and prosecution, respectively. The court rejected Grunbaum’s *488contention, saying, "Grunbaum used a different medium [and] the State produced two separate pieces of evidence showing the contrary — that eap did not rapidly deteriorate . . . (Emphasis supplied.) A commentator has suggested that the first "assertion is particularly interesting, since Grun-baum’s published work has shown that false eap positives are obtained on both mediums — cellulose acetate membranes and starch gel.”33 The commentator continued, "the court’s counting of evidence may have been wrong.”34 It appears that "it was not two pieces of evidence contradicting Grun-baum that were given, but rather one piece under two guises.”35 The "piece” of evidence was the "Denault” study that showed that eap isoenzymes can be correctly identified in dried laboratory-produced bloodstains for up to thirteen weeks.36

Grunbaum’s criticism that degraded samples cannot be reliably typed also appears to be based on a limited study:

Grunbaum took four different known samples, allowed a portion of each to dry under ambient conditions and heated the remainder for up to 48 h at 37°C. All samples were then subjected to electrophoresis and identified by four skilled technicians. Of the dried samples, two were correctly identified by all four technicians, a third sample was correctly identified by three technicians, with the fourth technician reaching no conclusion as to the isozyme. The fourth sample was correctly identified by one technician. Two were unable to identify it, and the fourth technician misidentified it.

*489Only one of the heated blood samples was correctly identified by all four technicians. One sample was correctly identified twice, misidentified twice, and not identified twice. The last sample was misidentified by three technicians and not identified by one. As the authors pointed out, the treatment of neither the bloodstains nor the heated liquid blood was extreme as compared to real life situations. The technicians were highly trained, yet they misidentified isozymes 7 out of 32 times, for an error rate of 21.8%. Furthermore they correctly identified the isozyme only 53.1% of the time.[37]

Despite the small sample used in Grunbaum’s research, Sensabaugh cited Grunbaum’s work as authority in an article in which he warned, "[t]he 'a’ isozymes are more labile than the 'b’ and 'c’ isozymes; thus in ba heterozygotes the more labile 'a’ isozymes may be lost, giving an electrophoretic pattern looking like 'b’ homozygote. ... It would be good to have a better understanding of how the biochemistry of this marker dictates its behavior.”38

The Denault study, which was also cited in Sensabaugh’s article, was a more thorough study.39

A perusal of the literature further clarifies the dispute in the field. A scientist writing before the Denault study appeared concluded "the [eap] enzyme is not particularly stable in dried bloodstains and hence they have to be not more than 2-3 weeks old for successful typing. Older stains can *490give spurious results.”40 The stain in the instant case was less than three weeks old. Another scientist who did a case study of eap type cb concluded it could be misread as type c. He did suggest, however, that an awareness of the pitfalls associated with degraded eap samples "will help to eliminate possible misinterpretations.”41

Despite these disagreements, if the only question about the reliability of electrophoresis of eviden-tiary bloodstains was the survivability of degraded samples, it would be questionable whether electrophoresis evidence should be excluded where the bloodstain is less than three weeks old. The most detailed independent study discussed by the scientists suggests degraded eap markers can be accurately read up to thirteen weeks. Before this study was written, other scientists believed eap markers could be accurately read up to two to three weeks, which was the length of time involved in the instant case. The main support for the defense’s critique of the studies are test results from too small a sample to carry much weight in the scientific community.

B

The second point of contention is the reliability of the thin-gel multisystem used in the instant case, which simultaneously analyzes three genetic markers, pgm, esd, and glo, on a single, thin-layer starch gel. Although other combination systems exist, the multisystem was designed by police scientists for police work; it allows the maximum amount of information to be drawn from electrophoresis of a small stain._

*491A specific question raised is whether the filter used in the test of the esd molecules has the unintended effect of compromising the analysis of the pgm and glo molecules. Because the results of the glo test were not interpretable in the instant case, the only issue is the reliability of the pgm results in the multisystem analysis.

The defense argues that the thin-gel multisys-tem is unreliable with respect to dried blood because the blood sample is too marginal to begin with to be accurately read after further diffusion. Once the electrophoretic separation has been conducted, a filter paper containing a chemical reagent is placed over the gel. The filter paper is meant to stain the esd molecules, but it also soaks up pgm molecules.42 Grunbaum says this "compromises” the pgm test because "the pgm molecules have diffused sideways, some have disappeared . . . [and] the intensity of the pgm bands are not the same as if they were stained first, before the esd.”43 Grunbaum said he could "deduce from the photographs [taken of the test] that a leaching out of the pgm has occurred and you can see it very well” in the instant case.44 The defense argues that the multisystem "aggravates” the problem inherent in analyzing degraded samples.45_

*492The defense further argues that no independent study verifying the reliability of the thin-gel multisystem has ever been published. No prosecution witness contradicted this argument. The developer of the thin-gel multisystem, Brian Wraxall, did conduct his own blind trials,46 but self-verification is not a sufficiently reliable procedure.47

*493The prosecutor’s response was to present witnesses who have done electrophoresis with other combination tests.48 Dr. Rachael Fisher and Dr. Harvey Mohrenweiser have used combination systems involving a thick-slab starch gel. Grunbaum distinguished the thick- from the thin-gel combination system. "They [those using the thick gel] slice their gel in such a way that they had several layers, like a layer cake, and they had fresh surfaces, and they stained only one for a given sys*494tem. So this was not a compromising system.” Testimony by at least one of the prosecution witnesses suggested there was some overlap on at least one of the combination systems he used. He did not think the overlap compromised the system: "[W]e will use the fluorescent stain for eap . . . [and] put the pgm stain on top . . . [but] because we have gone from a fluorescent to a visible stain . . . the two do not interfere with each other.”49 There was no further testimony by either side to resolve the dispute about the effect of the overlapping tests.

The prosecution also asked the scientists using the other combination tests why they believed the thin-gel multisystem was reliable. Their collective response could be summarized in the following comment by Dr. Rachael Fisher, "I have no reason to suppose it wouldn’t work.” They testified that they had seen no study demonstrating that the multisystem was unreliable. This line of reasoning would be adequate if the burden of establishing general acceptance of unreliability were placed on the defense. The burden of establishing general *495acceptance of reliability is, however, on the prosecution.

In sum, there are substantial unanswered questions respecting the reliability of Wraxall’s thin-gel multisystem. Conflicting expert testimony indicates that until independent verification tests have been conducted regarding the thin-gel multisys-tem, general agreement in the scientific community on the reliability of that multisystem is unlikely. A specific question left unresolved is whether the filter used in the test of the esd molecules compromises the analysis of the pgm molecules.

c

The reliability of blood degraded by dirt, gasoline, urine, sweat, and other possible crime scene contaminants is also at issue in the instant case. Electrophoresis for paternity testing and genetics research is not beset with these problems. The only scientists that have done electrophoresis of blood exposed to these contaminants are those with forensic experience.

Both witnesses for the defense, Grunbaum and Dr. Diane Juricek, testified that it is not possible to determine the reliability of electrophoresis of evidentiary bloodstains until the effects of crime scene contaminants are understood. Juricek said that for electrophoresis of evidentiary bloodstains to be accepted as reliable, scientists would have to study the effects of "common gasoline contaminants which appear on sidewalks, ddt, which can, you know, from spraying grass . . . appear .... There is [sic] also bacterial contamination possibilities. There are molds that could have an effect.” Although Juricek would not say for certain whether the contaminants would affect the electro*496phoresis, she said there was a "very strong theoretical possibility” that they would. Grunbaum testified "[t]here is just no way of knowing the degree of . . . the humidity, . . . heat . . . bacterial . . . [and], chemical contamination, and . . . this is a range that goes on beyond anybody’s imagination.”

Both witnesses testified that the reliability of electrophoresis of evidentiary bloodstains would not be established in the scientific community until controlled studies were conducted taking into account the possible contaminants present at a crime scene. Juricek said "[y]ou would have to check all of these different factors . . . singularly and then in combination . . . .” The studies would then have to be published and "verified independently.”

It appears from the record and a survey of the scientific literature that such comprehensive control studies have not been conducted.50

The prosecution relies instead on inferences drawn from tests performed on dried bloodstains prepared under ideal contamination-free conditions. The only publication referred to by name was the Denault study discussed earlier. Reliance on this study is curious given Denault’s own caveat.

[E]mphasis must be placed on the limitations of this study. It is intended as a starting point for future research. . . . Moreover, the tests were conducted on clean specimens free of impurities. It is realized that in actual practice serological evidence preserved under known and constant conditions is rare, and the specimens may be contaminated with impurities such as perspiration, urine, *497soil, and bacteria. These factors limit the application of the results of the study.[51]

When questioned about the proviso, the prosecution’s only expert with significant experience with evidentiary bloodstains commented, the cleanliness of the stains is not "as significant a problem as they think it is.” The prosecution emphasizes that no study has shown unreliability.

Prosecution witnesses testified about their experience with contaminants. Sensabaugh, relying on his own unpublished laboratory study, said that bacterial contamination would signal itself. He said the person interpreting the test would see "new bands appearing in odd positions . . . .” Fisher also testified that bacterial contamination would result in a "different activity, different position.” She suggested the contamination "will flag you . . . .”

Juricek has written, however, that bacterial contamination does not necessarily create easily excludable bizarre bands. "Many bacteria have been found to have Type 2 pgm, for example. Thus, Type 1 blood when contaminated by bacteria that have Type 2 pgm would be identified as Type 2-1 despite the use of starch gels and proper controls.”52 The result of the pgm test in the instant case was Type 2-1.

Fisher was more willing to recognize uncertainty with respect to unknown contaminants. When asked about soil, Fisher answered, "If it is contaminated with soil, I have no idea. It depends on what is in the soil.” In response to the judge’s comment that "your testimony is that if there was any problem with having to use it [electrophoresis of evidentiary bloodstains] in Court, it would have *498to come from an incompetent examiner who could not catch a flag that a contaminated sample is going to wave,” Fisher said, "There are obvious conditions that I have never tested and which nobody else has ever tested.” Although she followed up with the comment that "I can’t conceive of anybody [having problems], unless one is going to go around sprinkling the place with rare chemicals,” the testimony of another prosecution witness, Dr. Harvey Mohrenweiser, suggests that it is only fair to conclude that examiners will "catch flags” they are used to seeing. The court asked Mohrenweiser whether a competent examiner will "be able to separate the contaminated activity from the non-contaminated activity?” Mohren-weiser answered, the examiner "[s]hould recognize that there is a problem . . . they would be able to recognize and identify samples which are contaminated under sorts of conditions that we routinely operate. That’s part of the developmental procedure and recognize that it would be inappropriate to type . . . .” The only stains he had examined were those produced in laboratories. Because these stains were not collected under sterile conditions, there could be "some bacterial contamination.” Mohrenweiser and the other prosecution witnesses did not respond to the questions raised by the defense about the effect on electrophoresis of other likely crime scene contaminants such as chemicals and soil.

In sum, scientists do not agree what effect common crime scene contaminants may have on electrophoresis. They do not agree because comprehensive control tests have not been undertaken. The scientists testifying at trial had no experience with soil or chemical contamination and could only guess what effect such contaminants might engender. Although the scientists had some experience *499with the type of bacterial contamination found in laboratories, the bloodstains here were made during or following the commission of a crime and not under laboratory conditions.

VII

General agreement in the scientific community on the reliability of electrophoresis of evidentiary bloodstains has not been achieved because independently conducted validation tests and control studies have not been undertaken, and the results subjected to the scrutiny of the scientific community. Legal commentators have spoken of the need for comprehensive testing to establish the reliability of a new technique.53

The scientific tradition expects independent verification of new procedures. When other scientists analyze and repeat the tests, they counteract the dangers of biased reporting. It is scientists not responsible for the original research that confirm its validity.

Although electrophoresis has been generally accepted as reliable in the scientific community for many years, Wraxall’s multisystem test is a new technique. No independently conducted verification studies have been undertaken. Scientists evaluating the technique necessarily base their conclusions on the unpublished reliability study conducted by the multisystem’s developer. General agreément in the scientific community cannot be achieved on the basis of this type of testing alone. Independently conducted reliability tests on Wrax-all’s multisystem could, however, be undertaken without great difficulty. Such questions as whether the filter compromises the subsequent tests would *500seem to be readily resolvable through independent verification.

The absence of control studies measuring the effects of various contaminants on electrophoresis also stands in the way of general agreement in the scientific community. Although it is clearly not possible or necessary to measure every conceivable contaminant in the environment, or at least every contaminant a defense attorney can imagine, the effects of certain common contaminants such as soil and gasoline could and should be tested. According to the record, their effect is presently unknown. It is not clear whether they alter or remove bands and thereby destroy the reliability of the test or whether they just render the results uninterpretable or create an easily identifiable stray band.

The dangers of allowing implementation of an inadequately tested device are well-known. The paraffin test and the Daikon Shield are two familiar examples. The paraffin test was used by law enforcement agencies and introduced as evidence in court to establish that a suspect has recently fired a gun.54 "The theory behind the test was that the results established the presence of particles of nitrates . . . deposited on the hand by the gases of a discharged cartridge.”55 It was not until 1967, over thirty years after the test was accepted as reliable evidence in court, that the first comprehensive study was published conclusively showing "that many people who never fired a gun but whose profession, occupation, or happenstance brought them in contact with nitrates can be expected to yield positive reactions to the test.”56

*501The history of the Daikon Shield is also instructive. "With adequate testing, controlled studies and cautious marketing, [the manufacturer] could have discovered the increased risks which have been shown to be inherent in the Daikon Shield’s unique new design.”57 Instead, because of its desire to get the new product on the market as quickly as possible, "[t]he defendant relied entirely on [the researcher], who had not only the pride of invention, but also a personal incentive to bias his judgment. . . .”58

VIII

We conclude that the prosecution has failed to demonstrate general acceptance of the reliability of electrophoresis of evidentiary bloodstains by the scientific community.

We turn to the questions whether the error in admitting the electrophoresis evidence was harmless or whether we should remand for a new trial on a charge of second-degree murder.

The erroneous admission of electrophoresis evidence does not necessitate a retrial if it is clear the error did not prejudice the defendant. In the instant case, however, it is not clear the error was harmless.

At trial no one testified about the reliability of electrophoresis except Detective Stolorow. He presented a slide show demonstrating the technique and testified concerning its reliability. As previously discussed, he said that only 1.3 percent of the population has the protein subtype and abo combination found in the sidewalk stain and one-half of one percent has the subtype and abo combi*502nation found in the porch stain. He said that this meant that one of seventy-seven people would match the sidewalk stain and one of one hundred and seventy-seven people would match the porch stain. Parenthetically, cross-examination at trial did not reveal the absence of independent verification and control studies or other contradictory evidence subsequently raised by Grunbaum and Juricek during the evidentiary hearing.

The prosecution has argued that the admission of the electrophoresis evidence was harmless because of the "abundance” of other evidence.59 The other proof consisted of abo blood tests, fingerprints, and the testimony of Terrence Coleman.

No one contests the reliability of abo blood testing in general or the results of the abo test in the instant case. The porch and sidewalk stains were identified as blood group o. This has, however, little significance. As Stolorow testified at trial, forty-five percent of the population have type o blood. Young was not even the only suspect with type o blood.

The fingerprints found in the victim’s house are more significant. Twenty-three fingerprints were found, but eleven did not have enough points of comparison to be identifiable.60 Of the twelve remaining prints, ten of these were Lechtanski’s. The other two fingerprints which were found on the handles of a toilet and dresser drawer, were ultimately identified as Young’s. Questions were *503raised, however, concerning the reliability of the identification.

The fingerprints were compared to the fingerprints of Young on two separate occasions. On May 25, the police laboratory specialist was asked to compare the prints of nine suspects, including Young. The report listed the name of the suspects, with Young’s name at the top of the list, and concluded "Comparison was made of the latent prints and fingerprints with negative results.” The laboratory specialist’s explanation at trial of the negative results was that "the only thing I can think of is that I had the cards on my desk, that I had some that I hadn’t compared and some that I had compared. I either picked up two cards and stuck his card in with the one possibly above it, or I just put them in the wrong stack to begin with and apparently didn’t compare his card at the time.”

On June 13, however, the laboratory specialist was asked once again to compare Young’s fingerprints with the two prints found in the house. This time he concluded that the fingerprints "were made by the same person.” He said there were eight points of comparison between the thumb print found on the toilet handle and Young’s thumb print and more than twelve points of comparison between the fingerprint found on the bottom drawer handle and Young’s right little finger.

Experts generally require that "a minimum of eight identical ridge characteristics ... be found in both prints, though most experts prefer at least 10-12 concordances.”61 Although both prints meet the minimal requirements, the fingerprint from the toilet handle just clears the threshold. Of the two fingerprints, the print on the toilet handle is *504the most important. The reason is the toilet handle fingerprint would have probably been made recently, because "any print that would be put on there would be . . . destroyed by the next person that . . . flushes the toilet.” Fingerprints can last a long time, and defense counsel suggested to the jury that the second print might have been made at an earlier date. During cross-examination of a police officer, who testified that Young said he did not know Mitchell Lechtanski and had never been to his house, defense counsel asked "couldn’t Mr. Young have known him by simply a name Mickey?”62 The detective had not shown Young pictures of Lechtanski or the house.

An acquaintance of Young, Terrence Coleman, testified that the day before the murder he encountered Young on the street and asked what his plans were for the next day, May 16. Coleman testified that Young responded he was planning a robbery on the north side of town. Coleman said he met the defendant the next day and asked him about the murder of Lechtanski: "I asked him if he did it. And he said he did, and the only reason why he did do it because he was coming after him with a blunt object. . . .”

The reliability of Coleman’s statements depends on his credibility as a witness. Before the murder he had been convicted of driving away a motor vehicle and delivery of marijuana. Although he said he spoke to Young on May 16, Coleman did not tell the police what he knew until October 26, when he was in jail on charges of possession of stolen merchandise. When he testified at Young’s trial, there were charges pending against Coleman. Coleman said, however, that no promises had been made for his testimony._

*505As an appellate court, we do not independently evaluate this evidence. "[I]t is not the appellate court’s function to determine guilt or innocence. . . . Those judgments are exclusively for the jury . . . ,”63 Our responsibility is to determine how the error might have affected the jury’s decision. The inquiry is "what effect the error had or reasonably may be taken to have had upon the jury’s decision.”64

If it were clear that the erroneous admission of the electrophoresis evidence did not prejudice Young, the error would be harmless. We are, however, of the opinion that but for the electrophoresis evidence the jury may have had a reasonable doubt, and that evidence might have made the difference. We therefore remand for a new trial on the charge of second-degree murder.

(dissenting). The issue in this case is whether serological electrophoresis of evidentiary dried bloodstains 1 has achieved general scientific *506acceptance.2 As is frequently true, the point at which a court takes up analysis of an issue may mask the real concerns of the law and of individual justices. The question in this case is actually *507whether the Court is sufficiently convinced of the reliability of this process that it is willing to trust trial courts and juries to treat application of theory and technique in a particular instance in a manner consistent with procedural fairness and just results. This concern, whether articulated as the Frye3 standard or a different approach to admission of expert opinion, is the motivation for the assertion of an appellate court function when what is proposed to be offered is evidence of a new scientific procedure.4

I conclude that the record below establishes that both the theory of electrophoresis and its application to dried evidentiary bloodstains have gained general acceptance in the particular field in which each belongs. Frye v United States, 54 US App DC 46; 293 F 1013 (1923).

THE FRYE TEST

The Frye standard was first formulated by the United States Court of Appeals for the District of Columbia Circuit in 1923. In a case of first impression, the court found that the polygraph evidence offered by the defendant5 was inadmissible because the technique had not been sufficiently accepted as *508reliable by the relevant scientific community. In oft-quoted language, the court concluded:

Just when a scientific principle or discovery crosses the line between the experimental and demonstrable stages is difficult to define. Somewhere in this twilight zone the evidential force of the principle must be recognized, and while the courts will go a long way in admitting expert testimony deduced from a well-recognized scientific principle or discovery, the thing from which the deduction is made must be sufficiently established to have gained general acceptance in the particular field in which it belongs. [Frye, supra, p 47.]

Although critics have asserted that in the two-page opinion, the Frye court neither cited authority nor offered an explanation for adopting the general acceptance standard6 and that the standard may have been nothing more than dicta,7 the Frye test was soon accepted in most jurisdictions.

Since its adoption, the criticism of the Frye test has often been vehement. As one commentator noted:

Commentators have not been restrained in their criticism of the Frye test. See Moenssens [Polygraph Test Results Meet Standards for Admissibility as Evidence, in Legal Admissibility of the Polygraph (Ansby ed, 1975)] at 19 ("archaic”); 22 C Wright & K Graham [Federal Practice & Procedure § 5168] at 87 ("a 'sport’ ”); Conrad, Landmarks and Hallmarks in Scientiñc Evidence, in Sourcebook in Criminalistics 37, 38 (C Hormachea ed, 1974) ("antiquated on the day of its pronouncement”); Tarlow, Admissibility of Polygraph Evidence in 1975: An Aid in Determining Credibility *509in a Perjury-Plagued System, 26 Hastings L J 917, 923 & n 38 (1975) ("infamous”). [Giannelli, The admissibility of novel scientiñc evidence: Frye v United States, a half century later, 80 Colum L R 1197,1206-1207, n 59 (1980).]

Critics have focused on several problems with the Frye test. Most frequently, comments have been directed at its excessive restriction on the admission of relevant evidence, its tendency to fix standards when the evolution of scientific methodology may well make the standard outmoded in the future, and the implausibility of assigning many new scientific methods into well-defined areas. See Costley, Scientiñc evidence — Admissibility Fryed to a crisp, 21 S Tex LJ 62, 64-65 (1980); Giannelli, supra, 1208-1209.

An additional criticism of the Frye test is that there are no definite criteria to use to decide if there has been general acceptance. Because it is impossible to find unanimous agreement in any field, the courts have been hard-pressed to find the appropriate number of experts who must have accepted the technique as reliable. As one commentator noted:

It would be a mistake for the courts to wait for certainty and complete agreement. Neither science nor the law functions on such a premise. The inconsistencies found in the social world are varied and ever changing. In view of this, the courts should apply scientific principles while bearing in mind the fact that current reality means making decisions under uncertain and unstable conditions. [Costley, supra, p 67.]

See also Strong, Questions affecting the admissibility of scientiñc evidence, 1970 U Ill L F 1 (1970), and Giannelli, supra.

In response to the criticisms of the Frye test, *510several courts and at least one federal circuit8 have either abandoned the test entirely or severely limited its application. In a lead case, the Supreme Judicial Court of Maine ruled that novel scientific evidence would be admissible into evidence when "the testimony to be given is relevant and will assist the trier of fact to understand the evidence or to determine a fact in issue.” State v Williams, 388 A2d 500, 504 (Me, 1978). See also United States v Williams, 583 F2d 1194 (CA 2, 1978), cert den 439 US 1117 (1979); Whalen v State, 434 A2d 1346 (Del, 1980), cert den 455 US 910 (1982); State v Hall, 297 NW2d 80 (Iowa, 1980), cert den 450 US 927 (1981); Phillips ex rel Utah Dep't of Social Services v Jackson, 615 P2d 1228 (Utah, 1980); State v Catanese, 368 So 2d 975 (La, 1979); Watson v State, 64 Wis 2d 264; 219 NW2d 398 (1974).

Some states and federal circuits, while not specifically rejecting the Frye test, have modified it in cases where it would have posed an obstacle to the admissibility of otherwise relevant evidence. In Coppolino v State, 223 So 2d 68 (Fla App, 1969), app dis 234 So 2d 120 (Fla, 1969), cert den 399 US 927 (1970), for example, the Florida District Court of Appeals approved the admission of the results of tests formulated specifically for the case (test to detect the presence of succinylcholine chloride in body tissue), while specifically noting that Florida continued to adhere to the Frye rule. The record in that case only indicated that the reliability of the test was accepted by the state’s witness and not by those presented by the defense._

*511The Massachusetts Supreme Judicial Court modified the Frye Standard in Commonwealth v Lykus, 367 Mass 191; 327 NE2d 671 (1975), stating that general acceptance could be shown if it were only among those who would be expected to be familiar with its use. See also State v Souel, 53 Ohio St 2d 123; 372 NE2d 1318 (1978); People v Allweiss, 48 NY2d 40; 421 NYS2d 341; 396 NE2d 735 (1979); Ibn-Tamas v United States, 407 A2d 626 (DC, 1979); People v LaSumba, 92 Ill App 3d 621; 414 NE2d 1318 (1980), cert den 454 US 849 (1981); Ex parte Dolvin, 391 So 2d 677 (Ala, 1980); United States v Bailer, 519 F2d 463 (CA 4, 1975), cert den 423 US 1019 (1975); United States v Franks, 511 F2d 25 (CA 6,1975).

The continuing debate over the proper approach to novel scientific evidence is generated by the tension between a reluctant approach to novel evidence which may carry great weight for the factfinder and the principle that evidence which may be of value to the factfinder should be precluded only when necessary to carry out the dictates of a conflicting and overriding policy. I do not advocate that this Court abandon the Frye test. I would, however, reject the requirement that the reliability of novel scientific evidence must be established by disinterested witnesses, People v Barbara, 400 Mich 352; 255 NW2d 171 (1977), and People v Tobey, 401 Mich 141; 257 NW2d 537 (1977), that is, those whose livelihood is not "intimately connected with the new technique.” While it may be appropriate for the Court to use Frye for its intended purpose, I do not see a basis for the expansion of the rule in a direction which erects insurmountable obstacles to forensic evidence and thus needlessly impedes the admission of probative evidence. Such a rule establishes a "Catch 22” approach to admissibility in the face of a national *512trend toward a more liberal approach to the admissibility standard.9

Contrary to the contention that the approach of the majority is not a modification of People v Barbara and People v Tobey, neither case can be read as standing for the proposition that a foundation must be established by disinterested scientists, or that while "a certain degree of 'interest’ must be tolerated,” ante at 483, the testimony of a scientist whose livelihood is intimately connected with the new technique must be disregarded. Rather, in Barbara, the Court specifically noted that the credentials of the witnesses, "although outstanding for polygraph technicians, are not those of scientists.” Id., p 377. The Court also noted "[u]nder the present state of the art the general acceptance of the polygraph among psychologists and physiologists cannot be demonstrated, because such acceptance does not exist.” Id., p 390. While the Court stated that "one would want, where the task was to demonstrate general scientific acceptability, an acknowledgment of the value of the device and the techniques by disinterested scientists whose livelihood was not intimately connected with it,” id., p 376, Barbara did *513not stand for the proposition that the Frye standard could only be met by disinterested scientists.

Nonetheless, the Court in Tobey, supra, p 145, cited the rule of Barbara as follows: "[Gjeneral scientific recognition may not be established without the testimony of . . . 'disinterested scientists whose livelihood was not intimately connected with’ the new technique.” The Court in Tobey then applied the rule to the testimony of a police officer "but not a scientist” and to the testimony of a professor of audiology and noted that "[n]either Nash nor Tosi, whose reputations and careers have been built on their voiceprint work, can be said to be impartial or disinterested.”10 Id., p 146.

I do not quarrel with the Court’s conclusion in either Tobey or Barbara that the record was insufficient to establish general acceptance in the relevant scientific community. What I do take issue with is the majority’s conversion of the rationale of Barbara into a rule requiring the disqualification of the testimony of witnesses who indisputably have education, formal training, and background in the applicable scientific disciplines.

That a court may weigh the credentials and self-interest of a witness in determining whether a sufficient showing of reliability has been made is self-evident; it does not logically follow, as the majority suggests, that a witness’ self-interest compels this result.11_

*514A trial or appellate court could comfortably conclude that the testimony of technicians is insufficient to establish reliability. There is, however, simply no basis for placing wholesale the "disinterested expert/intimately connected” gloss on the Frye test. Indeed, as the trial court noted, the effect of such a test diminishes, as a matter of law, the testimony of Mark Stolorow, who is not a technician, but the holder of a bachelor of science degree and a master’s degree in forensic chemistry, is a codeveloper of the technique, and who has been working with it and teaching it since 1975. It also diminishes the testimony of James Kearney of the serology unit of the fbi who holds a bachelor of science degree in bacteriology and a master’s degree in microbiology, and who also has worked with and taught the procedure of electrophoresis of evidentiary dried bloodstains since 1978 (in a laboratory that does 8500 electrophoretic examinations a month, ninety percent of which are done by the multisystem).

The instant record is a paradigm of the problems created by such an approach. The testimony of those who have the most knowledge in the field, the forensic scientists, is regarded as suspect. The testimony of Drs. Sensabaugh, Grunbaum,12 and Juricek, all of whom may be said to have an advocacy interest in the issue, is accorded greater weight. The witnesses who have no financial interest in forensic phenotyping and no actual experi*515ence in the process are accorded the most deference in this variation of the Frye/Tobey/Barbara analysis. The result is that those scientists who know the most about the process are viewed, as a matter of law, as the least persuasive witnesses. The "livelihood . . . intimately connected with the new technique” test represents a regressive approach to scientific developments which, parenthetically, would have devalued the opinions of Jonas Salk, Albert Einstein, or Marie Curie, each of whose life work, livelihood, and standing in their professional community was intimately connected with a new scientific procedure.

THE THEORY OF ELECTROPHORESIS

Electrophoresis is the movement of charged particles through a buffered conducting medium by application of a direct current. The term isozyme is used to describe enzymically active blood proteins which can be identified by their relative mobilities in an electric field. After separation of the proteins into marker bands by application of a current, specific chemicals are applied to make the proteins visible. Biology Methods Manual, Metropolitan Police Forensic Science Laboratory, London, England (1978). The relative distance of the bands from a common origin is compared with known standards, and evaluated by established guidelines.

The results are then compared to population studies which show the known frequency of each factor in a given population. This produces a statistic which is representative of the percentage of the population that has that group and those factors in common.13 The more genetic markers *516identified, the smaller the population of persons who might possess a particular combination of factors.

APPLICATION OF THE THEORY TO EVIDENTIARY BLOODSTAINS

It is undisputed on this record that the theory of electrophoresis is generally accepted as reliable in the scientific community. Indeed all witnesses at the hearing on remand so testified.

There is ample basis in the record, even apart from defendant’s concession, that the theory of serological electrophoresis, and the application of the theory to laboratory-produced dried blood samples, has achieved general scientific acceptance as a reliable identification technique.14 Dr. Grunbaum testified at trial to these conclusions and also acknowledged that proficiency tests established that, as to evidentiary materials, "there are many competent analysts in the country that can do it [electrophoresis on evidentiary bloodstains] with a great deal of confidence.”_

*517In general, fewer marker systems are usable in stain analysis than in fresh materials. There are in excess of seventy-five markers in human blood, approximately a dozen of which are amenable to stain analysis. Whether a genetic marker can be typed in a stain depends on whether it persists in recognizable form in the stain material. This inquiry has required a determination of the life expectancy of particular markers in samples exposed to air and other possible contaminants. If the marker persists, then methods which produce reliable results with fresh blood will also give reliable results with stain material, Denault, Detectability of selected genetic markers in dried blood on aging, 25 J Forensic Sciences 479 (1980), provided the samples are carefully prepared and interpreted, Grunbaum, "Procedures for Phenotyp-ing of Genetically Controlled Enzyme and Protein Systems,” Handbook for Forensic Individualization of Human Blood and Bloodstains (1981).

The record establishes that the five genetic markers tested for in this case (eap, esd, glo, pgm, and Hp) persist in dried stains and that it is a common practice, accepted in both this country and the international scientific community, to transmit dried stains, on paper or cloth by mail. The testimony also revealed that these stains remain readable through the use of accepted protocols. See also Denault supra; Gaensslen, Source-book in Forensic Serology, Immunology, and Biochemistry, National Institute of Justice (1983). Gaensslen, supra, pp 431-432, reports fourteen studies on the survival of pgm enzymes in dried bloodstains and notes that all the samples could be correctly typed up to thirty-two days. Denault’s study reports the detectability of eap, pgm, Ak, and ada. The markers in all samples were unknowns to the investigators. Denault, supra. See also testi*518mony of Dr. Rachael Fisher, Michigan State University, and Dr. Harvey Mohrenweiser, University of Michigan, Department of Human Genetics. The literature also reports that in 1976 Wraxall and Eames reported a starch gel method for electro-phoretic typing of eap in stains and that "blind trial studies” indicated that the procedure was completely reliable. Gaensslen, supra, p 451. Dr. Rachael Fisher also testified that alterations from drying were not seen in pgm or ead and that the effect of aging on ada and eap could be reversed by Cleland’s reagent.

Thus, while it is true that blood degrades rapidly while drying, the literature and the record establish that markers which undergo alteration during drying are rejected for forensic use and that the markers in question in this case may be reliably identified by electrophoresis. Sensabaugh, Uses of polymorphic red cell enzymes in forensic science, 10 Clinics in Haematology 185 (1981).

The results of a Law Enforcement Assistance Administration study of dried stains were not published, but the study, conducted by Brian Wra-xall, is available in the National Criminal Justice Reference Service and the findings have been circulated to, and adopted by, a sizeable number of forensic laboratories. Thus, both the theory and its application to stains have been put into practice and relied upon for an extended period of time. Wraxall, Final Report, Bloodstain Analysis System, United States Department of Justice, 1978.

The literature and the record in this case indicate a continual effort to monitor results and to further reduce the possibility of error in electrophoresis of dried bloodstains. As a result of this effort, scientists and technicians are now well aware of procedures and indicators unique to dried bloodstain analysis, and genetic markers which undergo changes which will affect the typing re-*519suits are rejected for use by forensic scientists (and by geneticists for population studies).

Also, the forensic process itself includes the systematic evaluation of the stability of markers stored under various known conditions such as aging, humidity, chemicals, heating, and freezing. If the typing of a marker substrata system cannot be made reliable, then it is discarded. Sensabaugh, supra, p 198.

Because the principal effect of bacteria is on wet blood, contamination in a dried stain will signal itself as untypeable or will result in different band patterns. These patterns are evaluated by guidelines set forth in the literature. Biology Methods Manual, supra, pp 2-88 to 2-138. Bacteria, thus, manifests itself as a problem which is different from those normally seen and not comparable to the control marker. Forensic and nonforensic scientists routinely analyze air dried samples which have been contaminated by bacteria. One example of this process is illustrated by the method in which difficulties in the eap system have been addressed. See Williams & Shah, "Enzyme Patterns in Bacterial Classification and Identification,” Microbiological Classiñcation and Identiñcation, 1980, pp 299-318. The eap system is based on a genetic explanation of six phenotypes called A, ba, b, ca, cb, and c. The differences between these phenotypes on electrophoretic plates are a matter of band intensity. Misinterpretation of phenotypes is possible if these effects are not fully appreciated. The methodology, therefore, dictates the inclusion of appropriate controls of known phenotypes when using any phenotypic procedure, and particularly when attempting to diagnose b, cb, and c phenotypes. Gaensslen, supra, p 452.

Finally, and most significantly, the record indicates that the consequence of virtually all typing errors is a false exclusion. If one typing error is *520made on a true person, that person is excluded.15 The possibility of false inclusion is more remote.16

As the trial court correctly observed, the reliability of the particular sample in a given analysis is a distinct question from the reliability of the theory or the technique. The claim that the possibility of contaminants in a crime scene sample precludes a conclusion of the reliability of the technique or the theory of electrophoresis, is in reality a claim that there can never be a reliable forensic opinion,17 since as Dr. Grunbaum noted, the range of potential contaminants "goes on beyond anybody’s imagination.” While it is true that specimens may be contaminated with impurities, the literature indicates that different conditions of exposure cause variation in the persistence of typeable markers and that those markers that persist can be reliably identified. Although, as the majority notes, Dr. Juricek has testified and written that Type 2 pgm may contaminate Type 1 blood and lead to a Type 1 pgm being identified as *521a Type 2-1, the literature indicates that the only false positives that have been reported for pgm are in whole blood samples. Gaensslen, supra, p 433.

The effects of heating, humidity, aging, metals, chemicals, air born bacteria, and substrata have been investigated for the markers in question in this case. Culliford, The Examination and Typing of Bloodstains in the Crime Laboratory, Department of Justice, 1971, pp 108, 117, 120; Denault, supra; Biology Methods Manual, supra, pp 2-88 to 2-138. The application of electrophoresis to eviden-tiary dried bloodstains is generally accepted in the relevant scientific community and is reliable.

DISTINCTION BETWEEN RELIABILITY OF THEORY AND TECHNIQUE AND THE EVALUATION OF EXPERT OPINION PROFFERED IN A PARTICULAR CASE

If there is a question regarding the sample offered as the basis for the opinion in this case with regard to potential error in age of the marker, the fabric from which it was obtained, the particular test used or other factors going to the foundation of the expert opinion, these are factors to be evaluated by the trial court in its traditional function of determining admissibility, MRE 103, 104. If the test results are determined preliminarily to be admissible, facts affecting the weight and credibility of the opinion may, of course, be placed before the factfinder.

The effect of crime scene contaminants on a particular sample is a question that relates to the trial court’s duty to determine, as a preliminary issue of admissibility, whether the expert opinion offered in a given case with regard to a particular sample has a sufficient foundation to be relevant.18 If the testimony is admitted, questions concerning *522the reliability of the sample and the test employed are explored through cross-examination and go to the weight and credibility of the opinion.19

Because both the theory of electrophoresis and its application to evidentiary bloodstains are generally accepted in the relevant scientific community, I conclude that the testimony was properly admitted below.

THE MULTISYSTEM TECHNIQUE

While the issue of the particular technique used in this case was not within the scope of the remand order, and is not a proper subject for review by this Court, the majority deals extensively with the multisystem method as if that were the technique in question. Indeed, the observation in the majority opinion, p 491, that "[bjecause the results of the glo test were not interpretable in the instant case, the only issue is the reliability of the pgm results in the multisystem analysis,” highlights the fact that this is a question that goes to the discrete issue of admissibility and credibility of the opinion offered in this case.

It may, however, he noted that combination analysis, or a system that analyzes more than one enzyme at a time, is routinely used and accepted as reliable in the scientific community. The testi*523mony reveals that a combination method for the three markers tested simultaneously in this case, pgm, esd, and glo, is used in genetic research in the Departments of Pediatrics and Pathology at Michigan State University. A combination method involving two sequential markers on the same gel is also in use at the University of Michigan Medical School, and a combination method was used for three years at the Minneapolis Blood Bank.20 The literature establishes that there are a wide variety of multiple systems used to determine polymorphic enzymes on the same electrophoretic run and that these procedures are used by forensic and nonfor-ensic scientists. See Gaensslen, supra, p 432.

The particular multisystem method used in this case has been used by the fbi since 1979, is in use in more than one hundred crime laboratories in the country, and is taught to and used by students as a protocol at the University of California in Berkeley. It is also used for fresh blood testing in an independent laboratory available to any seeker of its services.

This method was tested by blind trial tests conducted in four separate laboratories. Final Report, supra. While the results were returned to and tabulated by Beckham Laboratories, to which Brian Wraxall was then under contract, it is not correct to suggest that this was not an independent study or that these tests were self-verifying. These blind trials were not "conducted” by Mr. Wraxall. Rather, five batches of six bloodstains, each of various ages, were sent to serologists in *524their own labs, for typing of eight genetic markers. Thus, while it appears that the examiners knew what they were looking for, the tests were blind in the sense that the examiners did not know the identity of the markers in the particular sample.21 Of a total of 912 readings, only one reading was incorrect. Wraxall, p B-7.22

*525Subsequent blind trial proficiency tests for laboratories using the multisystem and those using other systems were conducted by the Forensic Science Foundation from 1979 to 1983. The sum total error rate on individual marker analyses was 1.6 percent of 3107 total tests per protocol.23

I also disagree with my colleagues’ evaluation of Dr. Rachael Fisher’s testimony with regard to the multisystem. Dr. Fisher was qualified as an expert and is on staff at Michigan State University in the Departments of Pediatrics and Pathology. She testified that she had run the enzymes in hundreds of different systems in her research on inherited diseases, had read the manual for the multisystem, and offered her opinion that the system incorporated all the features that "are necessary to make it work.” She further stated that the protocols for ead, pgm, glo, eap, and Hp are recognized and, in reply to the specific question, "If you had a dried bloodstain or a limited supply of blood, could you use a thin gel method and develop all three on that one thin gel?”, she answered, "yes, one could be done certainly.”

In any event if, as Dr. Grunbaum opined, there was a misapplication of the particular technique in the Young case, in my judgment this is an issue to be resolved by the trial court. As the Ninth Circuit noted in United States v Gwaltney, 790 F2d 1378, 1382 (CA 9,1986):

*526To the extent Gwaltney complains of the application of the procedure in this instance, he does so in the wrong forum. Criticism of the application of a valid test in a particular instance bears on weight, not admissibility.

I would affirm.