¶1The defendant, Dana A. Daggett, was convicted of murder in the first degree by reason of extreme atrocity or cruelty. The principal claim of error the defendant makes is that evidence of tests performed on the victim’s deoxyribonucleic acid (DNA) and DNA extracted from blood in the trunk of the defendant’s automobile and blood found at the plant where defendant worked was improperly admitted at trial. The defendant contends that the error committed requires a new trial. The author of this opinion and Justice Nolan conclude that the evidence was not properly admitted; a majority of the quorum concludes that any error was not prejudicial. Therefore, we affirm.
¶2There was evidence of the following facts. On Saturday, November 5, 1988, the body of the victim was found along a path in a wooded area of Pembroke. The victim had been a prostitute working in Brockton. She had been stabbed thirty-seven times. The absence of blood near her body indicated that she had died elsewhere and then was transported to that spot. The only clothing on her body was a pair of socks.
¶3The defendant worked as a water pump operator at a water treatment plant in Pembroke. On Thursday, November 3, 1988, his shift began at 10 p.m. and ended at 6 a.m. on Friday. He was the only employee on that shift. The evidence presented by the Commonwealth tended to show that the defendant left the plant during his shift to solicit a prostitute, took the victim to the plant, killed her there, and then transported the victim’s body in the trunk of his automobile to the wooded area where the body was discovered.
¶4*349The defendant’s coworkers testified that, on Friday morning, an area of the pump room floor appeared freshly mopped. They testified they saw what they believed to be blood stains at various locations in the plant. Among other places, human blood was found on the stairs, the leg of a podium, a wooden chair, a table, the floor, and a mop in a bucket of water. Blood was also found on a mat in the trunk of the defendant’s automobile. Serologic tests other than DNA testing indicated that the blood was consistent with the victim’s blood, but not the defendant’s, and that approximately 4.5 per cent of the general population has blood consistent with these stains.
¶5Hairs from the trunk of the defendant’s automobile were consistent, in approximately thirty microscopic characteristics, with the victim’s hair. Pubic hair consistent with the defendant’s was found on the victim’s body. Paint chips taken from the trunk of the automobile were consistent with paint on the floor of the water treatment plant and paint chips found on the victim’s socks.
¶6A box containing charred clothing was found in the basement of the plant several days after the victim’s body had been discovered. The victim’s roommate testified that the victim was wearing these clothes when the roommate had last seen her. A charred metal barrel also was recovered at the plant. The Commonwealth presented evidence that the barrel, the ground near where it was found, and some of the clothing found in the box contained traces of an accelerant.
¶7In addition, the Commonwealth offered evidence of DNA testing by Cellmark Diagnostics laboratory (Cellmark) to prove that the blood found on the chair and the table in the plant, and on the mat in the trunk of the defendant’s automobile, was the victim’s blood. At trial, the Commonwealth’s expert witnesses testified to the process of DNA comparison testing and concluded that it was “highly likely” (or some other nonnumerical term) that the blood found on the chair, the table, and in the trunk came from the victim.
¶8The evidence of DNA testing was the subject of a five-day pretrial hearing conducted in April and May, 1990. The pur*350pose of the hearing was to determine whether the scientific theory and process underlying the forensic use of DNA technology is generally accepted by the relevant community of scientists.
¶9At the time of the Fryehearing, and at the time of trial, this court had not yet issued its opinions in Commonwealth v. Lanigan, 413 Mass. 154 (1992), and Commonwealth v. Curnin, 409 Mass. 218 (1991). In both of those cases, we held that evidence of DNA testing was inadmissible because the methods used by Cellmark to calculate the statistical probability of a random match were not generally accepted by the relevant scientific community.
¶10*352The Commonwealth argues that, even if it was error to admit the evidence of DNA testing, reversal is not necessary because the defendant was not prejudiced by the evidence. A majority of the quorum agrees.
¶11As described above, DNA tests were performed on blood samples recovered from the chair and table at the plant, and from the trunk of the defendant’s automobile. As regards the DNA testing of the blood found on the chair and table at the plant, we note that defense counsel conceded in her closing argument that the victim had either been killed at the plant or brought there afterward (thus implicitly conceding that the blood at the plant came from the victim). She characterized the evidence other than the DNA testing that the victim had been killed at, or brought afterward to, the plant as “compelling.” Given this concession, we can say confidently that the DNA analysis tending to show that the blood at the plant was the victim’s was merely cumulative of other overwhelming evidence and its admission was not prejudicial.
¶12The only other DNA evidence concerned the blood found in the trunk of the defendant’s automobile. Significant other evidence tended to show that the victim’s body had been placed in the trunk. Blood typing tests revealed that the *353blood on the trunk mat was consistent with the victim’s blood, but not the defendant’s, and that only 4.5 per cent of the population has blood consistent with those stains. Human blood, also consistent with the victim’s, was found on other items in the trunk, including weather stripping, a wheel guard, and ropes. Hair recovered from the trunk was consistent with the victim’s hair. Paint chips, found in the trunk of the defendant’s automobile and on the victim’s body were consistent with each other and paint on the floor of the plant. Also linking the defendant to the victim was evidence showing that pubic hairs consistent with the defendant’s were found on the victim’s body.
¶13We believe that the evidence of DNA testing performed on the blood from the trunk mat was cumulative of this other evidence. See People v. Barney, 8 Cal. App. 4th 798, 825-826 (1992). In light of the fact that the Commonwealth was not able to present strong evidence of a DNA match for the blood from the trunk,
¶14We turn to the other issues raised in this appeal.
¶15Motions to suppress. Four days after the victim’s body was discovered, the defendant was transported from the plant to the Pembroke police station shortly after he arrived for his evening shift. At the station he signed a consent form giving officers the authority to search his automobile and apartment and provided the officers with a statement of his activities during the preceding week. He moved to suppress the statement and the items seized, and a hearing was held on his motions. The record amply supports the judge’s denial of Daggett’s motions. Daggett does not direct us to any particu*354lar error in the findings contained in the judge’s memorandum of decision and order, and we perceive no error.
¶16Evidence of prior illegal conduct. The Commonwealth was allowed to present testimony (over the defendant’s objection) that approximately ten months prior to the victim’s stabbing, the defendant was arrested for soliciting a prostitute. On the night of this arrest, the defendant was assigned to the overnight shift at the water treatment plant. The Commonwealth argued that the evidence was admissible as evidence of the defendant’s ability, opportunity, and knowledge to commit the crime, since he worked the overnight shift at the time the victim was allegedly killed. To support its contention, the Commonwealth introduced records from the plant that showed that the defendant had made certain hourly recordings on both nights. This evidence tended to show that, despite his work schedule and the log records, the defendant nonetheless had the opportunity, ability, and knowledge necessary to solicit and pick up a prostitute in Brockton. The judge gave limiting instructions both at the time the evidence was presented and when he charged the jury.
¶17Although the question is a close one, we cannot say that the judge abused his discretion in allowing the Commonwealth to present this testimony. The judge instructed the jury that the probative force of the evidence was not that the defendant “has a bad character,” in that he picked up prostitutes. The judge told the jury that they could “consider the evidence to the extent it may be found to establish a plan or scheme or a particular way of doing something or the opportunity to do something.” See Commonwealth v. Otsuki, 411 Mass. 218, 235-236 (1986); Commonwealth v. Young, 22 Mass. App. Ct. 452, 456 (1986). See also P.J. Liacos, Massachusetts Evidence 420 (5th ed. 1981 & Supp. 1985). As long as it is admissible for a purpose other than impugning character, relevant evidence does not become inadmissible because it indicates a prior offense. Commonwealth v. Robertson, 408 Mass. 747, 750 (1990).
¶18Remarks in the prosecutor’s closing argument. The defendant contends that reversible error was committed when *355the prosecutor made two improper comments in his closing argument. The first comment concerned the interpretation of the DNA test results. The prosecutor made a reference to “lottery type numbers” that, given the earlier determination that no numbers would be presented, could be perceived as a fact not in evidence. See Commonwealth v. Connor, 392 Mass. 838, 853 (1984). However, given our conclusion that the DNA evidence was not prejudicial, we conclude that any error created by this comment was likewise not prejudicial.
¶19The defendant also claims prejudicial error in the prosecutor’s closing statements regarding the defendant’s prior arrest for soliciting a prostitute.
¶20Review under G. L. c. 278, § 33E, Although the defendant has not made any argument that we should set aside or modify the verdict under our extraordinary power of review in a capital case, see G. L. c. 278, § 33E (1992 ed.), we have reviewed the entire record in fulfilment of our statutory obli-
¶22We shall not in this case repeat a discussion of the basic processes involved in Restriction Fragment Length Polymorphism (RFLP) analysis of DNA, the basic scientific acceptance of which is not disputed. A description of RFLP analysis can be found in Commonwealth v. Curnin, 409 Mass. 218, 227-231 (1991).
¶23Recent cases from other jurisdictions confirm what we noted in Curnin, supraat 222 n.7, i.e., the importance to courts and to the scientific community of a valid statistical analysis of the likelihood of a match. See People v. Axell, 235 Cal. App. 3d 836, 866-867 (1991) (match means little without data on probability; calculation of statistical probability an integral part of process); Fishback v. State, 851 P.2d 884, 893 & n.18 (Colo. 1993) (match unaccompanied by statistical significance “essentially meaningless”); State v. Vandebogart, 135 N.H. 365, 381-382 (1992) (match “virtually meaningless” without statistical probability expressing frequency). Likewise, the National Research Council’s April, 1992, report entitled DNA Technology in Forensic Science (NRC Report), stresses that: “To say that two patterns match, without providing any scientifically valid estimate (or, at least, an upper bound) of the frequency with which such matches might occur by chance, is meaningless.” NRC Report at S-8.
¶24Justice Abrams, in concurrence, post at 357, cites Prater v. State, 307 Ark. 180 (1991), in which the court held that evidence of a match, stripped of all information regarding the likelihood of a match, would be admissible. The Pratercourt’s bald assertion stands in stark contrast to the conclusions reached in Cumin, supra, the cases cited above, and a host of others. See, e.g., United States v. Yee, 134 F.R.D. 161, 181 (N.D. Ohio 1991); People v. Barney, 8 Cal. App. 4th 798, 820 (1992). Moreover, it is a puzzling assertion in light of the fact that the court recognizes, later in *351its opinion, that “[o]nce two or more DNA patterns derived from loci that are known to be polymorphic (different among individuals) have been matched, the question arises whether this matching is coincidental. This is the point at which statistics begin to play a role.” Prater, supra at 198. Apparently, the Pratercourt did not believe that this role was critical to the admissibility of DNA evidence. In accordance with the view of the relevant scientific community, see note 4, infra, this court does. (We note that Arkansas does not employ the Fryestandard, and uses instead a more liberal relevancy test. Thus, the fact that the scientific community requires evidence of the significance of a match would not be dispositive.)
¶25The methods at issue in both Laniganand Cumin were those of Cellmark and appear to be essentially the same methods at issue here.
¶26The fact that the Commonwealth presented the evidence to the jury in nonnumerical terms does not alter this result. The Commonwealth has cited no authorities and presented no testimony either at the Fryehearing or at trial that the use of such terms is generally accepted by the scientific community in evaluating the significance of a match. Although the Commonwealth’s experts testified that, in their opinions, evidence of a match indicated that the source of the samples was “highly likely” (or some other such phrase), no one testified that such characterizations alleviate the concerns the scientific community has voiced regarding the evaluation of matches. See State v. Cauthron, 120 Wash. 2d 879, 907 (1993) (testi-many of nonstatistical opinion presented insufficient because it did not include the background probability information). See also State v. Alt, 504 N.W.2d 38, 52 (Minn. Ct. App. 1993), and materials cited (“[N]onstatistical opinion testimony . . . would not correctly reflect the nature of the DNA test results. . . . The scientific community considers the statistical frequency important in interpreting DNA test results, but does not consider the opinion of a match to be significant”).
¶27The concurrence mischaracterizes this opinion when it states, post at 356, “I do not agree that the DNA evidence offered by the Commonwealth in this case was inadmissible because the Commonwealth’s experts’ testimony concerning the background probability that the DNA matches in question were false was not presented numerically.” The point is not that this court should require a numerical frequency, but that the scientific community clearly does. If the relevant scientific community generally accepted some nonnumerical expression of statistical frequency, then this court would likely accept it as well. See Commonwealth v. Beausoleil, 397 *352Mass. 206, 218-219 (1986). There was simply no evidence presented below that indicated that this might be the case, and, as the cases cited above illustrate, there is a substantial indication to the contrary.
¶28The concurrence’s reliance on Praterv. State, supra, for the proposition that “not all courts have said that the likelihood must be stated numerically,” post at 357, is misplaced. Significantly, the Pratercourt was speaking not to the question whether a nonnumerical (as opposed to numerical) indication of likelihood was acceptable, but to whether any evidence of likelihood is necessary. See note 2, supra.
¶29The defendant argues that the improper admission of DNA evidence was not “harmless beyond a reasonable doubt.” As the Commonwealth notes, this standard applies only to constitutional errors. See Commonwealth v. Sires, 413 Mass. 292, 297 (1992). Because the error here was not a constitutional one, our inquiry is whether the improper admission of the evidence constituted prejudicial error. See Commonwealth v. Schulze, 389 Mass. 735, 741 (1983). Under this standard we consider whether “the error possibly weakened [the defendant’s] case in some significant way so as to require a new trial.” Id.For the reasons stated, we conclude that this is not the case here.
¶30Each of the experts testified to seeing a different number of “bands” in the trunk mat “lane.” The expert from Cellmark, Dr. Robin Cotton, testified that she would not declare a match because the results were inconclusive.
¶31We note also that, unlike Cumin, the jury were not exposed to claims that only one person (the defendant) in 59,000,000 would have the DNA components disclosed by the test. Cumin, supra at 222. Thus, the possible impact on the jury of the DNA evidence in this case was markedly weaker.
¶32The prosecutor stated: “Has the defendant ever done this before? Yes, he has, and arrested for it. It was on a night that he was working, a night that all of the records indicate that he was there all night. You’ve heard testimony that this defendant would make it a regular practice to leave for hours at a time and would come up to Brockton to visit his brothers, amongst other people. I suggest to you, that, again, is evidence linking this defendant to [the victim]. Is it conclusive? Of course not. What we’re talking about here is circumstantial evidence . . . .” *356gation to do so. We perceive no reason to alter the verdict of the jury.