delivered the opinion of the court:
Following a jury trial, defendant was convicted of murder (Ill. Rev. Stat. 1979, ch. 38, par. 9 — 1) and sentenced to an extended term of 75 years’ imprisonment. On appeal, defendant contends that the State produced insufficient evidence to establish his guilt beyond a reasonable doubt, and that he was denied his sixth amendment right to a fair trial by: (1) improper comment; (2) erroneous admission of irrelevant and hearsay “motive” testimony; (3) erroneous admission of irrelevant scientific and physical evidence; (4) improper references to weapons; and (5) improper jury instructions. For the reasons stated below, we reverse the judgment of the trial court.
The following facts were adduced at trial. At 6 p.m. on March 31, 1981, Frank Paul left his offices at 747 Devon Avenue, Park Ridge, Illinois, to dine at a restaurant across the street. At about 6:45 p.m. he returned to his offices, known as Interport, Inc. At about 7:15 p.m., the restaurant owner and two friends observed Frank Paul staggering toward the restaurant, drenched in blood. Paul’s last words were “I have been attacked.” Paramedics took Paul to Resurrection Hospital, where he was pronounced dead at 8:35 p.m.
The autopsy revealed seven stab wounds and ten cutting wounds. According to the pathologist, cutting wounds on the right forearm and wrist indicated that the victim attempted to defend himself by “raising his forearm, hand, or grabbing the knife.” The pathologist testified that the victim’s injuries could have been caused by any knife with a blade at least four inches long.
Park Ridge Police Officer Douglas Dostert arrived at the restaurant while the paramedics were still attending to Frank Paul. He went to Interport where he was joined by other officers. They found the front door locked and the rear door standing open. There was no sign of a forced entry; no one was inside. A telex machine was in the “on” position, as were some of the lights. The victim’s jacket and wallet were undisturbed on a hanger near the victim’s desk. An employee of Interport testified that nothing of value had been taken from the offices. Dostert saw a large amount of blood on the floor and carpeting.
The Park Ridge police began a search of nearby streets, alleys, sewers and roofs at 7:30 p.m. Assisted by Cook County sheriff’s police, the Chicago police canine unit, and the city’s fire and public works departments, they searched an 18-block area, a park and a shopping center by 5 a.m., April 1. They recovered a tweed and suede hat, a pen-type flashlight and one latex glove in the alley behind Interport. Officers also recovered a number of blood-stained items on *366the 1200 block of South Fairview Avenue, less than a block from the Interport offices. They found pantyhose on the sidewalk, a camel-colored sportcoat under a parked van, a knife in a catch-basin, and a second latex glove in a different catch-basin. The blood on the knife and second glove was wet; this glove.had.been cut with a knife in the palm and thumb area. A Park Ridge police evidence technician photographed and inventoried the items recovered from South Fairview Avenue and from the alley adjacent to Interport. Fingerprint impressions and blood samples were prepared from these items and from the Interport offices.
On April 7, 1981, investigators went to the Century Medical Labs at 8347 South Stoney Island in Chicago. There they saw the defendant and observed bandages on his left hand between the thumb and index finger. “Almost immediately” after investigators looked at the bandages, defendant placed his left hand in the pocket of his examining coat, where he kept it until the investigators left. Two employees of the First State Bank of Park Ridge, corroborated by bank security photographs, testified that they did not notice any cuts or bandages on defendant’s hands on the morning of March 31,1981.
On April 10, 1981, investigators executed a search warrant for defendant’s residence at 5025 West Dakin in Chicago. They spoke with defendant’s sister, Marlene Harbold, and asked whether she had noticed any injuries to her brother. She said that his left hand was bandaged, and that he had explained to her that he had cut his hand while using an electric saw to cut a door. She also said that she had seen a circular saw and a door, both with blood on them, in the basement.
The circular saw and door were photographed, seized and inventoried along with defendant’s automobile and several other items. When asked what was found during the search, Officer’ Terrence Conley said “*** a gg snub-nosed revolver.” After a hearing in chambers, the court sustained defendant’s objection and instructed the jury to disregard any reference to a weapon. However, Conley was questioned extensively concerning knives and sheaths found at defendant’s residence.
Pursuant to a subpoena, the defendant presented himself for examination by Dr. Joseph Danna on April 10, 1981. Dr. Danna observed two cuts; one on the left thumb, the other in the webbing between the left thumb and index finger. The cut in the webbing was several inches long and about one-third of an inch deep. Danna stated that the wounds were caused by a sharp, straight object and not by a circular saw. Based on the stage of healing, both wounds were be*367tween five and 14 days old, but the wounds might not have occurred simultaneously. Danna further opined that the wounds could have resulted in a loss of sensitivity and manipulation in the fingers and that the wounds should have been sutured. Defendant’s treatment, closing the wound with butterfly bandages, was not consistent with accepted medical practice.
Defendant’s sister, Marlene Harbold, was called as a court’s witness. She was in Chicago on March 29 and 30, 1981, to attend a confirmation party for defendant’s daughter. Frank Paul was a guest at the party, and she noticed no argument between defendant and Paul. A week later, on April 5, 1981, defendant called her at her home in Louisville and asked her to come to Chicago to care for his children as he was not feeling well. When she arrived on April 5, defendant appeared to be ill. The next morning she saw a band-aid on his hand and asked him what happened; he said, “it was just a nick, I nicked it on a saw.” Later in the week her brother told her that he had cut himself twice recently with a butcher knife. She admitted that she may have neglected to mention defendant’s remark concerning the butcher knife to investigators on April 10 and to the grand jury on April 13, but she explained that she first recalled the remark in November of 1981.
Three technicians from the Illinois Department of Law Enforcement (IDLE), Mark Stolorow, seralogy coordinator, Michael Podlecki, forensic scientist, and Mohammad Tahir, forensic scientist, testified for the State concerning their scientific analysis of blood samples prepared from stained physical evidence, the victim’s stained clothing and the bodies of defendant and victim. The samples were typed according to 10 systems of genetic characteristics or “markers.” For identification purposes, the genetic marker system works by process of elimination: each characteristic which is found in a given sample excludes all persons lacking that characteristic or having a different characteristic from having shed that blood. Stolorow testified that given the 10 groups of characteristics tested and the percentages of the population known to display each characteristic, “*** the probability of an accidental match is less than one in 500.” He explained that the fewer tests that could be performed on a given sample, the fewer people could be excluded and the greater the chance of an accidental match. He stated that, as contrasted with fingerprinting, blood typing technology does not permit the association of a blood sample with any one person.
Defendant’s blood tested as type O according to the conventional ABO typing system. The sample taken from the victim’s body was *368considered unreliable because five pints of universal type 0 blood was transfused into Frank Paul before he was pronounced dead. In addition, the vial containing the sample was broken when IDLE technicians received it. The IDLE experts “assumed” that the victim’s blood-stained shirt, slacks, shoes, and socks contained the victim’s blood. These stains were type A. The results of the remaining genetic marker tests were placed before the jury in the form of testimony and exhibits, but we see no need to catalog those results here.
All of the blood found near the scene was consistent with either defendant’s blood sample or stains on the victim’s clothing. The sport-coat contained both type 0 and type A blood. The glove recovered from the catch-basin had only one type 0 blood on it; the blade of the knife, only type A. Type 0 was found on the face of the circular saw blade, but no blood could be detected in the blade’s teeth. Based on spatter analysis, Tahir testified that the blood on the circular saw could not have been shed while the blade was moving. Tahir stated that blood spatters throughout the Interport offices indicated a violent struggle.
Motive evidence was introduced over defense objection. Frank Paul’s wife, Nancy, operated a Reedsburg, Wisconsin, airport which leased two airplanes from defendant. Airport employees testified that in 1979 and 1980 defendant telephoned and asked to speak to Nancy Paul when she was in Reedsburg. However, he did not telephone when she was away from Reedsburg, nor did he ask them about business. They further testified that defendant and Nancy Paul attended three or four airport-related public functions in Wisconsin; Frank Paul, the Paul children, and the Harbold children also attended these functions.
Gregory Wisniewski, an Interport employee, stated that defendant and Nancy Paul had been in Florida at the same time to purchase an antique automobile. On cross-examination, he recalled that Frank Paul had given him this information. Roy Palmer, an attorney who had represented defendant, testified that defendant and Nancy Paul arrived at Palmer’s apartment one evening in November or December of 1977, and remained there alone together after Palmer left. They were gone when Palmer returned at about 2 a.m. Defendant and Nancy Paul also attended a holiday season open house at Palmer’s apartment. Carol Yoshioka, an acquaintance, indicated that she and her date attended the theater and went to dinner with defendant and Nancy Paul one evening early in 1978. Over defense objection to hearsay, Yoshioka answered the question, “*** did Ralph Harbold and Nancy Paul ever make reference to the home that she owned up in *369Wisconsin?” and said, “Yes, they were talking about that they had been there.”
Gary Meisner testified that he flew Nancy Paul back to Chicago from Phoenix on March 31, 1981. Nancy Paul had told one of the airport employees that she wanted to move to Phoenix permanently. Meisner stated that over a four-day period immediately following his return he tried without success to reach defendant by telephone at his home and office. Defendant did not attend Frank Paul’s funeral.
Although he did not testify on his own behalf, nor did he call any witnesses, defendant was compelled to model the camel-colored sport-coat before the jury, and to display his left hand to the jury. Defendant introduced into evidence a study which contained updated population statistics associated with certain genetic markers, and a chart which incorporated the new statistics and the IDLE technician’s negative blood test results.
In closing argument, the prosecutor stated: “You can also look at the blood sample of the defendant, and look at the genetic markers across the board, and you will see that Ralph Harbold’s blood is on both rubber gloves, and on the pantyhose, it is on the left cuff of the sportcoat and it is also on the saw.” While discussing the chart introduced by the defense, the prosecutor stated: “What you see is that there is not one smallest piece of evidence that shows that Ralph Harbold is not guilty of this offense.” Although a defense objection to this comment was sustained, three similar statements were made, another objection was sustained and two more such comments were made before the prosecutor abandoned this line of comment.
In addition, the prosecutor argued that the motive testimony was “unrebutted and uncontradicted,” and corrected defendant’s characterization of Yoshioka’s testimony as placing Frank Paul with defendant and Nancy Paul at the Reedsburg residence. The prosecutor also stated:
“[I]n the hands of someone as competent as Mr. Schippers [defense counsel] *** [cross-examination] is as deadly as the knife in the hands of Ralph Harbold, or it is just like a scalpel in the hands of the plastic surgeon, to take this and then mold it into what he wants it to be. Just by asking questions, you can kind of transform things.”
The trial court’s instructions to the jury included Illinois Pattern Instruction (IPI), Criminal, No. 3.11 (2d ed. 1981), concerning impeachment with a prior statement. However, the court transposed the words “in this case” from the end of the first sentence to the beginning of the second. The court also included an instruction to the effect *370that defendant’s testimony should be evaluated in the same way as the testimony of other witnesses, although the defendant did not testify.
Following deliberations, the jury returned a verdict of guilty and defendant was sentenced to an extended term of 75 years’ imprisonment. Defendant appeals.
Opinion
Defendant first contends that the State produced insufficient evidence to establish his guilt beyond a reasonable doubt. Initially, we note that the evidence of defendant’s guilt is entirely circumstantial.
The jury’s determination of guilt will not be set aside by a reviewing court unless the determination is palpably contrary to the weight of the evidence or so unsatisfactory as to raise a reasonable doubt of guilt. (People v. Manion (1977), 67 Ill. 2d 564, 578, 367 N.E.2d 1313, cert. denied (1978), 435 U.S. 937, 55 L. Ed. 2d 533, 98 S. Ct. 1513.) Where a conviction rests solely upon circumstantial evidence, the guilt of the defendant must be so thoroughly established as to exclude every other reasonable hypothesis of innocence. (People v. Garrett (1975), 62 Ill. 2d 151, 163, 339 N.E.2d 753.) It is not required that the trier of fact be satisfied beyond a reasonable doubt as to each link in the chain of circumstances, so long as the evidence on the whole convinces the trier beyond a reasonable doubt of the guilt of the defendant. People v. Foster (1979), 76 Ill. 2d 365, 374, 392 N.E.2d 6.
Defendant argues that evidence of motive and blood type was inconclusive, resulting in a conviction based upon suspicion and probability. We consider issues of admissibility separately below, but we note here that to the extent defendant’s arguments go to the weight of evidence, he misconceives our role as reviewing court. We have carefully examined the record and analyzed the jury’s verdict in light of all the evidence. The State’s relevant and competent evidence permitted the jury to find or infer that the victim was stabbed to death in the course of a struggle on March 31, 1981; that the perpetrator was cut between the thumb and index finger and shed blood at the scene; that defendant’s hand was cut between the thumb and index finger some time after the morning of March 31 and before April 6, 1981; that defendant’s blood exhibited characteristics consistent with blood shed at the scene; that defendant knew the victim and had some connection with Nancy Paul; and that defendant displayed consciousness of guilt. We believe that the proof offered by the State, while not overwhelming, was not so unsatisfactory and unconvincing as to raise a serious doubt of defendant’s guilt.
*371Although the evidence was sufficient to sustain a conviction in a trial free from prejudicial error, our examination of the record reveals a number of trial errors which, considered together, deprived defendant of his constitutional right to a fair trial. Of course, overwhelming evidence would not lessen his entitlement to a fair trial (see People v. Weathers (1975), 62 Ill. 2d 114, 120-21, 338 N.E.2d 880), but the evidence in this case was close and circumstantial. “The fact that the evidence of [defendant’s guilt] is by no means of a conclusive character, requires us to consider the record to determine whether or not it is free from prejudicial error.” People v. Grizzel (1943), 382 Ill. 11, 20, 46 N.E.2d 78.
Defendant contends that he was deprived of his right to a fair and impartial trial by a pattern of improper prosecutorial comment. He asserts that the prosecutors undercut his presumption of innocence, cast aspersions on defense counsel’s methods, and mischaracterized the certainty of scientific evidence. The State maintains that the comments were waived, proper or cured by the trial court and that defendant was not unduly prejudiced.
In general, matters in evidence and reasonable inferences therefrom are proper subjects of comment to the jury. (People v. Warmack (1980), 83 Ill. 2d 112, 125-26, 413 N.E.2d 1254.) However, arguments which diminish the presumption of innocence are forbidden. (People v. Weinstein (1966), 35 Ill. 2d 467, 220 N.E.2d 432.) A reviewing court should assess arguments of counsel in the context of the whole trial, and should sustain a conviction unless it appears that improper comment substantially prejudiced the accused. People v. Tate (1970), 45 Ill. 2d 540, 545-46, 259 N.E.2d 791, cert. denied (1971), 401 U.S. 941, 28 L. Ed. 2d 222, 91 S. Ct. 944.
Here, the prosecutor repeatedly argued that no evidence showed defendant to be not guilty. The prosecutor stated: “What you see is that there is not one smallest piece of evidence that shows that Ralph Harbold is not guilty of this offense.” The trial court sustained the objection to this clearly improper comment. Again, the assistant State’s Attorney argued: “There wasn’t one single piece of evidence that went through the lab *** that in any way shows that Ralph Harbold is not guilty of this offense.” And again, “*** there isn’t one single thing around the entire area that shows he is not guilty.” And yet again, “*** I am going to bring before you all the evidence because when we bring all of the evidence to you I believe you can see there is nothing here that shows that he is not guilty.” Defendant’s objection to this comment was sustained, and the trial court admonished the jury that defendant need not prove his innocence. Thereafter, the as*372sistant State’s Attorney commented that, “the crime lab proved that absolutely nothing there showed that he is innocent,” and later stated, “Certainly, it doesn’t prove him not guilty.”
We can discern no purpose for this theme except to focus the jury’s attention on defendant’s failure to introduce evidence of innocence. The comments tended to shift the burden of proof to the defense and to diminish the presumption of innocence, and accordingly, were erroneous. (See People v. Giangrande (1981), 101 Ill. App. 3d 397, 402, 428 N.E.2d 503.) Furthermore, we believe that the repetition of the remarks after objections were sustained worked to the substantial prejudice of defendant. Our supreme court has held with respect to comments which undercut the presumption of innocence:
«*** fact that some of the remarks were repeated after an objection had been sustained by the court would indicate a disregard for the ruling of the court and the preservation of the constitutional rights of the accused, and constituted prejudicial error.” (People v. Grizzel (1943), 382 Ill. 11, 22, 46 N.E.2d 78.)
Accord, People v. Weinstein (1966), 35 Ill. 2d 467, 470-72, 220 N.E.2d 432.
We reject the State’s claim that the comments were invited by defendant’s argument concerning the inconclusive nature of the scientific evidence. In effect, the State asks us to countenance argument on the absence of proof of innocence because defendant attacked the strength of the State’s evidence, a position clearly repugnant to defendant’s presumption of innocence. The State’s position is particularly unpersuasive where defendant’s remarks were proper and based on testimony from the State’s experts, and where the prosecutor persisted in making improper remarks over sustained objections. We also reject the State’s contentions regarding waiver and cure. We believe that defendant’s objections and his specific reference to these remarks in his post-trial motion brought the matter to the attention of the trial court sufficiently to preserve it for review. (See People v. Clark (1982), 108 Ill. App. 3d 1071, 1077, 440 N.E.2d 387.) The constant repetition of this improper comment renders doubtful the effectiveness of the trial court’s admonitions. (See People v. Weinstein (1966), 35 Ill. 2d 467, 471, 220 N.E.2d 432.) We are unable to conclude that these comments which diluted one of the defendant’s fundamental protections were harmless.
Defendant’s complaints that the prosecutor cast aspersions on counsel’s methods and mischaracterized the certainty of scientific evidence are less compelling. While we question the comparison between counsel’s use of cross-examination and defendant’s use of the murder *373weapon, the context indicates a focus on the testimony of witnesses as opposed to the content of questions, and so the remark did not rise to the level of prejudicial error. Although a contemporaneous objection may have entitled defendant to a more correct statement (see People v. Giangrande (1981), 101 Ill. App. 3d 397, 402-03, 428 N.E.2d 503), the prosecutor’s comment that defendant’s blood “matched” blood found at the scene was not improper, because “match” connotes equality, similarity or suitability (see Webster’s Third New International Dictionary 1391 (1976)), and so understood, the remark was within the limit of testimony and inference. Similarly, the prosecutor’s statement that the genetic marker evidence showed defendant’s blood to be on the gloves, sportcoat and saw, while overstating the scientific evidence, could be inferred from all the evidence, and so did not constitute prejudicial error.
Defendant next contends that the admission of hearsay evidence violated his sixth amendment right of confrontation and deprived him of a fair trial. He asserts that Carol Yoshioka’s testimony to the statements of Nancy Paul denied him the opportunity to confront Nancy Paul and cross-examine her before the jury. The State argues that the statements were admissible as nonhearsay, not for the truth of the matter asserted, or in the alternative, as admissions, excepted from the hearsay rule. The testimony at issue appears in the record as follows:
“Q. [Assistant State’s Attorney] During the time you were having dinner at Nick’s Fishmarket, did Nancy Paul ever refer to any property that she owned or operated?
A. [Ms. Yoshioka] Yes.
Q. And what was that?
A. She said that there was a summer home and that she owned an airport that she was trying to bring business up there and develop the area.
Q. And during the course of that conversation, did Ralph Harbold and Nancy Paul ever make reference to the home that she owned up in Wisconsin?
A. Yes, they were talking about that they had been there.
Q. And do you recall what reference they made to in terms of when they had been there?
A. They had referred to the time there was some plumbing problems up there when they had been there.
Q. Plumbing problems?
A. Yes.”
The record belies the State’s position that the statements were *374nonhearsay. The nonhearsay purpose suggested by the State, to show that defendant and Nancy Paul held themselves out as a couple, was not accomplished by the fact of the statement; the jury must have believed that defendant and Nancy Paul spent time together in Wisconsin in order to infer from the statements that they held themselves out as a couple. The State’s nonhearsay purpose was accomplished by Yoshioka’s nonhearsay testimony to the participants, activities and circumstances of the evening; testimony to the statements added only the content of the statements. In addition, the manner in which the testimony was elicited, notably the prosecutor's repetition of “plumbing problems,” was calculated to impress upon the jury the truth of the matter asserted and the State argued that content of the statements to the jury in closing argument.
The record also negates the State’s contention that these statements were' admissions. A statement must be made or adopted by the defendant to be introduced as an admission of a criminal trial. (People v. Simpson (1977), 68 Ill. 2d 276, 281-82, 369 N.E.2d 1248.) Where a statement affecting defendant is made in circumstances affording an opportunity to reply, and where others similarly situated would naturally reply, defendant may be deemed to have adopted the statement by silence. (People v. Pfanschmidt (1914), 262 Ill. 411, 448, 104 N.E. 804.) This record is inadequate to support an argument that defendant either made or adopted statements concerning the Reedsburg residence. The form of the questions and answers failed to identify any statement to defendant. The casual circumstances of the conversation were not such that defendant or others similarly situated would naturally reply. Everyday conversation is rife with ambiguous, exaggerated and humorous expression, and the participants are seldom inclined to demand rigorous accuracy or to subject one another to critical scrutiny. We hold that the trial court erred when it overruled defendant’s objection to this hearsay, and in our opinion, this error deprived defendant of another substantial constitutional protection.
The sixth amendment to the United States Constitution provides in part: “In all criminal prosecutions, the accused shall enjoy the right *** to be confronted with the witnesses against him.” (U.S. Const., amend. VI.) The confrontation clause applies in State proceedings, as a matter of fourteenth amendment due process. (Pointer v. Texas (1965), 380 U.S. 400, 13 L. Ed. 2d 923, 85 S. Ct. 1065.) The Illinois Constitution provides in part: “In criminal prosecutions, the accused shall have the right *** to meet the witnesses face to face ***.” (Ill. Const. 1970, art. I, sec. 8.) These provisions are essentially coextensive, and seek to secure for the accused the opportunity of cross-*375examination. (People v. Tennant (1976), 65 Ill. 2d 401, 408, 358 N.E.2d 1116, cert. denied (1977), 431 U.S. 918, 53 L. Ed. 2d 229, 97 S. Ct. 2184.) The United States Supreme Court recently stated that the confrontation clause,
“envisions
‘a personal examination and cross-examination of the witness in which the accused has an opportunity, not only of testing the recollection and sifting the conscience of the witness, but of compelling him to stand face to face with the jury in order that they may look at him, and judge by his demeanor upon the stand and the manner in which he gives his testimony whether he is worthy of belief.’ [Citation.]
These means of testing accuracy are so important that the absence of proper confrontation at trial ‘calls into question the ultimate “integrity of the fact-finding process.” ’ [Citations.]” Ohio v. Roberts (1980), 448 U.S. 56, 63-64, 65 L. Ed. 2d 597, 606, 100 S. Ct. 2531, 2537-38.
Recognizing that face-to-face confrontation is not always possible, the court in Ohio v. Roberts (1980), 448 U.S. 56, 65 L. Ed. 2d 597, 100 S. Ct. 2531, set forth a two-part standard for conforming hearsay evidence with the goal of the confrontation clause:
“[W]hen a hearsay declarant is not present for cross-examination at trial, the Confrontation Clause normally requires a showing that he is unavailable. Even then, his statement is admissible only if it bears adequate ‘indicia of reliability.’ ” 448 U.S. 56, 66, 65 L. Ed. 2d 597, 608, 100 S. Ct. 2531, 2539.
In the instant case, the State sought to introduce Nancy Paul’s statements as proof of defendant’s motive. Therefore, the State was required to produce her at trial or show that she was unavailable. There is nothing in the record to indicate that the State made any effort to secure Nancy Paul’s presence at trial. Moreover, the record actively indicates that the statements were unreliable and prejudicial. The conversation took place four years before the witness testified, and the circumstances were not such as to assure the accuracy of the statements or to impress the witness with a full or accurate recollection. Indeed, the witness exhausted her recollection of the subject in the passage quoted above from direct examination.
As a result of the erroneous admission of this hearsay, the jury was exposed to the bare fact that defendant and Nancy Paul spent time together at a cottage in Wisconsin, proved by Nancy Paul’s words, through a witness whose lack of personal knowledge of underlying events prevented any qualifying or explanatory information *376from being revealed. The State thus accomplished precisely what the confrontation clause forbids: it presented Nancy Paul’s testimony against defendant and circumvented defendant’s right to cross-examine Nancy Paul. The presentation of motive evidence from the mouth of defendant’s purported paramour clearly prejudiced defendant, and viewed in the light of remaining motive evidence and other trial errors, we cannot say that this error was harmless.
Apart from hearsay, defendant argues that the motive evidence was irrelevant and should have been excluded in its entirety. Generally, any evidence which tends to show that the accused had a motive to kill the victim is relevant in a homicide case, but “such evidence to be competent must, at least to a slight degree, tend to establish the existence of the motive relied upon or alleged.” (People v. Gougas (1951), 410 Ill. 235, 238-39, 102 N.E.2d 152.) The State’s theory of motive in this case is that defendant admired the victim’s wife, Nancy Paul, and feared that she would leave the Chicago area. We agree with defendant that the State’s theory of motive is flawed.
The fact that Nancy Paul wanted to move to Phoenix could indicate motive only if defendant knew of her plans and had some reason to believe that killing Frank Paul would alter those plans. The State maintains that defendant’s knowledge may be inferred from the fact that one of the Reedsburg airport employees knew; the State makes no attempt in its brief to explain how killing Frank Paul could have altered Nancy Paul’s intention to move. However, the State offered this closing argument for the jury’s consideration:
“[Defendant] had a relationship with Nancy Paul. He had a crazed relationship because Nancy Paul was going to leave the Chicago area, and in despair he could do only one thing. The thing that was left to him was to kill Frank Paul to prevent Nancy from leaving.”
In People v. Gougas (1951), 410 Ill. 235, 102 N.E.2d 152, the State introduced evidence that the accused was the beneficiary of the victim’s insurance policy, but made no attempt to show that the accused knew of the policy. The court reversed defendant’s murder conviction, stating that evidence of the policy could do “little more than excite a suspicion of guilt.” (410 Ill. 235, 240, 102 N.E.2d 152.) Here, evidence that Nancy Paul wanted to move to Phoenix at least confused and distracted the jury, and we believe that the State’s use of the evidence excited suspicion of guilt.
The “Phoenix argument” was crucial to the State’s motive theory because the evidence failed to indicate a reason to kill Frank Paul in 1981. The evidence indicated that in late 1977 and early 1978, defend*377ant and Nancy Paul were seen together in a variety of social settings, that they were in Florida at some unnamed time, and that in 1979 and 1980, they attended airport-related public functions along with their families. This evidence was relevant because it tended to show some relationship from which the jury could have inferred motive. (See People v. Branion (1970), 47 Ill. 2d 70, 77, 265 N.E.2d 1, cert. denied (1971), 403 U.S. 907, 29 L. Ed. 2d 683, 91 S. Ct. 2213.) However, the probative value of the evidence suffered from its remoteness in time; evidence that the two spent time alone together predated the crime by three years or more. Our supreme court recently stated, “A love affair whose embers have long since cooled is not exactly a motive for murder. Its value is likely to be outweighed by the unfair prejudice against the defendant the evidence engenders.” (People v. Weaver (1982), 92 Ill. 2d 545, 562, 442 N.E.2d 255; see also People v. Holtz (1920), 294 Ill. 143, 160-62, 128 N.E. 341 (fact that incidents of marital problems occurred one to three years before crime was one factor in court’s conclusion that there was no evidence of motive).) The State bootstrapped the more remote motive evidence into the present with the Phoenix argument. In our opinion, defendant was entitled to have the existence of motive determined without this incompetent evidence and the accompanying inflammatory argument.
We must conclude that the introduction of irrelevant and hearsay motive evidence constituted substantial prejudice to defendant in this case. Evidence of motive is especially important where a conviction rests on circumstantial evidence. (People v. Holtz (1920), 294 Ill. 143, 155, 128 N.E. 341.) “[A]ny reasonable hypothesis of innocence is likely to be discounted progressively by degrees as evidence of possible motives compounds.” (People v. Weaver (1980), 90 Ill. App. 3d 299, 303, 412 N.E.2d 1353, aff’d (1982), 92 Ill. 2d 545, 442 N.E.2d 255.) We cannot say what the jury would have done if only relevant and competent motive evidence had been submitted, hence his right to a fair trial requires a new trial.
Defendant next contends that physical and scientific evidence should have been excluded because its probative value was outweighed by prejudice. He argues that the blood-stained items of physical evidence were not adequately connected to him and to the crime, and that the blood analyses rendered scientifically inconclusive results. The State rejoins that the physical evidence was adequately connected and that the blood test results were admissible even though inconclusive. Despite our reservations concerning some of the analytic results, we agree with the State that the evidence at issue was admissible. However, we find that testimony to the statistical probability of *378matching blood constituted plain error.
Physical evidence must be connected to both the defendant and the crime to be admissible, but the inquiry is one of relevance and not of weight. (People v. Free (1983), 94 Ill. 2d 378, 415-17, 447 N.E.2d 218, cert. denied (1983), 464 U.S. _, 78 L. Ed. 2d 175, 104 S. Ct. 200.) Accordingly, it has been held that an item is connected to the crime if it is suitable for the commission of the crime, and it is not necessary to show that the item was actually used. (People v. Miller (1968), 40 Ill. 2d 154, 159, 238 N.E.2d 407, cert. denied (1968), 393 U.S. 961, 21 L. Ed. 2d 375, 89 S. Ct. 401.) Similarly, a defendant was connected to weapons found at a public dump where “a person answering defendant’s general description had been seen in the vicinity of the dump after the killing.” People v. Ashley (1960), 18 Ill. 2d 272, 280, 164 N.E.2d 70, cert. denied (1960), 363 U.S. 815, 4 L. Ed. 2d 1157, 80 S. Ct. 1255, explaining People v. Collins (1940), 375 Ill. 253, 31 N.E.2d 295.
In this case, a sportcoat, pantyhose, penlight, gloves and knife were found within a block of the crime scene, within hours of the crime, stained with blood, many characteristics of which were consistent with blood shed at the scene. Keys found in the sportcoat’s pockets turned the locks at Interport, and according to experts, the knife was suitable for commission of the crime. The sportcoat, pantyhose and cut latex glove were connected to defendant in that several characteristics of the blood on these items were consistent with defendant’s blood. Further, the cut on the glove corresponded to the wound on defendant’s left hand. We hold that these items were sufficiently connected to one another, to the crime and to defendant so as to be admissible.
A more difficult problem is presented by expert testimony concerning the analysis of blood stains and whole blood samples, and the conclusions drawn therefrom. As a rule, expert testimony is admissible where the expert has knowledge or experience not common to laymen which renders his or her opinion an aid to the jury in determining a fact in issue. (People v. Jennings (1911), 252 Ill. 534, 550, 96 N.E. 1077.) A scientific opinion must be based upon principles which have gained general acceptance in the scientific community. (See People v. Baynes (1981), 88 Ill. 2d 225, 430 N.E.2d 1070.) Illinois courts have taken judicial notice of the scientific reliability of the ABO blood grouping system. (See, e.g., People v. Gillespie (1974), 24 Ill. App. 3d 567, 573, 321 N.E.2d 398.) In fact, in People v. Bush (1981), 103 Ill. App. 3d 5, 430 N.E.2d 514, the court took judicial notice of the ABO system’s reliability where expert testimony was based on a combina*379tion of ABO and Esterase D (EsD) typing. We would be remiss in our duty if we disposed of the issue of scientific acceptance on this authority.
Scientific acceptance of the ABO system provides little basis for accepting evidence of genetic marker detection by electrophoresis or opinions deduced from such information. The ABO system has been studied and refined since the turn of the century, and is widely recognized because accurate ABO typing is essential in order to administer blood transfusions safely. (See People v. Gillespie (1974), 24 Ill. App. 3d 567, 572-73, 321 N.E.2d 398, and sources cited therein.) By contrast, other genetic markers have only recently been isolated, do not affect transfusion compatibility, and are tested by electrophoresis, a technique entirely different from the one used clinically to determine ABO type. (See Robinson v. State (1981), 47 Md. App. 558, 574-75, 425 A.2d 211, 220; Forensic Science Handbook 340-42 (R. Saferstein ed. 1982).) There is some dispute among scientists as to the accuracy of results derived from electrophoretic testing of dried blood stains. (See State v. Washington (1981), 229 Kan. 47, 49-52, 622 P.2d 986, 988-91.) In an extensive review of scientific and legal authorities on the subject, one scholar has concluded that “the forensic tests for the detection of genetic markers in blood have not been proven to be reliable and should not be admitted into criminal trials.” Jonakait, Will Blood Tell? Genetic Markers in Criminal Cases, 31 Emory L.J. 833, 912 (1982) (hereinafter referred to as Jonakait, Genetic Markers).
In this case, genetic marker evidence seems to have been received on the basis of Mark Stolorow’s statement that these techniques are used in crime labs nationwide. Both of the cases we have found which consider the matter, State v. Washington (1981), 229 Kan. 47, 622 P.2d 986, and Robinson v. State (1981), 47 Md. App. 558, 425 A.2d 211, found scientific acceptance based on widespread use in crime labs. While this fact is certainly relevant to scientific acceptance (see e.g., People v. Jennings (1911), 252 Ill. 534, 546-49, 96 N.E. 1077 (fingerprint evidence)), we do not believe that use in crime labs alone can justify admission of evidence in the face of a bona fide scientific dispute. (Cf. People v. Baynes (1981), 88 Ill. 2d 225, 430 N.E.2d 1070 (polygraph evidence).) Not every useful investigative tool is necessarily admissible in a criminal trial. (People v. Monigan (1979), 72 Ill. App. 3d 87, 92, 390 N.E.2d 562.) One of the most important factors in Illinois’ rejection of polygraph analysis, the subjectivity of interpretation of test results (see People v. Baynes (1981), 88 Ill. 2d 225, 236, 430 N.E.2d 1070), is also involved in genetic marker analysis. (See State v. Washington (1981), 229 Kan. 47, 50-52, 622 P.2d 986, 989-91; Jona*380kait, Genetic Markers 893-95.) We believe that cautious consideration of this evidence is indicated before the reliability of ABO grouping is assumed to apply to it as a matter of law.
In addition to these general indictments against automatic recognition of genetic marker evidence, our examination of the record reveals several specific points of expert testimony which seem to conflict with other scientific authorities. For example, Stolorow stated categorically that all of the genetic markers have functions in the body apart from their use by crime laboratories, whereas Jonakait cites authority to indicate that science is presently unable to assign a precise function to each genetic marker. (Jonakait, Genetic Markers 849.) Stolorow also stated: “We know that there is enough genetic information in a person’s blood stain to isolate him alone in the population as the only possible donor,” but that “technology has not yet reached that stage.” This does not accurately represent current scientific understanding: scientists believe identical twins have the same blood groupings hence it may never be possible to “identify” blood in the same way that fingerprints can (Jonakait, Genetic Markers 833 N.2, citing Terasaki, Resolution by HLA Testing of 1000 Paternity Cases Not Excluded by ABO Testing, 16 J. Fam. L. 543, 549 (1978)), and some experts believe that the current approach to genetic marker analysis is inherently incapable of individualization. Jonakait, Genetic Markers 833-34, N.3, citing Sweet & Elvins, Studies by Crossed Electroimmunodiffusion on the Individuality and Sexual Origin of Bloodstains, 21 J. Forensic Sci. 498, 505 (1976).
Other inconsistencies in the blood evidence disturb us. Podlecki testified that there are six varieties of the Erythrocite Acid Phosphatase (EAP) enzyme; testimony in another case indicates that there are five varieties of this enzyme. (State v. Washington (1981), 229 Kan. 47, 49, 622 P.2d 986, 988.) The test results did not uniformly meet the expert’s expectations. A person with type A blood who received five pints of type 0 blood by transfusion would most likely exhibit both types, according to Stolorow, and the least likely test result would show type 0 alone. Yet when Frank Paul’s blood tested as type 0, the experts simply disregarded this result. Finally, the population frequency statistics, upon which all of the State’s experts based their opinions, came from an apparently unpublished study. In People v. Gillespie (1974), 24 Ill. App. 3d 567, 575-77, 321 N.E.2d 398, the court admitted published population frequency figures under the “Learned Treatise” exception to the hearsay rule. The rationale for admitting published statistical compilations does not extend to unpublished studies, and though Stolorow claimed personal knowledge of *381the study, we suspect that he relied upon others for certain aspects of sampling, analysis, compilation and validation of results.
We believe that this testimony tended to exaggerate the conclusiveness of the scientific techniques used in this case. We imply no criticism of the State’s experts, for the Illinois Department of Law Enforcement has helped to pioneer genetic marker detection in this country, and may well have a sound basis for zealous advocacy of these techniques. If defendant had raised these apparent inconsistencies, the experts might have qualified their testimony or given a satisfactory explanation. The defense corrected a flaw in the State’s evidence in at least one respect, by introducing published population statistics. We cannot hold that electrophoretic detection of genetic markers in field conditions is unreliable as a matter of law, but we believe that some questions as to scientific acceptance of the technique remain unanswered in this record and in the case law. Should the State elect to retry this case, we believe that the defendant and the people would be well served by a careful evaluation of this evidence, mindful to the oft-quoted passage from Frye v. United States (D.C. App. 1923), 293 F. 1013, 1014:
“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 principal must be recognized, and while 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 gainful general acceptance in the particular field in which it belongs.”
One glaring defect in the State’s evidence warrants consideration under Supreme Court Rule 615(a) (87 Ill. 2d R. 615(a)), which provides for review of plain error affecting substantial rights where the evidence is closely balanced, despite the defendant’s failure to bring the matter to the trial court’s attention. Mark Stolorow testified that “the chances of selecting any two people at random from the population and having them accidently have identical blood types in each one of these factors is less than one in 500, that is, what we call the probability of an accidental match is less than one in 500.” We believe that this testimony was irrelevant, beyond the proper scope of expert opinion, and highly prejudicial to defendant.
Illinois courts have not addressed the propriety of introducing statistical probabilities for the purpose of establishing a historical fact. The reasoning of the California Supreme Court in People v. Collins *382(1968), 68 Cal. 2d 319, 438 P.2d 33, 66 Cal. Rptr. 497, has been cited with approval in all of the subsequent cases which have considered the issue. Collins involved an interracial couple accused of robbery. The State’s case was mainly circumstantial, and in an attempt to bolster the witnesses’ uncertain identification, the prosecutor called a mathematician to the stand. The mathematician testified to the “product rule” which states that the probability of the joint occurrence of a number of independent events is equal to the product of the individual probability of the occurrence of each event. The prosecutor then provided the witness with individual probabilities for the characteristics in evidence. Combining these factors under the product rule, the prosecutor concluded that there was “one chance in 12 million that any couple possessed the distinctive characteristics of the defendants.” 68 Cal. 2d 319, 325, 438 P.2d 33, 37, 66 Cal. Rptr. 497, 501.
The Collins court reversed defendants’ convictions, stating:
“[T]he prosecution’s introduction and use of mathematical prob-' ability statistics injected two fundamental prejudicial errors into the case: (1) The testimony itself lacked an adequate foundation both in evidence and in statistical theory; and (2) the testimony and the manner in which the prosecution used it dis-' tracted the jury from its proper and requisite function of weighing the evidence on the issue of guilt, encouraged the jurors to rely upon an engaging but logically irrelevant expert demonstration, foreclosed the possibility of an effective defense by an attorney apparently unschooled in mathematical refinements, and placed the jurors and defense counsel at a disadvantage in sifting relevant fact from inapplicable theory.” 68 Cal. 2d 319, 327, 438 P.2d 33, 38, 66 Cal. Rptr. 497, 502.
The foundational defects in Collins were grave. The prosecutor in .that case invented figures which he terms conservative and applied them to mutually dependent variables to reach his conclusion. The questions we raise concerning the scientific acceptance of detection techniques are preliminary to the statistical flaws criticized in Collins. The're is empirical support for the independent occurrence of genetic markers, and there are published population frequency studies. (See Forensic Science Handbook (R. Safterstein ed. 1982).) Therefore, assuming that scientific acceptance can be shown, a statistically valid probability figure could be computed. Illinois courts have tacitly approved one application of the product rule by permitting experts to testify to the percentage of the population exhibiting certain combinations of blood characteristics. (People v. Bush (1981), 103 Ill. App. 3d 5, 430 N.E.2d 514; People v. Gilliespie (1974), 24 Ill. App. 3d 567, 321 *383N.E.2d 398.) Here, however, the State went much farther and expressed the similarity between the perpetrator and defendant in terms of the likelihood of a random match.
Our concern is more with the impact of the probability statistic upon the jury than the foundation for the testimony. The statistic was patently irrelevant. Probability theory is necessarily silent on the crucial question before the jury: Of the relatively small percentage of the population with consistent characteristics, which one, if any, committed the crime? (People v. Collins (1968), 68 Cal. 2d 319, 330, 438 P.2d 33, 40, 66 Cal. Rptr. 497, 504.) Jurors would be hard pressed to explain how the l-in-500 chance of an accidental match did not equate with a l-in-500 chance that defendant was innocent. (See Tribe, Trial by Mathematics: Precision and Ritual in the Legal Process, 84 Harv. L. Rev. 1329, 1355 (1971).) Of course, the statistic means nothing of the sort. Absent a sound basis to limit the number of possible defendants, the defendant here is but one of thousands of people who share these same characteristics. Legion possibilities incapable of quantification, such as the potential for human error or fabrication, or the possibility of a frame-up, must be excluded from the probability calculation.
In State v. Carlson (Minn. 1978), 267 N.W.2d 170, the supreme court of Minnesota considered random match probabilities of l-in-800 and l-in-4500 on the identification issue. Holding that the foundation was proper and “based upon empirical scientific data of unquestioned validity,” the court concluded:
“Testimony expressing opinions or conclusions in terms of statistical probabilities can make the uncertain seem all but proven, and suggest, by quantification, satisfaction of the requirement that guilt be established ‘beyond a reasonable doubt.’ ” (267 N.W.2d 170, 176.)
(Accord, State v. Boyd (Minn. 1983), 331 N.W.2d 480.) We believe that testimony to statistical probabilities encouraged the jury to disregard evidential risks traditionally weighed in determining guilt or innocence, and focused unfairly upon a numerical conclusion. (See People v. Collins (1968), 68 Cal. 2d 319, 330-31, 438 P.2d 33, 40-41, 66 Cal. Rptr. 497, 504-05.) As such, we find that the testimony violated one of the primary requirements of expert opinion, that the opinion be an aid to the jury. (See People v. Stapelton (1972), 4 Ill. App. 3d 477, 480-81, 281 N.E.2d 76.) In light of the closeness of this circumstantial case, we cannot say that this improper testimony, which gave a false impression of precision in the measurement of guilt, did not affect the jury’s deliberations.
*384 Defendant next contends that he was irreparably prejudiced by the prosecutor’s intentional misconduct in eliciting testimony concerning weapons found at his residence. The State maintains that the references were waived, proper, or cured by the trial court.
It is well settled that the prosecution may not initiate an inquiry into defendant’s evil character for the purpose of establishing guilt. (Michelson v. United States of America (1948), 335 U.S. 469, 93 L. Ed. 168, 69 S. Ct. 213; People v. Lindgren (1980), 79 Ill. 2d 129, 402 N.E.2d 238.) The rationale for the rule against character is that “it is said to weigh too much with the jury and to so overpersuade them as to prejudge one with a bad general record and deny him a fair opportunity to defend against a particular charge.” (Michelson v. United States of America (1948), 335 U.S. 469, 476, 93 L. Ed. 168, 174, 69 S. Ct. 213, 218.) In People v. Gregory (1961), 22 Ill. 2d 601, 177 N.E.2d 120, our supreme court granted defendants a new trial based on evidence of unrelated crimes coupled with improper argument. The court noted, as a corollary to the rule against character evidence, that the State’s Attorney’s duty to safeguard the rights of all the people extends to criminal defendants, and the prosecutor’s failure to delete prejudicial and improper matters from a statement constitutes a breach of that duty. 22 Ill. 2d 601, 604, 177 N.E.2d 120.
At trial in this matter, the assistant State’s Attorney asked a detective of 11 years’ experience what weapons were found pursuant to the execution of a search warrant. The detective responded: “we find [sic\ a .38 caliber snub-nosed revolver.” Defendant objected, and the trial court heard arguments in chambers. The assistant State’s Attorney offered proof stating “in his basement was a bullet-proof vest and two gas masks and I believe that is indicative of the type of person that Ralph Harbold is.” We find that the prosecutor violated his duty to safeguard defendant’s right to be judged solely with respect to the crime charge, and demonstrated disrespect for defendant’s right to a fair and impartial trial. (See People v. Clark (1983), 114 Ill. App. 3d 252, 256, 448 N.E.2d 926.) The issue was not waived. Defendant made his motion for a mistrial in chambers, and renewed the motion post-trial. Although the trial court admonished the jury to disregard any reference to a weapon, the admonition followed the reference by several minutes. Under different facts, even intentional misconduct such as this might be considered harmless. In this case, however, the prosecution’s consistent tactic of bolstering its case with irrelevancy militates against a finding of harmless error. We are not convinced beyond a reasonable doubt that this reference had no effect on the jury’s resolution of the issues.
*385We decline defendant’s invitation to comment upon the relevance of the knives found in the course of the search. We noted earlier that an item suitable for commission of the crime is usually admissible. (See People v. Miller (1968), 40 Ill. 2d 154, 238 N.E.2d 407, cert. denied (1968), 393 U.S. 961, 21 L. Ed. 2d 375, 89 S. Ct. 401. We note here that the subject was raised by defense counsel and that no objection was raised at trial.
Finally, defendant contends that the trial court erroneously instructed the jury. He asserts that the addition of the words “in this case” to the impeachment instruction indicated to the jury that Marlene Harbold’s testimony had been' impeached. We disagree. The words “in this case” were merely transposed from the end of the first sentence to the beginning of the second. The transposition had no substantive effect upon the instruction.
Defendant also contends that the trial court erred in instructing the jury as to defendant’s testimony, where defendant did not testify. Notably, the written instruction was stricken, and both oral and written instructions submitted to the jury properly apprised them that the defendant’s failure to testify should not be considered. Defendant submitted the instruction complained of, and defense counsel expressly waived any resulting error. We trust that this mishap will be avoided in a new trial.
In sum, we hold that the prosecutor undercut defendant’s presumption of innocence in closing argument, that the trial court erred in admitting hearsay motive testimony, that the prosecutor introduced irrelevant motive evidence accompanied by inflammatory argument, that testimony to an irrelevant probability statistic was improperly received, and that the prosecutor unfairly injected defendant’s character into the trial. We believe that defendant’s right to a fair trial was consistently compromised by irrelevancy and error which enhanced the State’s case. We cannot say beyond a reasonable doubt that defendant was not prejudiced by the cumulative effect of error in this close, circumstantial case. Therefore, the judgment of the trial court is reversed, and the cause is remanded for a new trial.
Reversed and remanded.
SULLIVAN and WILSON,* JJ., concur.