The defendant was convicted by a jury on August 20, 1981, of murder in the first degree for the fatal shooting of Robert W. LaMonica. The shooting occurred shortly after midnight on May 31, 1980, outside LaMonica’s Braintree apartment. The defendant was sentenced to a term of life imprisonment and now appeals his conviction. G. L. c. 278, § 33E.

Weichell claims that the trial judge erred in (1) denying his motions in limine to exclude certain evidence relating to motive; (2) granting the Commonwealth’s motion in limine to exclude evidence which tended to show that third parties had a motive to commit the crime; (3) refusing to exclude a “mugshot” photograph of the defendant’s profile; (4) permitting the Commonwealth to introduce in evidence a composite drawing; (5) allowing the Commonwealth to introduce in evidence an enlarged copy of a photograph of the defendant taken by the Braintree police at the time of his arrest; and (6) excluding photographs of the scene of the crime and the testimony of the photographer who took them. *64Weichell contends also that his conviction was against the weight of the evidence. He urges us to exercise our power under G. L. c. 278, § 33E, to order a new trial or otherwise to mitigate punishment. There was no error, and we find no reason to exercise our power under G. L. c. 278, § 33E.

1. Evidence. We summarize the evidence presented to the jury.

a. Motive. On May 18, 1980, Thomas Barrett and the defendant approached Francis Shea on a street in South Boston. Barrett and Shea had words and began to fight. They wrestled for several minutes until Barrett locked Shea in a “choke hold.” Shea “blacked out” but regained consciousness soon after. The fight attracted much attention, and several of Shea’s friends, including the victim, arrived. He helped Shea to his feet.

A heated argument developed between Shea’s friends on one side and Barrett and the defendant on the other side. Shea told Barrett that he would kill him. The defendant then stepped up to Shea and told him that if Shea killed Barrett, he himself would kill Shea and that “they’ll never find [your] body.” With these words, the defendant and Barrett turned away and left. Shea, LaMonica, and other friends of Shea, Dennis King, and Chuckie Carr, retired to the house of Shea’s brother. Shea was later treated at a hospital for lacerations on the right side of his head.

During the remainder of that day, LaMonica uttered to others several threats relating to Barrett and the defendant. He told Shea that he wanted to retaliate and said, “They picked on the wrong people this time. We are going to kill him.” Dennis King heard him say that they should “go after them. They messed with the wrong people.” LaMonica began work late in the afternoon. He returned to his apartment an hour and one-half after finishing work; he was “upset” and he had been drinking. He told his paramour, Maureen A. Connolly, that “[m]e and my friends, we’re going to get him, and we’re going to kill him.” Connolly testified that LaMonica was referring to both Barrett and the defendant.

*65The fight appears to have stirred up the friends of Shea and Barrett. On May 20, 1980, Barrett, the defendant, and a third person came to the house of Francis Shea’s brother, Richard. As Francis Shea watched them from the roof, the defendant told Dennis King and Richard Shea that he wanted “to know what Frankie wants to do. I have a brother dead, I have a brother in jail. I’m not going to wait. I’m going to act first.”

The next day Francis Shea saw the defendant and La-Monica arguing. He did not hear the words uttered, but testified that the defendant was “pointing his finger in Robert LaMonica’s face and stepping up and down the sidewalk.” A couple of days later, King, the Shea brothers, and Francis’s two sons came upon the defendant, Barrett, and a third man. Richard Shea challenged Barrett to a fight. The defendant turned around and replied, “No. Bring it down. We aren’t going to let this go.”

b. The shooting. LaMonica worked for the Boston Water and Sewer Commission. He worked from 4 p.m. to midnight. He would usually drive straight home from his job to his apartment, customarily arriving there between 12:15 a.m. and 12:30 a.m. He would turn off Faxon Street to park his car in a parking lot adjacent to his apartment building. Faxon Park is across from the entrance to the parking lot.1 LaMonica followed this routine on the morning of May 31, 1980. He parked and got out of his car. Four shots were fired, two of them hitting LaMonica. A bullet entered through his neck and penetrated the brain. A second bullet entered his back and lodged in his right rib cage. LaMonica died in the parking lot.

c. Identification. Shortly before midnight on May 30, 1980, John Foley, Jean Castonquay, Frederick Laracy, and Lisa Krause went to Faxon Park, after attending a drive-in movie together. Foley testified the group had been drinking and that he had consumed four or five beers during the *66movies. At 12:15 a.m., Foley was walking away from a wooded area of the park. He heard four “bangs” and saw a man run out of the parking lot and turn up Faxon Street to a waiting car. Krause screamed. The man looked toward the group briefly but continued running. Foley testified that he had a full-face view of the man for approximately one second as the man passed under a street light. Foley and Laracy went across Faxon Street to the parking lot where they found the body of the victim on the ground. The police arrived shortly thereafter.

Foley described to the police the man he saw running as being five feet, nine inches tall, 175 pounds, wearing jeans and a pullover shirt.2 He said that the man had dark curly hair, bushy eyebrows, and sideburns. He also stated that the man had a slightly crooked nose, “as if it had been broken.” At trial, he identified the defendant as the man he saw running that night.

Later that morning, Foley assisted Detective Wilson of the Braintree police department in making a composite drawing. After indicating that he could not draw a face by himself, Foley gave Wilson a general description. With the aid of an Identikit, Wilson and Foley assembled a composite. Foley examined the composite and asked for changes. Wilson then changed elements of the composite and put together a different face. Wilson used a pencil to alter the nose. After Foley altered the hair style, he declared that the composite “looks like him.” A photostatic copy of the composite was introduced in evidence at trial.

The next day, Foley was shown an array of nine photographs. He picked the defendant’s picture as “a pretty good likeness” of the man. Several months later, he again identified the defendant’s photograph out of the same array but which now included one additional photograph.

On June 12, 1980, two State troopers, Foley, and the victim’s two brothers drove through the streets of South Boston *67in a van. The LaMonicas gave directions, but did not speak to Foley. Eventually, Foley picked, out of a group of young men, an individual whom he thought was the man he saw running. The van was driven around the corner and passed the group for a second time. Foley stated, “That’s the guy.” A State trooper took a photograph of the individual which was introduced in evidence and identified as a photograph of the defendant.

Jean Gastonquay also testified that she heard four shots and saw a man running. At trial, she was unable to say whether the defendant was the man she saw. Moments later, she tentatively identified another person sitting in the back of the courtroom as the man.3 On three occasions Gastonquay was shown the same array of photographs as Foley, but was unable to pick out any one photograph. Instead, she picked out two or three photographs each time, always including that of the defendant. Neither Laracy nor Krause made any identification.4

d. The defendant’s case. At trial, the defendant’s counsel, through cross-examination, attempted to bring out whatever discrepancies existed in Foley’s testimony. He emphasized that Foley had indicated that the man he saw running had thick sideburns and bushy eyebrows. Foley admitted, however, that the defendant’s eyebrows were different. It also appears that the defendant did not have any sideburns. Despite some evidence to the contrary, the jury could have concluded that the defendant had curly hair at the time of the murder. The defendant also attempted to show that the lighting in the area was poor5 and that the *68identification process was unreliable. The defendant did not testify.

The defendant also sought to establish a defense of alibi. Three witnesses testified on his behalf. One witness’s testimony placed the defendant in downtown Boston until midnight. The other witnesses placed the defendant at the Triple O Lounge in South Boston at, or shortly after, the time of the shooting.

In rebuttal, the Commonwealth introduced evidence that the defendant could have left downtown Boston shortly before midnight and driven to LaMonica’s apartment by the time of the shooting. A trip to the Triple O Lounge from the victim’s apartment would have taken only another fifteen or twenty minutes. The Commonwealth also attacked the credibility of the two witnesses who placed him at the Triple O Lounge. Both were long-time friends of the defendant, and one was engaged to Thomas Barrett’s sister. The other witness failed to explain why he never came forward until one week before the trial.

2. Admission of the composite drawing. The composite drawing was prepared by John Foley and Detective Wilson at the Braintree police station on May 31,1980, the morning of the crime. Foley testified that he selected the hair style and other features from hundreds of plastic overlays; he then directed that pencil changes be made in the nose, and that he made pencil changes in the hair before he was satisfied. At the time the composite was prepared neither Foley nor Wilson had any prior knowledge of the defendant other than the events witnessed by Foley on the night of the murder.

The defendant argues that the composite sketch was inadmissible hearsay. We recently had occasion to consider the admissibility of a composite sketch prepared with the aid of a similar “Identikit,” but did not reach the issue because in that case the admission of the composite drawing could not *69have prejudiced the defendant. Commonwealth v. Blaney, 387 Mass. 628, 631 (1982). Here, if the composite was erroneously admitted, it is likely that the defendant would have been prejudiced. This is because of the emphasis placed upon the composite drawing in the Commonwealth’s closing argument,6 and because here the witness made his observations at night at a distance and for only a few seconds. In Blaney the witness observed the defendant from close proximity inside a well lighted store for a much longer period of time. In both cases, the witness made an in-court identification of the defendant and selected his photograph from an array exhibited to him by the police shortly after the crime. In both cases, the composite drawing bore a strong resemblance to the defendant and was prepared by a witness with the assistance of a police officer, neither one of whom had any prior knowledge of or acquaintance with the defendant. In this case, however, unlike the Blaney case, the evidence was limited to corroboration of the witness’s in-court identification of the defendant.

The defendant finds substantial support for his argument in Commonwealth v. McKenna, 355 Mass. 313, 326-327 (1969). There the court found error in the admission of a similarly made composite drawing. See id. at 327 (the *70drawing “could have been used to refresh [the maker’s] recollection but it had no standing as evidence of the truth or accuracy of the matter contained in it”). On this basis and for a number of other errors the court reversed McKenna’s convictions. Although there is language in McKenna that can be read as a broad-based condemnation of such composites as inadmissible hearsay, the admissibility of such a composite for the limited purpose of corroborating an in-court identification was not considered.7 The court has frequently held that extrajudicial statements to corroborate a witness’s in-court identification are not offered to prove the truth of the matter asserted and, therefore, are not hearsay. Commonwealth v. Repoza, 382 Mass. 119, 129-130 (1980). Commonwealth v. Sheeran, 370 Mass. 82, 87 (1976). Commonwealth v. Swenson, 368 Mass. 268, 273 (1975). Those decisions are representative of an intermediate position between the orthodox banning of all out-of-court statements by the witness and the admission for all purposes suggested by some modern scholars. See McCormick, Evidence § 251, at 603 (2d ed. 1972). One authority has stated the rationale for the rule as follows: “Ordinarily, when a witness is asked to identify the assailant, or thief, or other person who is the subject of his testimony, the witness’ act of pointing out the accused (or other person), then and there in the courtroom, is of little testimonial force. After all that has intervened, it would seldom happen that the witness would not have come to believe in the person’s identity. The failure to recognize would tell for the accused; but the affirmative recognition might mean little against him. The psychology of the situation is practically the same as when recent contrivance is alleged. To corroborate the witness, therefore, it is entirely proper ... to prove that at a former time, when the suggestions of others could not *71have intervened to create a fancied recognition in the witness’ mind, he recognized and declared the present accused to be the person. If, moreover (as sometimes is done), the person was then so placed among others that all probability of suggestion (by seeing him handcuffed, for example) is still further removed, the evidence becomes stronger. . . . That some modern courts are on record for rejecting such evidence is a telling illustration of the power of a technical rule of thumb to paralyze the judicial nerves of natural reasoning.” 4 J. Wigmore, Evidence § 1130, at 277-279 (Chadbourn rev. 1972). More recently, this court has held extrajudicial identification testimony to be admissible as substantive evidence of guilt even where there was no in-court identification by the person making the out-of-court identification. Commonwealth v. Vitello, 376 Mass. 426, 458 (1978). Commonwealth v. Fitzgerald, 376 Mass. 402, 408, 410 (1978). Commonwealth v. Torres, 367 Mass. 737, 738-739 (1975). See P.J. Liacos, Massachusetts Evidence 170, 171 (5th ed. 1982). These decisions are based at least in part on the realization that “[a] pre-trial identification is regarded as having equal or greater testimonial value than one made in court because the circumstances of the earlier identification often were less suggestive and because that identification occurred closer to the time of the offense.” Commonwealth v. Torres, supra at 739. Since the admission of evidence of out-of-court identification is no longer dependent on an in-court identification, admissibility can no longer be supported by the theory that the hearsay rule does not bar the evidence because it merely corroborates the in-court identification. Since juries may have trouble making the distinction between corroboration evidence and substantive evidence, the elimination of this distinction is salutary. Under the Federal Rules of Evidence and the Proposed Massachusetts Rules of Evidence (1980), a statement of prior identification is not hearsay if made by a witness who testifies at trial and is subject to cross-examination concerning it. Fed. R. Evid. 801(d) (1) (C). Proposed Mass. R. Evid. 801 (d) (1) (C) (1980). Paragraph C is consistent *72with Massachusetts law giving substantive effect to prior identification. Proposed Mass. R. Evid. 801 (d) (1) (C) Advisory Committee’s Note (1980). The statements of the witness that form the basis for the composite drawing here would not be inadmissible under our more recent decisions since they are statements of an out-of-court identification (“his eyebrows, nose, and hair looked like that”) by a witness who has made an in-court identification and is available for cross-examination. Since the statements of the witness that led to the creation of the composite are admissible, the composite which is prepared from the statements similarly ought to be admissible either because the com°posite retains the character of the statements that led to its creation or because the composite is not a statement within the meaning of the hearsay rule. See United States v. Moskowitz, 581 F.2d 14, 20 (2d Cir.), cert. denied, 439 U.S. 871 (1978). There a sketch made by a police artist from descriptions given him by two witnesses was held admissible under the Federal Rules of Evidence. Id. Although the majority in that case concluded that the sketch was admissible because it was not a statement within the meaning of Fed. R. Evid. 801(a), the view expressed in the concurring opinion was that the sketch was a statement that was not hearsay by reason of rule 801 (d) (1) (C). Under either view the composite sketch would be admissible under Proposed Mass. R. Evid. 801 (d) (1) (C) (1980).

There is no logical reason to permit the introduction of a witness’s out-of-court identification and to exclude statements identifying the various physical characteristics of a person perceived by the witness, or the composite of all those physical characteristics, which is no more than the sum of the parts perceived.8

As in the case of other convictions based upon eyewitness identification at trial following a pretrial identification, *73identification by composite will be set aside if the pretrial identification process was so impermissibly suggestive as to give rise to a substantial likelihood of irreparable misidentification. Simmons v. United States, 390 U.S. 377, 384 (1968). Commonwealth v. Mobley, 369 Mass. 892, 895 (1976). Here the defendant advances no argument that the preparation of the composite was so tainted, and the record is devoid of any evidence that the preparation of the composite was in any way tainted by suggestions of the police. The judge found that the procedures used in the preparation of the composite were not unduly suggestive.9 No issue of impermissible suggestiveness is presented on this record, even under the standard imposed upon us by G. L. c. 278, § 33E.

We therefore conclude that an “Identikit” composite sketch not shown to be prepared under suggestive circumstances is admissible as substantive evidence of identification.

3. Admission of evidence of motive. The defendant claims that the judge erred in denying his motions in limine to exclude certain evidence of motive.10 As a general rule, the Commonwealth is entitled to introduce all relevant evidence of motive. Commonwealth v. Borodine, 371 Mass. 1, 8 (1976), cert. denied, 429 U.S. 1049 (1977). Evidence is relevant if it renders the desired inference more probable than it would be without the evidence. Green v. Richmond, 369 Mass. 47, 59 (1975). Crowe v. Ward, 363 Mass. 85, 89 (1973). Evidence is to be admitted if it “tends to establish the issue” or “constitutes a link in the chain of proof.” Commonwealth v. Abbott, 130 Mass. 472, 473 (1881). See Poirier v. Plymouth, 374 Mass. 206, 210 (1978). Proof of motive need not be by direct evidence but may be based on inferences which could reasonably be drawn from *74the circumstances. Commonwealth v. Bonomi, 335 Mass. 327, 355-356 (1957). The probative value to be attached to such evidence is a matter for the jury, and evidence that “merely suggests rather than ‘clearly shows’ a motive for the crime may still be ruled admissible.” Commonwealth v. St. Germain, 381 Mass. 256, 271 (1980).

We consider first whether the evidence of the victim’s state of mind toward the defendant during the ten days prior to the victim’s death should be admitted in the absence of direct evidence that the defendant was even aware of the victim’s hostility toward him. Such evidence is admissible only if the jury could have reasonably inferred that the defendant knew of the victim’s state of mind. Commonwealth v. Borodine, supra. We conclude that the jury could have so inferred.

After the fight between Barrett and Shea, LaMonica told at least three people that he wanted to retaliate. The jury could have found that these threats were directed at both Barrett and the defendant. The events between the fight and the shooting support an inference that the threats had been communicated to the defendant. He played a prominent role at the fight and at the confrontations which followed. He told Richard Shea that he would “act first” and could appear to have anticipated an act of retaliation by Frank Shea and his friends. There was testimony that the defendant and the victim had engaged in a heated argument during the week before the murder. This evidence warranted an inference that the defendant and the victim had communicated hostile intentions towards each other. Thus, the evidence concerning the victim’s state of mind was properly admitted as tending to show motive, and the jury were entitled to determine its probative value.

The defendant also challenges on another ground the admission in evidence of the defendant’s statement that he intended to act first. The defendant argues that the prejudicial effect of that portion of the statement where he said, “I have a brother dead, I have a brother in jail,” substantially outweighed its probative value. We find no error.

*75Certainly, the prejudicial effect of the statement was far less than if the defendant had indicated that he himself had been in jail. Contrast Commonwealth v. Chalifoux, 362 Mass. 811, 815-816 (1973). The judge acted properly by instructing the jury to draw no adverse inferences against the defendant from this aspect of the statement. Commonwealth v. Jackson, 384 Mass. 572, 579 (1981). Moreover, evidence need not be excluded simply because it indicates that a defendant committed prior bad acts so long as it is admissible for other relevant purposes. Commonwealth v. Chalifoux, supra. The admission of this evidence was within the sound discretion of the trial judge.

The defendant also claims error in the exclusion of evidence which tended to show that other persons possessed a motive to commit the crime. The evidence was that in 1974 LaMonica had participated in a murder, and that the associates of the victim had threatened to even the score. The victim’s family and Connolly, his paramour, first told the police that they believed that LaMonica was murdered in retaliation for the 1974 murder. The evidence was excluded by the allowance of the Commonwealth’s motion in limine. See Commonwealth v. Hood, 389 Mass. 581, 595 n.5 (1983).

It is open to the defendant to introduce evidence that demonstrates that some other person committed the crime. Commonwealth v. Keizer, 377 Mass. 264, 267 (1979). But the evidence must not be too remote or too weak in probative value. Commonwealth v. Abbott, supra at 475. See Commonwealth v. Graziano, 368 Mass. 325, 329-330 (1975). In the instant case, the trial judge did not abuse his discretion in excluding the evidence. The prior murder was remote, and the defendant failed to tie this evidence to the facts of the case except in the most general way. Cf. Commonwealth v. Keizer, supra at 268.

4. Admission of the mugshot. The day after the crime, and again several months later, Foley picked the defendant’s photograph from an array of mugshots. During direct examination of Foley, the Commonwealth moved *76to introduce a sanitized set of the entire array to corroborate Foley’s identification. The defendant objected, arguing that only the front view of the defendant should be admitted. The judge overruled the objection and allowed the photograph of the defendant’s profile to be introduced.

The defendant’s argument is that the profile view lacked probative value in the circumstances of the case. He claims that Foley’s identification was based solely on a front view of the man running from the scene. It is clear from the record that Foley’s identification was based on both a front and profile view. The reliability of his profile identification was relevant. Foley’s description of the man at trial focused on the man’s “crooked” nose, a feature which was best seen from the side. The reliability of the profile identification was also important, given the duration of Foley’s opportunity to observe the man’s face from the front and the defendant’s claims at trial that identification process was suggestive and unreliable. There was no error in allowing the profile mugshot in evidence.

The defendant also contends that the judge should have severed the front and side views. We considered this issue in Commonwealth v. Blaney, 387 Mass. 628 (1982). In that case, we again noted the risk that use of double view police photographs raises an inference of prior contact with the criminal law and adhered to our view that the jury are “best left with the impression that any photographs used to identify the defendant were taken after his arrest on the charges for which he is being tried.” Id. at 639. While it will usually be better not only to sanitize such photographs but also to sever the two views, Commonwealth v. Lockley, 381 Mass. 156, 165-166 (1980), it was within the discretion of the trial judge to determine that the photographic array should go to the jury precisely as it existed at the time of Foley’s identifications. Commonwealth v. Blaney, supra at 639. We note that the judge employed every other reasonable means to limit any prejudice, and delivered a limiting instruction.

*775. Admission of the enlarged photograph. The defendant’s next claim of error is that the judge improperly admitted in evidence an enlarged black and white reproduction of a color photograph of the defendant taken at the time of his arrest.11 The judge correctly ruled that the best evidence rule does not apply to photographs. See Commonwealth v. Balukonis, 357 Mass. 721, 725-726 (1970), denial of habeas corpus aff’d sub nom. Allen v. Moore, 453 F.2d 970 (1st cir. 1972). Cf. Snow v. Massachusetts Turnpike Auth., 339 Mass. 620, 623 (1959). But cf. Proposed Mass. R. Evid. 1001, 1002 (1980). There was adequate evidence that the enlarged photograph was a fair and accurate representation of the defendant at the time of his arrest. There was no error.

6. Exclusion of night photographs and the testimony of photographer. The defendant sought to introduce in evidence six photographs of the parking lot at 196 Commercial Street at night. They were taken from various points in Faxon Park by a photographer retained by the defendant. One of the photographs was admitted over the Commonwealth’s objection after a Braintree police officer testified that it fairly and accurately depicted the lighting in the parking lot area on May 31, 1980. The defendant also offered in evidence the testimony of the photographer. He was to testify to the conditions under which the photographs were taken and that the photographs accurately reflected how the human eye would perceive a human figure at a distance of 175 feet. This testimony was excluded.

We believe that the judge could properly exclude the photographs. No witness testified that the five excluded photographs accurately represented the condition of the lighting at the time of the murder. See Commonwealth v. Sheeran, 370 Mass. 82, 87 (1976). Cf. Howe v. Boston, 311 *78Mass. 278, 281-282 (1942). Indeed, there was testimony on redirect examination that the photographs did not represent accurately the lighting on the night of the murder. Moreover, the verification of the one photograph by a police officer did not serve to verify the others. That officer testified that “your view could change dramatically” simply by moving several feet. The photographer also testified that he had no personal knowledge of the lighting conditions on the night of the murder. The photographs were not properly verified.

We also think that the testimony of the photographer was properly excluded.12 His testimony went to the issue of a witness’s ability to perceive and identify a human figure. This testimony is of a nature that is presumed to be within the common experience of a jury. See Commonwealth v. Jones, 362 Mass. 497, 501 (1972). It also appears that the photographer lacked the qualifications to testify how a human eye perceives a figure from a distance. Expertise in the area of photography does not qualify a witness to testify on the subject of human perception. Cf. Commonwealth v. Seit, 373 Mass. 83, 91-92 (1977). See P.J. Liacos, Massachusetts Evidence 104-106 (5th ed. 1982). The matter was within the sound discretion of the trial judge.

7. Review under G. L. c. 278, § 33E. The defendant asks that we set aside the verdict in this case as being unsupported by the weight of the evidence and not consonant with justice. The gunman who was lying in wait for this victim fired four shots hitting him twice. Although Foley, the principal identifying witness, had only a brief opportunity to view the killer, he did so while the killer was under a street light and had turned to face him. He participated *79in the preparation of a composite which bore a striking resemblance to the defendant on the morning of the murder when neither he nor the police had any prior knowledge of the defendant. Foley selected the defendant from a group after viewing hundreds of people in neighborhood parks and streets. The defendant was known to have quarreled with the victim and made threats against him. There was, therefore, sufficient evidence to warrant a conclusion by the jury that the defendant was guilty of murder in the first degree. We have considered this case on the law and the evidence and see no reason to grant a new trial or to direct the entry of a verdict of a lesser degree of guilt than found by the jury.

Judgment affirmed.

(dissenting). I find myself in agreement with most of the court’s able discussion of the issues raised by this case. I am compelled, however, to dissent from the court’s conclusion that a composite may be properly admitted as substantive evidence of identification.1 In Commonwealth v. McKenna, 355 Mass. 313 (1969), this court found error in the admission of an Identikit composite drawing, stating that the composite “had no standing as evidence of the truth or accuracy of the matter contained in it.” Id. at 327. In Commonwealth v. Blaney, 387 Mass. 628 (1982), we were asked to reconsider our holding in McKenna. The court found it unnecessary to do so, stating that the “admission of the composite drawing of the defendant could not have prejudiced him” in the circumstances of that case. Id. at 633. I cannot agree with the court’s conclusion that the admission of the composite in this case was proper and did not constitute prejudicial error. Unlike Blaney, in which I concurred in the result on the ground of harmless error, this case presents two distinguishing factors.

*80First, the witness’s opportunity to observe the defendant at the time of the crime was slight. Looking at the evidence most favorable to the Commonwealth, it is clear that the jury had before them evidence of an eyewitness identification based on a limited period of observation over a substantial distance at night by a witness who was under the influence of alcohol. In contrast, the identification made in Blaney was based on repeated observations of the perpetrator’s profile at close quarters in a store. In the circumstances of the instant case, we cannot say that the jury would not have relied on the composite drawing to determine if the defendant was the man seen running from the scene.

Second, the closing argument of the Commonwealth invited the jury to convict the defendant on the basis of the composite. During her closing argument, the prosecutor asked the jury to compare the composite with the photograph of the defendant taken at the time of his arrest. See note 6, supra. The strong reliance on the composite ensured that the admission in evidence of the composite drawing, if error, would not constitute harmless error.

I turn now to the question whether the admission of the composite was error. Foley and Wilson compiled the composite drawing with the aid of an Identikit. This kit consists of several hundred transparent overlays that can be assembled into a “composite overlay ‘sandwich.’” A.A. Moenssens & F.E. Inbau, Scientific Evidence in Criminal Cases § 17.03, at 666 (1978). Each overlay depicts a facial characteristic. The theory is that “because of certain consistencies in the structure of human likenesses, only four factors are necessary to construct the basic composite: age, height, weight, and one of 49 different hairlines. Other characteristics which may be added simply expand the versatility.” Id. at 667. A trained operator selects each overlay based on the description given by the witness. An initial composite is produced, and the witness may suggest changes until he is satisfied.

*81The degree to which composites are a reliable means of identifying the perpetrators of crimes has not been established. No evidence on this issue was put before the trial judge. The knowledge which presently exists on the subject suggest that composites may not be entirely reliable and may be less reliable than other means of identification. One study notes that the mental process involved in reconstructing a suspect’s face through the use of an Identikit is fundamentally different from the process of selecting a face from a photographic array. Cohen, Number of Features, and Alternatives per Feature, in Reconstructing Faces With the Identi-Kit, 1 J. Police Sci. & Ad. 349 (1973). The Identikit process which requires the witness to select a number of individual features introduces a critical element which is not present when a photographic array is used. Id. Witnesses may be incapable of accurately selecting the number of individual features required to assemble the composite.2 There is also some indication that the process of constructing a composite may taint a subsequent identification. E.F. Loftus, Eyewitness Testimony 150-151 (1979).

The question of reliability is central to our inquiry. Without some evidence of reliability, we are left without any means of determining whether the admission of composites furthers or frustrates the truth-seeking process of criminal trial. Substantial reasons exist to question the reliability of composites. I would therefore hold that the proponent of a composite should be required to lay the foundation of ad*82missibility by adducing evidence of the reliability of the composite process and also the reliability of the procedures used to produce the composite drawing being offered.

The admission of evidence which purportedly has some scientific basis of reliability can “create a substantial danger of undue prejudice and confusion because of its aura of special reliability and trustworthiness. ” United States v. Fosher, 590 F.2d 381, 383 (1st Cir. 1979). Evidence of a composite appears akin to various forms of newly developed scientific evidence and is subject to some of the same objections. See generally Commonwealth v. Vitello, 376 Mass. 426, 440-447 (1978) ;3 Giannelli, The Admissibility of Novel Scientific Evidence: Frye v. United States, A Half-Century *83Later, 80 Colum. L. Rev. 1197 (1980). In a sense, the process is based on scientific theories of physiognomy, perception, and memory. In my view, however, composites should not be subject to the rigors of the rule of Frye v. United States, 293 F. 1013 (D.C. Cir. 1923), for the reason that the process involved is not exclusively scientific in nature. This is so because the process includes the involvement of the percipient lay witness in the development of a particular composite.

A better standard to judge the admissibility of composites is found in traditional notions of relevancy. Evidence and expert testimony, which otherwise may be admissible under Frye or as an exception to the hearsay rule may still be objectionable if its “aura of special reliability and trustworthiness” is not commensurate with its actual reliability. City of New York v. Pullman Inc., 662 F.2d 910, 915 (2d Cir. 1981), cert. denied, 454 U.S. 1164 (1982), quoting United States v. Fosher, supra at 383 (even if hearsay report was admissible as a governmental record, trial judge could exclude report). United States v. Fosher, supra (expert testimony held properly excluded due to danger of undue prejudice and confusion). Marx & Co. v. Diners’ Club, Inc., 550 F.2d 505, 511 (2d Cir.), cert. denied, 434 U.S. 861 (1977) (statistical testimony held objectionable due to danger of “prejudicial overweight”). See generally 1 J. Weinstein & M. Berger, Evidence par. 403[04] (1982); Fed. R. Evid. 403; Proposed Mass. R. Evid. 403.4 Composites *84fall within this class of evidence. A jury are not in a position to examine critically the process of assembling a composite and may overestimate its reliability.

A related concern is the inherent suggestiveness of the method. A composite “represent[s] the effort of a trained police artist to transform a witness’s oral description into pictorial form.” Commonwealth v. Williams, 378 Mass. 217, 230 (1979). Commonwealth v. McKenna, 355 Mass. 313, 327 (1969) (composite is “a recording in graphic form” of witness’s statements). “The process is inherently susceptible to subtle and even unconscious suggestiveness by the police artist,” particularly when the witness is inarticulate or has “only an ill-defined image of the offender.” Commonwealth v. Blaney, 387 Mass. 628, 640, 642 (1982) (O’Connor, J., dissenting). To produce a composite, the witness’s mental image of the offender first must be matched with individual facial features. Only after a number of individual features have been selected and assembled is the witness able to determine whether his mental image of the offender comports with the composite. By that point, either the witness’s image may have been altered or the witness may be unable to decide which facial features are correct.

The problem of suggestiveness is particularly acute in two situations. First, in many cases the police may have identified a potential suspect. The danger then exists that the composite will be matched to the suspect rather than the suspect to the composite. Even the most conscientious police operator will have difficulty maintaining a neutral pose in such circumstances. While this concern is not present in the instant case, it may be present in other cases.

Second, any form of suggestiveness in the process of identification is problematic when the identification is based on an observation of limited duration. Clearly, the weaker the contemporaneous impression, the more likely the witness will be influenced by the identification process. Cf. Commonwealth v. Moon, 380 Mass. 751, 756-759 (1980) (period of observation lasted ten to twenty seconds in *85poor lighting); Commonwealth v. Botelho, 369 Mass. 860, 869-870 (1976) (limited period of observation over a substantial distance in dim lighting by witness influenced by alcohol); State v. Commeau, 409 A.2d 247 (Me. 1979) (period of observation lasted a few seconds at night from a distance of sixty feet); N. Sobel, Eyewitness Identification § 6.3(a) (1981). Contrast Commonwealth v. Ross, 361 Mass. 665, 667-669, 672 (1972) (witness obtained firm contemporaneous impression), judgment vacated, 410 U.S. 901, aff'd on remand, 363 Mass. 665, cert. denied, 414 U.S. 1080 (1973) (with dissents), habeas corpus granted sub nom. Ross v. Ristaino, 388 F. Supp. 99 (D. Mass.), aff'd, 508 F.2d 754 (1st Cir. 1974), rev’d, 424 U.S. 589 (1976); United States v. Moskowitz, 581 F.2d 14, 20 (2d Cir.), cert. denied, 439 U.S. 871 (1978) (witness had excellent opportunity to observe perpetrator); People v. Bills, 53 Mich. App. 339, 349 (1974) (same), remanded on other grounds, 396 Mich. 819 (1976). It will be precisely in those case, like the instant one, where the reliability of any identification is weak that a jury are apt to place the greatest reliance on a composite drawing. Thus, before the judge reaches the issue of the competency of the composite evidence under the hearsay rule and the opinion rule, he must make a determination in terms of relevancy, i.e., probative worth weighed against probable risks of undue prejudice or misleading of the jury. This process of weighing reliability and probative worth was not engaged in here.

Assuming that the trial judge has determined the composite evidence to be sufficiently probative to warrant admission, a third concern is that the composite is hearsay evidence. The hearsay rule renders extrajudicial statements inadmissible if they are offered to prove the truth of the matters asserted in the statements. Some courts have excluded composites, as has this court, on the ground that they are a form of inadmissible hearsay. Commonwealth v. McKenna, supra. Commonwealth v. Rothlisberger, 197 Pa. Super. 451 (1962).

We have stated previously that “[a]n extrajudicial identification made by a witness may be offered in evidence for *86three possible purposes: (1) for corroboration; (2) for impeachment; or (3) as substantive evidence of an identification, having.probative value.” Commonwealth v. Vitello, 376 Mass. 426, 458 (1978). When the extrajudicial identification is being offered to impeach a witness, or to rebut a claim of recent fabrication, the identification is not hearsay since it is not being offered to prove the truth of the matter asserted. Commonwealth v. Lacy, 371 Mass. 363, 370-371 (1976). Commonwealth v. Swenson, 368 Mass. 268, 274 (1975).5

Evidence of an extrajudicial identification introduced merely to corroborate an in-court identification has not been considered to be hearsay evidence. Commonwealth v. Repoza, 382 Mass. 119, 129-130 (1980). Our recent decisions on the admissibility of extrajudicial identification, however, have abandoned the distinction between substantive and corroborative evidence in the context of identification evidence.6 Commonwealth v. Vitello, supra at 458-459. Commonwealth v. Fitzgerald, 376 Mass. 402, 407-410 (1978). Commonwealth v. Torres, 367 Mass. 737, 738-739 (1975). These decisions reflect skepticism that *87a jury are able to maintain the distinction between an extrajudicial identification introduced for the purpose of corroboration and one introduced as probative evidence.7 Thus, the extrajudicial identification is as probative of the defendant’s guilt as an in-court identification, in both practical and legal effect. Commonwealth v. Blaney, supra at 641 (O’Connor, J., dissenting), and cases cited. As such, “[t]he evidence is admitted to prove the truth of the matter asserted outside of court, and is, therefore, hearsay. It is admitted despite the hearsay rule and as an exception to it.” Id.

This exception to the hearsay rule is premised on a practical assessment of the relative reliability of different methods of identification. The inherent suggestiveness of the courtroom setting and the passage of time serve to diminish the reliability of an in-court identification. A prior extrajudicial identification is therefore regarded as having testimonial value equal to or greater than one made in court. Commonwealth v. Torres, supra at 739. Commonwealth v. Locke, 335 Mass. 106, 112 (1956).

Evidence of a pretrial composite identification is distinguishable from the other types of pretrial identification evidence. Unlike extrajudicial photographic or in-person identifications, composites have not as yet been shown to possess a fair degree of reliability, let alone greater reliability than in-court identification. It cannot be said that they are significantly more reliable than an in-court identification. “Translating a mental image to a composite drawing provides opportunities for communication failure, error in reproduction, and extraneous influence that are not present when a witness can compare his mental image directly with either a photographic image or a person.” Commonwealth v. *88Blaney, supra at 642 (O’Connor, J., dissenting). See Eyewitness Identification, 18 Am. Jur. 2d, Proof of Facts 361, § 9, at 379-380 (1979). In the absence of evidence demonstrating their reliability,8 we should hold that composite drawings are inadmissible as probative evidence of guilt or as corroborative evidence9 and reverse the conviction.

(dissenting). I do not agree with the court’s holding that the composite was properly admitted in evidence. My views on the admissibility of composites in criminal trials were fully expressed in my dissenting opinion in Commonwealth v. Blaney, 387 Mass. 628, 640-643 (1982). Because I believe that admission of the composite was prejudicial error, I would reverse the conviction and remand for a new trial.