The defendant appeals from his conviction of larceny from a person after a trial by a jury of six in the Boston Municipal Court Department. The sole issue on appeal is whether the judge erred in denying the defendant’s motion to suppress the victim’s pretrial identification after concluding that the identification procedure was unnecessarily suggestive. We granted the defendant’s application for direct appellate review. We reverse.
We recite the facts found by the judge. On April 7, 1992, Leopoldino Goncalves1 was working at a parking lot on the corner of Traveler Street and Washington Street in Boston. After he finished work, at approximately 10:50 p.m., Goncalves walked across the street to use a public telephone that was located on Washington Street. Street lights provided the only illumination.
When Goncalves finished using the telephone, a white female with a limp approached him and asked him for a dollar. Goncalves told the woman that he did not have any money. A black male armed with a machete then approached. The man grabbed Goncalves’s wallet and at the same time the woman snatched money from Goncalves’s front pocket. The assailants discarded the wallet after removing the money. They left the area together in an automobile. Goncalves pursued them in his own automobile, but he lost sight of them in *460a public housing project. The entire incident described lasted only a few minutes.
Approximately forty-five minutes later, Goncalves went to the Area D-4 police station and reported the robbery. He described the male assailant as a twenty-seven to thirty year old black male, six feet tall with a medium build, weighing 170 pounds, and wearing a black cap, blue jeans, and a brown sweatshirt. Goncalves was shown about six books containing photographs of suspects, but was unable to identify his assailants.2 Goncalves then accompanied a police officer to view a group of potential suspects. Once again, Goncalves did not make an identification.
The day following the incident, four police officers arrived at Goncalves’s place of employment at approximately 5 p.m. They told Goncalves that they wanted him to view two suspects. Goncalves accompanied the officers. When they arrived at the location where the suspects were being held, Goncalves saw a group of six to eight people. Only one adult black male, the defendant, was present and a female with a limp was the only adult white female present. The two suspects were being “detained” by police officers but they were not handcuffed. The defendant and the woman were brought forward a few steps by the officers. Goncalves then identified the pair as his assailants. Goncalves based his identification in part on the fact that the clothing worn by the suspects was the same as that worn by his assailants.
The defendant possessed several characteristics that did not match Goncalves’s initial description of the male assailant. A booking photograph taken of the defendant at the time of his arrest, the day after the incident, shows that the defendant had a moustache. Yet Goncalves had never mentioned that the male assailant had a moustache. The booking sheet indicates that the defendant is thirty-seven years old and weighs 220 pounds, whereas Goncalves had described a man of approximately twenty-seven years in age, weighing *461170 pounds, with a medium build. Finally, at the time of the hearing on the motion to suppress the defendant was missing several front teeth. When describing his assailants to the police, Goncalves did not tell them that the male assailant had missing teeth.
The judge ruled that Goncalves’s identification of the defendant was tainted because it was made at an unnecessarily suggestive showup. The evidence presented at the motion hearing supports this conclusion. Although one-on-one confrontations are not per se excludable, they are disfavored because of their inherently suggestive nature. See Commonwealth v. Howell, 394 Mass. 654, 660 (1985); Commonwealth v. Barnett, 371 Mass. 87, 92 (1976), cert. denied, 429 U.S. 1049 (1977). Showups have been permitted when conducted in the immediate aftermath of a crime and in exigent circumstances. Commonwealth v. Barnett, supra at 92. See Stovall v. Denno, 388 U.S. 293, 301-302 (1967). The showup employed by the police in this case was conducted eighteen hours after the crime. It took place in the area of the housing project where Goncalves had seen his assailants drive the previous night; the defendant was brought forward from the group before Goncalves positively identified him; and the defendant was wearing clothes similar to those worn by the male assailant. Based on these facts, the judge was warranted in concluding that the identification procedure was unnecessarily suggestive.
Although the judge found the identification procedure unnecessarily suggestive, he found that the identification was admissible because it was reliable.3 In so doing, the judge relied on Appeals Court decisions which have adopted the “reliability test,” set forth in Manson v. Brathwaite, 432 U.S. 98 (1977), regarding the admissibility of identifications *462obtained through unnecessarily suggestive procedures.4 See Commonwealth v. Hicks, 17 Mass. App. Ct. 574 (1984). See also Commonwealth v. Riley, 26 Mass. App. Ct. 550, 553-554 (1988); Commonwealth v. Laaman, 25 Mass. App. Ct. 354, 362, cert. denied, 488 U.S. 834 (1988); Commonwealth v. Jones, 25 Mass. App. Ct. 55, 60 (1987); Commonwealth v. Key, 19 Mass. App. Ct. 234 (1985); Commonwealth v. Gordon, 6 Mass. App. Ct. 230 (1978).
Although the Appeals Court has applied the due process analysis set forth in the Brathwaite case, this court has never accepted the reasoning in Brathwaite as an accurate interpretation of the due process requirements of art. 12 of the Declaration of Rights of the Massachusetts Constitution.5 Whether we should embrace Brathwaite, as have the majority of other States, is a question we have left open. Commonwealth v. Melvin, 399 Mass. 201, 205 (1987). Commonwealth v. Correia, 381 Mass. 65, 81 (1980). Commonwealth v. Venios, 378 Mass. 24, 26-28 (1979). In cases involving an unnecessarily suggestive identification, we have adhered to *463the stricter rule of per se exclusion previously followed by the Supreme Court and first set forth in the Wade-GilbertStovall trilogy6 of cases. Commonwealth v. Botelho, 369 Mass. 860, 865-869 (1976). See Commonwealth v. Dinkins, 415 Mass. 715, 720-721 (1993); Commonwealth v. Smith, 414 Mass. 437, 442-443 (1992); Commonwealth v. Thornley, 406 Mass. 96, 98-99 (1989); Commonwealth v. Melvin, supra at 205.
The rule of per se exclusion, set forth in Commonwealth v. Botelho, supra7 states that the defendant bears the burden of demonstrating, by a preponderance of the evidence, that the “witness was subjected by the State to a confrontation that was unnecessarily suggestive and thus offensive to due process.” Id. at 866. If this is established, then the prosecution is barred from introducing that particular confrontation in evidence at trial. Id.8 As for other identifications the witness may have made of the defendant, “the prosecution is limited to introducing at trial only such identifications by the witness as are shown at the suppression hearing not to be the product of the suggestive confrontation — the later identifications, to be usable, must have an independent source.” Id. The prosecution must demonstrate the existence of an independent source by “clear and convincing evidence.” Id. at 868.
“In deciding whether a particular confrontation was unnecessarily suggestive, the judge is to consider ‘the totality of *464the circumstances surrounding it’ (Stovall, 388 U.S. at 302).” Commonwealth v. Botelho, supra at 867. Additionally, in determining whether a separate identification has a source independent of the unnecessarily suggestive identification, the judge considers the following factors: “(1) The extent of the witness’ opportunity to observe the defendant at the time of the crime; prior errors, if any, (2) in description, (3) in identifying another person or (4) in failing to identify the defendant; (5) the receipt of other suggestions, and (6) the lapse of time between the crime and the identification.” Id. at 869, quoting Commonwealth v. Ross, 361 Mass. 665, 671 n.2 (1972), judgment vacated on other grounds, 410 U.S. 901, aff'd on remand, 363 Mass. 665, cert. denied, 414 U.S. 1080 (1973).
The Commonwealth now urges us to abandon the per se rule of exclusion and, like the Appeals Court, follow the reliability test of Brathwaite. Under the “reliability” test, if a defendant demonstrates that an identification was unnecessarily suggestive, evidence of that identification is not per se excluded. Neil v. Biggers, 409 U.S. 188, 199 (1972). Instead, the court must determine whether the identification was, under the “totality of the circumstances,” nevertheless reliable. Id. at 198-199. In determining whether an identification was reliable, and thus there was no “substantial likelihood of irreparable misidentification,” the court must examine several factors. Id. at 199, 201. Manson v. Brathwaite, supra at 106. These factors are (1) the opportunity of the witness to view the criminal at the time of the crime, (2) the witness’s degree of attention, (3) the accuracy of the witness’s prior description of the criminal, (4) the level of certainty demonstrated by the witness at the confrontation, and (5) the length of time between the crime and the confrontation. Neil v. Biggers, supra at 199-200. Manson v. Brathwaite, supra at 114. “Against these factors is to be weighed the corrupting effect of the suggestive identification itself.” Id.
The case at bar presents us with the opportunity to establish our position with regard to the Brathwaite decision. We have carefully considered the matter and, for the reasons set *465forth, we conclude that we cannot accept Brathwaite as satisfying the requirements of art. 12. We conclude that art. 12 requires the application of the stricter per se approach described in Commonwealth v. Botelho, supra.
Our past resistance to the so-called reliability test reflects this court’s concern that the dangers present whenever eyewitness evidence is introduced against an accused require the utmost protection against mistaken identifications. There is no question that the danger of mistaken identification by a victim or a witness poses a real threat to the truth-finding process of criminal trials. Indeed, mistaken identification is believed widely to be the primary cause of erroneous convictions. See Gross, Loss of Innocence: Eyewitness Identification and Proof of Guilt, 16 J. Legal Stud. 395, 396 (1987); Note, Twenty-Years of Diminishing Protection: A Proposal to Return to the Wade Trilogy’s Standards, 15 Hofstra L. Rev. 583, 605-606 (1987); Grossman, Suggestive Identifications: The Supreme Court’s Due Process Test Fails to Meet its Own Criteria, 11 U. Balt. L. Rev. 53, 65 (1981). See also Wright v. United States, 404 F.2d 1256, 1262 (D.C. Cir. 1968) (Bazelon, J., dissenting) (“Since mistaken identifications are probably the greatest cause of erroneous convictions, we must require the fairest identification procedures available under the circumstances. With the stakes so high, due process does not permit second best” [footnote omitted]). Compounding this problem is the tendency of juries to be unduly receptive to eyewitness evidence. Manson v. Brathwaite, supra at 120 (Marshall, J., dissenting). See State v. Long, 721 P.2d 483, 490-491 (Utah 1986); Grossman, supra at 65. We have stated that “[t]he law has not taken the position that a jury can be relied on to discount the value of an identification by a proper appraisal of the unsatisfactory circumstances in which it may have been made. On the contrary, this court, like others, has read the Constitution to require that where the conditions are shown to have been highly and unnecessarily suggestive, the identification should not be brought to the attention of the jury.” Commonwealth v. Marini, 375 Mass. 510, 519 (1978).
*466These concerns were at the heart of the Wade trilogy of cases. The Wade Court acknowledged that “the vagaries of eyewitness identification are well-known; the annals of criminal law are rife with instances of mistaken identification.” United States v. Wade, 388 U.S. 218, 228 (1967). In declaring that the accused has a right to counsel at a postindictment lineup, the Court recognized that the presence of counsel was necessary because “[t]he trial which might determine the accused’s fate may well not be that in the courtroom but that at the pretrial confrontation, with the State aligned against the accused, the witness the sole jury, and the accused unprotected against the overreaching, intentional or unintentional, and with little or no effective appeal from the judgment there rendered by the witness — ‘that’s the man.’ ” Id. at 235-236. The “reliability test” is unacceptable because it provides little or no protection from unnecessarily suggestive identification procedures, from mistaken identifications and, ultimately, from wrongful convictions.
The Brathwaite Court examined three primary “interests” before holding that the per se rule should be abandoned in favor of the less protective “reliability” test. The first of these was the concern regarding the dangers presented by eyewitness evidence. Manson v. Brathwaite, supra at 112. The Court acknowledged that a witness’s recollection “can be distorted easily by the circumstances or by later actions of the police.” Id. While the per se approach addresses this concern, the Court stated, it “goes too far since its application automatically and peremptorily, and without consideration of alleviating factors, keeps evidence from the jury that is reliable and relevant.” Id.
We believe that Justice Marshall, dissenting in Brathwaite, had a more realistic view of the trial process when he stated:
“this conclusion totally ignores the lessons of Wade. The dangers of mistaken identification are, as Stovall held, simply too great to permit unnecessarily suggestive identifications. Neither Biggers nor the Court’s *467opinion today points to any contrary empirical evidence. Studies since Wade have only reinforced the validity of its assessment of the dangers of identification testimony.” Id. at 125 (Marshall, J., dissenting).
Indeed, studies conducted by psychologists and legal researchers since Brathwaite have confirmed that eyewitness testimony is often hopelessly unreliable.9 Permitting the admission of an identification obtained through unnecessarily suggestive procedures can only serve to exacerbate this problem. Furthermore, contrary to the Brathwaite Court’s unsubstantiated claim, the per se approach does not keep relevant and reliable identification evidence from the jury. Subsequent identifications shown to come from a source independent of the suggestive identification remain admissible under the per se approach. Commonwealth v. Botelho, 369 Mass. 860, 866 (1976). See People v. Adams, 53 N.Y.2d 241, 251 (1981). The per se approach excludes only the unnecessarily suggestive identification and subsequent tainted identifications. Id. As stated earlier, the court examines five factors in determining whether there was an independent source for subsequent identifications by the witness of the defendant. If, for example, the prosecution is able to demonstrate that the witness got a good look at his assailant and his initial description matches a description of the defendant, the court may conclude that there was an independent source and may admit evidence of any identification subsequent to the unnecessarily suggestive one.
The Brathwaite Court also discussed the public interest in deterring police from using identification procedures which are unnecessarily suggestive. The Court acknowledged that *468the per se rule is superior in promoting that interest because it provides greater deterrence against police misconduct. The Court nevertheless concluded: “The police will guard against unnecessarily suggestive procedures under the totality rule, as well as the per se one, for fear that their actions will lead to the exclusion of identifications as unreliable.” Manson v. Brathwaite, supra at 112.
To the contrary, it appears clear to us that the reliability test does little or nothing to discourage police from using suggestive identification procedures. One commentator has noted that “under Brathwaite, the showup has flourished, because the totality approach has failed to discourage this practice. As a deterrent to suggestive police practices, the Federal standard is quite weak. Almost any suggestive lineup will still meet reliability standards.” Note, Twenty-Years of Diminishing Protection, 15 Hofstra L. Rev. 583, 606 (1987). Indeed, an example of this result is seen in the instant case: the suggestion inherent in the showup procedure that was used to identify the defendant is plain. Furthermore, the showup was unnecessarily suggestive in that it was not conducted immediately after the crime or in exigent circumstances. See Commonwealth v. Barnett, 371 Mass. 87, 92 (1976). Yet the motion judge permitted the introduction of the identification based on his opinion that the identification was reliable. Rather than deterring unreliable identification procedures, the effect of the Biggers-Brathwaite reliability test has been, and would be in this Commonwealth, a message to police that, absent extremely aggravating circumstances, suggestive showups will not result in suppression. Whether or not to use a more fair and accurate identification procedure is, under that test, left to the officer’s discretion. See Grossman, supra at 60.
Finally, the Brathwaite Court considered the impact of the two tests on the administration of justice. It was here that the Court found what it considered to be the most serious drawbacks of the per se approach. However, it is also here, in our view, that the Court erred most. The Court opined: “Since it denies the trier reliable evidence, [the per se ap*469proach] may result, on occasion, in the guilty going free.” Manson v. Brathwaite, supra at 112. The inverse of this is probably more accurate: the admission of unnecessarily suggestive identification procedures under the reliability test would likely result in the innocent being jailed while the guilty remain free. See Manson v. Brathwaite, supra at 127 (Marshall, J., dissenting) (“[I]f the police and the public erroneously conclude, on the basis of an unnecessarily suggestive confrontation, that the right man has been caught and convicted, the real outlaw must still remain at large”). The Braithwaite Court disregards the wisdom of Justice Harlan when he wrote: “it is far worse to convict an innocent man than to let a guilty man go free.” In re Winship, 397 U.S. 358, 372 (1970) (Harlan, J., concurring).
Justice Grimes of the New Hampshire Supreme Court expressed this concern accurately when he wrote: “The dangers of misidentification are well known and documented. See United States v. Wade, 388 U.S. 288 (1967); Gilbert v. California, 388 U.S. 263 (1967); Stovall v. Demo, 388 U.S. 293 (1967). In these cases the court recognized ‘[a] major factor . . . [in] the high incidence of miscarriage of justice from mistaken identification has been the degree of suggestion inherent in the manner in which the prosecution presents the suspect to witnesses for pretrial identification.’ United States v. Wade, 388 U.S. at 288. . . . Considering the complexity of the human mind and the subtle effects of suggestive procedures upon it, a determination that an identification was unaffected by such procedures must itself be open to serious question. All things considered, the consequences of misidentification are so great and the risks so high that it is questionable if a totality-of-circumstances test provides either a sufficient deterrence against the unnecessary use of one-man showups or a sufficient protection against misidentifications which may result therefrom.” State v. Leclair, 118 N.H. 214, 217-218 (1978).
We note that there are essential policy differences between other aspects of an exclusionary rule and the rule supporting the per se exclusion of suggestive identifications and tainted *470trial identifications. The New York Court of Appeals summed up these differences in People v. Adams, 53 N.Y.2d 241, 250-251 (1981), when it stated: “The rule excluding improper showups and evidence derived therefrom is different in both purpose and effect from the exclusionary rule applicable to confessions and the fruits of searches and seizures. In the latter cases generally reliable evidence of guilt is suppressed because it was obtained illegally. Although this serves to deter future violations, it is collateral and essentially at variance with the truthfinding process (see, e.g., People v. McGrath, 46 N.Y.2d 12 [1978], cert. denied, 440 U.S. 972 [1979]). But the rule excluding improper pretrial identifications bears directly on guilt or innocence. It is designed to reduce the risk that the wrong person will be convicted as a result of suggestive identification procedures employed by the police.”
The Court of Appeals continued: “A reliable determination of guilt or innocence is the essence of a criminal trial. A defendant’s right to due process would be only theoretical if it did not encompass the need to establish rules to accomplish that end. Permitting the prosecutor to introduce evidence of a suggestive pretrial identification can only increase the risks of convicting the innocent in cases where it has the desired effect of contributing to a conviction. In most instances, where the witness is able to make an untainted identification in court, proof of the suggestive showup only serves to bolster the People’s case. However, if the jury finds the in-court identification not entirely convincing it should not be permitted to resolve its doubts by relying on the fact that the witness had identified the defendant on a prior occasion if that identification was made under inherently suggestive circumstances. Similarly, if the witness is unable to identify the defendant at trial the defendant’s conviction should not rest solely upon evidence of a pretrial identification made under circumstances which were likely to produce an unreliable result.” Id. at 251.
Thus, the Court of Appeals concluded: “Excluding evidence of a suggestive showup does not deprive the prosecutor *471of reliable evidence of guilt. The witness would still be permitted to identify the defendant in court if that identification is based on an independent source.” Id.
The reliability approach has the further effect of expanding a reviewing court’s powers because it allows the court to examine subjectively the circumstances surrounding the original crime on a case-by-case basis, instead of objectively examining the identification procedures employed by the police. Sherwood, The Erosion of Constitutional Safeguards in the Area of Eyewitness Identification, 30 How. L.J. 439, 457 (1987). See Note, Twenty-Years of Diminishing Protection, 15 Hofstra L.Rev. 583, 592 (1987). “By relying on the probable accuracy of a challenged identification, instead of the necessity for its use, the Court seems to be ascertaining whether the defendant was probably guilty.” Manson v. Brathwaite, supra at 128 (Marshall, J., dissenting).
The reliability test has been widely criticized by commentators,10 primarily because the test eliminates the protection essential to a fair trial. The reliability test hinders, rather than aids, the fair and just administration of justice by permitting largely unreliable evidence to be admitted directly on the issue of the defendant’s guilt or innocence.
This case presents an example of why we should not abandon the per se rule of exclusion and replace it with the reliability test. There is absolutely no evidence that the in-court identification of the defendant was the result of anything independent of the unnecessarily suggestive showup. For example, Goncalves’s description of his assailant, given to police just after the incident, did not match the defendant’s appearance, in part because the defendant possessed the unique feature of several missing teeth. Regardless of this fact, follow*472ing the showup Goncalves was able to “remember” that his assailant had missing teeth.11 Such flimsy evidence should not be permitted at trial. Only a rule of per se exclusion can ensure the continued protection against the danger of mistaken identification and wrongful convictions. Accordingly, we reject Brathwaite and affirm our confidence in the Botelho approach.
The verdict of guilty is vacated. The judgment of conviction is reversed.
So ordered.
The judge referred to the victim in his findings as Goncalves. However, the briefs before this court refer to the victim as Gonsalves. To avoid confusion, we will use the same spelling the judge used.
The record indicates that Goncalves looked at approximately 1,500 photographs while he was at the police station.
In concluding that the identification of the defendant was reliable, the judge cited Goncalves’s good opportunity to view his assailants, his certainty in identifying the defendant, and his rejection of hundreds of other photographs as well as suspects presented to him prior to his identification of the defendant.
This test, developed in Neil v. Biggers, 409 U.S. 188 (1972), and Manson v. Brathwaite, 432 U.S. 98 (1977), is sometimes also referred to as the “totality” test or the “totality of the circumstances” test.
States may decline to follow Brathwaite and its predecessor, Neil v. Biggers, supra, and adopt more stringent due process requirements under their own Constitutions or through their own Legislatures. Manson v. Brathwaite, supra at 118 (Stevens, J., concurring). Id. at 128-129 (Marshall, J., dissenting). See Commonwealth v. Henderson, 411 Mass. 309, 311 (1991); State v. Leclair, 118 N.H. 214, 218 (1978). See also Note, Twenty-Years of Diminishing Protection: A Proposal To Return To the Wade Trilogy’s Standards, 15 Hofstra L. Rev. 583, 596 (1987); Brennan, State Constitutions and the Protection of Individual Rights, 90 Harv. L. Rev. 489 (1977).
New York, for example, has interpreted its State Constitution as requiring the use of the pre-Brathwaite per se exclusion standard. People v. Adams, 53 N.Y.2d 241 (1981). See People v. Edmonson, 75 N.Y.2d 672, cert. denied, 498 U.S. 1001 (1990); People v. Sapp, 98 A.D.2d 784 (N.Y. 1983); People v. Tatum, 192 Misc. 2d 196 (N.Y. Sup. Ct. 1985). In addition, Utah, while applying a “reliability” test, does not explicitly follow Brathwaite and Biggers because it has concluded that the Supreme Court’s test does not provide adequate protection against mistaken identifications. See State v. Ramirez, 817 P.2d 774 (Utah 1991); State v. Long, 721 P.2d 483 (Utah 1986).
See United States v. Wade, 388 U.S. 218 (1967); Gilbert v. California, 388 U.S. 263 (1967); Stovall v. Denno, 388 U.S. 293 (1967). See also Simmons v. United States, 390 U.S. 377 (1968).
The rule of per se exclusion was also followed in Massachusetts prior to Commonwealth v. Botelho, 369 Mass. 860 (1976), in cases decided subsequent to the Supreme Court’s Wade trilogy. See Commonwealth v. Ferguson, 365 Mass. 1, 6-7 (1974); Commonwealth v. Ross, 361 Mass. 665, 670-676 (1972), judgment vacated on other grounds, 410 U.S. 901, aff'd on remand, 363 Mass. 665, cert. denied, 414 U.S. 1080 (1973); Commonwealth v. McGrath, 361 Mass. 431, 434-438 (1972).
“It would be open to the defendant, however, if he chose, to attempt to attack and weaken the prosecution’s case at trial by introducing the suggestive confrontation and arguing to the trier that it did corrupt any later identifications including the in-court identification.” Commonwealth v. Botelho, supra at 866.
See, e.g., State v. Long, 721 P.2d 483, 488-491 (Utah 1986); Gross-man, supra at 70-96; Gross, supra at 408-432; N. Sobel, Eyewitness Identification § 3.3(a) (Supp. 1992); L. Taylor, Eyewitness Identification (1982); Jonakait, Reliable Identification: Could the Supreme Court Tell in Manson v. Brathwaite?, 52 U. Colo. L. Rev. 511 (1981); E.F. Loftus, Eyewitness Testimony (1979); Wells & E.F. Loftus, Eyewitness Testimony: Psychological Perspectives (1984); W.R. LaFave & J.H. Israel, Criminal Procedure §§ 7.1, 7.4 (1984 ed. & 1991 Supp.)
See, e.g., Sherwood, supra, 30 How. L.J. 439 (1987); Rosenberg, Rethinking the Right to Due Process in Connection with Pretrial Identification Procedures: An Analysis and a Proposal, 79 Ky. L.J. 259 (1990); Note, Twenty-Years of Diminishing Protection, 15 Hofstra L. Rev. 583 (1987); Grossman, supra at 53. See also N. Sobel, Eyewitness Identification § 3.3(b) (Supp. 1992).
As noted earlier, Goncalves’s original description of the male assailant, given to police on the night of the incident, differed in several respects from the defendant’s actual appearance. The defendant weighed 220 pounds, while Goncalves described his assailant as weighing 170 pounds with a medium build. The defendant was thirty-seven years of age while Goncalves described his assailant as between twenty-seven and thirty years old. Finally, the defendant was missing several teeth and had a moustache but Goncalves mentioned neither of these characteristics in his original description.
At the hearing on the motion to suppress, when asked to describe what he remembered about his assailant’s appearance, Goncalves testified that he was in his late thirties, weighed about 200 pounds (“maybe more”), and had a large build. Goncalves also “remembered” while testifying that his assailant was missing some of his teeth.