The single issue in this case is whether the trial judge erred in excluding expert testimony offered to assist the jury in assessing the reliability of eyewitness testimony identifying the defendant as the man who, on July 10, 1976, robbed Alexander’s Supermarket in Dracut. The defendant appeals from his convictions of armed robbery and of assault and battery by means of a dangerous weapon. We affirm the convictions.
Although the armed robbery occurred in July, 1976, the defendant was not arrested and indicted until 1980. A trial in November, 1980, ended in a mistrial when the jury could not agree. The defendant was convicted at a second trial in October, 1981. The chief witnesses against him were Con*90rad Berube, assistant manager of the market, and Betty Southworth, the courtesy booth clerk. A salient fact in the defendant’s attempt to introduce expert testimony bearing on the reliability of the eyewitnesses’ identification of him is that Southworth had testified under oath at the 1980 trial and earlier at a probable cause hearing and before the grand jury that the person who robbed the store wore a short-sleeved shirt and had no distinctive features. At the time of the robbery the defendant had tattoos on his arms. The 1980 jury that could not agree on a verdict saw the tattoos. After the 1980 trial, Southworth, who was a sequestered witness at that trial, learned of the tattoos and shortly thereafter advised an assistant district attorney that she remembered that the robber was wearing a long-sleeved jacket. In February, 1981, the assistant district attorney, complying with the allowance of a motion for disclosure of exculpatory evidence, advised the defense counsel of Southworth’s changed “memory.”1
The defendant then moved for extra fees and costs for expert assistance to assess the eyewitness testimony. A Superi- *91or Court judge denied the motion because the defendant had not shown “what conclusions [the expert] would draw which would not be apparent to the average jury,” and, because the evidence would not be admissible, a litigant who could afford to pay would not hire the services of an expert. A Justice of the Appeals Court vacated the order denying the motion and indicated that a motion for extra fees and costs should be allowed. The Justice’s memorandum of decision regarded the defendant’s request as directed not at the general question of the reliability of eyewitness testimony but specifically toward explaining what effect post-event information concerning the defendant’s tattoos may have had on South worth’s current memory.
At the trial Berube identified the defendant as the man who approached him at 8:45 p.m. on July 10, 1976, as he was standing near the market’s courtesy booth talking to South worth. Berube said he was very nervous, too nervous to remember what the robber was wearing or any distinctive features of the robber. The gun in Berube’s stomach “looked like a cannon.” South worth then testified that the defendant was the man who robbed her. She said that the robber wore a dark-colored jersey and took a gun from underneath a jacket he was wearing. On cross-examination, she admitted that she had made prior inconsistent statements concerning the robber’s clothing. Before the grand jury she had testified that the man wore a short-sleeved jersey and pulled a revolver out from his shirt. She further admitted that, when she testified under oath in November, 1980, she said the man wore no jacket. On these occasions she made no reference to tattoos or other distinctive features. She then testified, on further cross-examination, that after the 1980 trial she overheard a conversation between the assistant district attorney and a police inspector and learned that the defendant had tattoos on his arms. Defense counsel pressed the cross-examination, developing the *92inconsistency and questioning the reason why Southworth “put a jacket on the man.”2
*93The night of the robbery Berube and Southworth went to the Lowell police station and looked at photographs. They selected a photograph of the defendant. While looking at numerous books of photographs, Berube and Southworth discussed various photographs, and together they identified a photograph of the defendant. Evidence of the photographic identifications and the circumstances of the identification were introduced through these two witnesses and police detectives.
The defendant presented alibi witnesses tending to show that on the night of the crime he attended a family party in Rhode Island. He exhibited his arms to the jury. The judge then conducted a voir dire to determine whether the testimony of Elizabeth Loftus, Ph.D., and Margaret Hagen, Ph.D., should be admitted. Dr. Loftus has substantial credentials as an experimental psychologist specializing in human memory and perception. She testified that at a high level of stress a person tends to have a reduced ability to remember what he observes. A person focused on a weapon, for example, has a reduced ability to identify the face of the person holding that weapon. She noted that the period *94of time between an event and one’s attempt to recall it, the retention interval, is significant because, as a matter of common sense, the longer that interval, the less accurate the memory. Dr. Loftus testified that individuals are susceptible to the influence of post-event information, information supplied to a witness after an event is over. The witness can incorporate post-event information in his memory, thus altering prior memory. She also testified that the introduction of post-event information four years after the event could easily influence memory for a particular detail. Where post-event information conflicts with what the witness has already recalled, the witness will attempt to resolve the conflict. One way to do that “is to have a reorganization in memory that allows for this conflict.” The witness is influenced “rather unconsciously” by post-event information and is not lying when the memory becomes altered. There is little relationship between the certainty and the accuracy of witness’s identification.
Dr. Hagen, an associate professor at Boston University in the department of psychiatry, testified on voir dire that the principles stated by Dr. Loftus concerning the acquisition, retention, and retrieval of information were generally accepted in the field.
The judge ruled that the expert testimony would not be admitted. He found Dr. Loftus to be “eminently well qualified as an expert in psychology; to wit memory and perception.” However, he concluded that “[t]he proposed testimony is not beyond the ordinary experience and knowledge of the average juror and would not aid jurors in their deliberations” (emphasis in original). He further concluded that “[t]he proposed testimony is not based on scientific evidence.” The defendant then rested and the case was argued to the jury.
Defense counsel presented the jury with a well-construe-ted argument that Berube’s and Southworth’s identifications of the defendant were wrong. She focused specifically on Southworth’s change of testimony about whether the *95robber was wearing a jacket.3 The jury, however, returned verdicts of guilty.
The issue of the admissibility of expert testimony offered to show the unreliability of eyewitness identification has received increased attention in State courts in recent years. State court opinions almost uniformly have upheld the trial judge’s exercise of discretion to exclude such testimony. See, e.g., People v. Plasencia, 140 Cal. App. 3d 853, 858-859 (1983); Dyas v. United States, 376 A.2d 827, 831-832 (D.C.), cert. denied, 434 U.S. 973 (1977); Jones v. State, 232 Ga. 762, 763-766 (1974); State v. Hoisington, 104 Idaho 153, 165 (1983); State v. Warren, 230 Kan. 385, 393-395 (1981); State v. Stucke, 419 So.2d 939, 944-945 (La. 1982); State v. Fernald, 397 A.2d 194, 197 (Me. 1979); State v. Helterbridle, 301 N.W.2d 545, 547 (Minn. 1980); State v. Porraro, 121 R.I. 882, 892 (1979); State v. Onorato, 142 Vt. 99, 104 (1982). Cf. State v. Galloway, 275 N.W.2d 736, 738-739 (Iowa 1979) (no abuse of discretion in excluding expert witness’s explanation that her opinion concerning the possibility of misidentification was based in part on a particular experiment performed by another); Hampton v. States 92 Wis. 2d 450, 454-459 (1979) (no abuse of discretion in not permitting a psychologist to give an opinion as to the reliability of a witness’s identification of the defendant, *96after he was permitted to testify to factors affecting human perception).4 Contra State v. Chappie, 135 Ariz. 281, 296-297 (1983).5 Some opinions have expressed concern about the aura of reliability that the jury may find in expert testimony bearing directly on their own fact-finding role. State court opinions generally note that the matter is within the jury’s knowledge and that the defendants’ rights can be protected by cross-examination and appropriate jury instructions. They give no particular attention to the circumstances of the eyewitness identification, such as whether a weapon was used, whether the witness was a victim of the crime, or whether the witness made a photographic identification. This is understandable because generally the expert opinions offered did not focus on the capabilities of the particular witness who had identified the defendant but rather dealt generally with problems of *97identification under stress and with the ability to recall events.
The Federal cases addressing the same issue uniformly affirm the trial judge’s exercise of discretion to exclude such testimony. See United States v. Thevis, 665 F.2d 616, 641 (5th Cir.), cert. denied, 459 U.S. 825 (1982); United States v. Fosher, 590 F.2d 381, 382-384 (1st Cir. 1979); United States v. Watson, 587 F.2d 365, 368-369 (7th Cir. 1978), cert. denied sub nom. Davis v. United States, 439 U.S. 1132 (1979); United States v. Brown, 540 F.2d 1048, 1053-1054 (10th Cir. 1976), cert. denied, 429 U.S. 1100 (1977); United States v. Brown, 501 F.2d 146, 150-151 (9th Cir. 1974), rev’d on other grounds sub nom. United States v. Nobles, 422 U.S. 225 (1975); United States v. Amaral, 488 F.2d 1148, 1152-1153 (9th Cir. 1973); United States v. Collins, 395 F. Supp. 629 (M.D. Pa.), aff’d, 523 F.2d 1051 (3d Cir. 1975). The basic reasoning in the Federal cases is the same whether they deal with matters arising before or after the effective date of Rule 702 of the Federal Rules of Evidence (concerning the admissibility of expert testimony). See Proposed Mass. R. Evid. 702 (1980). A trial judge could reasonably conclude in his discretion that expert testimony would not be likely to add to common understanding of the issue of identification. In short, the subject could reasonably be seen as being within the knowledge of jurors, and the testimony not apt to assist them in determining the identification question.
In Massachusetts, the issue of the exclusion of expert testimony concerning the reliability of eyewitness identification has followed the general pattern. We have upheld, as within the judge’s discretion, the exclusion of testimony of a psychologist and a psychiatrist concerning the capacity of identifying witnesses to observe accurately and to remember correctly. Commonwealth v. Jones, 362 Mass. 497, 502 (1972). In the Jones case, however, the trier of fact was a judge and thus the holding could be limited to deciding that the problems of identification were within the *98common experience of a judge, as opposed to a jury. In Commonwealth v. Middleton, 6 Mass. App. Ct. 902 (1978) (rescript), the Appeals Court found no abuse of discretion in the exclusion of expert testimony concerning factors affecting the reliability of eyewitness identification, including stress and the cross-racial nature of the identification. Recently, in Commonwealth v. Sowers, 388 Mass. 207, 215-218 (1983), we were careful to determine that the jury were not likely to have been misled by the testimony of an ophthalmologist that a legally blind, identifying witness could “identify” the defendant, in the sense of being capable of seeing him.
Our traditional view of what is a proper subject of expert testimony has been that, if other criteria are met, such testimony is admissible if, in the judge’s discretion, the subject is not within the common knowledge or common experience óf the jury. See New England Glass Co. v. Lovell, 7 Cush. 319, 321-322 (1851). See also Commonwealth v. Boyle, 346 Mass. 1, 4 (1963); Commonwealth v. Makarewicz, 333 Mass. 575, 591 (1956); Sargent v. Massachusetts Accident Co., 307 Mass. 246, 249 (1940). We have generally upheld the exclusion of an expert opinion based on a trial judge’s determination that the subject matter was within the common knowledge and experience of the trier of fact. See, e.g., Supreme Malt Prods. Co. v. Alcoholic Beverages Control Comm’n, 334 Mass. 59, 63-64 (1956); Johnson v. Orange, 320 Mass. 336, 338 (1946); Meehan v. Holyoke St. Ry., 186 Mass. 511, 514 (1904); Perkins v. Augusta Ins. & Banking Co., 10 Gray 312, 324 (1858). In recent years, without changing basic principles, we have focused on the question whether, in the wide discretion of the trial judge, the subject was one on which the opinion of an expert would have been of assistance to the jury. See Commonwealth v. Gaulden, 383 Mass. 543, 549 (1981) (expert testimony could be “of some assistance to the jury in understanding” how a lock worked); Bernier v. Boston Edison Co., 380 Mass. 372, 384 (1980) (“opinions useful to juries”); Commonwealth v. Fournier, 372 Mass. 346, 350 (1977) (the expert testimony *99gave the jury “appreciable assistance”); Commonwealth v. Boyd, 367 Mass. 169, 182 (1975) (same). See also 7 J. Wigmore, Evidence § 1923 (Chadbourn rev. 1978) (“But the only true criterion is: On this subject can a jury receive from this person appreciable help?”); Proposed Mass. R. Evid. 702 (1980) (“If . . . specialized knowledge will assist the trier of fact to understand the evidence or to determine a fact in issue . . .”).
The line between common experience and knowledge and matters known only to experts varies with time and circumstances. At any one time, the transition from one type of knowledge to the other is often gradual and cannot be defined precisely. Indeed, there are matters generally within the common experience of lay people that have specialized aspects known only to experts. See 3 J. Weinstein & M. Berger, Evidence par. 702[02] (1982). The decision to admit or to exclude particular expert testimony depends on the circumstances of the case. When the question whether expert testimony would aid the jury is close, the likelihood of prejudice from the admission or exclusion of that testimony is slight. It is not surprising, therefore, that appellate courts have given great deference to the rulings of trial judges in this area of the law of evidence. There have been, however, some occasions in which this court has found reversible error in a trial judge’s determination that a subject was or was not one on which expert testimony should be admitted. See e.g., Commonwealth v. Gardner, 350 Mass. 664, 666-667 (1966); Doherty v. Booth, 200 Mass. 522, 526 (1909).6
*100One of the most troublesome problems in the administration of criminal justice is the possibility that eyewitness identification testimony is wrong, given by a witness who sincerely but wrongly believes in the accuracy of his or her identification. Substantial constitutional protections have been developed in an attempt to reduce the likelihood of eyewitness misidentifications. See Simmons v. United States, 390 U.S. 377 (1968) (photographic identification); Stovall v. Denno, 388 U.S. 293 (1967) (showup); United States v. Wade, 388 U.S. 218 (1967) (lineup). Courts generally have relied on the cross-examination of identifying witnesses and on the final argument of defense counsel to develop for the trier of fact any weaknesses in identification testimony sufficient to create a reasonable doubt. Further, we have said that, if a defendant fairly raises the issue of mistaken identification, appropriate instructions should be given. Commonwealth v. Rodriguez, 378 Mass. 296, 302 (1979). We have thus placed reliance on jury instructions that a witness might honestly have been mistaken in his or her identification of the defendant and advising the jury to consider a number of factors in appraising identification testimony. Id. at 310-311.7
On the other hand, like almost all other appellate courts, we have not regarded expert testimony concerning eyewitness identification as a standard safeguard against misidentification testimony. Such testimony can prolong a trial, raise collateral issues, and significantly increase the cost of trial. However, these reasons alone would not justify the exclusion of evidence that would assist a trier of fact. The problem is more fundamental.
Experts testifying on the reliability of eyewitness identification may appear to be intruding into an area traditionally and exclusively the function of the jury. We look to the *101jury after an adversary trial to make the decision as to what testimony to believe. We permit, indeed require, the judge to instruct the jury concerning factors that bear on the reliability of eyewitness identification. Such a charge may be a checklist of relevant considerations, but it cannot be instructive on what weight to give those various factors in the circumstances. It is at this point that one may fairly contend that the jury would be aided by expert testimony. But that expert testimony, at least in this case and in most cases, deals with general principles, such as the fact that memories fade over time, that people under severe stress do not acquire information as well as alert persons not under stress, and that people tend unconsciously to resolve apparent inconsistencies between their memories and after-acquired facts. Obviously there are aspects of these general principles on which experts might make some contribution in particular cases. However, juries are not without a general understanding of these principles and, as the trial of this case demonstrates, they see the possible application of these principles in concrete circumstances. The jury have the opportunity to assess the witnesses’ credibility on the basis of what is presented at trial and not solely on general principles.
We conclude that trial judges should continue to have discretion in these matters. Thus we see no error of law in excluding the expert testimony. The case has not been made that the introduction of the testimony of Dr. Loftus would have assisted the jury in their difficult task. The general points of Dr. Loftus’s proposed testimony were covered by defense counsel’s carefully prepared cross-examination and final argument. The introduction of such testimony would have prolonged the trial and might have diverted the jury’s attention from its basic task.8 We decline to adopt a rule *102of law that would generally require the admission of expert testimony offered on eyewitness identification.9 We see no special circumstances in this case that would justify an exception for expert testimony on the effects of postevent information on memory.
Judgments affirmed.