GIBBS C.J. This is an application for special leave to appeal from a the applicant from his conviction on a charge that he entered as a trespasser a building with intent to steal therefrom. The main issue at the trial was whether the applicant was correctly identified as the man seen by a number of witnesses in circumstances that connected him with the crime. Each of the witnesses had first identified the applicant by looking at a folder containing about a dozen photographs, one of which was a photograph of the applicant. The photographs came from the possession of the police. The first witness, Senior Constable Beale, was shown the photographs soon after the crime was committed. The applicant had not then been arrested, although he was suspected of complicity in the crime. The other witnesses, Senior Constable Williams, Mrs Fedele and Mr Connell, were shown the photographs after the applicant had been arrested. No identification parade was held, although no valid reason has been advanced for failing to arrange such a parade in the case of the three witnesses last mentioned. At the trial, Senior Constable Beale, Senior Constable Williams and Mrs Fedele gave evidence identifying the applicant. Mr Connell gave evidence that he had been shown the photographs and had then identified a person concerned in the purchase of a car used in the crime. He gave evidence also that he had made a statement to the police indicating which of the photographs was that of the person concerned. It was open to the jury further to accept portion of his evidence in which, after having refreshed his memory from the statement, he swore that he had identified a particular photograph which was in fact that of the applicant, although he had no independent recollection of having done so. On the other hand the jury may not have been satisfied that at the trial Mr Connell was able to be sure which photograph he had identified, and the argument for the applicant was put forward on that basis. Evidence was given by Detective Sergeant Duggan that the photograph which Mr Connell had identified was that of the applicant. (at p397)

2. The first submission on behalf of the applicant was that it is improper and unfair for police officers to use police photographs for the purpose of obtaining evidence of identification for use in a criminal prosecution, when the person to be identified has been arrested or is in police custody or is suspected of having committed the crime. In such a case, so it was said, the proper course is to hold an identification parade, at which the witness can be asked to pick the accused from a sufficient number of more or less similar persons. It was conceded that the learned trial judge had a discretion to admit identification evidence based on police photographs, but it was submitted that in the proper exercise of his discretion he should have rejected the evidence. Alternatively, it was submitted that if such evidence is allowed, the photographs themselves should not be admitted. Finally, it was submitted that the evidence of Detective Sergeant Duggan, that Mr Connell identified the applicant, was inadmissible unless Mr Connell himself gave evidence at the trial that he had made the earlier identification. (at p398)

3. Evidence given by a witness identifying an accused as the person whom he saw at the scene of the crime or in circumstances connected with the crime will generally be of very little value if the witness has not seen the accused since the events in question and is asked to identify him for the first time in the dock, at least when the witness has not, by reason of previous knowledge or association, become familiar with the appearance of the accused. The reasons for this were explained in Davies and Cody v. The King [1937] HCA 27; (1937) 57 CLR 170, at pp 181-182 . In particular there is the danger that the witness will too readily come to believe, without any true recollection, that the man charged is the man whom he had previously seen, particularly if his own memory has become dim and there is some resemblance between the two men. The courts in England and Australia have long recognized the danger of acting upon evidence of identification made in those circumstances. It has accordingly become established practice for a witness to be asked to identify the accused at the earliest possible opportunity after the event, and for evidence to be given of that act of identification. Such evidence is, in practice, given not only by the person who made the identification but also by persons who saw it made. (at p399)

4. In theory the manner in which an accused was identified out of court goes to the weight rather than to the admissibility of the evidence. However, the objections to the evidence of an identification made of an accused person when he is in the dock are almost equally open to evidence of the identification of an accused person which is given by a witness who has been shown the accused alone and as a suspect, and in Davies and Cody v. The King [1937] HCA 27; (1937) 57 CLR 170 it was held that a conviction based on evidence of such a witness should be quashed as unsafe unless the identity of the accused was further proved by other evidence. The Court went on to say (1937) 57 CLR, at p 182 :

"Where that further evidence consists in or includes other witnesses whose identification has been of the same kind, the number of witnesses, their opportunities of obtaining an impression or knowledge of the prisoner and other circumstances in the case must be taken into account by the court of criminal appeal for the purpose of deciding whether on the whole case the possibility of error is so substantial as to make the conviction unsafe." (at p399)

5. The safest and most satisfactory way of ensuring that a witness makes an accurate identification is by arranging for the witness to pick out from a group the person whom he saw on the occasion relevant to the crime. If an identification parade is held for that purpose, it goes without saying that precautions must be taken to ensure that no prompting, suggestion or hint is given to the witness that any particular member of the group is the suspect. For example, it would be unfair and improper to show to a witness, before the identification parade was held, a single photograph of a person who was said to be the suspect, and it would be unsafe to act on evidence of identification given in those circumstances: Reg. v. Russell (1977) 2 NZLR 20, at p 27 . Indeed, where a suspect had been arrested, and it was intended to ask a witness to attempt to identify him at an identification parade, it would be unfair to show the witness, before the parade, a number of photographs including that of the suspect: R. v. Goss (1923) 17 Cr App R 196 ; R. v. Haslam (1925) 19 Cr App R 59 . On the other than it may be necessary for a police officer to show a number of photographs to a witness in an attempt to obtain information as to the identity of an offender; if such witness did identify the offender from a photograph, it would not necessarily be unfair for that witness later to be asked to select the offender from a group at an identification parade, but the fact that the witness had seen the photograph might affect the value of the later identification at the parade: see R. v. Dwyer and Ferguson (1925) 2 KB 799 . (at p399)

6. The value of holding an identification parade is not only that, if properly carried out, it provides the most reliable method of identification, but also that it is necessarily held in the presence of the accused, who is thereby enabled to observe, and later bring to light, any unfairness in the way in which the parade was conducted, or any weakness in the way in which the witness made the identification. However, as a matter of legal principle, it seems to me impossible to say that the admissibility of evidence of a prior act of identification depends on the fact that an identification parade was held. As a matter of law it would be equally admissible to prove that an identification was made by a witness who was shown a collection of photographs and selected one which he said was the photograph of the person concerned. There are, however, two grounds of objection to the proof of identification by means of police photographs. In the first place, the accused will of necessity be absent when the identification is made, and has no means of knowing whether there was any unfairness in the process or whether the witness was convincing in the way in which he made the identification. Secondly, the production in evidence at the trial of photographs coming from the possession of the police is very likely to suggest to the jury that the person photographed had a police record, probably for offences of the kind in question. (at p400)

7. For these reasons, it is most undesirable that police officers who have arrested a person on a charge of having committed a crime should arrange for potential witnesses to identify that person except at a properly conducted indentification parade. Similarly, speaking generally, an identification parade should, wherever possible, be held when it is desired that a witness should identify a person who is firmly suspected to be the offender. However, there is little support to be found in the authorities for the view that a conviction must necessarily be quashed if it is based on evidence that the accused was identified other than at an identification parade at a time when he had been charged or was definitely suspected, even though there was no valid reason why an identification parade could not have been arranged. The judgment of this Court in Davies and Cody v. The King [1937] HCA 27; (1937) 57 CLR 170 suggests that the proper approach is to consider whether the conviction can safely be sustained on the whole of the evidence. In England, although the courts constantly insist on the importance of holding an identification parade, the Court of Criminal Appeal in Reg. v. Seiga (1961) 45 Cr App R 220 dismissed an appeal against conviction in a case where photographs were used to identify the accused although he had been arrested and no reason was given why there should not have been an identification parade. In that case the witness who identified the accused in court had previously been shown, by a police officer, a group of photographs, including one of the accused. The Court of Criminal Appeal disapproved of the conduct of the police officer, but nevertheless sustained the conviction. In Australia, the question arose in Reg. v. Bouquet (1962) SR (NSW) 563 where the accused was identified by a witness who was shown a number of photographs after the accused had been arrested. An appeal against conviction was allowed on other grounds, but Sugerman J. (1962) SR (NSW), at p 568 said that the use of photographs instead of an identification parade goes to the weight and sufficiency of the evidence rather than its admissibility and may be specially significant when there is no other evidence identifying the accused. Wallace J. (1962) SR (NSW), at p 574 , at p. 574, said that the procedure of holding an identification parade should be followed except in special circumstances. The other Australian cases were not so directly concerned with this question. In Reg. v. Goode (1970) SASR 69 the accused was identified from photographs before his arrest, and no parade was subsequently held; the Court quashed the conviction because the warning given to the jury was insufficient. In R. v. Fannon and Walsh (1922) 22 SR (NSW) 427 and Reg v. Doyle [1967] VicRp 82; (1967) VR 698 , and also in the New Zealand case Reg. v. Russell (1977) 2 NZLR 20 , identification parades were held, but only after the identifying witness had identified the accused from a number of photographs produced by the police. In Reg. v. Doyle and Reg. v. Russell the photographs were shown to the witnesses during the investigation of the crime and before the arrest of the suspect; the report in R. v. Fannon and Walsh does not make it clear at what stage the photographs were shown. In all these cases the evidence of identification by means of the photographs was held admissible and the convictions were upheld. In Reg. v. Russell (1977) 2 NZLR, at p 27 , Richmond P. said:

". . . we respectfully agree with what was said in Reg. v. Doyle that evidence of identification by photograph is legally admissible and relevant. The real question in all cases is whether or not the trial judge ought to have exercised in favour of the accused his discretion to exclude admissible and relevant evidence on the ground that its prejudicial effect is out of proportion to its true evidential value, or on general grounds of 'unfairness'."

Finally, reference should be made to the Canadian case R. v. Bagley (1926) 3 DLR 717 . In that case, after the arrest of the accused, certain witnesses were shown a bundle of photographs, and later identified the accused at an identification parade. The conviction was affirmed. Macdonald C.J.A., in his dissenting judgment which is mentioned in Davies and Cody v. The King (1937) 57 CLR, at p 183 , said that he did not regard the question as one of admissibility of evidence but rather of weight, and that he would have sustained the verdict had the charge to the jury been sufficient (1926) 3 DLR, at p 719 . (at p401)

8. The authorities support the conclusion that I have reached, which is that, as a matter of law, evidence of an identification made out of court by the use of photographs produced by the police is admissible. However, a trial judge has a discretion to exclude any evidence if the strict rules of admissibility operate unfairly against the accused. It would be right to exercise that discretion in any case in which the judge was opinion that the evidence had little weight but was likely to be gravely prejudicial to the accused. In a case such as the present it seems to me proper for a trial judge, in deciding how he should exercise his discretion, to take into consideration that it is the duty of police officers investigating crime to take every precaution reasonably available to guard against the miscarriages of justice that can occur, and have in fact occurred, because of honest but mistaken evidence of identification, and that for this reason "only in exceptional cases should photographs be used at a stage when some particular person is directly suspected by the police and they are able to arrange an identification parade or some other satisfactory alternative means whereby the witness can be asked directly to identify the suspected person": Reg. v. Russell (1977) 2 NZLR, at p 28 . If the trial judge admits the evidence, and the accused is convicted, the true question for the Court of Criminal Appeal is whether having regard to the whole of the evidence it would be so unsafe or unsatisfactory to allow the conviction to stand that to do so would amount to a miscarriage of justice. In considering that matter the Court of Criminal Appeal also will keep in mind the importance of ensuring that the most reliable evidence of identification is obtained in every case. (at p402)

9. For the reasons I have given, I conclude that the evidence of the identifying witnesses in the present case was not inadmissible and that the learned trial judge was entitled to admit it in the exercise of his discretion. Once that evidence was admitted there was no reason to hold that the photographs themselves were inadmissible. There was no suggestion that there was anything objectionable about the photographs themselves - for example, they did not show the applicant wearing prison uniform or in handcuffs - and once evidence had been given that identification had been made by means of police photographs it does not seem to me to have been unfair to admit the photographs themselves, although again the trial judge had a discretion to exclude them if he considered that their prejudicial effect outweighed their probative value. (at p403)

10. The question whether it was admissible for Detective Sergeant Duggan to give evidence that Mr Connell had identified the photograph of the applicant is one of some difficulty. Where a witness who gives evidence in court identifying an accused swears that he identified the accused on an earlier occasion, the latter evidence is clearly admissible. It is relevant thar the witness identified the accused on the earlier occasion, and, since identification is an act of the mind, the evidence of the witness as to his own state of mind on the earlier occasion is original evidence and not hearsay. Where the identifying witness himself gives evidence of his own earlier act of identification, evidence as to that act by another person who was present at the time is also admissible. As I have already said, such evidence is constantly admitted in practice, and it is recognized by high authority that is is properly admitted: see R. v. Christie (1914) AC 545 . It is, however, not altogether clear what principle justifies the admission of the evidence of the person who observed another identify the accused on some occasion before the trial. If that evidence were admitted as tending to prove the fact that the earlier identification was made, the evidence would be hearsay, and would be admitted pursuant to some exception to the hearsay rule which has not yet been precisely formulated. Normally a person who identifies another will signify by words that the person whom he identifies is the person whom he saw on the occasion connected with the crime - he will use words to the effect, "that is the man" - but even if he does no more than point to the person identified that also expresses his state of mind just as much as words would have done. The evidence of an observer, whether of the statement or of the act which was equivalent to a statement, will be hearsay if its object is to establish that the statement, or the equivalent action, was true, that is, that the man identified was the man seen committing the crime or in circumstances connected with the crime. No doubt in some cases the fact that one person had identified another would be material in itself - perhaps because a reward had been offered to a person making an identification or because it was in issue whether the person making the identification had knowledge that the person identified had committed a crime; in those cases, evidence of what the person making the identification said while doing so would be original evidence. But where the question is whether the accused is the person who committed the crime, the materiality of statements identifying the accused is not simply that the statements were made, and evidence of the making of the statements is accordingly not original evidence. (at p403)

11. If the evidence of the prior act of identification is not tendered to prove that the identification was in fact made - if it is not tendered as an exception to the hearsay rule - the purpose of its admission must be to confirm the testimony of the identifying witness by showing its consistency with his previous act. (I am of course not dealing with those cases where the evidence may be admitted as part of the res gesta, or where the accused has adopted as true the statements made in identifying him.) In other words it seems to be admitted by analogy with the rule that allows evidence of complaints to be given in sexual cases or with the rule allowing proof of previous consistent statements to answer a suggestion of late invention. Wigmore on Evidence, 3rd ed. (1940), vol. IV, s. 1130, and Professor Goodperson in (1968) Cambridge Law Journal, p. 74 et seq. particularly at p. 83, treat the evidence as admitted by analogy with the latter rule. Of course, the evidence does not come within the principle which allows evidence of consistent statements to be given in cases of late invention, as that principle is ordinarily stated (see Nominal Defendant v. Clements [1960] HCA 39; (1960) 104 CLR 476 ) and if it is admitted on this basis must come within an extension to that principle. The question by what principle the admissibility of evidence of this kind is governed may not be of great importance when the identifying witness himself gives evidence of the earlier identification, for it is not in doubt that the evidence is admissible in such a case. However, if the identifying witness does not himself swear to the prior identification, evidence from other persons that he made such an identification can hardly be admissible if the ground of its admissibility is to show the consistency of his evidence. (at p404)

12. In my opinion the authorities are opposed to the view that evidence of an act of identification made out of court is admitted as hearsay. The question arose in R. v. Christie (1914) AC 545 . There the boy who had been assaulted gave evidence identifying the accused but he did not say in evidence that soon after the event he had identified the accused as the man who had assaulted him. The boy's mother gave evidence of this earlier act of identification. The majority of the House of Lords (Viscount Haldane L.C. and Lords Dunedin, Moulton and Reading) held that the mother's evidence was inadmissible. Viscount Haldane L.C. said (1914) AC, at p 551 that if the boy had given evidence that he had identified the accused, the mother's evidence would have been admissible; it would have been relevant "to shew that the boy was able to identify at the time and to exclude the idea that the identification of the prisoner in the dock was an afterthought or a mistake." Lord Moulton said (1914) AC, at p 558 :

"Identification is an act of the mind, and the primary evidence of what was passing in the mind of a man is his own testimony, where it can be obtained. It would be very dangerous to allow evidence to be given of a man's words and actions, in order to shew by this extrinsic evidence that he identified the prisoner, if he was capable of being called as a witness and was not called to prove by direct evidence that he had thus identified him. Such a mode of proving identification would, in my opinion, be to use secondary evidence where primary evidence was obtainable, and this is contrary to the spirit of the English rules of evidence."

See also per Lord Reading (1914) AC, at p 563 . Lords Atkinson and Parker took a different view; they were of the opinion that the evidence was admissible as primary evidence of the boy's act of identification. The view of the majority in R. v. Christie (1914) AC 545 was accepted as correct by the Judicial Committee in Teper v. The Queen (1952) AC 480 , where a witness gave evidence that an unknown passer-by had identified the accused; it was held that the statement did not form part of the res gesta and was inadmissible. Lord Normand said (1952) AC, at p 488 :

"The special danger of allowing hearsay evidence for the purpose of identification requires that it shall only be allowed if it satisfies the strictest test of close association with the event in time, place and circumstances."

He went on to cite from the passage of the judgment of Lord Moulton in R. v. Christie to which I have already referred. The words of Lord Moulton in R. v. Christie were again cited with approval in Sparks v. The Queen (1964) AC 964 . In that case the girl who had been assaulted, who did not give evidence at the trial, had said that "it was a coloured boy". Evidence of this statement was held to be inadmissible. Lord Morris of Borth-y-Gest said (1964) AC, at p 981 : "There is no rule which permits the giving of hearsay evidence merely because it relates to identity." (at p405)

13. These cases establish that evidence by a witness who observed an act of identification made by another person out of court cannot be admitted as hearsay. If the witness who made the identification gives evidence identifying the accused, the evidence of the person who observed the prior act of identification will be admissible to show that the identification made by the witness in court was not an afterthought or a mistake. If the witness who made the identification does not give evidence, obviously the evidence of the person who observed the prior identification will not be admissible for that reason. In Reg. v. Osbourne (1973) QB 678 , the Court of Criminal Appeal held that it was admissible for a police officer to give evidence that a woman had picked out the accused at an identification parade, notwithstanding that the woman said in her evidence that she could not remember picking anyone out at a parade. The evidence of the police officer in that case must have been admitted as hearsay evidence of the earlier act of identification; there was no other possible ground of admissibility. The Court did not clearly state the ground on which the evidence was admitted but said that it would be artificial to excluded it (1973) QB, at pp 690-691 . In Cross on Evidence, (4th ed.) 1974, p. 51; and 2nd Aust. ed. (1978), par. 2.23, it is said that as a result of this decision the views of Lords Atkinson and Parker in R. v. Christie (1914) AC 545 may now be taken to represent the law. In Reg. v. Collings (1976) 2 NZLR 104, at p 114 McCarthy P., speaking for the Court of Appeal of New Zealand, said that that Court would, if necessary, follow Reg. v. Osbourne, but in that case the evidence of the earlier identification was not in dispute and the identifying witness herself gave evidence of the earlier identification. With all respect I cannot regard the law as settled by Reg. v. Osbourne. On the contrary, with all respect, I consider that the evidence admitted in Reg. v. Osbourne was hearsay and inadmissible. I may add that in Reg. v. McGuire (1975) 4 WWR 124 , the Court of Appeal of British Colombia rejected evidence of earlier identifications made by witnesses who in the witness box swore that the person previously identified was not the person who had committed the crime. (at p406)

14. The position is, however, different when the identifying witness says in the witness box that he did on a previous occasion identify somebody as the person connected with the crime, but that he cannot now remember who it was that he identified. It is immaterial for this purpose whether the identification was made at an identification parade or by means of photographs. In such a case in my opinion evidence is admissible to prove who was the person thus identified. Such evidence would not be hearsay: it is not tendered to prove the truth of what the identifying witness asserted on the previous occasion. If the identifying witness has sworn directly as to his previous state of mind - for example that he was of the opinion that the person connected with the crime was the person portrayed in a photograph which he indicated on the earlier occasion - the evidence as to what photograph he indicated on that occasion is not in itself evidence of identification. Such evidence is material to show that a particular photograph was indicated, not that it was correctly indicated or that the person portrayed was the person who had committed the crime. The evidence of the observer of the earlier act of identification is in such a case admitted as original evidence. It explains and gives meaning to the evidence of identification given by the identifying witness in the witness box. (at p408)

15. We were referred to the decision of the Court of Criminal Appeal in Reg. v. McLean (1967) 52 Cr App R 80 . In that case G. made a mental note of a car number and later dictated something to C. who wrote a number down on a card. C. swore that he wrote down the number that G. told him. G. could not say what he had dictated to C. nor did he see what C. wrote. It was held that the evidence of C. (who produced the card) was hearsay and inadmissible. It is not necessary to consider whether the case was correctly decided for it may be distinguished from the present. In that case the identifying witness (G.) was unable to swear that he had told C. the correct number of the car. If he had given that evidence the case might be indistinguishable from the present. (at p408)

16. In the present case, on one view, Mr Connell did give evidence that he had identified a particular photograph (namely that of the applicant) as that of the person connected with the crime. On that view there is no doubt that the evidence of Detective Sergeant Duggan was admissible. However, on the view most favourable to the applicant, Mr Connell was unable to say in court who was the man, or which was the photograph, that he had identified but he did say that he had identified a photograph as that of the man connected with the crime and that he had made a statement to the police indicating which photograph was that of the man concerned. In these circumstances the evidence of Detective Sergeant Duggan as to which was the photograph indicated by Mr Connell was admissible as original evidence. (at p408)

17. For these reasons the challenge to the admissibility of the evidence fails. It was not argued that on the evidence properly admitted it would be unsafe or unsatisfactory to convict. (at p408)

18. In view of the nature of the questions raised I would grant special leave to appeal, but I would dismiss the appeal. (at p408)

STEPHEN J. This application involves problems of photo-identification, that procedure by which someone who has seen a suspected offender but cannot put a name to him is shown photographs of persons known to others, usually the police, and by recognizing one of them as the suspected offender, identifies him. It may be used in at least two ways: first, in the detection process, by helping the police to know whom they should arrest and charge; secondly, in the evidentiary process, to provide proof in court that the accused is in fact the offender. With its use in the detection process a court has no direct concern until that use of it intrudes, as it frequently will, upon its use in the evidentiary process. With its use in the evidentiary process courts will always be concerned. (at p408)

2. The accuracy of any identification of a stranger, seen once only, is likely to be affected by the fallibility of human perception and memory. When identification is attempted with the aid of photographs, there are introduced peculiar difficulties, due to the various ways in which photographic representations differ from nature: their two dimensional and static quality, the fact that they are often in black and white and the clear and well lit picture of the subject which they usually provide. The use of photo-identification in the evidentiary process involves three further factors of a quite different kind which apply only to its use in that process. Unlike the case of an identification parade, an accused whose identity as the offender is sought to be proved at his trial by evidence of previous photo-identification is likely to know nothing at first hand of the way in which the identifying witness earlier identified his photograph as that of the offender. He must rely upon cross-examination of prosecution witnesses for knowledge of the conditions of identification and of what safeguards were taken against error on the part of the identifying witness. Again, by what may be called the "rogues' gallery" effect, evidence that the police had in their possession and showed to the identifying witness photographs of the accused may often strongly suggest to a jury that the accused has a criminal record, perhaps even a propensity to commit a crime of the kind with which he is charged. Their production in evidence, or even reference to their existence, may then be highly prejudicial to an accused. Lastly, there is the "displacement" effect. Having been shown a photograph, the memory of it may be more clearly retained than the memory of the original sighting of the offender and may, accordingly, displace that original memory. Any subsequent face-to-face identification, in court or in an identification parade, may, on the identifying witness's part, in truth involve a matching of the man so identified with the remembered photograph, which has displaced in his memory his recollection of the original sighting. (at p409)

3. For the purposes of the police in the detection process, where these three factors do not operate, the use of photo-identification is of obvious value, despite its inherent defects. It offers to them a quick and convenient means, often the only readily available means, of having an eye-witness pick out the alleged offender from what may be a very large number of possible suspects. To deny its use to the police in the detection process, where it serves a useful, perhaps even an essential, purpose, would be to prejudice the attainment of one of the ends of justice, the detection and bringing to trial of offenders. (at p409)

4. Once used in the detection process, photo-identification tends inexorably to intrude upon the trial evidence. This occurs most commonly in the following way: the eye-witness who in the detection process has, by the use of photo-identification, put a name to a previously unidentified offender is very likely to be an identifying witness, perhaps the only available one, at the ensuing trial. Yet in that role his earlier association with photo-identification will tend to infect his evidence. Its displacement effect will operate to a greater or lesser extent upon any later identification which he makes in court or of which he gives evidence. To attack his evidence of identification the defence will wish to expose this effect. But cross-examination which reveals the earlier photo-identification will be likely to subject the accused to the rogues' gallery effect. That effect will be introduced even more directly should the prosecution itself be permitted to lead evidence of the photo-identification so as to support the evidence the witness gives of the identity of the accused as the offender. In the evidentiary process, then, photo-identification has serious disadvantages, in addition to that of its inherent unreliability. (at p410)

5. However, the unacceptable price of avoiding the intrusion of photo-identification and all its consequences into the trial would be the abandonment of its use in the detection process, the police forgoing all photo-identification in their detection of crime lest it disqualify a potential identifier from giving essential evidence of identity at a subsequent trial. There is a solution to the problem, based upon a compromise. Courts may discourage its use by police after the detection process has come to an end by treating that use as likely to disqualify a witness from giving evidence of the accused's identity: they may at the same time tolerate its intrusion into the trial evidence so long as its earlier use has been in the actual detection of crime. (at p410)

6. It is just such a solution that the courts have applied, attempting to reconcile the interests of the accused with those of the community by differentiating between the detection and the evidentiary process. Photo-identification thus remains available as an effective aid in the detection of crime, but will tend to be excluded from the trial evidence unless its intrusion be justified by its earlier use having occurred in the detection process. Where it is not excluded from the trial evidence the accused obtains some protection against its prejudicial consequences by the existence of that exclusory discretion, dependent upon a balance of prejudice against probative value, always vested in trial judges, who will also, where appropriate, warn the jury of the particular dangers of this form of identification. (at p410)

7. That this has been the solution arrived at by the courts is not, so far as I am aware, anywhere explicitly stated. But a survey of authority discloses how, case by case, the problems of photo-identification have been dealt with and the decisional pattern has evolved. (at p411)

8. It was in the English Court of Criminal Appeal in the first half of this century that the problems were first encountered and resolved. The reported cases are conveniently collected in Halsbury's Laws of England, vol. 11 (4th ed.), par. 363, n. 5. The need to distinguish between the use of photo-identification in the detection process and its use in the evidentiary process was early recognized, as was its necessary intrusion into the trial evidence if originally employed at the stage of detection. R. v. Palmer (1914) 10 Cr App R 77 ; R. v. Melany (1924) 18 Cr App R 2 ; R. v. Dwyer and Ferguson (1925) 2 KB 799 ; R. v. Wainwright (1925) 19 Cr App R 52 and R. v. Hinds (1932) 24 Cr App R 6 were all cases where the use of photo-identification occurred in the detection process. In all but Palmer the earliest in time, the distinction arising between use in the detection process and use in the evidentiary process was regarded as quite critical. On the other hand, in Goss (1923) 17 Cr App R 196 , in Seiga (1961) 45 Cr App R 220 and, at least as the Court of Criminal Appeal believed at the time of initially delivering judgment, in Haslam (1925) 19 Cr App R 59 , the use of photo-identification did not occur until the evidentiary process, after the accused was in custody. It had played no part in his detection. In these latter cases the Court stigmatized its use, describing it as "not a proper one" - Goss (1923) 17 Cr App R, at p 197 ; as "indefensible" - Haslam (1925) 19 Cr App R, at p 60 and as "unsatisfactory" conduct, of which the Court "disapproves" - Seiga (1961) 45 Cr App R, at p 224 . (at p411)

9. In the first group of five cases, where photo-identification originated in the detection process, it was, of course, with its subsequent intrusion into the trial evidence that the Court was concerned. In principle that intrusion was tolerated. But, although tolerated in principle, the Court in each case regarded the trial judge as nevertheless having a discretion to exclude evidence of photo-identification where the interests of justice and fairness to the accused required this to be done. In two of the cases, Dwyer and Ferguson and Wainwright, it concluded that in the circumstances evidence of it should have been excluded. The convictions were quashed accordingly. (at p411)

10. With this may be contrasted the Court's attitude in the second group of three cases. In each of these the Court rejected its use in principle when sought to be employed, for the first time, in the evidentiary process. In each case the conviction was quashed. (at p412)

11. The factors that, throughout all these cases, were regarded by the Court of Criminal Appeal as of peculiar significance were what I have called the displacement effect and the rogues' gallery effect. They were stressed in cases in each of the two above groups. They played a predominant role in the first of those groups in determining whether evidence of photo-identification, originating in the detection process and accordingly tolerated in principle when intruded into the trial evidence, should as a matter of discretion be permitted. (at p412)

12. In the early case of Varley (1914) 10 Cr App R 125 the rogues' gallery effect is referred to when it is said, of police photographs going before the jury, that "It is almost impossible that the jury will not be influenced by the fact - which they may infer - that the prisoner is known to the police" (1914) 10 Cr App R, at p 127 . Palmer (1914) 10 Cr App R 77 also refers to this prejudicial effect as do Goss (1923) 17 Cr App R 196 , Seiga (1961) 45 Cr App R 220 and Lawrenson (1961) Crim LR 398 . In Wainwright (1925) 19 Cr App R, at p 54 , it was said that, because of their prejudicial effect, it was "unheard of" that the prosecution should itself put forward as any part of its evidence in chief the photographs used in identification. In Dwyer and Ferguson (1925) 2 KB 799 the photographs were of a particularly prejudicial character, taken of the accused in prison garb; although the photo-identification had there been used initially in the detection process their prejudicial effect was such as to result in the conviction being quashed. (at p412)

13. The displacement effect of photo-identification was regarded as little less significant in the disposal of these criminal appeals. In Goss, one of the three cases in the second group, it was that effect which caused the Court to say, in quashing the conviction, that "it is not sufficient that what is done should be fair; it must be manifestly clear that nothing was done which might seem to be unfair" (1923) 17 Cr App R, at p 197 . In Dwyer and Ferguson the importance of this effect is also stressed, as it is in Haslam (1925) 19 Cr App R 59 and in Seiga. (at p412)

14. In a number of these cases, Palmer, Hinds (1932) 24 Cr App R 6 , Lawrenson and Seiga the Criminal Court of Appeal also considered what should be said in charging the jury, once evidence of photo-identification had in fact been admitted in evidence. The conclusion of the cases is, I think, that in addition to whatever warning may in any event be appropriate because evidence of identification is critical, the trial judge should also bear in mind the special need to prevent or counteract any particular prejudicial effect to which photo-identification may give rise. In some cases this may best be done by refraining from special comment lest it attract the jury's attention to prejudicial matter otherwise likely to escape its notice. Generally, however, some special warning should be given to the jury. (at p413)

15. The Australian cases involving photo-identification reflect the pattern set by the decisions of the English Court of Criminal Appeal. The only decision of this Court which is at all in point is Davies and Cody v. The King [1937] HCA 27; (1937) 57 CLR 170 and then only for its reference (1937) 57 CLR, at p 183 to "the embarrassments arising from the use of" police photographs as a means of identification and for its citation, in this context, of the dissenting judgment of Macdonald C.J.A. in the Alberta case of R. v. Bagley (1926) 3 DLR 717 and of the judgment of Ferguson J. in R. v. Fannon and Walsh (1922) 22 SR (NSW) 427 . Because of their citation by this Court, I begin by some reference to each of these two judgments. (at p413)

16. The dissenting judgment of Macdonald C.J.A. in R. v. Bagley (1926) 3 DLR 717 refers to and adopts the reasoning of the English Court of Criminal Appeal cases. It was a case in which photo-identification was used for the first time in the evidentiary process, not having been employed in detection of the crime. The Chief Justice adopted the English distinction between that use and its use in the detection process and was emphatic in applying in all its strictness what he described as "the English rule in this regard, or what is tantamount to a rule" (1926) 3 DLR, at pp 718-719 . He went on to say (1926) 3 DLR, at p 719 :

"Identification by photograph, before the arrest, fairly conducted for the purpose of ascertaining who should be arrested has been regarded by the Court of Criminal Appeal as not open to so grave an objection as that of the character given here. The decisions of that Court appear to me to go the length of holding that evidence such as is here complained of, particularly when the charge is unsatisfactory, is utterly worthless and that a conviction founded on it alone must be set aside (R. v. Haslam (1925) 19 Cr App R 59 ).".

His Lordship was there referring to the displacement effect, which was a significant factor in Haslam. In the concluding passages of his judgment the Chief Justice refers to what warnings should have been given to the jury, although, the evidence being worthless and his Lordship having apparently earlier accepted the submission that what occurred in this case involved "a substantial wrong and a miscarriage of justice", it is perhaps not easy to see how, on his view, the evidence could ever have been allowed to go to the jury. (at p413)

17. R. v. Fannon and Walsh appears to have been a case of photo-identification in the detection process. Certainly it was in that way that Ferguson J., speaking for the New South Wales Full Court, approached the question of its reception in evidence. How evidence of the photo-identification subsequently came to intrude into the trial evidence is described (1922) 22 SR (NSW), at pp 429-430 , where reference is also made to the displacement effect to which photo-identification may give rise. Speaking of that effect, his Honour says that to avoid injustice to the accused the fact of photo-identification should be disclosed to him, so that the defence may have the opportunity of using it to attack the evidence of identification. The judgment cites Palmer (1914) 10 Cr App R 77 and Varley (1914) 10 Cr App R 125 , stresses the risk of the rogues' gallery effect and concludes that, while evidence of photo-identification cannot be altogether prohibited, the leading of evidence of it in the prosecution case should always be examined to see whether its reception has done "substantial injustice" to the accused. (at p414)

18. Two other Australian cases deal in detail with photo-identification. In Reg. v. Doyle [1967] VicRp 82; (1967) VR 698 photo-identification had been used in the detection process and, as the Victorian Court of Criminal Appeal said (1967) VR, at p 704 its use at that stage was "eminently reasonable". The identifying witness had previously known the accused, so that there was, perhaps, little risk of his evidence being affected by the displacement effect. It was, rather, the rogues' gallery effect which came to the fore when evidence of the photo-identification was admitted at the trial. The Court acknowledged the need for care in the use of photographs so that "prejudice to the accused" might be avoided (1967) VR, at p 702 and, after citing a number of the English authorities, said (1967) VR, at p 703 that "every reasonable step is to be taken to prevent prejudice to an accused by the disclosure to the jury, by reference to photo-identification, that he is, or probably is, a convicted person". However that did not mean that "admissible and relevant evidence which may have an important bearing on conviction or acquittal" should be excluded (1967) VR, at p 704 . Whether an appropriate warning to the jury was called for was a matter for the discretion of the trial judge; in some circumstances such a warning might only highlight matters of prejudice. (at p414)

19. Reg. v. Goode (1970) SASR 69 was once again a case of the use of photo-identification in the detection process. Bray C.J., speaking for the Full Court of the South Australian Supreme Court, referred to the displacement effect of the use of photographs but also to the value, sometimes indispensable, of their use where a victim cannot name the criminal. However his Honour went on to stress the prejudicial dangers of photo-identification, saying that "the jury must not be led to the belief that the accused is known unfavourably to the police and therefore guilty" (1970) SASR, at p 80 . His Honour described as a delicate task that of determining what if anything should be said to a jury by way of warning in cases of photo-identification. (at p415)

20. Reg. v. Bouquet (1962) SR (NSW) 563 is a case which involved, but the decision in which did not turn on, the question of photo-identification. In it the police use of photographs occurred after the arrest of the accused and one member of the majority, Wallace J., may, in his very brief judgment, be adverting to that fact when he says that there were "other features of this trial of an unsatisfying nature" (1962) SR (NSW), at p 574 . It is significant that his Honour immediately goes on to express the view that, except in special circumstances, identification should always be by means of an identification parade. The judgment of the other member of the majority, Sugerman J., deals with the question of photo-identification only as it may have affected the weight and sufficiency of the evidence of the victim's "in court" identification of the accused. Soon after an attack on him he identified the accused as his attacker from photographs shown to him by the police, but it does not appear how evidence of this ever emerged at the trial. The appellant's relevant objection, at trial and on appeal, was confined to what was said to be the weakening of the "in court" identification by the fact of the earlier photo-identification. It did not concern any admission in evidence either of photographs or of the fact of their earlier use in identification. Sugerman J. disposed of the issue by referring to the considerable amount of other evidence which, quite apart from the "in court" identification, proved the accused's identity as the offender. (at p416)

21. Such of these Australian cases as squarely raise the question of photo-identification conform to the principles worked out in the English authorities. The same may be said of the recent decision of the New Zealand Court of Appeal in Reg. v. Russell (1977) 2 NZLR 20 . It was a case of the use of photo-identification in the early stages of the detection process and the particular facts of the case greatly reduced the risk of there being any displacement effect, as the Court observed at pp. 27-28. Thus, in the exercise of any discretion to reject evidence, the trial judge could properly assume that the usual doubts affecting the cogency of photo-identification had little application to the particular case. It was, accordingly, to the rogues' gallery effect that the Court had occasion to refer in detail. Having made the point that when photographs from police sources are shown to a jury there will necessarily be some positive disclosure "that the accused has some sort of police record", their Honours continued (1977) 2 NZLR, at p 28 :

"All of this adds up to the fact, as is recognised in the decided cases, that great care should always be taken with the use of photographs shown to anyone who may later become a witness as to the identification of a suspected person. Further, only in exceptional cases should photographs be used at a stage when some particular person is directly suspected by the police and they are able to arrange an identification parade or some other satisfactory alternative means whereby the witness can be asked directly to identify the suspected person. When photographs have been used it is quite clear, as was accepted in the present case, that in normal cases the Crown should not produce the photographs themselves as exhibits in the course of evidence in chief." (at p416)

22. In all the circumstances the Court concluded (1977) 2 NZLR, at p 29 that "we are not prepared to say that the evidence given at the trial in relation to the use of photographs should have been excluded by the judge either on the ground of unfairness by the police or on the ground that its prejudicial effect substantially outweighed its probative value." (at p416)

23. Reg. v. Russell (1977) 2 NZLR 20 not only provides an appropriate consideration of the factors involved in the exercise of discretion where photo-identification has been used initially in the detection process; it also explores the question, assuming that it is to be received in evidence, whether and in what circumstances it should be led by the prosecution in evidence-in-chief and how, if not so led, the accused is to be made aware that it has occurred so that he may, by cross-examination, test the identification evidence for the operation of what I have called the displacement effect. (at p417)

24. In a number of other common law jurisdictions where photo-identification has been considered the principles emerging from the English Court of Criminal Appeal decisions appear to have been followed. I refer to the decision of the Irish Court of Criminal Appeal in The People (A.G.) v. Mills (1957) IR 106 and to the decision of the Ontario Court of Appeal in Reg. v. Sutton (1970) 3 CCC 152 which is useful not only for the reasoning appearing in the judgment of that Court but for its citation of authority from other Canadian jurisdictions on the point. (at p417)

25. This survey of the cases discloses that when photo-identification is used after the detection process is over, that will in itself be a strong ground for excluding from the trial all evidence of identification by a witness who has been involved in that photo-identification. The police will in such a case already know, at the time of the photo-identification, the indentity of the "wanted man". Accordingly, the particular advantages which photo-identification offers as a means of supplying that information will no longer apply. Instead it will possess only its character of a second-best mode of identification, particularly prone to error and also productive of those well-recognized consequences so likely to be prejudicial to an accused. This will in most cases be reason enough, at least where some more reliable mode of identification such as an identification parade was available, to exclude from evidence the identification testimony of a witness whose testimony has been infected by contact with photo-identification. (at p417)

26. When, on the other hand, it is in the detection process that photo-identification was used, the witness's evidence of identification may be admitted in evidence unless an application of the familiar exclusory discretion, dependent upon the weighing of prejudicial effect against probative value, should require its exclusion. This will be so whether the witness's proposed evidence is confined to "in court" identification, includes evidence of recognition at an identification parade or extends to evidence of the photo-identification itself. However, the way in which the exclusory discretion is exercised will no doubt be affected by the extent to which the particular evidence proposed to be led is likely to be prejudicial to the accused. In judging of this it is not only the evidence to be led that must be considered. It will also be relevant to take account of what prejudicial material the defence may be forced to bring out in cross-examination in its efforts to combat the evidence led. A court will necessarily bring to any consideration of the exercise of the exclusory discretion a full appreciation of the now well-recognized prejudicial consequences that may so readily flow from the receiption into evidence of photo-identification. (at p417)

27. One further general observation is called for. Whether photo-identification was employed in the detection process or only after it ended will not depend upon whether at the time the accused was already in the hands of the police. No doubt, if he was, his availability for identification at a line-up may suggest the drawing of the distinction at that point. But it is the need of the police to know who is the wanted man which justifies both the use of photo-identification in the detection process and its intrusion into evidence. It will be the existence of that same need which will determine the point at which the distinction is to be drawn between the detection and the evidentiary processes. Once the police know who they are seeking, photo-identification loses its peculiar virtue while retaining in full its particular vices. And of course police knowledge of the identity of the wanted man is not the same as police custody of him. It is the former, once acquired, that should for this purpose mark the end of the process of detection. (at p418)

28. In the present case photo-identification was much used by the police. Four of the five witnesses who for one purpose or another identified the applicant first did so when shown an album of police photographs from which they selected a picture of him. What is significant is that in the case of no less than three of them this occurred after the detection process had ended, when the applicant had been arrested and was in custody. The police at no time conducted any identification parade and their failure to do so remains unexplained. They were apparently content to rest upon photo-identification followed by a subsequent "in court" identification. (at p418)

29. These, then, are clear instances of the use of photo-identification initiated for the first time in the evidentiary process. They illustrate the absence of any proper purpose justifying the adoption of such a procedure, as they do the prejudicial consequences to which it may lead: evidence of these witnesses' photo-identifications was led by the prosecution, the album of photographs was tendered in evidence and was made available to the jury. As was observed in the judgment of the Court of Criminal Appeal, it was "altogether unlikely that the jury would not have drawn the conclusion that the photograph was one of a man known to the police". (at p418)

30. In this case, where so much depended upon the question of identification, this should, in my view, be enough to require that the applicant's conviction be set aside. That the Court of Criminal Appeal did not, in the outcome, so regard the matter was due to the view it took of the photo-identification made by a further identifying witness, the policeman Beale. At the trial it was in relation to Beale's evidence that the question first arose about the admission of evidence of photo-identification. Once evidence of his photo-identification, including the album of photographs from which he made it, came to be admitted in evidence, it was thought that little or no additional prejudice was caused by the admission of the evidence of the three other photo-identifying witnesses. What was done regarding Beale's testimony in a real sense set the pattern followed in the case of other identifying witnesses. (at p419)

31. Because Beale's photo-identification was made before the applicant was in custody, their Honours regarded that factor as decisive. But for it, they said "a strong case could have been mounted for saying that the learned trial judge should have evinced his disapproval of the police procedure by rejecting the evidence of the identification and rejecting the tender of the photograph". However, the circumstances at the time of Beale's photo-identification, only some six hours or so after discovery of the crime, strongly suggest that the process of detection was then already at an end, that the police had decided who was their "wanted man" and that all that then remained to be done was to catch and charge him. The applicant's home had been searched, his alleged partner in crime had been arrested, the police were combing the neighbourhood for the man thought to be he, his photograph had been included in an album and taken to the scene to be shown to Beale. He was in fact arrested a few hours later. (at p419)

32. I would accordingly regard the detection process as then at an end, Beale's photo-identification being but a further instance of what appears in this case to have been a routine procedure of using photo-identification for evidentiary purposes while, for unexplained reasons, wholly ignoring the far more satisfactory procedure of identification parades. This impression is, if anything, strengthened by the curious incident some six days later, the applicant then of course being in custody, when Beale was asked at police headquarters to make a further photo-identification of the applicant, apparently once more using the same album of photographs. (at p419)

33. However, much of the police evidence is not now before this Court. It is conceivable that it contains some evidence showing that when Beale made his first photo-identification this was indeed done in the detection process, the identification being calculated, in some way not presently apparent, to assist the police in apprehending the applicant. (at p420)

34. In these circumstances I shall assume, contrary to all indications, that Beale's photo-identification was made in the detection process. Even so, I conclude that upon a proper exercise of the trial judge's exclusory discretion his evidence should have been rejected. On the one hand, as the Court of Criminal Appeal has said, to admit Beale's evidence was to invite the jury to conclude that the applicant had a police record. This was prejudicial in a very high degree, particularly, perhaps, in a crime of this nature, safebreaking, which juries would, I think, be prone to regard as the work of specialist criminals experienced in their field. As has recently been pointed out in the passages which I have cited from Doyle [1967] VicRp 82; (1967) VR 698 , Goode (1970) SASR 69 and Russell (1977) 2 NZLR 20 , to suggest that the accused has a police record is the very thing which courts should seek strenuously to avoid: yet in the present case it was, as the Full Court has said, "altogether unlikely" that this would not occur. On the other hand, the quality of Beale's evidence must greatly affect its cogency. He spoke of the applicant's photograph only as "reasonably similar, anyway" to the man he had earlier seen. That man was a stranger whom he had never seen before. His opportunity of observing him had been slight: he first saw him, reflected in the rear vision mirror of the police car, as the car driven by the man approached the police car from the rear; and again as the car passed the stationary police car, with the driver of the police car and the front seat passenger of the other car between Beale and the man: all this at a time when Beale did not know that there was anything to be remarked about the man or his car and when Beale failed to draw that car to the attention of his companion in the police car. He again saw him for a few seconds while the man was making his escape on foot, pursued at a distance by police. In these circumstances I cannot share the view of the Court of Criminal Appeal that the photograph of the applicant could be regarded "almost as though the witness Beale had drawn what his eyes saw" at the scene. Photo-identification can seldom, if ever, produce that result; it could certainly not have done so in this case. (at p420)

35. All this in my view required the discretionary exclusion of Beale's evidence of identification. The fact of its admission therefore provides no support for also admitting other evidence of photo-identification, undertaken by the three other identifying witnesses when the applicant was already in custody. (at p421)

36. However, even if, contrary to my views, Beale's testimony had properly gone before the jury, that would not, I think, have justified the reception of the evidence of the three other photo-identifiers, in each case undertaken after the detection process had ended. The inherent dangers and weaknesses of photo-identification would remain. Beale's photo-identification could neither gain cogency from nor lend it to that of others, undertaken at other times, in different circumstances and with quite disparate opportunities for observation. Nor would its admission in evidence deprive other such evidence of its prejudicial character: the jury's attention would be drawn not once but time and again to the fact that the applicant had a police record. This is the very consequence which courts have said that trial judges should guard against by refraining, where appropriate, from giving a jury special warnings of the prejudicial character of photo-identification, lest they thereby attract its attention to what might otherwise pass unnoticed. (at p421)

37. Although my conclusion differs in result from that of the Court of Criminal Appeal, the point of disagreement is, on analysis, quite narrow and I share, with respect, much of the substance of their Honours' judgment. That judgment deals at length with the dangers inherent in the use of photo-identification. Their Honours say that they would strongly disapprove were a tendency to develop of using photo-identification where an identification parade might instead have been held. They would, in that event, think it necessary for the courts to reject such evidence of identification either because of the prejudice inherent in the production of such photographs or as a means of thereby putting a stop to a practice having an "unsavoury and unnecessary element of unfairness about it". Their Honours elsewhere refer to what I have called the displacement effect, observing that a consequence of Beale having undertaken a photo-identification was that, as an identifying witness, his evidence was thereafter "infected" by it. (at p421)

38. The judgment of the Court of Criminal Appeal forms, in these respects, another link in that long chain of authority to which I have referred. With one exception, in none of these authorities which have turned upon photo-identification has a conviction been allowed to stand where photo-identification has occurred after the detection process has ended. The exception is the Alberta case of R. v. Bagley (1926) 3 DLR 717 and it is significant that it was to the judgment of the dissentient Chief Justice in that case that this Court referred in its citation of R. v. Bagley in Davies and Cody v. The King [1937] HCA 27; (1937) 57 CLR 170 . (at p422)

39. I would grant special leave, allow the appeal and order a new trial. (at p422)

MASON J. The applicant for special leave to appeal was convicted of burglary in the County Court at Melbourne. His application to the Court of Criminal Appeal for special leave to appeal was dismissed. It is from that dismissal that he now seeks special leave to appeal to this Court. (at p422)

2. The Crown case was that the applicant and one Keeley had in the early hours of 15th June 1977 entered premises of the National Bank at 593 Main Road, Belgrave with intent to steal. The principal issue at the trial was whether the Crown had proved that the applicant and Keeley were the persons who entered the premises with intent to steal. The Crown case rested heavily on identification evidence which, if accepted, placed the applicant in the vicinity of a Nissan motor car in Rickson Place, Vermont at about 7 a.m. on 15th June. The car was linked with the burglary by forensic evidence. The applicant was identified at the trial as having been in the vicinity of the car in Rickson Place by two police witnesses, Constables Beale and Williams. He was also identified by another witness, Mrs. Fedele, as having been in the neighbourhood of Rickson Place at 9 a.m. on that morning. The applicant's defence was that he had not been near Vermont on the day in question. The identification evidence which I have mentioned and other identification evidence given by one John Patrick Connell linking the applicant with the purchase of the Nissan two months before the burglary was strongly challenged at the trial. (at p422)

3. The application for special leave to appeal raises questions concerning the admissibility and sufficiency of this evidence. Each of the four witnesses whom I have mentioned was permitted at the trial to give evidence that he or she had selected a photograph of the applicant from a manilla folder containing a number of police photographs of various individuals. No witness was asked to identify the applicant from an identification parade. Constable Beale identified the applicant from photographs shown to him between 8 a.m. and 9 a.m. on 15th June, before the applicant was taken into custody after 2 p.m. on that day. The photograph which Beale identified was photograph No. 6 in a folder of police photographs shown to him (Ex. "D"). Constable Williams, Mrs. Fedele and Connell were not shown the manilla folder until after the applicant was arrested and taken into custody. (at p422)

4. As it is important that the questions sought to be argued should be considered in their precise factual setting, I shall begin by stating the relevant circumstances which I take from the judgment of the Court of Criminal Appeal. In the early hours of 15th June 1977 burglars entered the National Bank premises in Main Road, Belgrave by cutting a hole in the roof. They cut open two safes in the strongroom but were disturbed by the burglar alarm about 2 a.m. (at p423)

5. At about 5 a.m. Beale and Williams went to 5 Rickson Place and spoke to Mrs. English, said to be the de facto wife of Keeley. They searched the premises. There was no male in the house. They then returned to their car and continued their observations. (at p423)

6. At about 7 a.m. Constable Beale saw a light green Nissan sedan drive approximately half-way along Rickson Place. Mrs. English came running out of the house at No. 5 Rickson Place and spoke to the two occupants of the motor car. They thereupon left the car and fled separately. Keeley was intercepted by two policemen. He was carrying a bundle of clothes which forensic evidence suggested had been worn by one of the men who burgled the bank. The other man who left the Nissan was not immediately apprehended. (at p423)

7. At about 8.50 a.m. Mrs. Halifax, who was on duty supervising a school crossing in Stevens Road, Vermont South, opposite a primary school about a block away from Rickson Place, saw a man speak to the driver of a red Falcon car in the car parking area at the school. She then saw him collect a brown parka from the roof of a Volkswagen in the school yard and enter the red Falcon alongside the driver. The man was wearing a dark polo neck jumper or skivvy. She observed him as the car passed her at the school crossing. She described him as being between forty and fifty years of age with dark hair, but at the trial she was unable to identify the man. (at p423)

8. Mrs. Fedele, the driver of the Falcon, gave evidence at the trial that she had driven her child to the school where a man asked her would she drive him to the shopping centre at Vermont South as he had a flat battery. This she did. She testified that this man was the applicant. She had identified the person shown in photograph No. 6, Ex. "D" some days after 15th June 1977 as the man whom she had driven from the school to Vermont South. She made this identification straight away from a folder of police photographs shown to her. (at p424)

9. Connell, a motor car dealer who in March 1977 was carrying on business at Elsternwick, gave evidence that on 29th April 1977 at about 2 p.m. he saw his partner with two men. The two men were inspecting a Nissan motor car which was on sale. It was not disputed at the trial that it was the Nissan found in Rickson Place. (at p424)

10. On 16th June 1977, Connell was interviewed by the police and was shown folders of photographs. He identified two photographs contained in two folders as being photographs of each of the two men he had seen inspecting the Nissan. He stated at the trial that he could not recall whose photographs he had identified and he was unable to identify any person in court as being either of the two men. (at p424)

11. For the Crown it was contended that the result of the evidence was that after Connell had refreshed his memory by looking at a police statement taken from him on 16th June he stated that the photograph which he identified was photograph No. 6 in Ex. "D". This was not common ground. Detective Sergeant Duggan swore that one of the photographs which Connell had identified was photograph No. 6 in Ex. "D" and that this was a photograph of the applicant. (at p424)

12. Before Beale was called to give evidence, objection was taken by the applicant's counsel that the folder of photographs should be excluded on the basis that the prejudicial effect of the photographs outweighed their probative value. It was submitted that the danger of the folder of photographs lay in the fact that the jury must infer from them that the police knew the applicant and, worse still, that they considered him a professional burglar. The trial judge ruled that the prosecution could tender the folder containing photograph No. 6 which was said to be a photograph of the applicant. The prosecution was allowed to elicit from Beale the fact that within a short time of seeing the Nissan he had been shown a folder of photographs (Ex. "D") and that he had identified the person shown in photograph No. 6 in that folder as one of the men running from the Nissan. Beale, at the trial, identified the applicant in the dock as having been the driver of the Nissan. (at p424)

13. After Beale had given evidence and Williams was called to give evidence, but before his evidence was taken, the applicant's counsel objected to Williams being shown Ex. "D". The objection was on the ground that to do so would create a situation in which the prejudicial effect flowing from the production of the photographs would outweigh their probative value. At the time when Williams was shown the photographs the applicant was already in custody. The Crown conceded that an identification parade could have been arranged, though none was held. The objection was overruled. Subsequently, in the course of his evidence, Williams was shown the photographs. He identified photograph No. 6 in Ex "D" as a photograph of the applicant, one of the men whom he had seen running from the Nissan. He also testified that some five or six days after the events in Rickson Place he had identified that photograph as being a photograph of the man seen running from the Nissan. He identified a black skivvy found in the applicant's flat as the skivvy worn by the man who had escaped. (at p425)

14. Another police officer, Dempster, who had chased the man who had made a getaway from the Nissan, identified the applicant at the trial as the man whom he had chased. He testified to having made a similar identification at the committal proceedings. There was no evidence that Dempster had identified Alexander from a collection of photographs. (at p425)

15. Four grounds are assigned in support of the application for special leave to appeal: (1) that the evidence of identification based on the examination of photographs given by Beale and Williams should not have been admitted because the applicant was under suspicion when Beale made his identification and in custody when Williams did so, there being no reason why an identification parade could not have been held at those times; (2) that the photographs from which the witnesses made their identification should not have been admitted; (3) that the trial judge erred in exercising his discretion to admit the identification evidence of all five witnesses in that he failed to take into consideration the principle applied in Bunning v. Cross [1978] HCA 22; (1978) 141 CLR 54 ; and (4) that the evidence of Connell and Duggan as to the "out of court" identification made by Connell should not have been admitted once Connell was unable to make an "in court" identification of the accused. (at p425)

16. The problems which afflict identification evidence have their origin in four principal sources: (a) the variable quality of the evidence much of which is inherently fragile; (b) the use by the police of methods of identification which, though well suited to the investigation and detection of crime, are not calculated to yield evidence of high probative value in a criminal trial; (c) the consequential need to balance the interests of the accused in securing a fair trial against the interests of the State in the efficient investigation and detection of crime by the police; and (d) the difficulty of accommodating the reception of certain types of identification testimony to accepted principles of the law of evidence. (at p426)

17. Identification is notoriously uncertain. It depends upon so many variables. They include the difficulty one has in recognizing on a subsequent occasion a person observed, perhaps fleetingly, on a former occasion; the extent of the opportunity for observation in a variety of circumstances; the vagaries of human perception and recollection; and the tendency of the mind to respond to suggestions, notably the tendency to substitute a photographic image once seen for a hazy recollection of the person initially observed. (at p426)

18. The use of photographs by police, especially photographs of known or suspected criminals, is an essential aid to the detection of persons who have committed crimes. Yet the use of such photographs before a jury may tend to suggest that the accused is known to the police as a criminal who has committed offences of the kind charged. And, as I have said, once a witness has seen a photograph which he links with the person seen, he tends to substitute the photographic image for his recollection. (at p426)

19. Recognizing these dangers the English courts have tended to draw a distinction between an identification made in the course of investigating a crime, when the police may request a potential witness to make an identification from photographs, and an identification made after the accused has been taken into custody, when the use of photographs is frowned upon and the arrangement of an identification parade is urged as the course to be preferred. In the second situation the purpose of the identification is purely evidentiary; it is designed to produce evidence for use at the trial. (at p426)

20. Before I examine the questions which relate to the evidence of Beale and Williams there is another basic point to which I should refer, obvious though it is. Traditionally it has been accepted that a witness identifies the accused at the trial as the person whom he observed at the scene of, or in connexion with, the crime. This "in court" identification, sometimes described as primary evidence, is of little probative value when made by a witness who has no prior knowledge of the accused, because at the trial circumstances conspire to compel the witness to identify the accused in the dock. It has been the practice to reinforce this "in court" identification by proving that the witness had earlier identified the accused out of court in a line-up or by selecting his photograph from a collection of photographs (R. v. Fannon and Walsh (1922) 22 SR (NSW) 427 ; Reg. v. Bouquet (1962) SR (NSW) 563 ; Reg. v. Doyle [1967] VicRp 82; (1967) VR 698 ; see also Reg. v. Goode (1970) SASR 69 ; Cross on Evidence, 2nd Aust. ed. (1978), p. 54) though the propriety of proving the photographs has been challenged by the applicant. (at p426)

21. The admission of evidence of this kind has been justified by reference to analogy with the doctrine of recent contrivance (Wigmore on Evidence, 3rd ed. (1940), vol. IV, s. 1130). In Di Carlo v. United States (1925) 6F (2d) 364, at p 366 , Judge Learned Hand held that, even if it be thought to be contrary to the hearsay rule, that rule must yield to common sense. For my part, I see no violation of the hearsay rule, nor do I think it necessary to resort to the doctrine of recent contrivance to sustain the admission of the evidence. In my opinion an identification made out of court by a person qualified to make it is admissible in evidence, subject to qualifications later to be mentioned. This is because an identification out of court, being earlier in time and made under circumstances which involve a selection in the absence of any compulsion, is more likely to be reliable than an identification made in court. I note that the Devlin Report on Evidence of Identification in Criminal Cases draws attention to the deficiencies of dock identification and recommends that restrictions be placed upon it (see pars. 4.89-4.109). (at p427)

22. The applicant relies heavily on R. v. Wainwright (1925) 19 Cr App R 52 to show that police photographs and identification of an accused by means of them are not receivable in evidence. There, Hewart L.C.J. said (1925) 19 Cr App R, at p 54 : "it is unheard of that police photographs, and the identification of a defendant by means of them, should be put forward by the prosecution as part of its evidence in chief. It is manifest that this conviction cannot be supported." It seems that the basis of the decision was that the effect of informing the jury that the accused was identified by the use of the police photographs was to tell them that the accused had a criminal record. Consequently the decision seems to proceed on the view that the prejudicial effect of the evidence is to outweigh its probative value. (at p427)

23. This view underlies a number of English decisions, particularly those which draw a distinction between the use of police photographs during the investigation of an offence and their use after the accused has been taken into custody when identification should be made by a witness from a parade. Thus convictions based on identification made from police photographs shown to a witness after the accused has been taken into custody have been quashed (R. v. Goss (1923) 17 Cr App R 196 ; R. v. Haslam (1925) 19 Cr App R 59 ). See R. v. Melany (1924) 18 Cr App R 2 ; R. v. Dwyer and Ferguson (1925) 2 KB 799, at pp 802-803 ; Reg. v. Seiga (1961) 45 Cr App R 220, at pp 224-255 . See also the dissenting opinion of Brennan J. in United States v. Ash [1973] USSC 163; (1973) 413 US 300 (37 Law Ed 2d 619) , a decision which turned on the Sixth Amendment. These cases favour the use of an identification parade as the appropriate mode of making an identification. They disapprove of photographic identification as a substitute for it, or as a preliminary to it after the accused has been taken into custody. They stress the prejudicial effect before a jury of identification from police photographs and impliedly, if not expressly, they suggest that the prejudicial effect outweighs the probative value of the evidence. (at p428)

24. The reasons why an identification parade is to be preferred to identification from police photographs are expressed compellingly in the Devlin Report (par. 5.21):

"The object of the parade is to surround the accused with a number of people bearing a sufficient resemblance to him. At the time when the photographs are shown there is no definite suspect or accused and so it is impossible to collect photographs resembling him; indeed the object at this stage is not to collect resemblances but to collect possible suspects so as to see if one can be picked out. Then the photographic album does not afford the full inspection that is given on parade. Whatever regulations were made for its conduct, there being no suspect, neither he nor his solicitor can be there to see that they are observed. Finally, the production of the album would be bound to arouse the suspicions of the jury as to how it came into existence and about the sort of men who found their way into it." (at p428)

25. Despite the shortcomings of identification from police photographs, the English decisions concede that evidence of such an identification is admissible and is properly received if it was made before the accused became a definite suspect or before he was taken into custody. The concession that the evidence is admissible and properly receivable in such circumstances acknowledges that it has probative value and that its probative value may outweigh its prejudicial effect. If identification evidence of this kind is acknowledged to have probative value which outweighs its prejudicial effect, when the identification is made before the accused is taken into custody or before he becomes a definite suspect, there is no strong reason for denying the same value to evidence of the same kind when the identification is made at a later stage. An element of unfairness in failing to arrange an identification parade may intrude, but the balance between probative value and prejudicial effect in the eye of the jury will remain unaltered. (at p428)

26. Wainwright (1925) 19 Cr App R 52 and the approach taken in the English cases is plainly at odds with the Australian decisions. In Fannon and Walsh (1922) 22 SR (NSW) 427 , Bouquet (1962) SR (NSW) 563 and Doyle [1967] VicRp 82; (1967) VR 698 , evidence of the kind proscribed by Wainwright was admitted without adverse comment and the convictions were sustained. In Doyle, where Fannon and Walsh was applied, the Court of Criminal Appeal rightly observed that the evidence of identification from police photographs was admissible and relevant, though the trial judge none the less had a discretion to exclude it on the ground that its prejudicial effect outweighed its probative value. The Court went on to observe that, though the evidence involved some possibility of prejudice to the accused, it might be considered by the jury to have had important probative value in showing that the witness was able to pick out one man from a large group of photographs as the man who committed the crime. The evidence was on this basis correctly admitted. In my opinion the Court of Criminal Appeal in Doyle was right in thinking that identification from police photographs may have an important probative value which will outweigh its prejudicial effect, more particularly if the jury are so instructed as to minimize the risk that they may take into account an adverse inference as to the accused's antecedents. In Doyle the photographic identification was followed by identification at a parade. But this did not detract from the probative value of the photographic identification though, as the Court remarked, the converse was not true. (at p429)

27. On the other hand, in Bouquet (1962) SR (NSW) 563 where the photographic identification was followed by identification in court, there being no parade, the majority held that identification by photograph only in the first instance was not a ground for excluding that evidence but at most only went to its weight. It was pointed out that the question of its reliability might call for a special direction when there was no other evidence of identification. (at p429)

28. In my opinion the Australian cases demonstrate that identification from police photographs is admissible evidence, that its probative value may be important, that the judge has a discretion to exclude it when he considers its prejudicial effect outweighs that value, and that directions may be given to ensure that unfair use is not made of the evidence. I follow them in preference to Wainwright (1925) 19 Cr App R 52 and the other English cases. The approach taken in the Australian cases to which I have referred accords with what this Court said in Davies and Cody v. The King (1937) 57 CLR, at pp 180-181 , when it pointed out that "in each case the question must be, not whether the identification has been conducted with propriety and fairness, but whether upon the whole evidence as it in fact existed when it came to be laid before the jury, and having full regard to the treatment of the matter at the trial, the actual verdict ought not to stand because a miscarriage of the kind described occurred". (at p430)

29. There is something to be said for the view that the admission and reception of identification evidence of the kind now in question should be governed by special rules. The evidence, whether it be based on photographs or on a parade, is artificial in the sense that it is brought into existence by the police for the purpose of providing evidence at a criminal trial, allowing for those cases in which the purpose is detection or investigation of crime. Because it is evidence which may be inherently fragile and it may be influenced by suggestion, considerations of fairness indicate that the police should, wherever possible, arrange for an identification parade, when the accused has greater opportunity of knowing the circumstances under which the identification comes to be made. Rules of admissibility based on fairness could readily be devised, as indeed they have been devised by the Devlin Committee. But in the end I have come to the conclusion that at this time we would be best advised to adhere to the approach which this Court adopted in Davies and Cody [1937] HCA 27; (1937) 57 CLR 170 by giving attention to the whole of the evidence and the treatment of the matter at the trial and deciding whether a miscarriage has occurred whereby the conviction cannot safely stand. In so doing we will best achieve a balance between the interests of the prosecution representing the State and the interests of the accused. (at p430)

30. In my opinion, therefore, identification based on police photographs, as well as the photographs, is admissible in evidence, despite the suggestions made to the contrary in the English cases and in Reg. v. Russell (1977) 2 NZLR 20, at p 28 . However, in some cases the circumstances may be such as to show that it would be unfair to receive them. (at p430)

31. It is inherent in what I have already said that I reject the rigid proposition sought to be gleaned from the English cases that identification made from police photographs when the accused is in custody cannot be proved in evidence or ground a conviction. As R. v. Bagley (1926) 3 DLR 717 shows, there are some cases in which the police have little alternative but to resort to photographic identification after the accused has been taken into custody. And I should have thought it essential to the efficient investigation and detection of crime that the police should continue to be at liberty to ask a potential witness to make an identification from photographs, even though they have a particular suspect in mind. In cases of serious crime it would unduly hamper police investigations if they were compelled to disclose to a suspect that he was under suspicion by requesting him to participate in an identification parade. (at p431)

32. I acknowledge that in Russell, Richmond P. said that "only in exceptional cases should photographs be used at a stage when some particular person is directly suspected by the police and they are able to arrange an identification parade or some other satisfactory alternative means whereby the witness can be asked directly to identify the suspected person". This observation, as it seems to me, was designed to ensure that the police act with propriety; it was not a statement of law. For my part, I have, for reasons already stated, difficulty in accepting a rigid rule which distinguishes between identification made before and after the accused was taken into custody. Certainly I would be unwilling to accept a rule which excluded the use of photographs in relation to a person who was under suspicion but had not been taken into custody. (at p431)

33. The trial judge in the present case recognized that he had a discretion to exclude the evidence of Beale, and later that of Williams. He decided that the prejudicial effect of the evidence did not outweigh its probative value. The applicant, in attacking this exercise of discretion, relies, not only on the English decisions already mentioned, but also on the principle applied in Bunning v. Cross [1978] HCA 22; (1978) 141 CLR 54 . Bunning v. Cross provides no support at all for the applicant. There, the evidence was illegally and irregularly obtained. None the less it was held that its cogency required that it should have been admitted. Reference to Bunning v. Cross does not reveal the existence of a relevant factor omitted by the trial judge in exercising his discretion to admit the evidence. (at p431)

34. It has been doubted that evidence of an earlier "out of court" identification can be given by a person other than the witness when the witness himself cannot or will not identify the accused in court or does not depose to the "out of court" identification. In R. v. Christie (1914) AC 545, at p 558 Lord Moulton said:

"Identification is an act of the mind, and the primary evidence of what was passing in the mind of a man is his own testimony, where it can be obtained. It would be very dangerous to allow evidence to be given of a man's words and actions, in order to shew by this extrinsic evidence that he identified the prisoner, if he was capable of being called as a witness and was not called to prove by direct evidence that he had thus identified him. Such a mode of proving identification would, in my opinion, be to use secondary evidence where primary evidence was obtainable, and this is contrary to the spirit of the English rules of evidence." (at p432)

35. However, Lord Atkinson (with whom Lord Parker of Waddington concurred) stated the contrary view. He said (1914) AC, at p 554 :

"The boy had in his evidence at the trial distinctly identified the accused. If on another occasion he had in the presence of others identified him, then the evidence of these eye witnesses is quite as truly primary evidence of what acts took place in their presence as would be the boy's evidence of what he did, and what expressions accompanied his act. It would, I think, have been more regular and proper to have examined the boy himself as to what he did on the first occasion, but the omission to do so, while the bystanders were examined on the point, does not, I think, violate the rule that the best evidence must be given. His evidence of what he did was no better in that sense than was their evidence as to what they saw him do." (at p432)

36. The view expressed by Lord Atkinson was accepted and applied by the Court of Appeal in Reg. v. Osbourne (1973) QB 678 . There, evidence was admitted from a police officer to establish that two witnesses had identified the accused at an identification parade when each witness at the trial was unable to recollect having made such an identification and one of the two witnesses was unable to identify one of the accused at the trial. The Court held that there was no reason in principle why the evidence should not have been admitted. The Court proceeded according to the view, which in my opinion is correct, that the reception of such evidence does not violate the hearsay rule or the best evidence rule. It is the act of identification that is relevantly in issue. An observer of the act may give evidence of it. Obviously the weight to be given to this evidence varies with the circumstances; its worth partly depends upon what is said by the witness who makes the identification. If he denies having made the "out of court" identification, and gives reasons for departing from that "out of court" identification, evidence from a third party that he did so seems to have little value and should be rejected, as it was in Reg. v. McGuire (1975) 4 WWR 124 , not perhaps on the ground that it is not evidence at all, but on the ground that its probative value is so slight as to make it valueless. If, however, as here, the recollection of the identifying witness is hazy through lapse of time, the evidence of the third party may have value. (at p432)

37. Here, a period of two years ensued between Connell's identification of the accused from police photographs and the giving of his testimony at the trial. A reading of that testimony makes it plain that he was acknowledging that he had identified a photograph of the accused; he was asserting that his recollection was hazy, that he could not be certain that he had identified a particular photograph of the accused and that for the same reason he could not say that the accused was one of the two men he had seen inspecting the Nissan. Consequently the evidence of Connell and Duggan was correctly received. (at p433)

38. The conclusion which I reach, then, is that the identification evidence was properly received and that the challenge to the trial judge's exercise of discretion which resulted in the admission of the evidence fails. (at p433)

39. There was no challenge to the judge's directions to the jury on the subject of identification. He warned the jury that "there are dangers in acting on identification based on personal impression", instructed them to exercise considerable care and to subject the evidence to close scrutiny. He elaborated on the dangers and carefully itemized the criticisms made of the Crown evidence. There is, accordingly, no ground for thinking that there has been a miscarriage of justice or that the conviction cannot safely stand. (at p433)

40. However, the conduct of the police in failing to arrange an identification parade when it could have done so deserves strong censure. (at p433)

41. In the result I would grant special leave to appeal because the questions raised by the applicant are important, and I would dismiss the appeal. (at p433)

MURPHY J. The applicant contends that the trial judge erred in the exercise of his discretion, in allowing evidence to be given by Crown witnesses of prior identifications of the applicant made by them from photographs in the possession of the police, and in allowing the file of photographs used by the police for this purpose to be admitted into evidence. The applicant also contends that the trial judge erred in law in allowing evidence to be given by Sergeant Duggan of a prior out-of-court identification of the applicant by a Mr. Connell, on the grounds that it was hearsay and inadmissible or infringed the rule against prior consistent statements. The applicant claims that these errors brought about a miscarriage of justice. (at p434)

2. All the challenged evidence is admissible. Evidence of previous identification or recognition of the accused is admissible as evidence of facts tending to prove what had become an issue, that is the reliability of the in-court identifications of the accused, or else as tending to complete what is an incomplete in-court identification (which refers to and depends for its completion on evidence of the previous identification). (at p434)

3. A witness may give evidence that although he cannot now independently identify another whom he saw at the event in question, he has on some other occasion identified that other, by some representation (such as a photograph) or in person (at a line-up, or elsewhere). The proof of the earlier act of identification is then admissible to complete the in-court identification. The proof may be from his own recorded recollection (for example, from his note on the photograph - "This is the man I saw etcetera") of it may be by some other person, as in this case by Sergeant Duggan. Such evidence given by another to complete the earlier identification is not hearsay. Taken together, the evidence amounts to an in-court referential identification of the person at the event in question as the person in the noted photograph. If the witness does not give evidence that on the earlier occasion he identified (i.e. correctly identified) the person in the photograph as the person at the event in question (even though he cannot remember the details) then the evidence from others of his previous act of identification is hearsay and inadmissible. (at p434)

4. Mr. Connell's in-court identification was indirect and incomplete. In substance he stated that some days after the crime he had identified to the police a person whom he had observed some months earlier concerned in the acquisition of the car used in the crime, that he had done so by identifying the person as the person appearing in a photograph contained in a file of photographs. He could not in court independently recall which photograph he had identified, however he had at the time recorded his identification in a signed statement to the police referring to particular photographs. His in-court adoption or verification of that recorded identification was admissible as evidence of "past recollection recorded" (Wigmore on Evidence, 3rd ed. (1940), vol. III, s. 734). Also the evidence of Sergeant Duggan that the photograph identified by Connell was that of the accused Alexander was admissible to complete Connell's in-court identification. (at p434)

5. The other evidence of identification of the applicant by Crown witnesses, including the file of photographs, was also all admissible to support those witnesses in Court identification of the accused. (at p435)

6. However, the trial judge has a general discretion to exclude evidence which is legally admissible (Reg. v. Ireland [1970] HCA 21; (1970) 126 CLR 321 ; Driscoll v. The Queen [1977] HCA 43; (1977) 137 CLR 517 . This discretion may be exercised to exclude evidence which has been obtained unfairly or improperly or is the kind of material which experience shows has an adverse effect on the accused likely to exceed its probative value. Common experience demonstrates the risk of unreliability of identification based on a sighting of a person whom the witness did not know previously. Miscarriages of justice arising from mistaken identifications (even by numerous witnesses) are part of our folklore. The Report of the Committee of Inquiry Into The Case of Mr. Adolf Beck concluded that "evidence as to the identity based on personal impressions, however bona fide, is perhaps of all classes of evidence the least to be relied upon, and therefore, unless supported by other facts, an unsafe insufficient basis for the verdict of a jury" (1905, Cd. 2315, vii). (at p435)

7. In Reg. v. Sutton (1970) 2 OR 358, at p 368 the Ontario Court of Appeal referred to "the inherent frailties of identification evidence arising from the psychological fact of the unreliability of human observation and recollection". In The People v. Casey (No. 2) (1963) IR 33, at p 39 the Supreme Court of Ireland stated:

"We are of opinion that juries as a whole may not be fully aware of the dangers involved in visual identification nor of the considerable number of cases in which such identification has been proved to be erroneous; and also that they may be inclined to attribute too much probative effect to the test of an identification parade. In our opinion it is desirable that in all cases, where the verdict depends substantially on the correctness of an identification, their attention should be called in general terms to the fact that in a number of instances such identification has proved erroneous, to the possibilities of mistake in the case before them and to the necessity of caution. Nor do we think that such warning should be confined to cases where the identification is that of only one witness. Experience has shown that mistakes can occur where two or more witnesses have made positive identifications."

(See also Reg. v. Turnbull (1977) QB 224 ; and Mewett, "Notes and Comments: Identification", Criminal Law Quarterly, vol. 14 (1971), p. 20.) (at p435)

8. The general dangers in identification of a stranger are compounded when the first identification after the crime is from a photograph. The well-known "displacement" effect tends to reduce the reliability of later identification (see Australian Law Reform Commission Report No. 2 "Criminal Investigation", (1975), p. 56). In addition the "rogues gallery" effect can be highly prejudicial. Also photo-identification invariably occurs without the accused or his counsel being present to see whether the procedures used are fair. (at p436)

9. The practical necessity for use of photographs in the detection process therefore often produces dangers of unreliable evidence and prejudice in the trial process. The reconciliation of the societal interest in effective investigation with the societal and defendants' interests in fair trials may be generally achieved by the admission of evidence of photo-identification occurring during the detection process, and the exclusion of such evidence where the identification occurs after the detection process has been completed. (at p436)

10. The Australian Law Reform Commission reported their firm view "that the showing of photographs of a suspect to a witness after that suspect has been apprehended should be prohibited. Where a suspect is still at large and unidentified the showing of photographs is unavoidable. But this rationale ceases as soon as he is in custody: identification then, if necessary, should be by parade. The Commission recognises the necessity for an exception to this rule when the person in custody refuses to take part in, or consistently aborts, a parade, or where for some reason - for example a disfiguring injury - he is unable meaningfully and fairly to participate." (Report No. 2 "Criminal Investigation" (1975), p. 56; see also Report to the Secretary of State for the Home Department of the Departmental Committee on Evidence of Identification in Criminal Cases (The Devlin Report), 1976.) In general judicial discretion should be used to exclude photo-identification evidence obtained after the detection process. There may be exceptional cases, such as the absence of opportunity to use fairer methods, and those where, in the light of other evidence, for example, a demonstrably false alibi, the possibility of misidentification is slight. The general exclusion of such evidence removes the incentives on the part of police officers to use unfair methods when the opportunity exists to use methods which are fair, or at least fairer. (at p436)

11. In this case some photo-identification occurred during the detection process; evidence of this was properly admitted. Photo-identification was also used after the detection process was completed when apparently there was opportunity for use of fairer methods and no explanation was offered which would justify the procedures. (at p437)

12. The evidence of photo-identification which occurred after the accused was in custody, although admissible, should have been excluded, and failure to exclude it gave rise to a miscarriage of justice. The Crown has not established that no substantial miscarriage of justice has actually occurred (see s. 568 Crimes Act 1958 (Vict.)). Special leave should be granted. The appeal should be allowed. The conviction and sentence should be quashed and a new trial ordered. (at p437)

AICKIN J. I have had the advantage of reading the reasons for judgment prepared by my brother Mason. Those reasons demonstrate that, although the grant of special leave is warranted by the circumstances, the appeal should be dismissed. I agree with the order which he proposes. (at p437)

Application for special leave to appeal granted.

Appeal dismissed.