RESULT: Appeal allowed, judgment of acquittal entered.

The appellant was convicted of robbery. He had been identified by the victim from photographs. No identification parade was held. It was submitted that the appropriate test as to when evidence of such identification should be excluded in the exercise of the trial judge's discretion if at the time when they were shown to the witness it was reasonable and practicable, or reasonably open to the police, to request the suspected person to participate in such an identification parade.

(1) There is no rigid rule excluding evidence as to the use of photographs in relation to a person who is under suspicion but who has not been taken into custody. Nevertheless, a primary issue in most cases will be whether it was necessary at that stage in that particular case to show photographs to that particular witness in order to know who to arrest and charge. Until the police have such knowledge, the usual "Christie" discretion will need to be exercised in relation to the evidence of that witness. It is the necessity to use the photographs in such circumstances which plays an important part in the balancing exercise involved in the exercise of that discretion. The accused will have a harder task in persuading the exercise of the judge's discretion to exclude such evidence if the police did not then know that he was the person to arrest and charge, notwithstanding that it may at that time have been reasonable and practicable to have requested him to participate in such a parade. On the other hand, if it was reasonable and practicable for the police to hold an identification parade, and if the police did know at that stage that the accused was the person to arrest and charge, the Crown will have a harder task in resisting the exercise of the trial judge's discretion to exclude such evidence, but it remains a discretionary matter nevertheless and the appellant's more rigid proposition cannot stand.

Alexander v The Queen [1981] HCA 17; (1981) 145 CLR 395 discussed and followed

Regina v Russell [1977] 2 NZLR 20; Regina v Harrison (1987) 29 A Crim R 213; Regina v Browne & Ors (1987) 30 A Crim R 278; Regina v Dawson (1990) 47 A Crim R 458 discussed

Davies & Cody v The King [1937] HCA 27; (1937) 57 CLR 170; Rex v Fannon & Walsh (1922) 22 SR 427 referred to

Considerations which would incline a trial judge to permit evidence to be given of identification from photographs notwithstanding that the police already knew at that time that he was the person to arrest and charge discussed.

(2) The power of the trial judge to exclude evidence in accordance with the "Christie" discretion does not permit the judge, in assessing what its probative value is, to determine whether the jury should or should not accept the evidence of the witness upon which the Crown case depends. There is a distinction between that task of the trial judge and the task of the Court of Criminal Appeal in relation to a ground of appeal that the verdict was unsafe and unsatisfactory.

Doney v The Queen [1990] HCA 51; (1990) 171 CLR 207 followed

Domican v The Queen [1992] HCA 13; (1992) 173 CLR 555; M v The Queen (1994) 181 CLR 487; Regina v McKnoulty (1995) 77 A Crim R 333 referred to

HUNT CJ at CL: The appellant (Constantine Carusi) was charged with armed robbery, and alternatively with robbery. A District Court jury found him not guilty of the main charge but guilty of the alternative charge. Judge Gibson QC imposed a sentence of penal servitude for six years, consisting of a minimum term of three years and an additional term of three years.

The Crown case was that at about 1.50 pm on 20 May 1993, the victim (Mrs Patricia McHugh) left the ANZ Bank at Auburn with her employer's payroll of $12,723 in cash in her handbag. As she approached her car parked behind the bank, a man grabbed her bag and it appeared to her that he was holding a small black gun at the time. The incident was observed by two men, one of whom (the bank manager) then saw the offender driving a car from a nearby carpark. The car was registered in the name of the appellant's father. Mrs McHugh gave a description of the offender which fitted the appellant in rather general respects. She later identified the appellant as the man who had robbed her from one of two groups of photographs. When subsequently arrested, the appellant said that he had been at the Drummoyne RSL Club by himself from about 11.00 am to about 4.00 pm on that day. The records of that club and the evidence of another member established that a person by the name of "C. Carusi" had signed in at some time between 10.00 am and before midday on that day.

The appellant gave evidence that he had been at the RSL Club until 1.00 pm, when he had gone home and had stayed there all afternoon. He said that it had been an honest mistake on his part when he told the police that he had been at the club until about 4.00 pm. He had received telephone calls from three people during that afternoon. Two of those callers gave evidence on his behalf that they had made such calls, at about 1.00 pm and 2.30 pm respectively.

The principal ground of appeal was that the verdict was unsafe and unsatisfactory but, in order to consider that ground, it is necessary first to consider another ground of appeal, that the evidence by Mrs McHugh identifying the appellant from photographs should have been excluded. If that evidence were to be excluded the appellant would more easily succeed in obtaining a judgment of acquittal on the appeal, on the ground that any verdict of guilty without that evidence would necessarily be unsafe and unsatisfactory. The issue raised by this ground of appeal is one of some importance in the law of evidence as it stood prior to the commencement of the Evidence Act 1995.[1]

It is now beyond question that evidence of identification from photographs is relevant and therefore prima facie admissible.[2] Evidence of their use does, however, present very real dangers at the trial.[3] The first danger is that unfairness to the accused may result. As he is perforce not present when such an identification is made (as he must be when an identification parade is held), he is unable as effectively to litigate any weaknesses in the way in which that identification is made. Other dangers relate to the defects in the use of the photographs themselves.

The "rogues gallery" effect is well documented, in that the jury may conclude (rightly or wrongly) that the existence of the accused's photograph in police files means that he is unfavourably known to the police. There is also the "displacement" effect, by which the witness may subconsciously permit the photograph to displace his or her memory of the offender and thus lead that witness at any subsequent identification parade to match the accused person in the parade with a memory of the photograph rather than of the offender, whether or not it is the same person. Moreover, a photograph is only two dimensional, and it records what a person looks like in the one split second when that person may have been moving his or her features, and which may not therefore always provide a safe impression of what that person really looks like when seen in the flesh.

The result is that there is always the danger that the jury will give to evidence of an identification from photographs a greater weight than it deserves. When that is added to what has been described as the seductive effect of identification evidence generally,[4] the danger of the prejudice to the accused by its admission (for the reasons already outlined) is considerably heightened. Such prejudice can sometimes be lessened (but rarely removed) by appropriate directions to the jury. However, identification evidence (as with any other evidence) is subject to the common law "Christie" discretion, whereby the trial judge may exclude any evidence where its prejudice to the accused outweighs its probative value.[5] That discretion plays a large part in all identification cases,[6] and one test as to whether evidence of identification generally should be excluded is whether the quality of that evidence falls short of the point where its frailty or frailties cannot be cured by an appropriate direction or caution to the jury.[7] Such evidence is also subject to another common law discretion whereby evidence may be excluded where the use of the evidence would result in an unfair trial for the accused.[8]

The objection taken at the trial in the present case was that the evidence of the identification by Mrs McHugh from photographs should be excluded in the exercise of one or other or both of those common law discretions. On appeal, it was faintly submitted that the judge should have also rejected the evidence as having been improperly obtained contrary to the Commissioner's Instructions,[9] but this submission depended upon an interpretation of those Instructions which raises the same issue as must be resolved in any event in relation to the way in which the "Christie" discretion should be applied. It need not be pursued further here.

The argument on appeal in relation to the application of the "Christie" discretion to evidence of identification from photographs was more refined than that put at the trial. It emphasized the tension between permitting the police to use photographs for identification purposes during what has been called the detection process and the prejudice to the accused if evidence is given of such an identification because of the dangers which such evidence presented. It is where the line should be drawn which is the issue in the present appeal and which is of some importance. The appellant's submission, put generally, is that, as the appellant was already a suspect at the time when Mrs McHugh was shown the photographs, the evidence of her identification from those photographs should for that reason alone have been excluded. The evidence as to what was before the police may be stated very shortly.

Mrs McHugh was not shown the photographs from which she identified the appellant as the man who robbed her until 1 September, almost three and a half months after the robbery had taken place. The identification of the car used by the offender as being registered in the name of the appellant's father was, however, known to the investigating police the same day as the robbery. On the same day, the police also had the description of the offender given by Mrs McHugh which, as I have said, did fit the appellant is some rather general respects. So far as it is relevant for present purposes, she described the offender as being in about his early twenties, about five feet eight or nine inches tall, and clean shaven.

As a result of an examination of the car the following day, the fingerprints of one Tom Tamani were found on the outside of the driver's door window near the rear edge. We were informed that there were no signs of forced entry found in the car. According to the evidence given by the appellant at the trial, he was at the time of the robbery living with his father, although I note that his father (who did not give evidence) nominated a different address in his report made late on the day of the robbery that the car had been stolen. The father appeared to be living at the same address as the appellant when the appellant was arrested almost four months later.

In any event, evidence was given on the voir dire examination into the admissibility of the identification evidence that, by the time when the photographs were shown to Mrs McHugh, the police had some suspicion concerning the appellant because of what was described as his involvement with the car. The description of the offender given by Mrs McHugh necessarily excluded Tamani, who is a Fijian National with a very dark skin, he is six feet four to five inches tall and at the time he had a large dark moustache.[10] Her description was also inconsistent with the age of the appellant's father, and she did not identify the appellant's father when shown one of the two groups of photographs which included a photograph of him

Whether that evidence should be interpreted as demonstrating that the appellant was already a suspect at the time when Mrs McHugh was shown the photographs depends upon a consideration of five cases which are the only ones discovered by the careful researches of counsel where this issue has been discussed or at least referred to.

In Regina v Russell,11 the New Zealand Court of Appeal was concerned with an identification from a number of photographs (full face and side face) which had obviously been taken from police records. At the time when the photographs were shown to the only witness who saw the offender, the appellant was "an obvious suspect" (as it was put), and the Court said that he should at that stage have been asked whether he would take part in an identification parade. Two months later, the witness picked out the appellant in an identification parade.[12] The principle was stated thus:[13]

" ... 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."

The Court nevertheless declined, in "the particular circumstances of this case", to interfere with the jury's verdict or to say that the trial judge should have excluded the evidence in relation to the use of photographs either on the ground of unfairness by the police or on the ground that its prejudicial effect substantially outweighed its probative value.[14]

In Alexander v The Queen,15 the appellant was suspected by the police of complicity in the crime but not yet arrested when the first witness made an identification from photographs.[16] The appellant submitted that evidence of identification from a number of photographs should be excluded "when the person to be identified has been arrested or is in police custody or is suspected of having committed the crime".[17] In that case, there was no identification parade held. So far as the issue in the present case is concerned, Gibbs CJ said that -

" ... 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 ..."[18]

He said that an identification parade should, wherever possible, be held once the accused has been arrested or where he "is firmly suspected to be the offender".[19] He cited with approval the passage from Regina v Russell which I have already quoted, and he said that it stated the principle to be applied when exercising the discretion to exclude the evidence.[20] He added - following Davies & Cody v The King[21] - that the true question for the Court of Criminal Appeal on appeal from a conviction based upon evidence of identification from photographs when an identification parade could have been conducted 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.[22] The Chief Justice also said that, in considering the matter, the Court of Criminal Appeal will also keep in mind the importance of ensuring that the most reliable evidence of identification is obtained in every case.[23] He held that the trial judge had been entitled to admit the evidence.[24]

Stephen J accepted that the use of photographs by the police was proper in the detection process to enable them "to know whom they should arrest and charge",[25] but he pointed out that, where the witness to whom they were shown then gives evidence in the trial, his or her evidence will be detrimentally affected in the ways already discussed by having been shown the photographs.[26] Recognizing the tension between the need to use the photographs in the detection process and the unacceptable price to the public interest if their use were to require the disqualification of that witness's evidence at the trial, Stephen J said that the solution of the courts was to discourage the use of photograph identification 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: [the courts] 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."[27]

He accepted that such a solution had not, so far as he was aware, been anywhere explicitly stated, but he surveyed a large number of cases to demonstrate why he thought that that was the pattern which had evolved. Stephen J quoted with approval the distinction between the use of photographs in the detection process and their use in the evidentiary process as stated in a dissenting judgment in a Canadian case, which defined the former as:

" ... before the arrest, fairly conducted for the purpose of ascertaining who should be arrested ...".[28]

He also referred with approval to a statement in a South Australian case as to the value, sometimes indispensable, of the use of photographs "where a victim cannot name the criminal".[29] He too cited with approval the passage from Regina v Russell which I have already quoted.[30]

Stephen J went on to say that his survey of the case disclosed that, when a witness has been involved in photo-identification "after the detection process is over", there is a strong ground for excluding the evidence of that witness, as the police will in such a case already "know ... the identity of the `wanted man'".[31] His views as to the dividing line between the detection process and the evidentiary process should be quoted in full:

"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."[32]

When holding that the evidence of one of the witnesses should have been excluded because he had been shown the photographs at a time when the detection process was already at an end, Stephen J explained that this was so because "the police had decided who was their `wanted man' and that all that then remained to be done was to catch and charge him", as they had searched his home, and they were looking for the man thought to be the appellant.[33] He then said that he shared much of the substance of the judgment of the Supreme Court of Victoria in the case under appeal, and said this:[34]

"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'."

Mason J interpreted the English cases as permitting evidence of identification from photographs only where it was made "before the accused became a definite suspect or before he was taken into custody",[35] and he observed that such an approach was plainly at odds with the Australian cases, which permitted the evidence to be given subject to the discretion to exclude it where its prejudicial effect outweighed its probative value and where directions were given to ensure that unfair use is not made of the evidence. He followed the Australian cases in preference to the English cases.[36] Although he saw that there was something to be said for special rules to be devised, he (like Gibbs CJ) said that the approach to Davies & Cody v The King was to be followed 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.[37] He rejected the English view that evidence of an identification made from photographs must be rejected when the accused is already in custody or when he is already a "particular suspect" in mind.[38] He interpreted the statement in Regina v Russell which I have quoted as other than one of law, and he said that he 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,[39] but he concluded that the conduct of the police in that case in failing to arrange an identification parade, when it could have been, as deserving strong censure.[40]

Murphy J said that, in general, evidence of identification from photographs after the detection process has been completed should be excluded.[41] He did not expressly consider just where that line should be drawn, although he did quote from a report of the Australian Law Reform Commission "Criminal Investigation", which appears to draw the line between the time when the offender is still at large and unidentified and the time when he is in custody[42] - one which does not assist as to what is meant by the word "unidentified".

Aitkin J agreed with Mason J. I will return to Alexander's Case after referring to the remaining cases debated by counsel.

In Regina v Harrison,43 a decision of the West Australian Court of Criminal Appeal, the victim of a sexual assault made her first complaint to the police four days after the event with a description of the offender. She was shown photographs five days later and she identified the appellant. He was interviewed eleven days later, and he denied being present at the time but was detained by the police in order to permit the victim to see him (with many other people) leaving police headquarters. She identified him in the flesh. Burt CJ held, following the judgment of Stephen J in Alexander's Case, that the photographs had been shown to the victim during the detection process and not after it had ended.[44] The other members of the Court did not refer to the issue.

In Regina v Browne & Ors,45 a decision of this Court, the three appellants had been charged with various offences of violence against a guard in an armoured van in the course of an armed robbery. After one of them had been arrested, an off duty police officer, who had seen a man running through his back yard (which was in the vicinity of where the stolen getaway car had been abandoned by the offenders), made an identification from photographs of another of the appellants as that man. That appellant was arrested shortly afterwards. He submitted that, in the circumstances, an identification parade should have been held. This Court held that the photograph identification had taken place at a time when the police were still making their inquiries.[46]

In Regina v Dawson,47 also a decision of the WA Court of Criminal Appeal, the appellant is described as having already been a suspect[48] and as obviously a possible suspect in the mind of the police[49] at the time when the identification was made from photographs by the first witness and which lead to his arrest. Further witnesses then made an identification from photographs, and no valid reason was given for the failure to hold an identification parade.[50] Quoting from the judgment of Gibbs CJ in Alexander's Case, it was held by Malcolm CJ that the identification by the first witness had been made during the investigation stage.[51] Brinsden J, also quoting from the judgment of Gibbs CJ, said that the use of photographs was "clearly of an investigatory nature".[52] Wallwork J agreed with both judgments.[53]

The appellant in the present case has submitted that the appropriate test as to when the investigative stage has been completed, and thus the point after which evidence of identification from photographs should be excluded, is:

" ... when it is reasonable and practicable to request the suspected person to participate in an identification parade."

The test, it was said, is "whether it is reasonably open" to the police to conduct such a parade. If it is, then evidence of photograph identification made at that time should always be excluded.

Such a test is said to have been strongly suggested in Alexander v The Queen, following upon statements in that case that the investigative stage has been completed "when the police investigation has focussed on a person as being a definite suspect" and upon statements in Regina v Russell that it is "when there is a direct suspect". The adjectives "definite" and "direct" appear in those cases, but they are not themselves defined as giving any particular quality to the suspicion held. They seem to me to be of little assistance in determining in any particular case whether the investigative process has been completed. The appellant submitted that the "wanted man" referred to by Stephen J in the third of the passages from his judgment which I have quoted[54] - in the sentence:

"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." -

"is not the person whom the police intend to charge, but the person whom the police suspect, perhaps subject to the conduct of an identification procedure of some kind, as being the person responsible". Counsel continued: "The wanted man concept referred to by Stephen J is really another way of saying a person who is a suspect, and the suspect does not develop into the person who is going to be charged until the identification process is conducted".

To interpret the judgment of Stephen J in the way for which the appellant contends is to ignore the words by which he introduced it:[55]

"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."

I have added the emphasis to the word "know". Stephen J also quoted with approval the dissenting Canadian judgment which states that the purpose of their use was "of ascertaining who should be arrested".[56] This point is repeated by Stephen J in the third of the passages from his judgment which I have quoted, when he said that the point at which the distinction is to be drawn between the detection and the evidentiary processes is the existence of "the need of the police to know who is the wanted man", or "[o]nce the police know who they are seeking", and when he referred to "police knowledge of the identity of the wanted man".[57]

Those statements suggest very strongly to me that a suspicion of the nature suggested by the appellant - however that suspicion may be qualified by adjectives such as "definite" and "direct" - is not enough to bring about the completion of the detection process. The fact that the appellant concedes that such a suspicion may still be subject to some kind of identification procedure being conducted necessarily demonstrates, to my mind, that the detection process is not yet complete. Stephen J uses the word "ended" for a purpose, and it should be interpreted in its usual sense of finished or completed.[58]

It must, of course, be kept in mind that a person may be arrested without warrant wherever the arresting officer, with reasonable cause, suspects him of having committed an offence.[59] Stephen J is not saying that the detection process remains incomplete until the police know that the accused committed the crime. The issue which he has raised is as to whether the police know that the accused is the person whom they suspect with reasonable cause to have committed the offence. Where they so suspect that the offence was committed by a person whose description they have but whose identity they do not know, the detection process is completed only when they know that the accused is that person who is suspected upon that basis. I do not, therefore, interpret his judgment as supporting the appellant's submissions in this case.

To interpret Alexander's Case in accordance with the appellant's submission would also be to ignore the clear statement by Mason J (with whose judgment Aitkin J agreed) rejecting any rule which excluded the use of photographs in relation to a person who was under suspicion but who had not been taken into custody, and to ignore also the decision of Gibbs CJ that the trial judge in that case had been entitled to admit the evidence notwithstanding that the photographs had been shown to the witness at a time when the appellant was suspected of complicity in the crime. It would in addition be to ignore the recognition by this Court in Rex v Fannon & Walsh[60] that the use of photographs is still appropriate where somebody is under suspicion, when it described such use as -

" ... one of the most effective instruments for the speedy discovery of the guilty person and for the exoneration of others who might innocently be the object of suspicion."[61]

In Alexander's Case, Gibbs CJ thought that it was unclear from the report of Rex v Fannon & Walsh at what stage the photographs had been shown,[62] but Stephen J interpreted the approach of the judgment to the question of the admissibility of the evidence as demonstrating that they had been used in the detection process.[63]

The appellant drew attention to what had been said by various reports and in rules laid down elsewhere as to the limitations to be placed on the use of photographs in identification. In the UK Report "Evidence of Identification in Criminal Cases",64 known as the Devlin Report, it was said that photographs should not be shown to prospective witnesses unless it is strictly necessary for the purposes of the search; once one satisfactory identification has been obtained, the police "have a suspect", and the usual and proper way of confirming or eliminating the suspicion is by means of a parade.[65] The Code of Practice issued pursuant to the (UK) Police and Criminal Evidence Act 1984 relating to identification, Code D, provides that a witness must not be shown photographs or photofit, identikit or similar pictures if the identity of the suspect is known to the police and he is available to stand on an identification parade.[66]

The Canadian Law Reform Commission's Report, "Police Guidelines: Pre-Trial Eyewitness Identification Procedures",67 says that an identification parade is required if the police have "a particular suspect", even though they might not have sufficient evidence to justify an arrest, and that when the police suspect that "a particular person" may be responsible for a crime they should not show a witness photographs unless the suspect refuses to participate in an identification parade.[68]

It was submitted that these statements in the United Kingdom and in Canada are consistent with the decision of the High Court in Alexander v The Queen. I do not agree. To some extent, they could be said to be consistent with the judgment of Stephen J if that judgment is to be interpreted in the way for which the appellant contends, but I have already rejected that interpretation. The statements are inconsistent with the judgments of Gibbs CJ, Mason J and Aitkin J, which - subject to the discretionary considerations referred to earlier - permit evidence to be given of an identification from photographs notwithstanding that the accused person was then under suspicion.

Accordingly, I reject the appellant's submission that evidence of identification from photographs should always be excluded if at the time when they were shown to the witness it was reasonable and practicable, or reasonably open to the police, to request the suspected person to participate in an identification parade. The distinction drawn in the English cases between an identification made before the accused is arrested and one made after that event has already been decisively rejected in Alexander's Case.

In my judgment, there is no rigid rule excluding the use of photographs in relation to a person who is under suspicion but who has not been taken into custody. Nevertheless, a primary issue in most cases will be whether it was necessary at that stage in that particular case to show photographs to that particular witness in order to know who to arrest and charge. Until the police have such knowledge, the detection process (to adopt the language of Stephen J) is not complete, and the usual "Christie" discretion will need to be exercised in relation to the evidence of that witness. It is the necessity to use the photographs in such circumstances which plays an important part in the balancing exercise involved in the exercise of that discretion. The accused will have a harder task in persuading the exercise of the judge's discretion to exclude such evidence if the police did not then know that he was the person to arrest and charge, notwithstanding that it may at that time have been reasonable and practicable to have requested him to participate in such a parade. On the other hand, if it was reasonable and practicable for the police to hold an identification parade, and if the police knew at that stage that the accused was the person to arrest and charge, the Crown will have a harder task in resisting the exercise of the trial judge's discretion to exclude such evidence, but it remains a discretionary matter nevertheless and the appellant's more rigid proposition cannot stand.[69]

There are well recognized considerations which would incline a trial judge to permit evidence to be given of identification from photographs notwithstanding that the police already knew at that time that he was the person to arrest and charge - such as where it would be inappropriate in relation to the police investigation to reveal to the accused that he is under suspicion; where the accused has refused to participate in an identification parade or has failed to co-operate when the parade is conducted; where, by reason of the unusual physical appearance of the accused it is not reasonably practicable in the circumstances to produce a sufficient number of other persons of a roughly similar physical appearance to stand with the accused in such a parade; where the physical appearance of the accused is materially different to what it was at the time when he is alleged to have committed the offence; where the accused cannot within a reasonable time be located; where there are other exigencies of time which would make it impracticable to conduct an identification parade; where the witness is unwilling to participate in viewing the parade; and where the witness is unable reasonably so to participate. That list is not intended to be exhaustive.[70]

In the present case, in my opinion, the police did not know that the appellant was the person whom they should arrest and charge when the photographs were shown to Mrs McHugh. His only direct association with the robbery was that, because he lived with his father, he could possibly have had access to his car. Against that there was the report that the car had been stolen prior to the robbery - certainly, as it turned out at the trial, a highly suspicious report but the police were not to know that at the time. The description of the offender which Mrs McHugh had given did exclude the accused's father and Tamani, who were the only other persons suspected by the police (in the loosest sense of the word), but it did not very convincingly identify the appellant either. Those three were not the only persons who could have committed the robbery, and the police were by no means in a position to know that the appellant was the person to arrest and charge. It was, in my opinion, necessary at that stage to show the photographs to Mrs McHugh so that she could, if appropriate, point the finger at the accused.

Did the prejudicial effect of that evidence of identification nevertheless outweigh its probative value? It was made clear that the group of photographs from which Mrs McHugh identified the appellant as the man who robbed her were surveillance photographs taken only recently, so that there could have been no "rogues' gallery" effect upon the jury. There were some problems with the credit of Mrs McHugh, in the sense that the reliability of her evidence as to the description of the offender was attacked because of the manner in which she gave her evidence. I will have to consider those problems in connection with the principal ground of appeal, that the verdict was unsafe and unsatisfactory. The clear distinction between the task of this Court in relation to such a ground of appeal and that of the trial judge when exercising the "Christie" discretion must not be overlooked.

It was for the jury, and not for the trial judge, to determine the factual issues at the trial. Applying the ordinary tests, there was clearly a case to go to the jury based upon the evidence in chief of Mrs McHugh which (if that evidence were accepted) was a strong one, and it was for the jury to determine which parts of her evidence they accepted and which parts they rejected.[71] The power of the trial judge to exclude evidence in accordance with the "Christie" discretion does not permit the judge, in assessing what its probative value is, to determine whether the jury should or should not accept the evidence of the witness upon which the Crown case depends. The trial judge can only exclude the evidence of such a witness where, taken at its highest, its probative value is outweighed by its prejudicial effect; whereas this Court may use its supervisory powers to set aside a verdict where, the issue having been left to the jury, this Court is satisfied - on the whole of the evidence - that the jury ought nevertheless have had a reasonable doubt.[72]

Such a distinction is clearly drawn in the decision of the High Court in Doney v The Queen.73 It is also basic to that Court's decision in Domican v The Queen,74 by its insistence that the trial judge isolate and identify for the benefit of the jury any matters of significance which may reasonably be regarded as undermining the reliability of the identification evidence, and that directions be given to the jury that they are bound to take those matters into account in determining whether they will rely on that evidence.[75] That is why I leave to later the attack upon the reliability of Mrs McHugh's evidence generally so far as it depends upon the jury's acceptance of it. In accordance with the test earlier stated,[76] I do not consider that the frailties in the evidence of her identification of the appellant from photographs were such that they could not have been cured by an appropriate direction or caution to the jury. They did not require the evidence of the identification itself to be excluded from the jury.

Various submissions were also made that the judge misdirected himself when declining to exercise his discretion to exclude that evidence. I do not find it necessary to consider those submissions because, even if I were to conclude that any such misdirection had occurred, so that this Court would have to exercise its own discretion,[77] I would for myself conclude that there was in all the circumstances no basis for excluding the evidence of her identification from the photographs upon the basis of the "Christie" discretion.

It would have been an empty formality - it could even, in some circumstances, have been unfair - to have held an identification parade so far as Mrs McHugh was concerned because she had already viewed the photographs, and an identification of the appellant at such a parade would justly have attracted the criticism that she did so as a result of the displacement effect.[78] When the appellant was arrested, he declined to participate in such a parade, first at his father's insistence and later on his solicitor's advice, but he had already been told by the police that he had been identified from photographs and it is fair to assume that his refusal was dictated by the understandable concern that the value of any such parade had by that time become tainted. The other two witnesses never purported to identify the appellant, and it was not suggested that an identification parade should have been arranged for either of them. Upon the same basis, therefore, I am not satisfied that the absence of such a parade resulted in an unfair trial for the appellant. In my view, the police officers acted reasonably in not arranging one.

The evidence of the identification from photographs was in my opinion properly admitted, although it required substantial directions as to the dangers involved before the jury could act upon it. I return then to the principal ground of appeal, that the verdict was unsafe and unsatisfactory. The ultimate question which must be answered by this Court in relation to such a ground is whether - notwithstanding that as a matter of law there is evidence to sustain that verdict - this Court thinks that, upon the whole of the evidence, it was open to the jury to be satisfied beyond reasonable doubt that the appellant was guilty.[79] As the evidence of identification was properly admitted, I must now (in accordance with the authorities) make my own independent assessment of both the sufficiency and the quality of the whole of the evidence.

It should be said at the outset that the evidence of the appellant himself and that of the witnesses whom he called was not of such a quality that, as a result of such evidence having been given, I experience for myself any reasonable doubt in relation to the Crown case. The fact that the appellant signed in at the Drummoyne RSL Club at some time before midday on the day in question does not mean that he could not have driven to Auburn and been outside the ANZ Bank there at the time of the robbery at 1.50 pm. The strength of his case depended, rather, upon the evidence of the telephone calls which he said that he had received at his home at times which were necessarily inconsistent with the Crown case that he was at Auburn at the time of the robbery. His initial assertion to the police that he had remained at the club until 4.00 pm, although in answer to a question asked of him several months later, does not assist his own reliability as a witness.

The appellant said in his evidence that he left the club at "maybe" 1.00 pm, he walked home (taking five to six minutes) and he had lunch. Either during his lunch or after it, he said, he had received the first of these telephone calls, one from a female friend whom he denied was a girlfriend. When this person gave evidence, she described herself as a close personal friend of the appellant. She said that she had spoken to him on the telephone at around 1.00 pm, or roughly then. She had not been asked to recall this event until about twelve months later, but she said that she had recalled it because it had been on the day when the car was stolen. She had said to the police when interviewed that the appellant had told her when she spoke to him that his father's car had been stolen, but she said in evidence that this had been during another call later that day. The version which she gave to the police was an important inconsistency on her part.

The appellant said that the second call was received by him at 1.30 pm. It was from Tom Tamani, the man whose fingerprints had been found on the car owned by the appellant's father and used by the offender. The appellant said that Tamani had driven off in the car with his father earlier on the day of the robbery. He said that Tamani had telephoned him at 1.30 pm to inform him that his father's car had been stolen. Tamani was not called as a witness, and no explanation was given for his absence. The inference that his evidence would not have assisted the appellant's case is obvious. It was established that Tamani had told the police that he had not noticed that the car was stolen until 3.00 pm - that is, an hour and a half after the appellant said that the call had been made - and that it had been the appellant's father (rather than the appellant) whom he had told that the car had been stolen. That was an important inconsistency with the evidence of the appellant.

The appellant said that the third call was received by him at 2.30 pm. It was from his cousin by marriage, who was called to give evidence. The cousin said that he had telephoned the appellant at about 2.30 pm, after being interviewed by two police officers who were inquiring as to the whereabouts of the car owned by the appellant's father and registered at his (the cousin's) address, to tell the appellant that the car had been used in an armed robbery. It would seem that it was only at that stage (as soon as he had finished speaking to the cousin) that the appellant telephoned his father to inform him that the car had been stolen so that he could report it although, when pressed in cross-examination, the appellant did suggest that he "would have" called his father earlier as well, after hearing from Tamani that the car had been stolen. His father was not called as a witness, but it was established that his father had not reported the car stolen until 4.00 pm, and that he had told the police that it had last been seen at 11.45 am. That too was inconsistent with the evidence of the appellant.

The appellant's argument that the verdict was unsafe and unsatisfactory must therefore depend upon the lack of quality of the Crown evidence rather than the strength of his own. To a very large extent, this means the evidence of Mrs McHugh, the victim of the robbery. Only two other witnesses saw the offender, and neither purported to identify the appellant as that person. I will refer to their evidence later.

Before turning to the evidence of Mrs McHugh, it is necessary to emphasize again that this Court's task in determining whether it has, or whether the jury ought to have had, a reasonable doubt as to the appellant's guilt[80] is quite different to that of the trial judge in determining whether to exclude evidence in accordance with the "Christie" discretion because its prejudice to the accused outweighs it probative value. It was not argued at the trial that the evidence of Mrs McHugh as a whole should have been withdrawn from the jury upon the basis that its weaknesses were such that its probative value did not outweigh the prejudice created by the identification from photographs. The only argument was that that evidence of the identification itself should be excluded, and that argument took place before Mrs McHugh had given evidence. But, even if the argument had been extended to the whole of her evidence, that argument should in my opinion have been rejected because, if accepted, parts of her evidence had considerable probative value.

The appellant submitted that, in finding him not guilty of armed robbery and guilty only of robbery, the jury must have had a reasonable doubt as to the reliability of Mrs McHugh's evidence upon which the more serious charge necessarily depended. Her evidence was that, at the time when the offender grabbed her bag with one hand, his other hand was behind his back. It was only after he walked away that she saw a small black object in the hand which he had held behind his back which, she said, she thought was a gun with a barrel approximately four inches in length. In cross-examination, she said that to her it had looked like a gun, and later she said that she was one hundred per cent positive, one hundred per cent sure, that he was holding a gun. The bank manager also saw a black object of some sort in the offender's other hand, but he was unable to say what it was. The other witness did not give evidence of having seen anything in the offender's hand.

I do not accept that the jury's verdict of not guilty of armed robbery necessarily involves any doubt on their part as to the reliability of Mrs McHugh's evidence. During her cross-examination, the jury asked the following question:

"Does the victim have to see the weapon at the point when the money is handed over to constitute armed robbery?"

The judge told the jury that he would give them directions in due course in the summing up upon this issue, but in the absence of the jury he commented that basically the answer would be "yes". Unfortunately, no directions were in fact given upon this issue, but it is not difficult to understand how it may seem to the lay person that the more serious offence is the more serious because the victim knows, at the time when it is committed, that the offender is armed. It is unnecessary for the purposes of this appeal to decide whether or not that is the law, although it seems that it may not be.[81] It is sufficient to say that, in the light of the question which was asked, the jury may well have approached their task upon the basis that, as Mrs McHugh did not see the gun at the time of the robbery, the appellant was not guilty of the more serious charge.

Mrs McHugh did, however, have some difficulty with her evidence in relation to which hand the offender was holding what she was certain was a gun. In her evidence in chief, she said that the offender had grabbed the bag with his left hand, that his right hand was behind his back and that it was in his right hand that she later saw the gun. In cross-examination, she asserted that she was sure that he had grabbed the bag with his left hand but, when reminded that in the statement which she made on the day of the robbery she had said that the offender had grabbed the bag with his right hand, she was forced to concede that she was in error. The bank manager had not seen the offender grab the bag, but he did say that during the struggle he could see that the offender was holding the bag in his left hand and the black object in his right hand. There was no evidence as to whether the appellant was right or left handed.

It was of little importance whether the offender used his right or his left hand to grab the bag, and usually an inaccurate recollection of what happens in a highly stressed (and often confused) incident such as a bag snatch can be ignored. What is significant here, the appellant says, is the readiness of Mrs McHugh to assert a degree of certainty in her recollection where it was unwarranted.

Mrs McHugh initially gave evidence that her struggle with the offender had lasted approximately two minutes, and that she was looking directly at him during this time. She said that it may have been for a less time, but she was sure about her estimate of two minutes. In cross-examination, she denied that it could have been less than one minute, and she said that she was sure that it had lasted at least one minute. She was asked to sit during a timed period of one minute. When this had been done, she commented that it may be that the struggle had seemed like two minutes, but she remained sure that it was not less than one minute. The pedestrian, whose attention was drawn to the incident by Mrs McHugh's screams, estimated the time of the struggle to be very short, a matter of seconds.

The bank manager, who was inside the building, heard Mrs McHugh's screams and went to the window where he saw the struggle already in progress. He estimated that the struggle lasted thirty to fortyfive seconds, but he agreed that it could have been shorter by five seconds. There was then some discussion, and the judge indicated that he was interested to know how long the witness had been watching from the window. Counsel then appearing for the appellant asked the bank manager:

"Q. From the time that you heard the scream to the time when the robber broke free, how long was that do you say? A. Roughly half of that, like twenty, twentyfive perhaps."

It is reasonably obvious that the witness was answering the judge's question rather than that which he was in fact asked. He was asked to reconstruct the events in his mind, as if playing it back on a video, and to indicate how long it took whilst he was being timed. It is not clear which events he was reconstructing - that is, from the scream or from when he arrived at the window - but it was agreed that the reconstruction took only twelve seconds.

In my view, estimates of elapsed time (as with estimates of a person's height or age) are usually of little value when made only in retrospect. They are not like the impression made at the time and recalled as to the colour of a person's eyes or hair, or whether he or she had some particular characteristic like a broken tooth or a scar or something like that. Estimates of height (or age) made later have to be made in the absence of the person being described. It is even more difficult to make an estimate of elapsed time in retrospect. Reconstructions of events - particularly when done mentally - suffer from many problems which do not assist in obtaining an accurate estimate of the time which some event took.

The discrepancy between Mrs McHugh and the other witnesses here, however, was marked. It affected not only the time during which she had to observe the offender, but it too is another example of the readiness of Mrs McHugh to assert a degree of certainty in her recollection which was unwarranted.

There were two other very significant elements of the description of the offender given by Mrs McHugh which did not fit the appellant. The first was the colour of his eyes. She said that the offender's eyes were dark brown and quite big. In cross-examination, she said that there was no doubt in her mind that the offender's eyes were dark brown. The appellant gave evidence, which the Crown accepted, that his eyes were green. No suggestion was made that he was wearing coloured contact lenses at the trial, or anything of that sort.

The other significant description was of the offender having a dark olive complexion. Again, Mrs McHugh said that there was no doubt in her mind that his complexion was dark olive in colour. She agreed that, in her statement made the same day as the robbery, she had also said that the offender was Middle Eastern in appearance. She confirmed that this was so, and she agreed that, by Middle Eastern, she meant someone from The Lebanon or who is Arabic by extraction. The appellant gave evidence, again which the Crown accepted, that his father had been born in Italy. He was present at the hearing of the appeal, and it was clear to me that he has a darkish complexion, consistent with those from the south of Italy. It was not dark olive in colour. Mrs McHugh said that she was confident that she could distinguish between nationalities, and in particular between those of Italian and Middle Eastern appearance. The bank manager described the offender as being perhaps of Turkish background by reason of what he said were the offender's colouring and facial features.

For the Crown to have succeeded, the jury would have had to have been satisfied that these two elements of the description were merely errors in recollection and that there was no reasonable possibility that the offender was someone other than the appellant. It is difficult to see how Mrs McHugh could have been mistaken in relation to both matters in the light of the certainty with which she expressed herself, although (as it has been demonstrated) she was equally certain about other matters which are the subject of dispute. When the bank manager's description of the offender's Turkish appearance is added to that of Mrs McHugh, the reasonable possibility that the offender was someone other than the appellant is all the greater. I should point out that, when Mrs McHugh identified the appellant from the photographs, the colour of his eyes and of his skin would not have been apparent to her as the photographs were in black and white, not in colour - an issue to which I will have to return.

There were some discrepancies between the three witnesses as to other elements of the description of the offender. Mrs McHugh said that he was about five feet eight or nine inches (which is not too far out so far as the appellant is concerned), whereas the bank manager said that the offender was approximately six feet tall (which is substantially out), but - as I have already said - I regard such estimates made retrospectively as being of little value. The clothing worn by the offender is described in different terms by each of the three witnesses, but the differences are not of great significance in the overall picture.

What is significant is that, except for the rather general respects in which the description given by Mrs McHugh fitted the appellant (and no doubt many others) - as being in about his early twenties, about five feet eight or nine inches tall and clean shaven - there is, as the Crown conceded, absolutely no evidence which supports the accuracy of the identification of the appellant which Mrs McHugh made from the photographs, and there are at least two very significant inconsistencies between the appellant and the offender as described by her - neither of which would have been apparent to Mrs McHugh when she made that identification. The strength of the Crown case, as it eventually turned out, depended almost entirely upon that identification and upon the fact that the appellant could possibly have had access to the car used by the offender. That latter fact is of little weight unless the identification was a good one. I turn then to the circumstances under which that identification was made.

The identification was not made until almost three and a half months after the robbery had taken place. An unconvincing explanation was given by the police for the delay, but the relevant fact is that the delay took place. The photographs were in black and white and not in colour. Again, an unconvincing explanation was given by the police, but the relevant fact is that - despite the dependence of the description given by Mrs McHugh upon the colour of the offender's eyes, skin and hair - she was shown photographs (specially taken for the purpose!) where she was not given the assistance of colour and which, in two very significant respects, masked the difference in those colours between the offender as she described him and the appellant.

Mrs McHugh was asked the inevitable question in cross-examination as to whether she had selected the photograph of the person who looked most like the offender, and she replied:

"When I came to that photograph, that was the photograph I chose because I'm sure that in that photograph that was the person who robbed me. But I don't think I went any further when I saw that photograph."

She agreed that she had in fact looked at the photograph of the appellant twice. It was the fifth of ten photographs shown to her. She said that, even though she could not see the colours of the appellant's eyes and skin in the photograph, she was one hundred per cent sure that she was not mistaken. She said at first that it was not the jacket which the appellant was wearing in the photograph which drew her attention to him, it was his facial features. When she was reminded that she had at the time of the identification said that it could even have been the jacket which he was wearing at the time of the robbery, she conceded that, although it was more his facial features, the jacket would also have attracted her attention as it was the same type of jacket. In fact, the jacket in the photograph has a series of horizontal lines on it whereas her description of the offender's garment was that it was "checkered".

The photograph of the appellant which Mrs McHugh identified is taken from some distance away, and the enlargement is grainy in its definition. The appellant is holding his hand up to shield his face from the sun, and consequently most of his face is in shadow. Apart from showing him to have a somewhat fleshy appearance, it provides no clear indication at all of his features. That Mrs McHugh could be so certain in her identification in those circumstances is difficult to understand.

Reliance was also placed by the appellant upon Mrs McHugh's answer to the police officer's question at the time of the identification, as to whether she was sure that the appellant was the offender - "As sure as I can be" - but that statement must be interpreted in its context. She had immediately beforehand said:

"I'm positive in my own mind that he is the person who robbed me of the money at Auburn in May."

In that context, I do not think that it would be fair to conclude that the phrase "As sure as I can be" used by Mrs McHugh (an equivocal phrase in any event) was an indication of some doubt on her part.

But what remains is nevertheless in my view an unsafe basis for the verdict which the jury gave. It is a grave step to take to upset a jury's verdict in a case such as this, but it will not be the first time that the many dangers of identification evidence where the offender was previously unknown to the victim have led to an unsafe result, and it will not be the last. In the end, I conclude that, despite many suspicious features of the case, there is a significant possibility that an innocent person has been convicted.[82] In my opinion, the conviction should be quashed. In those circumstances, it is unnecessary to deal with the other grounds of appeal, but there are three particular matters to which reference should be made.

First, complaint was made that, although the trial judge was said to have known that the crucial issue in the case was the correctness of the identification made by Mrs McHugh, he failed in his preliminary directions to alert the jury to that crucial issue and to give them directions at that stage as to how they should approach such an issue. In my view, it would usually be very unwise for a trial judge to nominate such an issue in the preliminary direction as the crucial issue. Rarely does a trial turn out to produce a straightforward single issue, and on an appeal counsel for the appellant (who rarely appeared at the trial) will inevitably say that something else should have been mentioned as well. If counsel appearing for the accused at the trial wishes to narrow the issues at that stage, leave should be sought to make a statement to that effect at the conclusion of the Crown prosecutor's opening address. Trial judges should welcome the making of such a statement, and I understand that the Director of Public Prosecutions has encouraged prosecutors to consent to such a procedure being adopted.

Secondly, complaint was made that, when the judge was considering whether to exclude the evidence as to the identification of the appellant from photographs, he failed to take into account the provisions introduced by the Evidence Act 1995. It was conceded that, as the identification had taken place before that Act commenced, ss 114-115 of the Act did not apply to it.[83] It was nevertheless submitted that the provisions of those sections remained relevant as showing the quality of the evidence required to guard against people being wrongly convicted. That submission must be rejected. The fact that the Legislature has now imposed in that Act what are in some respects more stringent conditions of admissibility of identification evidence and at the same time expressly provided that those more stringent conditions do not apply to identifications which took place before the Act commenced cannot be twisted by any stretch of logic into the proposition for which the appellant contends. Those provisions were irrelevant.

Lastly, complaint was made that the judge failed to identify for the jury, in accordance with Domican v The Queen, a number of weaknesses in the evidence of Mrs McHugh which they were bound to take into account. It is not now necessary to deal with them individually, but such was the nature of some of the matters raised that it would appear to be necessary to draw attention once more to this Court's decision in Regina v Heuston,84 that the special "Domican" rule does not apply to matters of credit. The failure of a trial judge to deal in the summing up with matters of credit (in the sense of the honesty of the witness as opposed to the reliability of his or her evidence) must be judged in accordance with the rules applicable to the ordinary case. The "Domican" rule applies to matters of reliability, so as to ensure that the jury is made aware of the seductive effect of honest evidence relating to identification which may nevertheless be unreliable. Matters going to the honesty of the witness do not fall within it.

I propose that orders be made that the appeal be allowed, the verdict quashed and a judgment of acquittal be entered.

NEWMAN J: I agree with the orders proposed by Hunt CJ at CL, for the reasons which he has given.

IRELAND J: I too agree with the orders proposed by Hunt CJ at CL, for the reasons which he has given.

[1] Although the trial took place after the commencement of the Evidence Act 1995, the identification of the appellant from photographs took place before that commencement, so that ss 114-115 of that Act did not apply: Evidence (Consequential and Other Provisions) Act 1995, Sched 2, cl 9.

[2] Rex v Fannon & Walsh (1922) 22 SR 427 at 429-430; Regina v Doyle [1967] VicRp 82; [1967] VR 698 at 699; Regina v Russell [1977] 2 NZLR 20 at 27; Alexander v The Queen [1981] HCA 17; (1981) 145 CLR 395 at 399, 427, 434.

[3] The dangers of using photographs for identification have been discussed in the cases cited in the last footnote, as well as in Davis & Cody v The King [1937] HCA 27; (1937) 57 CLR 170 and Regina v Bouquet (1961) 62 SR 563, amongst others.

[4] Domican v The Queen [1992] HCA 13; (1992) 173 CLR 555 at 561.

[5] Rex v Christie [1914] AC 545 at 559, 564; Noor Mohamed v The King [1949] AC 182 at 192; Driscoll v The Queen [1977] HCA 43; (1977) 137 CLR 517 at 541; Phillips v The Queen [1985] HCA 79; (1985) 159 CLR 45 at 51. See now s 135 of the Evidence Act, which does not depart from the common law discretion: ALRC 26, vol 1, par 643.

[6] Alexander v The Queen (at 402-403, 417, 430, 433, 435); Regina v R (1989) 18 NSWLR 74 at 76; Doney v The Queen [1990] HCA 51; (1990) 171 CLR 207 at 212; Regina v Tugaga (1994) 74 A Crim R 190 at 193-194; Regina v Heuston (1995) 81 A Crim R 387 at 389. A number of cases in which the quality of identification evidence generally has, in the absence of an identification parade, led to a consideration of whether it should be excluded are collected in Regina v Pearsall (1990) 49 A Crim R 439 at 444.

[7] Scott v The Queen [1989] AC 1242 at 1259; Regina v R (at 79-80); Regina v Tugaga (at 196).

[8] McDermott v The King (1948) 76 CLR 501 at 506-507, 513-515; The King v Lee [1950] HCA 25; (1950) 82 CLR 133 at 150-151; Macpherson v The Queen [1981] HCA 46; (1981) 147 CLR 512 at 519-520; Cleland v The Queen [1982] HCA 67; (1982) 151 CLR 1 at 18, 31, 34; Phillips v The Queen (at 51); Foster v The Queen [1993] HCA 80; (1993) 113 ALR 1 at 8-10, 13-14, 19-20.

[9] Cf The Queen v Ireland [1970] HCA 21; (1970) 126 CLR 321 at 334-335; Bunning v Cross [1978] HCA 22; (1978) 141 CLR 54 at 74-75; Collins v The Queen [1980] FCA 72; (1980) 31 ALR 257 at 317; Cleland v The Queen (at 9, 34-35); Regina v Zorad (1990) 19 NSWLR 91 at 99; Regina v Edelsten (1990) 21 NSWLR 542 at 551-554; Foster v The Queen (at 20-21).

[10] Tamani's photograph is Ex F which, according to the evidence, shows what he looked like a week after the robbery.

11 [1977] 2 NZLR 20.

15 (1981) 145 CLR 295.

[21] [1937] HCA 27; (1937) 57 CLR 170. See p 410 of the judgment of Gibbs CJ.

[28] (at 413). The quotation is from Rex v Bagley [1926] 3 DLR 717 at 718-719.

[29] (at 415). The case was Regina v Goode [1970] SASR 69; the statement was made by Bray CJ on behalf of the Full Court at 79.

[33] (at 419). He also refers (at 422) to the detection process having "ended".

[42] (at 436). The reference to the Report is ALRC 2, at p 56.

43 (1987) 29 A Crim R 213.

45 (1987) 30 A Crim R 278.

47 (1990) 47 A Crim R 458.

[56] Rex v Bagley (at 718-719), quoted by Stephen J at p 413.

[59] Crimes Act 1900, s 352(2).

[69] These propositions have, generally, been distilled from the judgment of Mason J in Alexander v The Queen, with which Aitkin J agreed and which is consistent with the judgment of Gibbs CJ.

[70] In relation to identifications made after the commencement of the Evidence Act 1995, regard must be had to ss 114-115.

[71] Rex v Rothery (1925) 25 SR 451 at 461; Jayasena v The Queen [1970] AC 618 at 624; Haw Tua Tau v Public Prosecutor [1982] AC 136 at 150-151; Regina v Towers (1984) 14 A Crim R 12 at 15; Regina v R (at 81); Doney v The Queen (at 214-215).

[72] M v The Queen (1994) 181 CLR 487 at 492-495; Regina v McKnoulty (1995) 77 A Crim R 333 at 338-339.

73 [1990] HCA 51; (1990) 171 CLR 207.

74 [1992] HCA 13; (1992) 173 CLR 555.

[75] See also this Court's decision in Regina v Heuston.

[76] Regina v R (at 79-80); Regina v Tugaga (at 196).

[77] House v The King [1936] HCA 40; (1936) 55 CLR 499 at 504-505; Bloch v Bloch [1981] HCA 56; (1981) 55 ALJR 701 at 703; Regina v Alexandroaia (1995) 81 A Crim R 286 at 290.

[78] Alexander v The Queen (at 400); Regina v Browne & Ors (at 289); Regina v Pearsall (at 442); Regina v Tugaga (at 194-195).

[79] **M v The Queen (**at 492-495); Regina v McKnoulty (at 338-339).

[80] (See Regina v McKnoulty, at 338).

[81] In Miller v Hrvojevic [1972] VicRp 31; [1972] VR 305 at 306, Lush J held that to be armed with a weapon means something more than to be in possession of it; the weapon must also be available for immediate use as a weapon. In that case, the knuckle duster was in the dependant's left-side trouser pocket, where it could easily and rapidly be slipped into his hand. That decision, which was followed in Wilcox v Smith (James J, 2 June 1994, unreported), suggests that a person may be armed with a weapon without the knowledge of the victim, although this precise point does not appear to have been argued.

[82] I have here adopted one of the phrases used in M v The Queen (at 494).

[83] Evidence (Consequential and Other Provisions) Act 1995, Sched 2, cl 9.

84 (1995) 81 A Crim R 387 at 399.