Domican v The Queen [1992] HCA 13; (1992) 173 CLR 555, Edwards v The Queen (1993) 178 CLR 195, Pollitt v The Queen [1992] HCA 35; (1991) 174 CLR 558.

1 On 9 August 2007 the applicant was convicted on the murder of Mersina Halvagis at Fawkner on 1 November 1997. This is an application for leave to appeal against the conviction on a range of grounds. The facts of the matter are set out in the judgment of Ashley JA. For the reasons which follow, I would allow Grounds 5 and 6 and order that a new trial be had.

2 It is convenient to begin with Grounds 2 and 6 and to deal with them together. Under Ground 2, the applicant contended that identification evidence given by the witnesses Burman, Melnik and Weller was inadmissible, or should have been excluded by the judge in the exercise of discretion and, under Ground 6 the applicant contended that, if the evidence were properly admitted, the judge erred in the directions which his Honour gave to the jury concerning that evidence and that caused the trial to miscarry.

3 I reject Ground 2. I agree with Weinberg JA, for the reasons he gives, that the evidence was admissible and that the judge did not err in the exercise of his discretion in refusing to reject it.

4 I take a different view concerning Ground 6. I consider that the directions were inadequate. The law on the directions to be given to juries concerning disputed identification evidence appears to be clear. The High Court has laid down that a trial judge must warn the jury as to the dangers of convicting on such evidence and has said that, although the terms of the warning need not follow a particular form:

it must be cogent and effective. It must be appropriate to the circumstances of the case. Consequently, the jury must be instructed 'as to the factors which may affect the consideration of [the identification] evidence in the circumstances of the particular case'. A warning in general terms is insufficient. The attention of the jury 'should be drawn to any weaknesses in the identification evidence'. Reference to counsel's arguments is insufficient. The jury must have the benefit of a direction which has the authority of the judge's office behind it. It follows that the trial judge should isolate and identify for the benefit of the jury any matter of significance which may reasonably be regarded as undermining the reliability of the identification evidence.[1]

5 In this case, in my opinion, the judge did not comply with those requirements. His Honour gave the jury the general warnings prescribed by Domican. He instructed them that:

You must be cautious before you act on identification evidence and you must be very careful in considering that evidence. You all know from your everyday experience, ladies and gentlemen, that honest people can be wrong in identifying someone. That is also, ladies and gentlemen, the experience of the law and I tell you that fact, that honest people can be wrong in identifying persons and in the past, have been wrong in identifying persons. Further, because very often identification witnesses are honest and the question is not whether they are honest but whether they are reliable and accurate, that evidence can be seductive in the sense that they can look good but they can be wrong. So the issue with these identification witnesses is not, are they honest or not, but are they accurate and reliable or not? ... The experience of the law shows that and your experience of life would also show that and you must be very careful in considering identification evidence.

I will give you two further warnings, ladies and gentlemen. You must not add some outside fact to an identification witness's evidence and use that outside fact to bolster the identification witness's evidence. You look at the identification evidence itself as identification. Further, you must not add another identification witness to the one you are considering and use that other identification witness to bolster the one you are considering. Again, you examine the identification witness on a stand alone basis, are they accurate and reliable, or not? This is a further direction of law but I'm sure it coincides with your experience of life, ladies and gentlemen. In considering identification evidence, you consider all the factors which you consider could bear upon the accuracy and reliability of identification evidence. Obvious ones are the state of the lighting, the distance between the witness and the person observed, the length of time of the observation, the reason or the lack of reason for a witness to notice the other person or to notice their features, the fact that the witness had not seen the other person before and did not therefore recognise them, the lapse of time between the event and having to recall it and the circumstances of when the witness did recall it and in particular, whether in the circumstances of recall, not only was there a lapse of time and how much, but were there any interfering factors in the circumstances of recall; what technically is called displacement. That simply means that later images intervene between the event of the recollection and displace the original image in your mind and you remember the new image, not the original image because it has been displaced. They are all factors you look at, ladies and gentlemen, in considering identification evidence. That is the experience of the law and I am sure that is equally your common sense and experience of life, ladies and gentlemen.

Of course, once you have properly exercised caution and care in assessing identification evidence, you may act on the evidence as indeed the prosecution has submitted you should but always approach it with caution and care, ladies and gentlemen, bearing in mind what I have said.

6 The judge did not, however, direct the jury as to any of the matters specific to the case which might reasonably be regarded as undermining the reliability of the identification evidence, or even the matters referred to by counsel as affecting the reliability of the identification evidence; still less direct the jury that they were bound to take those matters into consideration in determining whether they could rely on the identification evidence.

7 As Vincent JA observed in R v Debs and Roberts,[2] general directions of the kind prescribed by Domican serve as an introduction designed to draw the jury's attention to the nature of the problem and some of the matters which may impact on the reliability of evidence of visual identification. But in view of what was said in Domican, the matter cannot be left there. A trial judge is required to address the particular evidence given in the trial and thus by reference to the evidence to isolate and identify for the benefit of the jury any matter of significance which may reasonably be regarded as undermining the reliability of the identification.

8 In this case, in my view, Domican required the judge to identify in relation to each identification witness individually each matter which it was apparent from the evidence was capable of undermining the reliability of the witness's identification of the applicant; to explain the possible significance of each such matter by reference to the evidence which bore upon it; and, in particular, by reference to the evidence, to explain the possibility that each witness's evidence was to a greater or lesser degree possibly affected by delay and displacement. Since the judge did not undertake that exercise, in my view his directions were deficient.

9 Admittedly, as Weinberg JA points out, defence counsel made extensive submissions to the jury on the factors which conceivably undermined the reliability of each identification witness's evidence, including a detailed explication of the displacement effect and how it may have operated in each case. Further, as can be seen in the judge's brief general directions on identification which I have set out above, his Honour also referred to the displacement effect and gave a brief explanation of its operation, albeit without reference to the evidence which bore upon that possibility. Additionally, towards the end of the charge, his Honour summarised all of the evidence, including the identification witnesses' evidence in chief and cross-examination; and, as well as that, the judge summarised counsel's arguments, including defence counsel's submissions as to the unreliability of the identification witnesses' evidence, and the possibility that their evidence was affected by the displacement effect. Arguably, the result of all that was sufficient to bring home to the jury the need to consider the factors which undermined the reliability of the identification witnesses and in particular to have regard to the displacement effect as defence counsel had requested them to do. Nevertheless, according to Domican, the applicant was entitled to have the benefit of a direction which had the authority of the judge's office behind it and, for that purpose, a warning in general terms even when coupled with reference to counsel's arguments is insufficient.

10 There are some cases where it is open to conclude that a deficiency in Domican directions has not been productive of a miscarriage of justice.[3] Debs and Roberts[4] was a striking example of that kind in which the strength of the Crown case, and the relative insignificance of the identification evidence, impelled the conclusion that the deficiencies in the Domican warning could not have made a difference. But this case is not of that type. Here, the identification evidence formed a significant, if not essential, part of the Crown case; and the jury's perception of the identification evidence was likely to vary profoundly according to their understanding of the factors which affected its reliability. In the case of two of the identification witnesses,

10 and to some extent also in the case of the third, there was a powerful displacement argument (based upon the influence of media publicity on the witness's identification of the applicant). But in the absence of a specific judicial direction on the point, the jury may have failed to give it the attention it deserved.

11 This was a long and difficult trial in which the applicant started with the extraordinary disadvantage that the jury was told that he had twice before been convicted of the murder of women and they knew that he had been judged by the media to be guilty of the murder of Miss Halvagis. Consequently, even assuming that the jury strove to act impartially, they are still likely to have regarded the applicant with a degree of suspicion and hostility which not even the clearest of directions could wholly overcome. They may also have viewed defence counsel, or at least his arguments on behalf of the applicant, with a concomitant degree of suspicion and hostility, and so been more than usually inclined to view the arguments on behalf of the accused as the dissemblance of a guilty man. At least, one may rationally suppose that the jury would have been disinclined to accord the defence arguments the attention which they warranted unless specifically directed by the judge to take them into account.

12 In such circumstances, the applicant was entitled to a trial 'conducted with all safeguards the law can provide'.[5] In this context, that meant jury directions which complied strictly with Domican. And in my view, it is not open to conclude that proper directions in accordance with Domican, with the full weight of the judge's authority behind them, could not have made a difference.

13 It follows, in my judgment, that the judge's failure to abide by Domican was productive of a miscarriage of justice.

Ground 4: Admissibility of Evidence of Lies as Evidence of Consciousness of Guilt

14 Under Ground 4, it is alleged that the judge erred in admitting, as evidence of consciousness of guilt, evidence that the applicant told several lies to the effect that he had never been to the Fawkner cemetery and did not have any relatives or other persons whom he knew buried in the cemetery.

15 The first lie was one alleged to have been told by the applicant to Detective Bateson on 12 February 1998, in substance that he had never been to the Fawkner cemetery and did not know anyone who was buried there.

16 The second lie was alleged to have been told by the applicant to a former priest, Patrick O'Brien, in April 1999, in substance that he had never been to Fawkner cemetery and did not have any relatives buried there.

17 The evidence on which the Crown relied to establish that the statements that the applicant had never been to the cemetery were lies was the evidence of the identification witnesses, Burman, Melnik and Weller. The evidence on which it relied to establish that the statements that the applicant did not know anyone buried in the cemetery and did not have any relatives buried in the cemetery was evidence given by John Saward that the applicant had told him in 1997 before the deceased's death that the applicant's grandfather was buried in the cemetery and evidence given by Margaret Newman that the applicant had told her on a date before 1 January 1998 that there was someone he knew who was buried there.

18 Defence counsel submitted before the judge below that the evidence should not be admitted because it was too imprecise in its terms and because it was impossible to say that, if the statements were false, they were lies told out of consciousness of guilt of the offence charged.

It is said by the prosecution that both those pairs of the statements can be shown to be lies, that is to say deliberately untrue as distinct from merely erroneous, on two differential bases. The first is the evidence before this jury of identification witnesses [Mrs] Burman, Mr Weller and [Mrs] Melnik who, if the jury accept, give evidence of the presence of the accused at the Fawkner Cemetery on the very day of the murder, 1 November 1997, but prior to the murder. If that evidence is accepted, that is evidence the jury is capable of acting upon that the accused indeed had been to the Fawkner Cemetery on 1 November 1997. The second category is said by the prosecution to be capable to be demonstrated as being a lie by reason of the evidence of two civilian witnesses, Mr John Saward and Ms Margaret Newman. Mr Saward, a carpenter, had a relationship with Ms Newman. Mr Saward gave evidence as appears at p.1052 that he and Mr Dupas who had struck up a friendship because of their common interest in woodwork were driving past the Fawkner Cemetery in the first week of September 1997 and, quoting Mr Saward's evidence at page 1052, 'He said his grandfather was buried in there somewhere. I asked him did he know where his grandfather was buried and he said he didn't.' That matter was again touched upon in cross-examination without traverse at pp.1055 and 1056. In relation to [Ms] Newman, she said that, on a date, unspecified but on the evidence the inference arises that it was prior to January 1998 at the latest, the accused said that 'there was someone there', someone buried in the Fawkner cemetery. She went on to say, 'I couldn't say for sure who it was'.

First, looking at the question of whether the statements can be demonstrated to be lies by independent evidence, I consider that it is quite clear that the statements are capable of being demonstrated to be lies by reason of the conflict between the clear and unequivocal statements of the accused to the police officer and the priest on the one hand and the antecedent statements of the accused to a Mr Saward, the woodworker, and his companion, [Ms] Newman. That is in relation to knowing his grandfather or a relative was buried in Fawkner Cemetery. As to never having been to Fawkner Cemetery, the evidence if accepted of the witnesses [Mrs] Burman, Mr Weller and [Mrs] Melnik contradicts that. Thus, there is a proper evidentiary basis for the jury, if it were satisfied, concluding that the statements of the accused to Senior Sergeant Bateson and Mr O'Brien were both untrue and deliberately so, that is to say, were lies and not merely errors.

Next, the lies are precise and coherent, not vague and uncertain. The circumstance that one is a relative and one is a category, in my view, is a matter of detail. Next, I consider they relate directly to a material issue in this case, that is to say, the presence of the accused at the Fawkner cemetery on the day of the murder. They are capable of showing a consciousness of the accused falsely to distance himself from the cemetery, that is, the cemetery in which the crime of murder was committed. The jury is entitled, properly instructed, to conclude that the statements were clear, were deliberately false, can be demonstrated by independent evidence to be false and were lies as contemplated in the line of authority commencing effectively with R v Lucas[6] and in its best known articulation, Edwards v R.[7]

There is a raft of authorities on the question of lies and consciousness of guilt, which is why I adjourned overnight to reread them ... I am satisfied that the criteria articulated principally in Edwards v R but developed in numerous cases up to the present, including R v Ciantar,[8] are properly satisfied in this case. Of course it is for the jury to decide what evidence they accept. My role is to ensure that they are not given a wrongful mode of reasoning and that is the function I am presently directing myself to.

20 It was contended in the applicant's written outline of argument that the arguments advanced below against the admissibility of the evidence should have been accepted. But counsel for the applicant did not advance any submissions in support of that contention other than to say that he relied on the arguments advanced below. In my view, the judge was right to reject the arguments for the reasons which his Honour gave.

Ground 5: Post Offence Conduct Amounting to Evidence of Consciousness of Guilt

22 In the course of final address, the prosecutor referred to a number of acts of post-offence conduct which in his submission were more than just coincidences and in combination with other evidence amounted to an overwhelming case of guilt:

... in the days after that murder of Mersina Halvagis you know Mr Dupas is getting new glasses. Maybe he wanted the particular glasses with the particular lenses that you've heard evidence about. But of course it's well known that a change in glasses can lead to a change in appearance, a different aspect. Is it just coincidence that within a few days, a handful of - just over a handful of days [after] the murder, Mr Dupas is getting new glasses?

As I say it's nothing particularly significant on its own but it has much greater significance when it's taken in conjunction with all the evidence and the same comment can be made about the evidence you heard about discussion that the accused man had with Ms D'Alberto, the mobile hairdresser. At a time both before and after 1 November 1997, as Ms D'Alberto told you, the accused wanted to change his hairstyle and over time he became more determined to change his hairstyle and what was the nature of the change? It related to the fringe. He wanted to have a part because, as you will see from that FACE image produced by [Mrs] Burman, one significant feature is the fringe, the hairstyle itself ... It's another piece of bad luck for Mr Dupas, there's evidence that he did change his hairstyle after the events of 1 November 1997. Certainly by February the following year.

There's more bad luck for Mr Dupas if he is not the murderer of Mersina Halvagis, because he had a family connection with the cemetery, the Fawkner cemetery. His grandfather is buried there. But he significantly, lies about that.

...

The point I am seeking to make is that all these matters just build up such an overwhelming case against Peter Dupas, and it [is] not just one piece of evidence, it's piece after piece after piece, which when added to the whole picture, provide an overwhelming case of guilt of the murder of Mersina Halvagis.

23 Under Ground 5, counsel for the applicant argued that the logical sub-text of the prosecutor's submission was that the applicant was conscious of his guilt of the deceased's murder and was seeking to change his appearance in order to avoid detection. Therefore, it was said, it was incumbent on the judge to direct the jury in accordance with Edwards v The Queen[9] that they were not to treat that conduct as evidence of consciousness of guilt unless satisfied that it occurred and that the only explanation for it was consciousness of guilt.[10]

24 I accept that submission. Despite the submissions of counsel for the respondent to the contrary, it appears to me to be clear that the purpose and effect of the prosecutor's submission was to persuade the jury that they should regard the applicant's conduct in purchasing new spectacles and changing his hair style as implied admissions of guilt and, in those circumstances, I consider that it was incumbent upon the judge to direct the jury in accordance with Ciantar as to the matters of which they needed to be satisfied before treating it in that fashion.[11] In my view, the judge's failure to give those directions amounted to a miscarriage of justice.

25 If this ground of appeal were the only ground of appeal made good, a question would arise as to whether the proviso could be applied. Since I have concluded, however, that ground 6 has been made out, in respect of which the proviso could not be applied, I do not stay to address that problem.

26 The thrust of Ground 7 is that the judge did less than was required in warning the jury of the dangers of acting on the evidence of Andrew Fraser; in particular, by failing to warn the jury of each specific matter which could reasonable be regarded as undermining Fraser's credibility.

27 In Pollitt v The Queen,[12] several members of the High Court expressed differing views as to how far a trial judge must go when warning a jury of the dangers of convicting on the basis of the evidence of a prison informer.

28 Dawson and Gaudron JJ (with whom Brennan J agreed on this point) said that there was no rule of law or practice which identifies the evidence of prison informers as requiring corroboration. But they added that it is always the duty of a trial judge to warn of the dangers of convicting on evidence which is potentially unreliable and that it would only be in an exceptional case that the evidence of a prison informer would not fall into that category. It followed, they said, that a trial judge should ordinarily warn the jury that it is dangerous to convict on the evidence of a prison informer unless corroborated by other evidence connecting or tending to connect the accused with the offence charged. They concluded, however, that the following warning given by the trial judge was sufficient for that purpose:

[The prison informer] is a person with a long criminal history. He is a person one would describe as being of bad character or, as [counsel] put it, a tainted person or certainly up to comparatively recently. When considering the evidence of such a person commonsense dictates that you scrutinize it very carefully.

It does not automatically follow from the fact that [the prison informer] has a criminal record and has been a person of bad character that his evidence will be untrue. Nevertheless, experience has shown that persons of that calibre may -- and I stress the word `may' -- have motives for not telling the truth. So when considering what weight you will attach to his evidence it is wise to look for independent evidence which corroborates his account, that is, evidence from some other acceptable source which implicates the accused man in the crime.

...

When you are dealing with such a person you must scrutinize his evidence very carefully when determining what weight you attach to it. You will also look to see if there is independent evidence you accept which corroborates the evidence of the witness in question. As I have said to you on a number of previous occasions such persons may have motives for lying. You must look at their evidence very carefully.[13]

29 Deane J, was similarly inclined. He held that, because the evidence of the prison informer in Pollitt was liable to be unreliable for a number of reasons which he stated,[14] it was necessary that an 'appropriate warning' be given. His Honour said, however, that:

It would be unwise to seek to formulate in the abstract the contents of an appropriate warning for every case where the prosecution places significant reliance upon the evidence of a prison informer. Such a warning is most likely to be understood by a jury and to be effective if it is moulded in the words of the trial judge to fit the circumstances of the particular case. The most that should be said is that there will inevitably be cases where the minimum necessary warning will be in unqualified terms which draw attention to the fact that the evidence is of its nature potentially unreliable, which stress the need for very careful scrutiny of it and which warn that it would be dangerous to base a conviction upon it.[15]

30 Deane J added that it would have been preferable for the judge to direct the jury 'in unqualified terms which draw attention to the fact that the evidence is of its nature potentially unreliable' for the reasons stated. But he concluded that, when considered in the context that defence counsel had not sought any further direction than was given, the direction which was given satisfied the requirement 'that an appropriate direction be given to warn the jury of the potential unreliability of the evidence'.[16]

31 Toohey J also considered that a trial judge must ordinarily warn of the risks of convicting upon the uncorroborated evidence of a prison informer. But his Honour was more specific as to what the warning should contain. He said that:

the warning to the jury should be couched directly in terms that it is dangerous to convict upon the evidence of a person who is a prison informer, that such evidence should be scrutinized with great care and that the jury must be satisfied beyond reasonable doubt as to the guilt of the accused, having regard to the potential unreliability of the evidence of a prison informer and the risk of a miscarriage of justice by attaching too much importance to it. The direction should be properly tailored to the circumstances of the particular case. This will ordinarily require the trial judge to refer specifically to those facts which are likely to render the evidence of the witness unreliable.[17]

32 Toohey J concluded, however, that, in the particular circumstances of the case, including the circumstance that defence counsel had not sought any further warning, the warning given was adequate 'to alert the jury to the dangers of acting on [the informer's] evidence'; or at least that 'No miscarriage of justice was likely to have resulted from that aspect only of the direction'.[18]

33 McHugh J went further than the other members of the court. He held that:

If accused persons are to be protected against the dangers arising from the evidence of prison informers and the fairness of criminal trials are to be ensured, juries must be directed as to the danger of acting upon such evidence. Ordinarily, that will mean that the jury must be directed that it is dangerous to act on that evidence where the account of the prison informer is not substantially confirmed by independent evidence. Thus, if the informer alleges that the accused has confessed to the crime, it will be dangerous to act on his evidence unless the making of the confession is substantially confirmed by independent evidence. If the informer alleges that he saw or heard the accused do or say something which incriminates the accused, it will be dangerous to act on the allegation unless the doing of the act by the accused or the making of the statement is substantially confirmed by independent evidence. Only in the most exceptional case, if at all, could the evidence of a fellow prisoner be regarded as independent evidence for this purpose. Furthermore, the jury should be warned not only as to the inherent or general danger of acting upon the evidence of a prison informer but, where appropriate, instructed to consider any specific matters which could reasonably be regarded as undermining the credibility of the witness. Finally, the trial judge should draw to the jury's attention the particular matters which could reasonably be regarded as confirming the evidence of the prison informer.[19]

34 But like the other members of the court, McHugh J concluded that the warning given was adequate in 'the very special circumstances' of the case (including the circumstance that defence counsel had not sought a further direction) 'to bring home to the jury the danger of acting on the evidence of [the informer]'.[20]

35 In this case, the judge gave the following warning:

Next, ladies and gentlemen, there is evidence in this case from Mr Andrew Fraser who at the relevant time was in gaol with Mr Dupas. I give you a direction of law in relation to the evidence of Mr Fraser as a person who was convicted of a serious crime and was in gaol. The experience of the law, ladies and gentlemen, and this direction also binds you, is that it is dangerous to convict on the evidence of prisoners who say another prisoner confessed to them. That is because of the obvious circumstance that first of all, the witness is a serving prisoner who has been convicted of a serious crime, second, the witness usually has something to gain or hopes to gain something such as early release and there is usually no other witness to the alleged admission.

So, ladies and gentlemen, the experience of the law is that such evidence from another prisoner is potentially unreliable and it is dangerous to convict on it. Therefore, ladies and gentlemen, you must scrutinise that evidence with great care before you act on it. Of course having scrutinised it with great care, including bearing in mind the directions of law I have just given you, you may act on that evidence as indeed the prosecution submits you should.

36 His Honour next summarised Fraser's evidence in chief, including reference to Fraser's testimony that the applicant had told him on more than one occasion that: 'There was no forensic evidence at Fawkner'; that: 'she [meaning the deceased] would not have seen [me] because [I] attacked her from behind'; and that the applicant had pantomimed killing the deceased. Then his Honour continued:

Well as you appreciate, ladies and gentlemen, it is strongly in dispute by the defence who say there was no pantomime, that is an invention of Mr Fraser. It is strongly in dispute that Mr Dupas ever said he'd 'left no forensics at Fawkner'. That's in dispute. But it wasn't put in dispute that Mr Fraser gave Mr Dupas advice on the Maher matter. It was not in dispute that in the Maher matter the question of forensics arose, the glove left at the scene.

So what the prosecution says is there is no dispute about that, there is discussion about the forensics in the Maher matter and the prosecution says the next step, which the defence denies, is after the discussion about the forensics in the Maher matter of the glove Mr Dupas added, 'I left no forensics at Fawkner'.

And what the prosecution relies upon is that it says that statement, 'I left no forensics at Fawkner' can only mean one thing, it can mean Mr Dupas was at Fawkner and was the killer. It is an implied admission that he was there at Fawkner and that he killed her. That is how the prosecution put that, ladies and gentlemen, and that is what is in dispute on that aspect of Mr Fraser's evidence.

I have gone through that ladies and gentlemen, in a bit of detail because it is a bit complicated to see what is in dispute and what is not. There is no dispute that there was discussion about forensics with Ms Maher and the glove. The dispute is did Mr Dupas go on [and] say, 'I left no forensics at Fawkner'? that is the issue between [t]he parties, ladies and gentlemen. And the prosecution says that is a clear implied admission that he had been at Fawkner and that he did the killing.

All right, that is the direction of law I give you in relation to the prison informant, that is to say Mr Fraser. I will of course take you through all the arguments about, to use [defence counsel's] expression, 'he [Fraser] was going for gold' et cetera later on but I am just stating for you what the law is and what the issue there in relation to, 'I left no forensics at Fawkner' is about.

37 As can be seen, the directions thus given satisfied the requirement of warning the jury that it is dangerous to act on the evidence of a prison informer. They did not, however, refer to the need for corroboration (because the judge had determined that there was no evidence capable of amounting to corroboration). Nor did they go to the length of informing the jury that they must be satisfied beyond reasonable doubt as to the applicant's guilt having regard to the potential unreliability of the evidence of a prison informer and the risk of a miscarriage of justice by attaching too much importance to it. They included three of the reasons identified by Deane J in Pollitt as to why a prison informer's evidence is unlikely to be reliable, namely: (1) 'that there is usually no other witness to the alleged admission' (which presumably would have conveyed to the jury that the accused was denied means of corroborating his denial of Fraser's testimony); (2) that 'the witness is a serving prisoner who has been convicted of a serious crime' (which presumably would have conveyed to the jury that Fraser was a witness of bad character); and (3) that 'a prison informer usually has something to gain or hopes to gain something such as early release'. But the directions did not condescend to the specific reasons which, according to the defence case, made Fraser particularly unreliable.

38 Shortly after the judge had concluded those directions, defence counsel took three exceptions to them. The first was, in substance, that the manner of the judge's delivery of the directions was not sufficiently authoritative. As defence counsel put it:

it's my respectful submission that your Honour did not disclose any real judicial authority when you were directing them on the difficulty with informers. It seemed, with respect, that your Honour's attention was rushed or delivery was rushed and not as firm as one would have liked at least from the defence point of view from the point of view of an authoritative statement from the Bench particularly, ...

... Your Honour's voice was fairly quiet and, with respect, you didn't catch the eye of the jury, it seemed your Honour was thinking above like that and - - -

...

Well I'd rather your Honour fix the jury's attention with both your tone of voice and the manner of presentation as your Honour's quite capable of doing with me from time to time.

39 The judge dismissed that with the observation that it was fortunate that the proceedings were videotaped, which I take to be a suggestion that, if one views the video tape, it will be seen that there is nothing in the point. I have viewed the tape, therefore, and I agree that it shows that there is nothing in the point.

40 The second exception was that the judge had not referred to corroboration and that if the judge were of the view there was no corroboration he ought instruct the jury that it was dangerous to convict on Fraser's evidence in the absence of corroboration and that there was no corroboration. After considering that matter over night the judge accepted that he should redirect the jury to the effect that there was no corroboration of Fraser's evidence and, therefore, that the jury would have to be satisfied beyond reasonable doubt on Fraser's evidence that the applicant had made the admissions to which Fraser deposed.

41 The third exception was that the judge had not tailored the direction to the circumstances of the case, as Deane J had said was necessary in Pollitt, inasmuch as he had not referred specifically to the facts which, according to defence counsel, were likely to render Fraser's evidence unreliable. Defence counsel identified those as being: (1) when in prison, Fraser believed he could obtain early release if he assisted the police; (2) Fraser's failure to mention the pantomime when first interviewed by police and his mention of it at the second occasion only after there was mention of a $1 million reward; (3) the $1 million reward itself; and (4) what counsel submitted was the inconsistency between the fact that the wounds which killed the deceased were wounds to the front of her body and the alleged pantomime which was of the applicant stabbing the deceased from behind. The judge ruled against that exception.

42 In the result, the only re-direction concerning Fraser's evidence was as follows:

Next, ladies and gentlemen, in relation to the evidence of Mr Fraser, the three implied admissions the prosecution relies upon, that he left no forensics at Fawkner plus the demonstration [pantomime] of the attack. To rely upon Mr Fraser's evidence in those respects for a conviction, you would have to be satisfied that that evidence is true. That is to say the words were said by Mr Dupas, he 'left no forensics at Fawkner' and that those words were true. That is because Mr Fraser's evidence is, if you like, stand alone and the prosecution relies upon it as an admission of guilt, three admissions of guilt, three times, he 'left no forensics at Fawkner' plus the demonstration and you would need to be satisfied beyond reasonable doubt of that evidence to act upon that in finding the accused guilty.

Next, ladies and gentlemen, as a matter of law and I direct you on this, there is no independent confirmation in the evidence of what Mr Fraser said Mr Dupas told him. So you assess Mr Fraser's evidence in that light and with that direction of law. Again, ladies and gentlemen, you have seen the witness give his evidence and be cross-examined, that is why you are here and bearing those directions importantly in mind you again decide whether you are satisfied beyond reasonable doubt on Mr Fraser's evidence that the accused made those admissions as stated by Mr Fraser.

43 Counsel for the applicant argued that both the warning and the redirection fell short of what was required. He submitted that the warning was deficient because it failed to identify the specific matters which made Fraser's evidence particularly suspect, and its effectiveness was further compromised by the way in which the judge had joined it with a re-telling of Fraser's evidence. According to counsel, that was likely to have assured the jury of the cogency of Fraser's evidence rather than emphasising its shortcomings. And the redirection was said to be ineffective because it came belatedly and was thus divorced from the body of the warning.

44 In my view, there is a degree of force in those submissions. Given the importance of Fraser's evidence in the Crown case, it would have been better if the judge had referred specifically to each of the matters identified as reasonably capable of undermining the reliability of Fraser's evidence, and specifically warned the jury of the need to take those matters into account in deciding whether they were persuaded beyond reasonable doubt that his evidence was true. It would also have been preferable if, when rehearsing the substance of Fraser's evidence in chief, the judge had referred as well to the principal defects in Fraser's reliability identified in cross-examination. Without the latter, there is some truth in counsel's complaint that recitation of Fraser's evidence tended to add to its credibility rather than highlight its deficiencies.

45 Nevertheless, it is apparent from the transcript of Fraser's evidence and final addresses that defence counsel laid great stress on Fraser's shortcomings as a witness and it seems to me that those weaknesses would thereby have been revealed to the jury. For example, in the course of Fraser's cross-examination, defence counsel established that Fraser knew that, if the applicant were convicted, Fraser stood a good chance of getting a substantial part of a reward of $1 million which had been offered by the government to anyone offering information leading to the applicant's conviction; Fraser had told the police when he was first contacted that he would not give evidence unless he got a substantial time off his sentence; it was only after Fraser was advised that the Crown would not consider reducing his sentence until and unless he had made a statement that he agreed to do so; it was only on the second occasion Fraser met police, after there was mention of the $1 million reward, that he disclosed that the applicant had told him that he had attacked the deceased from behind; it was only after the third visit that he told police that the applicant had pantomimed his killing of the deceased; Fraser had given incorrect sworn evidence in the course of a Basha inquiry, to the effect that he had not revealed to the Herald Sun newspaper while still in prison what he would say about what the applicant told him; and Fraser had told the police falsely that he had not revealed the contents of his statement to the Herald Sun newspaper.

46 Similarly, in the course of final address, defence counsel reminded the jury of the evidence which established that Fraser was a convicted drug trafficker and disgraced former solicitor who had taken unconscientious advantage of an innocent young woman to bring illicit drugs into the country; that he was motivated to give the evidence which he had, not out of concern for the deceased's family as he claimed, but because he saw it as a get-out-of-gaol card with a million dollar reward; that he had not mentioned the applicant admitting he attacked the deceased from behind until after there was mention of the $1 million reward; that he knew that the reward would only be available upon the applicant's conviction; that he had changed his evidence from what he had said at the Basha inquiry about not informing the press of the content of his statement; that he had not mentioned the pantomime to police until the third visit; and that, if the deceased were attacked from behind, as the applicant was alleged to have admitted, it was odd that she did not have one wound on the back of her body.

47 It is also to be observed that when directing the jury on the dangers of acting on Fraser's evidence, the judge said that he would come later to defence counsel's arguments pertaining to Fraser's unreliability and that, when summarising counsel's arguments, the judge specifically referred to a number of the matters which defence counsel had so identified in final address:

Finally, Andrew Fraser. ...[Defence counsel] said Mr Fraser is only interested in one person, Andrew Fraser, no one else. The only reason he pleaded guilty in the County Court to the drug charges those years ago was not because he was a good person but because his office was bugged and he was gone and that is why he pleaded guilty, not through any decency and it was a disgrace what he put that young girl through, who could have faced, theoretically, life in prison and certainly was in gaol for a month anyway. [Defence counsel] said in relation to that girl, it was a despicable and low act.

He then said that in relation to Mr Fraser, Mr Fraser's position is [that] he is in gaol, seven years with a minimum of five years before parole, an appeal had failed, his petition for mercy had failed and then a couple of years into that, Mr Scarlett [the investigating police officer] arrives playing a hunch. [Defence counsel] said this is his get out of gaol card and what's more collect one million from the community chest, referring to the old game of Monopoly where you have got to get out of gaol card and you got a community chest. He said that is what went across the mind of Mr Fraser who is only interested in Mr Fraser. Whether it was a ride for a ride or a ride there or a ride back, it meant the same thing, 'What's in it for me?'.

[Defence counsel] poured scorn on Mr Fraser's statement that he had forgotten about the demonstration [pantomime] the first time Mr Scarlett came down unannounced and only remembered it the second time[21] and said if that was true, you would never forget it. This witness said he forgot it, he did not forget it, he is making it up. You would not forget it.

48 Counsel for the applicant contended that it was not enough for the judge to refer to points which had been made in final address, and in support of that contention he invoked the rule in relation to identification evidence warnings that mere repetition of counsel's arguments is insufficient.[22] Counsel argued that, logically, that the same rule should apply to warnings as to the dangers of convicting on the evidence of unsatisfactory witnesses.

49 There is some force in that submission too. Inasmuch as the informing principle of both Domican and Pollitt warnings is the idea that it is necessary for a judge to draw to the jury's attention matters of which they might not recognise the significance, logic and commonsense suggest that, if it is necessary for the purposes of a Domican warning for a judge specifically to identify and warn of weaknesses which may reasonably be regarded as undermining the reliability of the evidence in question, it should also be necessary to do that for the purposes of a Pollitt warning.

50 As a matter of authority, however, it appears that it is not the case. For as has been seen, in Pollitt, which was decided some six months after Domican, the High Court regarded it as sufficient that the details had been dealt with in closing addresses and counsel had not seen fit to ask for a further direction.

51 It is true that, in this case, defence counsel did seek a direction that the jury take into account each of the weaknesses in Fraser's evidence. In that sense, this case is distinguishable from Pollitt. But as was pointed out in Domican, if justice requires that a warning be given, a judge's failure to give the warning cannot be excused on the basis of defence counsel's failure to ask for it. The absence of exception is simply a factor from which it might be inferred that defence counsel believed that the jury did not need any more assistance to comprehend the point, and from that to infer it was so. Conversely, as it seems to me, when one is dealing with a Pollitt warning, the significance of the fact that defence counsel has taken exception is that it provides a basis from which to infer that defence counsel believed that the jury did need more assistance to comprehend the point. It does not mean that one is bound to conclude that the jury were in need of that assistance.

52 In this case, given defence counsel's extensive cross-examination of Fraser on the matters which undermined his reliability, defence counsel's final address in which he rehearsed each of those matters, the judge's directions as to the danger in general of acting on the evidence of prison informers having motives to lie, and his Honour's recitation of defence counsel's arguments, I conclude that what was done was sufficient to bring home to the jury the danger of acting on Fraser's evidence.

53 Under Ground 8, it was contended that the judge's charge was unbalanced in two respects: first, because of failure to put adequately the defence argument that it was reasonably possible that the deceased had been killed by her fiancée, Angelo Gorgievski; and, secondly, because of failure to emphasise sufficiently the unreliability of Andrew Fraser's evidence in light of differences between his version of events, as portrayed in the pantomime of the killing which he alleged the applicant had acted out in prison, and the forensic evidence as to the manner in which the deceased was killed in fact.

54 I do not accept that contention. It appears to me that the judge dealt adequately with both matters. The judge's summary of defence counsel's argument concerning the possibility of Angelo Gorgievski's involvement was of almost two pages in length and was augmented by a full summary of Angelo Gorgievski's evidence in chief and in cross-examination. His Honour's recitation of defence counsel's argument as to the unreliability of Andrew Fraser's evidence was also thorough. It was more than two pages in length and was backed by a lengthy summary of Fraser's evidence in chief and in cross-examination.

55 In my view, the way in which the judge went about drawing attention to defence counsel's arguments and reminding the jury of the relevant evidence was adequate for the purpose of the trial. I do not think it reasonable to suppose that the jury would have been left in any doubt about the nature and effect of defence counsel's arguments or the features of the evidence on which they were based.

56 Counsel for the applicant relied upon the matters raised in support of Grounds 2, 4, 5, 6 and 7 to argue that, in sum, the verdict was unreasonable. Since I have concluded that the applicant has made good Grounds 5[23] and 6, it is strictly speaking unnecessary to consider this Ground. Mindful of the possibility that the matter may go further, however, and of the requirements laid down by the High Court in Cornwell v The Queen,[24] I make the following observations.

57 First, based upon my assessment of the totality of the evidence, I have concluded that, unless the jury were prepared to accept Fraser's evidence, they could not reasonably have been satisfied of guilt beyond reasonable doubt. In other words, even if persuaded by the identification evidence that the applicant was in the cemetery on the morning before the killing, and persuaded that the applicant lied as to not having been in the cemetery and not having a connection with someone buried in the cemetery, there would remain a reasonable doubt as to the applicant's guilt which acceptance of Fraser's evidence would be necessary to exclude.

58 Secondly, Fraser's evidence was not only the subject of credible attack but was given in a manner which, as I perceive it, did not engender confidence in its reliability. I cannot say, however, that a jury properly informed would be bound to reject it or even necessarily share my doubts about its veracity.[25]

59 Consequently, if proper directions had been given as to the matters the subject of Grounds 5 and 6, I accept that it would have been open to the jury acting reasonably to be satisfied of the applicant's guilt beyond reasonable doubt.

60 I turn finally to Ground 1, in support of which counsel for the applicant argued that the applicant's prior convictions for murder and rape, coupled with extreme pre-trial publicity of the fact and detail of those convictions and of allegations that the applicant was guilty of the murder of Ms Halvagis, made it impossible for him to receive a fair trial. On that basis, counsel submitted, the judge had erred in refusing to order that the trial be permanently stayed.

61 I agree with Ashley JA, for the reasons his Honour gives, that this is an extreme or singular case in the sense that, despite lapse of time, and no matter how careful and thorough may be the directions given to the jury, there was and for the foreseeable future there will remain a significant likelihood that any conviction of the applicant for the murder of Ms Halvagis would be affected by substantial prejudice and pre-judgment. That being so, I agree with Ashley JA that the balance of persuasion in R v Glennon[26] implies that the trial should have been stayed indefinitely.

62 The difficulty with that, however, is that the risk of prejudice and pre-judgment has been brought about by media publicity. Consequently, to grant an indefinite stay would be to recognise that the media has the capacity to render an accused unable to be tried. I do not consider that we should recognise that the media has that capacity. With respect, I adopt the view expressed by Brennan J in Glennon,[27] with which Dawson J agreed, that the community cannot afford to

acknowledge that the media has the capacity so to render an accused unable to be tried.

63 The result of that is to some degree unsatisfactory in that it leaves the applicant exposed to the unfairness of prejudice and pre-judgment. But the social imperative that an accused be brought to trial surmounts that consideration. And the consequent unfairness of that for the applicant is to be rationalised, as Brennan J reasoned in Glennon it should be, upon the basis that unfairness occasioned by circumstances outside the court's control does not make the trial a source of unfairness. The court must do all that it can to guard against the prejudicial effects of pre-trial publicity. But provided the trial is conducted with all the safeguards the law can provide, 'it is a trial according to law and there is no miscarriage of justice in a conviction after such a trial'.[28]

65 For the reasons I have given, I would allow the application and the appeal and order that a new trial be had.

66 Peter Norris Dupas is a notorious rapist and murderer. On 22 August 2000 he was convicted of the murder of Nicole Amanda Patterson at Northcote on 19 April 1999 and the details of the killing and the conviction were widely reported in the media. On 1 August 2001, his application for leave to appeal against conviction was refused and that was also widely reported in the media.

67 On 16 August 2004, he was convicted of the murder of Margaret Josephine Maher at Somerton on 4 October 1997 and the details of that killing and conviction

were widely reported in the media. In July 2005 his application for leave to appeal against that conviction was refused and that was also widely reported in the media.

68 On 11 September 2005 the applicant was charged with the murder of Mersina Halvagis at Fawkner on 1 November 1997 and, following a coronial inquest which lasted until 1 August 2006, which was also widely reported, and a four week trial in the Criminal Division of the Supreme Court, on 9 August 2007 he was convicted of the murder.

69 The applicant's modus operandi in the killing of Nicole Patterson was strikingly similar to the modus operandi employed in the killing of Margaret Maher. It consisted of multiple stab wounds to the victim's upper torso and post-mortem excision of her breasts. Consequently, evidence of the applicant's murder of Ms Patterson was admitted pursuant to s 398A of the Crimes Act 1958 (Vic) as similar fact evidence in the applicant's trial for the murder of Ms Maher.

70 The modus operandi used in the killing of Mersina Halvagis was to some extent also similar to the modus operandi used in the killings of Ms Patterson and Ms Maher, in that it involved multiple wounds to the upper torso; but it was not strikingly similar, inasmuch as her breasts were left intact. Thus there was no suggestion of admitting evidence of the other killings as similar fact evidence at the applicant's trial for the murder of Miss Halvagis. But evidence of the applicant's prior convictions - and, in an incidental but damning way, of material which implicated the applicant as the killer of Ms Halvagis - was admitted because it was inextricably linked to evidence of identification and evidence of a gaol-house confession. The judge was also of the view that the applicant's prior convictions were so notorious that it was better to face the problem head on and attempt to deal with it by appropriate directions rather than pretending that it did not exist.

71 The applicant now applies for leave to appeal, relying upon eight grounds,[29] against his conviction for the murder of Miss Halvagis. Because I agree with the reasons for judgment of Nettle JA, which I have read in draft, with respect to all the grounds other than ground 1, it is only necessary for me to address that single ground.

Ground 1: The Learned Trial Judge Erred in Failing to Order a Permanent Stay

72 Before the jury was empanelled, counsel for the applicant applied for the proceeding to be permanently stayed. The essence of his application was that pre-trial publicity concerning his client made a fair trial impossible - not simply at that time, but at any time. The publicity was to be found in a large number of articles in the popular press circulating in Melbourne, television programs which had been aired in Melbourne, in books available in Melbourne, and on internet sites. The publicity had extended over a period of some seven years preceding trial, although it had peaked at particular times. The publicity had a number of threads, often repeated.

73 The learned judge refused the application. Now the question is whether in all the circumstances - which include what transpired at trial - 'a miscarriage of justice ... occurred, and the [conviction was] unsafe, as a consequence of irreversible prejudice...'[30].

74 The applicant's solicitor swore an affidavit in support of the application. He identified seven internet sites,[31] approximately 120 newspaper articles[32] published over a period of seven years, and four books, all of which related either wholly or extensively to the applicant.

75 In support of the application, counsel for Dupas quoted extensively from the various publications. It is not clear whether the texts were formally put before the learned judge, but his Honour stated, when ruling upon the application, that he had read the articles and relevant excerpts from the books. It was not suggested at any point, I add, that what counsel read to his Honour was other than an accurate account of what had been said in the various publications.

76 It was also clear that the applicant had been referred to in a number of television programs, including an edition of the ABC program 'Australian Story', and that his image had been depicted in some of those programs. Indeed, such images were the foundation for some of the identification evidence eventually led at trial; and for other identification evidence which ultimately was not adduced. No evidence was led as to the number of relevant programs; but it was easy enough to infer the periods when such programs would have been most prevalent.

77 This is a synopsis of the pre-trial publicity, as revealed by the newspaper articles, books and internet sites:

(1) It was reported, and it was the case, that the applicant had been tried and convicted of the murders of two women. On 22 August 2000, he had been found guilty of the 19 April 1999 murder of Ms Nicole Patterson. On 16 August 2004 he had been found guilty of the 4 October 1997 murder of Ms Margaret Maher.

(2) It was reported that the killings of which he had been convicted had sexual connotations. The victims had been mutilated. Following his conviction for the murder of Ms Patterson, the applicant was described by the Herald Sun as 'a sex crazed predator', and as a 'serial rapist' who had been 'convicted of rape three times'.

(3) It was reported, and it was the case, that the killer - found by the juries to be the applicant - had used a knife when murdering Ms Patterson and Ms Maher. It was further stated that he had used or threatened to use a knife or similar implement when committing earlier offences.

(4) It was reported, and it was the case, that each of the applications for leave to appeal against the convictions for murder had been refused.

(5) Subsequent to his conviction for murdering Ms Patterson, the applicant was repeatedly implicated by the press in the murders of Ms Maher and Ms Halvagis; and as well in the murders of several other women. He was described in the popular press as 'a monster', and as a 'prime suspect' in the killing of other women. 'Prime Suspect' became an internet site. Whilst the language used was generally the language of 'suspicion' and of what the police 'believed', the unmistakeable gist of what was said, in my view, was that the applicant was guilty of the other murders.

(6) Also subsequent to his conviction for the murder of Ms Patterson, the applicant's full criminal history, which as I have said included convictions for rape, was published in the popular press - and not just once. There was a theme: how could it be that this violent repeat rapist had been allowed onto the streets so that he could offend again? For he was, in the words of the detective who had first arrested him for rape, 'evil personified'.

(7) When charged with Ms Maher's murder in late 2002, the applicant was further implicated by press articles as the killer of Miss Halvagis and two other women.

(8) Following his conviction in August 2004 for the murder of Ms Maher, the applicant was named by the press as a suspect in the killing of Miss Halvagis and another woman; and the question was posed: 'Is Peter Dupas a serial killer?'

(9) An article published in the Herald Sun on 12 August 2004, informing the public of the applicant's conviction for the murder of Ms Maher, named him as a suspect in the murders of Miss Halvagis and another woman. It further alleged that the applicant had mutilated two female corpses at the Austin Hospital in January 1969 (a matter which was never established against him).

(10) Subsequent to the announcement of a reward of $1 million for information leading to the conviction of the killer of Miss Halvagis, later increased to $1.5 million, the applicant was named as the prime suspect in the killing of Miss Halvagis and two other women.

(11) In September 2005, not long before an inquest commenced into the death of Miss Halvagis, the Herald Sun reported 'Killer accused on third case'.

(12) On 19 November 2005, The Age published an article headed 'Grave Secrets'. It set out, inter alia, the circumstances of the killing of Miss Halvagis. It listed the applicant's victims, naming Miss Halvagis and four other women, the four others including Ms Patterson and Ms Maher. That article was also placed on the internet.

(13) The applicant was identified as a serial killer on a number of internet sites, including Wikipedia. On Wikipedia there was published a full list of the applicant's previous convictions, and extracts from a statement made by Andrew Fraser, who was an important witness in the applicant's trial for the murder of Miss Halvagis.

(14) Another internet site, 'Monster Watch' stated that 'Mersina Halvagis was murdered in 1997 in Melbourne, Victoria by Peter Dupas'. To much the same effect was the website 'Serial Killer Central'.

(15) The applicant was identified as a serial killer in a book entitled 'Australia's Serial Killers'. In connection with the killing of Miss Halvagis, the author stated that 'The grave of [the applicant's] grandfather was just 100 metres away [from where Miss Halvagis was stabbed to death] and the disturbed rapist had been there often in the weeks before the murder'. The author further stated that although the applicant had denied killing any of the four women, including Miss Halvagis, 'the facts are damning'.

(16) In a book entitled 'The Encyclopaedia of Australian Murders', the applicant was treated under the heading 'The Mutilating Monster'.

(17) Another book, entitled 'Underbelly Four', written by a well-known crime reporter, John Silvester, dealt with the applicant under the heading 'Pure Evil', there setting out his life history including his previous convictions.

(18) The applicant's photograph, in connection with offences of which he had been convicted, and crimes in respect of which he was the suspected offender, was repeatedly published over the years in newspaper articles to which I have referred; and images of his face were also depicted on television programs.

78 According to counsel's submission at trial, there was an important relationship between the adverse publicity and the Crown's proofs. A number of witnesses who claimed to have identified the applicant in connection with Miss Halvagis' murder had done so by reference to newspaper photographs or television images, such photos or images accompanying publications which in substance indicted the applicant as a killer of women, including Miss Halvagis.

79 Further according to counsel's submission, despite the assumed robustness of juries and the faith that the courts have shown in them applying instructions given to them, there comes a point when a jury cannot do what the learned trial judge characterised in discussion as 'the psychologically impossible'. This was that case - an 'extreme or singular case'.

80 The prosecutor conceded that a jury was likely to know, even if not informed of the fact, that the applicant had been convicted of two murders. But it was wrong, he submitted, to think that the jury would recall the 'fine detail' of articles published years before. Material currently available on the internet would, however, need to be 'handled differently' - that is, by a specific warning. The prosecutor submitted that it was 'nonsense' that the point could be reached where a jury would be constrained by 'the psychologically impossible'. He further submitted that it ran in the applicant's favour that, by contrast with the murders of Ms Patterson and Ms Maher, Miss Halvagis had not been mutilated.

81 At the request of the learned judge, counsel for the applicant formulated the kind of warning which might be given before empanelment if his Honour rejected the stay application. The formulation included identification of the applicant as the murderer of Ms Patterson and Ms Maher, and included a proposed incantation that if any member of the jury panel -

... believes he or she is in possession of any information or opinion about the facts of this case that member of the jury panel should apply to be excused.

82 The prosecutor agreed that a pre-empanelment warning and direction should be given. It would need to say, he submitted, that merely having heard or read something about the applicant did not of itself disqualify that person as a juror. The influence of what had been heard or read must be such that 'in good conscience [the prospective juror] would be unable to render an impartial verdict based solely on the evidence presented in the trial and acting in accordance with [the judge's directions]'.

83 Although the learned judge reserved his decision whether to accede to the stay application, observations which he made before retiring to consider his ruling make it plain, in my opinion, that he had by then formed a provisional view that the application should be refused.

84 The learned judge delivered his ruling on 3 July 2007. He said this:

Essentially the defence admission (sic) is of the ubiquity and pervasiveness of the accused's reputation as a serial killer, is such that no fair trial can now be had. Submissions, both oral and written on behalf of the accused in support of the application, were made by [counsel]. They were excellent submissions and I commend him for them.

No permanent stay on the ground of irremediable prejudice has ever successfully been granted in this court. The power so to stay however exists see R v. Glennon (1992) 137 C.L.R. 592, per Deane, Gaudron and McHugh, JJ, at 623 to 624.

I turn first to the evidentiary considerations. The accused has twice previously been convicted in this court of murder. In August 2000, he was convicted of the murder on 19 April 1999 at Northcote, of Nicole Patterson, and was sentenced to life imprisonment with no minimum term. In August 2006, he was convicted of the murder on 4 October 1997 at Somerton, of Margaret Maher and again sentenced to life imprisonment with no minimum term. Both killings were of vulnerable women by knife attack and were characterised by extreme violence and brutality. The killing of Ms Halvagis was of a vulnerable young woman by knife attack and was characterised by extreme violence and brutality.

The two murder convictions of the accused received wide media publicity. He now stands trial for the murder of Ms Halvagis, and is entitled to a fair trial with the verdict given solely on the evidence led in the trial. That there has been a substantial amount of adverse publicity concerning the accused, is established by evidentiary material filed before me and which is the subject of formal exhibition and to an affidavit filed today. I received the evidentiary material yesterday although I have not yet formally exhibited because it was proper to receive it without impediment and as its prominence appeared on it face.

The accused's convictions for the murder of Ms Patterson and Ms Maher, are also notorious and on that basis, also are taken into account on this application. The evidentiary material falls into four categories, first, newspaper publications, at the times of the murders of Ms Patterson and Ms Maher and of the trials thereto, and in relation to alleged other actions, history or proclivities of the accused. Second, television and radio broadcasting likewise. Third, Internet and other electronic publication, likewise. Fourth, book publication, likewise. Generically I shall refer to the material in these categories as impugned material.

For the purpose of determining this application, I read all the newspaper, book and printed internet material relied upon by the defence and have viewed the ABC television program Australian Story, which went to air on 18 March 1999, concerning the offence charged including having purported re-enactments, and concerning the deceased and her family.

I refer to all the material as cumulative rather than discrete. Further, I take into account the fact of publicity over time, the effect of publicity over time and its percolating and pervasive effect. I also take into account as I have said, as a matter of notoriety, that the accused is known to be a convicted murderer.

The impugned material includes review in graphic detail of the deaths of Ms Patterson, Ms Maher and Ms Halvagis, their circumstances and similarities. It includes detailed review of the criminal and psycho-social history of the accused. It includes characterisation of him as a 'serial killer' and a 'monster'. It includes attribution to him, either as an actor or suspect, of the killing of other women as yet unsolved.

There is no doubt that the impugned material, if acted upon by a jury, would preclude his fair trial on the charge presently preferred against him.

85 His Honour described the criteria for determination of an application for a permanent stay by referring to the reasons for judgment of Mason CJ and Toohey J in The Queen v Glennon where their Honours said -

... a permanent stay will only be ordered in an extreme case (Jago (1989) 168 CLR at 34) and there must be a fundamental defect 'of such a nature that nothing that a trial judge can do in the conduct of the trial can relieve against its unfair consequences'.[33]

86 Then his Honour referred to other authorities, including Murphy v The Queen,[34] R v Long; ex parte Attorney General for the State of Queensland[35] and John Fairfax Publications Pty Ltd v District Court of New South Wales & Ors;[36] and as well to a dictum of Street CJ in Munday v R[37] in which the learned Chief Justice referred to his 'great faith in the multiple wisdom and balance reflected in the verdict of a jury'.

Because I consider that it is notorious that the accused has been convicted of the murders of Ms Patterson and Ms Maher I propose to give the jury directions about those two convictions. Although not every member of the jury will know of those convictions I consider some will and thus I shall give the jury appropriate directions which I shall foreshadow to the jury panel when inviting excuse applications. Depending upon future rulings as to admissibility of identification evidence the fact of those two murders may be introduced in evidence.

Applying the relevant legal criteria to the evidence before me including its cumulation and effect over time and the inferences deriving there from I am entirely unpersuaded that the accused cannot obtain a fair trial by jury on the charge of the murder of Ms Halvagis.

Indeed I am of the positive conclusion that the accused can and will receive a fair trial in its plenary sense. That is because of the inherently fair and robust jury system in this State which system will, in my view, preclude prejudice from the extraneous sources which I have characterised above affecting a deliberation or verdict. I have very responsible confidence that the jury, appropriately directed, will firewall its deliberations and verdict from extraneous considerations and from prejudice in this case.

That is for the following reasons: First, each juror will swear or affirm to give a true verdict according to the evidence. Second, the jury will be directed, with reasons therefore, to give a true verdict according to the evidence. Third, the jury operationally will observe and will inevitably be influenced by the care with which evidence is received and tested during the trial. Fourth, the jury will be assisted in its task by the nature of a jury trial, its methods of testing and of consideration and of analysis, its valuing of care and of scrupulousness and its conscientious commitment to fairness. Fifth, citizens in this community selected to act as jurors show, and historically have shown, a robust capacity and conscientious capacity to act on evidence and to put aside extraneous data and considerations and demonstrate an honourable commitment to fairness.

As to material currently available, particularly on the Internet, that material is easy of access in an electronic age but I am satisfied that empanelled jurors will comply with the direction of law I shall give them not to do their own research, not to have access to the Internet and to have regard only to evidence led in court.

As to the historical material referred to in the data placed before me I am satisfied that much of the detail of such material would not be known to jurors or be capable of recall by them.

None of the panel prospectively will know the case they are summoned for involves this accused.

Applying the criteria articulated by the authorities I have cited to the impugned material relied upon and the relevant considerations deriving from the material, I am satisfied that an empanelled jury will accord the accused a fair trial. (my emphasis)

88 The learned judge did not articulate how it could be that he considered it notorious (and would be known to at least some jurors) that the applicant had been convicted of the murders of Ms Patterson and Ms Maher, and yet that 'much of the detail of the historical material would not be known to jurors or be capable of recall by them'. The applicant's notoriety, after all, was the product of the relentless recapitulation of his prior criminal history, his murder convictions and the intimation that he had killed other women including Miss Halvagis.

89 But however that may be, except in the passage which I have highlighted his Honour said nothing about the way in which important aspects of the Crown case would inevitably call-up the applicant's prior murder convictions and the publicity which attended them; and of the inhibitions which this imposed upon the conduct of the defence case.

90 A jury was then empanelled. But a perceived problem of non-compliance with s 32(1) of the Juries Act 2000 (Vic) arose, and the learned judge reluctantly discharged the jury without verdict. For that reason, I need say nothing about the pre-empanelment warning which his Honour gave on that occasion.

91 A second jury was empanelled on 11 July 2007. Before empanelment, the learned judge instructed the jury as follows:

So let me tell you a bit about categories of application for excuse. The first category is this, and it's a pretty obvious one. If any of you know the accused or knew the deceased or know their families or know anyone connected with the case, you should say 'excuse', and that is because, ladies and gentlemen, for very sensible reasons persons who know anyone connected with the case should not be on the jury, and therefore if any of you know the accused, Mr Peter Norris Dupas, or knew the deceased, Ms Mersina Halvagis, or know their families or anyone connected with the case, when your number is called you should say 'excuse'.

However, ladies and gentlemen, the fact that you have heard something about this case before, in the media, or whatever, does not disqualify you from serving on the jury. We live in a media age where there is often publicity about cases, not particularly this case, but lots of cases, there's lots of publicity about, and the fact that you have seen publicity about a case does not disqualify you from being on the jury. However, if you actually know someone who is going to give evidence, or is connected with the case, that is a different thing, because that is personal knowledge, not just reading something in the media.

Finally ladies and gentlemen, let me tell you a little bit about what is actually required of you if you are selected as a juror, so you can bear this in mind in considering whether you seek to be excused. In a criminal trial, once a jury is selected, the judge and the jury have two functions. I am the trial judge. I decide any technical questions of law which might arise during the case, but I do not decide whether the person has been proved guilty or not guilty. The jury decides that. It is the jury who are the sole judges of the facts of the case, not the judge, not the barristers; the jury are the sole judges of the facts of the case. That is the job prescription, ladies and gentlemen, if you are selected as a juror.

So what is required of you as a juror? What is required is this, you are required to act as you would expect a judge to act, because each juror is a judge, a judge of the facts of the case. You are required to act fairly, calmly, without prejudice and solely on the evidence. That is what is required. You will be given every assistance, ladies and gentlemen, the jurors who are chosen, by the court in that function. You will be given every assistance in the following ways; I will give you any directions of law you need; that will be my job to assist you on that and I will. You will have the benefit of hearing witnesses called in front of you who will give evidence in the witness box, and they will be questioned in front of you, so you have the benefit of hearing the witnesses give their evidence and being tested by questions in court.

The barristers will address you to assist you in your task and at the end of the case I will give you a summary of the facts of the case. That is all designed to assist you ladies and gentlemen in your task as being a juror. You don't have to do it cold ladies and gentlemen, you will get plenty of assistance in being judges and plenty of assistance in being fair, acting fairly, calmly, without prejudice and solely on the evidence.

The next thing is this ladies and gentlemen. The principle you the jury must act upon and every jury in every case acts upon, is that you must give a true verdict according to the evidence. Each of you will swear or affirm, to give a true verdict according to the evidence. The evidence is what you will hear in the four walls of the court room, from the witnesses who are called in front of you and the exhibits which are tendered as part of the evidence. When you retire to consider your verdict, you can have the exhibits with you during your deliberations. That's what you have to decide the case on and solely on the evidence called in front of you. That means ladies and gentlemen, you have to put aside any outside influences, you have to put aside any previous publicity you might have seen in the media or in the newspapers or on television, you have to put aside previous publicity. You have to put aside outside influences, people wanting to talk to you about the case and things like that and decide the case solely on what you hear in court. You might say 'Really, that's pretty hard to do, how can I just segregate my mind from the outside world?' The answer is this: You will be very occupied in court with the evidence, you will find the evidence will compel your attention and you will be attentive to the evidence. You will soon be able to put aside any previous knowledge because you will be concentrating on the case in court, as you each swore or affirmed to do.

The next thing is, again, a jury will be required to act according to the principle of proof in my society in Australia, which is that the person who makes the allegation has to prove it. In this case the prosecution has charged Mr Dupas with murder, the prosecution has to prove it. The prosecution has to prove the charge of murder against Mr Dupas. It's not the other way around, Mr Dupas does not have to prove he is innocent, that is not the law, the law is and fairness and commonsense also are, if you make the charge, you prove it. The law is, and fairness and commonsense are, that the prosecution has to prove the charge beyond reasonable doubt. Mr Dupas does not have to prove anything. That's a principle you will have to act upon ladies and gentlemen and juries act upon it every week in this court, it's an important Australian principle of fairness, ladies and gentlemen.

The next in this job description ladies and gentlemen, is this: when are actually the juror in the case, you mustn't do outside homework, you must just confine yourself to the evidence in court. When you're a juror in the case, you mustn't do your own research, you must confine yourself to the evidence in the case, which means the jurors aren't entitled to look up old newspapers, go down to the library and look up old records, they must not look at the internet, they mustn't look at any previous research but concentrate solely on the evidence in court. When you think about it ladies and gentlemen, that's a very sensible rule, because a lot of things in the media and particularly a lot of things on the internet, are second or third or fourth-hand, half the time it's wrong. You can't tell whether it's right or it's wrong and you can't test it, so it's unreliable material, whereas in court, evidence is called in front of you, it is first-hand, the witness says 'I saw this'; 'I heard that', and it can be tested by questioning. The evidence in court has a quality which the internet and the newspapers never can have because the evidence in court is firsthand and can be tested. The internet and newspapers are not firsthand and you can't test them, so you must not as a juror, look at the internet during the case or look at old newspapers but concentrate on the evidence here in court. When you think about it again for a moment, ladies and gentlemen, it's just commonsense and fairness that you act on evidence in court and not outside materials.

Finally ladies and gentlemen, let me tell you this: as you know, the case in this court is that Mr Dupas is charged with the murder of Ms Halvagis in the Fawkner Cemetery on 1 November 1997. You may well have seen previous publicity about those allegations ladies and gentlemen, the fact that you've seen previous publicity about those allegations does not disqualify you from serving on the jury, you live in the world, if you see things in the media you put them aside. That's the first thing ladies and gentlemen, but you may also have seen allegations about previous publicity about Mr Dupas.

Mr Dupas in fact in the year 2000 was convicted of the murder of Nicole Patterson at Northcote, and in 2004 Mr Dupas was convicted of the murder of Margaret Maher at Somerton. Now that's not what he is here for, ladies and gentlemen. He is here charged with the murder of Ms Halvagis at Fawkner, but he did have those two previous convictions. Now as to those two previous convictions, ladies and gentlemen, which you also may have read about, as a juror the law requires that you must not in your deliberations reason that simply because a person has been convicted previously of a crime or crimes he is the kind of person who would commit this crime now charged, that is the murder of Mersina Halvagis at Fawkner on 1 November 1997.

So, ladies and gentlemen, when the jurors are selected you must not say to yourselves, 'well, Mr Dupas has been convicted twice before of murder and therefore he is the kind of person who would have murdered Ms Halvagis'. That is prohibited reasoning, ladies and gentlemen, and when you think about it for a moment it's simply fairness. It's very fair and very Australian that you don't jump the gun, you do give people a fair trial, and every accused person, including Mr Dupas, and every other accused person is entitled to a fair trial. So that's really a principle of fairness, ladies and gentlemen. It might sound hard at the start to do, but when you think about it it's simply a matter of fairness and you must as a juror act according to it.

Finally, ladies and gentlemen, the court does not ask impossibilities of people. If, despite the encouragement I have given you and despite your best efforts, you feel that as a juror you would be unable completely to put aside any previous publicity and give a true verdict according to the evidence, or if despite the encouragement I've given you and despite your best efforts you feel you would be unable to give the accused a fair trial on the evidence in this court, when your number is called you should say 'excuse' and I'll hear your application to be excused. But as I say, ladies and gentlemen, the fact that you've heard previous publicity about this case or about Mr Dupas does not disqualify you from serving as a juror in this case.

92 After the jury was empanelled, the judge gave this direction:

I am not a judge of the facts of the case. Only you 15 good people are the judges of the facts of the case. I am not a judge of the facts of the case. Counsel, the barristers, are not the judges of the facts of the case. You 15 good people are and you alone.

Because I am the judge of the law, ladies and gentlemen, any legal directions I give you, you must comply with. But I am not a judge of the facts and only you are the judges of the facts. So, as you are the judges of the facts, how do you act as a judge? As I said to you this morning, ladies and gentlemen, in the other court, you act as you would wish and expect a judge to act, fairly, calmly, without prejudice and solely on the evidence. Each of you has sworn or affirmed to give a true verdict according to the evidence. The evidence is what you hear in the four walls of this courtroom from the witnesses who will give their evidence in the witness box and the exhibits which attended as part of the evidence. That is what you judge the case on and that alone. The evidence from the witnesses here in court in front of you and the exhibits tendered as part of the evidence. During the trial, witnesses are called to give evidence and they will be questioned and tested by questions, that's the proper process of the court, ladies and gentlemen. You see that happening and you judge the case on the evidence led here in front of you in court, that's why you sit there and witnesses stand there. Exhibits are tendered, for example videos and things of that sort, and you are entitled to take those exhibits into your jury room when you are deliberating about your verdict at the end of all the evidence. That's what you decide the case on ladies and gentlemen and each of you has sworn or affirmed to decide the case on the evidence and the evidence alone.

A very important thing follows from that and it's this: you must not decide the case on anything outside the court, ...

The next thing is this, and this is very important in this case as in every case. Do not go and do your own homework or do your own research, don't go and look up old newspapers, don't go down to the local library, do not go and look at the internet, do not do any electronic searches about anyone connected with this case, that's very important ladies and gentlemen. You have sworn or affirmed to give a true verdict on the evidence ...

93 In his opening, the prosecutor referred to the identification of the applicant by Laima Burman, Horst Weller and Katica Melnik. As he explained it, to greater or lesser extent each of those witnesses had identified the applicant by seeing a photograph of the applicant in a newspaper or his image on television programs.

94 The prosecutor described the identification of the applicant by the three witnesses as 'important because it places the accused man at the cemetery on the very day.'

95 The evidence of those witnesses had a second significance. If accepted by the jury, it was the starting point of a second string to the prosecution bow: that the applicant had told lies out of consciousness of guilt. One alleged lie was that he had never been to the Fawkner Cemetery and that 'he did not or he was not aware of anyone that was buried there.' There was evidence that he had said such a thing. A second alleged lie was that he had not been to the cemetery and did not have any relatives buried there. Evidence was given that he had said such a thing and it was established that one of his grandfathers was buried there. If the evidence that he had said that he had never been to the cemetery was accepted, and the evidence of one or more of the identification witnesses was accepted, an inference that the applicant had told a deliberate lie was extremely strong. Where such an inference led would be the next question.

96 In the course of the prosecutor's opening, a further connection was exposed between the prosecution case and one of the murders of which the applicant had been convicted. According to the Crown case, the applicant had made relevant admissions to the former solicitor and convicted drugs offender, Andrew Fraser, whilst the two of them were in prison. One aspect of Fraser's evidence was that the applicant had said, in substance, that whilst there was forensic evidence to connect him with the killing of Ms Maher, he had left no forensic evidence at the scene of the killing of Miss Halvagis.

97 Pausing for a moment, three main threads of the prosecution case were (1) the identification of the applicant at Fawkner Cemetery on the day of the killing; (2) his alleged confession to Fraser; and (3) lies told in consciousness of guilt.[38] Each of those threads connected with the applicant's notoriety as a killer of women, and in that connection his effective indictment as the killer of Miss Halvagis.

98 Counsel identified the following issues: (1) the (un)reliability of the identification evidence; (2) the untruthfulness, as counsel made clear he would contend, of the confessional evidence to be given by Fraser; (3) the existence of other candidates as the perpetrator of the killing.

99 As could readily be forecast, important aspects of the evidence highlighted the connection between, on the one hand, the applicant's notoriety as a killer of women and his effective indictment as the killer of Miss Halvagis, and, on the other hand, the proofs in the case.

100 Mrs Burman was, I consider, by far the most substantial of the three identification witnesses. She informed the police of her observations at the Fawkner Cemetery on the morning of the killing very shortly after Miss Halvagis' death was reported. On 15 April 1999 she gave instructions which resulted in the preparation of a computer image of a man's face.[39] On 6 May 1999 she made a photoboard identification of the applicant as the person most likely to be the person whom she had seen at the cemetery.[40] On 16 August 2000, she saw an edition of the Herald Sun.[41] On the front page was a photograph of the applicant. She identified him as the man she had seen in the cemetery.

101 Page one of the Herald Sun of 16 August 2000 reported the conviction of the applicant for murdering Ms Patterson. It did so under the heading 'A Monster'. A sub-heading stated that the applicant was 'also the prime suspect in the Melbourne's graveyard murder mystery'. A photograph depicted the applicant in handcuffs, being restrained by another man. There were also photographs of Ms Patterson and Miss Halvagis, side by side. That is not the entirety of what was said. Later I will refer to other aspects of the page 1 article.

102 Counsel for the applicant was obviously placed in a very difficult situation when cross-examining Mrs Burman about her newspaper identification of the applicant. He must have wanted to suggest that the attendant text had influenced the identification. But the more counsel pursued that line, the more the applicant's past criminal history and his being implicated in the killing upon which he was then standing trial would be put before the jury.

103 The learned judge was alert to the problem. He said this to Mrs Burman in the absence of the jury:

I would ask you to do this, Ms Burman, I know it's not easy but I'd ask you do this, don't say what the newspaper said. Don't say what was said in the newspaper unless you are actually asked to say so by [counsel for the accused]. What was said in the newspaper isn't relevant for this jury's consideration so don't give what the headline was. Don't go into the detail of the other matters that were charged to Mr Dupas. Just answer the actual questions asked by the barristers and don't say what the headline was unless you're actually directly asked to say what it is. I know that's a bit difficult, Ms Burman, but there are a number of things which were in the newspaper which are not relevant for this jury ...

104 Counsel cross-examined to the effect that: (1) the witness, as at 20 August 2000, had been on notice that she might be asked to identify the man whom she had seen at the cemetery; (2) the photograph of the applicant showed him to be handcuffed, under restraint, and obviously under arrest; (3) the sub-heading referred to the applicant being the prime suspect in Miss Halvagis' murder; (4) there was a photograph of Miss Halvagis on the same page of the newspaper; (5) the sub-heading had caused her to think that the applicant might have been the man she had seen; (6) she had not rung the police straightway, rather mentioning her identification on 18 August when a policeman rang her.

105 This cross-examination went as far as counsel could sensibly have gone in suggesting that Mrs Burman had made a flawed identification - a suggestion which she strongly refuted. It avoided the worst excesses of the text being revealed. Even so, it linked a past conviction for murder, and an assertion that the applicant was believed to be Miss Halvagis' killer, with a key issue in the instant trial.

106 Horst Weller was the second identification witness. I consider that his evidence was weak. Its weakness is apparent from what I say about it below, and from other matters to which Nettle and Weinberg JJA refer in their reasons. But that does not touch its present relevance. Mr Weller gave evidence that one day he had seen a picture on a television program of a man named Peter Dupas. He said: 'I straightway realised that's a face I see in the cemetery'. He could not be sure when he had seen the image on the television. More than seven years after the killing, on 22 March 2005, he made a photoboard identification of the applicant, and a statement to the police.

107 Relevant cross-examination of the witness included the following:

Do you agree that you said to the police in your statement 'Since this day I have seen pictures of Peter Dupas on the television, and in the paper. I first saw him on the television one day and I recall the news saying something about killing a girl called Margaret Maher and there could have been another girl, and I also remember them saying he was suspected of the Fawkner Cemetery murder as well'. Did you say that to the police?---Yes.

Was that the truth?---Yes.

Did you say further 'I can't recall how long after the murder I saw this picture of Dupas on the television, at the time of seeing this picture of Dupas I straight away thought to myself 'I know this man's face'; is that right? ---Yes.

'I thought about it for a while and I recalled that this was the same man I saw at the Fawkner Cemetery'; is that so? ---Yes.

Did you say: 'I can't recall how long after seeing his face on the news, that I realised this but I think it was the same day'?---Yes.

But you'd been seeing lots of photographs of Mr Dupas around that time on the television and in the news, hadn't you? ---Yes.

108 The witness further stated that he read the newspaper every day. He accepted that he would have been recognising the applicant on a regular basis in the newspaper (in that connection, a number of newspaper photographs of the applicant had been put to him).

109 In the event, cross-examination necessary to provide one foundation for attacking the witness's identification of the applicant involved disclosure that such identification had followed the witness reading a newspaper article which linked the applicant with the murders of Ms Maher, perhaps another woman, and Miss Halvagis; and this against a background of reading the newspaper - inferentially the Herald Sun - every day.

110 Mrs Katica Melnik was the third identification witness. I consider that her evidence was weak. So much is apparent from what I say about it below, and also from the reasons of Nettle and Weinberg JJA. But again that is not the point of present relevance.

Mrs Melnik, the next day, did you hear on the news that a girl had been killed in the cemetery, is that right?---Yes, the next day, after next day, I'm not sure. I heard anyway.

Some time later did you watch the news on television?---Yes, I did.

Did you see a news broadcast which showed the picture of a man?---Yes.

What was the news broadcast about?---Was the news were that two - two girls were die - two blonde one girls and they showed they pictures and his picture and I recognised him.

Did the television say the name of the picture of the man? ---I think so, yes, they - yes, I think so, but I just recognise - not straight away, I was questioning myself, maybe (indistinct) but I have seen somewhere that man.

When you saw the television, what was your reaction to the picture of the man on the television?---It was my reaction I have seen somewhere and seen somewhere and I talk to my husband and he said just 'don't be stupid'.

Did you think about the man that you saw on the television? ---Yes.

At some stage, did you remember where you'd seen him?---Yes, cemetery, yes.

112 Cross-examined, the witness agreed that she had made a police statement on 29 April 2005 - that is, about seven and a half years after seeing a man in the Fawkner Cemetery on the afternoon of the killing.

Do you know whether it was the television at night or during the day?---The evening. Six o'clock.

You don't know what month it was?---No, I don't.

You don't know what year it was, is that right?---No. It was maybe quite long after when I - was quite long, I don't - when I saw him on television with those girls.

114 Once again, the challenge to this witness drew on the fact that her identification of the applicant, reported to the police long after the murder, was made by seeing an image on the television which accompanied a presentation - this is the inevitable import of Mrs Melnik's evidence - that described the applicant as having murdered two other women.

115 Fraser's evidence of the applicant's gaol confession had a number of components. Fraser did not mention each of them to the authorities when first interviewed. One thing which he did mention at the outset was the applicant's alleged 'no forensics' statement, first made (so it was said) in the context of discussion about the presence of forensic evidence against the applicant in the Maher matter; and, according to Fraser, repeated on two other occasions.

116 It turned out to be the fact there was no forensic evidence linking the applicant to Miss Halvagis' murder. That was capable of giving the alleged confession some credibility[42] - although the mere elapse of time between 1997 and the time when the applicant was said to have first made the critical statement to Fraser, without the applicant being charged,[43] strongly suggested an absence of forensic evidence; and although, had it turned out that there was forensic evidence, the prosecution could have contended with some force that the admission was no less credible, the applicant (despite his familiarity with the criminal law) being unaware of the reach of forensic investigation.

117 Be that as may, the evidence linked the applicant as the murderer of Ms Maher and Miss Halvagis; and implied, the killing of Miss Halvagis being later in time, that the applicant had learned from an earlier 'mistake'.

118 Fraser was vigorously cross-examined, and at length. His credit was impugned by reference to what were relatively easy targets. He made, in my view, a poor fist of dealing with the attacks on his credit.[44] It was put to him, against that background, that his evidence of the applicant's gaol confessions was a litany of lies.

119 Counsel did not mount a substantial attack on the 'no forensics' evidence. He did, however, shortly put it to Fraser that the particular evidence was a lie - no doubt to avoid any Browne v Dunn[45] criticism.

120 Counsel was placed in a difficult position. The particular evidence was potentially dangerous to the applicant's case. For that reason, substantial cross-examination might have been anticipated. But any such cross-examination ran the risk of reinforcing in the jury's mind the likelihood that the convicted murderer of Ms Maher and the charged killer of Miss Halvagis were one and the same man.

121 Finally with respect to the evidence, I note that the applicant stood mute.

122 The prosecutor made a very attractive final address in which he accumulated pieces of 'bad luck' for the applicant, if he chanced not to be the murderer. He referred, as was inevitable, to the circumstances in which the identification witnesses had identified the applicant. He submitted that the 'no forensics' statement attributed by Fraser to the applicant was 'a very clear indicator that Fraser is telling the truth'; and that Fraser's evidence was credible because of its complexity. He relied upon the lies which he contended had been told by the applicant, one of which depended upon acceptance of one or more of the identification witnesses.

123 Very early in his final address, counsel for the applicant made a plea for impartial, unprejudiced consideration of the evidence. This is what he said:

Now His Honour told you that Mr Dupas has two convictions for murder and he's told you that you must not reason that my client is the kind of person who would have murdered Ms Halvagis. In other words, you must not allow yourselves to become prejudiced against him. Well that's a tall order. It would be very difficult to imagine any of you inviting my client to dinner knowing that he had two prior convictions for murder but you must put prejudice out of your mind.

You must lift yourself to that task. No-one for the defence asks you to like Mr Dupas or to endorse his past but you must put prejudice out of your mind. Now I've been a criminal barrister for nearly 30 years and although my experience may be different to others I've never heard a trial judge tell a jury at the outset of the trial that the accused has two prior convictions for murder.

Normally a jury trial is a trial in ignorance of whatever prior convictions a person may have lest there be some who might think well if he did that he's the kind of person that would have done this. Now you have to then ask yourself why were you given this information by the trial judge before you were even selected? Well it was done because it would be stupid for the judge to assume that any of you or all of you were unaware of Ms Dupas and his prior history.

It would be stupid to pretend that none of you knew something about it or heard something about it in the past not because you're dishonest but just because in this State it would be unlikely to find a jury completely bereft of knowledge of Mr Dupas. So in speaking openly to you about these matters His Honour has trusted you and trusted you like no other jury to my knowledge has ever been trusted before.

He's confident that you will follow his directions and you must. You must bring in your verdict according to the evidence. You must suppress those natural feelings of sympathy and prejudice.

It might be easy to say in this case, well, he's nothing but a double murderer, let's convict him. What harm is there in that, even if I get it wrong, let's get on with it. Well or course, there'd be three things wrong with adopting that course. The first thing, it would be a breach of your oath, because you swore in effect, to bring in a verdict in accordance with the evidence you've seen and heard in this case. The second thing that would be wrong, it would be unjust or unfair to the accused person, and thirdly, it would not be fair to the community, not be fair to Mersina Halvagis and her friends and family.

124 Concerning the identification witnesses, counsel submitted that they were honest but that their evidence was unreliable. He identified what he asserted were inconsistencies between their descriptions of the man whom they had seen and the applicant. He referred to the problem called 'displacement'. He made this submission:

In this case, it seems that there was a picture of a man in handcuffs, under restraint by a prison officer in the newspaper, or on television, and surrounding that is newspaper commentary that this person has been convicted of one or two murders, and that he's strongly suspected of the murder of Ms Halvagis. If the police showed a person one photograph and said all that to him, you might wonder how fair that was and how the witness might well be induced to accept that as the face they saw, because after all, this is a terrible crime, and another characteristic of decent people is they want to help the police find the culprit.

125 He said, referring to the Herald Sun of 16 August 2000 -

First of all the headline says, 'Revealed now, killer convicted yesterday is also the prime suspect in Melbourne's graveyard murder mystery' and there, of course is Mr Dupas looking very similar to the picture in the photo folder and right beside him is a picture of Mersina Halvagis and if I zero right in on the small writing beside that - first of all, you see Mr Dupas is being held by what looks like a security man and that photo was unfortunately cropped but in the version of it produced by the Crown, he's got handcuffs on.

126 Concerning the evidence of Mrs Melnik, counsel made this submission:

... the Crown says you may conclude that she's obviously talking about some news item involving Mr Dupas and I concede that seems most likely but it would be very helpful for you to see the photograph that she recognised. None has been produced. Which one is it? But in any event, if it is a photograph of Mr Dupas, it's got the overriding prejudice that there were two girls who died in connection with it and I don't know whether there was also a commentary about whether or not Mr Dupas was suspected of the cemetery killing however she did say she had also seen him in the papers on the front page of the Sunday Herald Sun.

127 Counsel made a forthright attack on Fraser's credibility. He directed attention to what he claimed were improbabilities in Fraser's account, and inconsistencies between that account and other evidence. He said nothing about the 'no forensics' evidence, save to submit that the Crown had 'got it around the wrong way' in seeking to use the fact that there was no forensic evidence to buttress Fraser's credit.

128 The learned trial judge gave the jury a 'decide the case on the evidence' direction, which he tailored to fit the particular circumstances by saying -

Equally in this case_, you have been told that in the year 2000, Mr Dupas was convicted of the murder of Nicole Patterson at Northcote and in the year 2004, Mr Dupas was convicted of the murder of Maragaret [sic] Maher at Somerton. You have been told about those matters, ladies and gentlemen, as counsel has said to you because it is likely that some of you would have known that anyway just from general knowledge and it has been dealt with up front by you being told and also the Maher matter is referred to anyway in Mr Fraser's evidence. So you have been told of those matters, ladies and gentlemen. But you must not decide this case on other cases in the past._ You have only been told about them for the reasons I have just said but you must decide this case on the evidence in this case in relation to the accused and Ms Halvagis and her death at the Fawkner Cemetery on 1 November 1997.

The law requires, and this is a direction of law which binds you, ladies and gentlemen, that you must not in your deliberations reason that because a person has been previously convicted of a crime or crimes he is the kind of person who would commit the crime now charged, that is the murder of Mersina Halvagis at Fawkner on 1 November 1997.

The law requires that you must not reason that way and I direct you that you must not and anyway that is just fairness, ladies and gentlemen. You decide this case on the evidence relating to Ms Halvagis' death and the evidence relating to it, not prior cases and I am sure you understand that, ladies and gentlemen. I am quite sure you will act on it accordingly as I have directed you and as fairness will and I say no more about it, ladies and gentlemen. (emphasis added).

129 That direction very considerably understated the connection between the evidence in the case and the applicant's earlier convictions for murder, at least in that it said nothing about the publicity which accompanied those convictions and its relationship to the evidence of the identification witnesses, and thus to proof of one of the lies upon which the Crown relied.

130 The learned judge, with respect to the identification evidence, directed the jury that it should consider the question whether 'there were any interfering factors in the circumstances of recall; what technically is called displacement.' He did not identify any of the circumstances which might have suggested the operation of displacement in the case of the three witnesses. He did, later, faithfully recount the evidence of those witnesses, and the arguments which had been advanced for and against the reliability of such evidence.

131 Concerning Fraser's evidence, the judge gave a Pollitt[46] direction which is the subject of ground 7 of the Notice of Appeal. Nettle JA deals with that ground in his reasons, and, as I have said earlier, I respectfully agree with his conclusion. For present purposes it is enough to say that, when giving that direction, his Honour referred to Fraser's 'no forensics' evidence and 'pantomime' evidence, and in respect of the former made it clear to the jury that there was a dispute whether the applicant had made any such statement to Fraser.

132 Later, recounting Fraser's evidence, his Honour reiterated that it was disputed that the applicant had ever said, in substance, that he had left no forensics at the cemetery. Later still, he referred to the evidence of a police witness that there were 'no forensics', but that this fact, so far as the witness knew, had not been put into the public arena at any relevant time.

133 Finally in respect of Fraser's evidence, his Honour instructed the jury that there was no independent confirmation in the evidence of what the witness said the applicant had told him. Thus his Honour did not pursue his earlier analysis - which was to the effect that the fact that there were 'no forensics' was evidence capable of corroborating the alleged confession.

134 I turn to his Honour's summary of counsel's final addresses, only to notice that, concerning displacement, he said this -

What has happened, [defence Counsel] put, is what is likely to have happened here as these honest witnesses have got other material in their heads which has displaced the image that they originally had of the man in the cemetery and they have put the wrong two things together and he took you through all of that, ladies and gentlemen.

then summarising a number of points made by counsel in that connection.

135 The jury retired at 11.27am on 8 August 2007. It returned its verdict at 3.44pm on the following day.

136 It remains to mention that in his sentencing remarks the learned trial judge stated, concerning the stay application, that 'the jury, by its care and conscientious application to its task, and its intelligent address as demonstrated by its questions during deliberation, undoubtedly gave the accused a fair trial.' There was in fact no question. Rather, the jury asked to see a videotape of certain cross-examination of Fraser. It concerned the contact which he admittedly had with a reporter from the Herald Sun. That newspaper published a substantial article on the day after Fraser's release. It reported, in effect, that the applicant had made a confession to Fraser which was a lot stronger than an ordinary gaolhouse confession. In the evidence which the jury enquired after, Fraser admitted that, when questioned about his contact with the reporter in the course of a Basha enquiry, he had made a number of what he characterised as 'mistakes' and 'oversights'.

137 There are a considerable number of appellate decisions arising out of the refusal of a trial judge to stay a criminal proceeding on the ground that a fair trial could not be had either because of intense pre-trial publicity concerning the circumstances of that proceeding, or the revelation of the accused's prior convictions, or both. In Australia, they include The Queen v Glennon,[47] R v Glennon (No 2),[48] R v Dudko,[49] R v Long, ex parte Attorney General for Qld,[50] R v D'Arcy,[51] R v Ferguson; ex parte Attorney General (Qld),[52] R v Skaf[53] and R v McGee.[54] In England, decisions include R v McCann,[55] R v Taylor,[56] R v West[57] and R v Abu Hamza.[58]

138 There are also a number of decisions upon related questions; what I will call a second kind of case -

• In Murphy v The Queen,[59] one question, in a case which involved the notorious killing of Anita Cobby, was whether the trial judge had erred by refusing to adjourn the trial for a six month period.

• In Gilbert v The Queen,[60] the question was whether a misdirection had been of no effect because it could be assumed that the jury had not adhered to directions given by the trial judge.

• In John Fairfax Publications Pty Ltd and Anor v District Court of New South Wales & Ors,[61] one question was whether it had been a necessity to order that details of a man's conviction not be published because the man was to face further changes.

• In R v Thomas (No 3),[62] one question was whether, an appeal having been allowed, acquittal ought be ordered because the applicant would be unable to have a fair re-trial.

• In R v Vjestica,[63] one question was whether failure to order a change of venue had caused a miscarriage of justice.

• In General Television Corporation Pty Ltd v Director of Public Prosecutions (Vic) & Anor,[64] a question arose whether an injunction restraining the showing of the television series 'Underbelly' whilst a trial was proceeding - there being a commonality between characters in the series and persons involved in the trial - should be dissolved.

139 At a distance, but related in concept to the first kind of cases, are decisions dealing with other circumstances in which the propriety of an order staying or refusing to stay a criminal proceeding has arisen for consideration: Jago v The District Court of New South Wales & Ors,[65] Dietrich v The Queen,[66] Walton v Gardiner[67] and The Queen v Edwards & Anor.[68]

140 For present purposes, the first, and to a lesser extent the second, kind of cases are of particular relevance. Amongst the former, The Queen v Glennon is central.

141 In that case the accused man, who had a prior conviction for indecent assault of a young woman, faced charges of sexually assaulting young men. Before his trial, a radio presenter made serious allegations of criminal conduct and sexual impropriety against the accused. The previous conviction was mentioned. The presenter was convicted of contempt of court and ordered to be imprisoned. His conviction was upheld.[69] The proceeding for contempt and the appeals which followed conviction were apt to keep the content of the presenter's allegations before the public. Trial of the accused began some years after the allegations had first been raised; and, indeed, after the presenter had been released from gaol. The way in which the matter found its way to the High Court was unusual. A County Court judge refused a stay application. A proceeding was then brought by originating motion challenging the refusal. Crockett J was not persuaded that the trial of the accused would be unfair. The accused was then presented before another County Court judge. A still further stay application was made and refused, the judge holding himself bound to follow the conclusion of Crockett J. The trial proceeded and the accused was found guilty and convicted. There was an appeal to the Court of Criminal Appeal. It was allowed, and the court ordered that the proceeding be permanently stayed. From that decision, the Crown successfully appealed to the High Court.

142 Mason CJ and Toohey J said this in the context of deciding whether the Crown should have leave to appeal:

Apart from the unique case of Tuckiar v. The King, there has been no other instance in the judicial history of this country of an accused's conviction being quashed and a verdict of acquittal then entered on account of the potential prejudicial effect of pre-trial publicity. The decision of the Court of Criminal Appeal is all the more remarkable in that it rejects the discretionary judgment of Crockett J. refusing a stay in circumstances where that discretionary judgment appears to disclose no error of principle.[70]

As for public importance, there can be no doubt that the decision of the Court of Criminal Appeal has far-reaching consequences for the administration of justice. In contemporary society, where sensational media publicity presents very serious problems in ensuring that persons accused of criminal offences receive a fair trial, it is a matter of the utmost significance to determine whether the Court of Criminal Appeal was right, in the circumstances already outlined, in granting a permanent stay and thus a continuing immunity from prosecution. The importance of what is at stake in the present case was forcefully expressed by Southwell J. in the Court of Criminal Appeal:

'And so it is that this court is being invited, in effect, to create legal history by finding that in a large city, adverse media publicity must be held to have had the result that a person charged with serious offences will never be called upon to face trial. This, to my mind, is an invitation which ought not here be accepted.'[71]

Dealing with the merits of the appeal, their Honours said:

Because the Court of Criminal Appeal appears to have thought that Judge Neesham was practically, though not legally, bound to follow Crockett J.'s decision, the focus of the Court's attention was directed to that decision. It was common ground that the respondent's appeal could not succeed unless it were shown that Crockett J's decision was erroneous in accordance with the established principles governing appeals from discretionary judgments. Accordingly, it was for the respondent to show that Crockett J. acted upon a wrong principle, took into account some extraneous consideration, failed to take into account a relevant consideration or mistook the facts (24). If convinced of such an error but not otherwise, the Court of Criminal Appeal was entitled to set aside the decision of Crockett J. and exercise its own discretion.[72]

143 Although the issue thus framed had reference to House v The King,[73] their Honours made it apparent, I consider, that the circumstances of the trial would not necessarily be irrelevant to resolution of the issue. But as it turned out, the majority judges in the Court of Criminal Appeal, in concluding that the verdicts were unsafe and unsatisfactory, had not relied upon the way in which the trial had been conducted.

144 Concerning the possibility of a juror acquiring 'irrelevant and prejudicial information', which their Honours said was 'inherent in a criminal trial', Mason CJ and Toohey J said this:

The law acknowledges the existence of that possibility but proceeds on the footing that the jury, acting in conformity with the instructions given to them by the trial judge, will render a true verdict in accordance with the evidence. As Toohey J. observed in Hinch, in the past too little weight may have been given to the capacity of jurors to assess critically what they see and hear and their ability to reach their decisions by reference to the evidence before them. In Murphy v. The Queen, we stated:

'But it is misleading to think that, because a juror has heard something of the circumstances giving rise to the trial, the accused has lost the opportunity of an indifferent jury. The matter was put this way by the Ontario Court of Appeal in Reg. v. Hubbert: 'In this era of rapid dissemination of news by the various media, it would be naïve to think that in the case of a crime involving considerable notoriety, it would be possible to select twelve jurors who had not heard anything about the case. Prior information about a case, and even the holding of a tentative opinion about it, does not make partial a juror sworn to render a true verdict according to the evidence.'

To conclude otherwise is to underrate the integrity of the system of trial by jury and the effect on the jury of the instructions given by the trial judge.[74]

In our view, the approach adopted by the majority in the Court of Criminal Appeal, in finding that there was a substantial risk that the jury were aware of the applicant's prior conviction and that the verdict was unsafe and unsatisfactory on that score, disregarded the principled decision of Crockett J. The majority's reasoning was also based on materials which could not support the inferences drawn, took little, if any, account of the effect of the trial judge's instructions and disregarded the community's right to expect that a person accused of a serious criminal offence will be brought to trial. The last-mentioned consideration requires that a conviction be quashed as unsafe and unsatisfactory only where the materials justify a conclusion that there was a likelihood or substantial risk of prejudice arising from pre-trial publication where the conviction is challenged on that ground.[75]

... a permanent stay will only be ordered in an extreme case and there must be a fundamental defect 'of such a nature that nothing that a trial judge can do in the conduct of the trial can relieve against its unfair consequences'.[76]

145 In a separate judgment, Brennan J (with whom Dawson J agreed) observed that the Court of Criminal appeal 'would have had no jurisdiction to interfere with the verdicts of guilty unless it had found that a miscarriage of justice had actually occurred ...'[77]

146 His Honour eloquently stated the problem raised by pre-trial publicity prejudicial to an accused:

The problem is not new, but perhaps it has become more acute in recent times. One developing phenomenon is the holding of press conferences or the issuing of press releases by some law enforcement agencies after a person has been charged with a criminal offence, in apparent disregard of the risk that the fair trial of the person may be prejudiced unless the matter published is restricted to what this court in Packer v Peacock described as the 'bare facts' - by which the court meant '(but not as an exclusive definition) extrinsic ascertained facts to which any eyewitness could bear testimony, such as the finding of a body and its condition, the place in which it is found, the persons by whom it was found, the arrest of a person accused, and so on'. Sometimes the holding of a press conference or the issuing of a press release wears the appearance of corporate advertising of the work of the agency in solving a crime. Advertising of that kind is inconsistent with the impartial performance of the functions of a law enforcement agency in conducting or assisting to conduct a criminal prosecution. Another phenomenon which has contributed to the problem in recent years, especially in the media of television and radio, is the promotion of personalities who affect to convey the moral conscience of the community and to possess information, insights and expertise in exceptional measure. The image of some media personalities as informers of the public and moulders of public opinion is assiduously cultivated. When the belief is held that the public interest is served by publication of observations and opinions by media personalities on topics of contemporary relevance, publication may not always be restrained by the need to allow a fair trial for a person charged with the commission of crimes that have attracted public attention.[78]

147 But he also stated that even pre-trial publicity amounting to contempt of court did not mean that a trial must be aborted:

Administration of the criminal law cannot be made hostage to conduct amounting to contempt of court, even if the contempt be flagrant. If it were otherwise, the perpetrators of crimes which shock the public conscience, such as those charged in Murphy v R, would oftentimes go untried and unpunished, for pre-trial publicity prejudicial to an accused is stimulated by the notoriety of the accused and the heinousness of the crime. Yet it would undermine the criminal law's protection of society and its members to refuse to allow the law to take its ordinary course in these cases. The administration of criminal justice by the courts, which proceeds inexorably to its conclusion in each case, would be adventitious if trials could be halted by a punishable contempt.[79]

The law does what it can to protect the integrity of the criminal trial. In the forefront is the law relating to criminal contempt. If the protection given by that law should fail, the trial judge is given powers to adjourn the trial until the influence of prejudicial publicity subsides and is required to direct the jury that their verdict must be based on the evidence given before them on the trial and that, in reaching their verdict, they must disregard knowledge otherwise acquired and any revulsion against or sympathy for the accused. The trial judge may conduct the trial in whatever manner is appropriate (within the ordinary procedural constraints) to counter the effect of pre-trial publicity prejudicial to an accused. However, these protective mechanisms cannot guarantee perfect impartiality, as Mason CJ and Toohey J recognised in Murphy v R:

It may be said that there can be no guarantee that directions given by a trial judge in an effort to counter the effect upon a jury of media publicity will be successful. That is true just as it is true that there can be no guarantee that a juror may not have been influenced by other matters of which he or she has heard before the trial.

Of necessity, the law must place much reliance on the integrity and sense of duty of the jurors. The experience of the courts is that the reliance is not misplaced. In Munday v R, Street CJ repeated an unreported passage from one of his Honour's earlier judgments:

'... it is relevant to note that the system of jury trial is geared to enable juries to be assisted in every possible way to put out of mind statements made outside the court, whether in the media or elsewhere. There is every reason to have confidence in the capacity of juries to do this. Judges do not have a monopoly on the ability to adjudicate fairly and impartially. Every Australian worthy of citizenship can be relied upon to discharge properly and responsibly his duty as a juror. Particularly is this so in the context of being one of a number or group of others all similarly charged with this responsible duty. I have great faith in the multiple wisdom and balance reflected in the verdict of a jury.'

If the courts were not able to place reliance on the integrity and sense of duty of jurors, not only would notorious criminals or heinous crimes be beyond the reach of criminal justice but there would have to be a change in venue for many trials now held in circuit cities or towns where knowledge of the crime and of the alleged criminal easily acquires a wide currency outside the courtroom. Our system of protecting jurors from external influences may not be perfect, but a trial conducted with all the safeguards that the court can provide is a trial according to law and there is no miscarriage of justice in a conviction after such a trial.[80]

149 His Honour concluded that Crockett J had been correct not to stay the trial:

... either on the ground that the present case is not an 'extreme case' or on the ground - which, in my respectful opinion, is a ground better founded on principle and more realistic in practice - that the trial of the applicant, provided it was as fair as the court could make it, would produce no miscarriage of justice.[81]

150 Deane, Gaudron and McHugh JJ dissented in the result. It is desirable to note, however, two portions in their joint judgment. The first is their observation that:

... the question before the Court of Criminal Appeal fell to be answered in the context of knowledge of what had occurred at the trial and was, strictly speaking, a different question from that which had arisen before Crockett J.[82]

The central prescript of our criminal law that no person shall be convicted of a crime otherwise than after a fair trial according to law dictates that an accused is entitled to be protected from an unacceptable and significant risk that the effect of prejudicial pre-trial publicity will preclude a fair trial. Ordinarily, that risk will be obviated by appropriate and thorough directions and, if the circumstances also require it, a temporary stay for the minimum period adjudged necessary for the pre-trial publicity to abate. The balancing of the legitimate interests of the accused and the prosecution will, in almost every case, mean that if the proceedings are to be stayed at all, they should only be stayed temporarily and for the minimum period necessary. Nonetheless, one cannot exclude, as a matter of law, the possibility that an "extreme" or "singular" case might arise in which the effect of a sustained media campaign of vilification and prejudgment is such that, notwithstanding lapse of time and careful and thorough directions of a trial judge, any conviction would be unsafe and unsatisfactory by reason of a significant and unacceptable likelihood that it would be vitiated by impermissible prejudice and prejudgment. In such a case, a permanent stay may be granted.[83]

151 Two threads are discernible in Glennon. The first concerns the nature of the appeal to the Court of Criminal Appeal. The various judgments affirmed that Glennon must have shown a miscarriage of justice if he was to succeed in his appeal to the Court of Criminal Appeal. That is, now speaking generally, it is not enough for a court to be satisfied that the jury was likely to have embarked upon a trial imbued with prejudice against the accused. The judgments also show that the issue was not restricted to the question whether the exercise of discretion, considered at the time when it was made, was capable of being impugned. Rather, consideration was able to be given to the trial as it had unfolded.

152 The way in which the necessary approach has been worked out, in this State and in other States with similar appeal provisions, is exemplified by R v Glennon (No 2),[84] R v Long; ex parte Attorney General for Queensland,[85] R v D'Arcy,[86] R v Dudko[87] and the change of venue case of R v Vjestica.[88]

The question which, strictly speaking, was raised for the consideration ... was whether a miscarriage of justice had occurred, and the convictions were unsafe, as a consequence of irreversible prejudice due to sustained pre-trial adverse media publicity.[89]

154 In Vjestica, Maxwell P, with whom Buchanan JA and Whelan AJA agreed, observed that -

While the pre-trial ruling on change of venue is an exercise in risk assessment, the post-conviction appeal against a refusal to order a change of venue is a hindsight review. The appeal court must decide whether a miscarriage of justice occurred as a result of the refusal to order a change of venue.[90]

155 In New South Wales, the enquiry has followed a similar path.[91]

156 A somewhat differently structured enquiry has been undertaken in Queensland. Whilst recognising that in the end the question is whether a miscarriage of justice has been demonstrated, consideration has been given successively to examination of the exercise of the discretion for House v The King error, and then to whether the circumstances demonstrate a miscarriage.[92]

157 In all, it appears that if an exercise of the discretion was shown to be infected by House v The King error, it would carry an applicant a long way in showing that a miscarriage of justice occurred - although the presence of the one might not mandate a conclusion that the other was present. On the other hand, absence of demonstrated error in the exercise of the discretion might not mean, as the trial unfolded, that the subject matter of the exercise of discretion had not caused a miscarriage. Yet again, if the exercise of the discretion was not shown to have been infected by error - this necessarily focusing upon the time of its exercise - it might tend to show, notwithstanding the different temporal viewpoints - the improbability that refusal to stay the proceeding had in fact brought about a miscarriage.

158 The second thread which emerges from Glennon is the very high hurdle faced by an applicant who wishes to argue that the circumstances of his or her case are so extreme as to attract a conclusion that a miscarriage occurred which should result in the remedy of a permanent stay; and the reasons why that must be so. But it cannot be said, I consider, that there was consensus as to the height of the hurdle. The approach preferred by Brennan J, with whom Dawson J agreed, accepted that it would be enough if a trial was as fair as a court could make it. Mason CJ and Toohey J did not in terms go so far. It was implicit in their conclusion that there may be very adverse pre-trial publicity, and yet that refusal of a stay will - except in an extreme case - be justifiable, that a fair trial can encompass something less than a trial free of all prejudicial material.[93] Even so, their statement that a permanent stay will only be ordered in an extreme case admits of a possibility that the preferred approach of Brennan J would seem to deny. Then one goes to the dissentients, Deane, Gaudron and McHugh JJ. They stated that one could not exclude the 'extreme' or 'singular' case in which the effect of a sustained media campaign of vilification and prejudgment was such that, for reasons which they explained, any conviction would be unsafe and unsatisfactory and a permanent stay might be granted.

159 Mason CJ and Toohey J referred in Glennon to the 'unique case of Tuckiar v The King'.[94] In that case, counsel for the prisoner, after conviction, made a public statement in court that confessional evidence admitted against his client had been correct. There were deficiencies in the conduct of the trial which would ordinarily have led to an order for a retrial. But wide publicity given to counsel's statement, the High Court agreed, made a fair retrial impossible; and an acquittal was ordered.[95]

160 Tuckiar, then, is not one in which pre-trial publicity yielded a stay on prosecution; but it did preclude a re-trial.

161 Given the different paths of reasoning taken in Glennon, it is not possible to extract a binding ratio decidendi[96] concerning what I have called the second thread in that decision. As the High Court has interpreted the doctrine of stare decisis, it is impermissible to add the dissenting opinion of Deane, Gaudron and McHugh JJ to the obiter observations of Mason CJ and Toohey J in order to secure a majority in favour of the view that a permanent stay may be granted where the intensity of pre-trial publicity is such that, no matter what steps are taken within the bounds of ordinary procedural constraints, a conviction is likely to be substantially affected by prejudice and prejudgment. The most that one can say about the reasoning in Glennon, perhaps, is that 'the balance of persuasion'[97] lay in favour of that view, (although strictly it was obiter), and that the balance of persuasion in Glennon has been followed in subsequent cases.

162 In R v Long; ex parte Attorney General for Queensland,[98] Jerrard JA summarised pertinent considerations, in a passage with which I respectfully agree, this way:

Those decisions establish that considerations which are relevant on an application for a stay based on pre trial publicity include:

1. The extent and nature of the publicity, when it occurred, and the nature of the offence charged.

1. The legitimate public interest, and legitimate private interest(s) of a person charged with a crime, the witnesses, the victim of the alleged crime and their relatives, in the ordinary and expeditious process of prosecution to verdict of those charges.
2. That in this era of intense commercial publication of information about immediately current events, and easy electronic access to that, there can be no guarantee an individual juror may not have been influenced by pre trial publicity.
3. That recognition of that possibility requires judges to do what can be done to protect the integrity of the criminal process, including but not limited to punishment for contempt, adjourning a trial until the influence of prejudicial publicity subsides, ordering a change of venue for the hearing of the trial, ordering separate trials for different accused persons, and giving express directions to jurors that their verdict must be based on the evidence given before them on the trial and that in reaching that verdict they must disregard knowledge otherwise acquired.
4. That of necessity the law places much reliance on the integrity and sense of duty of jurors to comply with such directions and give a verdict based on the evidence led. Accordingly, it is necessary to show more than the possibility that a juror or jurors would have gained knowledge of prior convictions to support the argument that it was likely those jurors would or did ignore or disobey directions given.

6. That the necessary assumption that jurors understand and follow directions given by trial judges can give way to recognition that jurors' decision making is affected by matters of possible prejudice, where more is shown than the mere possibility a juror would have gained knowledge of inadmissible and prejudicial matters. It is in these cases that the discretionary exercise of the powers of the trial judge is critical, including the power of adjournment for a lengthy period.

1. That a permanent stay will be ordered only in an extreme case where there has been adverse pre trial publicity of such a nature that nothing a trial judge can do in the conduct of the trial could relieve against its unfair consequences. The need to maintain public confidence in the administration of justice, and the public interest in ensuring that the judicial processes are not abused and that trials are fair to the people charged, means that a permanent stay should be ordered when it is impossible to ensure that a fair trial could take place.
2. That the fact that adverse publicity is deliberately generated by those for whom the Crown should properly be held responsible may have the result that justice requires a permanent stay be granted.[99]

163 It remains, however, that there has not yet been a case in Australia where an appellate court has held that the refusal to order a stay, in the events which occurred, was productive of a miscarriage of justice. That is so, although, as Spigelman CJ has remarked,[100] applications have been made (and failed) 'in the most sensational of cases'.

164 The conceptual position in the United Kingdom is essentially the same.[101] But in that country, R v Taylor[102] stands as an instance where the Court of Appeal concluded that extreme pre-trial publicity was one reason why a guilty verdict was unsafe; whilst in R v McCann[103] the Court of Appeal held that a stay should have been granted because of things prejudicial to the accused man said in radio or television broadcasts during the trial itself.

165 In many decisions, in connection with what I earlier called the first and second kinds of case,[104] statements have been made that (1) the criminal law, at an appellate level, proceeds upon an assumption that jurors will understand directions given to them, and follow such instructions consonantly with their oath; and (2) it is the experience of judges that this is what jurors do in fact.[105]

166 One may accept those assumptions unreservedly, as I do, without discounting the prospect - considered in the cold light of the particular circumstances of a particular matter - that the jury's task was an impossible one. In Taylor, it was held that such a situation had arisen; likewise in McCann. The order made in Tuckiar proceeded from a conclusion that a fair re-trial could not be had. In General Television Corporation, this Court refused to interfere with the conclusion of the trial judge that a fair trial would be impossible if the 'Underbelly' program was permitted to be televised at the same time as the trial was proceeding.[106] See also observations as to the difficulties for judges - let alone jurors - which are posed by press agitation: Pennekamp v State of Florida,[107] Attorney General v BBC[108] and the extra-judicial remarks of Justice Cardozo in his 1921 address, 'The Subconscious Element in the Judicial Process'.[109]

167 I have kept steadily in mind the fact that the statement of principle which admits of the possibility that pre-trial publicity (in which I include publicity surrounding prior convictions of an accused) may be such that a miscarriage of justice will ensue if the prosecution is not stayed, or if a conviction is not set aside and the further prosecution stayed, has not been matched, in Australia, by the application of that principle. Tuckiar, sometimes cited in that connection, was not in fact such a case, although the consequence - in that there was no re-trial - was the same.

168 I have likewise reminded myself that only once, so far as I am aware, has the principle been applied in the United Kingdom.[110]

169 When a principle is said to exist, but no instance of its application can be found, the cynic might doubt its existence. But the balance of persuasion in Glennon is that it does exist, although it will only have application in the most extreme circumstances.[111]

170 Such a restricted application is consistent with, on the one hand, the very great importance that an accused have a fair trial (or at least as fair a trial as the circumstances permit); and, on the other hand, the heavy emphasis that has repeatedly been placed upon: (1) the presumed resilience of juries, and faith that jurors will act according to their oath; (2) the related assumption that jurors will abide the directions of the trial judge; (3) the great desirability that no accused person should be beyond the reach of the trial process. Were it otherwise, the more the abhorrent the crime, and the more the publicity attending it, the greater the prospect that a stay should be granted.

171 There are ways, in many cases, of adequately dealing with prejudicial pre-trial publicity, or the revelation of prior convictions, other than by permanently staying a criminal proceeding. Thus, first, pre-trial publicity may bring about a change of venue. That is common enough. But it is a far cry from an order permanently staying proceedings.

172 Second, pre-trial publicity from time to time leads to adjournment of proceedings for some period of time. Again, that is a far cry from a permanent stay.

173 Third, revelation during a trial that the accused has prior criminal convictions will often result in the jury being discharged; but it will not preclude a re-trial.

174 Fourth, in some cases - the trial of the applicant for the murder of Ms Maher was an instance - evidence is admitted on a similar fact basis which reveals that the accused offended in the past in a particular fashion. Even so, with appropriate warnings, trials proceed.

175 Fifth, it often enough happens that a conviction is overturned on appeal. Sometimes the conviction was attended by considerable publicity. Yet re-trials are conducted, and stays have been refused, even in cases of grossly prejudicial press reporting and misreporting.[112]

176 Sixth, it is sometimes implicit in the circumstances of an alleged offence that the accused is a person with a prior conviction. Such is the situation where a person is accused of an offence committed whilst in gaol. An example of an extreme case was that of Ronald Ryan, who was tried and convicted of killing a prison warder when escaping from gaol. It has not been said that the accused in a case of this kind is incapable of receiving a fair trial - although ample warnings must be, and are, given against impermissible reasoning in such cases.

177 Applying principle, and recognising that the criminal law is often able to accommodate the problems of adverse pre-trial publicity and revelation of prior convictions without staying criminal proceedings, should it be concluded that the present case involved such extreme circumstances that the refusal to grant a stay led, in the events which occurred, to a miscarriage of justice? And, if the answer be 'yes', should there be a stay - and if so, of what kind - on the further prosecution of the matter?

178 In my opinion, the first question should be answered 'yes'; and the second question should be answered by holding that the further prosecution of the applicant on the charge of murdering Miss Halvagis should be stayed unless and until a judge of the Criminal Division of the Supreme Court or a judge of this Court orders otherwise.

179 In my view, the circumstances of this case were extreme. With great respect to the learned trial judge, his ruling on the stay application shows that he failed to appreciate and take into account the intimate relationship between the large body of highly prejudicial material about the applicant (including his convictions for murder and the repeated intimation that he had murdered Miss Halvagis) and important aspects of the Crown case. As a corollary, when ruling, his Honour did not address the difficulties which defence counsel would inevitably face in presenting a proper challenge to that evidence.

180 Next, whilst I entertain no doubt that his Honour attempted to ensure a fair trial, so that a verdict would be rendered only according to the evidence in the case, I consider that the case had a concatenation of features which, as the trial unfolded, made achievement of that aim an impossibility. Thus:

(1) The applicant was a twice convicted murderer of women, killed by knife attack.

(2) The applicant was accused of killing the victim in the present case, a young woman, by a knife attack.

(3) The applicant's convictions were accompanied by a large amount of mass media publicity, repeated over the years, which accused him - in substance though not in form - of murdering Miss Halvagis; and which informed the public that he was not only a twice convicted murderer but a thrice convicted rapist.

(4) Three of the important threads of the Crown case harked back not simply to the previous murder convictions, but to the mass media publicity which attended them. The first was the identification evidence.

(a) Mrs Burman's identification of the applicant (which is not to downplay the significance of her computer image instructions and her photoboard identification of the applicant as the one of three men most likely to have been the man whom she saw in the cemetery) was made by reference to page 1 of the Herald Sun of 16 August 2000, with its photograph of the applicant handcuffed and being restrained, the photos of Ms Patterson and Miss Halvagis side by side, the headline 'A Monster', and the revelation that 'A sex killer convicted yesterday is also the prime suspect in Melbourne's graveyard murder mystery.'

Each of those matters was placed before the jury. The applicant's counsel had no choice but to bring that about. He did not place the entire content of page 1 before the jury, but one could not sensibly doubt that Mrs Burman had read at least the article on that page. It described the applicant as 'A sex crazed predator,' stated that the applicant had exercised his legal right to refuse to answer questions about the murder of Miss Halvagis, stated that he was also suspected of murdering Ms Maher (the suspicion later turned into conviction), stated that he had been convicted of rape three times, and informed readers that the applicant's grandfather was buried only 50 metres from where Miss Halvagis had been murdered.

(b) Mr Weller's identification of the applicant, made more than seven years after the killing, followed upon him recognising the applicant, as he claimed, on the television, in a news program which 'said something about killing a girl called Margaret Maher and there could have been another girl, and I also remember them saying he was suspected of the Fawkner Cemetery murder as well.' I have already said a deal about the content of relevant newspaper articles as the years passed, and I will not repeat it.

(c) Mrs Melnik's identification of the applicant followed her seeing a television program on an unidentified occasion 'maybe quite long after', which told of the killing of two women and which carried an image of the applicant. It can be inferred, having regard to the content of the program and to the fact that Mrs Melnik made her police statement in April 2005, that the program followed the applicant's October 2004 conviction of the murder of Miss Maher. Thus her identification was not only late-made, it was made in the context of her seeing an image of the applicant in a singular context - a context which she described to the jury.

(5) The second evidentiary thread was what was said to be the lie, told in consciousness of guilt, that the applicant had never been to the cemetery. It depended in the first place upon acceptance of one or more of the identification witnesses; and so it returned to the circumstances of the identification, which in turn drew attention to the applicant's past convictions for murder and his indictment, in the mass media, as the murderer of Miss Halvagis.

(6) The third evidentiary thread was the applicant's alleged gaol-house confession to Fraser. The 'no forensics' component of that evidence, as I have said, drew a link between the applicant's status as the convicted murderer of one woman and his status as the accused murderer of another woman.

(7) A feature of the earlier killings had been that the victims' breast areas were mutilated. In the present case, although most of the many wounds were to Miss Halvagis' upper chest, there was no such mutilation. But, particularly with there being so much interrelationship between the evidence in this case and the other murders, the applicant's counsel would have been foolhardy to cross-examine in order to draw the distinction. His client would have been revealed to be not simply a killer of other women, but a killer who had mutilated other female victims - in each case with a knife - in a truly horrendous way.

181 The learned judge concluded that the applicant's murder convictions were so notorious that at least some of the jurors must have known of them; in which circumstances, he concluded, it was best that they be brought to the jury's attention at the outset of the trial, so that everything would be in the open. The evidence given at trial shows that they must have come to the jury's attention in any event. They did so. When his Honour came to charge the jury, he warned them to deliver a verdict on the evidence, and to put the applicant's prior convictions out of mind. He then said, in effect, that Fraser's evidence meant that the applicant's conviction for murdering Miss Maher would have come to light anyway. But he said nothing about the evidence of any of the identification witnesses, nor about the consciousness of guilt lie, at that point. Neither did he say anything about the intersection between the 'verdict on the evidence' direction and the evidence in the case which depended to a greater or lesser extent upon the applicant's past convictions and the publicity which attended them. Although, in my view, the problem had by then become intractable, I consider that his Honour should have addressed those matters. His failure to do so adds just a little towards my conclusion that a miscarriage of justice eventuated.

182 Having read the entire transcript of the trial, I have the clearest impression that the interrelationship between the applicant's previous convictions for murder, the mass media publicity which attended them (including, in substance, his identification as the murderer of Miss Halvagis), and significant aspects of the evidence led for the Crown in the trial, made it impossible for the jury to put the pre-trial publicity out of mind. Its task of deciding whether the prosecution had established the applicant's guilt to the criminal standard had, in my opinion, become an impossible one. Further, applicant's counsel was, as I have shown, put in a forensically very difficult - it might fairly be said impossible - position at times. Complete cross-examination would necessarily have meant that the extent and nature of the applicant's criminal history, and the nature of the publicity which accompanied his past murder convictions, would be reinforced in the jury's mind. So counsel limited his cross-examination. Even so, some cross-examination which he did essay was apt to have had that effect.

183 I have not ignored the fact that there was evidence, unrelated to the applicant's past convictions and the publicity which attended them, which could have sustained a guilty verdict. Neither have I ignored the fact, of which limited use can be made in an appellate context, that the applicant stood mute. As to the former, the evidence falls into two categories: evidence capable in its own right of sustaining a conviction; and evidence not capable of doing so. Into the first category falls some of Fraser's evidence. Into the second category falls the computer image built up upon Mrs Burman's instructions, her photoboard 'similarity identification' of the applicant, and her evidence that the man she saw had a tear in his jacket at a particular position - a jacket with a tear in a similar place being found at the applicant's premises in 2000. The entirety of that evidence could have led to a conclusion that the applicant was in the cemetery on the morning of the killing, hours before it took place. Also in the second category was the lie, if so the jury concluded it was, that the applicant did not know anyone buried in the cemetery; and such evidence as there was that the applicant had changed his appearance subsequent to Miss Halvagis being killed.

184 It is one thing, however, to acknowledge that it was conceptually possible for the jury to have determined guilt without relying upon evidence affected by the applicant's past murder convictions, the publicity which attended them, and the repeated assertions that he was also the killer of Miss Halvagis. It is another thing to conclude that the jury's verdict was not substantially affected by the combination of knowledge of those convictions, that publicity, and the applicant's indictment as the killer of Miss Halvagis. It seems to me to be quite unrealistic to conclude that in arriving at its verdict the jury: (1) did not directly rely upon evidence, the likely reliability of which was intimately connected with those matters; (2) was not influenced - I do not mean consciously - whether to accept that evidence by reason of those matters; or (3) was unaffected by that evidence when deciding whether or not to accept other prosecution evidence.

185 So to conclude is not intended as a criticism at all of the jury. As I stated above, I do not suggest that the jury was consciously influenced to its verdict by the applicant's past convictions, the publicity which attended them, or his mass media indictment as the killer of Miss Halvagis. There is no reason to suppose that the jury did not do its best to comply with the judge's directions. But I consider that it was faced with an impossible task in attempting to exclude the impact of the matters to which I have referred.

186 In my opinion, whatever might be said about the situation when the learned judge made his ruling on the stay application, the way in which the trial unfolded establishes that refusal to grant a stay did result in a miscarriage of justice. The question becomes, what remedy should flow from the proven miscarriage? Strict adherence to what I perceive to be the balance of persuasion in Glennon would suggest that there should be a permanent stay. For the reasons already expressed, I consider that this is an 'extreme' or 'singular' case in the sense that, despite lapse of time and no matter how careful and thorough may be the directions given to the jury, there is and for the foreseeable future will remain a significant likelihood that any conviction would be affected by substantial prejudice and prejudgment. Put another way, if this is not the singular case contemplated by the authorities, I am unable to conceive of a case in the ordinary experience of the criminal law - I put to one side the special problems created by crimes against humanity - to which that description could apply.[113]

187 There being no binding authority on the point, however, and in circumstances in which the 'balance of persuasion' in Glennon is yet to be applied,[114] I am not prepared to hold that the applicant cannot ever be tried. The object of ensuring a 'fair trial' is a vital consideration. But it is not the only consideration.[115] There is also the vital importance of ensuring that those whose are charged with serious criminal offences are brought to trial.[116] With great respect, the view expressed by Brennan and Dawson JJ in Glennon, that the community cannot afford to acknowledge that the media has the capacity to cause a permanent stay, is a powerful consideration. As Brennan J said:[117]

If it were otherwise, the perpetrators of crimes which shock the public conscience ...would oftentimes go untried and unpunished, for pre-trial publicity prejudicial to an accused is stimulated by the notoriety of the accused and the heinousness of the crime. Yet it would undermine the criminal law's protection of society and its members to refuse to allow the law to take its ordinary course in these cases. The administration of justice by the courts, which proceeds inexorably to its conclusion in each case, would be adventitious...

188 In the absence of authoritative guidance from the High Court, I cannot be certain how one is to resolve the balance of persuasion in Glennon with the competing policy considerations essayed in that passage. But approaching the problem as a matter of logic and common sense, it appears to me that the question is essentially one of degree. On the one hand, it is fundamental that for an accused to have a fair trial the jury should reach its verdict by reference only to the evidence admitted at trial and not by reference to facts or alleged facts gathered from the media or some other outside source.[118] Accordingly, the law requires trial judges to conduct trials (within ordinary procedural constraints) so as to counter the effects of prejudicial pre-trial publicity.[119] On the other hand, the law recognises, at least implicitly, and despite statements to the contrary, that there is no guarantee that a trial judge will be successful in eliminating the prejudicial effects of pre-trial publicity. Thus, to some degree, the law tolerates the risk of prejudice in the interests of ensuring that a trial be had. To that degree, the concern that pre-trial publicity may render the trial unfair has been rationalized on the basis that 'unfairness occasioned by circumstance outside the court's control does not make the trial a source of unfairness'. So, within the far-reaching but imprecise boundary thus established, the law, having done what it can to guard against the prejudicial effects of pre-trial publicity, and although the results may not be perfect, accepts that a trial 'is a trial according to law and there is no miscarriage of justice in a conviction after such a trial'.[120]

189 On that analysis, the question is whether the circumstances of the particular case, including the circumstance that any unfairness has been occasioned by circumstances outside the court's control, define it as being inside or outside that boundary line. In a case which is outside it, the trial should be stayed. One might expect, in what will be an extremely rare situation, that the trial should be stayed permanently. But it seems to me that so extreme a remedy need not be mandatory. The object of any stay must be to prevent a fresh miscarriage of justice. The balance of persuasion in Glennon and the powerful considerations mentioned by Brennan J might be reconciled (though not perfectly) by fashioning a stay which admits of the possibility that circumstances may arise in the future where a sufficiently fair trial can be had.

190 That is the solution which I would adopt in this case. For the reasons which I have described, I consider that the judge's refusal to grant a stay resulted in a miscarriage of justice. I further consider that an unacceptable level of prejudice would attend any retrial of the applicant in the foreseeable future. Although it is to be expected that public recollection of the applicant and his crimes will dim with time, most of the problems of ensuring a fair trial are likely to endure. The fundamental difficulty is the connection between the evidence and the pre-trial publicity. Consequently, if the evidence to be given at any re-trial were substantially the same as that given at the trial in 2007, the problems of ensuring a fair trial would be substantially the same.

191 That said, however, even after this elapse of time one cannot be certain that new evidence might not come to light, or that some other new circumstance might not arise which the Crown would wish to contend makes a sufficiently fair trial possible.[121] Nor can it confidently be predicted that, in the face of the increasing and seemingly unassailable trend of interference with the due administration of justice by tendentious pre-trial publicity,[122] the law will not come to tolerate a higher risk of prejudice than that which is now tolerated. Thus it is conceivable, although it does not now seem likely, that the time will come when it is possible to afford the applicant a sufficiently fair trial, either because of a change in evidence or other circumstance or because of a reduction in acceptable standards of fairness. Either way, in those circumstances, there would be no sufficient reason why the applicant should not be put up again for trial.

192 In the result, and as I said earlier, I consider that any proceeding against the applicant on the charge of murdering Miss Halvagis should be stayed until further order of a judge of the Criminal Division of the Supreme Court or of a judge of this Court. But this is a minority opinion. Since I agree with the reasons of Nettle JA with respect to grounds 5 and 6, and thus respectfully disagree with what my brother Weinberg JA has said in that connection, it follows that the application should be granted, the appeal allowed, and a retrial had.

193 Courts have repeatedly referred, perhaps particularly in the last 20 years or so, to mass media outlets, individual journalists, media 'personalities' and law enforcement authorities so conducting themselves as to imperil the fair trial of accused persons. The passage from the reasons for judgment of Brennan J in Glennon which I cited at [146], although it did not deal with all pertinent situations, in my respectful opinion made the point very clearly. To his Honour's list may be added an apparent problem of law enforcement agencies, or individual members of such agencies, 'feeding' their suspicions or allegations to selected journalists and commentators with a view of publication.

194 It has been universally recognised by the courts that conduct of the kinds described has the potential to interfere with an accused's right to a fair trial. That right is given to every accused person. No lesser right is conferred because the media, or sections of it, adjudges an accused to be a bad person, or a very bad person, or even a monster. It is a right which exists in respect of each trial of an accused offender. An accused's prior convictions, even for heinous crimes, are ordinarily not disclosed to a jury in the instant trial, this assisting towards a fair trial. The right of which I have been speaking is not the less because victims and their families have rights. It is not the less because there is or may be money, circulation or ratings in 'fearless' or 'investigative' journalism. Neither again must it defer to 'the public's right to know', or an asserted 'freedom of the press'.

195 The approach of the courts, in the face of the misbehaviour which I described at [193], has largely been to explain why a fair trial, or a sufficiently fair trial, can be had in the particular case. Resort is had to - (1) the asserted ability of properly instructed juries to abide their oath, and deliver a verdict only according to the evidence - virtually regardless of the nature and extent of the prejudicial material; (2) the contention that the media publishes so much material of shocking import each day that shocking and prejudicial material published one day about an accused will soon be forgotten, or lose its sting; (3) the related proposition that so much shocking material is published daily in the mass media that the public will take any individual item with a grain of salt; (4) the proposition that summoning a large jury panel, and taking excuses after the panel has been well-apprised of the matter for trial, will result in the empanelment of a suitably unprej[123]ced jury;123 (5) the interest of the community that persons charged with offences, particularly serious offences, should stand their trial; (6) the dictum that there will be no miscarriage if the trial is as fair as can be had; and (7), the proposition that the law simply cannot accept that a trial which is arguably rendered unfair by reason of circumstances outside the court's control is not an unfair trial.

196 There is no problem, in principle or practice, with resort to those considerations where a stay application is just another arrow in the quiver. Such cases - and I have no doubt that they are the large majority - can readily be seen for what they are. The problem is the truly exceptional case. In that very rare case, in my opinion, reliance upon those reasons, singly or cumulatively, to justify a trial being held involves a too-ready acceptance that deplorable conduct of the mass media, individual journalists and commentators can be accommodated by the criminal justice system without doing unacceptable damage to a fundamental tenet of that system - that is, the right of an accused to a fair trial. If, in that very rare case, a trial cannot be had, the victim, the victim's family, and the public should be assisted by the court to understand why it is that the trial cannot be had. As a corollary of that observation, it is, I consider, long past time for the courts to adequately face up to and address the source of the problem, albeit that most often the excesses of the media will not preclude a trial being had.

197 I have had the advantage of reading in draft the reasons for judgment prepared by Nettle JA and Ashley JA respectively. Their Honours have concluded, each for their own reasons, that leave to appeal against this conviction should be granted, and the appeal allowed. With great respect, I am unable to agree. I would refuse leave to appeal.

198 In explaining why I differ from my brother judges, I will focus firstly upon the finding reached by Ashley JA that, although a permanent stay of the charge of having murdered Mersina Halvagis was not warranted (ground 1), any proceeding against the applicant on that charge should nonetheless be stayed until further order. I will then deal with grounds 2 and 6, which concern identification evidence and indicate why I differ from my brethren in respect of at least ground 6. Finally, I will deal with ground 5, which complains of the adequacy of the directions given as to 'post-offence conduct'. As regards the remaining grounds of appeal, I agree for the

reasons given by the other members of the Court that they are entirely without substance.

Ground 1: The Refusal to Grant a Permanent Stay

199 The applicant submitted that the trial judge erred in refusing to grant a permanent stay. It was submitted that the extent of the publicity that had attended his previous convictions for the murders of Nicole Patterson on 19 April 1999 and Margaret Maher on 4 October 1997 meant that he could not have received a fair trial, and that no period of delay could adequately have resolved the problem.

200 Ashley JA has set out the relevant chronology, and has canvassed thoroughly the principal authorities that bear upon this ground. That relieves me of much of the burden of dealing with those matters. There can be no doubt that, over a number of years leading up to the commencement of this trial for the murder of Ms Halvagis, the applicant had been the subject of a great deal of media attention, much of it sensational in nature and expressed in highly emotive terms. His name and photograph had appeared in the press, and on television. Descriptions of him as a notorious rapist and murderer had been widely publicised. Right from an early stage, he was identified in the media as a suspect in relation to the murder of Ms Halvagis.

201 The trial judge considered, but rejected, an application made on his behalf for a permanent stay. His Honour accepted that the applicant had been the subject of a great deal of media attention, much of it prejudicial in nature. However, he concluded that a jury, properly directed, could be trusted to decide whether the applicant's guilt had been established on the basis of the evidence led, and without regard to anything else they might have learned about him from other sources. Ground 1 complains that his Honour erred in rejecting the application for a permanent stay.

202 The problem of pre-trial publicity has received a good deal of attention in recent times. Sometimes that publicity has led to a court having to adjourn a trial for a period of months. In such cases, it is generally assumed that, over time, the memory that the public has of the subject of the publicity will fade. It is inevitable, however, that there will be cases where, by reason of the nature and extent of the publicity, even if memory has subsided, recollections of what the media has said will revive once the trial commences.[124] Plainly, the availability of such material on the internet complicates matters still further.

203 Even putting to one side those cases in which a trial is adjourned in order to ameliorate the problems of publicity, jurors are invariably told to confine their consideration to the evidence that is led in court. If an accused is charged with an offence that has previously attracted a great deal of publicity, jurors are specifically warned to put that out of their minds and to focus solely upon the evidence that they hear during the course of the trial.

204 It is a fundamental tenet of our criminal justice system that jurors decide cases according to their oaths, and that they obey instructions that trial judges give them. In Gilbert v The Queen,[125] McHugh J said:

The criminal trial on indictment proceeds on the assumption that jurors are true to their oath, that, in the quaint words of the ancient oath, they hearken to the evidence and that they obey the trial judge's directions. On that assumption, which I regard as fundamental to the criminal jury trial, the common law countries have staked a great deal. If it was rejected or disregarded, no one - accused, trial judge or member of the public - could have any confidence in any verdict of a criminal jury or in the criminal justice system whenever it involves a jury trial. If it was rejected or disregarded, the pursuit of justice through the jury system would be as much a charade as the show trial of any totalitarian state. Put bluntly, unless we act on the assumption that criminal juries act on the evidence and in accordance with the directions of the trial judge, there is no point in having criminal jury trials. It is of course true that, if a jury persists in returning a verdict that is contrary to law, the trial judge must accept it. But that only means in Lord Mansfield's words that, although '[i]t is the duty of the Judge ... to tell the jury how to do right ... they have it in their power to do wrong'.

In my respectful opinion, the fundamental assumption of the criminal jury trial requires us to proceed on the basis that the jury acted in this case on the evidence and in accordance with the trial judge's directions and that they would have done so even if manslaughter had been left as an issue, as it should have been left. In Spratt, Pidgeon J said, correctly in my opinion, 'that an appellate court must proceed on the basis that the jury have understood and applied the law in reaching a true verdict'.[126]

205 The fact that a prospective juror may have read or seen something in the media about a particular case has never been regarded as a sufficient basis, on its own, for excusing that juror from jury service. Jurors are expected to be able to distinguish between evidence, on the one hand, and rumour, gossip, and whatever else the media may have reported, on the other.

206 Ashley JA has analysed with meticulous care the reasoning of the various members of the High Court in The Queen v Glennon.[127] That case dealt with the circumstances under which an accused person might, conceivably, be granted a permanent stay in response to widespread antecedent publicity. It will be recalled that, in Glennon, the Victorian Court of Criminal Appeal had held, by majority, that by reason of publicity given to the accused by a radio broadcaster before trial, he could not receive a fair trial. That decision was reversed by the High Court.

207 Mason CJ and Toohey J, in a joint judgment, referred to the 'principled decision of Crockett J' who had rejected the application for a stay on an originating summons.[128] Their Honours said that the majority in the Court of Criminal Appeal had taken 'little, if any, account of the effect of the trial judge's instructions and disregarded the community's right to expect that a person accused of a serious criminal offence will be brought to trial'.[129]

208 Brennan J (with whom Dawson J agreed) emphasised that even when jurors are shown to have had knowledge of the accused and his or her alleged crime before trial, it would not follow that this has resulted in a miscarriage of justice.

209 After referring to what Crockett J had said in rejecting the application for a stay on the originating summons, Brennan J observed:

In my respectful opinion, his Honour's conclusion was clearly right either on the ground that the present case is not an 'extreme case' or on the ground -- which, in my respectful opinion, is a ground better founded on principle and more realistic in practice -- that the trial of the applicant, provided it was as fair as the Court could make it, would produce no miscarriage of justice.[130]

210 Even Deane, Gaudron and McHugh JJ, who dissented in Glennon, made it clear that in cases of this type, where a stay of some kind might be warranted, that stay should be temporary and for the minimum period necessary.[131]

211 There then followed the one dictum, upon which so much of the applicant's argument before us rested, that one could not exclude, as a matter of law, the possibility that an 'extreme' or 'singular' case might arise in which the effect of a sustained media campaign of vilification would be such that, notwithstanding careful directions by a trial judge, any conviction would be unsafe and unsatisfactory. In such a case, so the dictum said, a permanent stay might, in rare and exceptional circumstances, be granted.[132]

212 It goes almost without saying that, in determining whether to grant a stay based upon pre-trial publicity, a court must balance two competing interests. On the one hand, the right of an accused person to a fair trial is fundamental. On the other, there is a substantial public interest that those who are charged with serious criminal offences should be tried on those charges, and that their trials should proceed in a proper and orderly manner.

213 To my mind, there is a real question as to what weight should properly be given to what is after all only a dictum, and a dissenting one at that, by the minority in Glennon regarding the effect of an 'extreme' or 'singular' case, whatever that might entail. Plainly, that observation received no support from the other members of the High Court. Indeed, based on their analysis, Brennan and Dawson JJ would almost certainly have regarded it as heresy.

214 That is not to say that Glennon should be understood as undermining, in any way, the right of an accused person to a fair trial.[133] In this State, which makes no provision for trial by judge alone, that means trial before a jury of one's peers that is objective and unbiased. A trial judge faced with a situation where there has been extensive pre-trial publicity must consider whether, in all the circumstances, including the directions which will be given to the jury, they will bring an impartial mind to the issues to be decided. It is, as McHugh J so forcefully stated in Gilbert, a fundamental axiom of our criminal justice system that the jury will understand and obey the directions of a trial judge.[134]

215 Of course, the point made by McHugh J had been made many times before.

In Yuill v R,[135] Kirby ACJ observed:

Courts will assume that jurors, properly instructed, will accept and conform to the direction of the trial judge to decide the case solely on the evidence placed before them in the court ... There is an increasing body of judicial opinion, lately expressed, to the effect that whatever pre-trial publicity exists, jurors when they take on the solemn responsibility of the performance of their duties in the court room, differentiate between gossip, rumour, news and opinion which they hear before the case and the evidence which they hear in the court in the trial for which they are empanelled.

216 To the same effect is R v Milat,[136] where Gleeson CJ (with whom Meagher JA and Newman J agreed) had this to say:

The criminal justice system, of which trial by jury is an integral part, often has to function in circumstances of intense publicity potentially prejudicial to an accused person. Various mechanisms, including, where necessary, proceedings for contempt of court, are available to protect the integrity of the system. Ultimately, however, it is the capacity of jurors, properly instructed by trial judges, to decide cases by reference to legally admissible evidence and legally relevant arguments, and not otherwise, that is the foundation of the system.

217 In R v Dudko,[137] Spigelman CJ considered the question of pre-trial prejudicial publicity at some length. His Honour said:

Issues of this character have arisen on many occasions in which extensive publicity had been given to particular cases involving clear implications of guilt. Glennon involved a Roman Catholic priest who was convicted of sexual offences against a girl. Murphy involved the Anita Cobby murderers. Other cases include the case of Ivan Milat (see R v Milat, NSWCCA, 26 February 1998 (unreported)), the 'Mr Bubbles' child sexual abuse case (see R v VPH, NSWCCA, 4 March 1994 (unreported), the paedophile Phillip Bell (see R v Bell, NSWCCA, 8 October 1998 (unreported)), the child sexual assaults committed by a former member of the Queensland parliament (see R v D'Arcy [2001] QCA 325) and the Childers backpacker fire (see Long v R [2002] QSC 54). None of these cases satisfied the test of exceptionality, despite the intensity of the media publicity involved.

...

The authorities to which I have referred establish that even in circumstances of a crime of a high level of notoriety, where it would probably not be possible to select a jury panel who had not heard about the case and indeed who may have a tentative opinion, the trial must still proceed. Jurors who may have formed an opinion are not necessarily biased in the relevant sense. There is now a substantial body of judicial statements of the opinion that jurors accept their responsibility to perform their duties by differentiating between the evidence and what they may have heard before the trial.[138]

218 Recently, when confronted with an application for a permanent stay based on prejudicial pre-trial publicity, Kaye J in R v Mokbel[139] made a similar point:

In addition, it has long been the experience of trial judges in this State, and in this country, that jurors in criminal (and indeed civil) trials are particularly astute and conscientious in adhering to directions given to them as to their role, and as to the principles which they are to apply in determining their verdicts. It is important not to overlook, nor to underestimate, the capacity of juries and the increasing sophistication of them. The recorded cases are replete with statements by judges reflecting the experience of the common law courts that juries have consistently, for many decades, been particularly rigorous in the performance of their judicial roles.[140]

219 As Ashley JA points out in his judgment in this case, although there have been many occasions where accused persons have sought permanent stays based on prejudicial publicity, there appears to have been only one instance, Tuckiar v The King,[141] where something loosely akin to such an order has been made. In that case, the High Court, having quashed a conviction for murder, did not order a retrial because defence counsel had made public statements following the conviction to the effect that his client had admitted to him that a confession attributed to him was true. That case, of course, had nothing whatever to do with pre-trial publicity.

220 There are some instances of trial judges having granted permanent stays by reason of pre-trial publicity, but whose decisions have been overturned on appeal.[142] There are many other cases where, despite extensive pre-trial publicity of a highly prejudicial nature, no such applications were made or, if made, were rejected. I shall turn to some examples shortly. It may be useful, however, before doing so, to restate some of the principles which underlie the power to stay criminal proceedings.

221 In Australia, that power was first explicitly recognized by the High Court in Barton v The Queen.[143] That case raised the question whether the power of the Attorney-General to file an ex officio indictment was subject to judicial review. The answer was no, but the joint judgment of Gibbs ACJ and Mason J (with whom Aickin J agreed) made it clear that this did not preclude the exercise of a power to stay a criminal proceeding where necessary to prevent an abuse of process, and to ensure a fair trial.

222 Subsequently, in Jago v The District Court of New South Wales,[144] the High Court considered the circumstances in which that power could be exercised. Referring to a passage from the judgment of Wilson J in Barton, Mason CJ said:

To justify a permanent stay of criminal proceedings, there must be a fundamental defect which goes to the root of the trial 'of such a nature that nothing that a trial judge can do in the conduct of the trial can relieve against its unfair consequences'.[145]

223 The Chief Justice went on to say that the power to grant a permanent stay 'will be used only in most exceptional circumstances'.[146] The test of fairness to be applied involved a 'balancing process'. The community had the right to expect that persons charged with criminal offences would be brought to trial.

224 Brennan J was extremely critical of the idea that a permanent stay might be ordered merely because there had been extensive pre-trial publicity. His Honour observed:

A power to ensure a fair trial is not a power to stop a trial before it starts. It is a power to mould the procedures of the trial to avoid or minimize prejudice to either party. [147]

Obstacles in the way of a fair trial are often encountered in administering criminal justice. Adverse publicity in the reporting of notorious crimes ... adverse revelations in a public inquiry ... absence of competent representation ... or the death or unavailability of a witness, may present obstacles to a fair trial; but they do not cause the proceedings to be permanently stayed. Unfairness occasioned by circumstances outside the court's control does not make the trial a source of unfairness. When an obstacle to a fair trial is encountered, the responsibility cast on a trial judge to avoid unfairness to either party but particularly to the accused is burdensome, but the responsibility is not discharged by refusing to exercise the jurisdiction to hear and determine the issues.[148]

226 Brennan J was less inclined than the other members of the Court to take a broad view of the concept of abuse of process. His Honour observed:

An abuse of process occurs when the process of the court is put in motion for a purpose which, in the eye of the law, it is not intended to serve or when the process is incapable of serving the purpose it is intended to serve. The purpose of criminal proceedings, generally speaking, is to hear and determine finally whether the accused has engaged in conduct which amounts to an offence and, on that account, is deserving of punishment. When criminal process is used only for that purpose and is capable of serving that purpose, there is no abuse of process.[149]

227 Brennan J's concluding remarks regarding the remedy of a permanent stay should be noted. While recognising the serious delays which had frustrated the administration of criminal justice in New South Wales and in other States, and the prejudice thereby suffered by accused persons, his Honour rejected as inappropriate the remedy of a permanent stay simply to mark the court's disapproval of the failure of other branches of government to furnish the resources necessary to cope with an accumulation of criminal cases awaiting trial. He described the permanent stay as a 'radical discretionary power' which involved an abdication of the court's duty to try the case.

If it be said that judicial measures cannot always secure perfect justice to an accused, we should ask whether the ideal of perfect justice has not sounded in rhetoric rather than in law and whether the legal right of an accused, truly stated, is a right to a trial as fair as the courts can make it. Were it otherwise, trials would be prevented and convictions would be set aside when circumstances outside judicial control impair absolute fairness. ... If permanent stay orders were to become commonplace, it would not be long before courts would forfeit public confidence. The granting of orders for permanent stays would inspire cynicism, if not suspicion, in the public mind.[150]

229 Brennan J decried the notion, at that time widely prevalent, that the category of case in which the power to grant a permanent stay should be exercised was not closed, but the power was available 'whenever it would be unfair to the accused to permit the prosecution to proceed'.[151] Indeed, his Honour stated:

In practice, so broad a power does not fall far short of a power which is incompatible with the rule of law. Of course, one finds in these cases the qualification that a permanent stay will be ordered only in an 'exceptional' case, but that is an ineffectual qualification to place upon so broad a power.[152]

230 These days, the views of Brennan J would generally be regarded as mainstream. It is now rare for a permanent stay to be granted solely on the basis of delay. Even offences that go back many years are routinely tried.

231 Ashley JA has concluded, after a thorough and careful analysis, that the applicant's case provides one of those rare instances where it can be said that the jury would have found it impossible to carry out the trial judge's clear instructions, and therefore could not possibly have discharged their responsibilities according to law. In his Honour's view, no jury could have failed to be influenced in their deliberations by the knowledge, necessarily before them, that the applicant had previously committed crimes of the most heinous nature against other women. It therefore follows, according to his Honour, that the trial miscarried and that the appeal must be allowed. It also follows, his Honour says, that any further prosecution of the applicant on the charge of murdering Ms Halvagis must be stayed, at least until further order.

232 Both Nettle JA and Ashley JA characterise this as an 'extreme' or 'singular' case. Ashley JA says that despite the lapse of time, and no matter how careful and thorough the directions given to the jury, there is, and for the foreseeable future will remain, a significant likelihood that any conviction would be affected by substantial prejudice and prejudgment. He says that, having reached that finding, Glennon would suggest that any stay should be permanent. Yet he draws back from that conclusion, essentially on the basis that there is room for the view expressed by Brennan and Dawson JJ in Glennon that the community cannot afford to acknowledge that the media has the capacity to cause a permanent stay. He posits a situation whereby things may change, new evidence may come to light, or some new circumstance may emerge which would make a fair trial possible.

233 With the utmost respect, the prospect that something new may develop, which would enable this applicant to be tried again, for this or perhaps even any other offence, assuming for the moment that he did not receive a fair trial based on prejudicial publicity, strikes me as unrealistic. The factors that have led his Honour to conclude that the applicant could not, and therefore did not, receive a fair trial will in no way change in the foreseeable future. As a matter of practical reality, any stay that might be ordered will end up having to be permanent.

234 I acknowledge, of course, that the applicant and his counsel faced real difficulties in the present case. Whether or not those difficulties warrant the description 'extreme' or 'singular' is perhaps a matter of debate. However, even if that characterisation is apt, I would still hold that there was no basis for either a permanent stay, or one that would operate as such.

235 In arriving at that conclusion, I accept that the publicity that had preceded this trial meant that, no matter how the defence was presented, the jury were almost certain to have known something of the applicant's history. It was inevitable that one or more members of the jury would have been aware, from what they had seen or heard in the media, that he had previously been convicted of murder.

236 I also accept that the Crown case, as presented through Mr Fraser, meant that even if the pre-trial publicity had not had that effect, it was inevitable that the applicant's history would emerge. How else could Mr Fraser's evidence regarding the admissions supposedly made to him possibly be presented?

237 This meant that the applicant's counsel faced a number of unusual, but by no means unique, forensic challenges. The fact that he chose to meet those challenges head on, by having the jury told at the outset that his client had previously been convicted of the murder of two other women, was entirely understandable. That was a reasonable forensic choice.

238 However, there is nothing particularly remarkable, to my mind, about a case where, as a matter of necessity, the jury must be informed of an accused's previous history, including prior convictions. That occurs not infrequently, and even when the evidence regarding those prior convictions is highly prejudicial. Normally, this is dealt with by careful instructions being given to the jury as to how they must approach evidence of that kind. That happens as a matter of course when similar fact evidence is led. Sometimes such evidence is adduced even when it reveals to the jury the applicant's involvement in the most horrific of crimes.[153]

239 There are many instances where trials have had to be conducted against the background of the most extensive prejudicial publicity. That is no doubt unfortunate. However, it does not necessarily follow that those trials were in any relevant sense unfair. As McHugh J indicated in Gilbert, it must be assumed, at least when considering whether a case should be stayed, that a jury, properly instructed, will put to one side any pre-trial publicity and will focus solely upon the evidence led in the trial. It must also be assumed that the jury will obey any directions given by the trial judge as to how material that is potentially prejudicial can, and cannot, be used. It is no exaggeration to say that our criminal justice system depends, for its very existence, upon the acceptance of these assumptions.

240 There have been many recorded instances of accused persons having been acquitted despite extensive pre-trial publicity of the most prejudicial kind.[154] There are also many well-known examples of juries having acquitted despite the accused's prior convictions having been led before them.

241 The case of R v Wright and Haigh[155] provides a useful illustration of just this point. Robert Wright and Barry Quinn were charged on the first count of a five-count presentment with the murder of one K. The jury acquitted each of them on that count. Wright and Paul Haigh were charged on counts 2, 3, 4 and 5 with the murders of S, G, M and B respectively. Wright was convicted on counts 2, 3 and 4, but, significantly, acquitted on count 5. Haigh was convicted on counts 2, 3, 4 and 5.

242 The offences charged were among the most callous and brutal crimes imaginable. They included the cold-blooded execution of a mother and her young son, as well as the frenzied stabbing of a young woman. These murders were committed solely in order to cover up other crimes. Although all the counts of murder were tried together, and despite the fact that the jury convicted on most of them, it is important to note that they were still able to acquit on others. The very fact that they were able to do so provides a clear illustration of the ability of jurors to put to one side, when considering their verdict on a count of murder, even the knowledge that the accused has been shown to be a sadistic and psychopathic killer.

243 Other examples can readily be found of prejudicial publicity not having led to the grant of a permanent stay. There would be few cases that have attracted greater pre-trial publicity in this State than what was known as the Faraday School kidnapping. On 6 October 1972, two men, Edwin Eastwood and Robert Boland, entered the school armed with sawn-off shotguns. They forced the teacher and her six pupils, aged between five and 10, into a van. The van was driven off into a remote area in the bush. The kidnappers left a note at the school threatening to kill all the hostages unless a $1 million ransom was paid. The then Premier of Victoria, Dick Hamer, announced that the State Government would pay the ransom. The Education Minister, Lindsay Thompson, arrived at the scene and waited alone to deliver the money, but it was never collected. Fortunately, while the kidnappers were gone, the teacher and the children managed to escape.

244 In December 1972, Eastwood pleaded guilty to the kidnapping. He was sentenced to a lengthy term of imprisonment. He gave evidence against Boland, who was convicted by a jury in March 1974.

245 In 1976, Eastwood escaped from Geelong Prison. In February 1977, he kidnapped a teacher and nine pupils from the Wooreen State School in Gippsland. He also took three elderly women hostage. He demanded a ransom of $US7 million, guns, drugs and the release of a number of inmates from Pentridge Prison. However, one of the hostages escaped. Eastwood was eventually captured by the police, but only after a shootout.

246 Eastwood eventually pleaded guilty to 25 charges arising out of the second set of kidnappings. Yet, if the applicant's submissions in the present case are to be accepted, he was perhaps ill advised to do so. Eastwood had, by then, become known generally as the 'Faraday kidnapper'. Had he stood trial for the Woreen State School kidnappings, it would have been all but impossible to find 12 jurors who were unaware of his identity, and what he had previously done. That would have meant, on the applicant's case as argued before us, a permanent stay.

247 The idea that Eastwood could not have been tried for the Woreen State School kidnappings because of the enormous pre-trial publicity that had been given to the case must surely be regarded as unacceptable.

248 The same can be said of those cases where persons who have been imprisoned for horrific crimes are subsequently charged with offences allegedly committed by them while in prison. One has only to think of Craig Minogue, one of the Russell Street bombers, later tried for the murder of a fellow prisoner, Alex Tsakmakis, in the notorious H Division at Pentridge, to appreciate how strong must be the argument against the granting of a permanent stay solely on the basis of pre-trial publicity. Not only would the jury have known exactly who Minogue was, they would also have understood that his presence in H Division at the time of the killing spoke volumes about his then propensity towards violence. It is interesting to note that not even in that case was there any suggestion that Minogue could not be fairly tried for the murder of Tsakmakis.

249 The Walsh Street trial provides another illustration. That case concerned the murder of two Victoria Police constables, Steven Tynan and Damian Eyre. They were responding to the report of an abandoned car when they were gunned down in Walsh Street, South Yarra, in the early hours of 12 October 1988. Four men, Victor Peirce, Trevor Pettingill, Anthony Farrell and Peter McEvoy, were charged with murder. The case against them was opened on the basis that they were all violent criminals engaged in a war against the police, a reflection of the entire media coverage leading up to the trial. The prejudice was no doubt enormous, but the accused were all acquitted.[156]

250 The reason why these examples are put forward is not to show that juries invariably put to one side prejudicial material, and consider cases entirely upon their merits. Such a conclusion could not be drawn from a handful of cases, carefully chosen. What these cases may illustrate, however, is that it should not be too readily assumed that juries find it 'impossible', in difficult cases, to discharge their responsibilities in accordance with their oaths.

251 Without wishing to be facetious, on the applicant's case as argued before us, and applying the dicta of the dissentients in Glennon, Jack the Ripper, having been once convicted of the murder of one of his victims, would thereafter be immune from further prosecution. He would presumably be entitled to a permanent stay, based on pre-trial publicity and prejudice. One can only speculate about the possibility of trying Charles Manson, Ronald Biggs, or dare I say, Osama Bin Laden, faced with the view of the dissentients in Glennon.

252 Perhaps more to the point, if the applicant had been seen by a dozen witnesses of unimpeachable reliability to have murdered Ms Halvagis, and subsequently confessed to having done so, he would still, on his case, be entitled to a permanent stay in respect of that crime. No sensible system of criminal justice could possibly countenance such an outcome.

253 In my view, neither the pre-trial publicity regarding the applicant in this case, nor the fact that the jury were, of necessity, made aware that he had previously been convicted of murder, warranted the extraordinary course of granting a permanent stay of this trial. Nor does the fact of such pre-trial publicity, or the knowledge on the part of the jury of his previous convictions, warrant the quashing of this conviction. In that regard, I agree with Nettle JA as to the disposition of ground 1, though plainly my reasons are far apart from his. With great respect to Ashley JA, I disagree with his Honour's conclusion that even the more limited stay, which he regards as having been appropriate, should have been granted. There is nothing in any of the authorities to which my attention has been drawn that would come close to justifying such a conclusion.[157] In that regard, I include English cases as well as those decided in this country. The authorities show, overwhelmingly, that courts place significant faith in the integrity of juries, and in their capacity to understand and adhere to directions. For the reasons set out above, I am not persuaded that the applicant was denied a fair trial. I would therefore reject ground 1.

Ground 2: Admissibility of Identification Evidence.

254 As both Nettle JA and Ashley JA have explained, three witnesses, Laima Burman, Horst Peter Weller and Katica Melnik, gave evidence of having seen the applicant at the Fawkner Cemetery on the day of the murder. Counsel for the applicant sought, unsuccessfully, to have their evidence excluded in the exercise of the trial judge's discretion. It was submitted that, having regard to the circumstances in which each identification had taken place, it would be dangerous to allow that evidence to be led.

255 In particular, it was submitted that each identification was tainted. Two of the witnesses, Mr Weller and Mrs Melnik, had previously seen the applicant on television in connection with other murders of which he had, by then, been convicted. The third witness, Mrs Burman, had seen his photograph in the Herald Sun newspaper, together with commentary suggesting that he had murdered Ms Halvagis. In addition, it was submitted that, in relation to Mr Weller and Mrs Melnik, a number of years had passed between their sightings at the Fawkner Cemetery and their identification of the applicant as the person they had seen that day.

256 It was submitted on behalf of the applicant that none of the identification evidence had any real probative value and that such probative value as it had was significantly outweighed by its likely prejudicial consequences. In other words, it was sought to have this evidence excluded on the basis of what has come to be known as the 'Christie discretion'.[158] Of course, that discretion must be considered separately from other exclusionary discretions based upon unfairness[159] or public policy.[160]

257 It is well established that a trial judge has a discretion to exclude evidence if the strict rules of admissibility would operate unfairly against the accused. In Driscoll v The Queen, Gibbs J said:

The exercise of this discretion is particularly called for if the evidence has little or no weight, but may be gravely prejudicial to the accused ... [161]

258 In R v Sandford,[162] Hunt CJ at CL observed:

The probative value of the evidence, although relevant to that issue, must therefore have been slight indeed. The appellant says that such was its prejudicial effect that the judge should in those circumstances have exercised his discretion to exclude it: Rex v Christie [1914] AC 545 at 559, 564; 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. But what is the prejudice? Despite an apparently unshakeable misconception to the contrary on the part of some members of the legal profession, the mere fact that evidence tendered by the Crown is of little weight by itself does not require its rejection under this head of judicial discretion. Nor will evidence be excluded where the only prejudice which it causes is that it establishes (or tends to establish, or assists in establishing) the guilt of the accused. The 'Christie' discretion to exclude evidence is usually directed to evidence which although not itself probative (or only slightly probative) of guilt, is also probative of some other matter which may wrongly be regarded by the jury as probative (or strongly probative) of guilt - for example, propensity: Scott v The Queen [1989] AC 1242 at 1256-1257; Regina v Masters (1992) 26 NSWLR 450 at 479.[163]

259 It is important to note that the only basis upon which the identification evidence was sought to be excluded was the Christie discretion. However, even assuming that there was a question as to the reliability of some of that evidence, it is by no means clear that that would trigger the operation of that discretion.[164]

260 In R v Peirce,[165] Vincent J (as he then was) considered some of the authorities bearing upon the exercise of the Christie discretion in relation to evidence said to be unreliable. His Honour did so in the context of testimony sought to be led from an unindemnified accomplice of the accused. He accepted the theoretical possibility that such evidence could be excluded in the exercise of this discretion, but said that the occasions upon which that might occur would be few and far between. Normally the issues raised would be left to the jury as matters of credibility for their determination.

261 A similar conclusion was reached by the Victorian Full Court in Rozenes v Beljajev,[166] where it was said that there was no example, of which the Court was aware, of the actual exclusion of the evidence of an accomplice on the sole ground of its unreliability.

262 There are many instances in the cases of identification evidence that could be said to be suspect having nonetheless been admitted.[167] As will be seen, Domican v The Queen[168] provides a paradigm example.[169] The safeguard against the possible misuse by the jury of such evidence is normally the giving of appropriate directions by the trial judge.

263 In Festa v The Queen,[170] Gleeson CJ specifically rejected the contention that the possible weakness of identification evidence in a particular case meant that it should be excluded in the exercise of the Christie discretion. His Honour had this to say:

The argument that the evidence of the four witnesses should have been excluded turned upon what were said to be deficiencies in its quality.

The strength or weakness of evidence may depend in part upon the use that might be made of it. Mr Hill's selection of three photographs, including one of the appellant, of itself could not support a positive conclusion that the woman he saw was the appellant. But the evidence did not stand alone. And even if it only showed that the woman he saw was consistent in appearance with the appellant, that was a material fact. Similarly, the cogency of the evidence of the acts of identification at the Southport courthouse depended in part upon what was sought to be made of it. As positive identification of the appellant, it was weak. In fact, the evidence of Ms Ogilvie and Mr James did not amount to positive identification. But as evidence that the appearance of the appellant was consistent with that of the wigged female seen near the bank at the time of the Biggera Waters robbery, it was of some probative value.

Questions as to the admissibility of evidence may be related to, but are different from, questions as to whether the totality of the evidence in a case is sufficient to sustain a jury's verdict, or questions as to the warnings that need to be given to a jury about the use that may properly be made of the evidence. If evidence is of some, albeit slight, probative value, then it is admissible unless some principle of exclusion comes into play to justify withholding it from a jury's consideration. It is not enough to say that it is 'weak', and, as already mentioned, whether it is weak might depend on what use is made of it. The totality of the evidence may be such as to render a conviction unsafe. But that does not affect admissibility. And the jury may need to be warned that evidence, if accepted, only shows consistency of appearance between the person and the offender; a fact which may or may not be of much significance depending upon other matters. Evidence of blood sampling may be relevant and admissible, for example, even though, standing alone, it only establishes that it is consistent with the accused being the offender. Evidence may show that an accused was near the scene of a crime. Such evidence, on its own, does not show that the accused committed the crime. That does not mean it is of no probative value; in the end, it will have to be considered together with all the other admissible evidence.

For any one of a number of reasons, evidence of observations, including evidence of positive identification, may be made in circumstances which adversely affect its reliability. Those circumstances may be beyond anybody's control, or they may result, for example, from the way police have conducted an investigation. In Davies and Cody v The King this Court considered evidence of positive identification of an accused by a witness whose previous knowledge had not made him familiar with the accused, and who was first shown the accused, alone, as a suspect. The risk involved in identification made in those circumstances is obvious. The Court said :

'[I]f a witness whose previous knowledge of the accused man has not made him familiar with his appearance has been shown the accused alone as a suspect and has on that occasion first identified him, the liability to mistake is so increased as to make it unsafe to convict the accused unless his identity is further proved by other evidence direct or circumstantial. 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.'

That passage assumed the admissibility of the evidence, and accepted the possibility that, although standing alone the 'liability to mistake' of such evidence was apparent, in combination with other evidence, even other evidence of the same kind, it might sustain a conviction.

The decision of the New South Wales Court of Criminal Appeal in R v Bouquet was cited with approval by this Court in Alexander v The Queen. In Bouquet, the police had failed to conduct an identification parade, or to explain why one was not conducted, but had, instead, shown the victim a number of photographs, from which the victim selected a photograph of the appellant. The victim also made an in court identification of the appellant, both at the committal proceedings and at the trial. It was complained that the procedure adopted by the police in showing the photographs to the appellant was improper, and that the in court identifications were worthless. The failure to hold a line-up and the alternative procedure adopted was different from the course prescribed by police regulations. In that respect it was similar to what occurred at the Southport courthouse in the present case. The evidence, including the in court identifications, was held to be admissible. As to the photographs, Sugerman J said [29] :

'The use of photographs in this way, in lieu of a personal identification parade, goes to the weight and sufficiency of the evidence rather than to its admissibility ... '

Of all forms of identification evidence, one of the most notoriously dangerous is in court identification, which is usually performed in circumstances that strongly suggest the answer that is ultimately given. Even here, however, there is no absolute rule requiring rejection of such evidence; and there may be circumstances in which it is appropriate to allow it. In Alexander, Mason J discussed in court identification, which he said was 'of little probative value', in terms that accepted its admissibility. He went on to say: '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.'

The actual decision in Alexander was that, in a case where no identification parade was held, and witnesses, following the arrest of a suspect, identified him from photographs shown to them by police, the evidence of such photographic identification was admissible. Gibbs CJ said:

'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 of opinion that the evidence had little weight but was likely to be gravely prejudicial to the accused ... 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.'

It may be noted that the wording of the grounds of appeal in the present case is consistent with what was said by Gibbs CJ. The complaint is that the failure to exclude the evidence resulted in a miscarriage of justice.

...

There are two principal dangers associated with identification by means of selection from a group of photographs. These were discussed in Alexander. There is the inherent risk of error associated with suggestibility, and what is sometimes called the displacement effect. But there is also a risk of a different kind. The fact that the police have photographs of a suspect might convey to the jury the message that the suspect is a person with a criminal history. A similar risk arises where identification is made in circumstances suggestive of a criminal background, such as where a person is asked to attend a police station and look at a number of people reporting in compliance with bail or parole conditions. his is sometimes called the rogues' gallery effect. Because of the evidence as to the circumstances in which the photo-board shown to Mr Hill was prepared, that is not an issue in the present case. The first kind of risk concerns the probative value of the evidence. The second is a risk that the jury will draw an inference about a fact which, even if true, would ordinarily be excluded from evidence. In that connection, some care is needed in the use of the term 'prejudice'. Where it is present, a risk of the second kind is clearly a risk of unfair prejudice. It is a risk that a fact will be suggested which is of a kind that is ordinarily excluded from evidence in the interests of fairness to an accused. But prejudice does not arise simply from the tendency of admissible evidence to inculpate an accused. It is unfair prejudice that is in question. Where evidence is relevant and of some probative value, prejudice might arise because of a danger that a jury may use the evidence in some manner that goes beyond the probative value it may properly be given. If there is relevant prejudice of that kind, it lies in the risk of improper use of the evidence, not in the inculpatory consequences of its proper use. If it were otherwise, probative value would itself be prejudice. All admissible evidence which supports a prosecution case is prejudicial to an accused in a colloquial sense; but that is not the sense in which the term is used in the context of admissibility.

The evidence of the four witnesses named in grounds 1 and 2 was of some probative value. However, the trial judge had a discretion to reject it, in the interests of fairness to the appellant, if he concluded that its probative value was outweighed by the danger of unfair prejudice to the appellant. He was invited to exercise that discretion, but declined to do so. That was a decision that was open to him in the circumstances of the case, and his discretion has not been shown to have been affected by material error, or otherwise to have miscarried. And there has not been shown to have been a miscarriage of justice.[171]

264 In the present case, the trial judge rejected the application to exclude the identification evidence in the exercise of the Christie discretion. His Honour concluded that this evidence had significant probative value, and that this was not outweighed by any likely prejudicial effect.

265 With regard to the evidence of Mrs Burman, I agree entirely with his Honour's conclusion. That evidence was extraordinarily powerful. The same could not be said of the evidence of Mr Weller and Mrs Melnik. Nonetheless, it was open to the trial judge to decide that this evidence should be received, subject to the jury being given appropriate directions.

266 In the final analysis, the applicant bore the burden of persuading the trial judge that the Christie discretion should be exercised in his favour. I am not satisfied that his Honour erred in holding that that burden had not been discharged.

267 In addition, this ground of appeal involves a challenge to the exercise of a discretion. Accordingly, the principles laid down in House v The King[172] are applicable. As a general proposition, an applicant who challenges the exercise of a discretion must establish that the trial judge acted upon a wrong principle, or allowed extraneous or irrelevant matters to guide him, or mistook the facts, or failed to take into account some material consideration.[173]

268 There is nothing to suggest that his Honour fell into any error of that kind. It is clear that he applied the correct test in exercising his discretion. Indeed, so much was expressly conceded before us. It was not suggested that he took into account any irrelevant matter or that he failed to take into account some material consideration. In substance, the only argument advanced before this Court was that his Honour ought to have arrived at a different conclusion.

269 Absent any error of principle, and absent also any suggested mistake as to the facts, a challenge to the exercise of a judicial discretion clearly faces considerable difficulties. In my view, it was well open to his Honour to admit the identification evidence in this case. Ground 2 must therefore fail.

Ground 6: Failure to Warn Properly as to Identification Evidence

270 There can be no doubt that the identification evidence that was led in this trial was an important part of the Crown case. Accordingly, the trial judge was required to give the jury a strong warning of the dangers associated with evidence of that kind. He was also required to give that warning with the 'full authority of his office'.[174]

271 The evidence of identification took various forms. The most important witness, in that regard, was plainly Mrs Burman. She said that on the morning of 1 November 1997, which was a Saturday, she had gone to the Fawkner Cemetery, standing in for her son who regularly did voluntary work tidying up the area. She entered the cemetery through the main entrance in Sydney Road and went immediately to the Latvian section. She remained at the cemetery from about 8.00 am to 2.00pm.

272 At one stage, at about an hour to an hour and a half after she had first arrived, while she was cleaning the area, she saw a man. He was kneeling down at a grave about three or four plots away from her. The man approached her and asked if she worked there. She told him that she did. He said that he had just found the grave of his 'adoptive mother', whom he had 'never seen before'. He offered to show her its location and she went with him to see it. He complained that the top of the grave was untidy and asked her whether she would 'pull out all the flowers and whatnot'. She told him that she could not do that, but lent him a rake so that he could tidy it up himself.

273 Their conversation continued for some time. He told her that he had found his adoptive mother's grave through a 'family tree'. He added that he had located it through the cemetery office where he had seen the plans. They sat together on a bench and spoke for a time. He seemed very upset. She asked him if she could help him or, through her church, locate other relatives. She said that she would take down his name and telephone number, which she did. She recorded it in a diary. He gave the name 'John Roberts' and a telephone number, which, it later transpired, had one digit missing.

274 While they were seated on the bench, the man asked her twice whether she knew what was behind a hedge situated nearby. Eventually she returned to her own work. However, his story of never having seen his 'adoptive mother', and the other matters they discussed, caused her to be highly suspicious of him. She approached him again, and asked him to confirm that the grave at which he had been kneeling was indeed that of his 'adoptive mother'.

275 Mrs Burman described the man as having looked 'sort of teary and a bit disturbed'. She thought that he was in his late 30s. She said that he was about five foot seven inches tall, of medium build, and with fair hair. He had a fringe and his hair was of about collar length. He wore prescription glasses.

276 The man had on a light-coloured jacket which was half undone. Most of the time, his hands were in the pockets on the side of his jacket. She noticed that there was a rip near the right-hand pocket. He wore black tracksuit trousers with elastic at the bottom of the legs and white shoes. She last saw him either 20 minutes or about half an hour after they first met.

277 On the following day, Mrs Burman's husband told her about a news broadcast that he had heard that referred to a murder in the Fawkner Cemetery. As a result, she telephoned the police. They said that they would get back to her. Inexplicably, they did not do so.

278 The next time she spoke to the police was about a week later, when she returned to the cemetery. There was a police caravan there. She told them of the man she had spoken to on the day of the murder. On 15 April 1998, some five months or so later, she assisted a police expert in the production of a computer-generated image of that man. That image was tendered as Exhibit J.

279 On 6 May 1999, Mrs Burman was shown by the police a folder containing 12 separately numbered photographs (which became Exhibit K). She was asked if she could recognise the man that she had seen in the cemetery.

280 The transcript of her evidence regarding this matter reads as follows:

Do you recall now which photographs you referred to? --- Six, seven and 12.

What did you say about photographs six, seven and 12? --- They had similar hair, glasses.

Did you say anything in particular about any one of those three, six, seven or 12? --- Seven.

What did you say about No. 7? --- That I thought it was the most likely to be the same person.

281 Mrs Burman was then asked about a holiday that she had taken with her husband to Echuca in August 2000. She said that on Wednesday, 16 August, she and her husband went into a newsagency in Kyabram and saw that day's edition of the Herald Sun. On the front page, there was a photograph. She said that she recognised one of the persons shown in the photograph as the man she had seen at the cemetery.

282 Mrs Burman described how, when she saw the front page of the newspaper, she studied it carefully to make certain that it was the same person. She confirmed in her own mind that one of the two men shown in the photograph was indeed the man with whom she had spoken at length at the cemetery. The photograph that appeared on the front page of the newspaper was tendered as Exhibit L.

283 Under cross-examination, Mrs Burman confirmed that the photograph that she had picked out of the folder was indeed number 7 which was, of course, that of the applicant. She accepted that she might not have told the police, when she first saw that photograph, that it was most likely to be that of the man she had seen at the cemetery. She insisted, however, that that was her firm belief at the time. She recalled that she had told the police that the glasses worn by the man she had seen were different from those shown in the photograph because they were not tinted. She agreed that she had previously said, on another occasion, that she could not be '100 per cent' certain that the man depicted in the photograph was in fact the man with whom she had spoken at the cemetery.

284 Mrs Burman said that she had found the conversation with the man in the cemetery to be somewhat strange. For that reason, she recalled it clearly. She was questioned at considerable length regarding the description that she had given to the police. She agreed that she had told the police officer who assisted in the preparation of the computer-generated image that it was a '65 per cent likeness' of the man that she had seen.

285 Turning to the photograph that she saw in the Herald Sun on 16 August 2000, Mrs Burman agreed that it showed the applicant with handcuffs on, and being held securely by a man to his right. She agreed that it appeared that the applicant was under arrest. She also acknowledged that above the photograph appeared the words: 'Revealed, killer convicted yesterday is also the prime suspect in Melbourne's graveyard murder mystery.' A photograph of Ms Halvagis also appeared on the front page.

286 There followed lengthy questioning designed to elicit from Mrs Burman a concession that the photograph in the newspaper was so suggestive of the applicant's identity as the killer of Ms Halvagis as to render her entire identification worthless. She rejected that contention. She said that the applicant was 'definitely the person I spoke with in the cemetery'. She said that she was convinced of that fact without recourse to anything that the newspaper report contained. She insisted that her identification of the applicant was accurate. She maintained that she had been certain of her identification when she picked out photograph 7 in May 1999, although she conceded that she might not have made that sufficiently clear to the police at the time.

287 In re-examination, Mrs Burman reiterated her certainty that she was not mistaken.

288 The next witness who gave evidence of having seen the applicant at the Fawkner Cemetery on 1 November 1997 was Mr Weller. He had gone there with a friend, Patricia Rodriquez, who wished to visit her husband's grave.

289 Mr Weller said that he had been to the cemetery on previous occasions. On this day, they arrived at approximately 3.00pm. They went to the Roman Catholic section, where Mrs Rodriquez's husband was buried. According to Mr Weller, they saw a man who, for whatever reason, was staring at them from behind a bushy tree. He was described as having worn a dark-coloured parka and dark trousers. Mr Weller looked closely at the man, who was about three metres from him, noting that he wore silver-rimmed glasses. He appeared to be about 45 to 55 years of age, and slightly more than five foot seven inches tall. His hair was ginger coloured. He was of medium build, and had a 'bit of a belly'.

290 Shortly after that visit to the cemetery, Mr Weller was watching television when he saw the applicant featured on a news item. His evidence was that he 'straight away realised that's a face I see in the cemetery'. He added: 'as soon as I saw his face on TV, I realised ... I know that person, I know that face.'

291 On 22 March 2005, more than six years after Ms Halvagis's murder, Mr Weller was shown the folder of photographs (Exhibit K) by the police. He was told not to assume that the person in whom the police were interested was in that collection, but asked if he could recognise anyone. He said that he immediately pointed to photograph 7. He said that he was quite certain about the accuracy of his identification.

292 Under cross-examination, it was suggested to Mr Weller that he might be mistaken as to the date on which he saw the man at the cemetery. He insisted that it had been 1 November 1997, a date he recalled clearly because it was the Saturday a week or so after his father's death. He was asked why, having shortly afterwards seen the applicant on television, and having recognised him as the man at the cemetery, he had not immediately contacted the police. He said that he had been advised by his family and friends not to get involved.

293 Much of the cross-examination was directed towards casting doubt upon the reliability of Mr Weller's identification. It was suggested that, having seen the applicant on television, and in the newspapers, his identification of photograph 7 was worthless, because it was affected by what is commonly described as 'displacement'. However, Mr Weller insisted that he had had a clear view of the man, and remembered him because of his untoward conduct in staring at Mr Weller and Mrs Rodriquez for no particular reason. He said: 'I know that face, that's the face I saw in the cemetery.'

294 Mrs Melnik was the third and last of the key identification witnesses. Her evidence was that, on 1 November 1997, she had gone to the Fawkner Cemetery with two friends in order to tend to the grave of her parents who were buried in the Ukrainian section. She said that they had arrived at the cemetery at about 3.15pm. They had initially gone to the Ukrainian section, but then had moved to the Greek Orthodox section because that was where one of her friends wished to go. At that point, Mrs Melnik saw a man who appeared confused. He was wearing a grey or washed-out brown top. He appeared neatly dressed. He wore glasses. She noted that he kept his hands in his pockets. She said that his hair, which was dark blonde, appeared neat. She saw the man again shortly afterwards approaching a small white monument.

295 On the following day, Mrs Melnik heard the news that a girl had been killed in the cemetery. Sometime later, she saw a television news broadcast, featuring the applicant, in which it was said that he had killed two other women. She said that after thinking about the matter, she recognised him as the man she had seen at the cemetery. The two friends who had accompanied her to the cemetery had since died.

296 Under cross-examination, Mrs Melnik agreed that the first time she had made a statement to the police regarding this matter was in April 2005. Notwithstanding the passage of time, she insisted that her memory was good, and that she had not forgotten anything associated with the sighting of the man some seven or so years earlier.

297 It was put to Mrs Melnik that she might have been influenced in her identification by her knowledge that the man she saw on television had murdered two other women. She denied that, insisting that her identification was accurate.

298 I turn next to the trial judge's directions to the jury regarding the dangers of mistaken identification. His Honour said:

Next, ladies and gentlemen, I direct you as a matter of law on the question of identification evidence and this being a direction of law, also binds you. You know that in this case, there are three identification witnesses relied upon by the prosecution, Mrs Burman, Mr Weller and Mrs Melnik ... They are the witnesses, Mrs Burman, Mr Weller and Mrs Melnik, that the prosecution relies upon as identification witnesses.

I give you these directions of law, ladies and gentlemen, which bind you. You must be cautious before you act on identification evidence and you must be very careful in considering that evidence. You all know from your everyday experience, ladies and gentlemen, that honest people can be wrong in identifying someone. That is also, ladies and gentlemen, the experience of the law and I tell you that fact, that honest people can be wrong in identifying persons and in the past, have been wrong in identifying persons. Further, because very often identification witnesses are honest and the question is not whether they are honest but whether they are reliable and accurate, that evidence can be seductive in the sense that they can look good but they can be wrong. So the issue with these identification witnesses is not, are they honest or not, but are they accurate and reliable or not? That is the question for you, ladies and gentlemen, not for me, not for the barristers but for you and as I say, you must be cautious before you act on identification evidence. The experience of the law shows that and your experience of life would also show that and you must be very careful in considering identification evidence.

I will give you two further warnings, ladies and gentlemen. You must not add some outside fact to an identification witness's evidence and use that outside fact to bolster the identification witness's evidence. You look at the identification evidence itself as identification. Further, you must not add another identification witness to the one you are considering and use that other identification witness to bolster the one you are considering. Again, you examine the identification witness on a stand alone basis, are they accurate and reliable, or not? This is a further direction of law but I'm sure it coincides with your experience of life, ladies and gentlemen. In considering identification evidence, you consider all the factors which you consider could bear upon the accuracy and reliability of identification evidence. Obvious ones are the state of the lighting, the distance between the witness and the person observed, the length of time of the observation, the reason or the lack of reason for a witness to notice the other person or to notice their features, the fact that the witness had not seen the other person before and did not therefore recognise them, the lapse of time between the event and having to recall it and the circumstances of when the witness did recall it and in particular, whether in the circumstances of recall, not only was there a lapse of time and how much, but were there any interfering factors in the circumstances of recall; what technically is called displacement. That simply means that later images intervene between the event of the recollection and displace the original image in your mind and you remember the new image, not the original image because it has been displaced. They are all factors you look at, ladies and gentlemen, in considering identification evidence. That is the experience of the law and I am sure that is equally your common sense and experience of life, ladies and gentlemen.

Of course, once you have properly exercised caution and care in assessing identification evidence, you may act on the evidence as indeed the prosecution has submitted you should but always approach it with caution and care, ladies and gentlemen, bearing in mind what I have said.

299 In addition to providing this warning, the trial judge reminded the jury in considerable detail of the evidence given by each of the three identifying witnesses. In relation to Mrs Burman, he also reminded the jury of the evidence given by the police officer who, in April 1998, created the computerised image of the man she had seen at the Fawkner Cemetery.

300 I interpolate at this stage to say that it would be surprising if the jury did not regard Mrs Burman's evidence as much more cogent than that of Mr Weller and Mrs Melnik. Mr Weller claimed that he recognised the applicant as the man he had seen at the cemetery within a day or so of the murder, but did not make a positive identification from the folder of photographs until many years later. Nonetheless, it was not entirely without significance that he was able to pick out a photograph of the applicant from among a number of others, all of which showed men of quite similar appearance. Mrs Melnik, on the other hand, was not shown the collection of photographs. It is unlikely that the jury would have given her evidence great credence.

301 Mrs Burman's evidence was, as I have said, extraordinarily powerful. First, she had ample opportunity to observe from close range the man that she saw at the cemetery. She spoke to him at considerable length. His remarks to her were plainly somewhat bizarre, and therefore likely to have impressed themselves in her memory.

302 Despite the fact that Mrs Burman saw the man some hours before the murder of Ms Halvagis, she was sufficiently concerned about the manner in which he had behaved to contact the police the very next day.

303 In April 1998, Mrs Burman was involved in the production of a computerised image of the man she had spoken to at the cemetery. I have examined that computerised image closely. In my view, it bears a remarkable resemblance to the applicant. Certainly, the Crown invited the jury to come to that conclusion. I would be surprised if they had not also done so.

304 It was just over a year later that Mrs Burman was shown the booklet of photographs. She picked out photograph 7, from a group of 12 photographs of men of similar appearance, as being that which most closely resembled the man at the cemetery. It is of particular importance to note that Mrs Burman selected that photograph long before she had seen any image of the applicant on television or in the newspapers.

305 It is true, in one sense, that Mrs Burman did not make a positive identification of the applicant at that stage. She merely told the police that photograph 7 was 'the most likely to be the same person'. She also observed that the men depicted in photographs 6 and 12 had similar hair and glasses.

306 Nonetheless, the net effect of her evidence was that by May 1999, she had come very close to making a positive identification of the applicant as the man she saw at the cemetery in circumstances that seem to me to have lent enormous credence to what she had to say. She had done so at a time when she could not have been influenced, consciously or otherwise, by photographs of the applicant of the kind that were later displayed in the media.

307 The photograph that was published in the Herald Sun newspaper in August 2000, which Mrs Burman saw at that time, showed the applicant having been arrested, under the banner headline 'Monster'. It intimated that he was guilty not just of the two murders of which he had by then been convicted, but also that of Ms Halvagis.

308 Of course, once Mrs Burman had seen that newspaper article, with the accompanying photograph, there was ample scope for any subsequent identification that she made to be the subject of strong criticism. It could be argued that she was likely, thereafter, to have been influenced, at least subconsciously, towards identifying the applicant as the man at the cemetery. There was also the danger of displacement in any subsequent identification, based not just upon that photograph, but also upon her earlier identification of photograph 7.

309 Nonetheless, it must be said that Mrs Burman's evidence at the trial was forthright and impressive. When it was suggested to her, in cross-examination, that the Herald Sun article, which she had seen in August 2000, had caused her to 'firm up' her evidence of identification, she was emphatic in her rejection of that proposition. Indeed, she insisted that even at the time she picked out photograph 7, she had been certain of her identification. The jury had the inestimable advantage of being able to assess her credibility in that regard.

310 It is against this background that I turn to the adequacy of the trial judge's directions to the jury regarding identification evidence. Much of the law concerning this subject has developed in relation to what is described as evidence of positive identification of persons previously unknown. Such evidence differs from evidence of recognition,[175] and the principles that govern that form of identification.[176]

311 There is another category of what may be loosely termed 'identification evidence'. A witness may be unable to say no more than that the accused resembles the offender.[177] Such evidence is often described as 'similarity evidence'. Although evidence of this kind can assist in identifying an offender, the purist would say that it is not identification evidence in the strict sense. The witness does not positively identify the accused as the person previously seen, but merely says that he or she 'looks like' the offender.[178]

312 Another name for this type of evidence is 'circumstantial identification evidence'.[179] However, this ought not be confused with evidence of the more usual kind which simply permits an inference to be drawn that the accused committed the offence in question.

313 In the present case, there was 'similarity evidence' as well as evidence of 'positive identification'. The computerised image generated by Mrs Burman was 'similarity evidence'. So too, on one view, was her selection of photograph 7 as that which most resembled the man she saw, although her insistence that she had been certain at the time might be regarded as evidence of positive identification. Clearly, her evidence as to having recognised that man in the Herald Sun article was evidence of 'positive identification'. Both Mr Weller and Mrs Melnik gave evidence of 'positive identification'.

314 The authorities make it clear that where there is evidence of positive identification, and that evidence is significant in the context of a trial, the trial judge must instruct the jury as to how such evidence should be approached. Identification evidence of this kind is seen as inherently dangerous.

315 In Alexander v The Queen,[180] Mason J said:

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.[181]

316 The problem is that, despite its inherent fragility, positive identification evidence may be seen by a jury as particularly impressive, especially as it is often given by witnesses who appear both honest and convincing. Judicial experience has shown that such witnesses can be mistaken. Such evidence has led to miscarriages of justice in the past.[182] A warning is therefore necessary to prevent jurors, who may not fully appreciate this danger, from giving too much weight to evidence that may be flawed. These days, it is often referred to as a Domican warning.[183]

317 The warning must be strong and complete. It must be given even if a conviction could not be based on identification alone as, for example, where the identification, if accepted, would still provide only circumstantial evidence of guilt.[184] A warning must also be given even if there is ample additional evidence upon which the accused can be convicted and even if there is more than one identifying witness.[185] Mistakes can occur where two or more witnesses have each made positive identifications.[186]

318 That said, it is clearly understood that the terms of any direction must depend, to a considerable degree, upon the particular circumstances of the case. A trial judge's charge need not follow any rigid formula, but must be geared towards alerting the jury to any dangers lurking in the evidence that they might not otherwise appreciate. The strength of the warning that is required may depend to some degree upon the extent to which the Crown relies on the identification evidence. A stronger warning may be required when the witness had no previous acquaintance with the accused, or only a passing opportunity to make an observation.

319 It is important to understand that Domican does not require every conceivable aspect of identification evidence to be the subject of a warning. It is also useful to remember that in Festa, the High Court reminded trial judges that they ought be careful to ensure that any warnings that they gave did not rob identification evidence of all probative value.[187]

320 In the present case, the trial judge was obliged, as I have said, to warn the jury appropriately of the dangers of acting too readily upon the evidence of the three identification witnesses.

321 There is no doubt that the general directions that his Honour gave as to the dangers associated with mistaken identification met all the requirements laid down in Domican. The jury were told, in the clearest terms, that they had to scrutinise the evidence of identification in this case with special care, and to exercise particular caution before accepting it as reliable. They were told the reasons for this warning. They were directed to be aware of the risk that even honest and convincing witnesses might be mistaken in their evidence of identification.

322 The jury were also reminded to examine closely the circumstances in which each identification was made, including factors such as lighting, distance, and any obstacles that stood in the path of clear observation.

323 The cases say that where more than one witness gives identification evidence, the jury should be told that they still can all be mistaken.[188] Indeed, there is some authority for the proposition that it may be necessary, in such cases, to warn the jury that two unsatisfactory identifications do not support each other.[189]

324 However, the trial judge in the present case went much further than this. As has been seen, he instructed the jury that, as a matter of law, the evidence of each identifying witness had to be considered and evaluated in isolation, and without paying any regard to the evidence of any other identifying witness. As I have indicated, there is some authority for the proposition that this direction may have

been unduly favourable to the applicant.[190] The cases go no further than to suggest that the jury should be told that one unsatisfactory identification cannot support another. I should add that even that proposition has been doubted.

325 The trial judge directed the jury that not only did they have to consider each act of identification on its own, without regard to other acts of identification, but that when considering the accuracy of any single identification, they could not look to any circumstantial or other evidence that might support it. That direction certainly seems to have been unduly favourable to the applicant.[191] It was not supported by authority and I regard it as contrary to principle.

326 Domican establishes that the general directions of the kind that his Honour gave in this case must be backed by judicial authority. It is not sufficient simply to refer to arguments already advanced by counsel. The warnings must carry the weight of judicial office.

327 As regards the general directions given, the trial judge told the jury on more than one occasion that these were directions of law, and binding upon them. No criticism has been levelled at this aspect of his Honour's charge. Nor could it be.

328 The real point upon which this ground of appeal rests has nothing to do with the general directions that were given. It is said that his Honour's charge was deficient for one reason and one reason only, namely, that it failed to identify for the jury the factors that might affect the reliability of the identification evidence in the particular circumstances of this case.[192]

329 It was submitted on behalf of the applicant that Domican establishes that, in every case, a trial judge must isolate and identify any matters of significance which might reasonably be regarded as undermining the reliability of identification evidence, and draw attention to any weaknesses in the particular identification evidence given. It was submitted that his Honour had failed to do so.

330 It should be noted that there is at least one case, decided after Domican, which holds quite specifically that a trial judge is not obliged to point out every possible weakness that may be associated with identification evidence which is under challenge.[193] That case suggests that whether or not a particular weakness need be noted, and commented upon, will depend upon the facts of the individual case.

331 It was submitted that the trial judge's charge to the jury in the present case failed to remind them of at least two weaknesses that were said to be associated with the identification evidence that was led. These were, first, the delay between the sightings at the cemetery and the eventual positive acts of identification and, second, the 'suggestibility' associated with the manner in which those later acts of identification were done.

332 Bearing in mind that the jury must be instructed 'as to the factors which may affect the consideration of [the identification] evidence in the circumstances of the particular case',[194] and that their attention should be drawn to any weaknesses in the identification evidence[195] (with reference to counsel's arguments being insufficient to achieve that end),[196] the question now before this Court is, did his Honour's charge adequately meet these requirements?

333 In approaching that question, it must be remembered that Domican itself was a case in which the identification evidence was of a highly unsatisfactory nature. At the time that her husband was shot, Mrs Flannery, the sole identifying witness, had never met the appellant. Nearly nine months had passed before she formally identified him from a collection of photographs. Those photographs had been altered to show the appellant wearing a wig and false moustache in conformity with the description given of the person who fired the shots. By the time she saw the photographs, not only was the appellant a definite suspect, but Mrs Flannery had seen him on television a number of times. She had also allegedly seen him in the vicinity of her home.

334 There were other weaknesses associated with Mrs Flannery's identification. When she first saw the gunman, he was some distance away. Her vision was obscured and she had only a fleeting moment to observe him. Moreover, he was clearly disguised. Her first sighting of him took place after about 30 shots had been fired in her direction, and while she was in a state of shock.

335 To make matters worse, there were significant discrepancies between the description that Mrs Flannery gave to the police in her first statement to them, and what she later said about the offender's appearance. She was wrong about the make and colour of car that he was driving and the number of people involved in the shooting.

336 It was hardly surprising, in these circumstances, that the High Court held that the general warning given to the jury in that case regarding the dangers of identification was not sufficient. It did not bring home to the jury any of the real dangers associated with Mrs Flannery's identification.

337 It is against that background that one needs to understand the comments in Domican. In a passage from the joint judgment of Mason CJ, Deane, Dawson, Toohey, Gaudron and McHugh JJ (upon which the applicant largely relies), their Honours said:

Whatever the defence and however the case is conducted, where evidence as to identification represents any significant part of the proof of guilt of an offence, the judge must warn the jury as to the dangers of convicting on such evidence where its reliability is disputed. The terms of the warning need not follow any particular formula. But it must be cogent and effective. It must be appropriate to the circumstances of the case. Consequently, the jury must be instructed 'as to the factors which may affect the consideration of (the identification) evidence in the circumstances of the particular case'. A warning in general terms is insufficient. The attention of the jury 'should be drawn to any weaknesses in the identification evidence'. Reference to counsel's arguments is insufficient. The jury must have the benefit of a direction which has the authority of the judge's office behind it. It follows that the trial judge should isolate and identify for the benefit of the jury any matter of significance which may reasonably be regarded as undermining the reliability of the identification evidence.

...

Unfortunately, the learned trial judge's directions were not likely to impress upon the jury that the evidence of identification by Mrs Flannery had a number of weaknesses. His Honour told the jury that '(s)udden and unexpected acts of violence such as Mrs Flannery described in this case, can affect people caught up in the events in different ways.' He said that the terror of the occasion 'can serve to impress indelibly on the minds of some people the features of anyone they see involved in it. With other people the effect may be to obscure their judgment and their later recollection.' He also raised the possibility of honest mistakes being made in identification and the need for caution when dealing with this type of evidence. But apart from these statements, his Honour's directions did not refer to any matters concerning Mrs Flannery's evidence which would bring the attention of the jury to the many weaknesses in her evidence. To the contrary, some parts of his Honour's directions were capable of conveying the impression that her evidence of identification was reliable. After pointing out to the jury that '(w)hat it is important to know is whether the identification was aided by any suggestion made to the witness or any expectation on the part of the witness at the time when it was made', his Honour recited the evidence of Mrs Flannery concerning the steps by which she came to identify the appellant. His Honour then said:

'This is a case you might think where it was a recognition and realisation by the witness that 'that is the man' when it wasn't a circumstance in which there was any particular reason for her to believe that it was. It was a chance occurrence that she saw this man.'

This passage may have been understood by the jury, as it was probably intended to be understood, as meaning that Mrs Flannery's identification was more reliable than usual because she had no expectation that the person in the milk truck, in the orange Ford Falcon sedan and on television was the person who had shot her husband.

It is true that on two occasions in his summing up his Honour referred to the arguments of counsel for the appellant concerning the weaknesses in Mrs Flannery's evidence of identification. Thus, his Honour referred to an argument that the evidence of a surveyor showed that Mrs Flannery 'would not have had the clear view of the person sitting in the car that she said she had'. His Honour also said:

'As to Mrs Flannery, it was put by the defence that the circumstances in which she said that she saw the person with the gun in the car make her identification unreliable. The shock, her fear and the fact that on her own account the man was disguised, would make her identification unreliable It was also put that quite apart from the risk of genuine error, she may have believed that the shooter was the accused because of something Mr Eustace told her and so, the argument runs, she would have a reason for deliberately lying. You are also invited by the defence to reject her purported identification because she did not name the accused until September 1985 after previously telling the police from her first interview in May 1985 that she could not identify the person in the car.'

However, as we have already pointed out, mere repetition of counsel's arguments is an insufficient discharge of the trial judge's duty to draw the jury's attention to any weaknesses in the identification evidence. If the matters to which counsel has referred may reasonably be regarded as undermining the reliability of the identification evidence, the trial judge must direct the jury that they are bound to take those matters into consideration in determining whether they will rely on that evidence. Consequently, the learned trial judge's directions on the issue of identification were not adequate and constituted a misdirection.[197]

A trial judge is not absolved from his or her duty to give general and specific warnings concerning the danger of convicting on identification evidence because there is other evidence, which, if accepted, is sufficient to convict the accused. The judge must direct the jury on the assumption that they may decide to convict solely on the basis of the identification evidence. If a trial judge has failed to give an adequate warning concerning identification, a new trial will ordinarily be ordered even when other evidence makes a very strong case against the accused. Of course, the other evidence in the case may be so compelling that a court of criminal appeal will conclude that the jury must have convicted on that evidence independently of the identification evidence. In such a case, the inadequacy of or lack of a warning concerning the identification evidence, although amounting to legal error, will not constitute a miscarriage of justice. But unless the Court of Criminal Appeal concludes that the jury must inevitably have convicted the accused independently of the identification evidence, the inadequacy of or lack of a warning concerning that evidence constitutes a miscarriage of justice even though the other evidence made a strong case against the accused.[198]

339 Domican represents the most authoritative statement of the law in this area. It has, of course, been routinely followed by trial and intermediate appellate courts throughout this country.[199] A number of convictions have been quashed where its stringent requirements have not been met. However, as Domican also makes clear, there is no set formula to be followed when directing a jury as to the dangers associated with identification evidence. As I have said, it is still the law that the nature of the warning to be given in any particular case must be tailored to the particular circumstances of that case.[200]

340 J D Heydon, the learned author of Cross on Evidence, has suggested, in his discussion of Domican, that the warning will usually take the form of pointing out that the evidence of identification is critical (assuming that it is), that mistakes as to identification have occurred in the past, as the law recognises (but juries may not), and that if there is only one witness as to identification, the jury must be careful with regard to that witness's accuracy. The trial judge will also explain the reasons why such evidence may be unreliable, and the jury reminded that witnesses may be honest, and convincing, but still mistaken. The trial judge will also usually discuss the opportunities of identification available to the witness, including matters such as distance, light, and possible exposure to stress.[201]

341 When considering the challenge mounted to the adequacy of his Honour's charge in the present case, it is useful to bear in mind that each of the three identifications had its own potential weaknesses.

342 In relation to Mrs Burman, the only factor that might have cast any significant doubt upon her reliability was the risk of undue suggestion as a result of her having seen the Herald Sun article. However, as I have previously noted, that article had no bearing upon her earlier acts of identification. In addition, the circumstances under which she saw the man at the cemetery were as near to optimum as could be contemplated.

343 Putting to one side the strength of her evidence arising from the computer-generated image, and the identification of photograph 7, there is yet another factor that should be taken into account when considering this ground of appeal. That is Mrs Burman's evidence regarding the jacket that was worn by the man that she saw.

344 Mrs Burman described that jacket as having been of a 'light colour'. She said that she had observed a 'rip near the pocket' on the right-hand side. She had earlier given the same description to the police officer responsible for producing the computer-generated image. That image shows the jacket as being ivory or bone coloured.

345 There was evidence at the trial from a Detective Keighley who, on 22 April 1999, had searched a workshop at the rear of the applicant's house. He said that he had found there, on the bottom shelf of a cabinet situated against the northern wall, a rolled-up jacket. That jacket had, in fact, been worn by the applicant several days earlier when he murdered Nicole Patterson.

346 The jacket was tendered in evidence, over objection. Counsel for the applicant maintained that it was olive green. However, the prosecutor submitted that it was 'a sort of a khaki colour', and that it could be regarded as 'light coloured', consistent with Mrs Burman's description. Of critical importance was the fact that the jacket had a rip near the right lower pocket, exactly as she had described.

347 The trial judge permitted the jacket to be tendered. His Honour described the jacket as 'highly relevant as a matter of proof in the circumstantial case'. He ruled that its probative value far outweighed any possible prejudicial effect.

348 In substance, therefore, the jury were invited to consider the significance of a jacket found in the applicant's home that was said to match closely the description given by Mrs Burman. They were asked to consider, in particular, the coincidence associated with there being a rip to the right-hand side pocket.

349 In his closing address to the jury, the learned prosecutor said:

Another piece of bad luck for Mr Dupas, if he is not the murderer of Mersina Halvagis, is the evidence that you heard from Mrs Burman, that the man she spoke to in the Latvian section of the cemetery, was wearing a light coloured jacket with a rip near the right hand pocket. The jacket, Exhibit S, found at the accused's home at 19 Coane Street, Pascoe Vale, has such a rip in that position, a rip or tear near the right hand side outer pocket. That jacket was recovered from the workshop, from a cabinet in the workshop of the accused's home, on 22 April 1999. Mrs Burman had described that as part of the description that she gave to Mr Hardiman when she did the face image on 15 April 1998. Mrs Burman has referred to the jacket, the rip. At that time it's not as if the police have recovered a jacket and then it's able to be seen; she made comment to Mr Hardiman, as you heard in the course of his evidence, about that as part of the description at the time the face image, Exhibit J, was created on 15 April 1998.

True it is, as you know, Mrs Burman referred to a light coloured jacket. Mr Hardiman said that he received instructions from Mrs Burman that it was a bone or similar coloured jacket, and it might be said well, the jacket which you have seen held up by His Honour's tipstaff, is not bone. It may be said it's not a particularly light colour, but it's certainly not a dark colour like a black, and I suggest that your experience of life tells you that different people have different recollections about colours, because every jacket is some colour or other, and that is an impression, but a rip or tear near the outer right hand pocket is an unusual feature, likely to be remembered, likely to be remembered more clearly and more completely, than mere colour. It is a jacket with a tear near the right hand pocket. Is it just bad luck that Mr Dupas has such a jacket? True it is it's recovered much later, but it's part of the whole circumstances. The person that Mrs Burman sees at the cemetery that morning has a jacket with such a rip or tear, and so does the accused. It's another piece of bad luck if Peter Dupas is not the murderer of Mersina Halvagis.

350 Counsel for the applicant, in his closing address, argued that the jacket was green and not bone. He reminded the jury that there was no blood or DNA from Ms Halvagis upon it. He criticised the Crown for not having shown the jacket to Mrs Burman to see whether she could identify it. However, having been reminded that he had not put to any of the police witnesses that this had been done for an improper purpose, he subsequently withdrew that suggestion.

351 The jury were reminded by his Honour, in his charge, of the various arguments put by the prosecution and the defence in relation to the jacket. As I have said, they were permitted to inspect it from a distance, but did not have access to it in the jury room.

352 The evidence concerning the jacket, and in particular the tear near the right pocket, seems to me to have been particularly powerful. It provided strong support for the accuracy of Mrs Burman's identification.

353 As I have indicated, when the High Court in Domican spoke of the consequences of failure to comply with the warning requirements therein laid down, their Honours did so against the background of a set of facts far removed from those in the present case. They said that if a trial judge failed to give an adequate warning concerning identification, a new trial ordinarily would be ordered. They added that this would be so even when other evidence made a very strong case against the accused.[202]

354 If I thought that there was even a remote chance that the trial judge's warning to the jury regarding identification in this case might not have brought home to them, with the full authority of his office, the particular weaknesses associated with the evidence of the identification witnesses, I would be obliged, in accordance with Domican, to hold, subject to the proviso,[203] that there should be a new trial.

355 I note that in R v Debs,[204] this Court invoked the proviso in answer to an attack upon the adequacy of a charge relating to identification, but said that the charge itself was deficient. The main difference between Debs and the present case lies in the fact that Mrs Burman's identification, at least prior to her having seen the Herald Sun article, was of a quality that was vastly superior to that of the identifying witness in that case.

356 In addition, unlike what occurred in Debs, the trial judge in the present case did warn the jury specifically of the dangers associated with what is described as 'displacement', a matter that was of critical importance to the identification in Debs, but entirely overlooked in the trial judge's charge in that case. In speaking of the dangers of displacement, the trial judge in the present case did so with the full voice of his judicial office, and not merely by reference to the arguments of counsel. Accordingly, at least one crucial aspect of Mrs Burman's evidence that needed to be addressed was the subject of a full and complete warning.

357 As I have said, it should also be borne in mind, when considering the adequacy of his Honour's charge, that not every potential weakness associated with identification evidence must necessarily be highlighted. For example, consider the question of delay. It is true that the trial judge did not specifically remind the jury that some years had passed between the events at Fawkner Cemetery, and the positive identifications made by Mr Weller and Mrs Melnik. However, the matter of delay would have been perfectly obvious. It had been the subject of extensive cross-examination. Perhaps more importantly, delay, and the effect that this can have upon memory, is not a matter upon which any jury would need special direction. It is a matter of common sense that memory does fade over time.

358 Counsel for the applicant put to each of Mr Weller and Mrs Melnik that they must have been influenced, in making their identifications, by having seen the applicant on television, at least in circumstances where he was known by then to have murdered two other women. He put to Mrs Burman that she had been similarly influenced, in arriving at her positive identification, by seeing the photograph in the Herald Sun.

359 Both the prosecutor and counsel for the applicant addressed at considerable length on these issues. Indeed, counsel for the applicant spent a good part of this address focussing upon the dangers of displacement, the issue that he obviously regarded as of paramount importance in attacking the reliability of the individual identifications.

360 Of course, it is not sufficient for a trial judge to do no more than refer to the arguments of counsel. However, in this case, his Honour went much further than that. He warned the jury, specifically and in terms, of the dangers of displacement. He did so as a matter of law, and with the full weight of judicial authority behind that warning.

361 It might be said that, in a perfect world, his Honour could have gone further. He could have added, in perhaps a sentence or two, that the risk of displacement of which he had spoken applied to Mrs Burman's identification of the applicant from the Herald Sun because she had earlier identified him from the folder of photographs that she had been shown. He could also have added, in relation to Mr Weller and Mrs Melnik, a sentence or two to like effect. However, in my opinion, the fact that he did not do so did not cause this trial to miscarry.

362 I am in no doubt that the jury in this case were well aware of the need to be particularly careful with the identification evidence. His Honour gave them all the assistance necessary to enable them to understand just why that was so. It is no part of Domican, as I read what the High Court said in that case, that an intermediate appellate court should go over every word of the trial judge's charge to see whether it matches up precisely with the many judicial pronouncements that have been made in relation to identification evidence. Rather, the charge should be read as a whole, having regard to the real issues in the trial. The points upon which the trial judge should focus are those upon which the jury are likely to need assistance because of the risk that, without an appropriate direction, they may give too much weight to evidence that is, for whatever reason, suspect or potentially unreliable.

363 In this case, the trial judge's charge complied sufficiently with the requirements laid down in Domican. There was no risk, in my view, that the jury would have acted upon questionable identification evidence.

364 Finally, I would add that his Honour's warning to the jury to consider the evidence of each identifying witness on its own, and only act upon that evidence if satisfied beyond reasonable doubt as to its accuracy, was, in my view, unduly favourable to the applicant. It is true that 'poor quality' identification evidence, considered in isolation, cannot be improved by other evidence. It stands or falls according to its own strength. However, what other evidence can do is bolster the Crown case, which includes the identification evidence said to be flawed. In other words, even if of poor quality, evidence of identification may, together with other evidence, assist the Crown case.[205] His Honour's charge may have led the jury to disregard evidence which was in truth capable of providing support for the Crown case as a whole. In that regard, it was too favourable to the applicant.

365 Once the jury accepted that it was indeed the applicant whom Mrs Burman encountered in the cemetery on the morning of Ms Halvagis' death, it was but a short step from there to a conclusion that it was he who was responsible for her murder later that day. Not only was he at the scene of the crime a comparatively short time prior to the commission of the offence but, if Mrs Burman's evidence were to be accepted, he was behaving in a bizarre fashion and one that raised deep suspicions in her mind.

366 There was also a good deal of 'post-offence' conduct on the part of the applicant that tended to support the conclusion that, if he was at the cemetery on that particular morning, he was in fact the murderer. There were firstly the lies that he told, in May 1999, when he said to Father Patrick O'Brien, a retired Catholic priest he had known for some years, that he had never been to the Fawkner Cemetery. Moreover, on the Crown case, he falsely told Father O'Brien that he had no relatives buried there when, in truth, the position was that his grandfather was buried at the cemetery. Indeed, his grandfather's grave was only about 100 metres from the grave of Ms Halvagis's grandmother, which she had been tending at the time of the murder.

367 There was also evidence that, in early November 1997, within just a few days of the offence, the applicant asked his hairdresser to alter his hairstyle. He told her that he intended to go to Queensland. In addition, Father O'Brien gave evidence that, on one occasion, he noticed that the applicant was not wearing his glasses. This was unusual, and he asked him why. The applicant told him that he had broken them at work. Other witnesses were called to say that, on 7 November 1997, the applicant had ordered a new pair of glasses. On that day, he was seen to have a fresh cut to the left side of his face. He said that it had happened at work. However, there was no record of any such injury in the book maintained for the purpose of keeping track of such matters at his place of work.

368 It follows that, provided that the jury accepted Mrs Burman's evidence that the applicant was at the cemetery on the morning of the crime, the evidence taken as a whole pointed strongly towards the conclusion that it was he who had murdered Ms Halvagis. That was so irrespective of what weight, if any, the jury gave to Mr Fraser's evidence.

369 Finally, I note that no exception was taken to any aspect of his Honour's charge on identification. That is a matter which I regard as being of considerable importance.[206] Identification was obviously one of the central issues in this trial. Counsel on both sides were obviously well prepared to deal with that issue, and had become well versed in the law on this subject. Had there been any concern, after the trial judge's charge, that the jury might not have been made fully aware of the dangers associated with mistaken identification, it is inconceivable that the point would not have been taken.

Grounds 5: Post-offence Conduct Amounting to Consciousness of Guilt

371 It was submitted on behalf of the applicant that the various acts of post-offence conduct upon which the Crown relied in support of its case, including, in particular, his change of hairstyle and purchase of new glasses, should have been the subject of a direction as to consciousness of guilt, in accordance with Edwards v The Queen.[207] I reject that contention.

372 The jury were given detailed and appropriate directions regarding lies said to have been told by the applicant as to his never having been to Fawkner Cemetery, and not having any relatives buried there. For that reason, ground 4, which complains of the directions given regarding lies, is in my opinion devoid of any substance. I note that both Nettle JA and Ashley JA share that view in relation to ground 4.

373 The question to be considered in relation to ground 5 is whether a direction similar to that given in relation to lies also had to be given in relation to other post-offence conduct.

374 This Court has, in the past, intimated that an Edwards direction, or something closely akin thereto, should be given in relation to post-offence conduct, other than lies, where such conduct is relied upon as consciousness of guilt.[208] Thus, in R v Ciantar,[209] a case involving a hit-run motor vehicle accident, and a charge of culpable driving, it was held that evidence of flight from the scene, and subsequent ingestion of alcohol, could be led as consciousness of guilt. However, such evidence was not to be left to the jury unless the Crown precisely identified how, together with the other circumstances in the case, it was said to demonstrate a consciousness of having committed the offence with which the accused was charged, and not some other offence.

375 Ciantar was followed some months later in R v Cuenco.[210] That case concerned lies said to have been told by an accused. However, the judgment speaks in broader terms, positing, on one view, that any post-offence conduct upon which the Crown relies as evidence of consciousness of guilt, including, but not limited to lies, should be subject to an Edwards direction.

376 Both Ciantar and Cuenco concerned lies. In so far as Ciantar dealt with flight, it did not hold, in terms, that an Edwards direction must, in every case, be given. It certainly did not hold that such a direction must be given in every case involving post-offence conduct relied upon as circumstantial evidence by the Crown. Cuenco was solely concerned with lies. Any observations by the Court in that case regarding post-offence conduct, apart from lies, were, strictly speaking, dicta. To the extent that Ciantar follows an earlier dictum in Renzella, and purports to extend the detailed requirements laid down in Edwards to cases not involving lies, it does so only in a specific and quite narrow context. It recognises that there may be circumstances in which post-offence conduct is incapable of being shown to be probative of guilt of the charged offence as opposed to a lesser alternative, or another count on the presentment. In accordance with Ciantar, such evidence should not be left to a jury as evidence of consciousness of guilt.

377 In my respectful opinion, neither Ciantar nor Cuenco is directly in point so far as the present application is concerned. The Crown's reliance upon the applicant's change of appearance shortly after Ms Halvagis' murder had nothing whatever to do with any lies ostensibly told by him.

378 Unlike Ciantar, there is no question in the present case of any of the evidence relating to post-offence conduct being treated as consciousness of guilt of some offence other than murder. The change of appearance was nothing more than circumstantial evidence, of the most ordinary kind, upon which the Crown relied in order to strengthen its case. Where, as happened in this case, the jury were correctly directed as to the onus of proof and as to the principles governing circumstantial evidence, it was unnecessary, in my view, for an Edwards direction also to have been given. Such a direction could only have confused the jury. Moreover, the need for any such direction was obviated by the fact that no such direction was sought.

379 It is an indisputable fact that the form in which the Edwards direction must now be given has led to enormous difficulty in the conduct of criminal trials in this State.[211] Such a direction, and the attendant problems associated with the related Zoneff warning,[212] should be confined, so far as practicable, to the very special difficulties that are normally associated with lies. The Edwards direction, which is specifically tailored to those very special difficulties, is not well suited to other, broader, categories of circumstantial evidence.

380 If I am wrong about this, and the law in this State as it now stands is that an Edwards direction must be given in relation to any post-offence conduct relied upon by the Crown, I would have no hesitation in saying that the failure to give such a direction did not give rise to a substantial miscarriage of justice. I would therefore reject ground 5.

[1] Domican v The Queen [1992] HCA 13; (1992) 173 CLR 555, 562 (Mason CJ, Deane, Dawson, Toohey, Gaudron and McHugh JJ) (citations omitted).

[2] [2005] VSCA 66 [267].

[3] See, for example, R v Mendoza [2007] VSCA 120 [45] (Vincent JA).

[4] [2005] VSCA 66.

[5] See [63] below.

[6] [1981] QB 720.

[7] [1993] HCA 63; (1993) 178 CLR 193, 210-211 (Deane, Dawson and Gaudron JJ).

[8] [2006] VSCA 263; (2006) 16 VR 26, 55 [103] (Warren CJ, Chernov, Nettle, Neave and Redlich JJA).

[9] [1993] HCA 63; (1993) 178 CLR 193.

[10] See Woolley v R (1989) 42 A Crim R 418, 423 (Murphy, Marks and Gobbo JJ); R v Ciantar [2006] VSCA 263; (2006) 16 VR 26, 49 [74]-[87] (Warren CJ, Chernov, Nettle, Neave and Redlich JJA).

[11] R v Cuenco [2007] VSCA 41; (2007) 16 VR 118, 123 [15]-[16] (Nettle JA).

[12] [1992] HCA 35; (1991) 174 CLR 558.

[14] Which Deane J identified as being: (1) such evidence is easily concocted; (2) an accused is ordinarily denied corroboration of his denial of the alleged confession; (3) a prison informer is liable to be of bad character; (4) a prison informer may be motivated to fabricate evidence in the hope of deriving some benefit in terms of sentence, treatment or release on parole, or because of a variety of pressures which may easily arise in a prison but not be immediately apparent to a jury; (5) in that case, the informer claimed to remember the alleged confession 'nearly word perfect' years after the event; (6) in that case, the informer had an extraordinarily bad criminal record, was serving a life sentence, was at least partly motivated by the desire to receive favourable treatment from the authorities, and had received some.

[22] Domican v The Queen [1992] HCA 13; (1992) 173 CLR 555, 564 (Mason CJ, Deane, Dawson, Toohey, Gaudron and McHugh JJ).

[23] Though subject to the possible operation of the proviso.

[24] [2007] HCA 12; (2007) 231 CLR 260, 300 [103]-[105] (Gleeson CJ, Gummow, Heydon and Crennan JJ).

[25] Cf M v the Queen [1994] HCA 63; (1994) 181 CLR 487, 502 (Brennan J).

[26] [1992] HCA 16; (1992) 173 CLR 592.

[29] He abandoned ground 3, which complained that the evidence of the witness O'Brien had wrongly been ruled admissible.

[30] R v Glennon (No.2) [2001] VSCA 17; (2001) 7 VR 631, 657-658 (Winneke P, Ormiston JA).

[32] Published variously in 96 editions of the 'Herald Sun', 'The Age' and 'The Sunday Age' and two editions of 'The Australian'.

[33] [1992] HCA 16; (1992) 173 CLR 592, 605. His Honour misstated the source of the passage which he included in his ruling; but nothing turns on it.

[34] [1989] HCA 28; (1989) 167 CLR 94.

[35] [2003] QCA 77; (2003) 138 A Crim R 103.

[36] [2004] NSWCA 324; (2004) 61 NSWLR 344.

[37] (1984) 14 A Crim R 456, 457.

[38] There were two other threads: circumstantial evidence constituted by Mrs Burman's description of the jacket with a tear worn by the man whom she saw at the cemetery, a torn jacket of similar description being found at the applicant's premises when they were searched in 2000; and what were said to be the applicant's attempts to change his appearance after the killing of Miss Halvagis, such attempts being said to evidence consciousness of guilt.

[41] A copy of pages 1, and 4 -7 of this edition is annexed as Schedule A to these reasons. Mrs Burman, understandably, was not taken to pages 4 - 7. Their content was referred to by applicant's counsel on the stay application. They give the flavour of a good deal of what was written about the applicant over the years.

[42] Indeed, at one point the learned trial judge suggested that it might stand as corroboration of the alleged confession: See T1118; compare T1345-1347, and the discussion at T1349-1351.

[43] Sometime, on Fraser's account, between February 2002 and the early part of 2003. Other evidence showed that the applicant and Fraser were housed in the same protection unit between 22 January 2002 and 20 March 2003. Fraser's first statement to the police was made in mid 2005.

[44] So much is apparent from the transcript itself, and additionally from a videotape of his evidence which I have viewed.

[46] Pollitt v The Queen [1992] HCA 35; (1992) 62 A Crim R 190.

[47] [1992] HCA 16; (1992) 173 CLR 592.

[48] [2001] VSCA 17; (2001) 7 VR 631.

[49] [2002] NSWCCA 336; (2002) 132 A Crim R 371 (Court of Criminal Appeal of New South Wales).

[50] [2003] QCA 77; (2003) 138 A Crim R 103.

[51] [2003] QCA 124; (2003) 140 A Crim R 303 (Qld Court of Appeal).

[52] [2008] QCA 227.

[53] [2008] NSWCCA 303.

[54] [2008] SASC 328; (2008) 102 SASR 318.

[55] (1990) 92 Cr App R 239.

[56] (1993) 98 Cr App R 361.

[57] [1996] 2 Cr App R 374.

[58] [2006] EWCA Civ 1093; [2007] QB 659.

[59] [1989] HCA 28; (1989) 167 CLR 94.

[61] [2004] NSWCA 324; (2004) 61 NSWLR 344.

[62] [2006] VSCA 300; (2006) 14 VR 512.

[63] [2008] VSCA 47; (2008) 182 A Crim R 350 (Victorian Court of Appeal).

[64] [2008] VSCA 49; (2008) 182 A Crim R 496.

[65] [1989] HCA 46; (1989) 168 CLR 23.

[66] [1992] HCA 57; (1992) 177 CLR 292.

[68] [2009] HCA 20; (2009) 255 ALR 399.

[69] Hinch v Attorney-General (Vict) [1987] HCA 56; (1987) 164 CLR 15.

[70] The Queen v Glennon [1992] HCA 16; (1992) 173 CLR 592, 598.

[73] [1936] HCA 40; (1936) 55 CLR 499, 505.

[74] The Queen v Glennon [1992] HCA 16; (1992) 173 CLR 592, 603.

[84] [2001] VSCA 17; (2001) 7 VR 631.

[85] [2003] QCA 77; (2003) 138 A Crim R 103.

[86] [2003] QCA 124; (2003) 140 A Crim R 303.

[87] [2002] NSWCCA 336; (2002) 132 A Crim R 371.

[88] [2008] VSCA 47; (2008) 182 A Crim R 350.

[89] [2001] VSCA 17; (2001) 7 VR 631, 657-658 [62], (Winneke P and Ormiston JA).

[90] [2008] VSCA 47; (2008) 182 A Crim R 350, 354 [7].

[91] R v Dudko [2002] NSWCCA 336; (2002) 132 A Crim R 371, 374-375 [17]-[23] (Spigelman CJ).

[92] R v Long; ex parte Attorney General for Queensland [2003] QCA 77; (2003) 138 A Crim R 103, 144-145 [173]-[175], (Jerrard JA); R v D'Arcy [2003] QCA 124; (2003) 140 A Crim R 303, 306 [15] (Davies JA). The situation was different in R v Ferguson; ex parte Attorney General (Qld) [2008] QCA 227. That was a Crown appeal against a stay order, and only the exercise of the discretion fell for consideration.

[93] Or, as has been said, that a fair trial is not the same as a perfect trial.

[94] [1934] HCA 49; (1934) 52 CLR 335.

[95] Ibid 346, 347 (Gavan Duffy CJ, Dixon, Evatt and McTiernan JJ); 355 (Starke J). As a footnote to history, concerning Tuckiar's ultimate fate, see Cole, Hell West and Crooked (Angus & Robertson, 1988) p287.

[96] Great Western Railway Company v Owners of SS Mostyn [1928] AC 57, 73; Victoria v Commonwealth [1971] HCA 16; (1971) 122 CLR 353, 382.

[97] Federation Insurance Ltd v Wasson [1987] HCA 34; (1987) 163 CLR 303, 314.

[98] [2003] QCA 77; (2003) 138 A Crim R 103.

[100] John Fairfax Publications and Anor v District Court of New South Wales and Ors [2004] NSWCA 324; (2004) 61 NSWLR 344, 360 [59], see also R v Dudko [2002] NSWCCA 336; (2002) 132 A Crim R 371, 374-375 [19].

[101] R v Abu Hamza [2006] EWCA Civ 1093; [2007] QB 659, 679 [78], 682-684, [88]-[92].

[102] (1993) 98 Cr App R 361.

[103] (1990) 92 Cr App R 239.

[104] See [137] and [138] above.

[105] For example, The Queen v Glennon [1992] HCA 16; (1992) 173 CLR 592, 614-615 (Brennan J); R v Glennon (No 2) [2001] VSCA 17; (2001) 7 VR 631, 661-662 [67] (Winneke P and Ormiston JA); R v Dudko [2002] NSWCCA 336; (2002) 132 A Crim R 371, 375 [21] (Spigelman CJ); R v Long, ex parte Attorney General for Queensland [2003] QCA 77; (2003) 138 A Crim R 103, 106-107 [9] (McMurdo P), 142 [166] (Jerrard JA); Gilbert v The Queen (2000) 201 CLR 414, 420 [13] (Gleeson CJ and Gummow J), 425 [31] (McHugh J); John Fairfax Publications Pty Ltd & Anor v District Court of New South Wales & Ors [2004] NSWCA 324; (2004) 61 NSWLR 344, 366-367 [103]-[110] (Spigelman CJ); R v Thomas (No 3) [2006] VSCA 300; (2006) 14 VR 512, 520-521 [35]-[36]; General Television Corporation Pty Ltd v Director of Public Prosecutions (Vic) & Anor [2008] VSCA 49; (2008) 182 A Crim R 496, 511-512 [54]; R v Vjestica [2008] VSCA 47; (2008) 182 A Crim R 350, 361-362 [31]-[35] (Maxwell P).

[106] [2008] VSCA 49; (2008) 182 A Crim R 496, 512-513 [54]-[55].

[107] [1946] USSC 101; (1945) 328 US 331, particularly 356-358 (Frankfurter J).

[108] [1981] AC 303, 335 (Viscount Dilhorne).

[109] Published in 'The Nature of the Judicial Process' (Yale University Press).

[110] R v Taylor (1993) 98 Cr App R 361. R v McCann (1990) 92 Cr App R 239 addressed the problem of 'during trial' adverse publicity.

[111] In R v Ferguson; ex parte Attorney-General (Qld) [2008] QCA 227, the Queensland Court of Appeal said that 'The exceptional jurisdiction permanently to stay proceedings is truly residual in character, in the sense that it falls to be exercised only in those cases where the other legal safeguards of the right of the accused to a fair trial are not apt to secure that right ...'

[112] R v Skaf [2008] NSWCCA 303 is a case in point. In R v Ferguson; ex parte Attorney-General (Qld) [2008] QCA 227, the effect of the successful Crown appeal was that a trial would be had.

[113] The circumstances of R v von Einem (1991) 55 SASR 199, where Duggan J gave a ruling against a stay application in a post-Barton, Jago and Murphy, but pre-Glennon landscape, went close. The stay application was founded on four bases: (1) the conduct of the prosecution; (2) the magnitude of grossly prejudicial pre-trial publicity; (3) the proposition that a coronial inquest had been conducted ultra vires, and had generated further prejudicial material; and (4) delay. Duggan J ruled against the accused on each of those bases, the first three of which were interconnected. Specifically concerning the second basis, his Honour described media linkage between various offences and the accused's involvement as having been put in the language of 'theory and speculation'. He stated that he was satisfied that a properly instructed jury would be able to act on the evidence alone, ignoring that theory and speculation. He then referred to the publication of alleged 'similar fact' evidence given at the preliminary examination - it being clear that the accused would challenge its admissibility at trial. He concluded, 'eventually and not without anxious consideration' that the level of potential prejudice could be overcome so as to ensure a fair trial of the accused.

[114] Cf Farah Constructions v Say-dee Pty Ltd [2007] HCA 22; (2007) 230 CLR 89, 151 [134]-[135].

[115] Murphy and the Queen [1989] HCA 28; (1989) 167 CLR 94, 99 (Mason CJ and Toohey J).

[117] [1992] HCA 16; (1992) 173 CLR 592, 613.

[118] Murphy v The Queen [1989] HCA 28; (1989) 167 CLR 94, 98-9 (Mason CJ and Toohey J).

[119] The Queen v Glennon [1992] HCA 16; (1992) 173 CLR 592, 614 (Brennan J).

[121] For instance, in some jurisdictions trial by judge alone is permitted, albeit, at present, only with the consent of the accused: Criminal Procedure Act 1986 (NSW), s 132; Criminal Procedure Act 2004 (WA), s 118; Juries Act 1927 (SA), s 7; Supreme Court Act 1933 (ACT), s 68B.

[122] As to which see my remarks at [195] and [196].

[123] Although the jury might turn out to be unrepresentative of the panel as a whole.

[124] Bilal Skaf v R [2008] NSWCCA 303 [27] (Unreported, McClellan CJ at CL, Hidden and Howie JJ, 17 December 2008).

[126] Ibid [31]-[32] (citations omitted).

[127] [1992] HCA 16; (1992) 173 CLR 592.

[133] Jago v The District Court of New South Wales [1989] HCA 46; (1989) 168 CLR 23, 33-34 (Mason CJ).

[134] Bilal Skaf v R [2008] NSWCCA 303 [33] (Unreported, McClellan CJ at CL, Hidden and Howie JJ, 17 December 2008).

[135] (1993) 69 A Crim R 450, 453-4. This passage was cited with approval by Spigelman CJ in R v Bell [1998] NSWSC 570, 7-8 (Unreported, Spigelman CJ Abadee and Ireland JJ, 8 October 1998).

[136] [1998] NSWSC 795, 47 (Unreported, Gleeson CJ, Meagher JA, Newman J, 26 February 1998).

[137] [2002] NSWCCA 336.

[139] [2009] VSC 342R.

[141] Tuckiar v The King [1934] HCA 49; (1934) 52 CLR 335.

[142] See, for example, R v Ferguson; ex parte A-G (Qld) [2008] QCA 227.

[143] [1980] HCA 48; (1980) 147 CLR 75.

[144] [1989] HCA 46; (1989) 168 CLR 23.

[153] See, for example, R v Straffen [1952] 2 QB 911 and R v Smith (1915) 11 Cr App R 229 ('brides in the bath' case).

[154] In R v Mokbel [2009] VSC 342R [94] Kaye J made the same point. His Honour added that it was 'the common experience of trial judges that questions asked by juries and potential probity issues raised by them' through the judge's tipstaff 'regularly reflect a firm understanding by juries of their judicial function, and the seriousness with which juries regard and adhere to directions given to them by the trial judge'.

[155] [1983] VicRp 6; [1983] 1 VR 65.

[156] See also R v Bell [1998] NSWSC 570 (Unreported, Spigelman CJ Abadee and Ireland JJ, 8 October 1998) and R v Long [2003] QCA 77. In both of these cases, enormous pre-trial publicity, perhaps even more substantial than that in Glennon, did not lead to a permanent stay, and did not lead to an appellate court setting aside the conviction.

[157] See especially R v Abu Hamza [2006] EWCA Civ 1093; [2007] QB 659.

[158] R v Christie [1914] AC 545, 559. See also: Noor Mohamed v The King [1949] AC 182, 192; Driscoll v The Queen [1977] HCA 43; (1977) 137 CLR 517, 541 and R v Merritt and Roso (1985) 19 A Crim R 360, 377-380.

[159] McDermott v The King (1948) 76 CLR 501, 506-7; The King v Lee [1950] HCA 25; (1950) 82 CLR 133, 150-151; MacPherson v The Queen [1981] HCA 46; (1981) 147 CLR 512, 519-20 and Cleland v The Queen [1982] HCA 67; (1982) 151 CLR 1, 18 and 33.

[160] R v Ireland [1970] HCA 21; (1970) 126 CLR 321, 334-5; Bunning v Cross [1978] HCA 22; (1978) 141 CLR 54, 74-5 and Cleland v The Queen [1982] HCA 67; (1982) 151 CLR 1, 8-9.

[161] Driscoll v The Queen [1977] HCA 43; (1977) 137 CLR 517, 541.

[162] (1994) 72 A Crim R 160.

[164] Davies and Cody v The King [1937] HCA 27; (1937) 57 CLR 170, 181; Alexander v The Queen [1981] HCA 17; (1981) 145 CLR 395 and R v Tugaga (1994) 74 A Crim R 190.

[165] [1992] VicRp 17; [1992] 1 VR 273.

[166] [1995] VicRp 34; [1995] 1 VR 533, 548.

[167] R v Clark (1996) 91 A Crim R 46; Festa v The Queen [2001] HCA 72; (2001) 208 CLR 593; R v Stott [2000] QSC 273; (2000) 116 A Crim R 15; Kelly v The Queen [2002] WASCA 134; (2002) 129 A Crim R 363; and Al-Hashimi v The Queen [2004] WASCA 61; (2004) 181 FLR 383.

[168] [1992] HCA 13; (1992) 173 CLR 555.

[169] See also R v Jamal [2000] FCA 1195; (2000) 182 ALR 307 (admissibility of dock identification).

[170] [2001] HCA 72; (2001) 208 CLR 593.

[171] Ibid [12]-[23] (citations omitted).

[172] [1936] HCA 40; (1936) 55 CLR 499.

[173] See also Australian Coal and Shale Employees' Federation v The Commonwealth [1953] HCA 25; (1953) 94 CLR 621, 627.

[174] See generally Domican v the Queen [1992] HCA 13; (1992) 173 CLR 555.

[175] R v Turnbull [1977] 1 QB 224, 228; Smith v The Queen [2001] HCA 50; (2001) 206 CLR 650 [55], Trudgett v The Queen [2008] NSWCCA 62; (2008) 70 NSWLR 696 and R v Serrano [2009] VSCA 140.

[176] See R v Serrano [2009] VSCA 140 [33], citing Festa v The Queen [2001] HCA 72; (2001) 208 CLR 593 as authority for the proposition that in a case involving 'similarity', as distinct from 'positive identification', a less powerful direction than that stipulated in Domican would be justified.

[177] Festa v The Queen [2001] HCA 72; (2001) 208 CLR 593.

[178] Pitkin v The Queen [1995] HCA 30; (1995) 69 ALJR 612; R v Benz [1989] HCA 64; (1989) 168 CLR 110; Festa v The Queen [2001] HCA 72; (2001) 208 CLR 593 and R v Hassan [2004] VSC 84.

[179] R v Clune (No 2) [1996] VicRp 1; [1996] 1 VR 1 and Festa v The Queen [2001] HCA 72; (2001) 208 CLR 593, 610 (McHugh J). See generally, David Ross QC, Ross on Crime (3rd ed, 2007) [9.100].

[180] [1981] HCA 17; (1981) 145 CLR 395.

[182] R v Burchielli [1981] VicRp 61; [1981] VR 611; Domican v The Queen [1992] HCA 13; (1992) 173 CLR 555 and Festa v The Queen [2001] HCA 72; (2001) 208 CLR 593.

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

[184] R v Crupi (1995) 86 A Crim R 229.

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

[186] R v Burchielli [1981] VicRp 61; [1981] VR 611.

[187] Festa v The Queen [2001] HCA 72; (2001) 208 CLR 593.

[188] R v Turnbull [1977] QB 224 and R v Burchielli [1981] VicRp 61; [1981] VR 611.

[189] R v Burchielli [1981] VicRp 61; [1981] VR 611; R v Dickson [1983] VicRp 19; [1983] 1 VR 227 and R v Haidley & Alford [1984] VicRp 18; [1984] VR 229.

[190] Weeder v R (1980) 71 Cr App Rep 228; R v Haidley & Alford [1984] VicRp 18; [1984] VR 229 and R v Callaghan [2001] VSCA 209; (2001) 4 VR 79.

[191] R v Marshall [2000] NSWCCA 210; (2000) 113 A Crim R 190 [38] (Spigelman CJ).

[192] Kelleher v The Queen [1974] HCA 48; (1974) 131 CLR 534; Domican v The Queen [1992] HCA 13; (1992) 173 CLR 555; R v Debs [2005] VSCA 66 and R v Abbouchi; R v Allouche [2008] VSCA 171.

[193] R v Bint & Butterworth (1996) 187 LSJS 201.

[194] Domican v The Queen [1992] HCA 13; (1992) 173 CLR 555, 562.

[195] Kelleher v The Queen [1974] HCA 48; (1974) 131 CLR 534, 551 (McTiernan J).

[196] Davies & Cody v The King [1937] HCA 27; (1937) 57 CLR 170, 182-3.

[197] Domican v The Queen [1992] HCA 13; (1992) 173 CLR 555, 561-564.

[199] See, for example, Radford v R (1993) 66 A Crim R 210; The Queen v Glennon [1994] HCA 7; (1994) 179 CLR 1; R v Johnson [2001] VSCA 242; (2001) 126 A Crim R 395; Farrell v R (1998) 194 CLR 286; R v Dodd [2002] NSWCCA 418; (2002) 135 A Crim R 32; Johnson v Giumelli [2003] ACTSC 58; (2003) 175 FLR 467; R v Ngo (2003) 56 NSWLR 55; R v Rose [2002] NSWCCA 455; (2002) 55 NSWLR 701; R v Rivkin [2004] NSWCCA 7; (2004) 59 NSWLR 284; Al-Hashimi v The Queen [2004] WASCA 61; (2004) 181 FLR 383; R v Debs [2005] VSCA 66; R v Demiri [2006] VSCA 64; R v Mendoza [2007] VSCA 120; (2007) 173 A Crim R 157; R v Abbouchi; R v Allouche [2008] VSCA 171 and R v Serrano [2009] VSCA 140.

[200] R v Clune [1982] VicRp 1; [1982] VR 1; R v Dickson [1983] VicRp 19; [1983] 1 VR 227; R v Haidley & Alford [1984] VicRp 18; [1984] VR 229, 230-1 and R v Finn (1988) 34 A Crim R 425.

[201] J D Heydon, Cross on Evidence (6th Australian ed, 2000) [1400].

[202] Domican v The Queen [1992] HCA 13; (1992) 173 CLR 555, 565.

[203] In R v Debs [2005] VSCA 66, a deficient warning as to identification was nonetheless said not to warrant the setting aside of a conviction, presumably on the basis that the proviso could be invoked. The Court of Appeal in Debs did not speak in terms of the proviso, but it was plainly implicit in the reasoning of Vincent JA that this was what his Honour had in mind.

[205] Bilal Skaf v R [2008] NSWCCA 303 [71]-[79] (Unreported, McClellan CJ at CL, Hidden and Howie JJ, 17 December 2008).

[206] R v Serrano [2009] VSCA 140.

[207] [1993] HCA 63; (1993) 178 CLR 193.

[208] R v Renzella [1997] 2 VR 88, 91.

[209] [2006] VSCA 263; (2006) 16 VR 26.

[210] [2007] VSCA 41; (2007) 16 VR 118.

[211] See generally: Victorian Law Reform Commission, Jury Directions: Final Report 17 (2009) [3.71]-[3.81]. The Commission observes that since the mid 1990s, the Victorian Court of Appeal has heard at least 84 appeals which raised consciousness of guilt as an issue, the appellant having succeeded in 28 of those cases.

[212] Zoneff v The Queen (2000) 200 CLR 234.