Raymond George MORGAN v R

The Audi 35It is convenient to begin with the third and fourth grounds of appeal, relating to the RS4 Audi station wagon. It was apparently a high performance vehicle. There were two keys for it. As I have said, it was stolen from the driveway of the owners' home early in the morning of 11 January 2008. One of the keys, which was kept in a drawer of a table in the hallway of the house, was missing. There was no sign of forced entry to the house. 36There was evidence that for a period in early 2008 the appellant was staying at a house in Tempe, where his cousin lived with his partner and child. In the morning of 23 April 2008, the appellant was arrested near that home pursuant to a warrant issued after he failed to appear at a Local Court in relation to a traffic matter. He was searched, and in a bum bag which he had been wearing an Audi key was found. This was later established by the Technical Manager at Audi Australia Pty Ltd to be a key for the stolen car. 37On the day after the appellant's arrest, 24 April 2008, the house at Tempe was searched. In a bedroom which he was said to have been occupying police located a number of lock-picking instruments. A forensic locksmith gave evidence that some of those instruments were capable of opening the lock on the front door of the Mosman home where the car had been stolen, but it would be necessary also to use a tension wrench or a similar tool. No item of that kind was found during the search. 38In the early hours of 24 April the Audi was destroyed by fire in the garage area of a townhouse complex at Abbotsford Point. There was no evidence connecting the appellant with those premises, and he was in custody following his arrest at the time the car was burnt out. The effect of evidence of residents of the townhouses was that the car had been in the garage for some time, roughly a couple of months. It had stolen registration plates on it, but it was able to be identified by its VIN number.

Grounds 3 and 4 39The third ground is that the trial judge failed to direct the jury that, in order to find the appellant guilty of receiving the vehicle, they would have to be satisfied beyond reasonable doubt that he did not steal it. The fourth ground is that the verdict of guilty on the receiving count is unreasonable. 40As I have said, the counts of stealing and receiving the vehicle were in the alternative. However, at the trial the Crown did not rely upon s 121 of the Crimes Act 1900, and the trial judge did not direct the jury in accordance with it. Section 121 provides: " Verdict of 'larceny or receiving' Where, on the trial of a person charged with larceny, or any offence which includes larceny, and, also, with having unlawfully received the property charged to have been stolen, knowing it to have been stolen, the jury find specially that the person either stole, or unlawfully received, such property, and that they are unable to say which of those offences was committed by the person, such person shall not by reason thereof be entitled to acquittal, but shall be liable to be sentenced for the larceny, or for the unlawful receiving, whichever of the two offences is subject to the lesser punishment." 41Relevantly, what his Honour did say in the summing-up was this: "Count 2 is said to be 'in the alternative'. What that means is that if you find the accused to be not guilty of the first count, that is the car stealing count (count 1), then you will be asked to consider whether he is guilty or not guilty of the alternative charge. So you do not go on to consider count 2 unless you are satisfied that he is not guilty on count 1. You do not do them as a whole rolled up count, roll the two together. Look at count 1 separately. If, and only if, you are satisfied that he is not guilty on that, you then go on to consider count 2." 42Earlier, before the summing-up, his Honour had explained to the jury the notion of joint criminal enterprise, in relation to both the robberies and the stealing of the car. In the summing-up, after the passage which I have quoted, his Honour told the jury that they might be satisfied that the car was stolen but not satisfied that the theft was part of a joint criminal enterprise to which the appellant was a party. He continued: "You may not be satisfied ... on the inferential basis the Crown asks you to draw, from not just the unique ignition key but from the other matters as well. If you adopt that approach, and as I say you can have any one of a number of different ways of getting there, then and only then if you decide he is not guilty on count 1, then go to consider count 2, which is a receiving charge." 43The appellant was represented in this court by Mr Stratton SC, who did not appear at the trial. The thrust of ground 3, he explained, was that his Honour's directions fell short of conveying to the jury that to find the appellant guilty of receiving the Audi, they would have to be satisfied beyond reasonable doubt that he was not involved in stealing it (whether alone or as a party to a joint criminal enterprise to do so). 44Mr Stratton relied upon R v McCarthy and Ryan (1993) 71 A Crim R 395. In that case the accused were charged, among other things, with larceny of a motor vehicle. The Crown relied upon the doctrine of recent possession, but there was no alternative count of receiving. That being so, s 121 was not available. 45The trial judge had directed the jury about the concept of recent possession, explaining that it could give rise to the inference that the accused had either stolen the vehicle or had received it, knowing it to have been stolen. It was in the ensuing directions that his Honour was found to have fallen into error. Delivering the leading judgment in this court, Hunt CJ at CL summarised those directions in this way (at 399): "... the judge suggested first that it was sufficient if they considered it 'highly unlikely' that the accused had received the vehicle as receivers, then that it was sufficient if they merely preferred the conclusion that the accused had stolen the vehicle to one that they had received it knowing it to have been stolen, and next that, if they were unable to conclude beyond reasonable doubt one way or the other, the accused were entitled to be acquitted." 46The Chief Judge said that each of those directions was wrong. His Honour explained (also at 399) the relevant principles as follows (omitting references to authority): "The offences of larceny and of receiving are mutually exclusive ... . It follows as a matter of fundamental principle (and common sense) that, if there was any reasonable possibility that the particular accused being considered had received the vehicle knowing it to have been stolen, the jury could not be satisfied beyond reasonable doubt that he or she was guilty of its larceny ... ." 47Mr Stratton submitted that the converse is true of the present case. For the jury to have found the appellant guilty of receiving the Audi, they would have to have found that there was no reasonable possibility that he had stolen it. Put another way, they would have to have been satisfied beyond reasonable doubt that he was not the thief. Mr Stratton argued that this principle would not have been conveyed to the jury by his Honour's directions. 48No complaint was made about these directions by counsel for the appellant at the trial, and rule 4 applies. Mr Stratton pointed out that the same was true in McCarthy and Ryan , but the appeal by each of them against the conviction of larceny was allowed. That said, in the present case his Honour did tell the jury twice in the passage from the summing-up which I have quoted that they should consider the receiving count only if they were "satisfied" that the appellant was not guilty of the stealing count. The real complaint is that that direction required elaboration. In the event, I find it unnecessary to decide this ground because I am satisfied that ground 4, that the verdict of receiving is unreasonable, is made out. 49Mr Stratton did not argue that the evidence was not capable of establishing the appellant's criminal association with the Audi, either as thief or receiver. Rather, his submission was that it was not open to the jury to be satisfied to the requisite degree that he was a receiver, not a thief. The circumstantial case against the appellant, said Mr Stratton, was equally consistent with his having stolen the vehicle. 50The Crown prosecutor in this court, who also did not appear at the trial, relied on the lapse of time between the theft of the car in January 2008 and the appellant being found in possession of a key for it three months later. She noted that his Honour had given the jury a comprehensive direction about possession for the purpose of the receiving count and the counts of possessing the proceeds of crime. That lapse of time, she argued, supported a finding that the appellant had come into possession of the car some time after it had been stolen. 51The finding of the pick-locks in the room said to be occupied by him might be consistent with his having gained entry to the Mosman home and taken the car key, but she noted that the other implement necessary to gain entry, a tension wrench or similar tool, was not found. She also relied upon the fact that someone else must have been criminally involved with the car, because it was burnt out at a time when the appellant was in custody following his arrest the day before. 52On the other hand, it was the Crown case that the Audi was the vehicle used in connection with the robberies at the hotels, in which the appellant was involved. That evidence would place him in possession of the vehicle, or having access to it, only a matter of days after it was stolen. In any event, the fact that it was not until several months later that the appellant could be clearly linked to the car by his possession of the key for it is by no means decisive of this issue. It appears to me that the whole of the evidence concerning the vehicle is as consistent with his having stolen it as it is with his having received it. Certainly, the evidence leaves open the reasonable possibility that he stole it, either alone or in complicity with someone else. This court is as well placed as the jury was to make that assessment, and the jury could not reasonably have found otherwise. 53Accordingly, I am satisfied that the verdict of guilty of receiving cannot be supported by the evidence. If the jury had been instructed in accordance with s 121 of the Crimes Act , they may have brought in the special verdict for which that section provides. At the hearing of the appeal the question was raised whether, for that reason, a new trial should be ordered. However, the section applies to a trial in which both larceny and receiving are charged. The appellant having been found not guilty of larceny, that option is not available. I would quash the conviction on the second count, and enter a verdict and judgment of acquittal.

The robberies 54As I have said, the robberies at the hotels in Willoughby and Drummoyne both occurred in the early hours of 17 January 2008. They are the subject of the first, second and fifth grounds of appeal. The hotels are not far apart and the robberies were close in time. No witness identified the offenders at either hotel. It was the Crown case that the same two men, one of them the appellant, committed each offence, and that the stolen Audi was their means of transport.

Willoughby 55The robbery at the hotel at Willoughby took place at about 12.25am, after it had closed. The robbers smashed external doors to the hotel to gain entry. The manager, Mr Peter Brooks, was in his office and he called the police. He opened his office door after being ordered by one of the robbers to do so. 56He saw two men, one of them holding a sledgehammer. He described the man with the sledgehammer as "fairly large", "strong", "fairly solid." He was wearing a close fitting black balaclava, a tracksuit top and tracksuit pants. Mr Brooks described him as being not very tall, about 5'5" (170 cms). It was the Crown case that this man was the appellant. 57Mr Brooks described the other man as "very skinny" and about his own height, that is, 6'2" (184 cms). He was wearing tracksuit pants and a hooded top with the hood over his head. He could not see this man's face and thought that he was also wearing a balaclava. He did not recall that man holding anything. 58CCTV footage of parts of the incident was in evidence, together with some still shots from that footage. The two robbers can be seen in those exhibits. Footage obtained outside of the hotel at the time of the robbery depicted a car, and the Technical Manager of Audi Australia gave evidence that it appeared to be an Audi of the same model as the stolen vehicle. Enquiries by the officer in charge of the investigation revealed that no other red Audi of that model had been reported stolen between 1 January and 24 April 2008. 59Cash in an amount exceeding $45,000 was taken from two safes in Mr Brooks' office. It was bundled in certain quantities, depending on the denomination of the notes, and secured by rubber bands. $25,000 in cash, bundled in that way and secured by rubber bands, was found at the premises at Tempe during the search on 23 April. Mr Brooks gave evidence that money was generally bundled in this way in the hotel industry.

Drummoyne 60The robbery at the hotel at Drummoyne took place at about 12.50am on 17 January. It is unnecessary to describe the course of the robbery, except to say that two men entered the hotel and forced the manager to remove the cash from a number of poker machines and from the tills. A little over $13,500 was taken. The hotel was still open at the time. 61The manager, Mr Martin Schooling, described the first man he saw as being about his own height, 5'9", of medium build, and aged in his 30's. He was wearing a balaclava and a hooded top with the hood over his head. He was holding a crowbar. He saw a second man, whom he described as also about 5'9" tall and of a larger build than the other man, "probably a bit chubby." He had broad shoulders and a thicker torso. He was wearing a balaclava, a tracksuit top and pants, and black gloves. He was holding a sledgehammer. It was the Crown case that this second man was the appellant. 62The robbery was also witnessed by a security guard, Mr Abedinasser Elkordy, and two patrons, Ms Wendy Farrell and Mr Duncan Macpherson. Mr Macpherson did not give evidence, but his statement to police was admitted. All three of them described the two robbers, one of them holding a crowbar and the other a sledgehammer. 63Mr Elkordy described the man with the crowbar as about 160 - 165 cms tall, with a "mid-way" build. His head was covered and he was wearing gloves. He said that he could see this man's eyes, and thought that he looked Middle-Eastern or Italian. He described the other man, holding a "big hammer", as being taller and of a larger build. In evidence he said that he was more than 175 cms tall, although he had told investigating police that he was between 165 and 170 cms. 64Ms Farrell described the man with the crowbar as skinny, about 5'6" to 5'8" tall, wearing a black balaclava, a black "hoodie" and black gloves. She described the man with the sledgehammer as shorter and of a stocky build. In evidence-in-chief she said he was 5'6" to 5'8" tall, but in cross-examination she said that he looked about the same height as herself, 5'5". He was also wearing a black balaclava and black gloves. 65Mr Macpherson described the man with the crowbar as tall and fit looking, wearing a black ski mask on his face, a black hoodie and gloves. He described the man with the sledgehammer as shorter, also fit looking, wearing a black balaclava and black gloves. 66The hotel stands on the corner of Victoria Road and Lyons Road. At around the relevant time, Mr Alex Carlson left the hotel and waited for a bus at a bus stop on Victoria Road. He saw a red Audi station wagon, a high performance vehicle of a kind with which he was familiar, approach the intersection from Victoria Road and turn into Lyons Road. It apparently drove around a block, as he saw it again proceeding up Victoria Road and turning into Lyons Road. 67Some time later, he saw two men who, it might be inferred, were the robbers enter the hotel from Lyons Road. Shortly thereafter, the men emerged. The red Audi, which was stopped in a nearby laneway, reversed out and pulled up near the hotel in Lyons Road. The two men got into the vehicle and it was driven off quickly. 68Also at the relevant time, Ms Yvette Fenech was a passenger in a car being driven along Lyons Road towards Victoria Road. She saw a new model red Audi parked near the hotel. She saw a man running fast from the direction of the hotel towards the Audi. He got into it and it was driven off quickly in the opposite direction (towards Five Dock). 69She said that the man ran past an electrical box, and she estimated him to have been 30 cms taller than that box. The officer in charge of the investigation gave evidence that the electrical box was 1.46 metres high. On Ms Fenech's estimate, then, the height of the man was about 176 cms (roughly 5'9"). When describing the man to the police she had said that he was of slim build. 70Again, there was in evidence CCTV footage inside the hotel, together with still shots from that footage, in which the two robbers can be seen. There was also CCTV footage outside the hotel showing the movement of a red vehicle consistent with the observations of Mr Carlson and Ms Fenech. This evidence conveys that there was a third person involved in the robbery as the driver of the car.

Expert evidence 71An important part of the Crown case on the robberies was the evidence of Dr Maciej Henneberg, a biological anthropologist and anatomist. He had been appointed to the Wood Jones chair of anthropological and comparative anatomy at the University of Adelaide, and held a doctorate in biological anthropology from a university in Poland. He professed expertise in the anatomical comparison of images, known loosely as "body mapping." It is his evidence which is the subject of the first and second grounds of appeal. 72In August 2008, Professor Henneberg was asked by investigating police to conduct an anatomical comparison of CCTV images from both hotels of the person alleged to be the appellant, on the one hand, and images of the appellant obtained during a forensic procedure upon his arrest, on the other. He had access to the CCTV footage and stills from it, and photos and moving images of the appellant. He set out his comparison and conclusions in an expert certificate under s 177 of the Evidence Act , and this material was fleshed out at some length in oral evidence. 73In the presence of two detectives, he examined the CCTV images and noted the anatomical characteristics of the person said to be the appellant. He referred to that person in the certificate as the "offender" or the "person of interest." Only after he had done that did he examine the images of the appellant obtained by the police. He proceeded in that way to ensure that his observation of the characteristics of the offender was not influenced by his knowledge of the appearance of the appellant: as he put it, to guard against "displacement." He also examined images of two other persons as a "comparative background" to his observations, so as to increase the "validity of the analysis aimed as establishing anatomical similarity of a particular offender to a particular suspect." 74He set out his observations of the offender in the CCTV images in the certificate. In doing so, he used some technical terms which are sufficiently explained in that description. His observations were as follows: "Person of interest is an adult male of heavy body build. His shoulders and hips are wide. He has a prominent abdomen but his upper and lower limbs, especially in their distal segments, are not thick. This suggests centripetal pattern of body fat distribution. This pattern consists of the deposition of most body fat on the trunk while limbs remain relatively thin. His head and face were covered by a garment well adhering to the surface of the skin. This enabled me to make observations of the head shape, nose and face profile. His head is dolichocephalic (elongated) in the horizontal plane (viewed from above). His nose is wide and rather prominent while his face has straight profile (orthognathic). He is right-handed in his actions and carries himself straight." 75He set out his observations of the appellant in the images obtained by the police as follows: "Mr MORGAN is an adult male of darkish colour skin. He is heavily build (sic), but the bulk of his body fat is concentrated on the trunk while his distal limbs are not fat (see especially lower legs). His shoulders and hips are wide. He has straight posture. His face is orthognatous (straight profile, without protruding jaws). His nose is wide and rather prominent. Since during the forensic procedure no photographs showing his head from above were taken the assessment of his head shape is uncertain. Judging from head width shown on posterior (taken from behind) photographs and head length on profile photographs, his head seems to be dolichocephalic (elongated oval rather than a shape approximating a circle)." Images of the top of the appellant's skull later supplied to him confirmed his opinion about the shape of the head. 76In the certificate he expressed his conclusion in this way: "Based wholly or substantially on the above knowledge, I am of the opinion that there is a high level of anatomical similarity between the Offender and the Suspect (Mr MORGAN). My opinion is strengthened by the fact that I could not observe on the Suspect any anatomical detail different from those I could discern from the CCTV images of the Offender." 77Attached to the certificate was an explanation of his methodology. This included the following: "I use a method of anatomical examination. It consists of inspection of images in order to form an opinion on the individual characteristics of the body and the face of a specific person. Since individuals differ in a number of details, and since each set of photographs or digital images allows observation of a different number and quality of details, statistical analysis is not a practical approach. Statistical analyses provide reliable results in cases where anthropometric measurements can be taken precisely, or where standard photographs in prescribed body positions are taken at same (sic) angles. Surveillance materials and standard Police photographs are not of the quality allowing such analyses. It can be stated that the anatomical analysis of images is similar to recognition of persons on such images by lay observers, such as, for example, eye witnesses. The only difference is that observations are conducted by an experienced anatomist who has good understanding of shapes and proportions of various details of a human body." 78He stated that, in his "anatomical identification", he makes comparisons of individuals. He does not draw conclusions about a person's national, ethnic or geographic origin, saying, for reasons he identified, that for the purpose of individual identification statistical data on any particular racial group would be futile. He continued: "When making a statement regarding the degree of certainty or uncertainty of my specific individual identifications I take into account observable biological characteristics of particular individuals in their unique combination. For example, my statement that the identification is 95% certain is an estimate based on the fact that, due to the quality of images, I could not observe enough anatomical detail to reach certainty." To this last observation I shall return. 79Professor Henneberg's evidence was objected to and a voir dire examination was conducted, in which he gave oral evidence. While his certificate asserted "a high level of anatomical similarity" between the offender in the CCTV footage and the appellant, he said that he did not purport to identify the appellant as the offender. When he assessed the anatomical characteristics of the offender and his build, he had regard to other persons in the CCTV footage and inanimate objects, such as doorways or seats. 80While he recorded in the methodology set out in his certificate that he did not draw conclusions about a person's ethnicity, he did explain in evidence that ethnic groups might display certain bodily characteristics as a result of their geographical location and lifestyle. He said that most Australian Aboriginals had long thin limbs and that, when they put on weight due to a Western diet, it tended to be concentrated in the trunk, the process he described as centripetal fat distribution. He observed these features in the offender in the CCTV footage and, later, in the images of the appellant. It is apparent from the photos of the appellant in evidence that he is Aboriginal. 81He was asked in cross-examination about facial mapping. He considered that a loose expression, which he preferred not to use. What was involved, he said, was anatomical comparison and he added that, when that task is undertaken, there is no "qualitative difference between the face and the rest of the body." He denied that there are more identifiable features on a face than on a body. 82In his certificate he said that he subjected his "identification work" to peer review by persons "holding higher degrees in anatomical sciences." He also said that his evidence "was accepted" by courts in various Australian jurisdictions, citing particular cases in which he had been involved. In oral evidence, he said that he had undertaken more than fifty anatomical comparisons and provided expert certificates, usually to the police. In cross-examination, he gave this evidence: "Q. You don't know how many of those recognition or identification or similarity assessments you made are true and correct, do you? A. No I don't, I never take interest in the result. I provide my opinion and that's where it is. Q. There has been nothing to validate any of your opinions, has there, independently? A. I can't make that judgment. Q. Professor has there ever been any measurement or assessment of your ultimate opinions on any occasion? A. Well yes there were. Several of my opinions were peer reviewed by other individuals and although it's very inadequate or incomplete or maybe perhaps questioned, but a number of opinions expressed in certificates I provided were met with either admission or guilty plea of individuals whose comparisons I was conducting. Q. Have you ever been subject to a review where you have been provided with an array of CCTV images where the answer was known by your peers and asked to provide objectively your opinion as to whether that person was the person depicted as the suspect? A. No I wasn't, although I would welcome this. Q. You have never been tested in that way, have you? A. Yet again I don't recall, but probably I was." 83On the voir dire, the appellant led evidence from two other experts, Dr Richard Kemp, a forensic psychologist, and Mr Glenn Porter, a forensic scientist. Each of them had been asked to provide a critique of Professor Henneberg's expert certificate. Dr Kemp provided a report, and Mr Porter an expert certificate. Both of them gave oral evidence. 84Dr Kemp was a senior lecturer in forensic psychology at the University of Western Sydney. He had undertaken and published experimental research in both the theoretical and applied aspects of human face perception and person recognition. This included studying the factors affecting the accuracy of identification evidence. He acknowledged in evidence that he did not have a background in anatomy or anthropology. 85He noted that Professor Henneberg relied exclusively on a morphological approach to anatomical examination, and that he did not attempt to take measurements from the photographic images or to draw on any published data regarding the frequency of occurrence of particular anatomical features to estimate the probability that two sets of images showed the same person. His comparisons were made only with the naked eye. 86Dr Kemp had access to the CCTV and photographic material which Professor Henneberg had examined. He described the CCTV images as of relatively poor quality. He noted that the body of the relevant offender in those images was completely covered by clothing, so that no part of the face or body was directly visible. He questioned the professor's ability to make observations about the person's head shape, nose and face profile because the garment covering his head and face was "well adhering to the surface of the skin." He wrote in his report: "Professor Henneberg doesn't cite or present any evidence to demonstrate his ability to accurately discern face and head shape though (sic) a covering of clothing, or to do this by looking at relatively low quality video recordings. Professor Henneberg has no way to know the characteristics of the items of clothing, how the clothing adheres to the face and head, and whether the Offender is wearing anything else underneath the visible covering. Given this I believe that there is no scientific basis for the claim made by Professor Henneberg that he can discern aspects of the anatomy obscured by unknown items of clothing." 87To this Professor Henneberg responded in evidence that he had worked with the garment industry some years before, and had learned that in the "construction of clothing, so-called hang of garments is extremely important and it is related directly to body shape and posture." He added that "the force of gravity spreads the cloth on the body in a way that is inescapable because the force of gravity always acts and therefore certain aspects of body shape can be deduced from the study of images of cloth to body." 88Dr Kemp noted that Professor Henneberg did not attempt to make any statistical claim about the frequency of occurrence of the characteristics he observed in the offender in the CCTV footage and the appellant, so as to calculate the probability that two images of different individuals might show those characteristics. Yet, he said, such a probability assessment was implicit in the professor's conclusion that there was "a high level of anatomical similarity" between the offender and the appellant. Dr Kemp saw the effect of such a statement in this context as a suggestion "that these similarities are noteworthy and unusual." His own view was that they were "not uncommon in the adult population of Australia." He added that there was "no adequate statistical evidence available regarding just how common the possession of such a set of characteristics is, and as a result we have no way of knowing what conclusions we can draw from this observation." 89In an appendix to his report Dr Kemp summarised the findings of some research he conducted into the practice of producing facial mapping and body mapping evidence. Put shortly, his finding was that there was no scientific evaluation of its validity, reliability and error rate. He explained those terms in this way: "In scientific terms the validity of a technique is a measure of the extent to which a technique does what it is claimed to do. The reliability of a technique is the extent to which the technique would produce the same results when repeatedly applied by different operators or on different occasions. To be scientifically useful a technique must be demonstrated to be high in both validity and reliability. The validity and the reliability of a technique will partly determine its accuracy and the rate at which the technique will give rise to errors (the error rate). All measurement techniques have an associated error rate or level of accuracy. It is critical to the proper scientific use of any measurement technique that its reliability, validity and error rate are all known. If the error rate associated with a measurement technique is unknown then it is not possible to draw appropriate conclusions based on the measurements taken." 90He found that there did not appear to be a standardised procedure whereby the task of face or body mapping was undertaken. He also noted that the ability to compare images is affected by the quality and resolution of them, a matter explored in some detail in the evidence of Mr Porter, to which I shall turn shortly. In oral evidence he said that it had not been established "whether Professor Henneberg or anyone else can accurately discern the details he describes." He questioned the reliability of an attempt to discern characteristics from images of the quality of those in the present case. He added: "To make it clear I don't question for one moment Professor Henneberg's expertise as an anatomist, the question for me is that whether he or anyone else can make the judgments he's making on the basis of these images. It concerns me that there has been no attempt by anyone to test the validity and reliability of those kinds of observations." 91Mr Porter explained in his expert certificate that part of his work as a forensic scientist is concerned with "identification principles and methodologies," including facial identification from CCTV footage. He said in evidence that he had had wide experience in forensic photography over a period of years when he had been a member of the Australian Federal Police. He also acknowledged that he had no background in anatomy, biology or anthropology. He had the photographic material supplied to Professor Henneberg, and in his certificate he commented in some detail upon the CCTV footage. 92Like Dr Kemp, he found the footage not to be of good quality. He explained the problems with the photographic depiction of a person through distortion caused by the lens of a camera and the position of the camera in relation to that person. That latter problem arose in the present case because of the positioning of the CCTV cameras above the persons depicted and tilted down towards the ground. In his certificate he summarised his assessment of the CCTV footage in this way: "In summary, the photographic condition was generally found to be of low resolution and quality with problems with curvilinear distortion most likely caused by the lens design and rectilinear distortion caused by the camera focal plane not being parallel to the subject. These image artefacts would greatly affect the reproduction of shape including human forms. The photographic conditions are critically important when using the question (sic) images for photo-interpretation and identification purposes." 93In oral evidence he described the distortion, both curvilinear and rectilinear, as significant. In particular, the said that the effect of the distortion was a magnification of the upper body of the relevant offender compared to the lower body, which he described as a "keystoning effect." 94In his certificate, he challenged Professor Henneberg's determination that the offender in the CCTV footage was an adult male. He described that as a "wildly speculative assumption with no forensic science, imaging science or photo-interpretation basis." He added that there was nothing in the professor's expert certificate to suggest that he had given consideration "to the image distortion and other image artefacts presented in the CCTV footage." At the end of the certificate he expressed the following conclusion: "Prof Henneberg's propositions regarding his description of the POI limbs (thin), nose, face profile and head shape are dangerous assumptions that can not be supported by sound and knowledgeable photo-interpretation. Firstly, these features are not visible in the images, they are all covered by clothing (and of unknown clothing layers). Secondly, due to the level of significant image distortion (curvilinear and rectilinear) the human form represented in the CCTV images are (sic) significantly distorted and do not represent a true scale of shape and size. The level of image distortion renders any photo-interpretation of size and shape as unreliable." 95In response, Professor Henneberg said in evidence that he concluded that the offender was an adult male because, firstly, he did not look like a child and, secondly, the shape of the person "as revealed by the hang of clothing" did not have a bust or prominent waist or buttocks, whereas it had fairly large and wide shoulders and chest. He also noted that the person had a very small head in relation to the rest of the body, whereas in females the head is usually in a larger proportion in relation to the body. As to the quality of the CCTV footage, he said that it was neither better nor worse than a lot of CCTV images that he had analysed. He also said that he took into account distortion of those images in his analysis, adding that he always was conscious of photographic distortion and could observe and assess the amount of it. 96The trial judge admitted the evidence of Professor Henneberg. The first ground of appeal challenges that decision, and the second ground asserts that his Honour did not give any, or any adequate, reasons for it. It is convenient to deal with that second ground first.

Ground 2 97His Honour's reasons for admitting the evidence are not to be found in the daily transcript or, as often happens, transcribed or published as a separate judgment. Pre-trial issues, including the voir dire concerning the expert evidence, were dealt with over a period of several days before the jury was empanelled. Issues other than the admissibility of Professor Henneberg's evidence were also canvassed. 98At one stage during discussion with counsel, his Honour said that he would provide a "draft outline" of his reasons, apparently so that counsel might make submissions about them. In the event, on a subsequent day what was described as an amended version of those reasons was admitted as a voir dire exhibit. Why this procedure was adopted is unclear. However that may be, what was admitted was a twenty-four page document, entitled "Draft of matters to be taken into account on pre-trial rulings", in which his Honour expressed his reasons for decisions in relation to a number of matters that had been agitated. 99Among them were reasons for the admission of Professor Henneberg's evidence, under the heading "Identification: body mapping/identification." Over several pages the objections to the evidence were set out, and reference was made to the voir dire evidence of Dr Kemp and Mr Porter. His Honour referred to s 79 of the Evidence Act and relevant authority, to which I shall also refer later. He concluded that Professor Henneberg had requisite specialised knowledge and had made his methodology clear. He noted the critiques of that methodology by Dr Kemp and Mr Porter, but concluded that these were matters for the jury with the assistance of appropriate directions. 100At the end of the document his Honour set out a series of rulings, one of them headed "Identification: body and face types." The ruling was that the "Crown and Defence may lead evidence of similarity as to body and face types and the factual material on which that similarity evidence and the consequential opinions are based." 101Unusual as this procedure was, and notwithstanding the use of the word "draft" in its title, it appears to me that this document was intended to be the formal record of his Honour's reasons for his pre-trial rulings, including the admission of the evidence of Professor Henneberg. This ground is not made out.

Ground 1 102The first ground of appeal is that his Honour erred in admitting the evidence of Professor Henneberg to the effect that: (a) the appellant was similar in appearance to one of the robbers; (b) that the appellant was identical in appearance to one of the robbers. It will be seen that I would allow the fifth ground of appeal, that the verdicts of guilty on the two robbery counts are unreasonable. However, the first ground raises an issue of importance and most of the parties' submissions, written and oral, were devoted to it. It should be decided. 103For the purpose of this ground Mr Stratton referred to the evidence in the trial as well as that on the voir dire. Professor Henneberg gave evidence, and his expert certificate became an exhibit. Mr Porter gave evidence in the defence case. An amended certificate of Mr Porter, a somewhat expanded version of the original, also became an exhibit. Their evidence, while explored in greater detail, was broadly to the same effect as that which they had given on the voir dire. Dr Kemp also gave evidence in the defence case, but his report was not tendered and the focus of his evidence was different from that which he had given on the voir dire. I shall return to his evidence. 104However, a further expert was called in the defence case, Dr Meiya Sutisno, a forensic anatomist. Her expert certificate was admitted, and she gave oral evidence. Her evidence was also a critique of the evidence of Professor Henneberg. For the most part, it amounted to an endorsement, from the viewpoint of a forensic anatomist, of Dr Kemp's criticisms in his voir dire evidence. It also supported the views of Mr Porter. 105In her expert certificate Dr Sutisno had reviewed Professor Henneberg's findings in his certificate. Put shortly, her opinions were these:

• The professor's certificate "was found lacking information of the precise methodology or scientific protocol (either quantitative or qualitative)" for the observations of the offender which he used to arrive at the anatomical description of that person which he set out in the certificate.
• There was no explanation for "data interpretation", nor "whether or not he took into account the technological problems presented by the CCTV cameras and what measures he took to overcome these problems for accurate image analysis."
• The "accuracy and reliability" of his opinion that the offender in the hotels was the appellant was "certainly doubtful, as his opinion was deduced from merely visual observations, unsubstantiated by proper scientific quantitative or qualitative procedures." The small number of "general" features of the face and body of the offender, not constituting "unique identifying features," did not warrant the "high level of match" which he found between the offender and the appellant.
• He "failed to correctly process the images" of the offender and the appellant "for accurate comparative analysis." He did not assess the "resolution, camera angles, focal length, camera-to-subject distance and lighting conditions presented by the CCTV footage and the comparative images." The distortions produced by the CCTV cameras could "alter the true morphological characteristics of the body and facial features in both shape and size." Comparisons between the offender and the appellant should have been made only "from images of the same geometric plane or viewpoints." The forensic images of the appellant "were not sufficient for comparisons with the CCTV images" of the offender.
• It would not have been possible for him "to assess features (ie head shape, face profile and nose)" of the offender "which were not visible due to the balaclava worn over the head and face."
• Like Mr Porter, she challenged the professor's assessment of the offender as an adult male, suggesting that this was the result of "cognitive preconception whereby he subconsciously formed an impression of the fact to fit the case ... ."
• Because of the small number of features of the offender described, none of which was of "high identification value", it was inappropriate to "warrant a high level of identification match" between the offender and the appellant. The best that could be said is that the images were "inconclusive" on that question. 106Dr Sutisno was cross-examined about evidence she gave in a trial which came to be considered by the Court of Criminal Appeal in R v Tang [2006] NSWCCA 167, 65 NSWLR 681, a case to which I shall refer later. She acknowledged that she had not published any material for the purpose of peer review in relation to facial or body mapping. 107I have referred to Professor Henneberg's evidence on the voir dire about his contact with the garment industry. In the trial he said that he had run "two big anthropometric surveys for the garment industry, for clothing sizing." He also gave evidence, as he had on the voir dire, that he took into account distortion in CCTV images. He acknowledged that he had no formal qualifications in forensic photography, but said that he had experience in taking photographs in his professional work. He was familiar with curvilinear and rectilinear distortion. He said that they were easy to detect by examining whether there was distortion in the images of familiar objects, such as a door or a bookcase. He also had regard to whether there was similar distortion of the appearance of other people in the footage. 108In the course of Professor Henneberg's evidence-in-chief, his Honour received a note from the jury. It posed the question, "What proportion of the population would fit into the profile 'person of interest' taking into account an appropriate margin of error?" The Crown prosecutor put the jury's question to Professor Henneberg, and he answered: "Well the way to one can calculate this, although I have not, is to use combination of probabilities of the frequency with which each separate characteristic occurs, because all those characteristics are now occurring in the same person, and we do not have reliable statistics for Australian population describing the occurrence of those characteristics." 109He was invited to refer to the paragraph of his expert certificate, quoted at [74] above, in which he set out his description of the offender in the CCTV footage. This led to the following evidence: "A. Yes well I am referring to it but I will give the jury a very simple example. Lets assume for the purpose of this discussion that each of the characteristics described occurs in 50%, so one half, of Australian male population, it's an assumption that is making, that is unfavourable to me, in other words it will make the -- Q. It's a generous ratio. A. Yes, it's a generous ratio, yes. Q. Is that a fair way to quickly summarise it? A. Yes. So lets say heavy body build, we do combination of probability, 0.5 times 0.5 time 0.5, heavy body build 0.5. Shoulders and hips the same width so its one quarter. Prominent abdomen, one eighth. Distal segments of the limbs not thick, one sixteenth. Centripetal pattern is just a conclusion of the other traits described... ... ... I do not count centripetal pattern as an additional character because it is included in description of heavy body build and said distal limbs, and head shape is one 32. Face profile one sixty-fourth." Thus, albeit hypothetically, the incidence of the combination of features of the offender which he observed was whittled down to one sixty-fourth (1.6%) of the male population. 110Later in the trial, during the defence case, the jury's question was addressed again. In answer to it, Dr Kemp produced a document which, he said, Professor Henneberg had read and with which he had agreed. It became an exhibit. It was in the following terms: ... To answer this question we need to address two issues:

1. What is the frequency in the Australian population of the attributes noted by Professor Henneberg, and in what proportion of the population do these characteristics occur together in one individual?
2. What is the error rate associated with Professor Henneberg's technique? To take these issues in turn: What is the frequency in the Australian population of the attributes noted by Professor Henneberg, and in what proportion of the population do these characteristics occur together in one individual? We do not know the frequency with which these characteristics occur within the Australian population. Furthermore, we do not know whether the characteristic (sic) noted are independent of each other. The issue of independence is critical. If two features are statistically independent of each other, then the observation that an individual possesses one feature tells us nothing about the likelihood that this individual will also poses (sic) the second feature. In contrast, if two features are not independent of each other then they are said to be correlated and the observation that an individual possesses one feature will affect the probability that they also possess the second. For example, we know that in the Australia population earnings and gender are correlated, so the probability that a man earns more than $100,000 is much higher than the probability that a woman earns this much. In order to estimate what proportion of the population would share a particular set of features we need to know both the frequency with which each feature occurs in the population and the degree of correlation between all possible pairs of features. We do not know either of these things and therefore it is not possible to make a valid estimate of the proportion of the population who will share a particular set of features. In the absence of this information it would be very dangerous to attempt such an estimate, because even quite low correlations between features could have a large effect on our estimate.
3. What is the error rate associated with Professor Henneberg's technique? The error rate associated with a particular technique tells us how often the use of that technique will result in an error. In order to interpret the results of a technique we need to know the error rate associated with the use of that technique. For example, when using a particular test to detect cancer, we need to know how often the test results in an error and what proportion of these errors will be false positive errors (suggesting the patient has cancer when in fact they are healthy) and what proportion will be false negative errors (suggesting the patient doesn't have cancer when if fact they do). Without all this information we cannot interpret the results of the test. Information about the error rate of a procedure is obtained by evaluating the procedure under controlled conditions, for example testing patients who we know to have cancer and those we know to be healthy. No such evaluation of the procedures used by Professor Henneberg has ever been undertaken, so we do not know how often the application of this technique would result in an error. Given all this, the answer to the Jury's question must be that we are unable to say exactly what proportion of the population would fit the description provided by Professor Henneberg, and do not know the margin of error which should be applied. " 111However, to this was added another document which was admitted into evidence, and which expressed Professor Henneberg's view only. It read: "M Heneberg's answer to the Jury question. Anatomical characteristics occur with various probabilities in various populations. For example, strong body build in about 50% of Australian males. Dolichocephalic Head in 6% of Central Europeans (Henneberg 1976), but about 80% of Aboriginal Australians (Laron and Macintosh 1966). About 90% of people are right-handed. Were the anatomical traits completely independent of each other, the probability that specific traits occur in combination in the same person would be simply a product (multiplication) of individual probabilities. Correlations among traits can significantly affect calculation of combined probability. Since we do not have an extensive database for Australian population of all traits described in this case a precise enough calculation is impossible, while an estimate may turn out to be misleading. The only thing one can tell, that due to combination of various characters (sic) a probability of finding two individuals with several anatomical characters similar in the same way is low. I used the example of a person walking through a shopping mall trying to find two individuals of the same general anatomical characteristics like body build, head shape and face profile." 112When Dr Kemp gave evidence he was asked to comment upon the apparent contradiction between the agreed document and Professor Henneberg's document. He said that he saw a difference between the two, and added that the core of the agreed document was that it was not possible to estimate the frequency of certain physical features within the population "because we don't have the data available to estimate how frequently those characteristics occur within the population." He added: "I make an additional point, which I think is a very important one. It seems like a subtle statistical point, but actually it's critical - which is that even if we do think we know how frequently those features occur in the population, we would also need to know whether or not they're statistically independent of each other; ... essentially the critical question is if we note two characteristics, for example, of a person, if those two are so highly correlated that one always goes with the other, then noting two tells us no more than noting one. ... If they're completely independent of each other, then noting these two characteristics and estimating how frequently they occur in the population is useful, it adds something. Now, it's probably the case that most features have some correlation between them, that they're not independent, but we simply don't know, and that's the second critical point. So we don't know how frequently these features occur in the population, and furthermore, we also don't know whether they're independent of each other." 113In the light of this evidence, Mr Stratton examined the passage in the evidence of Professor Henneberg quoted above in which he made a hypothetical calculation of the proportion of the Australian population which might have the combination of characteristics he observed in the offender, upon the assumption that each of those characteristics occurred in half the male population. By adopting what is known in mathematics as the product rule, he arrived at a proportion of 1.6%. However, as Mr Stratton pointed out, that assumes that those characteristics were independent of each other. If there were a correlation between some of them, as one might expect, the incidence of that combination of characteristics in the general population would be significantly higher. 114Mr Stratton submitted that Professor Henneberg's answer to the jury's question, while acknowledging the significance of the difference between independent and related characteristics, was ambivalent and did not amount to a retraction of the hypothetical calculation he had made in oral evidence. Notable was his assertion in the document that the "only thing one can tell" was that "due to combination of various characters a probability of finding two individuals with several anatomical characters similar in the same way is low." 115Dr Kemp otherwise gave evidence in general terms about the lack of scientific research examining the validity, reliability and error rate of body mapping and of facial mapping, the latter being the particular expertise of Dr Sutisno. He also described some recent research using mock juries who were asked to determine if a suspect was a person depicted in video material. In that exercise they were presented with mock expert reports. The results conveyed that the mock jurors were heavily influenced by the expert opinions. As Dr Kemp put it in evidence, "their final decision is predicted not by the truth of the situation but by the expert evidence they hear." 116It was put to Professor Henneberg in cross-examination that there had been "no testing of the validation" of his opinions. To this he replied: "One of the tests includes examination in court. Well, there was occasionally. I had my colleagues or associates to assess the same CCTV and the same forensic procedure photographs independently. For example, Dr Carl Stephan did this on very few occasions but he did and it would be certainly good to have more of this happening." 117He was then taken to an occasion on which he was asked to compare images of the former politician, Ms Pauline Hanson with an image of a person said to be her published in a newspaper. He did so by examining the facial characteristics and expressed the opinion that he was "99.2%" sure that the person depicted in the newspaper was Ms Hanson. Putting it another way, he calculated that there was a 0.8% probability that two random people would have such an array of closely matching characteristics. Later in the trial, the Crown prosecutor produced material establishing that the newspaper had subsequently acknowledged that the person depicted was not Ms Hanson and published an apology to her. 118It is convenient at this point to deal with Ground 1(b) which challenges the admission of Professor Henneberg's evidence to the effect that the appellant was identical in appearance to the offender in the CCTV footage. As I have said, in evidence on the voir dire the professor denied that he purported to identify the appellant as the offender. He asserted no more than a high level of anatomical similarity. His evidence before the jury was to the same effect. 119Mr Stratton acknowledged this, but questioned whether the jury might have been left with the impression that the professor's evidence was, to all intents and purposes, an identification. He referred to the evidence already discussed in which, albeit hypothetically, the professor asserted that the combination of the characteristics of the offender would narrow him down to 1.6% of the adult male population. He also referred to that part of the professor's expert certificate which set out his methodology and, in particular, the passage quoted at [78] above dealing with his assessment of the degree of certainty or uncertainty of his "specific individual identification." That passage included the assertion, "... my statement that the identification is 95% certain is an estimate based on the fact that, due to the quality of images, I could not observe enough anatomical detail to reach certainty." 120Whatever one might make of that assertion, it is clear enough that it was a part of his description in general terms of his methodology. It was not directed to the case at hand. The professor said as much in cross-examination, describing it as a "general statement about methodology." In evidence-in-chief, he explained that he used images "to form anatomical descriptions of persons of interest, and of the suspect or suspects, in various cases. He added, "I never ... identify a person, I never state with 100% certainty that a person of interest and a suspect are the same person." 121I must say that I perceive some tension between that evidence and the manner in which the methodology is expressed in the certificate. Nevertheless, the case was conducted on the basis that the professor's evidence was an assertion of similarities, not of identification. His Honour instructed the jury accordingly by directions about which no complaint is made. Ground 1(b) is not made out. 122Let me turn, then, to ground 1(a), challenging the admission of Professor Henneberg's evidence to the effect that the appellant was similar in appearance to the offender. Mr Stratton's submission was that this evidence did not meet the test of the admissibility of opinion evidence set out in s 79 of the Evidence Act 1995. By that section opinion evidence is admissible if the person giving it "has specialised knowledge based on the person's training, study or experience," and the opinion is "wholly or substantially based on that knowledge." 123Mr Stratton referred to Tang (supra), in which the evidence of Dr Sutisno was at issue. She had given evidence of points of similarity between CCTV images of an offender and photographs of Tang, applying facial mapping and body mapping techniques. She expressed three related opinions: first, that the CCTV images and the photographs of Tang depicted the same person; second, that there was a level of support for this conclusion by the application of a six point scale derived from literature on the subject; and, third, that certain characteristics in the photographic material were "unique identifiers." 124Spigelman CJ, with whom Simpson and Adams JJ agreed, set out Dr Sutisno's methodology at [16] ff. She was dealing with CCTV footage in which the face of the relevant offender was exposed, but the quality of the images was such that a jury would not been able to make their own comparison of them with Tang's appearance. She was supplied with forensic photos of Tang. She noted individual facial and body features, a process known as morphological analysis, and compared measurements and dimensions of faces, known as photo-anthropometry. She used computer software to enhance the photographic images, together with a process of overlaying comparably enlarged images, known as photograph superimposition. She identified a number of individual characteristics, and it was these which she classified as "unique identifiers." 125The Chief Justice noted that Dr Sutisno had identified her area of expertise as a forensic anatomist specialising in facial identification, and that this expertise was reinforced by her formal research: [58]. His Honour found that this enabled her to detect similarities between the relevant images, placing her in a better position than the jury to compare them: [84]. The case against Tang was circumstantial, and evidence of his similarity to the particular offender in the CCTV footage was relevant: [83]. 126His Honour said at [120]: "... the evidence of particular similarities between the two categories of photographs of the accused and the [relevant] offender was admissible. The process of identification and magnification of stills from the videotape was a process that had to be conducted by Dr Sutisno out of court. Furthermore, the quality of the photographs derived from the videotape was such that the comparison of those stills with the photographs of the Appellant could not be left for the jury to undertake for itself. The identification of points of similarity by Dr Sutisno was based on her skill and training, particularly with respect to facial anatomy. It was also based on her experience with conducting such comparisons on a number of other occasions." 127However, one of the unique identifiers upon which Dr Sutisno had relied was the "relatively upright posture" of the person in the images and photographs: [29]. The Chief Justice noted that body mapping is an area of specialised knowledge "qualitatively different" from facial mapping: [57]. After an examination of the evidence at [58] - [79], his Honour found that Dr Sutisno did not have requisite expertise in body mapping. He concluded: "[134] Section 79 has two limbs. Under the first limb, it is necessary to identify 'specialised knowledge', derived from one of the three matters identified, i.e. 'training, study or experience'. Under the second limb, it is necessary that the opinion be 'wholly or substantially based on that knowledge'. Accordingly, it is a requirement of admissibility that the opinion be demonstrated to be based on the specialised knowledge. The Appellant invokes each limb. [135] With respect to the first limb of s 79, I have set out above the evidence of the nature of the specialised knowledge which Dr Sutisno said she had brought to bear in the formulation of the three opinions she expressed. There does appear to be a body of expertise based on facial identification. The detailed knowledge of anatomy which Dr Sutisno unquestionably had, together with her training, research and experience in the course of facial reconstruction supports her evidence of facial characteristics. [136] Nothing was presented to the Court which indicates, in any way, that Dr Sutisno's extension from facial to body mapping, with respect to matters of posture, has anything like that level of background and support. Specialist knowledge of posture can of course exist. ... But the foundation for admissibility must be lain. It was not lain in the present case. The so-called 'unique identifier' of posture was an essential element of Dr Sutisno's evidence of identity in the present case. ... [140] In the case of the Appellant the relevant evidence about posture was expressed in terms of 'upright posture of the upper torso' or similar words. The only links to any form of 'training, study or experience' was the witnesses' study of anatomy and some experience, entirely unspecified in terms of quality or extent, in comparing photographs for the purpose of comparing 'posture'. The evidence in this trial did not disclose, and did not permit a finding, that Dr Sutisno's evidence was based on a study of anatomy. That evidence barely, if at all, rose above a subjective belief and it did not, in my opinion, manifest anything of a 'specialised' character. It was not, in my opinion, shown to be 'specialised knowledge' within the meaning of s 79. 128His Honour dealt with the three opinions which Dr Sutisno had expressed at [88]. He described her evidence that the two bodies of photographs depicted the same person as "plainly an opinion." He saw the six point scale as having "no scientific basis," being "no more than a series of convenient labels, arranged in an ascending hierarchy, that state a conclusion." The characterisation of an identifier as unique, he said, "similarly has no scientific basis and is no more than an emphatic statement of an opinion." 129At [141] ff, his Honour held that each of those three opinions failed to meet the second limb of s 79 of the Evidence Act , that is, that they were wholly or substantially based on Dr Sutisno's specialised knowledge. As he put it at [146], "Facial mapping, let alone body mapping, was not shown, on the evidence in the trial, to constitute 'specialised knowledge' of a character which can support an opinion of identity." He noted that she claimed to have applied a "strict protocol" in arriving at those opinions, but did not explain the basis upon which such a protocol was developed. Indeed, she considered that information confidential and had not published it because she was in the process of patenting it: [154]. 130The Northern Territory Court of Criminal Appeal had occasion to examine evidence of Dr Sutisno in Murdoch v The Queen [2007] NTCCA 1, 167 A Crim R 329. In that case also Dr Sutisno had undertaken an exercise in facial and body mapping, comparing photos of the appellant with CCTV images and concluding that they depicted the same person. At [246] ff, the Court examined her evidence and arrived at the same conclusion as that in Tang : her evidence of similarities between the images was admissible but her positive identification of the appellant was not. As it happens, Professor Henneberg had been called for the defence at that trial and the court referred to his evidence, critical of that of Dr Sutisno, at [265] - [267]. Obviously, the admissibility of his evidence was not in issue. 131Mr Stratton took no issue with Professor Henneberg's expertise as a biological anthropologist and anatomist. He submitted, however, that the evidence did not meet the two-fold test of s 79 of the Evidence Act , as it did not involve any area of specialised knowledge based on the professor's training, study or experience, and his opinion was not wholly or substantially based upon that knowledge. 132Mr Stratton referred to the description of the offender in the CCTV footage in the professor's expert certificate, set out at [74] above. Stripped of the technical jargon, he argued, those observations were simplistic descriptions of the offender. Particularly was that so of the observations that the man was right handed and "carried himself straight", which were not observations which could be said to be based on any training, study or experience. He noted that the comparison between the images of the offender and the appellant were made purely by observation, and no measurements were taken. 133He relied upon the evidence of the witnesses in the defence case, both on the voir dire and at the trial, to submit that Professor Henneberg's opinions "did not have the indicia of a scientific opinion." He noted the lack of evidence about the validity, reliability and error rate of the professor's methods, together with the absence of any satisfactory evidence of peer review. 134Mr Stratton was critical of Professor Henneberg's oral evidence about the hypothetical incidence of a person sharing the characteristics of the offender in the Australian male population by the application of the product rule, arguing that that erroneous approach told against his qualification to give evidence in this area. The fallacy was exposed in the agreed document later furnished to the jury and, indeed, in Professor Henneberg's separate document. As to that latter document, Mr Stratton added, there was "no or very little logical or empirical basis" for the professor's estimates of the prevalence in the community of the features to which he referred. 135The Crown prosecutor submitted that Professor Henneberg did no more than give evidence of similarities in a manner expressly sanctioned by the Chief Justice in Tang . She emphasised that he did not purport to identify the appellant as the offender. She argued that Professor Henneberg was shown to have relevant expertise and, unlike Dr Sutisno in Tang , fully exposed his methodology for critical examination. The criticisms of his evidence by the other experts, she said, were properly matters for the jury to assess. She cited the observation of the Chief Justice in Tang that, for the purpose of s 79, the "focus of attention must be on the words 'specialised knowledge', not on the introduction of an extraneous idea such as 'reliability':" [137]. As she put it in oral submissions, s 79 does not create a "threshold test" for the admissibility of evidence in an area of specialised knowledge based upon "acceptance of scientific standards." All that is necessary is that the expert witness demonstrate a "knowledge which is specialised, namely, that it derives from an area beyond the experience of a lay person." 136As to the lack of measurement of the person in the CCTV footage and the appellant in the forensic photographs, the Crown prosecutor relied on Professor Henneberg's evidence that his analysis was qualitative, rather than quantitative. On the question of statistics, she submitted that the jury could not have been misled by the oral evidence the professor gave on that topic in the light of the agreed document and his own document later furnished. The effect of that material was that the incidence in the community of men sharing the characteristics of the offender could not be calculated and, indeed, in his expert certificate the professor had made it clear that statistical analysis was "not a practical approach." 137She argued that, in any event, Professor Henneberg's careful examination of the photographic material qualified him as an "ad hoc expert", relying on an observation of the Chief Justice in Tang at [120] and the cases to which his Honour there referred. That was the view taken of the evidence of Dr Sutisno of similarities between the person in the CCTV footage and the appellant in Murdoch : [296]. 138We were not referred to any appellant authority in which body mapping was subjected to critical analysis. The lack of research into the validity, reliability and error rate of the process, identified by Dr Kemp, is of concern. Professor Henneberg's use of the product rule in his hypothetical statistical calculation and his virtual identification of Ms Pauline Hanson as the person in the newspaper, which proved to be erroneous, are matters properly to be taken into account in assessing the reliability of his evidence as an expert. His assessment involves an observation of two sets of images and a comparison of anatomical features which he detects in them, without measurements and without the aid of technology such as computerised enhancement of the images and photographic superimposition, the methods adopted by Dr Sutisno in Tang . 139However, this is not the occasion to examine the science of body mapping or to undertake some general appraisal of Professor Henneberg's approach. As the Crown prosecutor rightly said, the question is whether he had specialised knowledge, beyond the reach of lay people, which he brought to bear in arriving at his opinion. That question fell to be answered by reference to the task which he undertook. 140That task was to make an anatomical comparison between relatively poor quality CCTV images of a person covered by clothing from head to foot with images of the appellant. Applying his specialised knowledge, Professor Henneberg claimed, he was able to detect not just a measure of similarity but "a high level of anatomical similarity" between the two persons. How he was able to do that when no part of the body of the offender in the CCTV images was exposed was, in my view, never satisfactorily explained. 141It may be that his experience enabled him to make appropriate adjustments for photographic distortion in the CCTV images. However, it is not apparent on the evidence how his undoubted anatomical expertise equipped him to take account of the clothing. Notably, he said that he could make observations of the "head shape, nose and face profile" of the offender, notwithstanding that his head was covered by a balaclava. True it is that that garment appeared to him to be "well adhering to the surface of the skin" and, as noted at [56] above, Mr Brooks at the Willoughby Hotel described the balaclava worn by the relevant offender as "close fitting." Nevertheless, this is no answer to the cogent criticisms of this part of the evidence by the other experts. 142Whatever might be said about Dr Sutisno's expertise in the light of the decisions in Tang and Murdoch , her critique of Professor Henneberg's evidence was, as I have said, consistent with those of Dr Kemp and Mr Porter. Among the points she made, noted at [105] above, was that it would not have been possible for the Professor to assess the features of the head of the offender, which were not visible because of the balaclava that person was wearing. Mr Porter noted that, generally, the features of the offender were not visible in the images because they were "all covered by clothing (and of unknown clothing layers)": [94] above. Dr Kemp challenged the professor's ability to "accurately discern face and head shape", noting that he could not know the characteristics of the clothing being worn, how the clothing adhered to the face and head, and whether the offender was wearing anything else underneath such clothing as was visible: [86] above. 143Professor Henneberg's evidence about his experience of the clothing industry, set out at [87] and [107] above, appears to be confined to the size and hang of garments, and their relation to "body shape and posture." This may mean that he has a facility in taking account of the clothing worn by someone in determining anatomical characteristics of the torso and limbs, although how he could assess the thickness of the clothing and whether there were multiple layers of it is not clear to me. However that may be, the evidence does not convey that his experience extends to the observation of anatomical features of the head and face of a person whose head is entirely covered by a garment such as a balaclava. 144Whatever might be made of the professor's observations of the offender's body shape through his clothing, his observations about the shape of his head and face were clearly vital to his conclusion that there was a high degree of anatomical similarity between that person and the appellant. It does not appear to me that those observations could be said to be based upon his specialised knowledge of anatomy. Generally, I am persuaded by Mr Stratton's submission that his description of the offender was "simplistic". It may well be that the jury would have required expert evidence explaining the effect of photographic distortion in the CCTV images. Subject to that, I am not persuaded that the comparison of the images of the offender with those of the appellant was a task which the jury would not have been able to undertake for themselves: cf Smith v The Queen [2001] HCA 50, 206 CLR 650. 145Indeed, with every respect to Professor Henneberg, I am of the view that his evidence raised the very problem about expert evidence in this area described by Dr Kemp at [115] above. It tended to cloak evidence of similarity in a mantle of expertise, described by Mr Stratton as a "white coat effect", which it did not deserve. 146Ground 1a is made out. This would lead to an order that there be a new trial of the robbery counts, unless the remaining evidence was insufficient to justify leaving those counts to a jury. However, I find it unnecessary to decide that matter because I am satisfied that, even with Professor Henneberg's evidence, the verdicts of guilty on those counts cannot be sustained.

Ground 5 147This brings me to ground 5, which is that the convictions on the robbery counts are unreasonable. 148In assessing the Crown's circumstantial case on the robberies, certain inferences arising from the evidence were not in dispute in this court. The conclusion that the two robberies were perpetrated by the same offenders was open to the jury. So was the conclusion that in each robbery the offenders used a red Audi of the same model as that stolen from Mosman some days earlier. The appellant's criminal involvement with that car, either as thief or receiver, was clearly established by his being found in possession of a key for it about three months after it was stolen. There was evidence that the Audi was the only vehicle of that colour and model reported stolen over the relevant period, although, of course, that is of significance only if the vehicle at the scene of the robberies was stolen. Someone burnt out the stolen vehicle after the appellant's arrest but, of itself, that evidence is equivocal as to whether the vehicle was destroyed to conceal its theft or its use in the robberies. 149The evidence that the appellant had access to the premises at Tempe searched by police was itself circumstantial, but that conclusion was also open and was not challenged before us. There police found $25,000 in cash, bundled in denominations and secured by rubber bands in accordance with the practice at the Willoughby Hotel. However, that was a practice commonly adopted in the hotel industry and, in any event, the money was found more than three months after the robberies. I should add that there was evidence, to which I have not so far referred, that the appellant had not been in full time employment over the three years prior to his arrest, and that early in February 2008 his partner and her mother were able to purchase a reasonably expensive car. 150On the other hand, there was no eye-witness identification of the robbers at either hotel. It is for this reason that the evidence of Professor Henneberg loomed large. However, in the light of the matters to which I have already referred and, in particular, the compelling critique of the professor's approach in the evidence of the other experts, I am satisfied that it was not open to the jury to find the high degree of anatomical similarity between the offender and the appellant for which he contended. At most, it established a number of similarities between the two which were unremarkable, individually or in combination. That being so, I consider that his evidence was of little weight. Nor do I consider that the jury, left to themselves, could have found any significant similarities between the CCTV images and the photographs of the appellant. 151Of course, such descriptions as the witnesses at the hotels were able to give of the robber said to be the appellant were confined to his height and build. They were of such generality as to be of little or no assistance to the Crown case. Mr Stratton noted, however, that the witnesses' estimate of the height of that man varied between 5' 5" and 5' 9", or 165 to 175 cms. Some of the forensic photographs of the appellant show him standing next to a scale, from which it appears that his height is a little over 180 cms but less than 190 cms, that is, of the order of 6' or a little more. Obviously, the witnesses' estimate of the height of the robber in question are unlikely to be accurate. However, to most of us a person of 6' or more is noticeably tall, and it is of some significance that no-one described the robber in that way. 152As Mr Stratton properly acknowledged, the evidence raises significant suspicion that the appellant was involved in the robberies and, certainly, points to his having been up to no good in some unspecified way in the period leading to his arrest. Nevertheless, on the robbery counts I am satisfied that the jury ought to have had a reasonable doubt. This was not a case turning upon the credibility of witnesses whom the jury had the advantage of having seen and heard. It turned primarily upon inferences from uncontested evidence, together with disputed expert evidence. As with the counts of stealing and receiving the Audi, this court is as well placed as the jury to assess that evidence. It is a case in which "a doubt experienced by an appellate court will be a doubt which the jury ought also to have experienced": M v The Queen (1994) 181 CLR 487, in the joint judgment at 494. 153Since preparing these reasons I have read in draft the judgment of Beazley JA. Her Honour has examined the evidence relating to the robberies in greater detail and has concluded that, absent the evidence of Professor Henneberg, there was an adequate circumstantial case against the appellant. However, in the light of the whole of the material, I remain of the view that the evidence does not rise above the level of significant suspicion. Of course, it is consistent with the appellant's guilt but, in my view, it does not prove it to the requisite standard. 154Accordingly, on each of the third and fourth counts I would quash the conviction and enter a verdict and judgment of acquittal. This outcome does not entitle the appellant to immediate release because he is not yet due for release in respect of the sentence passed upon him for the sixth count, possessing the proceeds of crime, which was not the subject of appeal. 155HARRISON J: I have had the advantage of reading in draft the judgments of both Beazley JA and Hidden J. I agree with Hidden J's detailed analysis of the appellant's challenge to the evidence given by Professor Henneberg and with his conclusion that the appellant must succeed on ground 1(a). However, I agree with Beazley JA that ground 5, in which the appellant contended that his convictions on the robbery counts were unreasonable and unsupported by the evidence, should be rejected and that in the light of the appellant's success on ground 1(a), the verdict should be quashed and there should be a new trial.

DISCLAIMER - Every effort has been made to comply with suppression orders or statutory provisions prohibiting publication that may apply to this judgment or decision. The onus remains on any person using material in the judgment or decision to ensure that the intended use of that material does not breach any such order or provision. Further enquiries may be directed to the Registry of the Court or Tribunal in which it was generated. Decision last updated: 12 December 2011