LORD JUSTICE TUCKEY: On 5th December 1977 in the Worcester Crown Court before Mr Justice Stephen Brown and a jury, Harold Percival Williams was convicted of murder and sentenced to life imprisonment. On 15th February 1979 this court (Lord Justice Waller, Mr Justice Lawson and Mr Justice Jupp) refused his application for leave to appeal against conviction. On 3rd August 2000 the Criminal Cases Review Commission referred his conviction to this court under section 9 of the Criminal Appeal Act 1995. Mr Williams died on 9th October 2000 but approval under section 44A of the Criminal Appeal Act 1968 has been granted by the single judge to his sister to continue the appeal. We shall refer to Mr Williams as the appellant.

Margaret Davies was murdered some time between 9.35 pm on Friday 21st and 2.30 pm on Sunday 23rd January 1977. The appellant was in a relationship but was not living with her at the time she was killed. He had a good alibi for the Friday evening up to about 11.50 pm and from 7.45 am the following morning. The Crown's case was that he killed the deceased between these times. The case against him was circumstantial. The time of death was therefore crucial. The Reference by the Commission has primarily been made to enable us to consider new evidence which is said to cast doubt on the evidence given at trial about the time of death.

The deceased lived at 76 Whittern Way in Hereford. At the time of her death she was 40. She was divorced, a heavy drinker and had had a number of relationships with various men. She met the appellant, then about 44, himself much married, at the end of 1975, and he moved in with her in early 1976. Their relationship was a stormy and violent one. However, on 3rd January 1977 he was made to leave after the deceased alleged that he had assaulted her. He was charged with assault occasioning actual bodily harm to which he later pleaded guilty and it was a condition of his bail that he should not contact or associate with the deceased and should live at his parents' address in Hereford, a couple of miles away. There was, however, continuing antagonism between the two of them after this separation.

At 9.35 pm on the Friday evening, a woman who knew the deceased well met her coming out of the fish and chip shop near to her house. The appellant was employed as a coach driver. On the Saturday morning at 7.45 his employer went to open up the garage when, to his surprise, he saw the appellant's car parked in the yard with its engine running. A few minutes later he went over to the car. Its engine was still running and the appellant appeared to be asleep. The employer opened the door and the appellant woke up with a start, saying he had just been having a doze.

On the Sunday the deceased's daughter let herself in to her mother's house at about 2.30 pm. The living room curtains were drawn. Her mother was dead upstairs on her bed. She was clothed, lying on her back and gagged. Her hands and feet were tied with one continuous piece of string. Her upper garments had been pulled up exposing her left breast. She had been stabbed. The police arrived at about 3.00 pm and they noticed that there were papers fanned out on the floor in the hall below a table. The rear door and the door leading to the front were locked from the inside. There was no sign of forcible entry and apart from the papers on the floor the house was undisturbed.

The home office pathologist, Dr Davies, arrived at 5.45 pm. He noted a ligature mark around the deceased's neck. The body was cold, with rigormortis fully developed. At 6.00 pm he took the body temperature and the room temperature. He noted marked signs of putrefaction. On turning the body blood issued from the stab wounds. He concluded that the cause of death was strangulation by ligature. There were six stab wounds and four puncture wounds on the upper part of the deceased's body which Dr Davies considered had been inflicted after death. One stab wound had divided the third rib which indicated use of considerable force. A duster had been stuffed firmly into the deceased's mouth and another duster had been tied around to hold it in place. Bruising to the back of the throat suggested that this had occurred before death. It is apparent from what we have said that whatever the cause of death, this was a violent and frenzied killing.

The Reference has been has focused on two aspects of Dr Davies' evidence: the time of death and the nature of the weapon used to inflict the wounds. In his autopsy report Dr Davies said that death appeared to have taken place about two days before he saw the body. The wounds had been inflicted with two types of weapon -- a sharp pointed knife and a bodkin-type weapon. We shall return to consider what Dr Davies said at trial about these two matters later in this judgment.

The appellant was interviewed by police on the Sunday afternoon at his parents' home, at about 4.30 pm. When asked when he last saw the deceased he replied not since he was put on bail. She had telephoned him at work but he had had nothing to do with her. He told the officers he had something to show them and returned with a pad of paper which included a reference to his magistrates hearing and a memo dated 20th January to the effect that the deceased had telephoned him at work. He said he made the note to show she kept trying to contact him. He presumed she wanted money for bills. He had paid a lot of her debts. He also supplied the police with a list of names and addresses of her associates. He gave an accurate account of his movements for the Saturday after 8.00 am. He then informed that the deceased was dead. He asked if the police were joking and said he did not believe it. He put his head in his hands and started to cry. After he had composed himself he said he loved her and asked "How did she die?" and "Was she on the bed?" He said he blamed the doctor for giving her pills. He said she could be a bitch when on the drink but at other times she was a lovely woman. He alleged that she had threatened to desecrate his son's grave.

He was taken to the police station and further interviewed that day and a number of times over the following days and weeks until he was charged with murder in April. In the course of these interviews he described his movements between leaving work on the Friday evening and the following morning. He said that after a meal and a visit to the Friar Tuck Chip Shop, he had driven at about 11.15 pm to a second chip shop in Grandstand Road which was on the other side of Hereford from where the deceased lived. He had then driven home to his parents' house, getting there at about 12.30 am. The police found no one at the Grandstand Road chip shop (which closed at midnight) who had seen him. When this was put to him the appellant named the assistant whom he said had served him, but this assistant was not on duty that night. Nevertheless, the appellant maintained his account of this visit through several interviews and up to trial.

The deceased's address book and other personal papers were found by police in the appellant's car. He said he had taken them because she had taken his address book and cheque book. But neither of these items were found in the deceased's house and an address book belonging to the appellant was found when the police searched his parents' house.

In the course of their enquiries the police searched a derelict car belonging to the appellant which he kept in the yard at work. A thorough search on 24th January revealed a table knife (exhibit 26) in the glove compartment. Six days later the boot was searched again and a fishing bag was found in or near to which was a bent knitting needle. These items had not been in the car earlier. In interview the appellant said that the knife had been in his car for a long time and he used the needle to pierce eels through the neck after he had caught them. No blood was found on the knife or the needle or on the appellant's clothing. The only scientific evidence to link the appellant to the killing was that a rib pattern in the blood on the deceased's body and on her bedspread could have been caused by the fabric of the trousers which the Crown alleged the appellant had been wearing that evening and which were found by police to have been freshly cleaned and pressed. The knots used to truss the deceased were those used by hop pickers. The appellant had worked as a hop picker.

At the trial, which lasted 14 days, the appellant was represented by leading and junior counsel. It is clear from the transcripts which we have seen that he was vigorously defended. It is common ground that the defence were assisted at trial by Dr, now Professor, Bernard Knight, a pathologist from whom they had a report which said:

We have seen the full transcript of Dr Davies' evidence. His estimate of the time of death was based on the deceased's body temperature and the putrefaction. He made it clear that, as he put it:

Dr Davies was also asked about the stab wounds and said that they could have been caused by the knife (exhibit 26) and the knitting needle. He maintained this opinion despite the fact that he had said in the autopsy report that the deep wounds had been made by a sharp pointed object and that his measurements of the length of these wounds was less than the width of the blade of a knife. He explained the latter discrepancy by saying that he had only used a foot rule, so his measurements might have been inaccurate.

The Crown called Dawn Mapp, a 16 year old schoolgirl at the time of the murder. She was a neighbour and had made a statement to the police in which she said she had seen the deceased in the street at 2.00 pm on the Saturday. However, on 28th March 1977 she made a further statement saying that having thought about it, it could well have been the Friday afternoon that she had seen the deceased. On the Saturday she remembered watching a film on television until about 3.30 pm and she would not have gone out of her house before she had finished watching the film. She gave evidence to the latter effect at trial, although we have not seen a transcript. The defence had both her statements and we do not know what questions they asked her, but it is apparent that they at least kept open the option that she was right in her first statement, although, as we have said, they had made the tactical decision to try and push the time of death back as far as possible.

At the close of the Crown's case the defence submitted that there was no case to answer. The judge disagreed. The defendant elected not to give evidence but made a statement from the dock. What he said was:

The judge summed the case up to the jury very fully and fairly. No complaint is now made about his directions of law. The jury convicted the appellant after a six hour retirement.

On the first appeal to this court leading counsel argued that the judge should have allowed the submission of no case to answer. The circumstantial evidence, it was submitted, did not establish guilt beyond reasonable doubt. It was, as counsel submitted, a lurking doubt case. Lord Justice Waller's fifteen page judgment makes it clear that the court gave full consideration to all the arguments which were raised about the quality of the circumstantial evidence. The court noted that the case depended not on just one small point of circumstantial evidence but on an accumulation of points. This led the court to the view that there clearly was a case to go to the jury and to say "that in the mind of each of us there is no lurking doubt."

So we turn to the new evidence upon which the Reference is based, the thrust of which is to show that the verdict was unsafe because the deceased may not have died until after 7.45 am on the Saturday. It comes in the shape of a lay witness, Mrs Hall, who says she saw the deceased alive on the Saturday morning, and evidence from the distinguished pathologist Professor Crane which it is argued casts doubt on Dr Davies' evidence as a whole, including his evidence about the time of death.

We heard the evidence from Mrs Hall without ruling on whether we should receive it. We admitted the evidence of Professor Crane without objection from the Crown and heard briefly from him. We were also asked to admit further evidence from Miss Dawn Mapp, now Johnson, but refused to do so. She made a further statement in September 2002 saying that her first statement was correct and that she had been tricked or pressurised by police into making her second statement. This is not the sort of evidence which this court should admit on appeal. It involves an admission that the evidence which the witness gave at trial was untrue - to put it bluntly, she had committed perjury. This court is now asked to admit further evidence from her to explain why it was that she perjured herself in circumstances where the senior police officers involved deny her allegations about how she came to make her second statement. At best if we had admitted her evidence we would have been left with a series of contradictory statements, all of which had been made under the usual caution about the need to tell the truth, one of which had been made on oath, and another of which had been made in hotly disputed circumstances. Analysed in terms of section 23 of the Criminal Appeal Act 1968 such evidence does not appear to be capable of belief and for that reason we refused to receive it.

The evidence of Mrs Hall did not fall into this category. Although there may be good technical section 23 reasons for refusing to admit her evidence based on the fact that for tactical reasons the defence might not have wanted to rely on her evidence at trial, we do not think this is a good enough reason for refusing to receive it on this Reference, for the simple reason that it appears the defence were not aware of Mrs Hall's evidence at the time. What Mrs Hall told us was that she was seen several times by police officers in the days following the murder and that she made a written statement to them. This was a handwritten statement and she does not recall being asked to sign a typed copy, which is the usual police practice. This may explain why the police have not been able to produce any early statement from Mrs Hall and why, apparently, the defence, who it is accepted had unlimited access to the relevant police records before trial, did not come across any such statement either.

Mrs Hall lived with her husband and four children three doors down from the deceased whom she saw in the street almost every day. She also had an exacting job as a bingo caller working Mondays to Fridays 1.30 to 3.00 pm and then 6.30 to 9.00 pm, plus Sunday afternoons. She did not work on Saturdays. She says that on the Sunday, after returning from work, she was unable to park because of police vehicles and general commotion. At some time, which she cannot now remember, she learned of the murder; nor could she remember, although it must have been much later, when she first became aware of the significance of when it was she had last seen the deceased. At all events, she told us that she was sure that she had seen the deceased on the Saturday. That morning she had just come out of her house and was going up to the shop. As she went towards the shop the deceased came down from the direction of the shop. The deceased was wearing a beige mac which had an attached hood with fur around it and was bent down. Mrs Hall thought the deceased was going home. They passed one another on the path. "We just said 'hello'. I went to the shop which was two minutes up the road." This was late morning "around 11.00-ish, something like that. I went to work at about 1 o'clock."

There are a number of features about Mrs Hall's evidence which call for comment. First, she clearly told us that she went to work after seeing the deceased, but she did not work on Saturdays. When this was pointed out Mrs Hall did not seem to remember what she had said about going to work and then maintained that she had not done so because it was a Saturday. Next, in the first written statement by Mrs Hall, which we have seen which is dated 9th October 1991, she said that she had seen the deceased on the Saturday some time during late morning, early afternoon, and "I had seen Margaret come out of the front door of her house, walk up her garden path, from where she turned left and walked up the street towards the shop ... I did not actually see her face. I was satisfied that it was Margaret because the person that I had seen leaving the front door of her house had been wearing the same type of coat that Margaret Davies always wore." This is a significantly different account from the one she gave us of passing on the deceased on the path somewhere away from but going towards her home. To be fair to Mrs Hall in a statement made last year she said that the earlier statement was wrong. But wherever or whenever Mrs Hall last saw the deceased, she did not actually see her face. Finally, Mrs Hall told us she saw the deceased almost every day. She does not know when she first became aware of the significance of seeing her on the Saturday.

We should not, however, make our final assessment of Mrs Hall's evidence without first considering Professor Crane's evidence about the time of death. He was first asked to consider this case in 1998. In a letter to the Commission, which he sent in April 1998, after commenting on how one could calculate the time of death by various methods and that he had used such methods in this case, he said:

In his reports and in his evidence yesterday Professor Crane stressed that a pathologist cannot give you a time of death with any degree of certainty. Body temperature is the best guide but it has to be taken accurately. Minor inaccuracy can significantly affect the assessment. It is taken rectally and this is often difficult. Mr Juckes, counsel for the appellant, argued that one could have no confidence in the temperature which Dr Davies took, given Professor Crane's other criticisms of his methods and conclusions to which we will come shortly. But we think there is no reason to doubt the accuracy of the temperature taken by Dr Davies. In a case such as this the approximate time of death is a vital piece of information to aid the investigation. An experienced pathologist, as Dr Davies was, will want to ensure that his assessment is based upon an accurate measurement and will have had much experience of overcoming the difficulties of taking the temperature.

So what it comes to is this. All three pathologists who have considered this case are agreed that it is unlikely that the deceased died after 6.00 am on the Saturday morning. Professor Crane yesterday said that he stood by the passage in his letter which we have quoted where he uses the words "very remote possibility, if indeed even that."

Looking at Mrs Hall's evidence in the light of the consensus between the pathologists, it is improbable that she saw the deceased at the time that she said she did, quite apart from her evidence suffering from the defects to which we have already referred. We make it clear that we do not think she is lying. She honestly believes she saw the deceased on the Saturday. But we think, for the reasons we have given, that she is mistaken about this.

At this point we must remind ourselves what this court's approach should be in a case where it has heard new evidence. That was clearly spelt out in R v Pendleton [2002] 1 Cr.App.R 441 at paragraphs 17, 18 and 19. Lord Bingham said that new evidence admitted by this court fell into three categories. The first was where the court concludes without doubt that the evidence cannot be accepted or cannot afford a ground for allowing the appeal. At the other extreme the evidence might conclusively lead to the appeal being allowed. In between lay the most difficult cases. Here, the court had to bear in mind that it was not and should never become the primary decision-maker. Its concern as an appeal court was with the safety of the conviction which was a decision of a jury. The Court of Appeal will have an imperfect and incomplete understanding of the full processes which led the jury to convict, so in making its assessment:

We must make this assessment in the light of all the new evidence and so there is one further aspect of Professor Crane's evidence to which we must refer. That is his evidence about whether the knife (exhibit 26) could have caused the six stab wounds found on the deceased's body. His view was that this was unlikely. The knife was not sharp or strong enough to have made them. The puncture wounds could have been made with the sharp tip of a knife, as well as by a needle.

Mr Juckes says that this evidence undermines the evidence which Dr Davies gave at trial, which added unjustified weight to the fact that the police had found exhibit 26 and the needle in the appellant's car. The impact of that evidence would have been considerably reduced if the jury had been told that it was unlikely that the knife had made the wounds.

We can well understand these submissions but Professor Crane's view has had to be formed without the obvious advantage of seeing the body. He has had to work from (as we understand it) photocopies of the mortuary photographs. Dr Knight was present at the trial. If he had really thought that Dr Davies had gone over the top about this, the defence could and would have called him. As it was, they were able strongly to challenge this part of his evidence for exactly the same reasons as those advanced by Professor Crane.

Mr Juckes further submitted that the scientific evidence about the trousers was very inconclusive and had been given in more assertive terms than such evidence would be given today. He also made a series of points similar to those made to this court in 1979 about the significance of the matters relied on by the Crown in support of their case. His overall submission was that this was a thin case to start with and the new evidence now made it one where this court should say that the conviction was unsafe. We have taken all these submissions into account in reaching our decision.

Like the court in 1979, we do not agree that this was a weak case. On the contrary, it was cumulatively a strong circumstantial case. The killer was known to the deceased. Either he got into the house with a key or she let him in. The way in which the deceased was killed suggested, as the trial judge said, "a remarkable degree of emotion, hatred or passion". The appellant had a strong and violent relationship with the deceased which had clearly not come to an end. The killer had obviously searched the deceased's papers in the hall but not been interested in anything else in the house. The appellant's address book was found at his parents' home and yet he had claimed that the deceased had taken it. His question "Was she on the bed?" when first told of her death was significant. His persistent lies about going to the chip shop in Grandstand Road were telling. This was very crucial if in fact he had gone from the Friar Tuck Chip Shop to the deceased's house. His explanation for the lie was unconvincing. The fact that he was found asleep in his car with the engine running the following morning, at a time when his employer was surprised to see him, added to the picture. There was no evidence that anyone else was suspected of the killing.

Against this background we do not think that the evidence of Mrs Hall or Professor Crane might reasonably have affected the trial jury's decision to convict. For the reasons we have given, supported as they are by the three pathologists' view of the time of death, we think the jury would, like us, have concluded that Mrs Hall must have been mistaken. Professor Crane's views about the knife would have added little to the evidence and argument which the jury already had.

We understand the strong feelings of those who believe that the appellant was wrongly convicted, but in this country we pride ourselves on our system of trial by jury and a jury's verdict should not lightly be set aside. The appellant was convicted after a full and fair trial, on what was strong circumstantial evidence. We do not think the new evidence would have affected the outcome. So we conclude that the conviction was safe and that this appeal must therefore be dismissed.