202302376 B1 - [2025] EWCA Crim 1036

On 16 December 1998, in the Crown Court at St. Albans (Mr Justice Gage), the appellant was convicted of the murder of Janice-Cartwright Gilbert on 28 February 1997.

Earlier, on 15 May 1998, in the Crown Court at Luton, the appellant had pleaded guilty to five counts of burglary and on 2 December 1998, before the same Court, he pleaded guilty to a sixth offence of burglary. On 17 December 1998, he pleaded guilty to 21 other offences on other indictments. On the same date, he was sentenced to imprisonment for life for murder and concurrent terms of imprisonment of varying lengths for the other offences.

On the 17January 2000, the full Court refused a renewed application for leave to appeal his conviction for murder.

On 1 March 2000, the appellant made an application to the Criminal Cases Review Commission (CCRC) for his case to be referred back to the Court of Appeal; this was rejected.

On 20 November 2017, the appellant made a further application to the CCRC. This resulted in a reference by the CCRC. On 15 July 2021, the full Court quashed the conviction for murder and lifted the order that the two offences of common assault should lie on the file (Counts 19 & 20 of from indictment 1 of 1997/1998) and ordered that the appellant be retried on one count of murder and be re-arraigned on two otherwise unrelated counts of common assault which were relied on as evidence of violent propensity.

On 19 June 2023, in the Crown Court at Aylesbury the appellant (then aged 50) was convicted of the murder of Janice-Cartwright Gilbert on 28 February 1997 for a second time. The appellant had entered guilty pleas to the two counts of common assault on 25 April 2023.

On 25 July 2023, before the same Court, the appellant was sentenced to imprisonment for life; the period of 16 years was specified as the minimum term under s.322 Sentencing Act 2020 for the offence of murder. He was sentenced to one week imprisonment concurrent for each common assault offence.

The two trials were very different. In the 1998 trial the prosecution relied on evidence from supposed experts who compared the marks left on the deceased when her killer stamped on her face, with the sole of the appellant’s right Nike trainer. At the first trial the forensic evidence stated, wrongly, that there was conclusive proof that the marks had been made by the appellant’s trainer. That evidence was the main basis of the prosecution case at that trial and its flawed nature explains why that conviction was quashed. That first trial was much shorter than the second trial.

Forensic evidence at the second trial stated that no blood or saliva had been found on the appellant’s trainers in thorough testing in both 1997 and 2021. The trainers did not appear to have been washed. Footwear mark analysis evidence, which was in effect agreed, showed that there was moderately strong support for the proposition that the marks on the face were made by the appellant’s right trainer. There was a visible pattern in the marks which was consistent with them being made by a size 6 training shoe of the type worn by the appellant. His shoes were not new, and there was wear on their soles which was consistent with some signs in the marks found on the face, but the definition seen in the skin was not clear enough to conclude that they were made by that shoe. The shoe was quite a common type.

The main difference between the trial in 1998 and the trial in 2023 was the prosecution reliance on the alleged confessions made by the appellant to Mr Dunne in June 1997 when they shared a cell at Bedford Prison. These had been available at the time of the first trial, but the prosecution decided not to rely on them. Mr Dunne was not called as a witness, and by the time of the second trial he was dead. The prosecution changed its mind. When he was alive he was said to be unreliable, but after his death he was put forward as reliable. These decisions were taken many years apart and by different people, so this is not a critical observation, and the prosecution were entitled to re-evaluate the case in the light of the quashing of the conviction and the direction for a retrial.

The original position taken by the prosecution does have consequences. Mr Dunne’s evidence was not the subject of a full disclosure process while everything was still available. He was then not cross-examined in 1998 when that material could have been put to him. If the conviction had been quashed and a retrial ordered after his death then the application to adduce his evidence as hearsay would have related to transcripts of his evidence and all the material which had been deployed would have remained available. Because he was regarded as unreliable in 1998 none of that happened, and the hearsay application related only to his witness statement.

A further consequence was that a great deal of evidence was adduced in the second trial which was designed to show that the confession was reliable, or alternatively that it was not. For example, in the first trial it had been accepted that nothing had been stolen by the murderer, but in the second trial the prosecution sought to prove that the appellant’s alleged confession that he had gone there to burgle and had in fact stolen some jewellery was true. Other evidence was examined in exhaustive detail to attempt to prove that, as he was said to have confessed, the appellant had been on his way to his sister’s at the time of the murder. There was no direct evidence about these matters, which only became prominent after the retrial was directed and a decision to rely on Mr Dunne was taken. This was well over 20 years after the event, and certainty on these matters was likely to be elusive. These attempts resulted in a long trial, and a very lengthy summing up, which runs to 230 pages of transcript before the jury retired to consider its verdict.