On 29^th April 2002, in the Crown Court at Preston, before Penry-Davey J and a jury, this appellant (then aged 18, having been born on 23^rd August 1983) was convicted of murder, attempted murder and possession of a firearm. For murder, he was sentenced to be detained pending the determination of Her Majesty’s pleasure with the specified minimum term of 12 years, less time spent on remand. Concurrent terms of 10 years and 7 years detention were imposed for the other offences. On 25^th May 2004, a renewed application for leave to appeal against conviction was refused by the full court (Clarke LJ, Jack J and Judge Fabian Evans): see [2004] EWCA Crim 1471. The appellant has since been released on licence.

In February 2005, an application was made to the Criminal Cases Review Commission (“CCRC”) who, by decision dated 29^th January 2007, determined that there was no basis on which the conviction could be referred to the Court of Appeal. At some stage, the Innocence Project and Pro Bono Unit attached to Cardiff Law School became involved in pursuing the matter on the appellant’s behalf and, in the light of new scientific evidence as to the significance of particles said to be gunshot residue which is reflected in R v George (Barry) [2007] EWCA Crim 2722, and a recent decision in relation to voice recognition evidence, R v Flynn and St John [2008] EWCA Crim 970, a further application to the CCRC was submitted.

Having obtained its own scientific evidence, on 8^th November 2013, the CCRC referred these convictions to this court pursuant to the provisions of s. 9 of the Criminal Appeal Act 1995 on the grounds that there is a real possibility of the court overturning the convictions, on the basis that the evidence of gunshot residue does not now attract the value attributed to it at trial, and therefore does not support the identification evidence. The appellant seeks leave to pursue further grounds of appeal relating to the admissibility of the voice identification evidence and the directions surrounding that evidence which followed.

On the evening of 25^th July 2001 at Miles Platting, Manchester, Daniel Dale was fatally injured and Darren Thomas was wounded in the hand by shots fired from the same gun.

The appellant was originally arraigned with three others: Ryan Brown, his brother Nathan Loftus, and Arron Cunningham. Before the trial, Cunningham pleaded guilty to possession of a firearm with intent to endanger life, possessing ammunition without a certificate, and assisting offenders. He went on to give evidence for the prosecution. Loftus changed his plea of not guilty to guilty of possessing a firearm with intent to endanger life. He was sentenced to 5 years in a young offender institution.

Brown, whose defence was alibi, was acquitted of murder and attempted murder but convicted of wounding with intent (later quashed as inconsistent with the acquittals) and possession of a firearm with intent (for which he was sentenced to 8 years’ detention, reduced to 7 years on appeal).

In short, the prosecution alleged that the appellant and Ryan Brown were responsible for the shooting, using a Walther PPK self-loading pistol which was recovered from the house of Cunningham. Cunningham said that he was minding the gun for Loftus, who had telephoned to say that the appellant would collect the gun. He described how a car arrived with the appellant driving. Brown got out of the car and collected the gun from him at the door of his house. The offence of possession of the firearm with intent to endanger life to which Loftus pleaded guilty was admitted pursuant to s. 74 of the Police and Criminal Evidence Act 1984. The defence of the appellant was one of alibi.

The evidence against the appellant at trial (which was entirely circumstantial) can be described as being based on four limbs or pillars, namely the factual background to the relationships between the various participants leading up to the shooting; the evidence of Cunningham; the evidence of voice identification; and the presence of gunshot residue on a coat found at the appellant’s home. The appeal relates strictly to the last two limbs but it is necessary to provide the context.

The background to the shooting was placed before the jury principally by way of a number of formal admissions by all three defendants; in addition, there was both CCTV and witness evidence all leading to the inference that the shooting was the outcome of gang rivalry.

In January 2001, Paul Ward was murdered. Between 23^rd and 25^th July 2001, a youth named Sheldon Keatings was tried at Manchester Crown Court in connection with that murder. Keatings was subsequently acquitted. It was admitted that the appellant was present at court on the 24^th and 25^th July and that on 24^th July, Keatings was punched in the face by Leon Critchley, a friend of Ward.

At about 4.30 pm on 25^th July, the appellant, Keatings and others left the court and travelled in two cars, a red Mazda and a silver Honda Prelude, to the Powerhouse Gym in Collyhurst. It was also admitted that the appellant, Keatings and others left the gym in the same cars and followed Critchley, who was riding a moped, to the New Allen Street area. The occupants of the cars were there threatened by a group of youths armed with various weapons. Critchley broke the driver’s window of the Mazda with his crash helmet. Thereafter, both cars then left the New Allen Street area and were driven to Ruskington Drive, Harpurhey. The prosecution case was that the appellant and Brown then acted together with others as part of a joint enterprise in the shootings of Dale and Thomas.

Turning to the evidence that Cunningham gave, two aspects of what he said were capable of linking the appellant to the shooting: these were his report of two telephone calls and his visual identification of the appellant. As to the first, it was admitted that at 7.26 pm and 7.27 pm on 25^th July, two telephone calls were made to Cunningham’s mobile telephone. In evidence, Cunningham said that both calls were from Loftus. In the first call, Loftus said that he was coming to get the “thing,” by which he meant the gun that Cunningham was keeping for him. In the second call, Loftus said that he was not coming, but the appellant was: the conversation (including the identification by name of the appellant) was held to be admissible.

In that context, the admission of the fact that Loftus had pleaded guilty was said to be relevant to this element of Cunningham’s evidence. The prosecution submitted that this was a necessary link in the evidential chain, which may be treated by the jury as confirming Cunningham’s evidence as to the involvement of Loftus. The appellant’s counsel submitted that it was irrelevant, or alternatively of only marginal relevance and should be excluded under s.78 of the Police and Criminal Evidence Act 1984. Penry-Davey J allowed the prosecution to adduce the guilty plea. He ruled that it was relevant to Cunningham’s account, which was very much in issue, and that no unfairness was caused by its admission.

Turning to Cunningham’s visual identification of the appellant, he said that 15 or 20 minutes after the second telephone call, a red car drew up outside his house. Cunningham said that the appellant was driving the car. In cross-examination, he accepted that he had a momentary glimpse of the driver and could not be sure that it was the appellant. It had been a black man who resembled him. In re-examination, he said that he had presumed that it was the appellant who was driving the car.

Brown came to the door and Cunningham handed the gun to him, wrapped in a plastic bag. Cunningham was certain about this identification. He had known Brown for a long time and saw him face-to-face. He later identified Brown at an identification parade. Brown got back into the front passenger seat of the car. As the car was driven away, Cunningham noticed two other black males sitting in the back of the car, but was unable to identify them.

Turning to the voice recognition, Stuart Shaw was among those called by the prosecution as eyewitnesses to the shooting incident. He said that on 25^th July he went towards New Allen Street. A burgundy Honda Civic car came down Osbourne Street. The car turned towards its side and two people in black got out from the passenger side at the front and the rear. Their faces were covered and they wore black gloves. They ran towards the group and raised their hands. Everybody ran. He heard one shot as he got to Keel Close. On Farnborough Road, he noticed the Honda Civic coming around the corner. He froze. It sped towards them and he shouted “Run.” He ran through the alleyway. He heard somebody shout loudly “You’re dead now.” He said that it was the appellant’s voice. He saw an arm raised and heard a gunshot, followed by Thomas shouting “Oh my hand”. He carried on running towards Nuneaton Drive and his aunt’s garden where he hid in the porch. He saw an arm and a gun, and then heard another gunshot. The gunman was roughly the height of the fence. He saw the shot hit Dale, who put his hand to his back.

In cross-examination, Shaw agreed that he and the appellant had both been at the same school together in May-June 1998. The appellant was not his friend. He had not mixed with the appellant since they left school and had never spoken to him. He had seen him about twice in four years but had not spoken to him on those occasions. He had once heard him talking outside a shop. He could not remember when that was and did not know what they were talking about. He told the police soon after the incident of what he then described as a coloured person’s voice, because he told the officer he was not sure about it. He could not be sure that it was the appellant’s voice. He agreed that in his statement he made no reference to the two men who got out of the car wearing gloves, but had referred to the gunman, whose arm he saw later, wearing gloves. He said that was more likely to be right. The gunman was probably over six foot in height. In re-examination, he said that he thought the voice was the appellant’s but was not sure about that.

Against that background the evidence of gunshot residue falls to be considered. When the appellant was arrested at his home on 21^st August, the police found a coat stored under the stairs. It was a black Henri Lloyd hooded jacket, size XL. The appellant asserted that it did not belong to him. Subsequent analysis found that it bore gunshot residue. David Collins of the Forensic Science Service (“FSS”) gave evidence for the prosecution. A transcript of his testimony is not available; we rely on his report and the summary of his evidence provided in the summing up by Penry-Davey J.

Mr Collins said that on the discharge of a firearm, a great deal of gunshot residue could come from the muzzle or the breech, depending on the construction of the weapon. The residue was a fine dust that could settle on persons or objects close to the firing gun. If it settled on clothing which was worn, it could disappear in a day or two, but if the clothing were left undisturbed the residue may be detectable for a long period. In his reports and witness statements, Mr Collins referred to finding two particles containing lead, barium, and antimony; one particle containing barium and aluminium on the front of the coat; and one particle containing barium and aluminium in the pocket. He found particles containing lead, barium, antimony and aluminium on the spent cartridges at the scene. He concluded that the coat had an association with a shooting incident, but it was not possible to establish a link with the shooting of Dale. The prosecution asserted that this was evidence supportive of the appellant having been the gunman.

The defence asserted that the particles could have arisen from sources other than the shooting. They had the benefit of expert evidence from Dr Renshaw who took the view that it would be unsafe to link the gunshot residue on the coat to the shooting of Dale. Further, a dummy cartridge taken from the appellant’s mother’s car could have been the source of the gunshot residue on the coat. In cross-examination, these propositions were put to Mr Collins who did not disagree with them: in the circumstances, they were not considered contentious and Dr Renshaw was not called to give oral evidence.

At the close of the prosecution case, both defendants made a submission of no case to answer. Counsel for the appellant submitted that both limbs of the well known test in R v Galbraith 73 Cr App Rep 124 were satisfied. Alternatively, the evidence was tenuous and inherently weak; none of the evidence individually or collectively amounted to a case which should be left to the jury. Dismissing the submissions, Penry-Davey J held that the state of the evidence was such that a jury properly directed could convict on the matters charged, and allowed the case to proceed.

The defence case was alibi: the appellant had gone to eat at Keating’s home. This was evidenced not only by his own account but also by Keating’s mother and her sister. Brown said that he went to McDonald’s restaurant before visiting friends; that evidence was corroborated by Jerome Barlow. In his closing speech, counsel for the appellant highlighted the weaknesses in the identification evidence, and asserted that the gunshot residue on the coat could be the product of secondary transfer.

In summing-up, Penry-Davey J provided the jury with directions of law which are not criticised. Although a circumstantial case which might have benefited from some description of the way in which such evidence should be approached along with its limitations, he addressed the four pillars of the prosecution case and did highlight the interdependency of the evidence.

In relation to the background, the judge both pointed to and summarised the relevant admissions and other evidence. Turning to Cunningham’s report of the telephone conversation, he said:

Penry-Davey J then dealt with the ‘other independent evidence’ and tied the various pillars of the prosecution case together. He said:

The judge then reviewed Cunningham’s evidence in detail and noted various inconsistencies and the extent to which his account changed when cross examined. He went on:

Passing on to Cunningham’s visual identification, Penry-Davey J gave an entirely appropriate Turnbull direction and drew attention to the particular weaknesses in Cunningham’s identification of the appellant. The appellant was not well known to him; he did not get out of the car; he only had a momentary glimpse; he could not be sure that it was the appellant; and he could not say anything about the length of his hair. He noted that there had not been an identification parade for the appellant who had consequently lost the prospect that, at such a parade Cunningham might have identified someone else, identified no-one at all or positively exonerated him.

As to the evidence of voice identification, Penry-Davey J said:

Thus, once again, the judge looked to the gunshot residue (described as such) as potentially corroborative of the voice identification.

Dealing with the gunshot residue, Penry-Davey J gave a measured account of the evidence underlining that the particles containing aluminium could have been produced from any cartridge case of which the expert had samples or from other ammunition that contains aluminium. He did not differentiate between the different particles (with different chemical constitutions) but went on to remind the jury that Mr Collins had said that the residue might have arisen:

The jury was then reminded that, when cross examined, Mr Collins had said that it was not possible to establish any specific link with the shooting of Daniel Dale; other potential sources for such residue could be blank firing guns and industrial nail guns; the dummy cartridge could be a potential source. The judge continued:

As mentioned above, the appellant’s renewed application for leave to appeal against conviction came before the full court on 25^th May 2004. His appeal was heard alongside that of Brown, who was given leave to appeal on the basis of the inconsistency of the jury’s verdicts.

The appellant’s appeal focused on the nature of the evidence at trial. Counsel submitted that Penry-Davey J erred in admitting Mr Loftus’ conviction; in admitting the evidence of Cunningham that Loftus had said that the appellant was coming to pick up the gun; in leaving the evidence of Cunningham to the jury (because of the circumstances and the absence of an identification parade); in failing to exclude the evidence of Shaw of voice recognition; and in refusing to accede to the submission that there was no case to answer. In dismissing the appeal, the court found that Penry-Davey J had correctly identified the legal principles and had decided each issue within the entirely legitimate bounds of his discretion. The evidence was such that the jury was able to consider it, subject to appropriate and robust direction. Such a direction had been given.

The appellant’s first application to the CCRC was dated 16^th February 2005 at which time no basis upon which to refer the conviction to this court could be discerned. The current application (received 23^rd July 2010) is primarily based on a scientific re-evaluation of the significance of gunshot residue generally as a result of which, on 19^th July 2006, the FSS issued guidelines on ‘the assessment, interpretation and reporting of firearms chemistry cases’ (“2006 Guidelines”). This document deals with the prevalence of small numbers of particles of gunshot residue with the result, so it is argued, that the number and type of particles of residue found on the coat were so small so as to be at or near the level at which they could not be considered to have evidential value. Parallels were drawn with the analysis in R v George (Barry) (supra). Having obtained fresh evidence in the form of an expert report prepared by Miss Angela Shaw, the CCRC decided to refer the case back to the court on this basis, reflecting that this court could conclude that the weight of the gunshot residue evidence was not such as supported the identification evidence.

This ground of appeal is formulated by the proposition that the gunshot residue evidence should not have been admitted before the jury. Alternatively, it is argued that Penry-Davey J failed to give what would now be an appropriate warning relating to the limited significance that could be attached to such evidence.

The CCRC declined to refer the convictions on the further ground that a re-analysis of the CCTV to demonstrate that the description of the witnesses was inconsistent with an assailant being the appellant. Nevertheless, the case having been referred, on his behalf, Mr James Wood Q.C. seeks leave further to argue that the conviction is unsafe because the voice identification evidence of Stuart Shaw should not have been given in evidence or, alternatively, following R v Flynn and St John (supra), which post-dated the appeal, the trial judge failed to give a warning appropriate to the voice identification evidence.

Mr Richard Whittam Q.C. for the Crown recognises that there has been a change of approach to evidence of gunshot residue but argues that this change does not necessarily determine the appeal. It is not suggested that gunshot residue evidence is of no value at all and, furthermore, it was clear that from the evidence at the trial that the presence of the gunshot residue could be explained in a variety of ways. As to voice identification, it is argued that the trial judge gave clear warnings as to its treatment. He emphasised the weakness of both that identification and the visual identification. It is submitted, therefore, that neither ground undermines the safety of these convictions.

Mr Whittam does not oppose the application to admit the evidence of Ms Shaw and there is no doubt that the 2006 Guidelines (updated in 2009 but with no material difference to this case) do provide new material as representing a change in scientific thinking on the significance of gunshot residue. Although we are both mindful of and share the cautious approach to fresh expert evidence (see R v Jones [1997] 1 Cr App R 86), in the circumstances of this case, we admit it and turn to consider its impact in this case and, in particular, the extent (if at all) to which it affects the safety of the conviction.

The 2006 Guidelines identify ammunition types, how particles are formed in the discharge of a firearm (including other possible sources of such particles such as fireworks, nail guns and brake linings) and the possibilities of secondary transfer. It classifies the number of gunshot residue primer particles into reporting levels of ‘low’ (1-3 particles), ‘moderate’ (4-12 particles); ‘high’ (13-50 particles) and ‘very high’ (greater than 50 particles) and contains the following advice in relation to reporting single particles and low levels of residue (at para. 9.5):

Against this forensic background, Ms Shaw analysed the findings (and re-examined the coat). She found that only the two particles on the coat containing lead, barium, and antimony could be said to be characteristic of gunshot residue. The two particles containing barium and aluminium were indicative of gunshot residue, but could also have originated from other sources, such as fireworks. An additional indicative particle containing barium and aluminium was found on the front of the coat. This also could not be said to be gunshot residue.

Applying the 2006 Guidelines, Ms Shaw considered that very little by way of interpretation could be applied to finding such low levels of gunshot residue, not least because of the lack of background data on residue in the external environment. She concluded that the particles could be related to a shooting; or to the dummy cartridge; or picked up unknowingly from the environment. In that regard, she noted that there had been shooting incidents on 30^th April 2001 (following which George was arrested but not charged) and at the Museum public house (in respect of which Ryan Brown and Kevin Faulkner were arrested but not charged). If George had been arrested by armed police officers, they could have had residue on their hands and clothing which could have transferred to his clothing. Suffice to say, it was not possible to conclude that the particles must be related to this particular shooting. We add only that it is unclear (but we doubt) whether there was any evidence of these other incidents before the court.

R v George (Barry) concerned the murder of Ms Jill Dando. A significant finding was the presence of a single particle of firearms discharge residue in the internal right pocket of a coat found hanging on the kitchen door of that appellant. It was a particle that contained the same constituent elements as discharge residue in a cartridge case found at the scene of the shooting and on the victim’s hair. It is clear that considerable significance was attached to this single particle. As was underlined in this court (see [2007] EWCA Crim 2722 per Lord Phillips of Worth Matravers CJ at para. 51):

That case was considered in R v Joseph [2010] EWCA Crim 2580 which was another murder which depended on a substantial body of circumstantial evidence: without seeking to be exhaustive, this included the recovery of the murder weapon and gunshot residue in a car which could be linked to Gavin Dean Abdullah, a man to whom that appellant could himself be linked through documents found at his home. Also in a bedroom at his home, there was found a single particle of what could be gunshot residue in a pocket of each of two motorcycle jackets and on a glove: they were of different types and concessions were made by the forensic scientist in that case as to secondary transfer such that the defence expert was not called. The case was left to the jury on the basis that the prosecution argued that the particle from the left glove was “capable of having come as discharge residue from the murder weapon” but also that the defence case was that “particles can easily be transferred from one surface to another. All it needs is a hand, a glove, a coat sleeve. Anything could transfer the particle”.

Although it was argued that the case was similar to George (Barry), this court rejected the notion that the jury had been misled. Pitchford LJ put it (at para 28):

Turning to the first ground of appeal, the appellant submits that this case is on all fours with R v George (Barry), where excessive weight was placed on a single particle of gunshot residue. The Crown submit that the better parallel is R v Joseph, where evidence of low levels of gunshot residue was admissible being specifically accorded appropriate (and not excessive) weight. As the appellant submits, there is only one more characteristic particle than in R v George (Barry). However, it is clear from R v Joseph that the task for this court is not merely quantitatively to apply the 2006 Guidelines, which have no force in law but are indicative only of the current state of the science.

In our judgment, there is no basis for challenging the decision of the trial judge to admit the evidence of gunshot residue and neither does the new evidence provided by Ms Shaw justify such a view. The fact that scientists have adopted a cautious approach to reporting low levels of residue (i.e. 1-3 particles) such that for that residue, on its own, no evidential significance can be attached to it does not mean that the evidence is necessarily inadmissible or irrelevant. Still less is that the case when (as here) there were in fact a total of four recovered particles, albeit that two are characteristic of gunshot residue and two indicative only (to say nothing of the additional particle found by Ms Shaw). The jury are more than able to assimilate evidence as to potential significance or lack of significance of recovered evidence, provided that there is an appropriate explanation of that potential significance, for example, by reference to what might occur in the environment or might otherwise be the consequence of entirely innocent contamination.

The importance of this point can be illustrated by reference to the forensic value of the absence of evidence. Whereas it is correct to say that absence of evidence is not the same as evidence of absence, the failure to recover anything that could even remotely be consistent with gunshot residue might provide a forensic argument supporting the proposition that involvement in the discharge of a firearm is disproved by the absence of particles that could be gunshot residue. The submission that the evidence now available demonstrates that the original forensic evidence should not have been placed before the jury is rejected.

Turning to the summing up, Penry-Davey J provided what appears to be a clear, balanced account of the evidence, underlining the various innocent explanations that Mr Collins had conceded during the course of his evidence for the presence of the particles on the coat found at the appellant’s home. As we have explained, he had specifically left open the possibility that the particles had no significance. On the other hand, however, he did not differentiate between the nature of or value to be attached to the different particles (whatever Mr Collins might have said) and referred to gunshot residue as potentially supportive of a conclusion of joint enterprise sufficient to justify reliance on Cunningham’s account that he was told that the appellant would collect the gun. Similarly, he referred to this evidence as potentially independent support for Shaw’s evidence to that effect that he recognised the appellant’s voice.

While we endorse Mr Whittam’s broad proposition that the change of approach to evidence of gunshot residue does not necessarily determine the appeal, had the present scientific concerns explained by Ms Shaw been available to the judge, we have no doubt that his directions would have been couched in terms of much greater circumspection and caution. The particles of gunshot residue may well be consistent with the appellant’s participation in the murder but, at the very least, the extent (if it got that far) to which they could provide positive corroboration would now have required much more detailed analysis of the science and the evidence.

The approach to the impact of fresh evidence (such as Ms Shaw provides) is identified in a well-trodden line of authority, ranging from Stafford v DPP [1974] AC 878 to the recent decision in Lundy v The Queen [2013] UKPC 28. Thus, in Dial and anorv State of Trinidad and Tobago[2005] UKPC 4; [2005] 1 WLR 1660, Lord Brown of Eaton-under-Heywood put the approach in this way (at para. 31):

The essential question remains whether, in the light of the fresh evidence, the convictions are unsafe or, as articulated in Lundy (supra) byLord Kerr (at para. 150):

In the context of this case, it is important to underline that Ryan Brown (whom Cunningham had also identified in circumstances more favourable to those which obtained in relation to the appellant) was acquitted and it is not fanciful to suggest that the evidence relating to what was called gunshot residue was seen by the jury as providing important independent support for the weak visual identification by Cunningham and weak voice recognition by Shaw (to say nothing of playing a part in justifying the conclusion of joint enterprise sufficient to rely on the evidence of Cunningham that the appellant had been named as one who was collecting the gun). The final plank of the prosecution case, namely the background events, provides critical context but is not probative of involvement in murder. In the circumstances, in the light of the new material, we are not prepared to conclude that these verdicts remain safe: the fresh evidence might reasonably have affected the decision of the trial jury.

Mr Wood also seeks leave to appeal on grounds relating to the admissibility of the evidence of voice recognition and the adequacy of the direction in relation to that evidence. As to admissibility, the issue was decided against the appellant in the first appeal and we see no reason to depart from that conclusion. Dealing with the approach of Penry-Davey J, although we recognise that the approach to such material is now identified in R v Flynn and St John (supra), we are satisfied that, in the circumstances of this case, the direction was sufficient: he indicated that increased caution was required with voice identification, and highlighted the specific weaknesses.

Having admitted the evidence of Ms Shaw, we have concluded that it might reasonably have affected the decision of the trial jury so that these convictions are no longer safe; in the circumstances, the appeal is allowed and the convictions quashed. In addition to expressing our gratitude to the Criminal Cases Review Commission, we pay tribute to the work of the Innocence Project and Pro Bono Unit at Cardiff Law School, which took up the appellant’s case and pursued it so diligently.