Dawes, R. v

In criminal proceedings, a trial judge should ordinarily determine identification-evidence admissibility from the disclosed material and submissions. A voir dire or trial within a trial is exceptional, particularly where alleged procedural failings can be explored before the jury and do not require preliminary fact-finding. Evidence elicited by the defence remains part of the case and cannot be ringfenced from legitimate prosecution use. Scientific evidence is not inadmissible merely because a DNA sample is small, mixed, has imperfect reproducibility or involves uncertainty about contributors. The question is whether a reliable profile can be produced. The jury determines weight, while the judge and advocates must explain likelihood ratios clearly.

The applicant had previously been convicted of murder, successfully appealed, and was convicted again after a retrial before HHJ Hillen at the Central Criminal Court in March 2020. He sought permission to appeal, challenging the admission of identification evidence arising from flawed police procedures and the admission of low-level mixed DNA evidence from latex gloves. The central issues were whether the trial judge should have conducted a voir dire, whether evidence elicited through defence cross-examination could be used by the prosecution, and whether the identification and DNA evidence were too unreliable to be left to the jury.

The application for permission to appeal was dismissed. The conviction was demonstrably safe.

1. The Court upheld the refusal to conduct a voir dire on the identification evidence. The judge had correctly applied the principles in AG Ref (No.2 of 2002) [2003] 1 Cr App R 21 (321), distinguishing between image comparison by an officer who had acquired special knowledge through extensive viewing and recognition by an officer sufficiently familiar with the defendant. The Court also approved the judge’s reliance on R v Beveridge (1987) 85 Cr. App. R. 255, R v Flemming (1988) 86 Cr. App. R. 32 and R v Martin and Nicholls [1994] Crim.L.R. 218. A trial within a trial on identification admissibility is rare. The judge may ordinarily decide the issue from the depositions, statements, disclosed material and submissions, unless admissibility depends on facts requiring determination.

2. The disclosed failures in the identification procedure, including the matters arising under Code D, did not provide a basis for excluding the evidence. Allegations of bad faith or collusion could properly be tested before the jury.

3. Evidence elicited in cross-examination, or obtained from witnesses tendered at the defence’s request, remains evidence in the case. It cannot be ringfenced from legitimate prosecution use, including re-examination and submissions. The resulting tactical risk did not make the trial unfair. The Court applied R v Barry George [2002] EWCA Crim 1923.

4. There was no fixed quantity threshold below which DNA evidence became inadmissible. Following R v C [2010] EWCA Crim 2578, the question was whether a reliable profile could be produced despite the low quantity. A mixed profile, imperfect reproducibility and uncertainty about the number of contributors did not themselves establish inherent unreliability. The reasoned expert opinion on contributors was admissible, and its weight was for the jury.

5. The complexity of a likelihood ratio concerned presentation rather than admissibility. Counsel and the judge had to ensure that the jury understood its meaning. HHJ Hillen had given clear directions on both the identification and DNA evidence, and the Court found no error in his rulings or summing-up.

• Court of Appeal (Criminal Division): application for permission to appeal dismissed; the conviction was found safe, [2021] EWCA Crim 760.
• Central Criminal Court: conviction entered after a retrial before HHJ Hillen in March 2020.
• Earlier appeal: the applicant had successfully appealed his original conviction, leading to the retrial.