Probyn v R

Fresh evidence does not render a conviction unsafe merely because, on contestable assumptions, it makes the prosecution scenario unlikely. The appellate court must assess the evidence alongside the whole trial case and ask whether it might reasonably have affected the jury’s decision, rather than decide guilt for itself.

Under section 23 of the Criminal Appeal Act 1968, evidence which cannot afford a ground for allowing the appeal need not be admitted. Where the unchallenged circumstantial evidence is overwhelming, fresh expert reconstruction evidence that does not make the prosecution account impossible will not disturb the safety of the conviction.

The appellant was convicted by a unanimous jury at the Crown Court at Bristol of murdering his wife, whose car was recovered from the River Severn with her body inside.

Following an earlier refusal of leave to appeal, the Criminal Cases Review Commission referred the conviction. The sole ground pursued concerned fresh agreed expert evidence about the vehicle’s movement across the riverbank before it entered the water. The appellant contended that the reconstruction supported an accidental entry and was inconsistent with his having propelled the car into the river.

The central issue was whether the fresh evidence could affect the safety of the conviction.

1. Appeal dismissed. The conviction was safe. The court heard the fresh expert evidence de bene esse but declined to admit it under section 23 of the Criminal Appeal Act 1968, because it could not afford a ground for allowing the appeal.

2. The correct question was whether the conviction was safe, not whether the court considered the appellant guilty. Applying the approach in R v Pendleton [2002] 1WLR 72, the court considered whether the evidence, if available at trial, might reasonably have affected the jury’s decision. Fresh evidence had to be assessed with the entire body of trial evidence, rather than in isolation.

3. The court accepted that the experts agreed that, if particular assumptions were correct, the car would have needed to travel at least 15 mph when it crossed the ridge and damaged its bumper. It would then have been unlikely to stop before the riverbank edge. Those assumptions were not secure. They depended on a survey conducted more than five months after the event, in an area prone to erosion and vehicle use, and on the debris having remained undisturbed for four days.

4. In any event, the reconstruction did not establish that the appellant could not have propelled the car into the river. It showed only an isolated improbability. The substantial circumstantial case remained untouched: the timing, open windows, extinguished lights, vehicle controls, injuries, seatbelt evidence, the appellant’s conduct and alleged lies, and the Adidas bag. Taken together, that evidence compellingly established that he had caused the car to enter the river.

5. R v Bowler was distinguishable on its facts. The fresh evidence did not suggest a real possibility that the deceased died otherwise than through the appellant’s actions. A later inconsistent account by the appellant was disregarded because it concerned credibility only and not the commission of the offence.

• Court of Appeal (Criminal Division). On a reference by the Criminal Cases Review Commission under the Criminal Appeal Act 1995, dismissed the appeal and held the conviction safe: [2005] EWCA Crim 2347.
• Full Court. Previously refused leave to appeal against the conviction, holding that there was no basis for concluding that the verdict was unsafe or unsatisfactory.
• Crown Court at Bristol (Auld J and jury). On 30 March 1994, convicted the appellant of murder by a unanimous verdict.