Patrick Radden Keefe's account of the London Met's police unit of "super-recognizers" is fascinating, but the use of their identifications in investigations or in prosecutions risks misidentification and wrongful conviction ("Total Recall," August 22nd). Since 1989, eyewitness misidentification has contributed to a seventy-one per cent rate of wrongful convictions, proved by DNA testing. Despite the unit's system of checks and balances, identifications by Met super-recognizers lack the basic elements of scientific inquiry: validation, peer review, and error rates. Even if super-recognizers are better at face recognition than most people, the hypothesis that they excel at matching a previously seen face to a target has not been tested in a clinical environment. Nor have there been any valid studies that independently establish error rates.