Editorial: Unsound science, verdicts?

Friday, Nov. 23, 2007 | 7:15 a.m.

The science behind an FBI forensic procedure that has produced damning evidence against defendants for 40 years was proved three years ago to be fatally flawed.

But the FBI waited a year to abandon the procedure. Moreover, it did not aggressively seek to review cases or notify courts until last week, and then only because it was under pressure from The Washington Post and CBS News' "60 Minutes."

Stories from a six-month investigation by the news organizations began appearing Sunday. Their disclosures were rooted in the work of William Tobin, an FBI metallurgist whose suspicions about a bullet-tracking procedure led him to spend years conducting experiments after he retired.

Tobin discovered that the procedure, known as bullet-lead analysis, produced unreliable results. After he began testifying as an expert defense witness, the bureau's then-lab director asked the National Academy of Sciences to examine the procedure.

Developed in the 1960s, the procedure evolved from the theory that a bullet's chemical makeup is nearly identical to that of other bullets made in the same batch. If a bullet fragment found at a crime scene could be matched to a box of bullets in a suspect's home, it could be the key piece of evidence determining guilt.

The National Academy of Sciences confirmed Tobin's findings in 2004, concluding, "Available data does not support any statement that a crime bullet came from a particular box of ammunition."

An estimated 2,500 cases over the past 40 years have included testimony about bullet-lead analysis. After the academy announced its conclusion, the FBI's response was limited to sending form letters - a year later - to police agencies and umbrella groups for local prosecutors and defense attorneys, the news organizations reported. And the letters expressed support for the "scientific foundation" of the procedure.

The FBI last week agreed to initiate reviews of all bullet-lead analysis testimony and alert courts and prosecutors to the true implications of the academy's findings.

We believe the probability is high that this flawed forensic procedure led to some wrongful convictions and that the FBI should have begun a review of past testimonies immediately after the academy's announcement.

In cases where this procedure amounted to key evidence, the statute of limitations for defendants to appeal their cases - usually two to four years - should be waived.