MELISSA BLOCK, HOST:
Advances in forensic technology are showing that what used to be considered clear-cut proof of guilt may not be. We're going to hear about a California case that illustrates this. The question: What do you do when an expert witness changes his mind because of better science?
Emily Green has the story.
EMILY GREEN, BYLINE: William Richards was convicted is serving 25 years to life for brutally murdering his wife. The evidence against Richards was mostly circumstantial; two different juries couldn't reach a verdict. A third trial was aborted because the judge recused himself.
At the fourth trial, the San Bernardino County prosecutor introduced, for the first time, testimony about a lesion on the victim's hand. Here is forensic dentist Norman Sperber describing an autopsy photograph he analyzed during that trial.
NORMAN SPERBER: It's kind of a reddish bruise, maybe a couple of inches wide. And there are little marks that look like the spaces between the teeth.
GREEN: Sperber told the jury that the apparent bite mark matched William Richards' unusual dental structure, one so unique he estimated just one or two out of 100 people might have it. Ten years later, another forensic dentist used new technology to correct a distortion in that picture.
SPERBER: If I had known that technology would help me be more accurate, I definitely wouldn't have testified as I did.
GREEN: Sperber now believes Richards could not have made the bite mark and questions if it's even human.
Similar cases are arising around the country. In Wisconsin and Texas, defendants have been exonerated or received a new trial. Last summer, the U.S. Department of Justice began reviewing thousands of convictions because of flawed forensic evidence.
In California, the state supreme court denied William Richards' request for a new trial, saying that Sperber's new analysis - a decade after the trial - didn't, quote, "unerringly point to Richards' innocence."
Richards' attorney, Jan Stiglitz, says the court set an impossibly high bar.
JAN STIGLITZ: We know that the linchpin in this trial was the bite-mark evidence. We now have experts who've come forward and all said this is not a mark that was made by Richards' teeth. And yet, Richards is going to spend the rest of his life in prison because he can't affirmatively prove that he didn't commit the crime.
JAN SCULLY: We need to have finality of verdicts.
GREEN: Jan Scully is a past president of the National District Attorney's Association and district attorney for Sacramento County.
SCULLY: There's always a new opinion or there might be a refinement in our forensic science areas. So, just because something new occurs doesn't mean that the original conviction somehow was not valid.
GREEN: The district attorney's office in San Bernardino, which prosecuted the original case, refused to comment for this story.
As more studies highlight major flaws with forensic evidence, challenges to convictions will continue to arise, says Georgia State University's Jessica Gabel. She says criminal appeals usually involve a so-called battle of the experts.
JESSICA GABEL: But when you've got the expert who testified for the prosecution who comes back and revisits the evidence and says, you know, the state of the science now tells me that my conclusions then were either incorrect or exaggerated or misleading - that is incredibly powerful in any given case.
GREEN: With state appeals exhausted, Richards is now appealing to the federal courts to review his case. But his lawyer says his best chance is getting clemency from the California governor.
For NPR News, I'm Emily Green. Transcript provided by NPR, Copyright NPR.