Before DNA testing became the gold standard in forensic science, hair analysis was often a prosecutor's trump card.
Developed by the FBI's vaunted crime lab, microscopic hair analysis - comparing a hair found at a crime scene with one from a criminal defendant - as described in polished, confident testimony by an FBI hair analyst, could seal a guilty verdict.
Now, an ongoing FBI hair-analysis review - preliminary results were announced April 20 - recommends a wholesale look at cases in which testimony about microscopic hair analysis contributed to a guilty verdict.
The FBI review of about 500 transcripts of trials before 2000, when DNA testing supplanted microscopic hair analysis, shows FBI hair analysts made erroneous statements or exaggerated evidence in 96 percent of cases.
The review has caused the FBI and the Justice Department to question the convictions of 284 people, including 21 in Pennsylvania, and take the unprecedented step of waiving court deadlines and procedural hurdles that could prevent those people from filing new appeals.
Legal experts say the analysis raises the possibility of a wave of hundreds or thousands more convictions being appealed nationwide as more transcripts are examined.
Fourteen people whose cases were cited in the April 20 report are now beyond the justice system's reach. They were among 35 sentenced to death, including five now on death row in Pennsylvania, based on hair analysis evidence. Nine of the 35 were executed; five died of other causes on death row.
"It's hard to comprehend the breadth, impact, and magnitude of this for those accused and the criminal justice system," said Center City lawyer Theodore Simon, president of the National Association of Criminal Defense Lawyers.
Simon's group partnered with the Justice Department, FBI, and Innocence Project of New York in the review.
"It's frightening to recognize that for every innocent person convicted and incarcerated, there's an actual perpetrator who is out there free," Simon added.
So far, the review has questioned the validity of the convictions of 284 people, including 21 in Pennsylvania, the second-largest group in the United States. Three cases were cited in Delaware, six in New Jersey.
"It's a big number," said Marissa B. Bluestine, legal director of the Pennsylvania Innocence Project, based at Temple University's Beasley School of Law, and it's a big challenge for the lawyers representing inmates who may have been wrongly convicted.
Individuals have 60 days to file a petition for review after the FBI notifies them that their convictions are questionable. Bluestine said she has heard from three or four inmates and lawyers for three or four more.
Bluestine said she is barred from revealing names under terms agreed to by the four organizations. She said she and advocates do not know how many people in the questionable cases are in prison.
And the problem is likely far larger than the 284 cases identified in the April 20 FBI report. Not included are thousands of cases nationwide in which hair analysis testimony was given by hundreds of state or local law enforcement personnel trained by the FBI in two-week long sessions over a quarter-century.
"The problem is that the FBI didn't just teach the analysis technique, they taught them how to testify," Simon said.
Microscopic hair analysis could never positively link two human hairs. At best, it could exclude a person on trial from being the source of the hair sample.
If the sample hair was not excluded, FBI analysts compared hair structure features and opined on whether the hairs could have come from the suspect.
It was when DNA came into use in criminal investigations that the subjectivity of microscopic hair analysis became clear.
By 2009, microscopic hair analysis - along with such forensic staples as analyses of bite marks and fibers - was labeled "highly unreliable" by the National Academy of Science in a study titled "Strengthening Forensic Science in the United States: A Path Forward."
In 2011, DNA testing exonerated a Washington man named Kirk Odom. Odom was 18 when he was convicted of rape and sodomy in 1981 and sent to prison. At trial, an FBI hair analyst testified that Odom's hair and hair samples from the crime scene "were indistinguishable" and described that as very rare.
The FBI took note of Odom's case, which was followed the same year by DNA exonerations of two other men. Each also had done more than 20 years in prison, and was convicted based on testimony from FBI hair analysts.
Two years later, the FBI and Justice Department, the Innocence Project, and the National Association of Criminal Defense Lawyers announced their "groundbreaking and historic agreement" to review about 3,000 trials from 1985 to 2000 in which an FBI hair analyst had testified and helped get a conviction.
The agencies got through 500 cases when they decided to go public and urge a wholesale review of convictions in which hair analysis was key evidence. They will continue to review the remaining 2,500 cases.
Beyond the size of the number of convictions called into question in the April 20 report was the systemic nature of the problem. All but two of 28 FBI hair analysts made erroneous statements on the witness stand or in lab reports - evidence used at trials in 41 states.
Amy Hess, executive assistant director of the FBI's Science and Technology Branch, said in the April 20 report that "such statements are no longer being made by the FBI," and that the Justice Department and FBI are "committed to ensuring that affected defendants are notified of past errors and that justice is done in every instance."
Peter Neufeld, codirector of the Innocence Project at Cardozo School of Law in New York, said the report documents an "epic miscarriage of justice," and urged creating national research and forensic standards "to prevent the exaggeration of results in reports and in testimony by crime lab analysts."
There is no framework for handling erroneous convictions based on hair analysis by FBI-trained state and local analysts.
"It will be up to the individual states to determine if they will follow [the Department of Justice's] lead in permitting these cases to be litigated," according to a joint statement that accompanied the FBI review.
Jeffrey A. Johnson, a spokesman for Pennsylvania Attorney General Kathleen G. Kane, said the office has received letters about hair analysis issues, but the cases were handled by county prosecutors. Kane's office has forwarded the letters to local district attorneys.
Richard W. Long, executive director of the Pennsylvania District Attorneys Association, said the group has not taken a stand on how to address the findings and implications of the FBI report.
Cameron Kline, spokesman for Philadelphia District Attorney Seth Williams, said homicide prosecutors do not believe that pre-DNA microscopic hair analysis was used in prosecuting city cases.
For most Pennsylvania police departments, state police labs provide the testing and forensic expertise needed to prosecute criminal cases.
Maj. William Teper Jr., director of the State Police Bureau of Forensic Services, said the lab relies on DNA testing and barred use of comparative hair analysis in 2009.
Even before 2009, Teper said, state police analysts were directed to take "a more conservative approach testifying about the significance of any association between a hair sample and the criminal defendant."
Teper said state police hair analysts did not testify that the results of a hair comparison implicated a defendant on trial.
"I think the problem was not so much the technique used, but the particular manner in which the FBI analysts gave testimony," Teper said.