Bill Richards came home from work on an August night in 1993 to find his wife lifeless and half-naked on the ground outside their San Bernardino, California home. It took three 911 calls before the police showed up at Richards’s property, but homicide detectives didn’t arrive until almost three hours later. Pamela Richards’s body lay outside all night during a time of year in which temperatures could reach the triple digits and the area faced a growing coyote problem. Exposed to the elements and desert wildlife, her body wasn’t processed by the homicide team until dawn.
The police later admitted they believed Bill Richards was guilty before they even arrived. While homicide detectives waited till morning to gather evidence and the body, animals ran freely around the crime scene and left Pamela’s body mangled. Despite his alibi, statements from neighbors and coworkers, and his insistence that he didn’t kill his wife, Richards’s guilt was signed and sealed, in large part because an “expert” testified on a single piece of forensic evidence that has now been debunked as junk science: the bite marks found on his wife’s body. Richards was quickly and confidently convicted of his wife’s murder.
Forensic science is a slow, unglamorous, and constantly changing field that is often anything but definitive. In fact, it’s such an amorphous discipline that new ways of testing forensic evidence — or debunking it — can flip old convictions entirely, in either direction. Junk forensic science, along with delays in evidence gathering and mistakes in processing, put Richards behind bars, but these problems are not unique or uncommon. From massive rape-kit backlogs to under-funded coroners’ offices, a national lack of resources and standards mean that such dubious science can easily fill the cracks.
This month, the Department of Justice announced $145 million in awards for research programs and grants to advance the field of forensic science and help prevent the mistakes and shortcomings that put people like Richards in prison. The money is set to give local jurisdictions resources to improve forensic analysis, but will fund research and development for “more accurate, reliable, and cost-effective and rapid methods of analyzing physical evidence.” While the funding is certainly a step toward preventing lapses in the quality of forensic investigation nationwide, it’s unclear whether it will patch the hole left by the Trump administration’s 2017 termination of Obama’s National Commission on Forensic Science, an initiative set up in 2013 to address this very problem.
During Richards’s 1997 murder trial, Norman “Skip” Skerber, a then-prominent forensic odontologist, testified under oath that he believed a bite mark on Pamela Richards’s hand was almost certainly made by her husband’s “highly unique lower dentition.” Skerber claimed that Richards’s bite pattern was so unique, similar ones only occurred in one to two percent of people.
Bite marks weren’t considered admissible evidence in California until 1975, when evidence of a bite mark in the cartilage of a murder victim’s nose was admitted into court. Dentists like Sperber started giving expert testimony on bite mark evidence in court almost immediately, and his work as a forensic dentist across the country contributed to the contested practice becoming widespread for decades. While forensic testing of saliva or blood left by biting would yield a more definitive DNA test, the absence of DNA evidence led many forensic dentists to look at bruise patterns left on victims and compare them with molds taken of suspects’ teeth. On top of the facts that flesh is malleable, people bruise differently, and bruises change shape and color rapidly (even in death), Michael Semanchik, the managing attorney for the California Innocence Project, told me that most bite-mark experts would only examine photos of the body — not the real thing.
Richards never stopped maintaining his innocence. In 1999, he became one of the first clients of the California Innocence Project. Two years later, the organization motioned the court to DNA test the murder weapon, Pamela Richards’s fingernails, and other pieces of evidence that had been ignored in favor of the bogus bite-mark claim made by Skerber. The DNA that the organization found revealed that, almost two decades later, Pamela was killed by a totally different man who was never caught.
Semanchik began working at the California Innocence Project while a law student in 2008. That year, Skerber recanted his testimony. He admitted that the statistic he had given about how “unique” Richards’s bite was wasn’t actually scientific. Richards’s legal team had additional dentists analyze the bite mark that had put him behind bars; ultimately the new dentists and Sperber concluded that they weren’t even sure the bite was human.
“It could have actually just been bruising left from the dogs or the coyotes that were there that night,” Semanchik said.
In addition to being driven by a science now accepted as all but completely unreliable, the Richards case was filled with errors and contamination that are sadly common. Contamination, both at the scene of the crime and during analysis, is one of the biggest factors that affects the veracity of forensic investigations, according to John Butler, a fellow and assistant to the director for forensic science at the National Institute of Standards and Technology. Butler told me that the lack of control that forensic experts face when obtaining uncompromised samples from crime scenes is a major factor in why forensic science isn’t as black and white as Hollywood might make it seem.
“That's just the reality of it,” Butler said. “There may not be evidence that can be found by the investigators or it may have been washed away in the course of something occurring outside. Environmental conditions could lead to a degradation of the sample that’s available to you, and you only have a partial set of information you can compare to that. That's where the biggest challenges lie.”
Butler says that contamination happens all the time, by both human error, and simply because it’s a fact of life. Because there aren’t uniform standards enforced in forensics labs across the country, samples all too often get switched or tampered with. In addition, people who live together are likely to have each others’ DNA all over each others bodies and possessions. Butler uses the example of a mother driving a car; even though a mother’s infant is not likely to touch the steering wheel of her car, the steering wheel will likely have both their DNA on it simply because the mother has touched both the baby and the steering wheel. The summation of these challenges, along with mistakes in interpretation and constantly burgeoning research are why even forensic science can’t solve most cases.
“The reality is that many cases never come to closure,” Butler said. “You may have 50 percent to 60 percent of cases that never get solved.”
In addition to the faulty bite-mark evidence, a single blue fiber matching Richards’s shirt was alleged to have been found under his wife’s fingernail. However, in photographs of Pamela Richards’s fingers taken before she was moved by police, no blue thread is visible.
“It's unclear whether or not the fiber was planted, or how it ended up under her fingernail, but it's not there until later down the road,” Semanchik said.
The new evidence and Skerber’s recanting of his testimony weren’t enough to get Richards out of prison. Expert testimonies like Skerber’s are treated like evidence; they’re used by lawyers to sort of connect the dots and interpret what an innocuous or minor detail found at a crime scene might mean. But Butler said the way analysts interpret and present forensic evidence to a jury is just another potential kink in the chain. “You may be able to measure data very well, but then extrapolating and making a stronger statement like, “This is definitely from that person,” he said, “and offering an opinion like that without having the data behind it to support it…that's one of the big challenges.”
And another huge stumbling block? Experts often get paid for their testimonies. As forensic odontology grew in popularity, dentists saw an opportunity. An odontology group formed in the Academy of Forensic Science, and lawyers started calling on dentists like Sperber to testify in their cases. A 1988 profile of Sperber in the San Diego Reader called him “the tooth detective,” venerating work he did as a forensic dentist, including bite mark analysis.
“Dentists started testifying and saying that they could, to the exclusion of all others, make these determinations,” Semanchik said. “They were getting paid big bucks for it. But I don’t think it was just a money thing — I think they thought they were doing the good of society.”
Semanchik and his team had to push for a bill to change the California state penal code to classify so-called expert testimonies that are recanted or disproved as “false evidence” in 2015. “‘False evidence’ includes opinions of experts that have either been repudiated by the expert who originally provided the opinion at a hearing or trial or that have been undermined by later scientific research or technological advances,” the criminal code now reads.
The “junk-science” legislation created a domino effect. Since 2015, 30 bite-mark related convictions have been overturned. “I think we can safely say that bite mark evidence is just junk,” Semanchik said.
New law enforcement tools have also helped cull some of the red-letter flaws of forensic investigation. In 1990, the FBI introduced the Combined DNA Index System, or CODIS. A digital catalogue of DNA samples from around the country, the database was meant to help local jurisdictions identify serial criminals who left DNA trails across state lines. However, a lack of training among law enforcement means that the CODIS system was and continues to be underused. But Butler thinks tools like CODIS and an overall improvement in the standards and scientific backing of forensic investigation will help prevent cases of people getting wrongfully convicted or let off. Still, even forensic efforts on the federal level are burdened with low reliability. In 2015, the FBI admitted that examiners in one of their forensic units had been giving flawed testimony in hundreds of trials in the ’80s and ’90s.
Exoneration organizations like the California Innocence Project, however, still have their work cut out for them. Semanchik said that major shortcomings exist in several forensic subfields that need to be addressed, particularly in infant death and arson cases, and at all levels of criminal justice.
“This goes to the top level of law enforcement and forensics in the United States, where we’ve just relied on junk-science to convict people,” he said. “And we don't have a better way of confirming the forensic science that we're trying to rely on when somebody's life is in the balance.”
Richards was released from prison on June 21, 2016. Semanchik described seeing Richards walk out of prison as one of the highs of his career. Though he joins the ranks of more than 2,300 exonerees in the U.S., Richards reportedly developed terminal cancer while incarcerated. He filed a lawsuit against the county in 2017, which is ongoing.