This is part two in a four-part series. Read part one here.

On May 4, 1692, the Rev. George Burroughs was arrested in Salem, Massachusetts on suspicion of witchcraft. The only physical evidence against Burroughs were bite marks found on some of the girls he was accused of recruiting to join him. Summarizing the research of historians on the ordeal in an article for the February 2014 newsletter of the New York State Dental Association, William James Maloney writes that at trial, “the defendant’s mouth was pried open and the prosecution compared his teeth with the teeth marks left on the bodies of several injured girls present in the courtroom.”

At the urging of notorious witch hunter Cotton Mather, Burroughs was convicted, sentenced to death and hanged. Two months later, the governor of Massachusetts called for an end to the witchcraft trials. He also prohibited the use of “spectral and intangible evidence” in criminal trials. Two decades later, Burroughs was declared innocent, and the colony of Massachusetts compensated his children for their father’s wrongful execution.

Nearly three hundred years later, in 1974, Walter Edgar Marx was convicted of voluntary manslaughter due in part to bite marks found on the nose of his alleged victim. The marks were found during an exhumation of the victim’s body more than six weeks after she had been autopsied, embalmed and buried. Three dentists testified for the state that they could match an impression made of the marks to Marx’s teeth. In 1975, a California appeals court upheld the conviction. That ruling has become enormously influential. In a 2000 article for the Albany Law Review, Seton Hall law professor and evidence expert Michael Risinger wrote that the Marx ruling “came to be read as a global warrant” for courts to admit bite mark evidence.

The Marx case effectively went around the prevailing standard for admitting forensic evidence: the 1923 case Frye v. United States, in which the U.S. Court of Appeals for the D.C. Circuit rejected testimony from a polygraph instructor who claimed that a rise in systolic blood pressure indicated that a suspect was lying. The appeals court ruled that in order to be admissible in federal court, scientific evidence or testimony must have “gained general acceptance in the particular field in which it belongs.” For the next 70 years, Frye was the model in federal court, and was subsequently adopted by nearly every state in the country. (The Supreme Court didn’t address the standard until 1993, with three rulings now known as the Daubert cases. The Daubert standard instructs judges to assess both the relevance of expert testimony and whether the testimony itself is reliable.)

In Marx, the judges actually accepted that there was no scientific research to support bite mark matching. There is “no established science of identifying persons from bite marks” and “no evidence of systemic, orderly experimentation in the area,” the court wrote. But the judges’ reasoning then took a peculiar turn. Because there was no science to analyze, the court declined to hold a Frye hearing. Instead, the judges simply invented their own test for evidence that wasn’t scientific, but was nevertheless presented with a science-like veneer. They found that because the trial judge saw the bite mark evidence and concurred that it seemed sound, that was good enough for them. (Marx was convicted at a bench trial, not a jury trial.) The appeals court judges wrote that the evidence was admissible because to not admit it would be to “abandon common sense.”

Three years later, another California appeals court relied on Marx to uphold bite mark evidence again . Bizarrely, that court explicitly referenced the “superior trustworthiness of the scientific bitemark approach,” despite the fact that the Marx opinion specifically acknowledged a lack of scientific research in support of the practice. From there, bite mark evidence began to get accepted simply by virtue of the fact that it had been accepted in other courts in previous cases. Thus began an established record of precedents. It also began an established record of wrongful convictions.

Bite mark matching then gained national notoriety in 1979 during the trial of serial killer Ted Bundy. The high profile nature of the case, the brutality of the crime and Bundy’s obvious guilt cast a public image of bite mark analysis as an emerging science that could put away serial killers and sex offenders, and the analysts themselves as heroic scientists who help put dangerous people behind bars. One of the analysts who testified in Bundy’s case was Lowell Levine. Bundy’s conviction launched Levine’s career. He became one of the most visible bite mark evangelists in the United States. In 1977, he wrote that a bite mark match “is as good as a fingerprint.” It was 20 years later that Levine’s analysis would lead to the wrongful arrest of Edmund Burke discussed in part one of this series.

Levine later served terms as president of both the American Board of Forensic Odontology (ABFO) and the American Academy of Forensic Sciences (AAFS), and became one of just a few dentists to make a full-time career of bite mark analysis. In a 2011 interview with CNN’s Anderson Cooper, Levine continued to defend bite mark analysis as “important and viable.” But when Cooper asked if there’s any way bite mark analysis can be reconciled with the scientific method, Levine replied with some candor: “I sure can’t think of it.” Yet Levine has testified countless times in court about his “level of scientific certainty” with respect to bite marks.

In a forthcoming law review article, Chris Fabricant of the Innocence Project and Tucker Carrington of the Mississippi Innocence Project look at how the Marx ruling affected bite mark admissibility. They found 16 court opinions from 12 states over the following 13 years after Marx that either relied on the decision, or adopted what Fabricant and Carrington call “the eyeball test.” All but three of those rulings noted the “scientific” nature of bite mark analysis, despite the fact that, again, not only is there no science to back up the claims of bite mark analysts, but the Marx decision explicitly acknowledges as much. One such ruling came in 1978 after the Arizona Supreme Court heard arguments to overturn a conviction based on bite mark testimony from Homer Richardson Campbell Jr., a ABFO-certified forensic odontologist. Campbell told the jury that that the odds of anyone other than the defendant leaving the marks he found on the victim’s breast were “eight on one million.” On cross examination, Campbell conceded that he didn’t compute those odds personally. Rather, they were a rough estimate of his memory of “articles written in the journals of the American Academy of Forensic Sciences.” In truth, there was no scientific basis for his estimation whatsoever. The court nevertheless found his testimony admissible, and upheld the conviction.

In 1987, Campbell’s testimony helped convict Joe Sidney Williams of a rape and murder in Texas. Williams was exonerated by DNA testing in 2001. Campbell’s bite mark matching also helped convict David Wayne Spence for the 1984 murders of three teenagers near Lake Waco, Tex. In fact, his testimony was the only evidence linking Spence to the scene of the crime. During Spence’s appeal, his attorneys showed the crime scene photos Campbell used to make his match to five other ABFO-certified forensic odontologists, along with dental molds from five people, including Spence. Only two matched the photographed bite marks to any of the dental molds, and both matched them to the mold of an uninvolved patient of the dentist who ran the test. The lead homicide investigator in the case told New York Times columnist Bob Herbert in 1997, “’My opinion is that David Spence was innocent. Nothing from the investigation ever led us to any evidence that he was involved.” Spence was executed by the state of Texas in 1997.

Campbell, now deceased, would go on to become a prominent advocate for bite mark matching. He also served as president of the American Academy of Forensic Sciences.

One particularly striking example of how bite mark matching was back-ended into the criminal justice system came in a 1986 ruling by the Court of Appeals of Wisconsin. The case was the first time bite mark evidence had been introduced in the state. The court’s ruling was one of “first impression,” meaning that there was no precedent, and that its ruling in the case would likely establish one. In 1983, Robert Lee Stinson had been convicted of raping and murdering an elderly woman. The only physical evidence linking him to the crime was the testimony of two bite mark specialists: Lowell Thomas Johnson and Raymond Rawson.

Rawson at the time was on the ABFO’s Bite Mark Standards Committee. He had also co-authored the organization’s original guidelines for bite mark matching. Johnson performed the initial analysis in the Stinson case by placing a mold of the suspect’s teeth over photographs taken of some marks on the victim’s body. Johnson concluded that the marks “had to have been made by teeth identical in all of these characteristics” to the teeth of Robert Lee Stinson. Rawson went next, and confirmed that the marks matched Stinson’s teeth “to a reasonable degree of scientific certainty.”

Stinson appealed the admission of bite mark evidence in his trial. The Wisconsin Court of Appeals ruled against him. The court’s opinion was forceful and authoritative.

A total of fourteen upper and lower jaw impressions were made from the bite marks found on Cychosz’s body. Because of the opportunity to examine so many bites, and the fact that some of the bites were so deep as to be three-dimensional, Dr. Johnson testified he was able to detect a repetition of some particularly unique features in several of the bites.

Dr. Johnson later performed a forensic odontological examination of Stinson. Following the examination, Dr. Johnson noted the following unique features: one of the central incisors was fractured and decayed almost to the gum line; the lateral incisor in the upper jaw was set back from the other teeth; all of the upper front teeth were flared; the lower right lateral incisor was worn to a pointed edge; the right incisor was set out from the other teeth on the lower jaw. Dr. Johnson used these features along with the arch of the mouth and the spacing, width, and alignment of the teeth to make comparisons with the bite marks found on the victim. After an exhaustive examination of the photos, models and tissue samples taken from Stinson and the victim, Dr. Johnson concluded, to a reasonable degree of scientific certainty, that the bite marks on the victim were made by Stinson.

The jury also heard from Dr. Rawson who concluded, based on the workup Dr. Johnson performed on both the victim and Stinson, that Stinson had inflicted the bite marks on the victim. In Dr. Rawson’s opinion the evidence in the case was overwhelming and he stated that “if we have four or five teeth that we are able to examine, then we can say that there is no other set of dentition like that.” In this case, Dr. Johnson was able to identify seventy-five individual tooth marks in various combinations of between five and eleven teeth.

In the end, the Wisconsin Court of Appeals concluded that the reliability of the bite mark evidence in the case was sufficient “to exclude to a moral certainty every reasonable hypothesis of innocence.”

And yet the court was wrong. Stinson spent 23 years in prison before DNA testing exonerated him. Stinson never bit the victim. All of the argumentation about set-back incisors, flared teeth and the arch of the mouth, all of that evidence that screamed guilt — to a moral certainty no less — it was all nonsense. Yet the court never made any effort to correct its mistake. As Fabricant and Carrington point out in their article, State v. Stinson is still the controlling precedent for bite mark evidence in Wisconsin. That the man whose name appears in the case was actually innocent doesn’t seem to matter.

Six years later, after the ruling in the Stinson case, Raymond Rawson helped convict another innocent man. In 1992, Ray Krone was convicted of murdering Phoenix waitress Kim Ancona. The only physical evidence linking Krone to the crime scene were bite marks left on the victim that two forensic odontologists, one of them Rawson, said could only have been inflicted by Krone. The highly-publicized trial and the nature of the bite mark testimony earned Krone the nickname, “the Snaggletooth Killer.”

In 1995, Krone was awarded a new trial due to a legal technicality. The following year he was convicted a second time, again due to testimony from Rawson, who declared a pattern on the victim’s bra to be a “scientific match” to Krone’s teeth. Krone spent 10 years in prison, including some time on death row, before he was exonerated by DNA evidence in 2002.

By 1988, the West Virginia Supreme Court noted in an opinion that bite mark matching had been so “generally accepted” in American courtrooms that a Frye analysis was no longer necessary. (In that particular case, a bite mark analyst had determine that a tooth mark in a roll of paper towels was a “perfect match” to the defendant “to the exclusion of all other individuals.”) By the time that ruling came down, 21 state appellate courts had accepted bite mark analysis, without a single dissenting opinion. There still wasn’t a shred of scientific evidence to back any of it it up. Instead, all of these courts had relied on the rulings of prior courts, going all the way back to Marx. In some instances, these state courts adopted what Fabricant and Carrington call the Marx court’s “eyeball test.” In others, the state courts falsely claimed that Marx had already validated the science of bite mark matching, or they cited opinions that had falsely stated the same, thus relieving them of the need to do an analysis themselves — a phenomenon Fabricant and Carrington call the “echo chamber effect.”

“Most of the time when doing one of these analyses, the only thing a judge will ask is, ‘Have other courts allowed this?’” says Michael Saks, a law professor at Arizona State University who has written extensively on the intersection of law and science. “If the answer is yes, then they’ll figure out a way to let it in. Or they’ll decide that if the government is paying a person to do this analysis, it must be legitimate. That’s a far cry from an analysis of its scientific merit. But it doesn’t seem to matter.”

More state courts followed West Virginia’s lead, accepting bite mark matching based solely on the fact that other courts had already accepted it.

By the early 1990s, the bite mark matching business was booming. One of the most prolific expert witnesses of that era was a Hattiesburg, Miss., dentist named Michael West. Using yellow goggles and ultraviolet light, West claimed to have pioneered a new method of bite mark analysis that allowed him to find and analyze bites that no one else could see, not even other trained forensic odontologists. Conveniently, West said that his process couldn’t be duplicated or recorded. He called it “the West Phenomenon.”

West struck the right balance of brash arrogance and aw-shucks charm to win the trust of Mississippi jurors. That made him hugely popular with prosecutors. His services were in ever-increasing demand, particularly in Mississippi and Louisiana, even as his claims grew more preposterous.

In one case, West claimed to have matched the bite marks in a half-eaten bologna sandwich to the defendant. The jury convicted. (The conviction was overturned on appeal when defense lawyers discovered that the autopsy report recorded a partial bologna sandwich in the stomach of the victim.) In 1991, West claimed to have found bite marks in an exhumed body that had been buried months earlier, even though police and investigators never noticed the marks when the body was fresh. He then said he could match the bite marks to the woman’s husband, Anthony Keko. West and (the also controversial) Mississippi medical examiner Steven Hayne didn’t photograph or preserve the incriminating marks. They claimed to have tried to preserve one of them, but say they inadvertently destroyed the sample by storing it in the wrong type of solution. So West was testifying from memory. The jury still convicted.

Despite these incidents, by 1994 West had testified in several states, was testifying frequently in Mississippi and had been elected coroner of Forest County, Miss. In 1996, West told the American Bar Association Journal that he had testified in 55 cases. He had also recently been certified by an Ohio court as an expert in “splash patterns.” That ABA Journal profile was critical, and noted the mounting skepticism in the legal community about West’s claimed expertise.

Yet judges continued to certify West, and prosecutors continued to utilize him. One Louisiana prosecutor told the publication, “I’m quite confident in the guy . . . I think he makes one heck of a witness.”

In September 1990, 3-year-old Courtney Smith was abducted from her bedroom in Noxubee County, Miss. She was then raped and murdered. Michael West was called in to perform a bite mark analysis, and he claimed to match marks he found on the child’s body to Levon Brooks, the boyfriend of the girl’s mother. In January 1992, Brooks was convicted of murder and sentenced to life in prison.

Four months later, in May of 1992, 3-year-old Christine Jackson was abducted from her room, raped, and murdered. Jackson lived just a few miles from where Smith had been murdered. Local law enforcement officials again focused their investigation on the boyfriend of the victim’s mother, in this case Kennedy Brewer. West again claimed to have found bite marks on the victim, and again claimed he could match the marks to the prosecution’s main suspect, to the exclusion of anyone else. Brewer was convicted in 1995 and sentenced to death.

Today, the ABFO disclaims Michael West, calling him a rogue whose methods were never embraced by mainstream analysts. “Things have changed since the 1990s,” says Peter Loomis, the ABFO president. “This is an evolving field.We want to do the right thing.”

It’s true that in 1995, the ABFO gave West a one-year suspension. To date, it is the only time the ABFO has ever disciplined one of its members. (West protested by resigning from the organization.) But though that suspension was in effect during Kennedy Brewer’s trial, it still didn’t prevent West from testifying. And up until that point, West was an ABFO-certified forensic odontologist, despite the fact that he was already regularly giving testimony well outside the constraints of reality.

By the time the Brewer case made its way to the Mississippi Supreme Court in 1998, the problems with West were well-known. Yet the court still upheld his testimony, explicitly writing that West “clearly” had the “knowledge, skill, experience, training, and education necessary” to testify as an expert witness. The next year, the same court also upheld the conviction of Levon Brooks. This time, the court explicitly took judicial notice of bite mark analysis, writing, “We now take the opportunity to state affirmatively that bite mark identification evidence is admissible in Mississippi.” Just one justice dissented. Among the authorities the court cited in that opinion: the Wisconsin Supreme Court’s Stinson decision.

Also in 1992, West’s testimony helped convict Eddie Lee Howard of killing 84-year-old Georgia Kemp. In his closing argument, Mississippi District Attorney Forrest Allgood made reference to the growing criticism of West, but he told the jurors that this was merely a testament to West’s brilliance. West, Allgood said, was a great mind of science who was merely ahead of his time. “Whether we like to think so or not, the progress of mankind has been carried forward on the backs of people like Michael West,” Allgood said. “The church threatened to burn Copernicus because he dared to say that the planets didn’t revolve around the earth. So it was with Michael West.” (Allgood also seems to have mistaken Copernicus for Galileo.) The jury convicted Eddie Lee Howard, and sentenced him to death.

In 2001, West was further exposed, this time by Christopher Plourd, the attorney for Ray Krone, the Arizona man convicted by Ray Rawson. Angered by what had happened to his client, Plourd set out to demonstrate the illegitimacy of bite mark analysis. Plourd had a private investigator send West photos of the bite marks on the woman Krone was wrongly convicted of killing. The investigator also sent West a dental mold of his own teeth, a retainer check and a made-up cover story. He asked West if he could match the crime scene photos to the dental mold he had enclosed. West not only confirmed the match, he sent back a 20-minute video in which he confidently explained why only the person whose teeth were represented in the dental mold could have left the marks on the victim. That of course was impossible.

Five years later, the Mississippi Supreme Court heard the appeal of Eddie Lee Howard. The court was now aware of all of West’s exploits, including Plourd’s “proficiency test.” Incredibly, the court still upheld West’s testimony, explaining that “just because Dr. West has been wrong a lot, does not mean, without something more, that he was wrong here.”

Since then, videos have emerged of some of West’s bite mark examinations. In them, he is shown repeatedly jamming suspects’ dental molds into the skin of the alleged victims. Forensics specialists have said that at minimum, the videos depict gross malpractice and reveal West to be tampering with evidence. But some experts, like Michael Bowers (see part one for more on him), say the videos show West actually creating the bite marks he would later claim in testimony were inflicted by the suspect.

In 2007, two years after that Mississippi Supreme Court’s ruling in the Howard case, Kennedy Brewer and Levon Brooks were exonerated. DNA taken from the young girls’ bodies in both cases were finally run through the state database. The DNA profile from both cases was the same; it matched Justin Albert Johnson, a man who lived near the scene of both crimes. He later confessed.

Johnson was initially a suspect in the first murder before West matched the bite marks to Levon Brooks. Not only did West’s bite mark matching wrongly imprison two innocent men for 15-plus years, he may well have allowed Johnson to remain free to rape and murder Christine Jackson.

Recently, even West himself said he no longer believes in the validity of bite mark matching. (Though in typically odd fashion, he still defended his own bite mark testimony.) But that doesn’t mean the state’s courts and prosecutors are overturning convictions that West was a part of. In the summer of 2011, Mississippi Attorney General Jim Hood claimed he had opened an investigation into West. Months later the assistant district attorney in charge of that investigation said it had consisted of no more than a “Westlaw search” of relevant cases — the legal equivalent of a Google search. Three-and-a-half years later, nothing else has come of that investigation. Instead, Hood and his subordinates continue to argue that the defendants still in prison due to West have already used up their opportunities to challenge his credibility. That is, they either tried to challenge West or bite mark evidence in general and lost, or they never tried to challenge, and now it’s too late.

Just last month, Hood’s office filed another brief in the Eddie Lee Howard case which astonishingly cites the Mississippi Supreme Court’s rulings in the Brooks and Brewer decisions. The brief points out that the state’s supreme court has “unequivocally held on direct appeal that Dr. West was qualified to testify as an expert in forensic odontology,” and that “bite-mark evidence is admissible in Mississippi.” Legally speaking, Hood’s office is correct. Brewer and Brooks may have been innocent, but they weren’t exonerated by a precedent-setting court opinion. So just as Stinson is still the law in Wisconsin, the Mississippi Supreme Court’s rulings in Brewer and Brooks are still good law in Mississippi — and the state’s controlling law on bite mark evidence.

If he wanted, Hood could release Howard tomorrow, or even agree to a new trial without using West’s testimony. Instead, Hood’s office is essentially arguing that the people convicted based on the testimony of a man Hood himself has said is untrustworthy should be kept in prison on a technicality. Or possibly even executed. Currently, two men are still on death row due primarily due to bite mark analysis performed by Michael West — Howard in Mississippi, and Jimmie Duncan in Louisiana.

Though bite mark analysts like West and Rawson are now either disclaimed or quietly ignored by the ABFO and the community of bite mark analysts, their impact on the field is hard to overstate. Rawson, as noted, was a former president of ABFO who helped write the first guidelines for bite mark analysis. And both West and Rawson were as prolific at authoring articles on their methods for forensic journals and odontology textbooks as they were at testifying in court.

The ABFO may now try to distance today’s bite mark analysts from men like Rawson, Campbell and West, but those figures wielded enormous influence in the field during the era when the courts were issuing precedent-setting opinions about admissibility. (Rawson is still a member in good standing.) And that influence persists. As noted earlier, West is still considered a reliable expert by the Mississippi Supreme Court and the office of Mississippi Attorney General Jim Hood. And it isn’t just in Mississippi. In 1994, John Kunco was convicted of rape due to bite mark analysis that was based on the methods first pioneered by West. That conviction was upheld by a Pennsylvania judge in 2011.

More to the point, the ABFO still embraces members who have participated in more recent wrongful arrests or convictions. Some of them hold or have held leadership positions within the organization. The ABFO has never sought to discipline or file ethics complaints against those members. Instead, as this series will explore tomorrow, the group’s leadership has focused on ruining the people who have helped expose those wrongful arrests and convictions. Bite mark matching in America began with a literal witch hunt. Its proponents are engaged in a figurative one today.