Thirty-one exonerations have come from a re-examination of cases based on the forensic comparison of bite marks.

The unsupported comparison of such bite marks left in human skin during rapes, murders and other violent attacks should be totally thrown out of forensic science, a dentist and researcher writes in the latest Journal of Forensic and Legal Medicine.

While the reconsideration of bite-mark evidence has changed their general acceptance, the backlash against it should be even more complete, writes C. Michael Bowers, of the Ostrow School of Dentistry at the University of Southern California.

“It remains only for courts to view the accumulated disconfirming evidence on bitemark identification and the lack of supporting empirical evidence through the lenses of the law’s demanding standards,” writes Bowers. “When they do, bitemark identification will join the ranks of other deceased forensic science.”

Bite marks first became universally accepted after a 1974 murder of an elderly woman in her California home. She had been sexually assaulted with a knife, and had “an elliptical laceration of the nose” potentially coming from the killer’s teeth. But the analysis and comparison of the marks to the dental molds taken of Walter Edgar Marx was only completed after the victim’s body had been buried for nearly two months—and the remains were exhumed by court order. The evidence was upheld by state appeals courts—and it became the de-facto beginning of widespread use of bite-mark evidence in the United States, according to the new analysis.

But quantifiable tests of the comparisons were not undertaken for decades, according to Bowers.

The first accurate transfer-of-teeth tests were conducted upon cadavers at the University at Buffalo Dental School beginning in 2007, which continued through 2013.

In the interim, the National Academy of Sciences issued its watershed report “Strengthening Forensic Science in the United States: A Path Forward” in 2009. That report took especial aim at bite-mark comparison.

The criticisms have continued. The Texas Forensic Science Commission also issued a condemning report in 2016, based on one of those 31 exonerations, that of Steven Mark Chaney, who was convicted in 1989 of killing a drug dealer and his wife. And the President’s Council of Advisors on Science and Technology, or PCAST, issued a report reiterating many of the complaints outlined in the 2009 NAS report.

The bite-mark gatekeepers themselves also found problems.

The American Board of Forensic Odontology undertook its own tests of bite-mark comparison in 2015. Of the three fundamental questions on the bite-mark identification test, the 39 examiners only agreed at a rate of 90 percent or more in eight of the 100 cases presented, according to Bowers. Just the next year, the ABFO published new standards limiting the bite-mark testimony of its members.

But that is not enough, Bowers writes. Bitemark comparison should go the same way of some 20th century relics of forensic science including phrenology, footwear comparison, hypnosis-aided eyewitness statements, bullet lead comparison and hair comparison analysis, the dentist writes.

Human skin is malleable and changing, and thus impressions of teeth may always be inaccurate—and should therefore not be admitted to courtrooms as identification criteria.

“Even if ‘blinded’ evidence analyses were made mandatory, the lack of established inclusion/exclusion criteria for bitemark comparisons along with the physical vagaries of skin are permanent obstacles for success,” he writes. “Hundreds of cases of bitemark identification used personal confidence as a substitute for empirical data and acceptable statistical testing. This terminology describing confidence is now not supported within the scientific professions and are unacceptable in today’s courts.”