A federal judge has ordered Alabama to retry Charles McCrory or release him. For testifying experts, the opinion is less about odontology than about what happens when a discipline’s foundation gives way beneath a conviction.
On September 21, 2026, Chief Judge R. Austin Huffaker Jr. of the U.S. District Court for the Middle District of Alabama granted habeas relief to Charles McCrory and gave the State 180 days to commence a new trial or let him go. McCrory has been in custody since 1985. The physical evidence that put him there consisted of two small marks on his wife’s arm.
The Case in Question
Julie Bonds McCrory, 24, was found dead in Andalusia, Covington County, on May 31, 1985. The couple were separated and a divorce was pending, though her date book suggested Charles might be moving back in. Investigators recovered no blood on his clothing or in his apartment. There was no confession, no DNA, and no motive the State could point to with confidence.
What the State did have was Dr. Richard Souviron, the Florida forensic dentist who had matched wound patterns to Ted Bundy’s dentition in 1979 and was, by 1985, about as close to a celebrity as forensic odontology produced. He told the jury the two marks matched Charles McCrory’s teeth. McCrory, then 26, a tech worker and volunteer paramedic, was convicted and sentenced to life.
Souviron signed an affidavit on December 11, 2019 recanting: he no longer believed the individualized comparison testimony he gave was reliable or proper, and he would not offer the same today. He also confirmed something the jury never weighed — he never examined Julie’s body. He worked from photographs.
In 2021 the Covington County District Attorney offered McCrory time served. Thirty-seven years in, he could walk out that week if he admitted killing his wife. He turned it down. “What am I going to do, make up a story?” The postconviction hearing went forward, Souviron recanted on the stand and went further, testifying that bite mark comparison is inappropriate for criminal cases generally. Other experts agreed with him. The State called no rebuttal witness and offered no contrary evidence.
The state courts denied relief regardless. The Alabama Court of Criminal Appeals reasoned that jurors could simply have looked at the photographs and drawn the comparison themselves. The Supreme Court denied certiorari on July 2, 2024, drawing a statement from Justice Sotomayor.
Judge Huffaker had little patience for the lay-comparison rationale. The finding “blinks reality in light of the evidence,” he wrote. “If an expert could not testify about bitemark identification because it is junk science, a jury should not have been allowed to consider it either. It makes little sense that a lay jury could permissibly form a conclusion that no forensic expert could.” He found the state court’s ruling both an unreasonable application of clearly established federal law and based on an unreasonable determination of the facts — the demanding two-part showing AEDPA requires, and one federal courts rarely find satisfied. On inherent prejudice he was blunt: “The non-bitemark evidence from which the jury could convict McCrory was thin.”
Why the Discipline Was Formed on Wet Sand
Bite mark identification rests on three premises, and the field never established any of them; the first is that the human anterior dentition is unique. NIST’s scientific foundation review, finalized in March 2023 after examining more than 400 publications, found no study establishing that uniqueness and no population data on how often particular arrangements of front teeth recur.
The second is that whatever uniqueness exists transfers faithfully to skin. It does not. Skin stretches, moves under the bite, swells, bruises, and begins healing immediately. The pattern an examiner photographs days later is not the pattern the teeth made.
The third premise is the one that should trouble every practitioner: that examiners can reliably read what did transfer. A 2016 study involving board-certified odontologists found that participants could not agree on whether a given injury was a human bite mark at all, never mind who produced it. The discipline could not reach consensus on the threshold question.
None of this was sudden. The National Academy of Sciences said in 2009 that there was “no evidence of an existing scientific basis for identifying an individual” from bite marks. PCAST concluded in 2016 that the method lacked foundational validity and rated the prospects of it ever achieving validity as low. The Texas Forensic Science Commission became the first regulator in the country to recommend a moratorium, in February 2016.
Who is objecting, and who is not
It would be convenient for prosecutors if the opposition were confined to the defense bar. It is not. The polemisicts are the National Academy of Sciences, the President’s Council of Advisors on Science and Technology, NIST’s metrologists, academic statisticians, a state forensic science regulator, and — increasingly — the odontologists themselves. Souviron is the most visible defector but not the only one.
The Innocence Project has documented 39 wrongful convictions and indictments involving bite mark comparison. Keith Allen Harward served 33 years in Virginia before DNA identified another man in 2016. Robert Lee Stinson served 23 in Wisconsin, exonerated in 2009 when DNA and independent dental review both inherently went against the original testimony. Eddie Lee Howard came off Mississippi’s death row in 2021.
A residual group of practitioners still defends a narrowed version of the work: exclusion-only opinions, no individualization language, the discipline rebranded as pattern injury analysis. The more consequential resistance, though, comes from prosecutors whose arguments are procedural rather than scientific — finality, AEDPA deference, harmless error, the claim that recantation is mere impeachment. Sotomayor identified exactly that pattern in her statement.
What testifying experts should take from it
Rule 702, as amended December 1, 2023, now makes explicit that the proponent must establish each requirement by a preponderance, and subsection (d) asks whether the opinion itself reflects a reliable application of the method. Volume of experience does not answer a question about error rate.
Note also where the failure actually occurred. Souviron’s own 1985 report cautioned that if these two marks were the sole means of identifying the perpetrator, “this is not in the best interest of justice.” The caveat existed. It simply did not survive the transition from report to witness stand. Overstatement, not fabrication, is what does the damage.
Finally, your testimony has a half-life. Changed-science statutes in Texas, California and a growing number of other states allow convictions to be reopened when a field moves beneath them. Record what you examined, record what you did not, and state your limits in language a transcript will preserve.
McCrory is now in his late sixties. Whatever the retrial produces, the arithmetic is settled: forty-one years, two marks, one opinion.
By E Price
- Expert Toolkit: 7 Tools freely accessible to members > Independence & Conflict Tracker, Rule 72 Check, Daubert Preparation Tool, Testimony & Deposition Log and more.
- Join WitnessDIrectory.com here from $230 a year.
Sources in researching this article:
- McCrory v. Alabama, Sotomayor statement respecting denial of certiorari, July 2, 2024
- Bitemark Analysis: A NIST Scientific Foundation Review (NIST IR 8352)
- NIST: Forensic Bitemark Analysis Not Supported by Sufficient Data
- CBS News: Charles McCrory, teeth marks and the 1985 murder case
- Hoodline: Alabama bite-mark ruling, September 2026
- Innocence Project: Description of Bite Mark Exonerations
- Innocence Project: Texas Forensic Science Commission moratorium
- https://www.supremecourt.gov/DocketPDF/23/23-6232/305960/20240327084838648_2024.03.27%20McCrory%20Reply%20to%20BIO.pdf
