DELPHI, Ind. — “This is not settled science.”

Tania Brief, a senior staff attorney with The Innocence Project, is talking about a type of commonly used forensic examination accepted by law enforcement and courts for decades: tool mark analysis.

It is how the Indiana State Police Laboratory says it confirmed an unfired 40-caliber cartridge was ejected out of a handgun of Richard Allen. The unfired round was found between the bodies of Abigail Williams and Liberty German. This is the only piece of publicly disclosed evidence linking the accused killer to the crime scene.

In the probable cause affidavit used to charge Allen with two counts of murder, it is stated as fact, “(an) upsent .40 round between the bodies of Victim 1 and Victim 2 was forensically determined to have been cycled through Richard Allen’s Sig Sauer Model P226.”

To understand how such a match is determined, we talked with Chris Monturo who has 26 years of experience and a forensic consulting and testing company in the Cincinnati area.

Monturo explained while ammunition is generally made of soft metals like brass, copper and lead, firearms are made of much harder metals.

“The extractor is steel. The ejector is steel. Those are going to scratch that cartridge case as it gets thrown out of the gun,” explained Monturo.

He added, “When you dive deep and look deep into the marks, microscopically they be different from gun to gun.”

FOX59 Investigates asked how confident he was that he could match a cartridge to a specific gun. Monturo responded that he’s “absolutely confident” because of his training and experience.

There are lots of studies which back up Monturo’s certainty. A spokesperson for the U.S. Department of Alcohol, Tobacco and Firearms pointed us to several, including one published this year in the Journal of Forensic Sciences. In it, a trio of Federal Bureau of Investigation scientists gave 172 certified tool mark examiners cartridges and guns and asked them to look for matches.

The reported instance of false-positive errors (where a cartridge and a gun were incorrectly matched) was 0.933%.

But Brief said research backing tool mark analysis is flawed.

“The problem is that you can have a million studies that purport to show something, but if those studies are not properly designed, they’re not meaningful.”

The Innocence Project, along with an array of university scientists, point to issues with research supporting tool mark analysis. They include small sample sizes, standards set by tool mark examiners themselves, and often studies are financed by law enforcement agencies that benefits from positive tool mark matches.

Monturo flatly denies tool mark examiners are under pressure to produce cartridge-gun matches.

“My findings are my findings. Quite frankly, I get paid the same whether it IDs the gun (or not).”

But The Innocence Project is pressing on against tool mark evidence. During its 30 years of existence, it has used DNA to exonerate 375 people who were wrongfully convicted. An examination found flawed forensic evidence played a role in 51% of those erroneous convictions.

“The question for courts is whether or not the science is current and reliable,” said Brief.

An example of that effort is an appeal underway in Maryland of the murder conviction of Kobina Ebo Abruquah. Thirteen university professors are challenging the testimony of a tool mark examiner that linked bullets found in the murder victim to Abruquah’s gun.

The filed brief states studies supporting the analysis are “well below thresholds of scientific validity.”

This dispute looks likely to playout in the upcoming Allen murder trial. In a press release last week, Allen’s attorney dismissed the prosecution’s theory of a “single magic bullet” said to be matched to their client’s handgun. The release goes on to state, “it is safe to say that the discipline of tool-mark identification is anything but a science.”

Out of respect for the gag order in place in the Richard Allen case, FOX59 Investigates did email Allen’s attorneys and the Indiana State Police, not for comment, but to inform them this story was about to be posted.