In the tiny central Texas town of Comanche, in the same courtroom where he was convicted nearly three decades ago of murdering his wife, Joe Bryan and his attorneys finally got the opportunity this week to make their case that the 77-year-old former high school principal deserves a new trial.

Over three days, they presented a succession of witnesses who testified that the forensic evidence used to convict Bryan of his wife’s 1985 murder was faulty, and that new evidence pointed to a possible alternative suspect in the killing.

The hearing — which held out the tantalizing promise of winning Bryan a fresh chance to prove his innocence — drew dozens of spectators who packed the hard-wooden benches inside the Comanche County courthouse. Lining the two front rows were Bryan’s family members, who wore matching red t-shirts emblazoned with the words “Justice For Joe.”

Behind them, Clifton residents who had not seen Bryan since the 1980s crowded in beside law students, reporters, and top attorneys with the Innocence Project of Texas, who were on hand to assist with Bryan’s case. Penny Blue, the sister of Joe’s late wife, Mickey, sat by herself in the back of the courtroom.

Two retired prison employees, former Texas Department of Criminal Justice Lieutenant Frank AuBuchon and ex-prison chaplain Carroll Pickett, were also in attendance. AuBuchon had driven all the way from Missouri to show his support. Bryan had worked for both men at the Walls Unit in Huntsville, Texas, in the late 1980s and early 90s, and they had come to believe in his innocence. “My conscience told me to be here,” AuBuchon said.

Bryan sat just a few feet away, behind the defense table, in a black-and-white striped jail jumpsuit. He was barred from speaking to anyone other than his lawyers, but every so often, he surveyed the courtroom for familiar faces, or glanced over his shoulder at his older brother, James, and smiled.

In May, Bryan’s case was the subject of a two-part investigation by ProPublica and The New York Times Magazine, which questioned the accuracy of the bloodstain-pattern analysis used to convict him and highlighted the scant training of some analysts, who nevertheless testify in criminal cases as experts.

At the time of Bryan’s arrest, he had no criminal record. He and his wife, Mickey, a fourth-grade teacher at Clifton Elementary School, were both beloved members of their community, and by all accounts, had a happy marriage. In the days surrounding Mickey’s murder, Bryan was attending a principals’ convention in Austin, 120 miles from Clifton, where the crime occurred. He has always maintained that he was in Austin, asleep in his hotel room, when Mickey was shot in their bed.

Last month, Bryan’s case commanded the attention of the forensics community when the Texas Forensic Science Commission — which investigates complaints about the misuse of forensic testimony and evidence in criminal cases — announced that the blood-spatter analysis used to convict him was “not accurate or scientifically supported.” Spurred by the Bryan case, the commission had already moved to end the practice of allowing law enforcement officers with minimal training in bloodstain-pattern interpretation to testify in Texas, stipulating that such analysis must be performed by an accredited organization if it is to be allowed in court. The decision is expected to prompt other states to follow suit, as the influential commission’s reforms often do.

This week’s evidentiary hearing was the culmination of years of efforts to have Bryan’s case re-examined, and the legal proceedings unfolded with all the intrigue and plot twists of a high-profile murder trial. One of the country’s most prominent bloodstain pattern analysts would discredit a former student who was a key witness for the prosecution. The former district attorney who prosecuted Bryan not once but twice would claim that he could not remember basic facts of the case. And four women would give powerful testimony that raised questions about whether a now-deceased former Clifton police officer could be the real killer.

The potential of the hearing to serve as a referendum on the use of flawed forensic testimony in the courtroom also drew Michael Morton, who has become a powerful symbol of Texas’s efforts to reform its criminal justice system. Morton spent nearly 25 years in prison for the murder of his wife, Christine, before being exonerated in 2011 when DNA analysis implicated the actual killer. “Joe Bryan’s situation dredges up old memories for me, because I can’t help seeing myself there, in his shoes,” Morton said. “We were both convicted because of misapplied science and character assassination.”

Morton’s presence was a reminder that the Bryan case — which is still being investigated by the Texas Forensic Science Commission — has far-reaching implications. The hearing would prove to be not only about the corrosive power of junk science in the courtroom, but about the resistance of a justice system to reconsider criminal convictions, even one in which a man has already served more than 30 years in prison.

Day One: Blood Spatter

Defense attorney Jessica Freud began by trying to show that some of the most damning testimony in Bryan’s trial and retrial — testimony that came from former police detective Robert Thorman — was simply wrong.

Thorman had just 40 hours of training in bloodstain-pattern analysis when he was called in to work on the Bryan case. Bloodstain interpretation is a forensic discipline whose practitioners regard the drops, spatters, and trails of blood at a crime scene as treasure troves of information that can sometimes illuminate the choreography of the crime itself.

The most important piece of evidence in the case was a blood-speckled flashlight that Mickey’s brother, Charlie Blue, found in the trunk of Bryan’s car four days after the murder. A state crime lab chemist in 1985, working before the advent of DNA analysis, had only been able to determine that the blood on the flashlight was type O, which corresponded to Mickey, as well as nearly half the population.

Thorman’s testimony helped the state not only connect the flashlight to the crime scene, but explain away the inconsistencies of the state’s evidence, including the fact that no blood was found in the interior of Bryan’s car. Thorman testified that the flecks of blood on the flashlight lens were “back spatter,” which is indicative of a close-range shooting. With the help of prosecutors, he wove a narrative that suggested the killer was holding the flashlight in one hand when he shot Mickey with a gun he held in the other, and that he cleaned himself up, changing his clothes and shoes in the master bathroom, before fleeing the scene.

At the hearing, Celestina Rossi, a senior crime scene investigator with the Montgomery County Sheriff’s Office, near Houston, called Thorman’s testimony “false” and “unscientific,” adding, “It’s the most horrifying thing I’ve ever read in my life.”

Rossi had previously reviewed the Bryan case at the behest of the Texas Forensic Science Commission. When she took the stand in Comanche, she eviscerated Thorman’s analysis of both the flashlight and the crime scene, arguing that the detective had misstated scientific concepts, used flawed methodology, and incorrectly interpreted evidence. She highlighted claims he made that she said had no grounding in science — such as his contention that human blood dries in three to seven minutes and possesses “its own characteristic geometric patterns.”

She also attacked several critical components of the state’s case. Regarding Thorman’s testimony that the killer must have cleaned up in the master bathroom before leaving the Bryan home, Rossi explained that there was no evidence to back up this claim. “The first thing I would do if I thought there had been a cleanup is I’d test the sink and countertop and shower … to see if I could pick up any trace amount of blood,” she said. Thorman, she added, had failed to do so.

Rossi also tried to discredit the prosecution’s theory that the flashlight was held in the killer’s hand when Mickey was shot. If that were true, she argued, the spatter on the flashlight would have a “similar orientation” as it radiated backward. Instead, she explained, the spatter appeared to come from different directions.

On cross-examination, Rossi said that she could not eliminate the flashlight from the crime scene. But she could not tie it to the scene, either.

Most of her findings were backed up by the state’s expert in bloodstain-pattern analysis, Tom Bevel, who said that Thorman gave inaccurate testimony and was not qualified to be an expert. Bevel, who wrote the discipline’s seminal textbook, faulted Thorman for numerous errors, adding that he believed the prosecution’s key witness “doesn’t understand how to properly do this.”

Bevel disagreed with Rossi on one fundamental issue: he said the pattern of blood on the flashlight could be the result of “back spatter” — the signature of a close-range shooting. “I can’t rule that out,” he testified.

Bevel’s critique was particularly powerful because it was Bevel himself who had taught the 40-hour blood-spatter course that Thorman took in 1985. Bevel was never asked if he had concerns that one of his own students had so grossly misunderstood the concepts he taught or that others might do the same. (Bevel’s consulting firm continues to teach 40-hour classes in bloodstain-pattern analysis to police officers around the country for about $700 per person.) Instead, he laid the blame on Thorman for agreeing to testify as an expert before he had taken more advanced coursework.

Students who complete his introductory class, Bevel said, are given a warning before they head back out into the world: “You are dangerous as hell right now.’”

---

Day Two: The Special Prosecutor and the Alternate Suspect

With their challenge to the blood-spatter testimony behind them, Bryan’s attorneys turned to another critical issue: the state’s use of a special prosecutor who was hired and paid by the victim’s brother, Charlie Blue. Under Texas law, a district attorney cannot hand over the control of a case to a private lawyer. If the defense could show that former Bosque County D.A. Andy McMullen had ceded control, it could serve as grounds for a new trial.

The day started off with a series of statements by McMullen, who the defense called to the stand. The former prosecutor said he recalled little about the Bryan case — one of the highest-profile cases of his career, and one that he had described at Bryan’s retrial as “the most brutal crime in the history of Bosque County.”

Yet on Tuesday, McMullen, who is now in private practice, said that he could not remember the year Mickey was murdered; what role Charlie Blue had played in the case; the reason Bryan’s first conviction was reversed; and other basic details, like the names of two prior Clifton police chiefs.

Despite knowing that he would be called to testify, McMullen said that he did not review the record, only a few letters he had written. As Bryan sat before him listening intently to his testimony, McMullen was dismissive of the defense’s efforts to question him about the decades-old case. “I haven’t thought about it in years,” he said.

When asked about the infamous case of Judy Whitley, a 17-year-old Clifton high school student who was raped and murdered four months before Mickey Bryan, McMullen said, “I remember hearing the name. Was that a murder case?”

At times, his answers drew audible gasps from the spectators.

But McMullen easily remembered details about the case that were not favorable to Bryan, recalling, for example, the presence of two plastic particles on the flashlight lens that investigators believed had the same characteristics as fragments of the birdshot shells found at the crime scene.

Bryan’s attorneys had discovered memos in the D.A.’s files describing meetings that were held at the office of the special prosecutor, Garry Lewellen, with all the key law enforcement officers in the case — from the lead investigator to the state crime lab chemist — that McMullen did not attend. They also found memos drafted at Lewellen’s office that outlined possible motives and trial strategy.

Under questioning, McMullen was adamant that he had always maintained control of the case. He testified that it was his idea, in fact, to bring Lewellen on to help him. “[Texas Ranger Joe] Wilie recommended Lewellen to me,” he testified.

It was an assertion at odds with Blue’s testimony at both trials that he had retained Lewellen’s services himself (Lewellen is now deceased). Hoping to rebut the former prosecutor’s claims, the defense called Wilie to the stand, and asked the now-retired lawman if he had recommended Lewellen to the former D.A. “No, ma’am,” Wilie told Freud. “I didn’t know Garry Lewellen before this case.”

Some of the most troubling testimony came Tuesday afternoon, when a parade of witnesses testified about the predatory behavior of Dennis Dunlap.

A former Clifton police officer who hanged himself in 1996, Dunlap was investigated posthumously in connection with Judy Whitley’s murder. Clifton police subsequently determined that he was Whitley’s killer. Bryan’s attorneys believe that Dunlap is a viable alternative suspect in Mickey’s murder. Several years before his death, Dunlap told his wife that he had been with Mickey on the night she died.

On the stand, several of Mickey’s peers described being pursued by Dunlap. Linda Liardon told of a leering encounter with the police officer that left her fearful for her safety. Susan Ellis recounted how Dunlap attempted to follow her home one night, trailing her down a succession of Clifton streets in his police cruiser, until she was finally able to shake him. Cindy Kissner testified that Dunlap stalked her and threatened to kill her in the mid-1980s. “I lived in fear of what he was going to do to me,” she said, explaining that the Clifton police department did not take her repeated complaints about Dunlap seriously.

Judy Whitley’s older sister, Patricia, testified that Dunlap remained fixated on their family even after Judy was killed, constantly inquiring about the progress of the murder investigation and insisting that he be a pallbearer at the teenager’s funeral.

As each woman told her story, it became obvious how few investigative leads were pursued back in 1985. Wilie recounted escorting Dunlap to a polygraph test in the weeks that followed Whitley’s death. When the test came back inconclusive, Wilie said he told Dunlap, “Inconclusive doesn’t mean you did do it, or you didn’t do it.” Wilie nonetheless appeared to drop his investigation into Dunlap afterward. Inexplicably, Dunlap was not pursued as a suspect in the Whitley murder until after his death, despite circumstantial evidence that pointed to his involvement.

Freud showed Wilie a handwritten note he jotted down in 1985 that suggested that Dunlap, who stepped down from the Clifton police force before Mickey’s murder, was still in the area at the time of her death. She did not press him on why he had not pursued the law enforcement officer more aggressively.

Freud also noted that 33 years after Whitley’s murder, the rape kit in her case, which could answer important questions for both Whitley’s family and Bryan’s, remains untested. The kit was sent to the Southwestern Institute of Forensic Sciences in 2012.

---

Day Three: The Pursuit of Justice and DNA Testing

As the defense wrapped up its arguments about the improper use of a special prosecutor, Freud brought in a local legal legend, Baylor Law School professor Brian Serr. Both Freud and Bosque County D.A. Adam Sibley had been taught by Serr in law school, and they treated him with deference as they questioned him.

A towering, silver-haired attorney with a booming voice, Serr captured spectators’ attention with testimony that brought into focus what was at stake. “I tell my students that the single most important decision they’ll ever make is whether they are going to try to put an American citizen in prison,” he began.

Serr spoke of “the awesome power of the state when used to deprive someone of their liberty,” and the unique duty a prosecutor has “to seek justice, not a conviction.” Lewellen, he argued, had an inherent conflict of interest from the start, because he was being paid by someone who wanted to secure a conviction. “Lewellen was hired by Mr. Blue, who also hired a private investigator, who just happened to be present when some incredibly powerful evidence was discovered,” he said. “That’s not a prosecutor who’s representing the public. It’s inconsistent with prosecutorial ethics. That troubles me. In 32 years, I have not run across an arrangement like this before.”

Serr repeatedly questioned the logic of the D.A.’s office in bringing the case against Bryan to trial in the first place. “If you never prosecute anybody, yes, the guilty person gets away,” he said. “But you compound that problem when you prosecute an innocent person on flimsy evidence. Now you have two injustices.”

He also questioned the key piece of evidence used against Bryan. “There’s this magical flashlight that just appeared, without which there is no conviction, without which Mr. Bryan is not in prison for the past thirty years,” he said. “It doesn’t sit well with me.”

During cross examination, Sibley referred to the possibility that DNA testing could provide more clarity. “What if Mickey Bryan’s blood comes back on that flashlight?” he said.

Even the presence of Mickey’s blood on the flashlight could not definitively prove Bryan’s guilt, Serr explained, since Bryan did not have possession of his car in the days leading up to Blue’s discovery. “Blue happens to find the flashlight in front of the private investigator he’s hired?” Serr said. “I don’t think I’ve ever see anything like that before. It just looks funny. It all looks funny.”

“We also look at the amount of circumstantial evidence,” Sibley said. “Two juries looked at that.”

“What concerns me is the evidence that is entirely absent,” Serr said. “There were no financial problems, no infidelity, no turbulence of any kind in the marriage. Nobody saw Bryan leave the hotel. No one saw him return to the hotel. None of the neighbors heard or saw his car pull up [outside his home]. Nothing puts him at the scene, except for the flashlight in his car. If you were going to shoot your wife, why would you use a gun that would make such a mess? It’s more consistent with an intruder entering the house.”

In the end, Serr said, “There wasn’t proof beyond a reasonable doubt.”

Other than Bevel, the prosecution did not put on any witnesses. At the end of the hearing, Judge Doug Shaver declared an extended recess so that new DNA analysis of the flashlight — which both the prosecution and the defense have agreed to — could be performed.

The hearing is expected to resume in the early fall when DNA results will be available. At that time, lawyers will give closing arguments. Shaver will then make a recommendation about whether Bryan should receive a new trial. His findings will go to the Texas Court of Criminal Appeals, whose justices will be the final arbiters.