A hearing to determine whether Joe Bryan should be granted a new trial came to a dramatic conclusion on Monday with a surprise, eleventh-hour admission from the expert witness whose testimony had proved critical in convicting the former high school principal of the 1985 murder of his wife, Mickey.

“My conclusions were wrong,” retired police Detective Robert Thorman wrote in an affidavit introduced by the defense of the bloodstain-pattern analysis he performed. “Some of the techniques and methodology were incorrect. Therefore, some of my testimony was not correct.”

Bloodstain-pattern analysis is a forensic discipline whose practitioners regard the drops, spatters and trails of blood at a crime scene as clues, which can sometimes be used to reverse-engineer the crime itself. Thorman had only 40 hours of training in the discipline when he was called in to work on the Bryan case. His testimony about a blood-speckled flashlight that Mickey’s brother found in the trunk of Bryan’s car four days after the murder made the state’s tenuous theory of the crime seem plausible.

At Bryan’s trial in 1986 and then again at his 1989 retrial, Thorman testified that tiny flecks of blood on the flashlight could only be “back spatter” — a pattern that indicated a close-range shooting. What connection the flashlight had to the crime, if any, was never clear; in 1985, a crime lab chemist found that the blood on it was type O, which corresponded not only to Mickey but to nearly half the population. But Thorman effectively tied the flashlight to the crime scene, going so far as to say that the killer had likely held the flashlight in one hand while firing a pistol with the other.

Bryan had been attending a principals’ convention in Austin, 120 miles from where the murder occurred in Clifton, Texas, in the days surrounding the murder. He has always maintained that he was in Austin, asleep in his hotel room, at the time of the crime.

During both trials, Thorman also helped explain away one of the biggest holes in the state’s case: no blood was ever found in the interior of Bryan’s Mercury, though the prosecution alleged that Bryan fled the messy crime scene in his car. Thorman provided an explanation for this, asserting that the killer had changed his clothes and shoes in the master bathroom of the Bryan home before making his escape.

“In no way did I lie in my report or testimony,” Thorman stated in his affidavit, which is dated Sept. 13. “I was doing what I thought was correct as a result of my training at the time.” Thorman, who is now 80, did not specify which parts of his testimony had been incorrect.

The affidavit was read aloud on the stand by Montgomery County crime scene investigator Celestina Rossi, who had previously challenged the accuracy of Thorman’s analysis. Prosecutors, who glanced at each other repeatedly as she read the affidavit, appeared taken off guard.

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. The series highlighted the scant training of some law enforcement officers, who nevertheless go on to testify in criminal cases as expert witnesses.

In July, 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.”

Then, on Aug. 20, the evidentiary hearing got underway in the central Texas town of Comanche. For three days, Bryan’s attorneys presented a succession of witnesses who testified that the bloodstain-pattern analysis used to convict Bryan was faulty, and that new evidence pointed to an alternative suspect in the killing, a now-deceased Clifton police officer named Dennis Dunlap.

On Aug. 22, Judge Doug Shaver declared an extended recess so that DNA testing could be performed on the flashlight.

At the hearing on Monday, Brent Watson, a DNA analyst with the Texas Department of Public Safety Crime Lab in Waco, took the stand to testify about the results of his examination. Watson explained that he tested six minute stains on the flashlight. Five of those stains came back negative for the presence of blood. The sixth stain tested positive, but whose blood it was could not be determined.

A partial DNA profile on the lens of the flashlight yielded little information. Whether or not Mickey was a contributor to that profile, Watson said, was inconclusive. “Degradation had an effect on my ability to obtain complete DNA profiles,” he said of the 32-year-old evidence.

Watson, however, was able to glean more information from a partial DNA profile he developed by swabbing the handle of the flashlight for DNA. “Joe and Mickey were excluded as contributors,” he said.

Jessica Freud, one of Bryan’s attorneys, seized on these uncertainties, asking Watson, “Is the state of Texas able to show that Mickey’s blood or DNA is present on the flashlight?”

“No, it is not,” Watson replied.

Watson went on to testify about another notorious piece of evidence from Bryan’s two trials: a pair of Bryan’s discarded underwear, which was found in the couple’s master bathroom wastebasket during the 1985 investigation. At that time, a state crime lab analyst determined that the underwear was stained with semen that matched Bryan’s blood type, and she testified to this fact at both trials. But current testing protocols, Watson said, showed no evidence of semen on the underwear.

The semen-stained underwear had been used to great effect by the prosecution at both of Bryan’s trials. At Bryan’s first trial, the Bosque County district attorney at the time, Andy McMullen, had argued that it was “evidence of a kind of perverted behavior.” The prosecution had always insinuated, Freud told the court, that Bryan had been aroused by the act of killing his wife.

By the close of the hearing, there appeared to be little left of the state’s case. No one could say whose blood was on the flashlight, how that blood had been deposited or when the blood was left there. Nor was it any clearer who had put the flashlight in the trunk of Bryan’s Mercury and when. No semen was found to be present on the underwear. Thorman’s bloodstain-pattern analysis had been debunked.

In his closing arguments, the current Bosque County DA, Adam Sibley, focused on two plastic particles that were found on the flashlight lens. At both trials, an investigator and a crime lab technician had testified that the particles possessed the same characteristics as birdshot shell fragments found at the crime scene. These particles, Sibley said, “proved the flashlight was at the scene of the crime.” Sibley noted that neither Tom Bevel, the state’s bloodstain-pattern analyst, nor Rossi, the defense’s expert, could “eliminate the flashlight from the scene.”

When Freud rose to address the courtroom, she ticked off the lengthy list of questionable evidence and odd circumstances that had allowed Bryan to be convicted, twice, of his wife’s murder: The bad science. The flashlight that had been allowed into evidence even though no chain of custody had ever been established. The underwear that was not actually stained with semen. The special prosecutor paid for by Mickey’s brother, Charlie Blue, even though Blue, because he discovered the flashlight, was the case’s star witness. “It not only gives the appearance of impropriety, it is improper,” Freud said.

She argued that jurors in Bryan’s trials had never had heard compelling evidence that suggested Dunlap, who hanged himself in 1996, could have killed Mickey. After his death, Clifton police launched an investigation in which they determined that he had killed a 17-year-old high school student named Judy Whitley just four months before Mickey’s murder. During that inquiry, an ex-wife of Dunlap’s told investigators that he had boasted of being with the principal’s wife on the night she died. Freud played a brief portion of that interview aloud, and spectators in the crowded courtroom strained to listen.

“There is nothing left,” Freud said of the state’s case, in closing. “There is nothing left here.”

A new trial, she said, should be devoid of faulty bloodstain-pattern analysis and irrelevant evidence, like the underwear; it should not be tried by a special prosecutor paid for by the victim’s family; and it should “permit a full and fair presentation of all of the exculpatory evidence to a jury.” She added: “If they’re going to retry him, make them do it the right way. They’ve never done it the right way.”

Before she concluded, she turned to Shaver. “Joe was in Austin,” she said, her voice rising with emotion. “He never left. He is innocent. Always has been.”

The case will be on hold until Nov. 9, when the state and defense will present their written conclusions about the hearing to Shaver. The judge will then recommend to the Texas Court of Criminal Appeals whether Bryan should receive a new trial. That court’s justices will be the final arbiters.

Many of Bryan’s supporters who crowded the courtroom on Monday — from his former high school students to ex-Clifton Record editor Leon Smith, who has spent decades chronicling Bryan’s case — were disappointed that a potential resolution would be further postponed. Bryan had listened attentively throughout the hearing, sitting behind his attorneys in a black-and-white prison jumpsuit, but he looked deflated at the news, his shoulders sagging at the thought of yet another delay. Bryan, who suffers from congestive heart failure, will turn 78 on Friday.

“How much longer will he have to wait?” said the Rev. Carroll Pickett, who served as the prison chaplain for Texas’ death row for more than a dozen years, as he stood outside the Comanche County Courthouse. “I’ve known Joe for 31 years, and I have yet to hear any evidence that he committed this crime.”