On a Monday morning in late September, I arrived at a house in a gated subdivision in Alabama and asked for James F. Cooper, a retired agent with the federal Bureau of Alcohol, Tobacco, Firearms and Explosives. A tall, sturdy man in his 70s came to the door a few minutes later. His white hair was in a slightly overgrown crew cut; he wore athletic clothes and navy blue Crocs. “What can I do for you?” he asked, stepping outside.
I wanted to talk about an old arson case he investigated in 1992: a fatal fire at a small, one-story house in Old Hickory, Tennessee, just outside Nashville. A 24-year-old woman named Lorie Lee Lance had died in the blaze. Her boyfriend, Claude Francis Garrett, was arrested for setting the fire. He swore he was innocent. But two separate juries convicted Garrett of murder, first in 1993 and then again in 2003. Cooper was the star witness for the state.
Cooper recalled the case. He also remembered my previous attempts to reach him about it, for a story I published in 2015. The case was fairly unusual, Cooper said. As a federal agent, he did not generally work local arson cases, but he’d been called by the Nashville Metro Police Department about a suspected homicide early in the morning. It was February 24, 1992. Cooper could still describe the scene, along with most of the story told by Garrett: After a night of drinking with Lance and her stepfather at a local bar, he had awoken to the house on fire. Garrett yelled for Lance and ran with her toward the front door, he said, but she turned back toward a room at the other end of the house, where she was later found dead from smoke inhalation.
At trial in 1993, Cooper gave expert testimony to bolster the state’s theory against Garrett: that he was an abusive boyfriend who locked Lance in the back room, poured kerosene throughout the house, lit it on fire, and left her to die. The jury found him guilty and sent him to prison for life. But it would not take long for doubts to emerge. Garrett’s original conviction was overturned when he discovered that the trial prosecutor had concealed a police report in which a key witness said the back-room door had been unlocked.
The investigation into the fire was looking increasingly like a relic from another age.
More significantly, in the decade between the first trial and the retrial, the field of fire investigation had radically transformed. Old assumptions about arson and fire behavior were debunked and new investigative methods were adopted. The so-called pour patterns found at the scene would come to be regarded as junk science. By the time of Garrett’s 2003 retrial, the investigation into the fire was looking increasingly like a relic from another age, resting on techniques that had long been discarded. Nevertheless, Cooper defended his findings on the stand and the jury sent Garrett back to prison.
Cooper did not wish to revisit the case when I first sought to interview him in 2014. Neither did the two prosecutors who tried Garrett, although both were at least willing to briefly discuss the case. One of them — Jon Seaborg, who handled the retrial — acknowledged the challenge of outdated forensics. “As the science changes somebody needs to pay attention to it,” he said, adding “I don’t know how you do that.”
Regardless, Cooper was the person I had been most anxious to reach. Although he retired before the 2003 retrial, it was unfathomable that he would have remained unaware of the sea change in fire investigation since 1992. Yet he had repeatedly rebuffed my attempts to speak to him, even rejecting a list of questions sent via a colleague at the ATF. Now, as we spoke on his doorstep in Alabama, I repeatedly offered him an envelope containing a pair of scientific reports on the fire in Old Hickory. Written in 2016 by a group of renowned fire scientists who had reviewed Cooper’s investigation, the reports were firm in their conclusions: There was no evidence to support a determination of arson in Garrett’s case. “The central piece of evidence of the use of an accelerant is now recognized as a myth,” one of the authors explained. “A modern fire investigator would not find that this fire was incendiary.”
Cooper did not want the envelope. “I stand by my report,” he told me. He did not appreciate other experts second-guessing his work — especially people who had not worked the scene themselves. “My rule of thumb is ‘Were you there?’” he said. It echoed Cooper’s testimony during cross-examination in 2003. “If I’m proven wrong, I will admit I am wrong,” he said. “But on this one, no sir. I was there. I saw it with my eyes.”
Convictions Stuck in Time
In the 3 1/2 years since I first wrote about Garrett’s case, several people have been exonerated in old arson cases. Their cases are included in the National Registry of Exonerations, which tracks cases involving false or misleading forensic evidence. William Amor was acquitted in an Illinois retrial earlier this year, more than two decades after being convicted of killing his mother-in-law in a 1995 fire. Adam Gray was exonerated in 2017 for setting a fire that killed his upstairs neighbors in Chicago in 1993. And Herbert Landry was exonerated last year of trying to burn down his apartment complex in 2006; investigators found “pour patterns” like those at the scene in Garrett’s case. Other arson defendants have had their convictions overturned or reduced without being declared innocent. They are not included in the registry, making a tally hard to come by. Among them is Leticia Smallwood, who was finally released in Pennsylvania this year, after more than four decades in prison on dubious arson charges.
Despite the inescapable reality that flawed fire investigation methods once sent innocent people to prison, most states have taken no systematic steps to revisit old arson cases. A more common response has been to sweep wrongful convictions under the rug. In the case of Angela Garcia, which I wrote about in 2017, the Cuyahoga County Prosecutor’s Office spent months delaying an evidentiary hearing that would likely have dismantled their case against her, only to suddenly offer her a deal if she pleaded guilty right then and there. Despite swearing her innocence for the fire that killed her two daughters, Garcia tearfully took the deal. She is slated for release in 2022.
Now Garrett, too, has a shot at release. This year marked his 25th in prison, making him eligible for a parole hearing. On October 8, he will go before the Tennessee Board of Parole at Riverbend Maximum Security Institution in Nashville. In theory, he has a decent chance: Garrett has been what is often described as a “model inmate,” with a clean disciplinary record and a long list of people willing to vouch for his character. But such things can only take Garrett so far. In Tennessee, as in most states, parole hearings are often little more than a referendum on the original offense, no matter how much a person may have changed behind bars. For those who insist upon their innocence, there’s another dilemma: the expectation that there be a display of remorse for their crime.
Garrett has always said he will not apologize for a crime he did not commit. But forcefully invoking his innocence before the board could backfire.
Garrett has always said he will not apologize for a crime he did not commit. But forcefully invoking his innocence before the board could backfire. A safer route would be to rely on the numerous letters sent on his behalf, which argue that Garrett is a worthy candidate for release who has served his minimum sentence and deserves a chance to thrive on the outside. Still, he wants the parole board to understand the injustice of his case. Among the people Garrett has asked to speak at the hearing will be one of his most passionate advocates, Stuart Bayne, a veteran fire investigator based in East Tennessee. Bayne was the defense expert at Garrett’s 2003 retrial. For him, the hearing is a chance to correct an egregious wrong that has plagued him for over 15 years. In his letter to the parole board, Bayne called Garrett’s case “a classic example of injustice.”
“Tremendous advances in the understanding of fire behavior have occurred since 1992,” Bayne explained. Enclosed with his letter were seven flash drives, one for each member on the board. They contained his own reports on the case, a statement from Garrett, and the pair of expert reports I tried to show to Cooper at his house. If anyone is in a position to take them seriously now, it’s the board. “Please, please Mr. Chairman, distribute one to each member,” Bayne wrote. “Please, please, review the files.”
“Seeking Justice Through Fire Science”
From the moment I began looking into the Garrett case in 2013, it was impossible to miss the parallels to a more famous case — that of Cameron Todd Willingham, executed in Texas in 2004. The fires in their respective cases had occurred just two months apart — and both men were convicted largely on circumstantial evidence. In both cases, neighbors who initially described panicked behavior at the scene would later come to believe that it was just a larger deception. But most compelling was the fact that the evidence in both cases included “pour patterns” that had been disastrously misinterpreted.
I knew from Willingham’s case that burn marks once associated with arson could actually be the result of a long-misunderstood phenomenon called “flashover” — a transition phase during which a room’s contents simultaneously ignite. The physical evidence left by a “post-flashover fire” includes burn marks and patterns that form depending on factors like oxygen and ventilation, but which were once believed to be evidence of an ignitable liquid. After meeting Bayne in 2013, he gave me repeated lessons in such fire scenarios. At one meeting, he had me watch a clip from a short film by the National Fire Protection Association called “Countdown to Disaster.” It showed how quickly and dramatically something like a smoldering cigarette dropped on an upholstered chair could lead to a conflagration. It also illustrated his own theory of how the fire in Garrett’s case started — a cigarette left unattended on a loveseat in the living room after a night of drinking.
Like Willingham’s case, the fire in Old Hickory had occurred at a moment when fire investigation was on the cusp of a revolution. In 1992, the National Fire Protection Association published “NFPA 921,” which set forth a new set of guidelines for fire investigators, applying the scientific method to fire scenes rather than relying on investigators’ observations and experience. At first, “nobody accepted it,” Hurst said. Instead, professional fire investigators spent the next several years looking for ways to get around it. If Garrett’s 1993 trial was too early for the teachings of “NFPA 921” to have been absorbed, by 2003, they were more firmly established. Cooper would have known about them by then — “if he wanted to know,” Hurst said.
“It’s a case in which the investigation was a typical piece of crap.”
Hurst’s own foray into criminal cases came in the mid-1990s, when he testified in the trial of a woman named Sonia Cacy. She had been convicted of murdering her uncle in a fire and was facing a retrial. Like Bayne in Garrett’s case, Hurst was unable to sway the jury, and he was scarred by the outcome. “I had to get her out,” Hurst said. When Cacy went up for parole, Hurst gave a presentation to show that the fire scenario had been impossible. The parole board granted her parole and she was released, although it would take another 20 years for her to be exonerated.
Cacy’s case was highly publicized. But it was the Willingham case that broke into mainstream consciousness, Hurst said. “Every fire investigator in the country knows Willingham didn’t do it. Everyone,” he told me. It helped attract new fire experts to his cause, including chemists and engineers — a rarity among fire investigators, most of whom come from firefighting or law enforcement backgrounds. It also helped shift the perception of defense experts in arson cases, who were generally dismissed as “high-priced defense whores,” according to Hurst. “It was an unsavory sort of profession. How dare you work against the noble police? The noble fire marshals? The noble ATF?”
Hurst passed away just a few weeks after The Intercept published my story about Garrett. By then, the article had reached a group of experts who review old fire cases pro bono. They convened remotely under the banner of the Tetrahedron Committee, a loose consortium of fire scientists and veteran investigators started in 2007. The title is a reference to the four factors that combine to generate fire: fuel, heat, oxygen, and a chemical reaction. The committee’s motto is “Seeking Justice Through Fire Science.”
In April 2016, Craig Beyler, a respected Maryland-based fire engineer — and the author of a famed report on the Willingham case — produced an assessment of Garrett’s case on behalf of the Tetrahedron Committee. It echoed what Bayne had explained to me over and over again beginning in 2013: The investigation had been fatally flawed. The house was filled with furniture, paneling, and materials that were critical to explaining how the fire started and spread. Yet Porter, the fire marshal investigator who was first on the scene, did not bother to note this evidence. “As was common in the day, he simply removed all the contents of the room and hosed out the room to display the floor damage pattern,” Beyler wrote. “He treated the remains of the room contents as an obstruction to viewing the floor, rather than as evidence to be studied. All the contents were simply thrown out into the yard.”
The same month, a renowned fire scientist named John Lentini submitted an affidavit to Garrett’s federal public defender. Lentini drew from the history he lays out in his textbook, “Scientific Protocols for Fire Investigation,” to show how Cooper was emblematic of the initial resistance to “NFPA 921.” Though there had been a marked shift around 2000, “when this case was tried in 2003, some fire investigators, including Agent Cooper, still believed that by looking at the shape and texture of burning on the floor, they could infer the presence of ignitable liquids, even if subsequent laboratory analysis failed to reveal the presence of any such residues,” Lentini explained. Indeed, while “Cooper repeated the phrase ‘pour pattern’ over and over in his testimony,” no kerosene residue had been found in the flooring samples.
Lentini’s affidavit also contained evidence that was a revelation to Garrett — and which he believes should exonerate him once and for all. In 2013, researchers published a study called “Forensic Analysis of Ignitable Liquid Fuel Fires in Buildings,” sponsored by the National Institute of Justice. “What they learned is that ignitable liquids only burn for a very short time in fires, and do not cause the kind of charring found on the floor in the Garrett residence,” Lentini wrote. The photos from the scene showed the charring to be too deep, Lentini explained. In other words, it “could not have been caused by an ignitable liquid. Agent Cooper’s testimony on this point was scientifically unsupportable and erroneous.”
As I listened to Cooper’s recollections of the fire in Garrett’s case, I was struck by his remarks about pour patterns, which sounded far less cavalier than what I recalled from the trial transcripts. On the stand in 1993, Cooper had not only insisted that the marks were proof of a liquid accelerant, he had also claimed to be able to tell the difference between a deliberate pour and a spill. Now Cooper explained the need for caution. He recalled a different fire scene where he discovered what he believed to be pour patterns, only to be told by a forensic chemist that the marks were the result of a varnish that had been applied to the floor. Regardless, Cooper remained confident about the ones in the Garrett case. “I don’t make my determination by one thing, like pour patterns,” he told me. He looked at the totality of the evidence.
Cooper retained a trait that is a trademark of certain old-school fire investigators: a deep belief in his own instincts. “Can you tell from looking at a house where a fire started?” he asked me. No, I said. “See, I can,” he responded. “I have the advantage.” In reality, determining a fire’s point of origin requires far more than visual analysis. But as with any expert witness, the credentials of an ATF agent can impress a jury no matter how flawed their testimony. In the Angela Garcia case, Cleveland prosecutors struggled to win a conviction — her first two trials ended in hung juries — until they bulked up their witness list at her third trial, adding an ATF agent who insisted he could see a pour pattern in a photograph. “I don’t care what the NFPA says.”
Tunnel vision would certainly help explain Cooper’s sloppy investigation. “I don’t remember if I ever interviewed Garrett,” he said. He didn’t. Nor did he speak to the firefighters at the scene. It was not even clear how he became convinced that the latch to the back door had been locked. Neither the door nor the lock was removed or studied, let alone presented to the jury as evidence. If Cooper had scrutinized this crucial piece of evidence, Lentini wrote in his affidavit, he would almost certainly have found that the latch had been in an unlocked position. “Agent Cooper was shockingly uninterested in the fact that the fire left carbon deposits on the sliding bolt showing that the bolt was slid to the right at the time of the fire,” he wrote.
Cooper offered no explanation for why he did not study the latch or the door. And he flatly denied saying something that had jumped out at Bayne when he first started studying the case: that Garrett likely singed his face from leaning down to light the kerosene in the living room. Cooper had compared it to lighting a gas grill, according to transcripts of his trial testimony — “sometimes it won’t ignite and you stick your head down there to see and you’ve got this open flame that comes back with a POOF!” Such a “flash back” of kerosene vapors was “impossible,” Lentini wrote.
“He’s right,” Cooper told me. He could not imagine having made that claim, he said. I told him it appeared in the transcripts, but he protested that he did not remember. “I would be crazy to say that.”