Arson science questioned in two convictions

August 5, 2012

By Frank Green


Two men who claim they are innocent of arsons that killed their children are hoping scientific advancements in fire investigation in recent decades will clear their names.

Davey J. Reedy, 57, is on parole after serving 20 years for the 1987 arson deaths of his two children in Roanoke; and Michael L. Ledford, 36, is serving a 50-year sentence for the 1999 arson death of his 1-year-old son in Stuarts Draft.

Prosecutors remain convinced the two are guilty as found by juries beyond reasonable doubt.

Critics disagree and contend authorities mistakenly relied on what they believed was legitimate science. Some experts believe the two investigations were flawed and that the evidence does not show the two fatal blazes were the result of arson.

Reedy has a clemency petition before Gov. Bob McDonnell, and a clemency petition is being prepared for Ledford. However, unlike a DNA case where proof can be definitive, innocence claims in arson cases are tough to win.

"When you can prove some other dude done it you're in much better shape. With fires, it's not 'some other dude done it,' it's 'nobody done it,' and that's very difficult to prove once you're convicted," said John J. Lentini, a fire investigation expert and consultant.

Nevertheless, more than a half dozen people convicted of arson, usually involving deaths, have been cleared across the U.S. in recent years.

Shortcomings in fire investigations recently gained national attention in the Texas case of Cameron Todd Willingham, executed in 2004 for the 1991 arson deaths of his three children in a conviction critics now say relied in part on little more than folklore.

A four-person panel of the Texas Forensic Science Commission concluded in 2010 that state and local authorities used flawed science in determining the blaze was arson.

Stephen Olenick, an expert who studied Ledford's case, wrote in a report that, "If the Willingham situation should have taught us, the fire investigative community, anything, it is that we should want to get it right, no matter when."

Lentini, author of the book "Scientific Protocol for Fire Investigation," said it is clear that many of the rules of thumb for fire investigation during the 1980s and 1990s have turned out to be based on anecdotal evidence at best and "witchcraft" at worst.

"The ability to determine where a fire started is even more in doubt (now) than it was in the?'80s and?'90s," Lentini said. "The more you learn, the less you know."

* * * * *

Ledford, 36, is in his 12th year behind bars for the Oct. 10, 1999, fire that started a few hours after the family celebrated his son's first birthday. The fire was contained to an area surrounding an upholstered chair in the living room of their two-bedroom apartment in Augusta County.

It broke out after Ledford left the apartment to run errands and to stop by the fire station where he was a volunteer firefighter. His son, Zachary, was killed by smoke inhalation and his wife, Elise, was severely burned but lived.

Authorities first classified the cause of the fire, which largely burned itself out, as undetermined, but later decided it was arson committed by Ledford.

For 4½ hours of a six-hour interrogation, Ledford was adamant he had nothing to do with the fire. He confessed after he was falsely told he failed the polygraph examination.

The Innocence Project says false confessions played a part in 25 percent of the roughly 300 cases across the U.S. where convictions were proven wrongful by DNA testing. In Virginia, that includes Earl Washington, who was wrongfully sentenced to death for a rape and murder. His death sentence was commuted, and he was later given a full pardon.

In his confession, Ledford said he dropped a lit candle on the chair as he left the apartment, which neighbors said was 8:25 to 8:28 p.m. Ledford was seen at a gasoline station filling his car at 8:35 p.m.

At about 8:45 p.m., neighbors saw Elise, wakened by a smoke alarm, in the living room. At 8:47 p.m. two 911 calls — one from Ledford's apartment believed made by Elise — were received reporting the fire, and equipment was sent to the fire at 8:50 p.m.

However, Richard J. Roby of Combustion Science & Engineering Inc. — hired by Ledford's defense for $15,000 — testified that the fire could not have been set the way Ledford confessed it had and that all possible accidental causes had not been ruled out.

Reached last week, Roby still believes that is the case.

"Michael Ledford could not have started the fire because he had left his apartment approximately 20 minutes before the fire was observed and reported by a neighbor, and by his wife," he said.

"The small amount of fire damage caused by this fire and the eyewitness testimony of a neighbor who saw the fire through the apartment window and reported it to 911 prove that the fire was of much shorter duration than 20 minutes," he said.

Roby, a former firefighter with a doctorate from Stanford University, was a professor at Virginia Tech, where he taught courses in combustion, heat transfer, wood mechanics, internal combustion engines and similar areas from 1986 to 1992.

The company revisited the case in 2010, and its report concluded: "New scientific advancements over the past 10 years of modeling toxic gas generation, window breakage, and smoke alarm activation all provide further support for the conclusion that this was not an intentionally started incendiary fire."

But A. Lee Erwin, the commonwealth's attorney for Augusta County who prosecuted Ledford, said Ledford's trial expert testified to much the same thing.

"We had an expert and they had an expert to contradict ours, and the jury believed ours," he said. The Virginia State Police investigator was a graduate of arson investigation schools. An expert with an insurance company testified that the fire started in the chair.

The trial transcript shows Ledford's marriage was stormy and that the lives of his wife and baby were covered by insurance policies totaling $75,000.

Erwin has no doubts about Ledford's guilt. "I thought the evidence was overwhelming. We had an eyewitness who saw him drive and park near the house and wait until the fire was discovered."

John Bennett Allen, an independent, pro-bono advocate for the wrongly convicted, is preparing the clemency petition for Ledford. He said Ledford was at the firehouse when he heard the alarm and drove back there himself in his car.

The evidence in the case, including the chair, nearby items and a lamp — that Ledford said had not been working properly and may have caused the fire — were discarded and not available for Ledford's defense to examine or test.

Allen said he believes that photographs taken at the scene show substantial, compelling evidence there was an electrical fire.

* * * * *

In the Roanoke arson case, shortly after 6 a.m. on Aug. 10, 1987, a fire broke out in the kitchen of Reedy's home occupied by Reedy, who was asleep on a couch, and his 4-year-old daughter and 2-year-old son.

Reedy said he discovered the fire and attempted to save his children, but was overcome by heat and smoke and jumped through a window to escape. He was hospitalized for smoke inhalation, burns and lacerations. His children died from smoke inhalation.

The Roanoke Fire Department concluded the fire was set by pouring and igniting gasoline on the back porch and kitchen areas. Reedy was convicted in 1988 of arson and two counts of homicide.

Reedy, who had a drinking problem at the time of the fire, has always maintained he was innocent. He first claimed his ex-wife may have set the fire and later, with the support of some experts, contended that it started accidently.

Over the years, his appeals and a petition for a writ of actual innocence have all failed.

A neighbor — who had had disagreements with Reedy — testified that Reedy threatened several times when drunk or distraught to burn his house down with his children and himself rather than allow his wife to win custody in a divorce battle.

Another neighbor testified Reedy told her his children were not in the burning house but conceded that when he said the word, "gone," he may have meant he thought they died.

In a 2004 affidavit by Dr. Daniel P. Harrington, the director of emergency psychiatry at the University of Virginia Hospital in 1987, said that when Reedy was taken to the burn unit, he was given heavy pain medication that inhibits the brain's "executive function."

"The patient tends to be loosely conversational, often 'incriminating' himself," said Harrington, and the records show it was clear Reed did not know his children were dead until the following day when a psychiatrist told him.

Key evidence against Reedy were tests at the state forensics laboratory that determined there was gasoline on Reedy's shirt and on a wood floor.

One of Reedy's lawyers, Thomas J. Bondurant Jr., a former federal prosecutor, contends the only evidence of arson were the test results that detected gasoline and the fire marshal's subjective interpretation of burn patterns.

Reedy has had a clemency petition before governors since 2003. In a 2008 letter to then-Gov. Timothy M. Kaine, his lawyers argued that "Reedy's entire conviction was premised upon the conclusion that there was an arson in the first place."

Bondurant said he does not believe authorities acted with any ill intent — they simply relied on what they believed to be legitimate science in 1988.

In 2006, Lentini studied the chemical analysis prepared by a state forensics expert and concluded the test results did not prove there was gasoline on the shirt or the floor.

"Even by the standards used in 1987, this should have made the identification of gasoline in these two samples suspect," Lentini wrote.

Reedy's lawyers say the fire marshal saw "pour patterns" and concluded the fire was fed by a flammable liquid. He sent samples to the state laboratory for testing, telling the lab the fire was arson and asking them to look for petroleum-based products.

Science now shows such burn patterns could have just as easily been the result of an accidental fire, experts say.

Ray Ferris, then an assistant commonwealth's attorney who helped prosecute Reedy, said that years ago he was shown information challenging the burn patterns and new science that was not the standard back when Reedy was prosecuted.

"I found that information to be interesting, I certainly found it would have made for good discussion at trial if we'd had that kind of information," said Ferris, now a criminal defense lawyer and a member of the Roanoke City Council.

However, he said it was only a one-sided presentation of the flaws in that science.

"I will tell you that I never had any doubt when I prosecuted the case, I have not had a doubt since prosecuting the case," he said. "We presented the commonwealth's evidence to the jury in a way that was fair to Davey Reedy."

Ferris said, "It was a circumstantial case, it was a very complicated case, but (with) the totality of the circumstance and the totality of the evidence, I drew the conclusion that we prosecuted the right guy."

He said that part of what convinced him of Reedy's guilt at the time was the science.

But, he said, "Just because the defendant produces an expert that says the fire marshal was wrong about the pour pattern and they produce an expert to say this wasn't really gasoline — I'm not convinced by that without hearing the other side of the story."

Bondurant, with Gentry Locke Rakes & Moore, argues that it is not a question of prosecuting the right guy. The question, said Bondurant, was whether there should have been an arson prosecution at all.

Reedy, sentenced to two life terms, was paroled in 2009. "The Virginia Parole Board, (which) paroles nobody, let him out," he said.

"He never showed remorse. He said he was innocent to the last day," and still the parole board released him, Bondurant said.

Bondurant wrote McDonnell again on Jan. 23 requesting a meeting about Reedy's case but has yet to receive a response.

"Everyone knows there's something wrong with the conviction, just nobody wants to do anything about it because there was two dead children involved," Bondurant contends. Among those with apparent concerns about the case is McDonnell.

In 1999, while a state delegate representing Virginia Beach, McDonnell wrote then-Gov. Jim Gilmore about Reedy's case telling him there was "a substantial amount of evidence to warrant further investigation."

"I would ask that you look over the material and consider appropriate action on behalf of Mr. Davey Reedy, if he has filed a request for a pardon or other intervention from your office," McDonnell asked Gilmore.

fgreen@timesdispatch.com (804) 649-6340

Truth in Justice