Another arson conviction challenged

Scientists questioning investigative practices that were used for years

By Jack Nicas
Globe Correspondent / September 8, 2010

E-mail this article

Invalid E-mail address
Invalid E-mail address

Sending your article

Your article has been sent.

Text size +

In early 2001, James Hebshie’s Taunton convenience store was struggling. He had tried selling it without luck. Then, on a Saturday in April, seven minutes after the security system recorded him leaving, the building broke out in flames.

Five years later, Hebshie was convicted of arson. Prosecutors pointed to his potential motives and testimony from state investigators who said the fire began in his store.

Today, one of the nation’s leading fire scientists is questioning the verdict, the latest in a string of Massachusetts arson convictions now under scrutiny.

“This fire did not start where the government’s witnesses said it started,’’ John Lentini said in an affidavit filed as part of Hebshie’s motion to set aside the conviction, which awaits a decision in US District Court. “The methodology used to determine the [fire’s] origin was outdated.’’

Lentini, who has investigated more than 2,000 fires and become a national voice against poor fire investigative practices, said state investigators misinterpreted evidence that showed the fire probably started in the basement. “The government’s hypothesis of the way the fire spread violates the laws of physics,’’ he said.

For the past decade, fire scientists have been testifying in courts across the country, questioning old arson convictions with new science. The movement has launched a high-stakes battle, pitting scientists against prosecutors, who dispute the need for new trials.

This summer the debate hit Massachusetts.

In the past several months, three of the nation’s top fire scientists have questioned at least three Massachusetts arson convictions, including the 1983 life sentence for Victor Rosario of Lowell, the subject of a Globe investigation in June.

The scientists — fire engineers who scientifically test and study the outcomes of fires — say that years ago investigators regularly pointed to certain burn patterns as irrefutable evidence of arson. But scientific tests have since shown that those same patterns also appear in accidental fires.

Those misconceptions plagued most fire investigations before the field’s first scientific manual was published in 1991 and continued until the widespread adoption of the science a decade later, the scientists said.

Over the same period, the number of structure fires ruled arson in Massachusetts steadily declined.

From 1984 to 2001, the number of structure fires ruled arson fell more than 70 percent, while total structure fires remained largely stable.

In 1984, roughly 10,600 structures caught fire and investigators ruled 2,133 arson; in 2001, there were about 10,200 structure fires, of which 618 were ruled arson.

(A state fire marshal’s office spokeswoman cautioned against comparing post-2001 statistics, which show a similar decline, because of a change in reporting methods.)

State Fire Marshal Stephen D. Coan attributed the decrease in arsons to more fire education, visibility of law enforcement, and successful prosecutions. Coan said the data simply shows a job well done.

“It sends a strong message: Massachusetts is not the place you want to engage in a crime of arson,’’ he said last month in an interview. “We’re always open for ways to improve, but I think we really set the benchmark for the way fire investigations should and can be conducted.’’

But the scientists said the decrease shows that as investigators used more science, they ruled fewer fires arson.

“There were a lot of accidental fires determined to be arson that weren’t,’’ Lentini said. “I don’t know any other way to interpret this dramatic of a decline.’’

It was a case out of Texas that turned the national spotlight on the flawed practices of old arson investigations.

Cameron Todd Willingham was executed in 2006 for setting his home on fire in Texas, killing his three daughters. Two years later, after an independent panel of five fire scientists concluded the investigation and many others of that era were based on “a collection of myths,’’ Texas’s new Forensic Science Commission tapped Dr. Craig Beyler to reexamine the case.

In his report, Beyler, who also reviewed the Rosario case for the Globe, called Texas investigators’ testimony “beyond belief in the context of fire investigation as an applied science.’’

Barry Scheck, cofounder of the Innocence Project, which nominated the Willingham case for review, said states have a moral imperative to scrutinize convictions that took place before the adoption of the national investigative standards.

“Once a jurisdiction recognizes that [the national standards are] the accepted scientific approach to arson investigations and further recognizes that [they weren’t] being used, and instead hellacious conventions were . . . they have an obligation to look back at old cases,’’ Scheck told the Globe.

Arizona, Nebraska, and Oklahoma recently passed resolutions urging the review of questionable arson convictions, but no such legislation has been filed in Massachusetts. Coan said reviews are unnecessary.

“There would be no reason to go back . . . and look at old cases in the 1980s,’’ Coan said. “That’s just not something we’re engaging in.’’

Rosario was convicted of setting a 1982 fire that killed eight people in Lowell. At the time, a state investigator said the fire pattern “was indicative of a flammable, burning fluid flowing on the floor — and burning intensely, especially [in] little crevices, the cracks in the floorboards,’’ he said. “[T]he burning of the threshold and the bottom edge of the door had to be caused by a flowing, flammable liquid.’’

But the national manual asserts, “Burning between seams or cracks of floorboards or around door thresholds, sills, and baseboards may or may not indicate the presence of an ignitable liquid.’’

A Lowell investigator repeatedly testified “alligatoring,’’ charring that resembles alligator skin, was a definitive sign of arson. But the manual discredits the theory: “The presence of large shiny blisters (alligator char) is not evidence that a liquid accelerant was present during the fire.’’

Coan and two state investigators declined to comment on the Rosario case pending an expected appeal. They said they were confident in procedure post-1991, and have no reason to believe state investigators used flawed practices before then.

The other two Massachusetts cases stem from fires that occurred years after misconceptions were widely rooted out. Fire scientists who reviewed the recent cases said investigators made a more common present-day mistake: misidentifying the point of origin.

In the aftermath of the Taunton fire, Trooper David Domingos, an investigator out of Coan’s office, concluded the fire started in Hebshie’s store. But this year, Lentini concluded Domingos got it wrong — his point of origin did not jibe with evidence of flames raging behind walls on the other end of the room.

Two years later, Domingos concluded a fire in Carver began at a pile of lumber in the basement. The homeowner’s then-boyfriend, Kenneth Davis Jr. of Rockland, was later convicted of arson; he was released from prison last month.

But in a recent review of the case, Dr. Gerald Hurst, a fire science pioneer, and two other fire scientists concluded Domingos and other investigators not only destroyed the fire scene by removing materials, they also disregarded a “beaded’’ wire — a broken wire common in accidental fires — hanging above the determined point of origin.

Sergeant Paul Horgan, a state fire investigator, has defended his colleagues, saying the campaign to retry arson convictions has gone too far.

“I think some of these cases they’re digging up from the past, they’re looking for whatever they can to dredge these cases up without really having the evidence,’’ he said. “It’s getting out of control.’’

But Hurst said in the scores of cases he has examined, the lack of legitimate evidence is often on the prosecution’s side, as it is in the Carver fire.

“I don’t see the evidence here, quite frankly,’’ Hurst said, discussing Davis’s conviction. “I would call this an undetermined fire, and you don’t prosecute for an undetermined fire. You obviously might set an arsonist free by not doing it, but the flip side of that is not tolerable.’’

Jack Nicas can be reached at