The wildfire community has closely followed the case of Van Bateman, a Forest Service fire management officer and one-time incident commander of an elite national incident management team. Prosecutors charged that in 2004, Bateman set two unauthorized fires on the Coconino National Forest south of Flagstaff, Ariz., to burn brush and give his crew a bit of work. The fires barely burned 22 acres, but they publicly exposed a long-standing and more or less accepted brush-clearing practice.

"I'm not lily-white on this," said Bateman, who admitted to setting the fires when confronted with global positioning data from his cell phone. He claimed that his actions were sanctioned by tradition, and more than 50 fellow firefighters sent letters to support him. "I'm saying ... I was doing my job. Did I obtain the proper authorization? I did not. But I wasn't trying to start an arson fire. I was trying to clean this piece of country up. I would be shocked if there's anybody who has spent their career in forest management who hasn't done this."

Eventually, the arson charges against Bateman were dropped, but a federal judge sentenced him to 24 months behind bars for starting the technically illegal blaze.

The prosecution of Oyler was a far more complex and challenging effort. Not only had no one ever been convicted of murder for setting a wildfire, the case against Oyler hung by a thread: The notion that a series of fires started with stick matches, cigarettes and rubber bands was the work of a single person. It helped that Oyler's DNA was present on ignition devices for two of the non-lethal fires, that all the fires were close together and near Oyler's home, and that the series evolved from small and ineffective blazes to larger and more destructive ones, one sign of a serial arsonist.

When the prosecution and defense finally rested, the jurors filed into their deliberation room, and a sheriff's deputy locked the door behind them. The room was small and sparsely furnished. A table with chairs occupied most of the space and whiteboards hung on a wall. For nearly two months, jurors had been forbidden to talk with family and friends or even each other about the disturbing narrative that had unfolded in the courtroom. Now they were free to speak and argue over interpreting the evidence.

The jurors introduced themselves, selected a foreman and took a quick vote to see how matters stood: Nine hands went up for guilty, three for not guilty. The foreman, Don Estep, cast one of the not-guilty votes, he said later, to make sure the process continued to the discussion stage. But he soon joined the majority.

Two women, identified here as Stephanie and Amanda (not their real names), could not accept the prosecution's theory that one man had set all the fires. Three other women, by chance seated at the opposite end of the table, confronted them day after day, using photographs, documents and maps from the trial. And day after day, the vote split 10 to 2.

Stephanie "had had a hard life and saw Raymond Oyler as an underdog, and she was an underdog too," says Janis McManigal, an events coordinator, who was one of the three who confronted Stephanie and Amanda. But Amanda, she says, "had trouble seeing the pattern, the connections" between the fires and the similarities of the ignition devices. She repeatedly asked, "How could one person have started all these fires?"

Exhibits studded with yellow memo notes went up on the walls. Scribbles and connecting arrows criss-crossed the dry-erase boards. "Everyone was congenial, it was a nice bunch of people, though I had to cool it down once or twice," Estep recalls. "There was a lot of venom in that room," says McManigal. Finally, in their fifth day of dispute, the jurors separated the 23 fires into groups, based on the weight of the evidence for each. Stephanie and Amanda agreed that Oyler was guilty of starting some fires but refused to go along on three relatively inconsequential ones, for which there was virtually no evidence.