WASHINGTON -- One night in autumn 2001, as the U.S. reeled from the worst act of bioterrorism in its history, Bruce Ivins was alone in his cluttered Fort Detrick, Md., office, scrubbing phones, walls and furniture.

For colleagues, this was proof of the anthrax scientist's attention to safety. From a distance of seven years, it might be evidence of his guilt.

Reuters

Bruce Ivins, shown in a 2003 photo, took an overdose of painkillers and died in an apparent suicide last week. The FBI said the government scientist was close to being charged in 2001's deadly anthrax attacks.

Like the detective in Agatha Christie's play "Mousetrap" who turned out to be the murderer, Dr. Ivins played a haunting dual role in the anthrax mystery, federal law-enforcement agents say. He was part of the team that examined the poisoned letters. Investigators say he implicated other scientists and submitted incomplete samples to throw them off-track. The Pentagon even adopted tips he offered to guard against the dangers of experimenting with anthrax.

The scientist's daily presence at the alleged scene of the crime is fuel for the continuing debate as to whether the U.S. government has its man. Friends and colleagues say the conundrum makes it impossible to differentiate between clues left by Dr. Ivins the investigator and Dr. Ivins the suspect.

Dr. Ivins killed himself last week after being informed that the Federal Bureau of Investigation planned to charge him in the murders of five people who handled the anthrax-laced letters. His lawyer said the FBI's case, made public Wednesday, is flawed and that Dr. Ivins would have been cleared in court had he lived.

Dr. Ivins himself seemed to think he might be two people, a sign of mental imbalance he largely hid from colleagues, according to court documents. "Other times it's like I'm not only sitting at my desk, I'm also a few feet away watching me," he wrote in an email to an unidentified friend in April 2000.

Colleagues held a memorial for him Wednesday at Fort Detrick. More than 200 colleagues packed its chapel while the FBI, 50 miles south, unveiled its case against him. Col. John P. Skvorak, commander of the army institute, saluted Dr. Ivins's "openness, his candor, his humor and his honesty," according to Dr. Ivins's lawyer, Paul Kemp, who wrote the remarks down.

In the days after attacks, suspected samples of anthrax began pouring into the U.S. Army Medical Institute of Infectious Diseases. In Room 19 of the bacteriology division, Dr. Ivins's office was jammed with desks for six more staffers where once there had been just Dr. Ivins and perhaps one other.

"We were scared," says Col. Arthur Anderson, who at the time was chief of laboratory medicine in the diagnostic-systems division. Many of the scientists were seeing for the first time the weaponized version of a microbe they had worked with for years.

They were also fascinated, particularly Dr. Ivins, who talked repeatedly about the refinement of the spores sent with the Oct. 9 letter to then-South Dakota Sen. Tom Daschle. The spores nearly floated out of the envelope.

"The stuff just came out without any prompting," he told Gerald Andrews, his boss at the time.

Dr. Ivins, his colleagues said, argued that al Qaeda was responsible. "He was very passionate about this," former boss Jeffrey Adamovicz said. "He was very agitated." In these conversations, Dr. Ivins dwelled at one point on a purported link between Florida victim Robert Stevens, a photographer for American Media, and an apartment rented to 9/11 ringleader Mohammad Atta, Dr. Adamovicz said. (The FBI discounts that as an explanation.)

To his colleagues, this kind of talk wasn't unusual. Dr. Ivins was expressing a theory commonly held in the lab and he often got emotional in the pursuit of scientific debate.

That winter, the FBI asked Dr. Ivins to take his first and only lie-detector test, according to a law-enforcement official. The polygraph was part of the bureau's vetting of investigators. The FBI hasn't released the results. Dr. Ivins retained his role in the investigation.

In April 2002, researchers noticed an anthrax-laced deposit on the outside of a flask outside the biocontainment area. The contamination spawned an investigation and a 361-page Army report, during which Dr. Ivins admitted his unauthorized office cleanup.

He told Army investigators that he had cleaned his office the previous fall, and then again without permission in April, because "I had no desire to cry 'Wolf!' " and blame someone else for the spill. The Army cleared him and adopted his recommendations to improve "cleaning inside the suites and maybe surveillance."

By this time, all of the scientists in the bacteriology division were under the FBI's investigative microscope, people working there at the time said. One after another, they submitted to a 3½-hour polygraph test. Dr. Ivins "was in the safety zone" because he had already passed his polygraph, Dr. Andrews said. Dr. Ivins was never tested again, a law-enforcement official said.

A siege mentality began to build in the division, and Dr. Ivins shared in his colleagues' resentment. "Ever since they started calling back the people in the bacteriology division for polygraphs, they all sort of got quiet about things," Col. Anderson said. "Bruce wasn't the only person who stopped talking."

Meanwhile, Dr. Ivins developed hypotheses about other suspects. In a search of his house, the FBI found an email in which Dr. Ivins names two fellow scientists, providing 11 reasons for their possible guilt. He sent the email from a personal account to his Army account. It isn't clear whether Dr. Ivins was puzzling over the case or whether he was plotting to divert suspicion.

He grew more critical of the investigation when it began pointing toward him in late 2006. In lunchtime conversations, he told colleagues he thought the FBI might be trying to set him up, Dr. Adamovicz recalled. Dr. Ivins began poking holes in efforts to link him to the attacks, saying a positive DNA match between the anthrax in the letters and that stored at Fort Detrick would mean little "because those labs are shared," Dr. Adamovicz said.

This spring, Dr. Ivins told a colleague, W. Russell Byrne, that he was planning to retire in September. Dr. Byrne asked Dr. Ivins to let him know about the retirement party.

Dr. Ivins responded: "I don't know if I'm going to have one," Dr. Byrne recalled.