WASHINGTON — In early 2002, federal agents who were hunting the anthrax killer were trying to winnow a suspect list that numbered in the hundreds. They knew only that they were looking for someone with access to the rare Ames strain of anthrax used in research labs around the world. Profilers said the perpetrator probably was an American with "an agenda."
The powder-laced letters, which killed five people, contained no fingerprints, hair or human DNA but they did offer one solid microscopic clue: The lethal spores in the powder were dotted with genetically distinct variants known as morphs.
So agents set out on an arduous task: Collect samples from Ames anthrax cultures around the world, sort through them and find one with morphs that matched the attack powder. Then they'd have a line on where the murder weapon was made and, perhaps, the identity of the killer.
Bruce Ivins, an Army scientist at Fort Detrick, Md., had a good idea where the inquiry was headed. In the months after the attacks, he'd schooled federal agents in the intricacies of anthrax, explaining how the telltale morphs can arise from one generation to the next.
In April 2002, Ivins did something that investigators would highlight years later as a pillar in the capital murder case that was being prepared against him before he committed suicide in 2008: He turned over a set of samples from his flask of Ames anthrax that tested negative, showing no morphs. Later, investigators would take their own samples from the flask and find four morphs that matched those in the powder.
Rachel Lieber, the lead prosecutor in a case that will never go to trial, thinks that Ivins manipulated his sample to cover his tracks. "If you send something that is supposed to be from the murder weapon, but you send something that doesn't match, that's the ultimate act of deception. That's why it's so important," Lieber said.
However, a re-examination of the anthrax investigation by "Frontline," McClatchy and ProPublica turned up new evidence that challenges the FBI's narrative of Ivins as a man with a guilty conscience who was desperately trying to avoid being discovered.
Records recently released under the Freedom of Information Act show that Ivins made available a total of four sets of samples from 2002 to 2004, double the number the FBI has disclosed. And in subsequent FBI tests, three of the four sets ultimately tested positive for the morphs.
Paul Kemp, Ivins' lawyer, said the existence of Ivins' additional submissions was significant because it discredits an important aspect of the FBI's case against his client. "I wish I'd known that at the time," he said.
THE SEARCH FOR SAMPLES
To understand how investigators eventually came to see almost everything Ivins did or said as proof of his guilt, you have to return to the fall of 2001.
The FBI wasn't equipped to handle deadly germs, so the attack powder was rushed to Fort Detrick, the home of the U.S. Army Medical Research Institute of Infectious Diseases.
From the beginning, Fort Detrick researchers played a prominent role in the inquiry. Ivins was among the most voluble, offering advice and a steady stream of tips about co-workers, foreign powers and former employees who might have carried out the attacks.
Investigators quickly recognized they were in an awkward situation. Any of the scientists could be the killer. Agents canvassed the tight-knit laboratory, inviting the researchers to finger their colleagues. "We were heroes in the morning and suspects in the afternoon," recalled Jeffrey Adamovicz, at the time the deputy chief of the Bacteriology Division, where Ivins worked.
At 8:45 a.m. on Dec. 16, 2001, Ivins typed out an email to colleagues offering to provide Ames strain "for genetic analysis or sequencing by whomever." He offered a sample of the original Ames anthrax taken in 1981 from a Texas cow and a collection of spores sent to Fort Detrick in 1997, mostly from the U.S. Army base in Dugway, Utah. Seven years later, prosecutors announced that they were certain the attack powder had been grown with germs from the Dugway flask Ivins was offering for scrutiny.
John Ezzell, a USAMRIID scientist at the time who assisted the FBI, said in an interview that it was likely that Ivins didn't think the technology could distinguish among Ames variants. But the record suggests otherwise.
On Jan. 23, 2002, Ivins gave an FBI agent a detailed tutorial on how to spot morphs in anthrax colonies. He also volunteered the names of two people who had the "knowledge and character" to have prepared and sent the letters while explaining that he'd never worked with powdered anthrax.
Ivins then showed the agent photos of anthrax morphs, and told them that "DNA sequencing should show the differences in genetics," the mutations that make morphs grow differently. Ivins had good reason to understand the biology of morphs: One of his best friends at the lab, Patricia Worsham, had published a pioneering paper on the subject several years earlier.
In suggesting that the FBI use morphs to catch a killer, Ivins was proposing some cutting-edge science. No one had ever attempted to genetically fingerprint morphs, and the researchers advising the FBI weren't even sure it could be done. Such genetic detective work, now commonplace, was in its infancy. Today, this technique is recognized as a precise, but not foolproof, method of identification.
A few weeks later, Ivins gave several people the sort of evidence he seemed to be suggesting they collect. He provided a sample to a colleague who wanted to look at the spores under a microscope. Then, on Feb. 27, Ivins drew anthrax from his flask, which he labeled RMR-1029, and provided it to investigators who were assembling the FBI's worldwide library of anthrax. If prosecutors are right, the murderer had handed over his gun for testing.
But then the narrative took a strange twist. Perhaps deliberately, perhaps by chance, Ivins placed the spores in the wrong type of glass vessel. Investigators rejected the sample and told him to try again.
Some time in the next few weeks, prosecutors contend, Ivins figured out for the first time that the morphs might trap him. Until then, they assert, he'd assumed that the anthrax in his flask was pure and therefore without morphs. But Paul Keim, the scientist who helped the FBI identify the attack strain, said it seemed implausible that Ivins thought his spores were morph-free. The Dugway culture included dozens of separate batches, most of which were grown at the Utah Army base in fermenters, an ideal breeding ground for morphs. Ivins, Keim said, was likely to have understood this.
In April 2002, Ivins prepared a third sample from RMR-1029. This time, his lawyer says, he plucked a sample using a technique called a "single colony pick," a method biologists use to maintain purity when growing bacteria. Ultimately, this sample tested negative for the morphs. Prosecutors say they're not even sure that the sample Ivins submitted came from the flask. If it did, they say, he obstructed justice, since their subpoena instructed scientists to capture diverse samples of spores that would be sure to reproduce any morphs. Ivins told investigators he'd followed standard procedures for microbiologists when he sampled just one colony. Investigators eventually seized and tested the germs Ivins turned over to his colleague for microscopic examination and found they tested positive for the morphs. Separately, they stumbled across a duplicate first submission from February: the material that had been rejected. It, too, was positive.
A CURIOUS CLEANUP
In late April 2002, investigators confronted Ivins about reports that he'd been furtively testing for anthrax spores in his office and other areas outside the "hot suites," the sealed rooms where researchers worked with deadly pathogens.
Ivins said that that was true and volunteered that he'd also conducted cleanups in the lab not once, but twice — in December 2001, when he bleached over areas he'd found to be contaminated, and again in mid-April, when he conducted a search for errant anthrax spores.
These acts violated the lab's standard procedure, which called for the safety office to investigate and clean up any contamination.
Ivins offered curious explanations. He said that in December, he was trying to address the worries of a junior technician that sloppy handling of the attack powder had spread deadly spores through the lab. In April, against the advice of his supervisor, he launched his own tests after two researchers accidentally spilled a small amount of anthrax in the hot suite.
Asked why he didn't inform safety officials of the possible dangers, Ivins told investigators he didn't want to disrupt the FBI inquiry or alarm colleagues.
Whatever his motivation, subsequent tests showed Ivins had a point about failures to contain anthrax in the labs. His sampling showed that tiny amounts of anthrax of various strains had somehow seeped out of the hot suite and into his office. The Army ordered an investigation into why the lab's safety procedures weren’t followed.
Despite Ivins' puzzling behavior, investigators wouldn't focus on him for years.
The anthrax inquiry was following another course, and had zeroed in on a virologist named Steven Hatfill. A blunt character who boasted of his years in Rhodesia, Hatfill had a penchant for publicity, holes in his resume and an unpublished novel that featured a Palestinian terrorist who attacks Washington with the bubonic plague. The evidence against him was entirely circumstantial.
Investigators remained on Hatfill's trail until late 2006. By then, they'd searched his home, deployed anthrax-sniffing dogs and even emptied a pond, from which they removed a suspicious contraption. It was a turtle trap. No evidence of anthrax turned up. (Hatfill sued the government and received $5.8 million to settle the case).
A RAID ON FORT DETRICK
In early April 2004, Ivins was asked to help the FBI collect a complete set of cultures from Fort Detrick. Earlier, FBI agents had found 22 vials of anthrax that hadn't been turned over. On April 6, a lab assistant found a test tube of material that appeared to have been removed from Ivins' flask.
The assistant gave the germs to Henry Heine, a colleague of Ivins' who happened to be in the building. Heine said he checked with Ivins, who told him to send a sample from the tube to the FBI. In an April 6 email, Ivins thanked Heine, acknowledging that the anthrax "was probably RMR-1029."
Heine views this moment as a sign of his colleague's innocence, pointing out that Ivins willingly turned over a sample he thought had originated from his flask. In an interview, Heine said there were no cameras in the building, that FBI agents weren't monitoring the search and that Ivins easily could have prepared the sample himself and tampered with the evidence.
A day later, investigators seized Ivins' flask, locking it in a safe double-sealed with evidence tape.
What happened next raises questions about the reliability of the FBI's method for detecting morphs. The bureau separately ordered tests on Heine's sample and a second one drawn from the same test tube. Records show conflicting results, one negative and one positive.
Does this mean the FBI's tests for morphs is unreliable?
An FBI scientist said Ivins had told investigators the anthrax in the refrigerator had been diluted. This perhaps made the morphs undetectable in testing, said the scientist, who was made available to discuss the matter on the condition of anonymity.
Heine said the sample he sent wasn't diluted.
"We can only go by what Bruce told us," the FBI scientist replied.
Heine said he sent the FBI at least two additional samples from RMR-1029 that Ivins had shared with him. He said the FBI later told him both had tested negative for the morphs. The FBI scientist said the bureau could find no record of this.
THE FEDS CLOSE IN
In September of 2006, the FBI assigned Edward Montooth to lead the anthrax inquiry. Montooth looked at the evidence through fresh eyes, and his attention quickly focused on the background and conduct of Ivins. By December, he told FBI director Robert Mueller that Ivins had emerged as the prime suspect.
Investigators saw mounting evidence that the Fort Detrick scientist was hiding something.
They learned about his lifelong obsession with the Kappa Kappa Gamma sorority and discovered that the Princeton, N.J., mailbox into which the letters were dropped was just yards from a KKG office. He mailed packages under assumed names, and his emails expressed fears that he was paranoid, delusional or suffering from a split personality. "When I get all steamed up, I don't pout. I push Bruce aside, then I'm free to run about!" one read.
The genetic evidence seemed persuasive. Investigators had tested 1,059 Ames samples from U.S. and foreign labs and found only 10 with three or more of the morphs that genetically matched the letter powder. All traced back to Ivins' flask.
The FBI identified 419 people at Fort Detrick and other labs who could have had access to the material. Prosecutors say each was investigated and cleared of possible involvement.
All of Ivins' actions in the early days of the investigation were reinterpreted as signs of his guilt. Investigators recovered a portion of his first sample from RMR-1029 — the test tubes that had been rejected because they were the wrong kind — and found that they contained the incriminating morphs. They contrasted that with the second sample from 2002 — no morphs — and saw it as proof that Ivins had learned before submitting it that the Ames strain could be traced to his flask.
Asked about the April 2004 sample turned up by "Frontline," McClatchy and ProPublica that tested positive, prosecutor Lieber said it could be easily explained. Ivins had no choice: FBI agents were swarming through Fort Detrick looking for missing anthrax.
The unauthorized cleanup of the lab, the Justice Department said in its report last year, reflected a "guilty conscience."
"The evidence suggests that Dr. Ivins knew where to swab because he knew where he had contaminated the building," prosecutors wrote.
Even Ivins' defenders had questioned his actions at the time. "I said, 'Bruce, do you realize how bad this looks?' And he was a little bit puzzled," Adamovicz said. "I said, 'This makes you look suspicious because it looks like you're trying to hide something. Bruce of course denied that he was trying to hide anything. And again in my view I don't think he was trying to hide anything. I think he couldn't keep a secret if he had to."
The Army's report on the incident didn't portray it as nefarious. It confirmed the presence of Ames anthrax and two other anthrax strains in Ivins' office. Ames also was found near a "pass box" through which Ivins had transferred one of the letters to the hot suite, the men's changing area and Ivins' office. But investigators were unable to nail down an original source for the contamination. They determined that it was unrelated to the minor spill in the hot suite and speculated that it could have resulted from the handling of the anthrax letters.
Investigators executed a search warrant at Ivins' home and office. They found guns, a shooting range in his basement and Tasers. But swab after swab taken from every conceivable nook and cranny found not a single spore from the attack powder. Lieber said that was to be expected with a microbiologist trained to handle dangerous germs.
Claire Fraser-Liggett, a key genetics consultant for investigators, found such a dismissal troubling. "You think about all the efforts that had to go into decontaminating postal facilities, and the volatility of those spores and the fact that they were around for so long," she said. "I think it represents a big hole, really gives me pause to think: How strong was this case against Dr. Ivins?"
Ivins sat down for detailed interviews with prosecutors in early 2008 and volunteered a series of damaging admissions with his lawyer present. He acknowledged making long drives at night while his wife slept and calmly recounted his obsession with Kappa Kappa Gamma, blindfolding and bondage. Sometimes his answers were incoherent, FBI summaries show. He couldn't explain, for example, why he spent so many late nights in the lab in the weeks before the letter attacks.
By the summer of 2008, Ivins was coming apart. He told his group therapy session, which he'd begun attending in recent months, that he was planning to get a gun so he could kill his enemies. The FBI searched his home again and seized several guns, bulletproof vests and 250 rounds of ammunition.
Ivins was briefly committed and then released. On July 26, he took an overdose of over-the-counter medication. Three days later, he was dead at 62. Neither he nor the prosecutors would ever have their day in court.
A week later, Justice Department officials called a news conference to describe their evidence against Ivins as some in Congress called for an independent investigation of the case.
"We believed that based on the evidence we had collected, we could prove his guilt to a jury beyond a reasonable doubt," U.S. Attorney Jeffrey Taylor told reporters at the time. "Based on the totality of the evidence we had gathered against him, we are confident that Dr. Ivins was the only person responsible for these attacks."
Prosecutors and investigators patiently laid out the case against Ivins. At its heart, they said, was the revolutionary science that had improbably traced the attack powder to a single flask.
"RMR-1029 was conclusively identified as the parent material to the anthrax powder used in the mailings," the Justice Department wrote in its summary of the case.
Fraser-Liggett, who did some of the pioneering genetics work for the investigation, remains agnostic.
"This was not an airtight case, by any means. You know, I think that, for an awful lot of people, there is a desire to really want to say that 'Yes, Ivins was the perpetrator. This case can reasonably be closed. And we can put this tragic chapter in U.S. history behind us,' " she said. "But I think part of what's driving that is the fact that, if he wasn't the perpetrator, then it means that person is still out there."
(McClatchy collaborated with the investigative newsroom ProPublica and PBS' "Frontline" to produce this series. Gordon works for McClatchy. Engelberg works for ProPublica and Wiser and Gilmore are with "Frontline.")