A decade ago, as Earl Washington Jr. neared his execution date, a leading DNA expert first suggested an analyst in the vaunted Virginia state crime lab might have erred in the case.
The lab's director, Paul Ferrara, rejected the criticism as unfounded.In April, when a second expert hired by Washington's lawyers questioned another round of tests, Ferrara dismissed him as a "hired gun" and rebuffed calls for an outside review.
Several months later, three other experts--this time not paid by the defense --reached the same conclusion. The lab's analyst, they said, had misinterpreted the evidence, but Ferrara again balked at an outside review.
"I'm not going to admit error when there is none," Ferrara said in a recent interview at the highly regarded Richmond facility, the first state lab to build a database linking evidence from unsolved crimes to suspects through their genetic profiles.
Within days of that statement, the lab experienced another first. On Sept. 30, the governor of Virginia ordered an audit of the lab's work on the Washington case.
That it took a governor's edict to force one of the nation's most-respected labs to allow such a review illustrates the broader problems undermining confidence in the nation's crime labs.
Revelations of shoddy work and poorly run facilities have shaken the criminal justice system like never before, raising doubts about the reputation of labs as unbiased advocates for scientific truth.
The far-reaching crime lab scandals roiling the courts are unlike other flaws in the criminal justice system--the rogue prosecutor, the incompetent defense attorney, the unscrupulous cop--because for years the reputation of the labs had been unquestioned.
But the consequence of lab errors, whether due to incompetence, imprecision or fraud, is frequently the same--an innocent person behind bars.
A Tribune examination of the 200 DNA and Death Row exoneration cases since 1986--including scores of interviews and a review of court transcripts and appellate opinions--found that more than a quarter involved faulty crime lab work or testimony.
In recent years, evidence of problems ranging from negligence to outright deception has been uncovered at crime labs in at least 17 states. Among the failures were faulty blood analysis, fingerprinting errors, flawed hair comparisons and the contamination of evidence used in DNA testing.
Scandal also has hit the FBI crime lab, long considered the nation's top forensic facility.
In the mid-1990s, a lab whistle-blower touched off a broad inquiry over allegations of improper handling of evidence. It led to the firing of several lab officials and the overhaul of protocols and procedures.
In May of this year, an FBI analyst, Jacqueline Blake, pleaded guilty to a misdemeanor charge of making false statements about following protocol in some 100 DNA analysis reports.
Though the FBI said its review found no wrongful convictions resulting from her work, the Justice Department's inspector general concluded that the lab's failure to detect her misconduct "has damaged intangibly the credibility of the FBI laboratory."
Blake was dismissed from the lab and last month was placed on 2 years of probation.
Veteran lab directors around the country contend the exposure of such scandals is evidence that labs are policing themselves.
In most cases, however, lab problems have come to light only after defendants have challenged their convictions.
"Virtually every major lab scandal has been broken by a post-conviction DNA exoneration," said Barry Scheck, a founder of the Innocence Project, a non-profit legal clinic that has helped exonerate dozens of inmates.
Given the sheer volume of cases that labs handle, the discovery of even a single flawed analysis raises the prospect of re-examining hundreds, if not thousands, of cases.
In many jurisdictions, the task of re-evaluating that many cases is so daunting that authorities have declined to conduct broad audits, despite evidence that analysts have committed errors or engaged in fraudulent practices.
One of their well-placed fears: that uncovering additional problems in a lab would spawn lawsuits or unravel an untold number of convictions.
Two of the nation's highest profile crime-lab scandals--involving analysts Fred Zain in West Virginia and Joyce Gilchrist in Oklahoma--resulted in the exonerations of at least 10 defendants, millions of dollars in settlements and broad reviews of hundreds of their cases. Both were accused of falsifying test results and giving false testimony; both denied any wrongdoing.
Earlier this year, in response to the DNA exoneration of a man who served 13 years in prison for rape, the city of Cleveland appointed an independent special master to review all the casework and, if necessary, retest the evidence handled by one analyst, and conduct a random audit of others in the lab.
But such a response has been uncommon.
In Texas, Gov. Rick Perry has rejected a plea from Houston's police chief to halt executions of inmates convicted in Harris County until the scope of problems at the police crime lab can be determined.
Two inmates from Harris County have been executed in recent weeks since Chief Harold Hurtt announced the discovery of 280 boxes of evidence from at least 8,000 Houston cases spanning 25 years. The boxes contained everything from clothing and weapons to a fetus.
Even before the latest crisis, revelations of incompetent analysts in the police lab's DNA section forced Houston authorities to shutter it. The new questions cover everything from firearm identification to blood typing in a jurisdiction that has sent 75 people to the death chamber, more than most states.
Earlier this year, Boston police admitted that two fingerprint examiners had linked Stephan Cowans to the 1997 shooting of a police sergeant, even though a later review found that the comparison, in the words of Massachusetts Atty. Gen. Thomas Reilly, "wasn't even close."
An outside consultant then conducted a broader examination of thousands of prints in the Boston police fingerprint unit, "and the only error he found was in the Cowans case," said police spokeswoman Beverly Ford. Last week, the department turned over all fingerprint examinations to the state police until it can get its own lab accredited.
In Montana, the state Supreme Court narrowly voted last month to dismiss a petition seeking an independent audit and retesting of evidence in hundreds of cases handled by a former state crime lab examiner whose erroneous hair-comparison testimony contributed to three wrongful convictions.
Among other revelations, the scandals have exposed the lack of independent oversight and the often-ineffective standards governing the labs that analyze forensic evidence.
Lab directors contend that the American Society of Crime Laboratory Directors' accreditation board, which will review Virginia's work in the Washington case, are sticklers for quality and accuracy. But even some tough-on-crime politicians question the effectiveness of the board's review teams.
"Everyone boasts that their labs are certified by them," said James Durkin, a former Cook County prosecutor and former Republican state representative. "I believe they are more of a fraternal organization than an authoritative scientific body."
Reputation under fire
Walking through the Virginia state crime lab in Richmond--the centerpiece of the state's four-lab system--Paul Ferrara is visibly proud of its accomplishments.
It has long been a leader in DNA testing. In June, the lab, which is not affiliated with a law-enforcement agency, passed a crime-solving benchmark: its 2,000th "cold hit," linking unsolved crimes to suspects through its DNA database. Last month, a newspaper clip touting the achievement still sat on a table in Ferrara's office.
But the Washington saga and criticism of the lab's testing in other cases have threatened to undermine the lab's reputation.
"They're very protective of the reputation that they're the gold standard for crime labs," said Betty Layne DesPortes, a Richmond defense attorney and jurisprudence section chair of the American Academy of Forensic Sciences.
Of the 200 DNA and Death Row exonerations in the U.S. since the mid-1980s, few have loomed larger or have been as controversial as the case of Earl Washington Jr.
It began in 1982, when Rebecca Williams was found stabbed 38 times in her Culpeper, Va., apartment. Authorities claimed Washington confessed, though his attorneys said police took advantage of the mildly mentally retarded farmhand to get a false confession.
At trial, Washington maintained his innocence, but he was convicted and sentenced to death.
Doubts about his guilt persisted. In 1993, as Washington faced execution and as DNA testing began to gain wider acceptance in forensic labs, then-Gov. Douglas Wilder ordered testing to help resolve the question.
The work was done by Jeff Ban, a top DNA analyst in the lab and now a member of a panel of scientists that sets national DNA standards. Ban performed tests on a vaginal swab taken during the autopsy on Williams and on a semen stain on a blue blanket recovered from her apartment.
Ban reported that Washington was excluded as the source of the semen stain on the blanket but said the result of his analysis on the vaginal swab was less clear. He said that while he had found a genetic trait that did not belong to Washington, the victim or the victim's husband, he could not eliminate Washington as a potential source of semen.
On Jan. 14, 1994, nine days before the state was scheduled to execute Washington, Wilder commuted Washington's sentence to life in prison, saying the test results raised a "substantial question" about Washington's guilt.
Wilder, who was leaving office, declined to pardon Washington altogether because the DNA test that failed to exclude him did not erase all doubt about his involvement in the crime.
Henry Erlich, a developer of the DNA test that Ban used, then examined Ban's work at the request of Washington's defense lawyers. He said that Ban had misinterpreted the results--and that Washington should have been excluded.
Six years later, prompted by the still-lingering doubts over Washington's guilt and continued advances in DNA testing, then-Gov. James Gilmore ordered the lab to conduct another series of tests on the evidence.
Again, Ban did the work. Again, his tests sparked debate.
According to Ban's report, the tests revealed a genetic profile on the blanket that did not belong to Washington.
The unknown profile was plugged into the state's DNA database. It linked to Kenneth M. Tinsley, who previously had been convicted of rape in Illinois and whose genetic profile had been added to the Virginia database after he was convicted of a rape there and sentenced to life in prison, according to authorities.
Ban further reported that he was unable to obtain a genetic profile from a slide made from the vaginal swab--although at Washington's trial, a medical examiner had testified there was an abundance of sperm on the slide.
Even more puzzling were the results of his tests on a second, similar slide. Not only did Ban exclude Washington, but he excluded Tinsley and, according to his report, turned up two additional unidentified genetic profiles.
The exclusion of Washington was enough for Gilmore to grant him a pardon--just as Ban's earlier test was enough to prompt Wilder to commute his death sentence. After 17 years in prison, more than nine of them on Death Row, he was set free.
That did not settle the matter, though.
Tinsley's DNA was detected by the lab on the blanket. But because Ban said he did not find it on the slides, authorities did not prosecute Tinsley, leaving the case open. The test results prompted some police officers to continue saying they believed Washington was involved.
Duplicate slides were sent to Dr. Ed Blake, a DNA expert, who was working for Washington's attorneys. His tests isolated only Tinsley's genetic profile, he said, and conclusively eliminated Washington.
In his report, Blake called Ban's work "logically incoherent. ... The result is so nonsensical and so covered with red flags that it should never be published."
Blake's assessment gave new ammunition for Washington's lawyers, including Peter Neufeld of the Innocence Project, who are suing police in federal court. Neufeld called for an audit of Ban's work and, more broadly, all of the lab's DNA work.
Ferrara stood behind Ban, dismissing Blake because he was hired by the defense and because he believed Blake's analysis had no merit.
Three other DNA experts, at the request of a Virginia newspaper, reviewed Ban's reports. They all agreed that his work was troubling and warranted further scrutiny.
Again, Neufeld called for an independent audit of Ban's work.
Ferrara again refused.
In an interview with the Tribune, Ferrara said it is possible for two scientists to come up with different test results because no two samples are alike--although Ban and Blake tested slides created from the same swab.
"As far as we're concerned, there is no error at all except in the minds of [critics]...," Ferrara said. "When you're on the top of the heap you are going to have someone trying to knock you down."
The accreditation board requires that when a lab identifies an analytic error, it must do a review and recommend corrective action, with the possibility of a broader review of either a single analyst or a particular section.
But the requirement is triggered only when a lab admits error.
And Ferrara said in the interview that he had no intention of doing so. Soon after, though, Gov. Mark Warner stepped in and ordered him to seek a review.
"The governor believes that an outside investigation will help maintain the lab's reputation," a Warner spokeswoman said, "and help maintain confidence in our criminal justice system."
Any crime lab scandal raises serious concerns. But nowhere are the stakes higher than in Harris County, Texas, because of how many people it sends to the state's busy execution chamber.
So the recent news that police had found 280 boxes of evidence in a police property room shook the legal community in Houston, the county seat. Defense attorneys want to know what evidence is in the boxes and if it can help their clients on appeal.
In belatedly announcing the discovery, Hurtt, Houston's new police chief, urged Gov. Perry to halt all executions of inmates from Harris County until the evidence could be reviewed. The chief plans to ask the Houston City Council to appoint a special investigator to conduct a review.
"We don't believe in our heart that this is going to jeopardize any case. We are hopeful, anyway," said Hurtt's spokesman, Alvin Wright. "But we want to make sure justice is served correctly."
Beyond the lost evidence and faulty DNA analysis, the Houston lab has seen its firearm analysis come under scrutiny.
In the case of Nanon Williams, convicted of a 1992 drug-related murder, a Houston firearms examiner testified the victim was shot with a .25-caliber bullet.
That was the same caliber of Williams' gun.
Six years later, the same firearms examiner reviewed the case again and determined that the bullet was a .22-caliber from a co-defendant's gun. The firearms examiner acknowledged that he had never tested that weapon.
The same firearm examiner testified against Johnnie Bernal, who was sentenced to death for a 1994 murder. At trial, the examiner said the bullet in the murder came from a gun that police seized from Bernal when they arrested him.
But the examiner conceded he cleaned the gun barrel and fired it 25 times before he could declare a match. Firearms examiners typically fire a weapon two or three times before they declare matches.
The lawyer for Dominique Green is citing concerns about testing at the lab in seeking a delay in his execution, which is scheduled for next week.
Green was convicted for his role in the October 1992 murder of a Houston man during an armed robbery. Two other men implicated in the crime testified against Green and received sentences of 10 and 30 years.
Prosecutors charged that Green was the gunman, saying a semiautomatic handgun believed to have been used in the murder was found in a car Green had been riding in. A veteran Houston firearms examiner, Ray Klein, was the prosecution's last witness and provided key testimony.
Klein testified that he test-fired the gun, then examined the fired bullet under a microscope to tell if the unique markings left on the bullet as it traveled through the barrel matched those from the bullet taken from the victim.
Green, while admitting he took part in the robbery, denies firing the fatal shots, said his lawyer, David Dow, a law professor and the director of the Texas Innocence Network. Dow said a review of Klein's comparison could settle the question of whether the gun found in the car was used in the murder.
"Our concern is that it's the same lab. It's the same technique. And it's proved to be unsound in other Harris County death penalty cases," Dow said.
Roe Wilson, an assistant Harris County district attorney, said she was confident in the lab's work and dismissed the defense's attempt to delay the execution. "They're basically trying to use the problems that have arisen in the Houston crime lab," Wilson said. "And those problems don't apply in this case."
A legacy of questions
It is unlikely that anyone had more influence on the Montana state crime laboratory than founder Arnold Melnikoff. The question is whether that influence has tainted the lab's work.
Director of the lab for nearly two decades, Melnikoff left in 1989 to become an analyst for the Washington state police crime lab. But 15 years after his departure, questions have been raised not only about his work on specific cases, but also about his legacy as a lab director.
Three Montana men he helped convict of rape have been exonerated by DNA, and Melnikoff's testimony in those cases has been discredited. His new employer, the Washington state lab, fired him in March after an audit of his toxicology work there criticized his lab procedures and accused him of inflating his testimony to help prosecutors.
The question looming over the Montana crime lab was framed this year in a petition filed with that state's Supreme Court seeking a broad review of the more than 200 cases Melnikoff had handled.
The petition, filed by the Innocence Project and joined by five former Montana Supreme Court justices, urged the review not only because of Melnikoff's errors, but also because of his influence on the lab.
"If `juicing' the testimony, offering unprofessional statements, and making scientifically unsupportable claims was his gold standard, we must infer that this is the standard of practice that he conveyed to his employees," the petition asserted. "Many of these staff now hold supervisor positions at the lab."
The lab's current administrators say Melnikoff trained none of their present lab employees. Whether he nonetheless shaped the culture of the lab may remain unknown: The petition was dismissed days after it was filed. The Supreme Court said it did not have jurisdiction to order the review.
With a bachelor's degree in biology from Northern Illinois University and a master's degree in chemistry from the University of Montana, Melnikoff established the state's first crime laboratory, in Missoula, in 1970.
He began by handling chemistry and toxicology tests and gradually expanded into other disciplines such as blood typing and hair and fiber analysis. He analyzed samples of suspected drugs and substances in suspected arson cases, according to the petition.
He guided the lab through the first two decades after it was started in a building that once housed a vaudeville theater.
Over the years, Melnikoff developed a unique manner of testifying about hair evidence, using a statistical analysis he said was based on his own examinations. He said he had analyzed hair in 500 to 700 cases and found matches between unrelated people only a few times.
Based on that analysis, Melnikoff testified that the odds of a person other than a particular defendant having the same hair were 1 in 100. In some cases, he extended his analysis further--using odds of 1 in 10,000.
Melnikoff used his statistical estimation in the prosecution of Chester Bauer in 1983, Jimmy Ray Bromgard in 1987 and Paul Kordonowy in 1989. All were convicted. All have since been exonerated.
In the wake of the exonerations, a panel of forensic scientists was organized by the Innocence Project; they unanimously criticized Melnikoff's statistical analyses as scientifically unfounded and wrong.
Through his lawyer, Melnikoff declined to comment for this article. He is appealing his dismissal.
After Kordonowy was exonerated last year, the Innocence Project sought an independent audit and retesting of every hair case in which Melnikoff was involved.
In response, Montana Atty. Gen. Mike McGrath ordered a review of 270 prosecutions in which Melnikoff testified. Officials read through trial transcripts and case files, but McGrath declined to order retesting of any evidence or reopen any cases. The state crime lab's advisory board concurred.
Challenging those decisions, the project and former judges filed their petition with the Supreme Court in August, seeking an independent examination and retesting of those cases.
Bill Unger, director of the lab, said in an interview that lab officials did not urge McGrath to retest the evidence in part because the cases were so old and in part because there is a backlog of pending cases waiting to be analyzed.
"Timeliness was part of the reason," he said. "My biggest nightmare is workload. If the legislature would give us a couple of positions, it would be fine."
Though Unger said no current lab employees were trained by Melnikoff, the petition alleged that the problems go beyond him and involve other lab analysts who "have followed Melnikoff's lead."
Specifically, the petition focused on Julie Long, an analyst at the Montana lab since 1980.
In a 2002 case, the lab was asked to analyze the underwear of a Bozeman woman who said she had been sexually assaulted. Lab personnel reported that "no indications of seminal fluid were detected."
But when defense attorneys hired Blake to examine the evidence, he found semen and sperm. Blake called the crime lab's failure to find the genetic evidence "mind-boggling."
Long, who was the lab's head of quality assurance and control, said the evidence was missed because no microscopic examination was conducted--a practice that has long been standard at most labs in the country.
"It wasn't in our protocol," she said. "We have since changed that."
The defendant, Joshua Stephen Warren, was acquitted after more than 500 days in jail.
`We're the white hats'
Many crime lab officials contend the scandals that have unfolded around the nation demonstrate how tough they check themselves.
"We hold ourselves to a much higher standard. ...," said Don Plautz, who worked for 24 years as a lab director and in other posts with the Illinois State Police crime labs before retiring in 2002. "The bulk of problems I've dealt with in forensic science have been identified by the fellow forensic scientists."
But many crime labs are fiercely resistant to letting outsiders check to make sure that is true.
Janine Arvizu knows her way around a lab. She once ran the Navy's program to audit the commercial and government laboratories that the Navy uses, and she managed an analytical lab for the Department of Energy in Idaho.
When she started her own consulting firm and began working with defense attorneys in recent years to check the quality of crime labs, she was stunned.
"Their attitude, you don't encounter elsewhere: We work for the good guys. We're the white hats," Arvizu said. "They're very uncomfortable with this idea of independent oversight, which is a fundamental precept of laboratory quality assurance."
She has audited written records and other data from dozens of forensic labs across the country. Only one of those labs allowed her inside its doors--and only because a judge ordered the lab to let her in.
A fundamental principle of quality assurance is that the higher the stakes--in health and safety--the more checks and controls a lab should have. So Arvizu puzzled for a long time about why the standards seemed so relatively low for crime labs, given that people's liberty is at stake.
She concluded there is little natural constituency, at least not outside the walls of American prisons.
"There's no upswelling of people who feel they're at risk from failures by crime labs," Arvizu said. "It will take the son of a federal judge to be wrongfully convicted on the basis of flawed forensics to make the kind of quantum improvement in forensic quality standards that needs to happen."
Not all crime labs have been defiant in the face of questions about their work. One possible model for dealing with lab scandal is developing in Cleveland.
In 2001, Michael Green was released after spending 13 years in prison for a rape that DNA tests proved he did not commit. Another man has since confessed to the crime.
At issue in Green's prosecution were blood-typing tests conducted by Cleveland police lab technician Joseph Serowik. This year, Cleveland officials acknowledged that Serowik's lab work was flawed.
In settling a wrongful-conviction lawsuit filed by Green, the city agreed to pay him $1.6 million. More important, perhaps, officials agreed to the appointment of independent experts to review more than 100 of Serowik's cases dating to 1987.
"If there are other Michael Greens out there, we want to know who they are," said Subodh Chandra, a former federal prosecutor who is the city's law director.
"As a public official, you can view this simply as an issue of financial liability and trying to avoid embarrassment," Chandra added, "or you can view it as trying to improve the criminal justice process."
Scheck, who also is president of the National Association of Criminal Defense Lawyers, sees progress in Cleveland's response and in recent legislation passed by Congress.
One measure, the Paul Coverdell Forensic Science Improvement Grant Program, would legally obligate crime labs seeking the grants to have an independent auditor in place to conduct "investigations into allegations of serious negligence or misconduct substantially affecting the integrity" of lab analysis. Grant amounts for each state could range from tens of thousands of dollars to $800,000.
The provision would mean some labs that have not done audits after allegations of errors would be barred from receiving the money. "Certainly that's our intent," said Sen. Jeff Sessions (R-Ala.), chief sponsor of the Coverdell Act. "If you get this money, you do have to have an accountability process."
In Baltimore County, for instance, officials pledged a review of analyst Concepcion Bacasnot's cases after DNA tests in 2003 exonerated an inmate whom her work helped convict of rape. In another case, she said she did not understand the science involved in her work. She later resigned saying it was for personal reasons.
Baltimore officials identified nearly 500 cases that Bacasnot was involved in. But saying they didn't have the money or staffing for it, they decided against a review, leaving it up to defendants and their lawyers to investigate cases themselves.
Most of the attention and tens of millions of federal dollars, though, have gone toward easing the backlogs of unanalyzed DNA samples that have overwhelmed many crime labs. Last week, Gov. Rod Blagojevich announced Illinois' share of that money, more than $3.2 million.
Earlier this month, when Congress passed the Innocence Protection Act, it included a provision allowing states to spend funds on forensic analysis other than DNA testing--but only when there is no DNA backlog.
"Fingerprints, fiber tests, ballistic tests on bullets and cartridges, chemical and drug analysis and others are far more frequent and often just as important to justice in a particular case as DNA," said Sessions, who argued for the amendment.
"Clearly," he added, "we need to be working hard to improve our entire forensic science system, not just one small aspect of it."