Posted Sep 02, 2004 04:55 pm CDT
There was no other physical evidence tying Bowling to the crimes—no DNA, no fingerprints, no blood, no hair or fibers, no gunshot residue.
In fact, the only direct evidence of Bowling’s guilt was the FBI examiner’s testimony about the matching bullets and the word of a police informant who testified that Bowling had confessed to the two murders while they were housed together in the same jail.
Bowling contends that he never even spoke to the informant, who had struck a deal with prosecutors to trade testimony in Bowling’s state trial in exchange for a favorable disposition of the federal mail fraud charges then pending against him.
And now the FBI examiner’s testimony in the case is being called into question by a recent study that raises serious doubts about the validity of the agency’s so-called comparative bullet-lead-analysis evidence.
The study, released in February by the National Academies’ National Research Council, found that while the FBI’s scientific method for comparing bullets was generally sound, its examiners have sometimes overstated its importance in court and played down the likelihood of a false match.
There is no disputing the fact that trace amounts of certain elements in bullet lead can be precisely measured. The controversy centers on how the FBI has interpreted that data.
For decades, FBI examiners operated on the twin assumptions that every batch of bullet lead was compositionally uniform throughout and that no two batches were compositionally alike. So if two bullets were found to have the same concentrations of the same elements, the reasoning went, those bullets must have come from the same batch of lead.
In the past few years, however, critics have begun to question those assumptions. Some researchers have found that different samples from the same batch of bullet lead can have different elemental compositions. Other research has confirmed that different batches of bullet lead can have the same elemental composition. Still other research has shown that bullets from the same box of ammunition can have different elemental compositions, some of which can be attributed to bullets from different melts and some of which can be attributed to intramelt variability.
“This is not a case in which the critics of a forensic practice are merely pointing to a lack of research data to validate an assumption,” says University of California at Davis law school professor Edward J. Imwinkelried, who co-wrote a critical law review article about bullet lead evidence last year. “This is an even more troublesome situation in which there are available data at odds with both of those assumptions.”
Retired FBI metallurgist William A. Tobin, a longtime critic of the bureau’s bullet matching technique, and Imwinkelried’s co-author on the law review article, says he feels vindicated by the committee’s findings.
“They validated almost every single criticism I have of the practice,” he says.
In its report, the council recommended a number of changes in the FBI’s procedures. They include improved training and oversight, new statistical methods for analyzing bullet lead data, and a limit on how far experts should go when they testify about the possibility of a match between a crime scene bullet and a bullet traceable to a suspect.
The FBI, which has been using bullet comparison evidence since the 1960s, had asked the council for advice on how to analyze bullet lead and present the findings in court in a sound, scientific manner. To conduct the study, the NRC put together a committee of 14 experts in chemistry, metallurgy, statistics, forensic science and the law.
Kenneth O. MacFadden, the independent consultant who chaired the committee, says this finding is a key point of the whole study: Just because two bullets match does not necessarily mean they came from the same box of ammunition.
MacFadden says comparative bullet lead analysis, known as CBLA, can be a useful forensic tool, but does not have the unique specificity of techniques such as DNA typing, which can be used as stand-alone evidence.
“[CBLA] can never be used for anything other than circumstantial evidence, and an entire case can’t be built around it,” he says.
Case Western Reserve University law professor Paul Giannelli, who served on the committee, likens the significance of CBLA evidence to the discovery of a Nike shoe print of a certain size found at a crime scene.
If a suspect were found with the same type and size of shoe, “It would be admissible and it would be relevant, but it wouldn’t be dispositive because there may be thousands, if not hundreds of thousands, of people out there with that same size and type of shoe,” he says.
Here’s how bullet lead analysis works: bullets found at crime scenes are tested for trace amounts of seven different elements, including arsenic, tin, copper and silver. Those findings are compared to a similar analysis of bullets traceable to a suspect. If the crime scene bullets and the suspect’s bullets are determined statistically to be analytically indistinguishable for each of those elements, FBI examiners have historically concluded that the two samples probably came from the same source. By the FBI’s own estimate, such evidence has been used in about 2,500 cases over the past three decades, including about 500 trials.
The technique has proved especially useful in shooting cases in which either no gun is recovered or a crime scene bullet is too small or too mangled to compare the marks left on it by its passage through the barrel with the marks made on a test bullet fired from a suspect’s gun.
Bowling was convicted and sentenced to death in 1992 for the murders of two gas station attendants during a pair of early-morning robberies in Laurel County, Ky., in early 1989. He was arrested after a shooting incident at a third gas station three days after the second murder.
At his trial, ballistics evidence linked a gun found by the side of a road on which Bowling had traveled to the bullets used in the first two shootings. Then, FBI lab examiner Donald Havekost took the stand. Havekost testified that he believed every bullet could be traced back to its own manufacturer and to the specific batch of lead from which it was made. That was possible, he said, because when bullets are made from molten batches of lead, there is a uniform concentration of lead along with the alloy elements in every batch.
Havekost also testified that the FBI maintained a database to keep track of the different chemical compositions of bullets it had tested over the years. And he said that the only time he ever found two matching bullets and checked them against the FBI database was when two crimes were related.
According to Havekost’s testimony, one of the bullets from the first murder and one of the bullets from the second murder had the same chemical composition, meaning they originated from the same manufacturer’s batch of bullet lead. Havekost also testified that three of the bullets that killed the second victim, three of the bullets found at the scene of the third shooting, and 16 of the 24 bullets found in a partially filled box of ammunition at Bowling’s home all matched. In addition, two of the bullets that killed the second victim also matched one of the bullets at the scene of the third shooting and three of the bullets in Bowling’s ammunition box, he said. And one of the bullets at the scene of the third shooting matched five of the bullets found in Bowling’s home.
The FBI examiner’s testimony made it sound as if each batch of bullet lead is unique, although it is not, according to assistant public advocate Susan Martin, one of Bowling’s appellate lawyers. Havekost’s testimony also made it seem as though each batch of lead produces a small enough number of bullets such that a match between two bullets sounds more significant than it actually is, she says.
“The potential for a jury to give undue emphasis to the significance of a match between a defendant’s bullets and bullets found in the victim is so great it ought to be excluded altogether,” Martin says. “It’s very disturbing to think this type of evidence is being used to convict people, and it’s most disturbing to think that my client is on death row because of it.”
Retired commonwealth attorney Thomas Handy, who prosecuted Bowling, says the bullet lead evidence was an important element of his case. He also says he has no reason to doubt the accuracy of the bullet lead testing done by the FBI.
But Handy says he thinks the strongest evidence against Bowling was the testimony of the third gas station attendant, who positively identified Bowling as the person who had initiated a violent assault on him that was consistent with the two murders.
“That alone wouldn’t have justified [Bowling] being convicted of the two murders, but when the evidence is taken as a whole, I think it was a very strong case beyond a reasonable doubt that he was the killer,” he says.
In its report, the committee found that while CBLA evidence has been admitted for more than 30 years, there have been fewer than two dozen published appellate opinions on the technique. The overwhelming majority of those have been homicide cases, at least some of which have been capital cases. And since there are few federal homicide statutes, the evidence has most commonly been used in state prosecutions.
All but one of the courts that have addressed CBLA admissibility have admitted it, often as only one piece of evidence in cases in which numerous other items of evidence have been introduced at trial, the committee determined. In a few cases, however, the courts have indicated that CBLA evidence played an important part in securing a conviction.
The published cases also reveal a wide variety of interpretative conclusions in trials in which CBLA evidence has been introduced, according to the committee. Experts in some cases have testified only that two bullets were “analytically indistinguishable,” although it wasn’t clear in those cases whether that was the only conclusion rendered at trial. In other cases, however, experts concluded that two samples could have come or did come from the same source or batch of lead.
The testimony of some examiners went even further, referring in a number of cases to a specific box of ammunition, which usually contains 50 but sometimes as few as 20 loaded cartridges. Sometimes, examiners testified only that two bullets could have come from the same box of ammunition. But other times, examiners testified that two bullets probably came from the same box or “must have come from the same box or from another box that would have been made by the same company on the same day.” One expert went so far as to testify that two bullet specimens “had been made by the same manufacturer on the same day and at the same hour.” Brown v. State, 601 P.2d 221 (Alaska 1979).
In recent years, the committee notes, the testimony of FBI examiners appears to have become more conservative, which coincides with a 2002 FBI publication suggesting that examiners should limit their conclusions to the effect that two bullets “likely originated from the same manufacturer’s source (melt) of lead.”
Still, the opinions in some cases indicate that certain prosecutors and judges don’t understand the limitations of such evidence. In United States v. Davis, for example, an FBI expert testified that the bullets recovered from the scene of two bank robberies were analytically indistinguishable from a box of bullets found in the defendant’s car, that such a finding is rare, and that “the bullets must have come from the same box or from another box that would have been made by the same company on the same day.”
But the St. Louis-based 8th U.S. Circuit Court of Appeals, in a 1996 ruling affirming the defendant’s conviction on charges of armed bank robbery and using a firearm during a crime of violence, wrote that the “evidence made it more probable than not that the expended bullets originated from the cartridge box found in the [defendant’s car].” 103 F.3d 660.
At least one federal court has recently excluded CBLA evidence on the ground that it doesn’t meet the standard of admissibility for scientific evidence set by the U.S. Supreme Court in Daubert v. Merrell Dow Pharmaceuticals Inc. 509 U.S. 579 (1993). In United States v. Mikos, the defendant in a murder case had moved to exclude an FBI expert’s testimony that bullets recovered from the victim’s body and bullets found in a box in the defendant’s car likely came from the same manufacturer’s batch or source of lead.
U.S. District Court Judge Ronald A. Guzman of Chicago noted that the expert had been allowed to testify about such evidence for years. But in a Dec. 9, 2003, opinion, he concluded that the expert could testify only that the composition of the bullets recovered from the body is indistinguishable from the composition of the bullets found in the defendant’s car. “There is no body of data to corroborate the government’s expert’s further opinion that from this finding it follows that the bullets must or even likely came from the same batch or melt,” he wrote. No. 02 CR 137 (N.D. Ill.).
Other courts are also beginning to restrict CBLA evidence. At the 2001 trial of Miguel Trujillo—who had been charged with the 1999 murder of his lawyer—Albuquerque, N.M., district court judge Neil Candelaria allowed an FBI examiner to testify that a bullet recovered from the victim’s body may have come from the same melt as bullets found in the defendant’s son’s room, but would not allow her to testify that they came from the same box. State v. Trujillo, No. D-0101-CR-2000-284.
In its report, the committee said that some valid conclusions can be drawn about bullet lead composition, but that any court testimony should be carefully qualified. For example, because of variations in the bullet manufacturing process, there is inadequate data to support statements that a crime scene bullet came from, or is likely to have come from, a particular box of ammunition or that it was manufactured on a given date. And because very limited information exists on where bullets are distributed, FBI examiners should not testify as to the probability that a crime scene bullet came from a defendant.
However, any attempt to predict the composition of bullets from the bullet manufacturing and packaging process over extended periods of time is unreliable and potentially misleading, the committee found, because the process varies from one manufacturer to the next, and even among production runs by the same manufacturer. For example, bullets may be boxed immediately or as customer orders come in. And bullets from different lead melts may be mixed before they are boxed and shipped.
Moreover, the composition of a lead melt may change because many manufacturers add scrap lead to the melt at random times in the production process. And information on where bullets are distributed once they leave the manufacturer is unavailable, making it impossible to determine the probability that analytically indistinguishable bullets are more likely to be found in one geographic area than another, the committee found.
The committee did conclude that there exists in the bullet manufacturing process a “compositionally indistinguishable volume of lead,” or a CIVL, which might produce as few as 12,000 and as many as 35 million .22 caliber bullets. The committee also found it was possible that bullets from different CIVLs may by coincidence be analytically indistinguishable. It recommended that FBI experts acknowledge these facts in their testimony and lab reports.
Legal experts say the findings are likely to result in a significant number of new legal challenges to the use of bullet lead evidence in the future.
“The committee’s report gives any defense attorney paying attention a blueprint for bringing a Daubert or a Frye challenge to the admissibility of such evidence,” says University of California Hastings College of the Law professor David Faigman, an expert on science and the law.
Faigman also says the courts are likely to follow the committee’s recommendations and limit FBI examiners’ testimony to the more modest claims it identified as being scientifically supportable.
But Faigman says he doubts the committee’s findings will result in the overturning of many convictions because a petitioner would have to show that the outcome of the case would have been different without the bullet lead evidence.
“That’s a very difficult standard to meet,” he says.
Other legal scholars say the report also could lead to new defense challenges involving other types of evidence in which experts attempt to turn what essentially are class characteristics into individual characteristics.
“People like me have been waiting a long time for things like tool marks, firearms and bite marks to be put under the microscope,” says James Starrs, a professor of law and forensic science at George Washington University.
Imwinkelried sees two possible benefits arising from the release of the committee’s report. In the short term, conscientious bullet lead examiners will probably be much more circumspect in the way they express their opinions on the witness stand. In the long term, the report should give bullet lead analysts the incentive to do the kind of rigorous research they should have been doing all along, he says.
Barry Scheck, co-director of the Innocence Project and president of the National Association of Criminal Defense Lawyers, recently commended the National Research Council study. But he also said that if the FBI’s written and unwritten protocols for CBLA had been subject to the same rigors to which the scientific community holds itself, any number of tainted convictions could have been avoided.
Scheck called for a review of every case in which bullet lead evidence has played a part. And because so much forensic science enters the courtroom before the scientific community at large has adequately assessed its validity, he also called for the creation of a national institute for forensic science. Its role, he said, would be similar to the part played by the National Institutes of Health in assessing the adequacy of medical treatments and drug efficacies.
“The FBI, which rightfully claims leadership in advancing new forensic technologies, is not equipped to validate itself, any more than any university or private research lab can validate its own research,” he said.
FBI officials could not be reached for comment. But in a prepared statement released at the time the NRC report was issued, the bureau noted that the committee had found that its current analytical instrumentation is appropriate and is the best available technology with respect to both precision and accuracy for the elements analyzed. It also acknowledged that the committee’s recommendations included suggestions to improve statistical analysis, quality control procedures and expert testimony regarding bullet lead evidence.
“The basis of bullet lead compositional analysis is supported by approximately 50 peer-reviewed articles found in scientific publications beginning in the early 1970s,” the statement said. “Published research and validation studies have continued to demonstrate the usefulness of the measurement of trace elements within bullet lead.”
The release of the committee’s report may have come in the nick of time for Bowling.
Bowling’s claim of ineffective assistance of counsel, based on the CBLA testimony, was raised and rejected in the Kentucky state courts in his post-conviction attack on his murder convictions and death sentence. The same claim was also raised in his 2003 habeas corpus petition in federal district court in London, Ky.
The district court dismissed the petition on procedural grounds but will let Bowling refile it after certain pending proceedings are completed in state court. He appealed that dismissal to the 6th U.S. Circuit Court of Appeals, based in Cincinnati, where his petition is now pending on procedural issues.
But when Bowling’s petition is returned to the district court for substantive review, his lawyers say they will be in a position to supplement his appeal with the findings from the committee’s report, which they intend to do. If he loses there, he will have one more appeal to the 6th Circuit. His last shot—and it’s a long one—would be a discretionary appeal to the U.S. Supreme Court.
Still, lawyers haven’t given up hope.
“The committee’s report does what hadn’t been done but should have been done [at the time of Bowling’s trial] in 1992, which is to show that the whole scientific premise on which the bullet lead evidence is based may be fatally flawed,” Martin says. It may be, and it may not.
What’s clear to Hastings professor Faigman is this: The committee didn’t conclude that the basic technology was either irrelevant or wholly unreliable. “It simply says that the claims for what the technology indicated was possible have been overstated.”