Why we cannot rely on firearm forensic "science"

Tyrone Jones is serving a life sentence, in part because of a microscopic particle that Baltimore police found on his left hand. At his trial for murder in 1998 the crime-lab examiner gave evidence that the particle was residue from a gunshot. He claimed Jones must have held or fired a gun shortly before his arrest. Jones denies this and still protests his innocence. His defence team is appealing the conviction, claiming that the science of gunshot residue (GSR) analysis is not as robust as the prosecution claims.

Now, a New Scientist investigation has found that someone who has never fired a gun could be contaminated by someone who has, and that different criminal investigators use contradictory standards. What's more, particles that are supposedly unique to GSR can be produced in other ways. Forensic scientists often testify that finding certain particle types means the suspect handled or fired a weapon. Janine Arvizu, an independent lab auditor based in New Mexico, reviewed the Baltimore county police department's procedures relating to GSR. Her report concludes: "The BCPD lab routinely reported that gunshot residue collected from a subject's hands 'most probably' arose from proximity to a discharging firearm, despite the fact that comparable levels of gunshot residue were detected in the laboratory's contamination studies." The BCPD did not return calls requesting comment.

Some specialists argue for a more cautious approach. "None of what we do can establish if anybody discharged a firearm," says Ronald Singer, former president of the American Academy of Forensic Sciences and chief criminalist at the Tarrant county medical examiner's office in Fort Worth, Texas. Peter De Forest of John Jay College of Criminal Justice in New York goes further. "I don't think it's a very valuable technique to begin with. It's great chemistry. It's great microscopy. The question is, how did [the particle] get there?"

GSR analysis is commonly used by forensic scientists around the world. In Baltimore alone, it has been used in almost 1000 cases over the past decade. It is based on identifying combinations of heavy metals in microscopic particles that are formed when the primer in a cartridge ignites. The primer sets off the main charge, which expels the bullet. There is no standardised procedure to test for GSR, but the organisation ASTM International, which develops standards that laboratories can look to for guidance, has developed a guide for performing the technique that was approved in 2001. This states that particles made only of lead, barium and antimony, or of antimony and barium are "unique" to gunshot residue. The particles are identified using a scanning electron microscope and their composition analysed using energy-dispersive spectrometry.

But recent studies have shown that a non-shooter can become contaminated without going near a firearm. Lubor Fojt sek and Tom s Kmjec at the Institute of Criminalistics in Prague, Czech Republic, fired test shots in a closed room and attempted to recover particles 2 metres away from the shooter. They detected "unique" particles up to 8 minutes after a shot was fired, suggesting that someone entering the scene after a shooting could have more particles on them than a shooter who runs away immediately (Forensic Science International, vol 153, p 132).

A separate study reported in 2000 by Debra Kowal and Steven Dowell at the Los Angeles county coroner's department reported that it was also possible to be contaminated by police vehicles. Of 50 samples from the back seats of patrol cars, they found 45 contained particles "consistent" with GSR and four had "highly specific" GSR particles. What's more, they showed that "highly specific" particles could be transferred from the hands of someone who had fired a gun to someone who had not. This doesn't surprise Arvizu. "If I was going to go out and look for gunshot residue, police stations are the places I'd look," she says.

Scientists using the technique are aware of the potential contamination problem, but how they deal with it varies. In Baltimore, for example, the police department crime lab's protocol calls for at least one lead-barium-antimony particle and a few "consistent" particles to be found to call the sample positive for GSR. The FBI is more cautious. Its protocol states: "Because the possibility of secondary transfer exists, at least three unique particles must be detected...in order to report the subject/object/surface 'as having been in an environment of gunshot primer residue'." So a person could be named as a potential shooter in Baltimore, but given the benefit of the doubt by the FBI.

Even worse, it is possible to pick up a so-called "unique" particle from an entirely different source. Industrial tools and fireworks are both capable of producing particles with a similar composition to GSR. And several studies have suggested that car mechanics are particularly at risk of being falsely accused, because some brake linings contain heavy metals and can form GSR-like particles at the temperatures reached during braking. In one recent study, Bruno Cardinetti and colleagues at the Scientific Investigation Unit of the Carabinieri (the Italian police force) in Rome found that composition alone was not enough to tell true GSR particles from particles formed in brake linings (Forensic Science International, vol 143, p 1).

At an FBI symposium last June, GSR experts discussed ways to improve and standardise the tests. The bureau would not discuss the meeting, but special agent Ann Todd says the FBI's laboratory is preparing a paper for publication that "will make recommendations to the scientific community regarding accepting, conducting and interpreting GSR exams".

Singer maintains that the technique is useful if used carefully. "I think it's important as part of the investigative phase," he says, though not necessarily to be presented in court. But he adds: "There are people who are going to be a bit more, shall we say, enthusiastic. That's where you're going to run into trouble."

(From issue 2527 of New Scientist magazine, 23 November 2005, page 6)



(And don't forget your ration of Wicked Thoughts for today)