The four locked steel boxes, each about the size of a coffin, were unusual enough to attract the attention of authorities as they were being shipped through Singapore in 2010. X-rays of the boxes revealed they contained elephant tusks, tightly packed and wrapped in fabric bags. Interpol was notified, and the boxes, still locked and unopened, were shipped to the forensic laboratory Interpol uses for such cases—the U.S. Fish and Wildlife Service Forensics Laboratory in Ashland, Oregon.
When the boxes were opened at the lab, which describes itself as “the only lab in the world dedicated to crimes against wildlife,” investigators found more than a thousand pounds of tusks. They began their standard forensic investigation, including DNA analysis, but the question of where the tusks originated remained. Clearly they were from African elephants, but traditional forensics wasn’t pointing to a specific area in Africa where the poachers and smugglers were at work.
Moving Beyond Traditional Forensics
Aware of these limits, the investigators called in David and Paul Stoney, two long-time forensic researchers who run Stoney Forensic in Chantilly, Virginia. They have long been advocates of an investigatory technique called “very small particle (VSP) analysis,” in which careful examination of dust on an object can help trace it back to its original location. Their approach also allows authorities to compare two or more objects to determine if they have, at some point, been in the same place.
The brothers have a successful history of using such analysis to assist investigators working for U.S. government agencies, local police departments and, as in the elephant tusk case, international authorities.
“We got out to the fish and wildlife lab and helped them unpack [some of the tusks],” Paul Stoney said. “We took dust samples and then we got out of the way and let them continue their work. Our processes don’t interfere with anything else that could go on forensically.” The Stoneys took particle samples from the fabric bags and from the bottom of the shipping containers.
Their first level of analysis looked at the largest of the particles, mostly plant tissue, insect parts, tusk fragments and pieces of cardboard. Then the fine particles were analyzed, including DNA sequencing of botanical particles. By classifying the particles by type, including geological, ecological and “human activity signals,” the Stoneys were able to first eliminate 91 percent of the area of Africa, or 36 countries, as a source of the tusk shipment. Of the remaining 12 countries, their analysis eliminated 72 percent of the area, allowing Interpol investigators to focus on specific portions.
They then proposed another level of dust analysis that would have allowed them to further narrow possibilities for the likely source of the shipment. Based on the progress of their investigation, Interpol investigators chose not to do that level of analysis, but if they had, Paul Stoney explained they could have reduced the number of possible sites to the point where each remaining site could be sampled and compared to the particles on the tusks.
The process could lead to the one site that exactly matched,” he said.
The brothers have been developing their understanding of how populations of very small particles (VSP) can be used in forensic investigations for decades. In a 2012 study, one of several of their VSP projects supported by National Institute of Justice awards, they note, “very small particles are ubiquitous in our environment and are virtually ignored by forensic science. These particles range in size from an order of magnitude smaller than conventional trace evidence, down to the molecular level.”
“We move about in a soup that is a combination of VSP that provides an extraordinary, largely untapped resource for forensic associations and source attribution,” they conclude.
A hint of how they developed a detailed database that allows them to match a population of particles to a specific location can be found in a glass bookcase at one end of their office. More than a decade ago they did extensive work in the Middle East, and their bookcase shelves are lined with books with titles such as: Wildflowers of Saudi Arabia; Trees of Pakistan; and Flora of Syria, Palestine, and Sinai. There is an 11-volume set about the flora of Turkey, and a Ph.D. thesis on the differentiation of flowers in Turkey.
“This is what remains of our references,” Paul Stoney said as he motioned toward the shelves of books. “It’s probably a $200,000 collection, but it’s only valuable for what we were doing. We read and screened and had a botanist go through the material. These are the books that we individually picked out because they contain the best, most definitive information, and these books are our datasets.”
Similar data about the flora and geology of Africa helped narrow down the likely sources of the dust from the tusks. The dust tells a story, and the Stoney brothers looked at the mix of mineral, botanical, zoological, microbial and anthropogenic particles and compared that mix to their datasets.
Constantly expanding the dataset eventually allows for the creation of a profile of the dust that can almost literally take investigators to the scene of a crime,” Paul Stoney said. “One key in analyzing particles is knowing what question you’re trying to answer and adjusting the particle analysis to get the answer."
Using some of his Middle East experience as an example, he noted that if we want to ask if you were in Pakistan this morning, and did your plane come from Pakistan, that’s one type of question. “But if I know a man came from Pakistan and I want to find the room where he is holding the hostage, that’s another type of question. There is enough information in those particles to be able to answer all of those questions, but knowing how much information I need, and don’t need, to answer the question depends on the problem.”
David Stoney believes matching the depth of particle analysis to the problem can be difficult for analysts in forensic laboratories who are not accustomed to doing particle combination analysis. He said, “If we start looking at a lot of different kinds of particles, not just a handful, and we leverage the part of modern chemistry that lets things be done very quickly on very small amounts, then we can analyze these small particles that are present in every single case, not just hair or fiber or paint cases.”
“The trick,” he said, “is thinking about the entire population of particles on a piece of evidence, something that isn’t typically done in forensic labs. What I’m talking about does not mean you necessarily analyze the particles the way a chemist would. If you give a chemist one type of particle and say, ‘I want to know exactly what this is,’ the chemist would do that particle by particle in a way that is very expensive.”
“We are saying, let’s look at thousands and thousands of particles and let’s leverage the techniques that can go very quickly, and let’s leverage the computational methods, the statistics that can look at lots and lots of particles. In that way we can make it apply to more cases, and we can make it cost effective.”
Potential Uses of VSP Analysis
Apply this methodology—known as particle combination analysis—to traditional crime scenes, and you can significantly increase the value of trace evidence in criminal investigations. In a “what if” paper published by Stoney Forensic, the researchers put forward scenarios to demonstrate the value of VSP analysis.
In the first scenario, a vehicle containing illegal immigrants, weapons and drugs is stopped near the U.S. border with Mexico. A wipe of dust from clothing, weapons, and drug packaging is done. When the particle profile is compared against a database of similar profiles developed in previous arrests and property seizures, border patrol officials can determine if an established smuggling route was used, or if a new route, perhaps even a tunnel, was used. If it was a tunnel, the particles could reveal its location by reflecting the immediate environment (rural, agricultural, developed), and nearby landmarks, such as specific types of vegetation, rock outcroppings, and commercial activity.
In a second scenario, a female victim of human trafficking escapes and shows up at a clinic in Baltimore. She doesn’t speak English, but dust from her personal belongings link to other individuals and cases, pointing to possible suspects. Dust from her clothing “provides information about where she has recently been held, narrowing the search to less than 10 percent of the surrounding area.”
The 2012 study, “Probative Value of Small Particles on Fibers,” conducted in partnership with South Dakota State University statistician Cedric Neumann, focused on a common type of trace evidence—carpet fibers. But unlike traditional forensic investigations, it did not focus on the fibers themselves. Instead, the researchers used a scanning electron microscope with X-ray microanalysis (SEM/EDS) to analyze the hundreds to thousands of particles found on each fiber. In the paper, they describe their work as “a fundamentally different approach to trace evidence analysis.”
Working with carpet fibers gathered from 90 residences by crime scene technicians in nine jurisdictions across the United States, the Stoneys showed that the particles could be collected, characterized and used to “associate” two or more items.
Their current research, also supported by an NIJ award, broadens the study of VSP from carpet fibers to particle profiles found on handguns, cell phones, drug packaging and ski masks. The items, evidence from closed cases, were collected from the San Diego County Sheriff’s Office Regional Crime Laboratory and are being analyzed with the SEM/EDS system. The objective is to see if the particle combination analysis works on different evidence types in different circumstances, and if it provides an investigative tool that is currently absent in forensic labs.
As the use of trace evidence wanes in courtrooms due to the Daubert standard, which is used by judges in federal and many state courts to assess whether a scientific expert’s evidence is truly scientific, David Stoney believes VSP analysis will allow a resurgence in trace evidence admissibility. “VSP analysis,” he said, “responds to the demands of more critical Daubert-type considerations because we can actually compute numbers and do the statistics that will show a particular measured value to the associations we get.”
“It used to be that, ‘I believe,’ was good enough for the criminal justice system when expert witnesses testified,” he said.
But given the challenges to such testimony in the wake of the 1993 Daubert ruling and the 2009 National Academy of Sciences critical assessment of scientific standards in forensics, judges now insist that there be valid scientific data to support expert testimony. “This is a way to take the traditional kind of trace evidence and look at it a different way,” David Stoney said of the VSP work. “We can actually give a measurement and test the validity of the data.”
About the Author
Jim Dawson is a science writer with Leidos, a federal contractor, on assignment at the National Institute of Justice, U.S. Department of Justice.