On a clear, cool Monday morning in the fall of 1998, a worker mowing beneath a highway billboard in Orange County, N.C., spotted a pair of white sneakers. Hesitating, he moved closer. Then he saw the bones.
It didn't take long for detectives to swarm the area, cordoning off the badly decomposed body. What they had found upset them all: The remains belonged to a child. An autopsy concluded the remains were from a boy between 10 and 12 years old, but there were no reports of a missing child of that age in the area.
Twelve years later, the police still hadn't identified the boy, so Tim Horne, the detective in charge of the cold case and a father himself, approached anthropologist Ann Ross at North Carolina State University to see if there were any new methods to help identify him. Ross agreed and suggested they try an unorthodox approach — determining the child's ancestry.
Anthropologists study the ancestral origins of skeletons to track the movement and relationships among populations of people. Forensics experts use ancestry to help classify unidentified remains. Yet both groups typically refrain from analyzing the ancestry of the remains of children. And for a good reason — there has been no good way to do so.
Calipers, the Swiss Army knife of forensics, are a simple measurement device used since the time of the Greeks and Romans. They are great for measuring linear distances on a skeleton, but are not good for comparing adult skulls to children skulls: The size differences between small children's skulls and full-grown adult skulls are non-comparable. Researchers typically do not try to identify the ancestry of child skeletons since they can't compare them to adult skeletons.
In 2005, Ross teamed up with Dennis Slice, a professor of scientific computing at Florida State University, on a grant from the National Institute of Justice to develop a way for anthropologists and medical examiners to use "geometric morphometics" to determine the sex and ancestry of unidentified remains.
Geometric morphometrics is the three-dimensional analysis of the shape of a surface. Using this mathematical technique, size can be removed from the equation, focusing instead on shape. Ross recently published results using the technique on 24 skulls at the University of Pennsylvania Museum of Archaeology and Anthropology. She demonstrated — for the first time — that there are no significant shape differences between teen and young adult crania, suggesting that the adult shape of a face is attained years earlier than researchers previously thought. This new mathematical technique, therefore, can identify the ancestry of younger skulls.
In January, Ross and Slice debuted 3D-ID, a simple Java program that allows anyone with a computer and a digitizer to determine the ancestry of a skull. It is freely available at 3D-ID.org. "All you have to do it collect the data, plug it in, and it spits an answer out at you," Ross says.
The program is designed to be widely applicable for medical examiners and anthropologists alike who may not be familiar with the complicated math of geometric morphometrics. "It's just a matter of getting people used to using a digitizer" rather than calipers alone, Ross says.
In Ross' lab at North Carolina State, Sarah Cunningham, a graduate student in anthropology, demonstrates the program. Leaning over the cast of a skull held in place with clay on a metal stand, Cunningham places the slim metal tip of a digitizer — a pen-like device that records x, y and z coordinates in space — to the bridge of the nose.
"You have to make sure you get the right point," says Cunningham, craning her neck to reach another spot at the back of the skull.
Once Cunningham has entered 32 data points, 3D-ID will compare the information with a reference population of more than 1,000 individuals from various ancestries that Ross and other researchers have collected over time from museums, medical schools and even study collections dug up from old cemeteries. "As the reference population gets bigger, we're getting better," Ross says.
In addition to applications in forensics, 3D-ID will expand the ability of anthropologists to study past civilizations: They can now use the remains of children to determine what a population looked like in a specific area, instead of relying solely on adult remains.
The technique may also have clinical applications, Ross says. There is currently a lack of population-specific standards for reconstructive surgery in children, she says, and using 3D-ID data, clinicians should be able to develop standards for reconstructing a child's facial structure based on his or her ancestry.
When Horne invited Ross to try the technique on the remains of the John Doe boy, it was by far the youngest child she had ever attempted to classify. Using geometric morphometics, Ross determined that the boy was of Mesoamerican descent, a region extending from the middle of Mexico into Central America.
Knowledge of his ancestry has led the police to believe the child was perhaps an illegal immigrant, which was maybe why they have uncovered so little information on his disappearance over the last decade.
"I was thrilled, to say the least," Ross said. "They've been working on this for a long time." With that knowledge in hand, the search for the boy's identity continues.