The Skeleton Crew (19 page)

What's more, data entered into the NCIC wasn't easy to use. At the Virginia Beach conference where I first met Todd Matthews, one of the speakers was a tall man with close-cropped graying hair, a cleft chin, and a trim mustache.
Harry E. Carlile Jr., who had taught
the ins and outs of the FBI to recruits for more than a decade, explained to the primarily law enforcement audience his agency's “cold hit” system. Every morning, the massive NCIC computer in Clarksburg
churns through gigabytes of data seeking to match
details about the missing and the unidentified, and spits out something called $.M. (“dollar M”—the
M
is for “missing”) reports.

The $.M. reports look like they are generated by a monkey at a keyboard: strings of seemingly random characters followed by a number. The number represents a point total of items that matched or came close to matching, such as a missing person's date of birth falling within the range of an estimated age for an unidentified body. You have to learn how to read it, Carlile said.

The printouts are sent each day to law enforcement agencies around the country, but they are so incomprehensible, I wasn't surprised to hear that at least one police department discarded the $.M. reports unread, like junk mail. Carlile professed to be amazed by this, and distressed that many detectives never even saw the $.M. reports. Even if detectives had seen them, from what I heard, few could have spared the time to follow up on the potentially dozens of cold case leads.

And the NCIC's matching abilities were far from perfect. At one point a Seattle dentist decided to figure out if the records stored in the NCIC were doing any good at all for the missing and unidentified. Washington state had led the country in passing a law requiring investigating agencies to collect dental records for people missing more than thirty days, resulting in the highest numbers of dental records on file for missing persons anywhere in the country.

Gary L. Bell, one of the trained forensic
odontologists who coded dental records for the missing throughout the Pacific Northwest, knew that in 1982 several badly decomposed bodies and skeletons had turned up in the Seattle area. Ten years later, in 1992, Bell tried an experiment with four test cases: the dismembered remains of a woman found a hundred miles east of Seattle; the partially decomposed remains of a woman found in the woods forty-five miles from Seattle; the partial skeleton of a twelve-year-old girl; and a human cranium found near a river. Like all forensic specialists, Bell expected that dental records were the best hope for identifying these nameless corpses, and in fact this turned out to be true. Washington state investigators had used a U.S. Army dental matching system to positively identify all four bodies, suspected victims of “Green River Killer” Gary Ridgway, who pleaded guilty in 2003 to murdering forty-eight women between 1982 and 1998.

Yet, when Bell submitted dental records from the four sets of remains
and corresponding victim details into the NCIC, as he later reported in the
Journal of Forensic Sciences
, the NCIC failed to match any of them. There had been rumors that at least one of the four victims had been listed as missing in the NCIC system for “some time,” Bell wrote, and had never been matched to her own remains, even though a perfect dental match for her existed.

In fact, the NCIC's online searches—the searches that generated the daily $.M. reports—had never produced even one positive identification in the state of Washington.

Bell's experiment called attention to the fact that not only was the NCIC system ineffective for matching the missing and unidentified, it was poorly populated with data. Of the 73,000 missing persons entered into the NCIC at the time, less than 2 percent had dental records in the system.

Ultimately software is no match for human discernment in connecting clues. In another case,
the family of a Richmond, Virginia, runaway,
believing the boy was with another family member, didn't report him missing until a certain date. By the time they did, he had been dead for three days. Despite the almost perfect alignment of physical details between the missing boy and the body, NCIC computers never recognized the two as a potential match because of the disparity in dates. A web sleuth insisted the two were a match, and DNA confirmed it.

When Matthew Hickman of the Bureau of Justice Statistics set out in 2004 to figure out just how many unidentified remains existed in the United States, he heard an earful from medical examiners and coroners about the NCIC. “Entering data into that thing is a nightmare,
if medical examiners could even get access
to it” was the gist of the complaints.

But Bobby Lingoes had access.

Bobby committed to memory the details contained in the official NCIC report about the skeleton found in the Quincy woods. After work, he rode his mountain bike a half mile over Quincy's asphalt streets to Thomas Crane Public Library, which had something the police station didn't have: Internet access. It was there he first stumbled upon websites
such as the Doe Network. Little kids reading picture books near him grew wide-eyed when they glimpsed on Bobby's screen the photographs of the very pale people with the staring eyes. Despite Bobby's law enforcement career, he was shocked himself by the quantity of unidentifieds. He scribbled down facts listed on the Doe Network about missing individuals, then took his notes to work the next day to check against the NCIC database.

From his training on the use of the NCIC, he knew about a technique called off-lining that had been used
in the aftermath of the 1995 Oklahoma City bombing.
Fragments of the explosive-filled van Timothy McVeigh detonated in front of the Alfred P. Murrah Federal Building flew over a ten-block radius. Among them was a shard of the vehicle's transmission, on which investigators deciphered a partial vehicle identification number, or VIN. Having that unique identifier, investigators searched for a vehicle with those digits in its VIN that had been reported missing.

Bobby realized that if he came across information about a body with certain unique characteristics—an unusual tattoo or an artificial leg—he could search through individual records for only those details, instead of looking at broad categories such as age or race, which couldn't always be determined accurately from a decomposed corpse or a skeleton anyway. He looked, as he put it, for dead people to “speak” to him, beckoning to him from the grave with distinctive clues to their identities.

Such off-line queries couldn't be run independently from the Quincy police station. Bobby needed a warm body at the NCIC to plug in parameters that would transform the search process from a handful of buckshot to a heat-seeking missile. In 2001, Bobby's boss gave him permission to work off-line, and almost at once things began to happen.

Perusing the Doe Network, Bobby spotted
a Jane Doe who had turned up near Waco,
Texas. Seeking help from citizens, the Texas Department of Public Safety had posted information online about the young woman, who had suffered a severe head injury after a train hit her on tracks off Farm to Market Road 308 in the town of Elm Mott, population 190, on July 28,
1993. After days in a coma, she had died in a local hospital and been buried in a cemetery for indigents in McLennan County. She had no identification, but the word “Tonk” was tattooed on her left shoulder.

Working with an NCIC technical information specialist, Bobby queried the system: “Show me a missing person who is five foot four, eye color brown, with a tattoo on her left arm.” Within minutes a name popped up on the NCIC technician's screen in Clarksburg. He called Bobby at work, sounding a bit incredulous. “We have a possible match.” Angela Marie Parks was twenty-three, had two children, and had disappeared from her home in Bowling Green, Kentucky. Parks, who tended to vanish for extended periods of time, had not been in contact with her family since 1992.

Bobby was cautiously optimistic. He and the FBI submitted the tip to Waco and Bowling Green police. Meanwhile, he went to work on another train-related death from the Texas Department of Public Safety website, this one from 1992 involving a man in the town of Victoria. This victim, another transient, was thought to have gone by the name Kelly.

With the help of a Doe Network volunteer from Los Angeles named Vicki Siedow, a private investigator who had access to resources available only to licensed investigators, Bobby found a Social Security number for a missing Michigan man by the name of Kelly Zeazical. A set of fingerprints matched those of the train victim.

Fingerprints also sealed the file on Angela Parks. Her prints proved that the Elm Mott Jane Doe and Parks were one and the same. “Bingo!” Bobby said. “Oh, wow, it works. Let's do it some more.” (At the time, the Doe Network had been in existence for around two years. Angela Parks and Kelly Zeazical were among its first “solves.”)

Soon afterward, a mental health counselor named Sheree Greenwood, owner of an embroidery store in the rural central Massachusetts town of Warren, signed on with the Doe Network after sixteen-year-old Molly Bish, a classmate of her son's, disappeared from her post as a lifeguard at a local pond. (The remains of the stunning blond girl, who had been abducted, murdered, and dumped in the woods five miles from her home, wouldn't be found until 2003, following the largest, most expensive missing-­person search in Massachusetts history.)

Sheree silk-screened T-shirts at her shop, which caused her to zero in on a Baltimore case:
unidentified skeletal remains that had been discovered
clad in a red T-shirt depicting an Indian chief.

In December 2000, a worker moving an abandoned tractor-trailer on the eight hundred block of Roslyn Avenue in the industrial southeast section of Baltimore discovered human bones, the T-shirt, a pair of jeans, and white size-seven tennis shoes. Forensic anthropologists determined the bones belonged to a white woman, around four-ten to five foot four, 120 to 140 pounds, between thirty-eight and forty-five. She might have been strangled. She had likely been dead around a year. But without a name, the police had little to go on. They had the tattered red T-shirt, which said “Wynn Family Reunion 1997” and listed names—James, Dock, Jessie, Nick, Chief, Corinia, Bessie—and the words “children of Joseph and Estelle Wynn.”

Although Sheree didn't live near Quincy and hadn't met Bobby Lingoes, she had heard about Bobby's off-line successes and thought he might be intrigued. Bobby agreed that that T-shirt seemed to be shouting out clues. But its message was opaque.

Of course, there was no guarantee that if the meaning of the T-shirt could be deciphered, it would have any special significance to the woman who died wearing it. She might have picked it up in a secondhand shop or found it in someone's trash. Baltimore police had questioned local Native Americans, but no one knew of a family named Wynn or anything about a reunion at an unspecified location held three years earlier. Bobby's first NCIC off-line search produced no results. He plugged the names on the shirt into the motor vehicle database. Nothing.

He called Vicki Siedow in LA.

More than a decade later, Vicki Siedow was delighted to revisit the case when I called. Vicki, with shoulder-length blond hair and a wide smile, was smart, outspoken, and energetic, but she would never strike it rich as a private eye. In those days she took on pro bono cases such as the one with Bobby, and others for parents of missing kids. Meanwhile, her own kids ate grilled cheese and her electricity got shut off. Vicki was a sucker for a challenge, and as she saw it, real-life mysteries beat crossword puzzles any day.

Vicki had looked at the graphic of the Indian chief and the names on
the T-shirt. A family reunion wasn't just for celebrating the current generation, she said to herself. People celebrated their ancestors, people of note in the family tree.

She went on Ancestry.com and within a couple of hours had found an earlier generation of Wynns that included most of the individuals listed on the shirt. Sifting through names of survivors in obituaries, she traced history forward to a Thomas Wynn, a member of a Native American tribe based in Lumberton, North Carolina, and a chemical engineer for the U.S. Department of Energy who now lived and worked in Oak Ridge, Tennessee.

Thomas Wynn was surprised and impressed that someone had tracked him down based on an old T-shirt. He said his sister, Lola Wynn Haskins back in Lumberton, had designed it. Dozens had been handed out to a sizable crowd at the 1997 North Carolina family reunion. Vicki called Lola Haskins, who recalled that one attendee had given a T-shirt to a visiting girlfriend. Police tracked down the girlfriend, who, it turned out, had given the shirt to a Baltimore woman named Crystal Wright.

Crystal had roomed with a Brenda Wright, no relation, an addict and a prostitute who had not been seen in some time. Brenda Wright's brothers had been searching for her for months, showing people her photograph, posting flyers, calling hospitals and morgues.

With that information and a DNA sample from family members, Baltimore detectives identified the body as that of Brenda Sue Wright, age forty-six. The job still remained to figure out who had killed her. But identifying her body was a crucial step, a masterful bit of detective work on the parts of Bobby Lingoes, Sheree Greenwood, and Vicki Siedow.

A 2008 TV show featured the case
and quoted a Baltimore detective who sounded appreciative of the web sleuths' involvement. “Naturally, by all means we did use any help we could get,” he said on camera. Bobby didn't remember it that way. He recalled it was like pulling teeth to find out whether the body had been identified; he had to call the city's mayor before the cops would talk to him. “They really didn't give a shit about responding to the Doe Network or civilians helping them,” he said. “Police have that attitude, ‘I'm a cop, and you're not.'”

Bobby's next three cases—solved like clockwork in 2002, 2003, and 2004—were more akin to his first two, the Texas train casualties.