On June 25, the remains of a child were found in a trash bag on the shore of Boston’s Deer Island. It wasn’t clear how long she’d been dead, but there was some decomposition. Investigators estimated that the unidentified girl, “Baby Doe,” was about 4 years old. Among the clues found at the scene: pollen on the child’s pants and blanket.

The investigation quickly hit an impasse. The Boston police contacted the National Center for Missing & Exploited Children (NCMEC), and asked if pollen analysis of the evidence would be possible. NCMEC—which served as a facilitator throughout the case—contacted Andrew Laurence, the head palynologist of U.S. Customs and Border Protection. Palynology is the study of pollen grains and other spores.

I’d called Laurence after reading about a cold case: A woman was murdered in Baltimore County in 1976 and, nearly 40 years later, still lacked an identify. The Woodlawn Jane Doe—she’d been found near the Woodlawn Cemetery—had been asphyxiated. This caught my eye: “Over the years, Baltimore County Police have had several items analyzed, recently ordering a pollen analysis to narrow down her identity.”

The analysis of the pollen, I learned, was performed by Laurence, who works out of a Houston lab. As it happened, the Baltimore case was off limits, but he could talk about his own profession and its lonely practitioners. And what he had to say was illuminating.

* * *

After the attacks of September 11, 2001, the federal government approached Vaughn Bryant to, as he put it in a recent email, help “identify and find the terrorists involved.” Bryant, the director of the palynology laboratory at Texas A&M University, is an unusual man. He’s investigated the origins of counterfeit honey; he’s even traced the first recorded kiss to India, in approximately 1500 B.C. He agreed to work for the government in a part-time capacity, and in 2010 he began doing forensic pollen studies, mostly of illegal drug shipments, but also on “persons of interest” caught by the Department of Homeland Security.

Bryant’s work was successful enough that DHS wanted a full-time forensic palynologist. Bryant trained his student, Andrew Laurence, who was hired by U.S. Customs and Border Protection in 2011, while still in grad school and at work on his dissertation. (U.S. Customs and Border Protection falls under the umbrella of DHS.)

The work, as Laurence describes it, is mindboggling. There are 380,000 different species of plants on the planet, each with its own unique pollen type. As Laurence put it, “Plant composition changes depending on where you are. Even if the same types of plants are growing in different areas, the abundance of each plant may be different. For example one place may have 70 percent pine, 20 percent oak, and 10 percent grass while another area may have 40 percent pine, 40 percent oak, and 20 percent grass. Same plants, different composition.” For example, he said, “in the United States, the number of pine trees relative to other plants decreases traveling east to west as the Eastern Woodlands transitions into the Temperate Grasslands.”

Each region has its own unique pollen print generated from those plants. “Think of it,” says Laurence, “as a fingerprint for a region.” And it’s a relatively indestructible fingerprint. Even if you put clothes through a washing machine, a pollen print remains.

It’s all quite seductive. It’s for good reason that forensic palynology—the use of pollen and spores to solve criminal or civil cases—is not exactly obscure. It has played a recurring role on crime shows (Bones, for example); it was pivotal to the widely-publicized investigation of Ötzi the Iceman, and it even helped to convict Bosnian war criminals. As an investigative tool, forensic palynology has proven its use for decades.

Why, then, is 29-year-old Laurence the only full-time forensic pollen analyst in the United States? It’s not as if he can’t use some assistance. When asked how many cases, ranging from smuggling to homicide, he’s worked on during his three and half years of full-time employment with U.S. Customs and Border Protection, Laurence said, “Oh, geez. Hundreds? At least 160.”

* * *

The most famous forensic-palynology case, and one that Laurence looks to as a landmark, is the first. In 1959, in Austria, a man disappeared and, despite the lack of a body, was presumed murdered. The investigators had little physical evidence, but they had a suspect with a muddy pair of boots. As recounted in Criminal Psychology and Forensic Technology, a local palynologist found in the boots a 20-million-year-old fossil of hickory pollen, from a tree that no longer grew in Austria. But there was a small area, on the Danube River, where the pollen grain had been absorbed into the environment. Presented with this information, the suspect disclosed the whereabouts of the body.

Since then, forensic palynology has been embraced by law enforcement in the United Kingdom and, with particular enthusiasm, in New Zealand. Dallas Mildenhall, a white-haired New Zealander in his 70s, is among the world’s leading forensic palynologists. Mildenhall achieved prominence in 1983, when he consulted on the case of Kirsa Jensen, a vanished a schoolgirl, and solved it.

“I was able to show that a particular individual that the police were suspicious about was the person who was involved in that girl’s disappearance,” he said recently. The case received so much attention in New Zealand that police used him for other cases. Mildenhall has since worked on 300 to 400 cases internationally, spanning the theft of a lawnmower to murder. For 30 years, he’s collaborated closely with Bryant and the U.K.’s Patricia Wiltshire—all giants in the field.

From an American vantage point, New Zealand’s embrace of pollen analysis is impressive. Whereas the U.S. only uses pollen analysis as an investigative tool—it has yet to meet the Daubert standard—in New Zealand, it is admissible as evidence in court.

But the long-term outlook for forensic palynology isn’t so rosy, as funding has been scarce. “This funding problem is quite serious,” says Mildenhall. “I used to work on about 30 or 40 cases a year in New Zealand. I’m now lucky if I work on two or three.” The work is expensive, so palynologists are called in only if the initial investigation doesn’t yield results. This is troublesome because, ideally, the palynologists should collect samples as soon as possible, to avoid contamination. If it’s collected too late, it’s sometimes not admissible.

* * *

Forensic palynology is arguably a victim of its own success. “I am swamped with the amount of samples that come in,” Laurence says. “I can’t even come close to staying on top of just what U.S. Customs and its constituents provides.” And there isn’t a fleet of aspiring forensic palynologists on the horizon to populate the nation’s crime labs. Palynology takes years of training. A Ph.D. is the industry standard, and a background in botany, ecology, pharmacology, and, depending on your specialty, geology is required. For forensic palynology, tack on geography and climatology.

After all that work, graduates tend to gravitate toward the money. “Most palynologists are actually geologists. They’re working for the oil industry,” he said. “There’s not a whole lot of palynologists in the world to begin with, and even less doing forensics.”

Vaughn Bryant told me that a number of federal agencies want to develop a “black box” that would be run by a computer—the idea being “you could put in a pollen sample from something and bingo, the black box identifies all the pollen types and sends the information to the computer for interpretation … all in a matter of minutes.” So far, he said, the U.S. government has spent “several million or more” trying to develop this. He emailed:

I have been involved from the very beginning with different government agencies and I told all of them up front in the beginning “THIS WILL NOT WORK.” Nevertheless, they continue to throw money at this concept. What they ought to be doing is training more people like me to do this type of work…

In 2012, U.S. government agencies held a meeting in Washington, D.C., with the major agencies that deal with drug importation and terrorism, including the Department of Homeland Security. The agencies agreed a centralized pollen database was a necessity.

The next year, I learned, the government allocated approximately $4 million to build the database. Mark Bush, a professor of biological sciences at Florida Institute of Technology who had experience building pollen databases, was chosen to head the project. From the outset, Bush recalled, he was adamant the database should not simply be a cache of images. “You’ve got a bunch of unknowns and uncertainties,” Bush told me. “One is, is that really the right grain? Does the name match the grain? What’s the provenance of the pollen?”

Bush says the pollen in people’s collections—pollen-reference collections are not uncommon in this field—either built by the collector, or inherited, are potentially problematic. The surface patterns or size and even shape of pollen grains change, depending on the mount they were put on and how long ago. “A lot of the grains from collections that we have from, say, the 1950s, ‘60s and ‘70s, those grains may have almost doubled in size,” he said. “They look a bit like a drowned body. They are bloated and smooth and don’t look fresh.”

Bush’s solution has been to standardize the mounting technology; he used a single protocol. He also went to top-flight herbaria and built materials only from the voucher herbarium specimens. (Here it’s worth mentioning: The $4 million outlay from the government funded not just the collecting and databasing of pollen; it also paid for morphological keys and extensive modeling for geolocation. “It goes beyond just some pictures of pollen grains,” said Bush.)

The project is now in the third year of a three-year phase, which ends in July 2016. They’ve photographed 3,300 grains—largely from the near-Tropics—and described them in the database. The aim to is to amass 4,000 by the end of the funding period and, ultimately, 10–12,000 grains.

As for Mildenhall, he told me there are papers coming out about “the automatic pollen identification and classification, using electronic means.” But there’s a significant caveat. “At the present rate of progress,” he said, “we’ll be able to electronically identify pollen grains, fern spores, and fungal spores sometime toward the end of the next century.”

* * *

In September, thanks in part to the assistance of Laurence, Baby Doe was given a name: Bella Bond. An investigation revealed that her death was the result of being punched in the abdomen. Her mother, Rachel Bond, and her mother’s boyfriend, Michael McCarthy, were charged in September—McCarthy with murder, Bond with being an accessory after the fact. Both pleaded not guilty, and will be in court on November 19 for a probable cause hearing.

Laurence has moved on to other cases, but this one sticks with him. He received the samples in July and started processing them immediately. “I vacuumed the clothing samples. I did a chemical wash of the hair to try to remove pollen grains from it,” he recalled. Processing took a single day. Then, he said, “there was five or six days of scope time, me sitting down at microscope, counting and analyzing the pollen grains.” Finally, he took three or four days to write and research the report, which he sent to NCMEC.

He found the pollen contained multiple species of cedar, which “is not normal, really, anywhere, especially in the northeast. Only the cedar-of-Lebanon can grow there. If you have more than one type, that means it’s kept alive by someone, most likely in an herbarium or an arboretum.”

“Baby Doe had played among the pines and oaks of New England,” reported The Boston Globe.

[S]he was dusted with traces of privet hedges and cedar-of-Lebanon, which are not native but are often planted in the suburbs. The soot mixed in with the pollen told investigators her surroundings were urban. Somewhere near Boston, they concluded.

Thanks to Laurence, the police realized the child was local. He considers the case his proudest achievement: “It’s the gratification of seeing the end result, of having an impact.”