• Science
  • Crime

It Might Be Impossible to Get Away With Crime Some Day

14 minute read

Somewhere in the 20,000 genes and 3.2 billion base pairs that make up the genome of Ted Kaczynski lie the genetic codes for madness. It wouldn’t be easy, even today, to tease out those genes, and it was even less possible in 1996, when the man better known as the Unabomber for the 16 bombings he carried out over 17 years was at last apprehended.

There are a number of locations on the human genome that have been implicated in the development of schizophrenia, and Kaczynski has been variously diagnosed as schizophrenic or at least schizoid or schizotypal — lesser related forms of the condition. Whatever his diagnosis, you don’t wind up in a tar paper shack in a remote corner of Montana mailing out package bombs to strangers if you don’t have at least a few genetic wires crossed.

As it turned out, even if DNA science has not helped diagnose Kaczynski, it did help the FBI nab him. After The New York Times and Washington Post published the Unabomber’s 35,000 word manifesto in exchange for a pledge from the to “desist from terrorism,” Kaczynski’s younger brother David thought he recognized something familiar in the manifesto’s writing style. He called the FBI to report his suspicions and investigators compared the DNA in saliva traces from the envelopes that had been sent to the Times and the Post to others Kaczynski had sent his family. A match was confirmed.

“It was a very limited amount of DNA; it was a very low level test,” says Jenifer Smith, a 23-year veteran of the FBI who worked on the Unabomber case. “But it certainly indicated that he was in the category of people who could not be excluded. That became part of the probable cause that allowed the agents to serve search warrants on his property.”

It was an investigative coup, a triumph of patient, dogged law enforcement and yet, as Smith says, it was comparatively primitive stuff. The forensic science of 1996 was a blunt instrument compared to what’s available in 2017, and Smith, more than most people, should know. Today she is director of Washington, D.C.’s Department of Forensic Sciences (DFS), a gleaming new laboratory in a gleaming new building that conducts forensic work not just for the Metropolitan Police Department police but for numerous other offices and law enforcement agencies, including the District Fire and Rescue, the U.S. Attorney’s Office and the Public Defender Service.

The 220 investigators in the 351,000 sq. ft. lab—which shares space with the Office of the Chief Medical Examiner—worked on 8,576 cases in 2016 alone, benefiting from technology that has grown exponentially in just a generation, from fingerprint and bullet identification to crime scene investigation to forensic chemistry to digital forensics—the latter a category of criminal investigation that barely existed in 1996. All of this is helping law enforcement achieve its most fundamental goal: preventing lawbreaking when they can, and catching the lawbreakers when they can’t.

“The forensic science we have today,” says Karen Wiggins, a 25-year veteran of the District’s Metropolitan Police Department and now director of the labs within the DFS, “makes it likelier that if somebody is committing a crime, that person is going to be apprehended and convicted and will go to jail.”

If the science has advanced across the entire landscape of forensics, it is DNA technology that has arguably made the most progress. The ability to use genetic sequencing merely to put potential perps into buckets of suspects who can’t be excluded has given way to an ability to make far-more solid identifications of specific individuals. There is only a 1 in 64 billion likelihood that a pair of unrelated people will carry closely matching DNA, though since tests aren’t perfect, that’s not the same as saying there’s only a 1 in 64 billion chance of a suspect being misidentified. Still, things have progressed enough that done right, genetic testing can effectively make a case. But that business of doing it right isn’t easy.

During the O.J. Simpson murder trial, just the year before the Unabomber case was solved, prosecutors relied principally on a form of DNA testing called restriction fragment length polymorphism (RFLP), which involves cutting strands of DNA at particular spots and analyzing the length of the snips. In the same way that measuring two different people from, say, the knees to the ankles will typically yield different results, the length between certain target spots on a DNA strand will differ from person to person. If the strand length in a bit of DNA evidence left at a crime scene matches the strand length in a suspect, it’s good news for investigators—and bad news for the perp.

Still, such evidence hardly meets a beyond-a-reasonable-doubt standard, partly because other people might have the same matching strands and partly because RFLP requires a large, reasonably pristine sample to work well. The blood used in the Simpson case was recovered in small quantities from clothes, a car and elsewhere; worse, it was a mix of the suspect’s blood and that of both victims’.

“The best you can say in a case like that,” explains Smith, is “‘Well, I can exclude all other people, but I can’t exclude these three.'”

Things improved at about the same time as the Simpson and Unabomber cases were playing out with the introduction of polymerase chain reaction (PCR), which allows investigators to begin with a very small DNA sample and reproduce the sequences over and over, providing more genetic material to study. Still-newer DNA analysis kits make it possible to study those samples in much more detail. Mixed stains, for example, can now be automatically separated by gender, which is especially helpful in cases of sexual assault, since those overwhelmingly involve a male assailant and a female victim.

DNA technology is making hair samples easier to study too. While hair is typically a poor subject for genetic analysis because it contains very little nuclear DNA—or DNA drawn from the nucleus of the cell—it does contain a lot of DNA from the mitochondria, the tiny energy-generating organelle within the cell’s body. Mitochondrial DNA is inherited exclusively from the mother, which means it will be identical in anyone descended from that one woman. Again, that does not makes for a definitive genetic identification of a suspect; judging by mitochondrial DNA alone, David Kaczynski could have been just as guilty as his brother Ted. But it does help narrow the pool of possible perps way down.

All of this DNA data is now being widely shared among law enforcement agencies. Both nuclear and mitochondrial gene sequences gathered at crime scenes and collected from suspects by local, state and federal investigators are regularly uploaded into a national, searchable FBI database known as the Combined DNA Index System (CODIS).

If there is a weak spot in the CODIS concept, it’s that while science may always advance, law enforcement budgets don’t, and a lot of DNA evidence never even gets processed, much less reliably uploaded into a national database. Rape kits—which are used to collect evidence from victims in the immediate aftermath of an assault—are among the most powerful tools available to get sex offenders off the streets. Currently, however, there is an estimated backlog of 175,000 kits on the shelves of police labs nationwide awaiting processing. Nonprofit groups like End the Backlog are raising money to address this problem, but progress is slow and many offenders remain at large.

If DNA is all about the numbers, refining the probability of a match down to as many decimal places as possible, fingerprint and bullet science offers much less certainty. In both cases the challenge is all about recognizing patterns—the whorls and dots and bifurcated ridges in a fingerprint, or the unique pattern of scratches and grooves the interior of a gun barrel leaves on a bullet as it passes through.

Microscopes and even mere magnifying glasses long ago made it easy to see and photograph those telltale marks, but what came next was always painstakingly slow: looking really, really closely at lots and lots of fingerprints or bullet patterns on record to see if you could find a match. “We routinely searched a fingerprint against a million cards already on file,” says Barbara Evans, an FBI veteran who started at the Bureau in 1971 and is now fingerprint supervisor at the DFS. “You’re actually manually flipping through thousands of cards.”

Actually you were manually flipping through them. Today, computers have taken on a lot of that work, with another databank—the Automated Fingerprint Identification System (AFIS)—conducting massive searches of thousands or even millions of prints far more efficiently than humans could. AFIS was first installed in the 1990s and immediately slashed the time it took to find a match, and more-powerful computers have made things faster still—down to as little as three minutes for a search that would once have consumed days.

Yet another databank called the National Integrated Ballistic Information Network (NIBIN) does the same work for investigators looking for bullet matches, and the DFS ballistics division provides a lot of the imagery that gets uploaded into the database. In a glassed-in display room, the lab shows off upwards of 5,000 guns that District police have taken off the streets over the the years—sometimes going back a very long time.

“We have a revolver that shoots ball and caps,” says Wiggins, with evident institutional pride.

When guns are freshly seized, they are often fired on an indoor rifle range at the lab and the bullets are collected so their distinctive markings marks can be scanned. If a pristine bullet is required—one that is not flattened by impact with a surface as most bullets are—the gun is fired in a long water tank, which provides enough resistance to cause the bullet simply to slow and stop without making impact with anything except the bottom of the tank as it gently falls.

Forensic scientists are getting better not only at analyzing and reading evidence, but at finding it in the first place, thanks to improved technology for investigating crime scenes. In a basement garage at the DFS headquarters, a burgundy SUV sits off by itself in a parking bay. It was once the scene of a sexual assault and while the perp has long since been tried, convicted and incarcerated, his impounded car is still being used to train law enforcement officers.

To look at the interior of the car with the naked eye is to see nothing in particular. But to look at it in a particular shade of blue light while wearing a particular shade of orange goggles is to see it in a wavelength of precisely 454 nanometers. And in that light, ugly things happen—as spots on the carpet and ceiling light up brightly. Almost any body fluid will fluoresce at 454 nanometers, but in this case the fluid is semen—though not from the original crime; that was all removed in swatches for evidence. In this case it was planted by trainers to see if new recruits will know where to look and what they’re looking for.

“People will forget the ceiling,” says Greg Greenwalt, unit manager of the crime scene sciences division. “But if a hand swings upward during a struggle, evidence can be there.”

Blood, easily the most abundant of body fluids, is the only one that doesn’t glow in the blue-orange filter, but it does show up as a ruddy brown. The left rear bumper of the old SUV is smeared with dry blood—bovine blood, Greenwalt stresses, to prevent biohazards—and once you’ve learned what to look for it’s easy to spot.

“In the past, we would literally use a D-cell flashlight with an orange filter,” he says. “The glass technology, the bulb technology, the light technology has all jumped.”

Crime scene investigation is taking advantage of something approaching virtual reality too. DFS investigators now work with 360-degree lasers that take the precise measurements of an enclosed space, from the center of a room to all other points in the room, measuring them within 99.9% accuracy. These measurements are then married with equally thorough photographs of the room, providing a total immersion experience for jurors who can see where a crime was committed without the need for site visits.

Finally, and most innovatively, is the progress that’s been made in digital forensics. Just a little over twenty years ago, if a suspect had any digital presence at all, it was mostly locked inside the hard drive of a desktop computer. Yes, there was an Internet, but it was slow and pokey and not much good for even the most imaginative of bad guys. A generation later, there’s all manner of sophisticated cybercrime and that gives investigators both more policing to do and more tools to use to catch the perps.

“It’s what we call the digital landscape,” says Tracy Walraven, lead forensic scientist in the digital evidence division. “Everything is interconnected, everything has an IP address. Even when you’re digging through just one digital device like a cellphone, the information could also be on a tablet or in the cloud.”

Walraven once worked in digital technology both on Capitol Hill and at NASA and spends a good deal of her time investigating hacking and phishing and other more familiar crimes. At the moment, however, she is also paying attention to something that should be beyond such criminality: toys.

On a work bench in her DFS lab sit both a Furby doll and a Barbie doll, entirely innocent things except that they’re also interactive Furbies and Barbies. A child can engage with them, talk to them and share information with them, which is fine as far as it goes. But the toys in turn can engage with other devices, and from there the information can get out to the world.

“We’re looking at this from the creepy person’s perspective,” Walraven says. “We know how hackers think and how child molesters look at these toys as bait to lure.”

So far she has not found any evidence that the toys are being exploited that way and she suspects that if she does, the companies will respond quickly to make the necessary fixes. For now, she believes, the best approach involves public service announcements to warn parents of the possible risks. On July 17, the FBI issued just such an alert, explaining what makes interactive toys vulnerable and how best parents can protect their kids.

Twenty-first century forensics are by no means perfect, and the DFS, like other law enforcement operations, has learned that lesson painfully. Just last March, the lab was rocked when news broke that at least 150 cases of firearms identification were being reviewed for possible errors. The month before there was a similar problem in the lab’s public health division when it was revealed that at least nine pregnant women were erroneously told that they had tested negative for the Zika virus when in fact they were infected. In both cases, it was the DFS’s own internal quality control that caught the problems, but it was a black eye for the lab nonetheless.

Smith, as the DFS chief, knows that she will earn applause for what the lab does right and will be judged responsible for what it does wrong—and it’s that second part that preoccupies her most. “Once you cut a corner, once you cross that line, you can affect the reputation of your lab, of yourself, of your discipline,” she says. “That continually forces us to be the best at our game.”

That struggle is an ongoing one—and the cunning of the perps themselves ensures that it always will be. But there’s no denying that the crime-fighting science is better than it’s ever been. In the eternal arms race between the people who enforce the laws and the people who break them, it’s the good guys who increasingly have the edge.

This video was produced in partnership with Discovery.

The story of the Unabomber is the subject of a new Discovery series, Manhunt: Unabomber, premiering Aug. 1 at 9 p.m. E.T.

More Must-Reads From TIME

Write to Jeffrey Kluger at jeffrey.kluger@time.com