The trucks keep arriving from all over the country at an office park in Northern Virginia, each containing hundreds of envelopes marked with one word written on red tape: “Evidence.”

Just after 9 on a weekday morning, Kim Freeman pulls her hair back from her face and walks from her cubicle to a large storage room. “I’m here for the two,” she tells a desk attendant.

Two envelopes. She put them on reserve the day before, and now she scans one to make sure the contents label is correct. “Sexual Assault Evidence Collection Kit,” it says.

Another day has begun at the Bode Technology laboratory in Lorton. Another slow excavation of a national problem.

The problem is a backlog of untested rape kits, tens of thousands of them, each representing an alleged assault. In every case, evidence was collected from victims after their attacks, but it was never analyzed. Instead it sat on shelves for years, even decades.

There is no federal law requiring rape kits to be tested or tracked, and only a handful of states have enacted their own legislation. Staffing and money shortages have contributed to the backlog; a single kit can cost more than $1,000 to process. Sometimes, if a victim had already identified her attacker, police might not have processed her kit — though it could have contained DNA linking the assailant to other unsolved crimes.

Whatever the reasons, the backlog kept growing as kits were overlooked, ignored and forgotten. Recently, as attention on the issue has increased, that has begun to change. Twelve hundred kits were uncovered in Colorado Springs. Four thousand in Dallas. Twelve thousand in Memphis. And 6,600 in Houston, 5,000 of which were processed last year by Bode, one of several private companies contracting with local jurisdictions in a kit-by-kit effort to bring the backlog down to zero.

Freeman goes into the lab and opens the first kit. The victim was a minor, she reads from a report. The assault happened nearly a decade ago. She keeps reading. “It seems there were two attackers,” she says.

The kit contains cotton swabs, placed in separate envelopes, each with a label: Oral. Vaginal. Rectal. Bite Mark.

She cuts off pieces of the swabs, depositing them into test tubes after checking for discoloration. Sometimes if the assault was violent, the cotton swabs might be not white, but a dull blood color.

“I treat them all the same,” she says, believing that the best way to do her job is to focus on the science in front of her and not speculate on the stories behind it.

She picks up the next kit.

This victim, too, was a minor when the assault happened.

This kit, too, has a series of cotton swabs. Eventually, these cotton swabs will be treated with chemicals, put through machines, translated to DNA ladders on paper printouts. But now comes the cutting, the most intimate step in the process. Freeman opens the envelope containing the samples, which are stained a dark, reddish brown.

“Okay,” she says. “Okay.”

* * *

The kits Freeman is looking at came from California, where in 2011 a district attorney named Nancy O’Malley launched an investigation to learn how many kits had accumulated in her jurisdiction. O’Malley had heard of a backlog problem elsewhere in the country, but she didn’t anticipate finding anything so close to home, not where she had long prided herself for her county’s work on sexual-assault prosecution.

The numbers came back. There were 1,900 untested kits. The majority were from 2004 or later. Two were from 1983. O’Malley was stunned.

She booked a ticket to New York, to learn how that city successfully eliminated its 17,000-kit backlog in the early 2000s, and then she set out to see how her jurisdiction could do the same thing. New York’s kits had been processed at Bode, and O’Malley’s county decided on a public-private partnership as well.

Testing in other jurisdictions has yielded results: New York’s arrest rate for sexual assaults went from 40 percent to 70 percent after its backlog was cleared. When Detroit tested 1,600 of its backlogged kits, the city came back with 127 potential serial rapists, according to the Joyful Heart Foundation, which advocates on issues related to the backlog.

The California partnership with Bode is new, but O’Malley has experience telling rape victims that their kits have yielded DNA matches.

Some of them cry, and some of them shut down in shock, she says. One woman’s attacker had worn a mask. When she learned there was a name to match to her faceless assailant, the woman realized she’d known him all along. He was a friend.

“Imagine sitting at home and getting a phone call saying: ‘This is the police department. There was a hit made, and we think we have a match,’ ” O’Malley says.

This is what she thinks about while she waits for the results to come back from Bode: the victims who, in many cases, don’t know their kits were never processed. O’Malley wonders whether a DNA hit will make them relieved, or force them to mentally relive their attack, or whether, after all this time, the victims will just be angry that it took so long.

As soon as the DNA profiles come back from the lab, they can be uploaded to a national database and O’Malley’s team will start prosecuting crimes.

* * *

Lab analysis can take weeks. There is a reason kits are expensive to process; from start to finish, dozens of procedures must be carried out in a tedious and specific order. Intake is followed by sample-cutting. Cutting is followed by “extraction,” in which male cells are separated from female cells.

Back in the lab on another day, Natalie Morgan, Bode’s director of forensic casework, begins extraction on a batch of samples from the Midwest. She runs 30 test tubes at a time through a whirling centrifuge that forces male DNA to the bottoms of the tubes. She uses a pipette to transfer the top matter to different containers labeled with corresponding numbers. It’s repetitive work, and it takes hours.

She adds the final chemical needed to complete the extraction process. It’s the same chemical that hairstylists use to give women perms. It makes the lab smell like a hair salon.

She finished a batch recently. The victims were mostly children. That was just chance, she assumes; she doesn’t know how which cases end up in which batches. Some of the children were as young as 5. Morgan has a young daughter at home. Those samples stayed in her head for a while.

After the extraction step comes quantification, to determine how much DNA is present, then amplification, in which a DNA sequence is multiplied for easier analysis. And then the paperwork of sorting through the results. Each step is designed to purify the DNA and distill it down to something simple: a yes or a no. Yes, there is male DNA present. Yes, there is tangible evidence of an assault.

A week after Freeman finished cutting the cotton swabs from California, an analyst named Daniel Curtis pages through binders of paperwork in his cubicle, preparing the analysis for a batch from a Southern state. He uses an algorithm to calculate the odds that a suspect’s genetic material would be the same as someone uninvolved with the crime: 1 in 330 quintillion, 1 in 27 sextillion. This kind of certainty is the closest thing to an ending that happens in the lab.

Sometimes there is not a yes or no: There are trace amounts of DNA, but not enough for a firm result. Or, perhaps, the sample got too old and degraded before it could provide an answer either way.

* * *

Three weeks after Freeman first requested the envelopes from the evidence room, she goes to her cubicle and lays out a notebook ripped to a fresh sheet, ready to examine the data on the batch from California.

She boots up her computer and adjusts the settings so DNA profiles begin to appear. On-screen, each profile looks like a heart-rate monitor, with spikes peaking above a straight line. Taller spikes represent the presence of more DNA.

Some of the California samples, Freeman discovered earlier in the process, contained no male DNA, or too small of an amount to analyze.

Now, as Freeman looks at her computer monitor, other samples do reveal male DNA, but DNA that might belong, according to case notes, to a consensual partner the woman reported having. Still other samples reveal multiple assailants.

Freeman opens the next file. It’s one she remembers cutting.

The DNA ladder for this case appears on her screen, and Freeman silently examines it.

“Aha!” she says. “This is good.”

The ladder has tall, distinct peaks.

“This is a single-source male profile.” She leans in for a closer look. “You couldn’t ask for a prettier profile.”

She checks the case notes for the file. It seems that a suspect’s DNA was also submitted. If the DNA matches, then an attacker has been identified. Three thousand miles away, there could be a prosecution, a conviction, a moment of closure.

An ending, she thinks as she closes the file, and then she moves on to the next one.