Reprinted with permission from ProPublica.

British chemists toiled with a tripod-shaped bond of nitrogen and oxygen molecules linked by carbon and hydrogen they referred to as “research department explosive” — a substance one and a half times as powerful as TNT, but so delicate it had to be mixed with beeswax to be stable and pliable enough to fit into warheads. Even then, it wasn’t good enough. Only 70 tons could be made in a week. Defeating the Nazis would require more.

In 1941, American chemists accomplished what their British counterparts could not. John Sheehan and Werner Bachman, University of Michigan researchers, worked with a team of government scientists to invent a new chemical process that made it possible to manufacture what Sheehan described as “super-explosives.” Best, enormous quantities could be churned out quickly — 500 tons a day, an assembly line for destructive might.

The Americans called the new formula RDX, and it transformed weapons overnight. RDX enabled the bazooka — the world’s first hand-held anti-tank rocket launcher — to pierce armor. RDX was packed into 10,000-pound underwater bombs dropped by British airplanes to blow up German river dams and disrupt the country’s hydropower in the critical Dambuster campaign. It was even surreptitiously soaked into firewood that would later explode in the furnaces of German locomotives.

By many estimates, RDX was critical to victory in World War II. It also spawned the greatest period of military manufacturing — and perhaps the largest arsenal — in the history of the planet. For most of the last 74 years, a single industrial plant in rural Tennessee served as America’s RDX factory and it produced as much as 40 million pounds of white crystalline powder each month that fueled the vast carpet bombing of the Korean peninsula, and then later, America’s involvement in Vietnam.

RDX is “a mighty instrument,” Sheehan wrote, “that may well have transformed the nature of modern warfare.”

But RDX, a powerful triumph of military ingenuity, has had an unwanted second life — as an unusually persistent pollutant poisoning the American homeland.

At bomb-making plants and ordnance testing ranges across the United States, RDX has spread into the soil and contaminated water supplies. The first signs of trouble emerged at Army bomb-packing plants in Tennessee, and then near Grand Island, Nebraska, where drinking water aquifers were found to be contaminated with the explosive. By 1990, the U.S. Environmental Protection Agency had designated RDX a possible human carcinogen. Near Salt Lake City, people who grew their food with RDX-contaminated water alleged it caused their cancers. Then RDX damage was identified in Virginia, California, and even on one of the East Coast’s iconic vacation destinations, Cape Cod.

The Pentagon nonetheless never stopped manufacturing with RDX or test-firing the weapons made with it on U.S. soil. As a result, the number of bases — and their surrounding communities — that face environmental threats from RDX has only continued to grow. In 2013, for example, residents near Fort Jackson in Columbia, South Carolina, who live near a U.S. Army grenade range where more than half the Army’s soldiers train, were told the Army had found RDX in their drinking water wells. In 2015, the Department of Defense settled a lawsuit after contamination was found in public drinking water 143 miles downstream from the Army’s RDX manufacturing facility in Kingsport, Tennessee. According to the court complaint, the plant was dumping more than 68 pounds of pure RDX directly into the nearby Holston River each day. And just this summer, high levels of RDX were identified in the groundwater on the site of an old missile factory near Los Angeles that was supposed to have already been cleaned up.

In all, previously unreleased federal EPA and Defense Department data obtained by ProPublica shows there is RDX contamination at more than 65 military installations across the country.

Harry Craig, one of the EPA’s foremost authorities on explosives and Defense site cleanups, said RDX was the “number one” challenge facing the EPA’s office of federal facility enforcement. Worse, he said, was the fact that all these years later the scope of the dangers posed by RDX was still “not fully recognized.”

From the first reports of RDX contamination on American soil, the Pentagon has either ignored or actively sought to discount RDX’s threat to public health and the environment, according to ProPublica’s review of thousands of pages of EPA and Pentagon documents and the accounts of more than 23 current and former officials and lawmakers. When confronted with evidence of the risk posed by RDX, the Pentagon tried to get its bases exempted from almost any kind of environmental oversight, Congressional records, transcripts and Pentagon documents show. When that failed, the Department of Defense financed an array of studies that minimize evidence of RDX’s potential to cause cancer and other health problems, and then pressed the EPA to relax its classification of RDX’s cancer threat — and the environmental cleanup standards that are recommended along with it, Pentagon and EPA records show.

“Where they believe their core interests are threatened, they become quite adversarial,” said Robert Sussman, who was senior policy counsel to the EPA in the Obama administration. “The chain of command is behind them and they are able to bring to bear a lot of pressure within the government.”

ProPublica’s broader investigation of the Pentagon’s handling of its environmental liabilities makes clear that denial and avoidance are key elements of a playbook its officials have employed for decades. The Pentagon, for instance, has continued to dispose of munitions with open burning, insisting it is necessary and safe long after other countries abandoned the practice in favor of cleaner methods. And the Pentagon routinely hires private contractors to handle its toxic cleanups, looking the other way or claiming not to be liable when those companies have proven to be incompetent or corrupt.

With RDX, the Pentagon’s long and bitter behind-the-scenes combat with the EPA endures to this day. In 2013, the EPA appeared ready to declare RDX not just a possible, but a “likely” carcinogen, an outcome that would have costly and complex implications for the Pentagon. The Pentagon’s existing obligations to address thousands of contaminated installations are already daunting, with costs that could top $70 billion. If EPA officials were to decide that RDX “likely” causes cancer, it could lead to new regulations and demands for more thorough cleanups, adding significant costs. Some military officials have said those additional costs — several cleanups have cost $1 billion — could cut into spending on weapons and soldiers.

And so over the last four years, the Pentagon has campaigned heavily against any upgrading of RDX’s cancer risk.

The effort appears on the verge of a striking victory — not only avoiding any upgrade of RDX’s cancer risk, but actually leading the EPA to downgrade its estimate of RDX’s cancer potency.

Ronald Melnick, a former senior toxicologist at the National Institutes of Health, has been observing the EPA’s scientific review of RDX and submitted comments to the agency. Melnick said he believes the pressure exerted by the Pentagon through comments and presentations, and the RDX research it has recently funded, have prompted the EPA to reverse a long-held position justified by the scientific findings over decades.

“I was surprised and shocked,” Melnick said about the EPA’s latest assessment.

In an emailed response to questions, the EPA rejected Melnick’s suggestion. “No individual, agency, or stakeholder group has had undue influence in the development of the peer review draft,” a spokesperson wrote, referring to the agency’s chemical assessment process.

In interviews and in emailed questions, four Pentagon officials also said that they believed the Defense Department has acted in good faith. The Army’s director of toxicology, Mark Johnson, rejected assertions that the Army’s research has been motivated by anything other than advancing the science. And Lucian Niemeyer, the assistant secretary of defense for energy, installations, and environment, said in a statement that “the Department of Defense is committed to meeting environmental cleanup requirements while protecting human health, safety and the environment.”

ProPublica’s examination of the battle over RDX lays bare an enduring predicament: The military’s role in protecting America comes at the price of being one of the nation’s most prolific polluters.

“As long as we fight wars and we send men and women into combat, we have an obligation to them and to the country to train them realistically. And one of the costs of that training is to create some environmental damage,” said Alberto Mora, a former general counsel for the Navy, and now a senior fellow at the Harvard Kennedy School of Government. “I always felt this was a tradeoff that the country needed to make and would want to make.”

Few Americans, though, have any idea such a bargain was ever struck.

THE U.S. ARMY BEGAN to grasp that its use of RDX could cause enormous problems when it found explosives in the groundwater seeping underneath the fence line of the Cornhusker Army Ammunition Plant, eight miles outside the town of Grand Island, Nebraska.

Cornhusker was run by the Quaker Oats Ordnance Corp., a subsidiary of the cereal company, since it began operating in 1942. With so many American men fighting in Europe, women ordnance workers, or WOWs, built the bombs they dropped. They received crystalline RDX by the trainload, then stirred it in vats mixed with TNT and tamped it into the tops of 5-foot-tall bomb shells destined, as the local newspaper put it, “for Hitler.”

Explosives residue collected on the tables, floors and equipment, and on the workers’ clothing. The dust was hosed down and the clothing laundered before so much accumulated that the waste itself could explode. But wastewater, running in a frothy pink mixture that is a telling sign of the explosives, streamed uncontrolled from the plant’s processing units and was allowed to seep into soil or pool in large, unlined pits dug at some 58 sites across the 19-square-mile property. For more than 30 years, the plant operated this way, its production surging to produce mines and artillery shells for the fighting in Korea and Vietnam.

Then, in 1980, the Army tested the grounds at Cornhusker and found that RDX and other explosives had settled more deeply into the soils than had been understood, reaching the aquifer used for drinking water. The Army knew this was likely not isolated contamination. “Pink water” had been seen running from other Army plants that manufactured ordnance, in states from Tennessee to California. At the Iowa Army Ammunition Depot, residents reported seeing raccoons with dyed fur after the animals had waded in nearby reddened streams.

But if the Army was only starting to come to terms with the physical spread of RDX contamination, it had long known of the compound’s potential toxicity. Army documents show that as early as the 1970s, Pentagon medical experts suspected that RDX might cause cancer, and by 1979 the Army had begun experiments to test the long-term effects of RDX on mice and rats.

Still, even after detecting what the U.S. Centers for Disease Control staff later described in a 1992 report as “high concentrations” of explosives in central Nebraska’s groundwater, the Army kept its problem there quiet. After it idled weapons assembly at the plant in 1981, it continued to let laborers working for a private company that leased part of the property drink from a 4-inch pipe drawing well water at the site.

“All of us drank that water and washed in it, for years and years,” said Dennis Mudloff. He was in his late 20s when he first went to work at the plant in 1981, and he said he felt invincible. He worked around a warehouse called the “nitrate building” — for the explosives that were once processed there.

“We slurped it all day long,” he said of the water.

When the Grand Island newspaper wrote about contamination problems at the base in 1982, the Army minimized the threat and said it could take more than a century for the pollution to reach the city of Grand Island.

But what the Army was learning about RDX was more ominous. RDX, it turned out, does not dissolve in water, quickly degrade underground, or cling to soil particles that can keep it in place and limit its spread. Instead of becoming diluted over distance — like many other pollutants — it remains concentrated, and then travels quickly.

“It’s widely used, persistent, and mobile,” said the EPA’s Craig. “It doesn’t go away.”

By 1983, the Army’s monitoring of wells at Cornhusker showed that a plume of pollution containing RDX was already stretching over some four square miles from the nitrate building, and like a blot of ink spreading on wet paper, it was bearing down on the 35,000 people who lived in Grand Island at the pace of 827 feet each year.

The Army’s medical researchers had come up with their own estimate for how much RDX might be safe for people, and by 1985 their testing showed that water in more than 236 homes near Grand Island contained as much as three times that limit. The Army quickly agreed to build out Grand Island’s municipal drinking water system, taking hundreds of private wells tapped into the aquifer off line. Residents who lived out of reach of city water were advised to drink bottled water.

For the EPA, warning people away from poisoned water sources wasn’t enough.

The agency, formed in 1970, had struggled in its early years to enforce brand new statutes and regulations meant to protect public health and the environment. But by 1985, the EPA had grown muscles, and was newly empowered with a broad array of enforcement powers — hazardous waste laws, the Safe Drinking Water Act, the Clean Water Act and more. It didn’t take long for the EPA to realize the military’s role as environmental offender.

The Pentagon, the agency learned, was responsible for legions of disastrously contaminated sites across the country, sites that by sheer number would soon dwarf the liabilities of any other single entity. There were the artillery testing grounds, packed with unexploded ordnance; the chemical weapons ranges; and the rocket fuel and airplane sites, saturated with solvents and fire retardants. The damage at the sites was so serious that in 1984 the EPA amended the rules of its Superfund cleanup program — the powerful 1980 law that allows federal authorities to take jurisdiction over the highest-priority contamination sites in the country and mandate their cleanup — to include military sites. Then it listed 36 of them.

The emboldened EPA added Cornhusker’s RDX groundwater plume to its list of federal Superfund sites in 1987.

RDX’s threat as a powerful neurotoxin affecting brain and nervous system development was at the same time becoming better understood. American troops in Vietnam had experienced epileptic fits when they ate C-4 explosives containing RDX as a sort of gruesome initiation rite. A Chinese medical journal reported that a person who accidentally ate RDX had died. Then five RDX factory workers in Tennessee also succumbed to seizures after being exposed to high levels of RDX in dust. And in 1986, researchers in Arkansas reported on a 3-year-old child who had violent seizures after he was exposed to RDX on his mother’s work boots.

But it was the experiments the Army had begun to conduct on mammals in 1979 in a medical research lab at Fort Detrick, Maryland, that raised the strongest warnings about what long-term exposure to the explosives could mean. That year, the Pentagon had toxicologists feed high doses of RDX to rats and mice. Then they watched the animals’ response over a period of two years.

This type of research had rarely been done and it posed a question that had no simple answer: How much RDX exposure was too much? Given the potential enforcement powers of the EPA, and the severity of the pollution, “they needed to establish permissible levels in drinking water,” one of the studies’ authors, Barry Levine, said later in a court deposition.

The Army experimented on 850 animals. Of the mice receiving larger doses, roughly half died — so many that researchers lowered the maximum dosage midway through the test. Before they died, as the dosage increased, the RDX made the rats and mice agitated and aggressive, so much so that it became difficult for researchers to differentiate between the lesions — on genitals, eyes and organs — caused by RDX poisoning and those resulting from fighting in the cages. The hearts of the animals became enlarged, and their eyes grew discolored, then opaque.

But it was the response among the female mice that drew the most concern. Of the animals receiving moderate to heavy doses of RDX, one in six had grown tumors on their livers, roughly half of which were malignant — a “statistically significant” and worrisome sign that RDX could mutate genes and cause cancer in people. There was also evidence that both male and female mice developed carcinomas — cancerous tumors — in their lungs.

The studies were never peer reviewed or published. But after the Army shared its final reports with the EPA in 1984, work described as “the gold standard of research” by one EPA scientist, the agency intensified its oversight of explosives pollution.

In 1990, the EPA classified RDX as a “possible human carcinogen,” a formal warning that a contaminant is potentially dangerous and deserves more study. It also revised its suggested limits for how much RDX was safe in water, determining that the level the Army had applied at Cornhusker was 17 times too lenient. Instead, the EPA suggested that only two parts per billion of RDX was safe, giving it a danger score stricter than other deadly pollutants, including benzene and the herbicide atrazine.

Residents of Grand Island believe the growing awareness about RDX came too late. Mudloff began experiencing health symptoms his doctors could not explain. First, he became chronically fatigued. Hanging Christmas lights would lead him to wheeze. Then even minor math — addition and subtraction — proved difficult.

“My mind was getting worse. I could not think,” Mudloff said. Even the Mayo Clinic, which he visited in 2008, was stumped. Then a neurologist diagnosed myoclonus, a spasmodic, jerky contraction of muscles that has forced him to walk with a cane. Still, the diagnosis could not account for what was happening to his brain, or the sluggishness of the muscles in his tongue, which makes him speak as if he’s had a stroke. Mudloff’s doctor blamed his illness on his exposure to military explosives including RDX.

Mudloff’s former co-workers and neighbors, he said, have fared worse. One by one, he said, the men he worked with in the 1980s took disability, suffering from exhaustion and odd mental degradation. One died of kidney cancer, another of lymphoma. These illnesses were never tied to RDX or any other pollution — no one investigated them. But they put the town on edge.

The Agency for Toxic Substances and Disease Registry, a division of the CDC, came to Grand Island to examine widespread reports of illness across the city. It couldn’t pin health effects directly to RDX. The agency’s report was inconclusive about the ways people were exposed, and ambiguous about when and for how long.

But the agency did declare that the pollution posed a serious health threat and made public a new element of the threat: The explosives were likely making their way to people not only via local water supplies, but through meat, fish and produce.

ABOUT AN HOUR SOUTH of Salt Lake City, the Wasatch Mountains transition from rugged peaks to a sloping field of sediment, forming a fertile plain outside the small city of Mapleton, Utah.

There, in the mid-1990s, Marilyn Petersen was planting a garden of squash and beans and tomatoes. Petersen was a member of Mapleton’s city council, and would later become its mayor. At home, she was an avid gardener, tending a large plot outside her stately brick, nine-bedroom ranch home on a quaint country lane lined with aspens and pearly white fences. The neighborhood was wealthy — full of estates with swimming pools and stables and dazzling mountain views. But many of the residents shared a more humble practice: They grew their own food.

Neither Petersen nor her neighbors knew what RDX was at the time. Nor did she know that for the 15 years since she had purchased her home, the water she drank and used to irrigate her plants contained enormous amounts of it — far more of it than was considered safe.

Where the Spanish Fork Canyon spills into the Utah Valley, a short ways south of Petersen’s property, the Trojan explosives plant had been either manufacturing or recycling bomb material since the 1940s. Its latest owner — Connecticut-based Ensign-Bickford — was the recipient of more than 1,600 contracts with the Department of Defense, worth hundreds of millions of dollars.

For at least 20 years — and probably for decades longer — the Trojan plant discharged its waste, including pure RDX, pink water and nitric acid, into ponds and an unlined ditch running downhill from the plant. Near the facility’s outfall, an irrigation canal cuts north along the foothills, using gravity to route Spanish Fork river water north past Mapleton so the river can be tapped by farmers throughout the valley. The water carried RDX, and other chemical compounds associated with or derived from it, toward Mapleton and into the groundwater aquifer beneath Petersen’s property, according to plaintiff’s experts in a lawsuit neighbors filed against Ensign-Bickford.

Petersen began eating vegetables grown on farmlands that had been flood irrigated with water from the ditch in 1980, and drilled a new well into the polluted aquifer in 1983. In 1984, at the age of 41, she was found to have stage IV non-Hodgkin’s lymphoma, a cancer of the immune system.

Soon after, three of her neighbors received the same diagnosis. There was Charles Bates, who lived a block south, and Howard Ruff, a nationally known investor who had advised Presidents Carter and Reagan, who lived six blocks to the east, and Glenn Allman, a Brigham Young University professor, who lived two miles north. A similar cancer that affects white blood cells was diagnosed in Cherie Hunt, who lived two blocks southeast, and 17-year-old Stacy Broadbent, two blocks to the southwest, was found to have lymphoblastic leukemia. Besides being members of the same privileged community, they all grew their own food and they all lived less than 1,500 feet from the irrigation ditch.

For 13 years, Petersen fought her cancer, charging through rounds of chemotherapy and enjoying brief respites of remission only to see her illness rage back. The suspicion that the Trojan explosives plant might be to blame emerged over time. Anecdotally, cancer in the area seemed to be on the rise. The Trojan plant itself had had a reputation for problems — explosions, a deadly accident and a large spill in 1986 that made the local newspaper.

Then, in August 1997, Petersen received a letter from Kendall Robins, Ensign-Bickford’s manager at the Trojan plant, explaining that the company had found significant levels of RDX in the local aquifer. “Please don’t be alarmed,” Robins wrote, assuring her that the levels detected were safe. But if “you use the water from your well for culinary purposes, we ask you to contact us immediately.”

Petersen and the five other families ultimately sued Ensign-Bickford — filing two separate cases — in federal court, blaming the company for their cancers. In the years before the cases were settled, extensive and acrimonious court proceedings hashed out the most thorough investigation into the chemistry and dangers of RDX ever completed. The records from Petersen’s case — including her account, internal Ensign-Bickford documents, and extensive expert testimony — provide a rare window not just into what happened in Utah, but the possible liabilities the Pentagon faced at RDX-contaminated sites elsewhere.

It turns out that Ensign-Bickford, and the companies that operated the Trojan plant before it, were well aware that they had caused extensive groundwater pollution years before they told Marilyn Petersen, and before she ever drew a sip or sprinkled a drop from her well.

As early as 1979, company documents show, Mallinckrodt, the company that operated the plant before Ensign-Bickford acquired it, had detected nitrate contamination related to explosives in water wells offsite. In 1980, after reports of more waste being dumped into ditches and leaking from unlined pits, an EPA official stated that the contaminants that leaked could be expected to contain explosives. By 1981, the company was theorizing that the contaminants were seeping through the ground and into the irrigation ditch.

“I regularly saw the waste water from that operation leave the production building and flow in a downhill ditch,” Lawrence Bradshaw, who worked with RDX production at the plant for 19 years, said in testimony. Bradshaw described poor maintenance and constantly overflowing equipment. “The waste water … contained RDX waste and TNT waste. … It was red in color.”

Then, in 1986, the liner for a 45,000-square-foot, 12-foot-deep waste pond of nitric acid ruptured, sending hundreds of thousands of gallons of acid and other contaminants from the production of the explosives into the soil and the ditch nearby. After the spill, Ensign-Bickford tested the water in its monitoring wells and detected RDX at 10,700 parts per billion — some of the highest levels ever recorded in the environment and about 5,000 times what the EPA thought was safe. A few years later, a consultant hired by the company warned its executives that the contaminants could be consumed through irrigated vegetables.

None of this was disclosed to Petersen or the other residents of Mapleton, who continued to water their fields and drink from their wells.

Instead, the company continued to quietly study its problem, and strategize about ways to minimize the fallout. Ensign-Bickford hired consultants who outlined the enormous liability the company could face, and the “negative publicity” it would undoubtedly endure depending on how the contamination was made public. “A serious threat exists to the long-term water supply of the city of Mapleton,” read handwritten meeting notes from the Trojan Corp., which Ensign-Bickford owned, dated 1990. The notes said Mapleton would be forced to shut its groundwater wells and could lose a significant portion of its water supply, leading to “media attention, increased pressure from regulatory agencies and potential lawsuits.”

“Trojan Corporation,” the notes state, “faces implication as the probable cause.”

By 1994, the company had tested and documented the existence of RDX in the private wells of Petersen and Ruff, having collected samples under the guise of routine water monitoring, but it still hadn’t informed them of the results, according to the plaintiffs’ statements. (In some documents Ensign-Bickford denies it documented unsafe levels of RDX. The company did not respond to requests for comment.) Trojan was able to keep the contamination secret in part because the EPA’s guidelines for RDX concentrations in drinking water were not a legal, enforceable limit. The agency had never specifically formalized regulation of RDX, and so the pollution was not in violation of the law.

As evident as Ensign-Bickford’s cover-up was, however, Petersen and the other plaintiff families still had to prove that the explosives had made them sick. Petersen’s lawyers built their case not on RDX itself, but by focusing on the chemistry that results as RDX degrades in the ground, and how those unusual compounds are then absorbed into plants consumed as food.

What they found casts new light on the nation’s RDX problem, suggesting that even as RDX fades away, the chemicals in it will remain lethal, perhaps becoming even more so as they break down.

In 1999, the plaintiffs hired a professor specializing in microbiology and epidemiology from the University of Nebraska Medical Center to examine how RDX might be connected to the specific types of cancers found in Mapleton. For years, the doctor, Dennis Weisenburger, had studied how non-Hodgkin’s lymphoma develops and how environmental exposures may play a role in it.

Weisenburger examined the chemical breakdown of RDX into groups including what are called hydrazines, and another called nitrosamine compounds. These descendant compounds can be more prevalent and more dangerous than the RDX itself, he warned. He noted that nitrosamines in particular are “some of the most potent carcinogens.” (Formaldehyde, another breakdown product, has since also been classified as carcinogenic.)

Hydrazines were already classified by the EPA as a “probable human carcinogen” based on animal studies. One related breakdown compound is so potent it has long been used to induce colon cancer in rodents for research purposes, often on the basis that a single dose will cause tumors. Another variation has been specifically linked to non-Hodgkin’s lymphoma in mice.

The nitrosamines — and their broader family of related compounds — are more difficult to analyze because there are many, and some are extremely dangerous. More than 300 varieties are believed to cause cancer, Weisenburger noted in his testimony. One type, for example, led to tumors in mice with a single minuscule dose of just one-tenth of a part per billion. The variations derived from RDX are not well studied, but based on his research and published papers, “one should assume that they have carcinogenic potential, until proven otherwise,” Weisenburger wrote.

Both nitrosamine derivatives and hydrazines were found in the water wells of Petersen and the others. And both types of compounds, Weisenburger noted, have been directly linked with the same type of non-Hodgkin’s lymphoma that befell the Mapleton residents.

Years earlier — as concerns about residents’ health mounted — Utah’s health department investigated a potential cancer cluster in Mapleton, but concluded that while leukemia cases appeared unusually common, they could not be linked to pollution from explosives. The agency found no apparent cluster of non-Hodgkin’s lymphoma. Weisenburger argued that the state had diluted the sample pool by including an area far away from the Ensign-Bickford plant, and that it had erroneously divided two types of lymphoma into separate categories when they should have been counted as one. He also found that Petersen and Ruff’s cases had not been counted in the study. When he combined the incidents of chronic lymphocytic leukemia and non-Hodgkin’s lymphoma and added in Petersen and her neighbor, the numbers were stark: Mapleton had twice as many cases of non-Hodgkin’s lymphoma and three times as many cases of leukemia as would have been expected.

Petersen’s legal team believed that she and her neighbors got their cancers by eating food grown with explosives-saturated water and by drinking that water. For years, researchers, including those working for the Pentagon, had been investigating the potential for plants to absorb RDX and other explosives, often with the goal of using those plants to remediate, or suck the contaminants out of polluted soils. That RDX and its chemical derivatives accumulated in plants and fish was well accepted.

“Plants need nitrogen for protein, just like we do,” said Terry McLendon, a former University of Texas professor of biology and an ecological risk assessor who has worked with the Department of Energy on contamination in plants at Los Alamos, and on other federal projects. “And RDX is a nitrogen-based substance.”

But accumulation in plants also meant that the vegetables concentrate the chemicals and amplify the exposure to anyone who eats them far beyond what it would have been if they’d simply swallowed polluted water. The question was how much did the chemicals accumulate, and how much had Mapleton residents consumed?

The group hired McLendon’s consulting company to find out, and the calculations he presented to the court were striking. Based on previous research, he calculated that a carrot grown with water containing the amount of RDX measured in Petersen’s well, for example, would have up to 286 times as much RDX in it as the water it was grown with. Spinach would contain 55 times as much RDX, the researchers found.

The chemicals are most concentrated in the roots and the leaves of plants. So tomatoes might be less of a risk than lettuce or potatoes. Petersen meticulously accounted for her home-grown diet over the years, calculating that she’d eaten at least one potato, on average, each day since she’d moved into her home, as well as thousands of beans, carrots, squash, cucumbers and more. McLendon calculated that she’d consumed more than 100 mg of pure RDX and its derivatives each year.

The defense team criticized Weisenburger’s and McLendon’s conclusions for making what it described as theoretical leaps beyond what could be proven. But before the case was settled in 2002, the arguments made in the Mapleton lawsuit were subjected to a special kind of scrutiny: The federal district court judge held a so-called Daubert hearing to evaluate the credentials of the experts and the scientific approach they took in offering their testimony. And after weeks of deliberation and an extensive review of the evidence, the judge validated the plaintiffs’ approach, ruling that their experts’ analysis that hydrazines and nitrosamine compounds were likely to cause non-Hodgkin’s lymphoma were valid, and that the methods used to quantify residents’ exposure were legitimate.

It takes significant evidence to meet that legal standard, said Laurie Ashton, the attorney who represented the Mapleton families. “You can’t just be a crank,” she said.

What appeared to be missing, though, was more conclusive research along the lines of the studies that Levine had done for the Army in 1984, a controlled experiment exposing live mice and rats to RDX doses. Levine, deposed in the Mapleton case, said the Pentagon was extremely unlikely to ever sponsor such a study again.

Ensign-Bickford’s settlement with the six families was for an undisclosed amount of money — Rodney Petersen, Marilyn’s husband, said their share amounted to $1.8 million — but the company did not acknowledge responsibility for any harm they experienced. Marilyn Petersen died of her illness in 2004. Glenn Allman, Charles Bates and Cherie Hunt died before the lawsuit was settled.

“We have seven kids and they were also exposed,” Rodney Petersen told ProPublica during a recent visit to his home near the Spanish Fork Canyon. “So the end of the story may not be written yet.”

CAPE COD IN MASSACHUSETTS depends on one large aquifer for the vast majority of its drinking water. That underground body of water is shaped like a lens, with a high point that approaches the surface at its peak. There, at its most vulnerable, it is separated from the grassy fields and thick woods of the western Cape by a relatively thin layer of porous sand.

Directly above this high point in the aquifer is the 14,000-acre target area of a vast firing range where Army National Guard troops have for decades launched ordnance and munitions, testing their power and precision on the training fields of the Massachusetts Military Reservation.

In the mid-1990s, the EPA discovered RDX in that single source of drinking water, a supply that supports more than 520,000 people during the Cape’s busiest summer months.

The roots of the Massachusetts Military Reservation — as the sprawling 21,000-acre military bases of Camp Edwards and Otis Airfield were called — date back to 1911, when the Cape was remote and far less inhabited.

Now known as Joint Base Cape Cod, the bases have served as critical training grounds for the Pentagon since World War II. The southern half has been overseen by the Air Force, which ran an experimental aeronautical missile program there in the 1960s and classified nuclear missile preparations in the 1970s. Meanwhile, the Army has long used the ranges of the northern base for artillery practice.

The bases also have had confusing lines of ownership and authority. They were once federal land, then were transferred to the state for its National Guard, but with portions, including the test ranges, leased back by the Army. That, in turn, muddled the question of which jurisdiction’s environmental laws apply and who enforces them.

“The net result is an atmosphere in which no one knows who is in charge,” Lt. Col. William FitzPatrick wrote in an analysis for the Army Environmental Policy Institute in 2001.

Serious environmental problems had been a concern at the bases prior to the EPA’s RDX finding. Otis Airfield was declared a Superfund site in 1989. And for decades the people of Cape Cod wondered which environmental blight — from pesticides to smog — might be responsible for high rates of cancers there. In 1991, the Boston University School of Public Health even noted the risk of air pollution around Camp Edwards’ bombing ranges, finding that lung and breast cancers near there were unusually common.

The prospect that the Army’s firing ranges were contaminating the drinking water posed a new problem, one with significant implications. Not only were far more people at risk than in any of the other known cases of RDX groundwater contamination, but in Massachusetts, RDX couldn’t be written off as coming from messy, outdated practices at the nation’s weapons manufacturing sites. It appeared to be coming from active training that was still underway.

One thing stood in the way of the EPA finding out for sure: Because of the continuing Superfund cleanup at Otis Airfield, the EPA and the Army had already legally agreed to the boundaries of the cleanup area long ago, as the law requires. Camp Edwards’ firing ranges and the new RDX contamination lay outside that cleanup boundary. The Army refused to voluntarily expand the boundaries, arguing that the ranges were still active, and that an environmental investigation would interrupt training of troops. The RDX coming from active ranges could be subject to different environmental laws, but those wouldn’t apply until the firing ranges were closed — something EPA lawyers say the Army postponed even for areas that hadn’t seen action in decades — or until the contaminants seeped off the military’s land and actually began dripping out of people’s taps.

As a result, the Army had kept the EPA from testing on the northern part of the base’s ranges for years. And when confronted with mounting concerns about the RDX in groundwater there, the Army sought to keep it that way.

“They were using every argument in the book,” said one EPA official familiar with cleanup effort at Camp Edwards.

Such friction was hardly new. The Pentagon had for years argued that “sovereign immunity” clauses protected the military from prosecution and fines under environmental law. In the spring of 1992, the Supreme Court ruled on a similar argument by the Department of Energy, confirming that the Pentagon was correct; environmental agencies couldn’t fine it for past violations of hazardous waste statutes. The move took the teeth out of enforcement at both the federal and the state level, because without fines there were no consequences for refusing to clean up contaminated sites.

Congress ultimately responded a few months later with the 1992 Federal Facilities Compliance Act, restoring a large portion of the EPA’s oversight authority over federal agencies’ handling of hazardous waste. But it didn’t come easy.

“DOD fought it, for six years, over three Congresses,” said Richard Frandsen, who served as chief counsel for environment issues for the House Energy and Commerce Committee for more than 30 years. “We had all 50 attorneys general, Republican and Democrats, and we finally got it passed.”

Ever since, the relationship between the Pentagon and the EPA has remained uneasy, often dysfunctional.

In 1997, the EPA began pressing the Army to measure the amount of RDX and other contaminants in the groundwater near its firing ranges. EPA officials hoped to identify the source of the contamination and asked the Army to survey its grounds for unexploded munitions. According to an internal EPA document obtained by ProPublica, the “EPA received no meaningful response.”

The EPA didn’t have military expertise or much knowledge of the physics and chemistry of explosives; it depended on Pentagon officials for such information. In what EPA officials viewed as an effort to quell their concerns, the then head of the Department of Defense’s explosive safety board filed a technical memo with the agency promising that the risk of contamination from unexploded munitions was “virtually zero.” The memo said that unexploded shells lying in the ground don’t leak or break, that they are essentially secure “museum quality” vaults of metal. For a short time, the document convinced EPA senior managers that the RDX must have come from a different source, and the agency relaxed its order.

But soon after, EPA lawyers came to believe that the Army had deceived them. Another Army weapons expert told the EPA that the Army had been studying the fate of buried munitions and knew at the time the memo was prepared that unexploded ordnance in fact does break down and ultimately leak its contaminants.

“They knew it was false,” said William Frank, a senior EPA attorney who worked in enforcement over Department of Defense facilities, including at Camp Edwards, for 25 years before his retirement. The technical memo “was given to EPA to influence its implementation and oversight in an enforcement matter,” Frank said. A short while later, Army officials would discover a large quantity of “extremely corroded” ordnance that included a rotten, rusted 155 mm shell on the Massachusetts fields with a hole worn through and raw RDX exposed to the open air.

To federal environmental officials — and even some defense experts — the Pentagon’s efforts at deception and delay seemed to reflect concern over the broader, national implications of the pollution. If RDX was indeed spreading in the water beneath a test bombing range, it meant the scope of the Pentagon’s RDX liabilities across the country could be monumentally larger than previously thought. Munitions containing RDX had been fired on ranges at thousands of American military sites; at many of those sites old unexploded munitions were still strewn across the land, slowly degrading over time.

“They don’t want to get stuck with the bill,” said Rick Stauber, a leading munitions expert who has worked under contract with the Army to identify unexploded ordnance for most of the last three decades, including in Massachusetts.

A few months later the EPA issued a second order forcing the Army to suspend its live-fire exercises while it investigated the groundwater plume. The agency’s regional administrator at the time, John DeVillars, announced in May 1997 that the RDX pollution in the Cape’s aquifer was an “imminent and substantial endangerment” to the public, an unusual declaration that gave the EPA jurisdiction under the Safe Drinking Water Act, and the emergency authority to stop dangerous pollution still taking place, regardless of its nature or cause.

DeVillars didn’t mind taking on an interagency battle and was of the mind that “the bigger the opponent, the bigger the fight,” according to colleagues.

His move outraged Pentagon officials, who felt the EPA had overstepped its bounds. Generals and high-ranking EPA officials, including the agency’s assistant administrator, convened in a Pentagon conference room in Arlington to hear the Army’s pleas. But in the end, the EPA upheld DeVillars’ “imminent endangerment” order.

The issue had by then escalated beyond environmental concerns. With the interruption of Army National Guard training, the core missions of the Defense Department and the EPA were colliding. The Army asserted that any restrictions on training would endanger soldiers and compromise national security. If the EPA could disrupt training for environmental reasons in Massachusetts, it could interfere with the Pentagon’s mission anywhere on U.S. soil.

“There was zero trust, and it was because of what I was wearing, the camouflage,” said Lt. Col. Joseph Knott, who ran the Army’s cleanup at Cape Cod once it was underway. “The military is like, ‘We’ve got to train our soldiers so when they go fight, they come home.’ That seems pretty obvious.”

From the EPA’s perspective, the spread of RDX directly threatened the health of the very Americans the Army was fighting to protect overseas.

“They felt like EPA was encroaching on their mission of protecting the country,” said one senior EPA official with knowledge of the clash. “We felt our mission was more important than their mission.”

The environmental testing the EPA had pressed for, in the meantime, had yielded the results the agency had most feared. Nineteen separate groundwater plumes emanated from 10 distinct ordnance targets or gun positions on Camp Edwards, spanning much of the 14,000 acres. RDX levels were as high as 43,000 parts per billion in the soil, 21,000 times greater than the EPA’s advised lifetime human health exposure threshold and four times the highest levels ever detected near Mapleton, Utah. Perchlorate, a rocket fuel, as well as lead, cadmium and other significant explosives such as TNT, nitroglycerin and HMX — a byproduct explosive of RDX — were also spreading from the target ranges.

In firing munitions, as many as one in five fail to detonate, and the EPA believed that a significant amount of the contamination was coming from unexploded bombs that eventually leaked their contents. EPA scientists were tracking the plumes as they moved from the base. The drinking water in towns including Mashpee, Bourne, even one day Barnstable were all at risk.

Still, according to internal EPA legal documents, the Army National Guard “was unwilling” to clean up the contamination. The disagreement boiled down to a matter of perception of risk, since the contamination, at the time, remained within the bounds of Camp Edwards.

The resistance was based on concerns that bowing to EPA authority would set a precedent for interrupting military training at sites across the country, as well as fears about what cleaning up the groundwater beneath a site like Camp Edwards might cost, according to several military officials and contractors familiar with the confrontation. A cleanup would be a decades-long process, involving scraping millions of tons of soil off the land, and then pumping water from underground, treating it, and reinjecting it into the aquifer.

Every year, the Pentagon gets $3 billion to $4 billion in funding from Congress to support its environmental work. But that money is specifically targeted for work under the Superfund and hazardous waste laws designated by the Federal Facilities Compliance Act of 1992. When the EPA declared Cape Cod’s water problem an “imminent and substantial endangerment” under the Safe Drinking Water Act, it triggered an entirely different form of response, one that had to be paid for directly out of the base’s operations budget. Cleanup funds were scraped together out of monies meant for the bases’ water and fire services, power generation and maintenance. “Some of these other pieces of the pie were going to get cut back to put the cleanup on steroids,” Knott said.

In early 2000, the EPA issued its most substantial demands in a third order, laying out a specific plan for cleaning up the sites. It involved gathering and then safely disposing of unexploded munitions material, and gradually treating contaminated groundwater to reduce the size of the plumes. Though that cleanup has since progressed with the Army’s cooperation and has slowed the spread of contaminants, RDX ultimately made its way into public groundwater beyond the base’s boundaries in 2010. In 2011 the Army announced a plan to scrub the aquifer, systematically pumping that groundwater up, treating it, and injecting it back underground until the RDX and other military-related contaminants are nearly gone. That process continues today.

The Pentagon lost its battle with the EPA at the Massachusetts Military Reservation. But it came away with a renewed commitment to fend off liability for RDX contamination.

In a speech delivered at an explosives safety seminar in 2000, the Army’s then deputy assistant secretary for the environment, Raymond Fatz, said that at least 20 other military bombing ranges in the U.S. lay directly over sole-source drinking water aquifers, just like on Cape Cod, and warned that because RDX would be so prevalent at these sites, “this has the potential of being a huge problem.”

Pentagon officials feared that if it continued to be held liable under federal environmental regulations, its cleanup costs could bankrupt the Defense Department’s environmental programs, according to the accounts of several lawyers and environmental contractors who have worked with the Pentagon. A typical DOD cleanup cost around $15 million to $22 million. The cleanup at the Massachusetts Military Reservation alone — which continues today — will wind up costing the Pentagon nearly $1 billion.

The military’s budget, though huge, was not infinite and Pentagon officials also worried that protecting the environment would get in the way of training and protecting troops.

“The only other option they had was to go to Congress and try to change the statute,” said Frank.

And that is what the Pentagon did.

As the U.S. went to war in Iraq in 2003, top Pentagon officials saw a political climate that might yield the ultimate reprieve. Defense Secretary Donald Rumsfeld, with a small group of Pentagon lawyers, argued that in order to preserve “readiness” for the fight against terrorism, the Pentagon would need to be freed once and for all from EPA oversight. One of the nation’s largest polluters should be allowed to operate largely outside of the scope of the law, especially, the Pentagon’s lawyers noted, when it came to pollution from explosive chemicals, which included RDX.

“There were those in the department who thought that this was a great opportunity to try to remove environmental impediments,” said Robert Taylor, a former general counsel to the Department of Defense who describes himself as a reluctant architect of the strategy.

The Pentagon proposed amendments to six of the nation’s most important environmental laws that would largely exempt Defense Department lands with ongoing operations from regulation: the Comprehensive Environmental Response, Compensation and Liability Act, or Superfund; the Resource Conservation and Recovery Act, which governs hazardous waste; the Clean Air Act; the Marine Mammal Protection Act; the Migratory Bird Treaty Act; and the Endangered Species Act. The Pentagon negotiated over five specific pollutants, including RDX and perchlorate. The final language of the bill specifically named “munitions constituents,” which include all chemicals from explosives, as substances it sought to redefine as noncontaminants outside the scope of waste handling.

“It was stunning, what they were trying to do,” Frandsen said. “They were basically trying to take all enforcement away.”

The Pentagon argued that after the 9/11 attacks, environmental cleanups would come at the expense of the safety of U.S. forces in Iraq and Afghanistan.

“Preparing America’s military forces for battle,” Raymond DuBois, the deputy undersecretary of defense for installations and the environment, told a House energy and commerce subpanel in 2004, “is critical.”

Others were more hyperbolic, arguing that exemptions were necessary for the sake of “winning the war on terror,” and “protecting Americans from deliberate attacks that would kill millions of our fellow citizens,” as Rep. Christopher Cox, R-California, then chair of the Select Committee on Homeland Security, said.

But this strategy, too, largely failed.

Since exempting the Department of Defense from hazardous waste laws effectively would have transferred the cleanup burden to local businesses and municipalities attempting to turn old military lands into economically productive places, not even reliable Pentagon allies in Congress felt able to support such a step. There were more than 5,000 contaminated military installations and 900 Superfund sites linked to military operations nationwide — enough that almost all members of Congress had such sites in their districts. They rejected the Pentagon’s pursuit of exemptions from the major water and waste statutes.

John Dingell, the longtime Democratic congressman from Michigan who chaired the House committee on energy and commerce and the committee on investigations, and was ranking member during the exemption hearings, put it simply:

“Nowhere has a single set of legislative proposals had so much audacity — and so little merit.”

IN 2012, 22 years after it first issued a cancer warning for RDX, the EPA launched a toxicological review to re-examine the risk the chemical posed to people.

To health professionals observing the process, the agency seemed poised to strengthen its cancer warning for RDX. Using Barry Levine’s original study, the EPA had already calculated what it calls a cancer slope factor for RDX, an in-depth quantitative process that attempts to predict the specific dose of a chemical that will cause cancer and which is normally done only for a chemical already believed to pose a serious cancer threat.

In 1998, the agency had added RDX to a list of candidate contaminants for concern — a regulatory status that reflects a risk to public drinking water and is often a precursor to regulation. In 2008, it listed RDX among contaminants to be monitored by water utilities across the country. And as it released formal documents to begin its review in 2013, the EPA summarized its case for RDX’s cancer-causing properties starkly: In the best studies ever conducted, RDX was linked to two different types of cancer, in two sexes among two different species of animals. That checked every box in the agency’s formal criteria for classifying a toxin as a “likely” human carcinogen.

The EPA has a lengthy formal process for reviewing toxins, and it’s aimed at identifying which substances are dangerous to humans and how much people can be exposed to. First, EPA scientists assemble a dossier that includes a thorough assessment of published research, a literature review that can consider hundreds of papers, and a proposed methodology for weighing the evidence. Then those documents are made public, and agency officials incorporate the feedback as they write early internal drafts.

The EPA then shares a more developed draft with other federal agencies — in this case including the Department of Defense — and revises the document further based on their response. Only then is a version released to the public, and only then does an outside advisory committee of experts get to review the draft. The process is repeated, including public comment periods, with the outside scientists submitting criticism before the EPA issues its final assessment.

The Department of Defense told ProPublica that it had anticipated the EPA’s review long before it was launched. By the time the EPA released its initial dossier in 2013, the Pentagon had already funded a number of new scientific studies that raised and amplified doubts about whether RDX caused cancer, or posed any health threat. One of these studies was even paid for out of the very same Pentagon environmental cleanup program responsible for addressing some of the lands contaminated with RDX. These studies made up much of the new research the EPA considered.

One article argued that the models used in peer-reviewed papers to predict how much RDX persists in the organs of mice that ingest it were not reliable. Another showed that mice genes didn’t mutate after RDX exposure, suggesting cancer was a less likely threat.

“Looking at the merits of the science was absolutely something that the department felt that it should do to make sure that the science was well developed and objective,” said Taylor, the former Pentagon general counsel.

In 2006, Gunda Reddy, an Army Ph.D. toxicologist working at Aberdeen Proving Ground in Maryland, was part of a team that re-examined Levine’s original RDX study using the original data and samples. Reddy concluded that the increase in cancer Levine observed was less pronounced than previously thought. Reddy, who has published nearly 100 scientific papers, called the evidence of a cancer risk “equivocal.”

Separately, Reddy force-fed RDX to baby pigs to learn how they digested the explosive and then co-published a study that found that RDX had not accumulated in the pigs’ livers, prompting skepticism that it could cause tumors there. In 2011, a researcher from the Naval Medical Research Unit raised doubts about whether RDX was the cause of noncancerous prostate inflammation observed in rats. Though she, too, had merely reanalyzed old data, she concluded that RDX could be safely ingested by people in far higher doses than the EPA had suggested.

That these studies were funded by the Pentagon, which had a specific stake in how RDX was classified, did not make the research less trustworthy, EPA toxicologists and others involved in the review said. The EPA’s Craig points out the studies helped fill a void in overall knowledge about RDX’s effects — especially their noncancerous effects — and several were validated through peer review.

But scientists directly involved in the review process said the Pentagon shaped the outcome of its studies by the questions its researchers asked, or chose not to ask. The persistent focus of the Pentagon studies on uncertainties in existing research, their aversion to repeating the original rodent studies, and the consistent findings that RDX was less dangerous than previously thought, sparked skepticism.

“You can always get a burger your way,” said Frank, the former EPA federal facilities enforcement attorney. “The scientific process itself has been under attack by the Department of Defense for many years.”

Just as Levine predicted in his testimony in the Utah case, the Defense Department researchers never repeated the original graduated-dose study of RDX on mice to observe whether it caused malignant liver and lung tumors in higher numbers.

Reddy, for his part, wrote in a PowerPoint presentation he and two co-authors made about the EPA review for the Army Public Health Center that if the EPA didn’t loosen its RDX standards based on his research, “training and testing activities will be adversely affected, adversely affecting military readiness.” Furthermore, if “artificially low” environmental standards were set for RDX, “significant resources will be spent [on] cleanup costs associated with unnecessary remediation.”

Military assessments from scientists trained in toxicology, who are not supposed to have an interest in the outcome of their research, are “frankly, highly unusual,’’ said the EPA’s Craig. Reddy did not respond to a request for comment about his statements; it is unclear how he determined what effect changing RDX standards would have on military readiness.

In an emailed response to questions, Mark Johnson, director of toxicology for the Army Public Health Center at Aberdeen Proving Ground, said the notion that the Pentagon had tailored its research on RDX to sway the EPA’s review — or that it was part of an agency-wide campaign to avoid responsibility for RDX — was “simplistic’’ and incorrect.

“There has never been a strategy to manipulate RDX regulation by deciding not to do studies,” he wrote. “Our interest has always been to provide as much information as possible so that the best possible science can be used to reduce uncertainty associated with establishing safe levels of exposure for decision-making.”

J.C. King, the Army’s director of munitions and the chief official responsible for Army explosives cleanups, said the Army simply wants to make sure taxpayer funds are not wasted on unnecessary environmental work.

“That’s our obligation,” he said in an interview at the Pentagon in July. “We’re spending your money. And we spend your money as wisely as we can.”

The Pentagon-sponsored research bolstered the military’s longstanding argument that the health and environmental dangers of RDX are unproven.

By the time the EPA completed the first internal draft of its RDX review in 2014, meant for deliberation between executive agencies, it indicated it was considering a looser and more ambiguous categorization that is less likely to lead to stringent regulation. Instead of declaring RDX a likely carcinogen, it was now prepared to say RDX was merely “suggestive” of carcinogenicity.

Two years later, when that assessment was finally released for review by the science advisory committee — made up of 26 prominent toxicologists, epidemiologists and cancer doctors from American universities — the less serious characterization of the risk immediately raised concerns.

“Why lower cancer risk designation?” asked George Cobb, a member of the EPA’s RDX advisory board and chairman of the environmental science department at Baylor University. Cobb wrote in late 2016 comments that the liver and lung tumors in the mice found by Levine nearly three decades ago alone warranted the more serious warning. “The hazard identification should be … indicative of higher risk,” he wrote.

“The case for this classification presented in the document is not strong,” wrote Stephen Roberts, another member of the RDX advisory board and a professor of public health at the University of Florida. The “likely” characterization, he wrote in 2016, “fits the data for RDX.”

The EPA contends that its overall rating of RDX is more or less unchanged, and that the term “suggestive” reflects the current state of the research.

“Choosing a descriptor is a matter of judgment and cannot be reduced to a formula,” an EPA spokesperson wrote to ProPublica in response to questions emailed to the agency. “EPA’s conclusions are driven by the scientific evidence and risk assessment methods available at the time of assessment development. The strength, reputation, and influence of the [EPA’s environmental risk program] is founded on its scientific integrity, highest caliber of scientific process, and rigorous peer review.”

The EPA declined to allow its advisory committee members to speak with ProPublica. The EPA also has not responded to three public records requests about the board’s meetings and communications, the first of which ProPublica filed with the agency more than 13 months ago.

But other public comments and records from the EPA’s RDX meetings show that from 2013 to 2016, the Pentagon and organizations and agencies friendly to it, including the American Chemical Counsel (an advocacy group that generally lobbies against chemical regulations) and the Office of Management and Budget (which approves the Pentagon’s cleanup spending) either recommended looser standards, pushed the EPA to include more studies that found no negative effects from RDX, or advocated a “weight of evidence” approach.

This would mean considering all studies on the subject and giving each more or less equal weight, regardless of their quality. Since the Pentagon had funded so many studies that found RDX to have less health risk, this approach was more likely to lead the EPA to downgrade RDX’s risk profile.

The EPA, in a statement to ProPublica, insists that the quality of studies is not ignored in this approach. But critics argue that the scales are better balanced when the EPA evaluates the research and allows the best studies to have the most influence — or at least protects them from being dismissed as aberrations.

“If you base it on weight of evidence, and you stack it with a bunch of negative studies, you are going to win,” Melnick, the former NIH toxicologist who submitted comments to the EPA (and does not sit on its advisory committee), told ProPublica. “even if the negative studies are not very good.”

In his formal comments submitted to the EPA, Melnick called the agency’s representation of the Army’s research “a misleading justification” for its decision to downgrade, warning the decision will “protect polluters rather than protecting U.S. citizens.”

Of particular concern to Cobb are the known cancer risks of RDX’s breakdown chemistry, the nitrosamines that were the subject of the Petersen lawsuit in Mapleton, Utah. The EPA’s review for RDX, he noted, hardly takes them into account. In the comments he submitted to the agency, Cobb described the Pentagon-funded research as “rehashes of old studies in attempts to decrease toxicity profiles of RDX.”

The only reliable way to answer the cancer question, Cobb insists, is for the EPA or the National Institutes of Health to fund and conduct their own repeat of Levine’s original 1984 research. “This should be done before any diminution of toxicity characteristics,” Cobb wrote in 2016.

Pentagon officials say they never repeated the study because of how much it would cost. According to Johnson, such a study “could be worthwhile” but would take five years to complete and cost $2 million to $3 million. “There has never been a strategy to not repeat” Levine’s studies, he wrote to ProPublica. But “currently there is no funding to support it,” he said, in the Pentagon’s annual budget, which has reached $585 billion.

In September, the EPA’s peer review advisory committee submitted its final comments for the EPA’s consideration. The committee consented to the EPA’s “suggestive” cancer hazard description, but raised numerous other concerns about the potential for underestimating RDX’s risk.

In a letter addressed to EPA Administrator Scott Pruitt, the committee warned that the EPA was overstating its confidence in some of the data, and recommended an “uncertainty” score three times higher than the EPA’s figure, noting the risk, for example, that “repeated exposures to RDX have cumulative effects” on brain function.

The committee criticized the completeness of the EPA’s data on RDX, saying that it “does not capture all of the potential adverse outcomes, or their severity.” It warned that the EPA’s suggested human exposure limits might not account for the fact that even low doses of RDX could cause behavioral and developmental problems, and didn’t consider other factors — including that the offspring of rats exposed to RDX also showed traces of RDX in their brains.

The EPA is now incorporating the advisory committee’s comments into a final assessment of RDX’s toxicity, which is not expected to be made public until sometime next year.

Jane Caldwell, a senior environmental health scientist in the EPA’s toxicology program until she retired last year, said she had concerns with the EPA’s handling of the RDX question dating back years. Indeed, in 2014, she wrote a memo describing the agency’s reassessment work as recklessly incomplete, saying it had glossed over or ignored important signs that RDX caused rare tumors and carried serious health risks.

In an interview this month, Caldwell said the EPA’s decision to “downgrade,” as she put it, RDX’s cancer status was ultimately a political choice. When it came to RDX, she believed senior agency staff overrode the agency’s toxicologists “to try to avoid pressure from the DOD.”

“It was obvious that it should have been one way, and all of a sudden it went another way,” said Caldwell, who worked at the EPA for 26 years. “The understanding was that from the top of the organization, probably with DOD influence, they had chickened out.”

One EPA scientist directly involved in the final stages of the review process offered a detailed sequence of events before what Caldwell called the EPA’s final capitulation.

The scientist, speaking on condition of anonymity for fear of reprisals from EPA management, said that at least 10 EPA staff toxicologists and statisticians had raised concerns about the significance of rare tumors in the RDX studies, and had made a case for calling RDX a “likely” carcinogen. And upon hearing those concerns, in a high-level meeting at EPA headquarters before the public review draft was released, a dozen EPA branch chiefs and managers reached a consensus to take that step.

Then, a few days later, the decision was reversed.

“DOD wasn’t going to let it go without a fight, and the EPA wasn’t interested in that fight,” said the scientist.

Sitting at a dark cherry wood table in a small conference room outside her office at the Pentagon one day last July, Maureen Sullivan, the deputy assistant secretary of defense for environment, safety and occupational health, would not say whether the Pentagon was happy with the EPA’s conclusions or if she had pressed for a specific outcome.

Though she is the top official overseeing the agency’s tens of thousands of cleanup sites, and administering a $4 billion annual environment budget, she said that she hadn’t given the implications of RDX regulation much thought.

“Honestly, I haven’t asked my staff to do an assessment of that, of what it means to us,” she said.

Coda

IN 1982, JOHN SHEEHAN, one of the scientists responsible for developing RDX, received a kind of lifetime achievement award from the University of Cincinnati and the American Chemical Society. For decades, Sheehan had been haunted by misgivings about RDX’s lasting consequences.

To him, his invention had been a “mixed blessing.” RDX had helped win wars, but it had also helped increase the human toll of those wars, among combatants and civilians alike.

“We tend to assume that we can contain the destructive effects of new weaponry more than history justifies,” Sheehan said in accepting his award.

The chance that RDX’s destructive effects could include damage to the American environment was just emerging when Sheehan gave his speech. Thirty-five years later, the EPA is still wrestling with just how much of an environmental and health peril RDX poses and how to regulate it.

But the EPA is an agency undergoing a radical remaking. In the 11 months since President Donald Trump was inaugurated, the EPA has dramatically scaled back its role as a regulator of dangerous chemicals of all sorts. The changes suggest the agency is less likely to move aggressively to take on Defense Department pollution, including RDX, former agency officials say. Pruitt is considering a relaxation of cleanup standards at Defense-related Superfund sites and chose Nancy Beck — a former American Chemistry Council executive who took the Pentagon’s side on RDX’s risk level — to lead the EPA office that will determine RDX regulations.

For much of the Obama administration, Mathy Stanislaus oversaw the EPA division tasked with hazardous waste management and restoration of lands once used as military facilities. He grew familiar with the Pentagon’s power and effectiveness in fending off EPA efforts to fix toxic leftovers related to its explosive weapons programs, including RDX.

In the current environment, Stanislaus said, a fight that once was difficult might now be impossible.

“What leverage does EPA have to move forward?” asked Stanislaus. “I would say it’s very limited.”