BETRAYED BY THE VA Thomas Alan Bowman Sr. asked his country for help. He got silence. Sign the demand →
Dear Doug Collins

The Mustard Gas Test Subjects: Gassed by Their Own Military, Sworn to Secrecy, and Then Asked to Prove It

They were American soldiers and sailors, most of them teenagers, and the gas that burned them was American. Between 1942 and 1945, the Army and Navy used more than 60,000 servicemen as human subjects in secret chemical-defense tests of protective clothing, ointments, and gas masks, exposing them to mustard gas and Lewisite. At least 4,000 of them went through the most severe tests — locked in gas chambers, or sent into fields saturated with the agents. The experiments were kept off their service records, and the men were ordered never to speak of them, under threat of dishonorable discharge and military prison. No follow-up care was provided, although the long-term harm was already known by 1933. This page is that record — the tests, the 1993 report, the 1994 regulation, the promise kept for only 610 of some 4,000, and the last remedy in 2017.

The service and the promise

The Second World War has been called, in the Institute of Medicine’s phrase, “the unfought chemical war.” Both sides stockpiled chemical weapons and prepared extensively, though they were never used in battle between the great powers. The United States knew these agents from the First World War: sulfur mustard, the blister agent of 1917 and 1918, and Lewisite, its arsenic-containing counterpart. If enemy gas came, troops would need gear that worked — and the services concluded animal studies were no substitute for testing it on men.

VA describes it plainly: the Defense Department recruited “volunteer soldier” subjects to evaluate protective clothing, ointments, and equipment; in 1942, VA’s history records, program managers took formal authority to recruit and use them.

What “volunteer” meant is part of the record. The Institute of Medicine committee that examined the program in 1993 wrote that although the subjects were called “volunteers,” it was clear from the official reports that recruitment “was accomplished through lies and half-truths.” Men were offered inducements — extra leave was the common currency — and told partial stories or none; some were told they would test summer clothing. The soldier’s basic bargain was broken at recruitment: they were not told what they were being exposed to, or that it would afterward be treated as though it never happened.

What happened

By war’s end, over 60,000 servicemen had been used as human subjects (the Institute of Medicine’s figure; VA gives nearly 60,000), across a wide range of exposures:

At least 4,000, the Institute of Medicine found, underwent high-concentration chamber or field tests — the full-body exposures, some without protective clothing — at more than a dozen Army and Navy locations.

Three accounts stand for the rest. Nathan Schnurman was a 17-year-old sailor in 1944, recruited — per his Senate testimony — to test Navy summer clothing for a three-day pass. He was instead locked in a gas chamber and exposed to mustard gas and Lewisite. During his sixth exposure, violently ill, he twice asked over the intercom to be released; the answer, he testified, was no — the experiment was not complete. He passed out in the chamber. Charlie Cavell was 19 in 1945 when, NPR reported, he volunteered for a “special program” in exchange for two weeks’ vacation. He and eleven others were locked in a heated chamber at the Naval Research Laboratory while mustard agent was piped in, the humidity raised to intensify its effects. After an hour, six men were released; Cavell and five others were kept a second hour, then ordered back to barracks still wearing their gas-saturated uniforms. Blisters rose where he sweated most, to the size of half-dollar coins. “We had ceased to be volunteers,” he told NPR. Rollins Edwards, an African-American Army soldier, was led with a dozen others into a wooden gas chamber where mustard gas and Lewisite were piped in. “It felt like you were on fire,” he told NPR. “Guys started screaming and hollering and trying to break out. And then some of the guys fainted.” Edwards also crawled daily through fields saturated with mustard gas. Seventy years later his skin still broke out and fell off in flakes where he had been burned; for years he carried a jar of the flakes, to prove what had happened to him.

Edwards’s presence there was not incidental: NPR’s 2015 investigation found evidence that some experiments grouped subjects by race: African-American soldiers tested to see, in Edwards’s words, “what effect these gases would have on black skins”; Japanese-American soldiers used as proxies for Japanese troops; Puerto Rican soldiers singled out, including in Panama; white enlisted men used as controls. A study NPR obtained under the Freedom of Information Act described 39 Japanese-American and 40 white soldiers exposed over twenty days in the spring of 1944. The Pentagon acknowledged NPR’s findings in 2015; it had never before spoken about the race-grouped tests.

Then came the part that governed the next fifty years. All of the men in the chamber and field tests, and some in the patch tests, were told never to reveal the nature of the experiments — secrecy sworn under threat of dishonorable discharge and military prison (Cavell named Fort Leavenworth), per NPR’s reporting — and the experiments were kept off their official military records, leaving most with no proof. Underneath, the records were as thin as the secrecy was thick: surname-only listings for some 2,900 Naval Research Laboratory chamber subjects (GAO, 1993), and Army records not kept to identify participants.

And there was no follow-up — none. The Institute of Medicine stated it in its strongest terms: “Most appalling was the fact that no formal long-term follow-up medical care or monitoring was provided for any of the WWII human subjects, despite knowledge available by 1933 that mustard agents and Lewisite could produce long-term debilitating health problems.” Chamber subjects were sworn to secrecy and released on leave, some still blistered, without instructions for care. The committee’s causal findings — cancers of the nasopharynx, larynx, lung, and skin; chronic respiratory disease; serious eye disease; acute nonlymphocytic leukemia for nitrogen mustard — are set out in the ledger below. The same absence of follow-up extended to the Bari harbor casualties in Italy — a separate story in this section.

What the institution did afterward

For roughly forty-five years, the institution did essentially nothing, because the men kept the secret — almost to a man, for forty years or more (Institute of Medicine). Those who did file claims met the secrecy coming back at them: the GAO recorded that until 1992, VA had approved only 13 of 145 claims arising from the tests.

The break came from the subjects themselves. Schnurman sued in federal court in 1979; the suit failed, under a legal doctrine that, as NPR summarized it, protects the military from suits for service-related injuries. He and his wife, Joy, spent two decades compiling evidence. In 1990, Schnurman’s group went public (NPR). In 1991, federal officials for the first time admitted the military had conducted mustard gas experiments on enlisted men. VA Deputy Secretary Anthony Principi said on CBS’s 60 Minutes that year, as NPR quoted him: “They should be praised for what they did. And as far as the secretary and I are concerned, we’re about to do right by them.” VA also loosened its documentation requirements that June and commissioned the Institute of Medicine review, published as Veterans at Risk in 1993. Declassification was layered rather than a single act — many documents declassified shortly after the war, others restricted for decades — and on March 9, 1993, Deputy Secretary of Defense William J. Perry released the men from their secrecy vows. The reckoning, when it came, came in a cluster:

When the experiments were declassified in the 1990s, NPR reported, VA promised to locate about 4,000 men used in the most extreme tests and compensate those permanently injured. In more than twenty years, VA attempted to reach just 610 of them, with a single letter. Brad Flohr, a VA senior adviser for benefits, said the rest could not be found because the military’s records were incomplete: “There was no identifying information. No Social Security numbers, no addresses, no … any way of identifying them. Although we tried.” NPR then did the work itself: its research librarian Barbara Van Woerkom located more than 1,200 of the men in two months, from the VA’s own list and public records, and NPR published a searchable database of 3,900 named subjects. NPR also documented routine denials of qualifying men. Cavell was one: in 1988 he obtained his Naval Research Laboratory records — his full name, his time in the chamber, the gas level — and submitted them to VA. Denied anyway, he was granted only after NPR inquired, on the same information that had sat in his file for decades; officials told NPR they could not explain the delay.

Where it stands now

The last remedy came seventy-two years after the tests, built around one man. Arla Harrell enlisted in the Army in 1945 at 18. At Camp Crowder in southwestern Missouri, his daughter Beverly Howe told reporters, mustard agents were dabbed on his skin and he was placed in a gas chamber without protections; a 2012 Army Corps of Engineers report later confirmed chemical warfare training and gas-mask tests in three rooms it called “gas chambers” at Camp Crowder, with mustard gas present at the site. Harrell lived with lung problems and skin cancer; VA denied his claims again and again — a fourth denial came in April 2016, when he was 89 and in a nursing home. Senator Claire McCaskill took up his case and carried his name into law: the Arla Harrell Act, signed in August 2017 as part of a broader veterans education bill, requires VA to re-examine all previously denied WWII mustard gas and Lewisite claims, puts the burden on VA and the Defense Department — not the veteran — to prove or disprove exposure, presumes full-body exposure in reconsidered cases unless an agency proves otherwise, with revamped claims handling and mandated reports to Congress. By February 2018, VA reported 165 denied claims identified from 140 veterans and nearly $1 million awarded to 17 — Harrell among them; the law had told him, in McCaskill’s words, “We believe you.” He died in 2018, at 90 — believed, at last, by the government that had gassed him at 18.

For anyone still living: § 3.316 remains in force on the same terms, other conditions decided case by case, and NPR’s database remains the public instrument the government never built — 95 percent of its men already dead by 2015 (Columbia Journalism Review).

That is where the fight ends: not with a ceremony, but with arithmetic. McCaskill’s office estimated in 2017 that perhaps a couple of hundred subjects were still alive. The men’s own claims are closed by mortality. What remains belongs to their survivors — claims a widow or dependent may still pursue where a service-connected condition caused or contributed to a veteran’s death — and to the record itself: a program run in secret, documented so poorly that the institution later pleaded the missing records as its excuse, and answered, in the end, for 610 letters, one regulation, one database built by journalists, and one act of Congress named for a man who died the year after it passed.

The ledger

The accounting, in three lines — what the record shows has been done, what it shows has not, and what is still being sought.

What has been done

What has not been done

What is being sought

The record, on video

The spoken record is thin: the men were forbidden to speak for half a century, and most never did so on camera. What survives is the record of the fight at the end: the McCaskill report’s video, NBC on the benefits battle, CNN on the Harrell case, and St. Louis on Harrell’s death, a year after his law passed.

The report. Companion to Senator McCaskill's 2016 report *Cruel and Unusual Service*: 60,000 tested, some 4,000 given full-body exposure, held to secrecy for nearly half a century. (Senator Claire McCaskill.)

The benefits battle. NBC Nightly News, July 2016: veterans still fighting VA for benefits decades after the deliberate exposures. (NBC Nightly News segment, posted by Senator Claire McCaskill.)

The Harrell fight on CNN. Jake Tapper's *The Lead*, June 2017, examines McCaskill's fight for the veterans exposed to mustard gas, weeks before the Arla Harrell Act passed. (CNN segment, posted by Senator Claire McCaskill.)

The namesake, gone. KMOV St. Louis, June 2018: Arla Harrell dies at 90, the year after the law bearing his name finally won him the benefits denied him since the 1940s. (KMOV segment, posted by Senator Claire McCaskill.)

Sources

Part of Betrayed, Not Forgotten — this site's accounting for the service members and veterans their own institutions failed. Videos are embedded from their publishers; nothing here is re-hosted.