More than sixty thousand American soldiers were used as chemical test subjects during World War II. Some of them were sorted by the color of their skin.

Rollins Edwards was nineteen or twenty, a Black draftee with the 1329th General Service Regiment, when officers walked him and about a dozen other men into a wooden room at Camp Claiborne, Louisiana, during training, and locked the door behind them. Then they piped in gas.

The agents were sulfur mustard and lewisite, the blistering weapons that had scarred a generation in the last war. Lewisite burns on contact, and Edwards said it felt like being on fire. Men screamed. Some fainted. They pushed at the walls of a room built so they couldn't get out. The mustard worked slower — its damage would keep surfacing for days — but by the time the door opened, the experiment had the reactions it was built to measure.

I found Edwards while I was researching something else, and I kept going back to one line from his Library of Congress interview. He had understood the rule he was living under, and he put it in about as few words as it can be put: you do what they tell you to do, and you ask no questions. He was a soldier. Refusal, as he understood it, could be punished. The obedience was the whole mechanism — it was what made the experiment possible in the first place.

None of this was hidden from the men it happened to. They knew, and their families knew, and carried it for decades. It was hidden from everyone else — filed, classified, kept out of the textbooks. And when the records did surface, they showed something the official memory of the war has no place for: Edwards was a data point in a comparison. The Army wanted to know, in its own words, what these agents did to Black skin — held as a separate question from what they did to white skin.

The design

A medical historian named Susan L. Smith went through the surviving records and found at least nine government-funded projects from the 1940s built on the premise that race changed how a body took a chemical weapon. African American, Japanese American, and Puerto Rican soldiers were each measured against a control group. The control — the baseline, the "normal" against which the others were scored — was white enlisted men.

The racial theory underneath this was mainstream science, respectable enough to fund and to design experiments around: federal money, the Army's own medical apparatus, the era's textbook assumptions about human difference. Skin tone became a variable you could chart.

And the charting had a use. The research was partly a search for what the planners called an ideal chemical soldier — men who could be shown to withstand gas better than others. Resistance would not have earned those men safer duty. It would have earned them the worst of it, on the theory that their bodies could take what others' could not. Black soldiers already made up seventeen percent of the personnel assigned to Chemical Warfare Service units, far above their share elsewhere in the War Department. Segregation had already sorted these men. The experiments put the sorting to one more use.

The comparison reached past one group. David Bessho grew up looking at an Army commendation on the wall at home, a column of more than three dozen Japanese American names. His father, Louis, had been a subject. The Army, his father told him, wanted to know whether the weapons would act on Japanese bodies — again, the Army's framing — the way they acted on white ones. Behind that question sat a plan the men never saw: a 1945 study for the planned invasion of Japan, "A Study of the Possible Use of Toxic Gas in Operation Olympic," which estimated that a gas campaign could inflict something on the order of five million casualties.

The scale

The full program was vast, and most of it was quieter than a locked chamber. The National Academies later counted more than sixty thousand servicemen used as subjects in chemical-defense testing. Many got small patch tests — a drop of agent on the forearm, a note on the reaction the next day.

The worst tier was smaller. About four thousand men went through chamber or field trials with heavy exposure. Some of the chamber runs carried a name the researchers used without any apparent flinch: man-break tests. Men went in wearing gas masks and treated suits and stayed one to four hours, often in rooms held at ninety degrees and sixty-five percent humidity to stand in for the Pacific. Afterward some kept the suits on for four to twenty-four more hours, so that the next morning their skin could be read for the redness that showed the vapor had come through the cloth.

Panama took it outdoors. On San Jose Island in the Canal Zone — the jungle standing in for the islands where American troops expected to fight — the Army ran more than a hundred experiments. Juan Lopez Negron, a Puerto Rican soldier, remembered planes spraying mustard from above while the men moved through the growth below. The soldiers had uniforms. The animals staked out beside them had none; he remembered the dead rabbits. He was burned and sick almost at once and spent three weeks in a hospital. In the large-scale field runs, bombers laid between 125 and 550 tons of sulfur mustard over a target area, and men walked patrol patterns through it afterward, dropping to the dirt on command.

Feeding all of this took four arsenals turning out close to 175 million pounds of sulfur mustard, plus some forty million pounds of lewisite. The country built an industry around a weapon it was mostly firing at its own soldiers.

The file

What happened after the tests is the part that decides whether any of it can be known at all.

The men were ordered to keep silent, and most of them did for more than forty years. The experiments were left out of their service records. So a man could carry scarred skin and ruined lungs and hold no document that connected the damage to what had been done to him in uniform.

That absence held for a long time. The federal government did not publicly acknowledge the experiments until the early 1990s, when a National Academies report laid out the scale of them; the racial selection took longer still to reach daylight, through Smith's work and the reporting that followed. The VA eventually named some conditions it would treat as linked to the exposure. But the paperwork fight ground on for years past that. John Tedesco, a Navy veteran tested at Great Lakes in 1944, spent decades short of breath and filed claim after claim; a letter from a man in his own company, swearing the two of them had been gassed together, did not move the denials. What finally moved the VA was the veterans themselves — men like Arla Harrell, a Missouri soldier who spent much of his life trying to prove what had been done to him. The 2017 law that made the agency reconsider those old claims carries his name.

Here is the thing I keep going back to. The Army measured these men with real care. It logged the redness on their skin the next morning, sent the numbers up the chain, and kept the file. What it did not keep was the man's own claim to that file. The body was documented and the person was not. Decades on, a veteran could stand at a benefits window as living evidence of an experiment his own government had run on him, and be told the one line that would have tied the two together was missing.

Rollins Edwards carried the proof on his own skin for the rest of his life — the burns still surfacing seventy years on, when a reporter finally sat down to listen. Whether it happened was never the hard part. The hard part, across forty years and a mountain of denied claims, was whether the sentence that said so would ever be allowed onto the page.

Sources: Rollins Edwards's account draws on his Library of Congress Veterans History Project interview and NPR's 2015 investigation (Caitlin Dickerson) into the race-based selection. Program scale, agent production, and the San Jose Island trials follow the National Academies' 1993 report Veterans at Risk: The Health Effects of Mustard Gas and Lewisite; the race-comparison studies follow Susan L. Smith, "Mustard Gas and American Race-Based Human Experimentation in World War II" (Journal of Law, Medicine & Ethics, 2008). The Operation Olympic gas study is documented in Polmar and Allen, "The Most Deadly Plan" (U.S. Naval Institute Proceedings, 1998). The postwar claims fight and the Arla Harrell Act draw on the Congressional Research Service (R45205) and reporting on the 2017 Harry W. Colmery Act (P.L. 115-48).