The men did their jobs. They stood at their stations on warships in the open Pacific while, in some trials, aircraft and barges released clouds drifting toward and across their decks. What a portion of them did not know — what the Navy would not confirm for forty years — was that among the things being sprayed in the broader test series were live chemical warfare agents, including sarin and VX, two of the most lethal nerve poisons ever synthesized. This was Project SHAD, Shipboard Hazard and Defense, and the sailors were the instruments.

SHAD sat inside a larger Cold War effort called Project 112, run out of the Deseret Test Center in Utah from roughly 1962 to 1973. The official goal was defensive: to learn how vulnerable U.S. ships and troops were to a chemical or biological attack and to develop ways to detect, decontaminate, and survive one. To get real data, the planners used real agents and real people. Across the program, roughly 5,000 to 6,000 U.S. servicemembers — most of them Navy and Army, some Marines and Air Force — were involved in dozens of trials. Of those trials, a documented subset used actual chemical agents: not only sarin and VX but also other toxic compounds and biological simulants, with caged rhesus monkeys staked out on barges as living dosimeters alongside the men.

For decades the official posture was denial and silence. Veterans who came home with unexplained illnesses — and who half-remembered strange clouds and decontamination drills — were told, in effect, that there was nothing to know. The turn came in the late 1990s and early 2000s under congressional and Department of Veterans Affairs pressure, when the Department of Defense was finally compelled to declassify and release a series of formal fact sheets — eventually dozens of them — naming individual SHAD/112 tests, the agents used, the ships involved, and the dates. These are official Pentagon documents admitting to the exposures. The cover-up didn't end because the government confessed; it ended because it was forced to publish its own paperwork.

The evidence trail is unusually clean for a story like this, and that is exactly what makes it damning. You can read the DoD fact sheets. You can read the resulting independent study: in 2007 the Institute of Medicine (now the National Academy of Medicine), commissioned by the VA, published "Long-Term Health Effects of Participation in Project SHAD," which accepted as established fact that the tests occurred and that servicemembers were exposed to chemical and biological agents. The argument by that point was no longer whether sailors had been sprayed — it was what it did to them.

And here the honest reading gets uncomfortable for both sides. The 2007 Institute of Medicine study, and a follow-on analysis, compared SHAD veterans against similar veterans who hadn't been in the tests and found no statistically significant difference in overall mortality or in most measured diseases. In other words, the large-scale epidemiology did not demonstrate that exposure produced a clear pattern of long-term harm. The Pentagon leans on that finding hard. But the same studies were explicit about their own limits: many tests used simulants or low doses, the surviving exposure records were incomplete, individual dose data often didn't exist, and the cohort was small enough that real but modest effects could hide inside the statistical noise. "No proven population-wide signal" is not the same as "these men were unharmed," and it is certainly not the same as "this was acceptable."

Because the part that no statistic can launder is the consent. Sarin and VX are not crowd-control irritants; a pinhead of VX on bare skin can kill. The animating principle of post-Nuremberg medical ethics — and of the U.S. military's own later rules — is that you do not expose human beings to weapons-grade toxins as test subjects without their informed, voluntary agreement. By the accounts of many SHAD veterans, that agreement was never sought and the true nature of the trials was never disclosed; they learned what they had been part of from declassified fact sheets decades later, after some had already buried the question with their doctors.

So the unresolved questions are not really about toxicology. They are about the file itself. SHAD/112 was kept secret for forty years and only surfaced under sustained external pressure — which means the public timeline starts at the moment the government chose to stop hiding, not at the moment it had nothing left to hide. How many trials, agents, and exposed personnel are fully captured in the released fact sheets, and how many of the original records were degraded, lost, or never written down? When an institution sprays its own people with nerve agent and then needs four decades and an act of Congress to admit it, the safest assumption is not that the published account is complete. It's that it's the part they were finally willing to show.