Eleven patients, one funeral, twenty-six years of silence.

September 20, 1950.

The United States began building biological weapons in 1943, at a small airfield outside Frederick, Maryland, that the Army took over and renamed Camp Detrick. The work went on through the war and did not stop when the war did. By the end of the decade, the emphasis had shifted. The question was no longer only what could be built. It was what could be done to an American city by somebody else who had built the same things.

Nobody had an answer, because nobody had the numbers. There were agents, and there were ways of dispersing them, and there were ranges in Utah where clouds were released over open ground and measured as they moved. A city is not open ground. A city has hills and glass and steel, streets that channel wind, ventilation systems that pull outside air into buildings and push it through offices, and several hundred thousand people breathing at various rates in various rooms. What a live aerosol did once it drifted into a place like that was a matter of estimate. The only way to stop estimating was to release something over a real city and go count it afterward.

Whatever was released had to satisfy two conditions. It had to behave in the air the way a real agent behaves, which meant the particles had to fall in the same size range, small enough to stay aloft for hours and small enough to reach the deep part of a lung. And it had to be countable when the test was over, which meant a technician had to be able to look at a culture plate days later and know the organism on sight.

Serratia marcescens met both. It is a rod-shaped bacterium found in soil and standing water, and the strains available at the time grew in colonies of a deep red that nothing else on the plate resembled. It had been used before as a tracer, in plumbing studies and in hospitals, to follow the path of contamination through a building. It was used that way because the medical understanding of the period held it to be harmless to healthy people. The second organism in the mixture was Bacillus globigii, a spore former in the same size range as anthrax, also regarded as safe.

San Francisco was chosen for its geography and its weather. The city sits at the end of a peninsula with the Pacific on one side and the bay on the other. The prevailing wind comes off the ocean. In late summer and early fall, cold sea air meets warm air over the interior and fog moves in through the gap at the Golden Gate in the evening and sits on the city until the following midday. Material released offshore into that air would be carried inland and held there for hours.

In the third week of September a Navy vessel took station off the coast. Spray equipment was rigged on the deck. Over six days the ship ran lines across the wind and released the mixture in stretches of about an hour, timed against the fog. Nothing about it was visible from the shore, and the sound of a ship's machinery two miles out is the sound of any other ship.

The city did what it did every week. There were something over three-quarters of a million people in San Francisco in 1950. They worked the shipyard at Hunters Point and the canneries and the downtown offices, rode the streetcars, sent their children to school, opened their windows in the afternoon and shut them when the fog came. The test was classified. No public health department was told, no hospital was told, and no notice was given before, during, or after.

Forty-three sampling stations had been placed around the bay before the ship sailed. Each drew air through a collector. The collectors were taken in and cultured, and over the following days the red colonies came up on the plates. They came up in the downtown blocks and along the waterfront and across the bay and well inland, at distances beyond what the planners had expected the cloud to cover. Counts at some stations were high enough that the men reading them went back and checked the method.

The finding was that a city of that size could be blanketed with a biological aerosol released from a single ship standing offshore, and that nearly every person living in it would take in thousands of particles. The report went into a classified file.

The other half of what happened that autumn happened four miles from the waterfront, and the two halves would not be set next to each other for twenty-six years.

Stanford University Hospital was in San Francisco then. The medical school did not move down the peninsula to Palo Alto until the end of the decade. In the last days of September the hospital laboratory identified Serratia marcescens in a patient sample. That was not a result anyone there expected to see once. Over the months that followed it came back eleven times, in eleven patients, most of them with urinary tract infections.

The physicians could not account for it. This was the organism you used to trace where other things went, not the organism you found at the end of the trace. They looked for a common source inside the hospital, a piece of equipment, a solution, a procedure, and they did not find one that explained the pattern. The cluster was strange enough that they wrote it up and published it in the medical literature the following year, describing an outbreak of infection caused by a bacterium that was not supposed to infect anyone.

One of the eleven was a retired man named Edward J. Nevin. He was seventy-five years old. He had been admitted for surgery on his prostate, and the operation itself went the way it was meant to go. Afterward, he developed an infection that would not clear. The bacteria moved into his bloodstream, and from the bloodstream they settled on a valve in his heart. That condition is endocarditis, and in 1950 the drugs available against it were limited and the organism causing it was one nobody had experience treating. He was in the hospital for weeks. He died on the first of November, 1950. His family was told he had died of an infection following surgery.

The San Francisco trial was not the last of its kind. Over the next nineteen years, the Army ran open-air tests using stand-in organisms in other American places. Bacillus globigii was released into the ventilation system of the Pentagon. It was released along highways and at airports and in bus terminals. In 1966, men carrying light bulbs packed with the spores dropped them onto the tracks in the New York City subway and then rode the trains, taking samples, while the pressure of the moving cars pushed the cloud from station to station through the tunnels. The results in New York were similar to the results in San Francisco. The aerosol went further and faster than expected.

President Nixon ended the offensive biological weapons program in November 1969. The open-air testing stopped with it.

The San Francisco test became public knowledge in 1976, through newspaper reporting on the Army's testing history. In March 1977, a Senate subcommittee held hearings on biological testing that had involved human subjects. Army witnesses appeared and described the scope of what had been done. The number given for open-air tests conducted between 1949 and 1969 was 239. The San Francisco operation was one line in that accounting.

Edward Nevin's grandson was a lawyer in San Francisco. He read the coverage, went and pulled his grandfather's hospital records from 1950, and filed suit against the United States.

The case was tried in federal court in San Francisco in the early 1980s. The government argued three things: that the strain recovered at the hospital was not the strain that had been sprayed, that no causal link between the test and the eleven infections could be established at that distance in time, and that the decision to conduct the test was a policy judgment of the kind the government cannot be sued over. The court ruled for the government. The ruling was upheld on appeal.

No charges were brought. No agency was penalized. The records of the operation were declassified and are public, and the sampling data from the forty-three stations can be read by anyone who wants to look at what the fog carried into the city that week.

Serratia marcescens was selected for the test in part because it was believed to be harmless. It is now classified as an opportunistic human pathogen, a documented cause of pneumonia, bloodstream infection, and endocarditis, and American hospitals report outbreaks of it every year.