Nobody at the General Electric Research Laboratory set out to make weather. The company’s own history of Project Cirrus, written by Barrington Havens in 1952, traces the work back to a request from the Chemical Warfare Service in 1940: Irving Langmuir and Vincent Schaefer were asked to study how gas-mask filters caught smoke, and to make smokes to test them with. They learned how to make particles of extraordinarily uniform size by heating an oil, mixing the vapour with air and quenching it suddenly. When a form letter came round in August 1941 asking whether anyone could make a white screening smoke to cover whole areas against bombing, they tried the same method on a large scale, and the Army used the generators through the war.

The next request was precipitation static. Aircraft flying through snowstorms in the Aleutians charged up to a quarter of a million volts, and the corona discharge drowned their radios. Langmuir and Schaefer had no ideas on the subject, the history says, “except that it had to do with weather,” and they went to the best place they knew for weather, the observatory on the summit of Mount Washington in New Hampshire, where the winter temperature averages minus four or five Fahrenheit and the wind sixty miles an hour. Schaefer went up several times in the winter of 1943, and discovered that anything exposed there was at once coated with ice, because the cloud sweeping the summit was supercooled water. The Army Air Forces were interested in aircraft icing, so in 1944 the two of them turned to that, with Langmuir calculating what fraction of the droplets in a cloud would hit a surface and Katharine Blodgett helping him feed the sums to a differential analyser. By late 1946 the icing contract had run out and the pair had become, by their own admission, absorbed in a question that had nothing to do with the Army: why a cloud well below freezing sometimes snows and sometimes does not.

Their answer, from the summit, was that a supercooled cloud from which no snow falls simply has no effective nuclei in it. Put ice crystals into such a cloud and they must grow fast at the expense of the droplets and fall out. The question was what made crystals appear.

The box

Langmuir was away in California for three or four months in 1946, and in his absence Schaefer made what Langmuir later called “some beautiful experiments.” He had been supercooling drops on polystyrene surfaces to minus twenty degrees, and knew that was not a cloud. So he took a home freezing unit of the kind used for food, lined it with black velvet, shone a light down into it and breathed into it. The moisture condensed into a fog of droplets just like a cloud, at about minus twenty-three degrees, and no ice formed. He dusted into the box every substance he could think of, and almost never got a crystal.

One July day the chamber was not cold enough, and he put a large piece of dry ice in to bring the temperature down. “In an instant the air was full of ice crystals,” the history says, and they persisted after he took the dry ice out. A tiny grain did the same. A sewing needle dipped in liquid air and passed once through the fog did the same. It was not the carbon dioxide; it was the cold. Schaefer measured the temperature at which the cloud changed to ice of itself, minus 38.9 degrees, give or take a tenth, and from then on anything at minus forty or colder would serve.

Bernard Vonnegut, who had come to the laboratory from the Massachusetts Institute of Technology in the autumn of 1945 and had been supercooling metals to keep out of Schaefer’s way, was asked by Langmuir that autumn to help count the crystals a pellet made. It occurred to him that a substance whose crystal was shaped like ice might serve as a nucleus without any cold at all. He went through the tables of crystal structure, picked three candidates out of more than a thousand compounds, lead iodide, antimony and silver iodide, and dropped them into the box. The powders did almost nothing. A silver smoke drawn from an electric spark did a great deal but did not last; silver iodide vaporised and condensed into a smoke worked completely. His first failure with the powder, it turned out, was impurity. By his calculation a few pounds of silver iodide smoke could give every cubic inch of air over the United States a nucleus.

In August 1946 Langmuir worked out on paper what dry ice would do in a real cloud: how fast the pellets fell and how long they lasted, how many crystals they made, how big the snow would be. He also showed that freezing the cloud’s water would warm it, and that the warmed air would rise and carry the nuclei through the rest of the cloud, so that seeding along lines a mile or two apart should nucleate a stratus deck within half an hour. That calculation is the hinge of everything that followed: it is why a few pounds could be expected to do the work of a storm.

The aeroplane

On 13 November 1946 a Fairchild was rented at the Schenectady airport, with Curtis Talbot as pilot, and Schaefer went up to look for a cloud. He found one over Pittsfield, Massachusetts, about thirty miles to the east, at 14,000 feet and minus twenty degrees. His laboratory notebook for the day is quoted in the history:

“Curt flew into the cloud and I started the dispenser in operation. I dropped about three pounds (of dry ice) and then swung around and headed south. About this time I looked toward the rear and was thrilled to see long streamers of snow falling from the base of the cloud through which we had just passed. I shouted to Curt to swing around, and as we did so we passed through a mass of glistening snow crystals!”

They made another run through the unseeded part of the cloud, with Schaefer letting three pounds more out through the window so that the suction of the passing air carried it off, and watched “draperies of snow which seemed to hang for 2-3000 feet below us” and the cloud drying up “very similar to what we observe in the cold box in the laboratory.” “I turned to Curt and we shook hands as I said ‘We did it!’” Back at the airport Langmuir rushed out: from the control tower of the laboratory he had seen the draperies appear within two minutes of the radio call that the run was starting.

Three things were learned at once. The veil of snow that appeared below the cloud could not have fallen from it so quickly; it had formed where the pellets fell through air saturated with respect to ice but not to water. A supercooled cloud could therefore often be seeded from just below. And on 23 and 29 November, when Schaefer seeded isolated cumulus clouds and photographed them from the ground every ten seconds, the whole of each cloud turned to ice within five minutes, but the air inside an orographic cloud is replaced in about as long, so that experiments on small cumulus were “usually of little interest, for the effects last but a few minutes.” On 21 November he seeded a valley fog, and found he could either thicken it by making more crystals than there were droplets or clear it by making just enough.

The flight that counted was the fourth, on 20 December. The sky was overcast, and by nine in the morning the Weather Bureau at Albany was forecasting snow for seven that evening. At about noon Schaefer dropped some twenty-five pounds of granulated dry ice into the lower part of the overcast, a pound or two to the mile, about a thousand feet above its ragged base, at 7,000 to 8,500 feet, with a two-pound bottle of liquid carbon dioxide for good measure. Supercooled drizzle had been falling and evaporating before it reached the ground; flying back along the line he found that where he had seeded it was snowing and where he had stopped the drizzle began again. He made three more runs on the same line and then went down to 4,000 feet to see where the snow was reaching the ground. At 2:15 that afternoon it started to snow in Schenectady and at many places within a hundred miles, at about an inch an hour for eight hours, the heaviest snowfall of the winter.

“While the seeding group did not assume it had caused this snowstorm,” the history says, “it did believe that, with weather conditions as they were, they could have started a general snowstorm two to four hours before it actually occurred, if they had been able to seed above the clouds during the early morning.” The company’s press release did not claim the storm. “The coincidence of the two events did cause some independent speculation over the possibility of cause and effect.”

The contract

That speculation went straight to the general counsel. The work was no longer under any government contract; the last had expired in June. C. G. Suits, the laboratory’s director, had already taken the results to the company’s officers, who had agreed that weather modification was remote from anything the laboratory or the company did, that it was possibly of great importance to the country, and that it should be pushed forward, under a government contract if one could be had, since the work would need weather data, aeroplanes and flight equipment that only the government could provide. A proposal had gone to the Signal Corps’ Evans Signal Laboratory at Belmar, New Jersey, on 20 September, at the invitation of the chief of the Air Weather Service. Now Suits took the snowstorm to Vice President R. E. Luebbe.

The history gives the outcome in two sentences that read like a minute. “It was recognized that the possibility of liability for damage from cloud-seeding experiments was a very worrisome hazard in this new form of cloud experimentation. Since such a threat to the share owners’ money would not be balanced by any known gain to the Company’s products or business, there was great reluctance to incur risks of uncertain but potentially great magnitude.” The remedy followed: “It was considered particularly important for this reason that any seeding experiments be conducted under government sponsorship. No further seeding flights were made until such sponsorship was provided.”

The Signal Corps contract, W-36-039-sc-32427, for a “research study of cloud particles and cloud modifications,” began on 28 February 1947, with the Office of Naval Research as joint sponsor and the Air Force providing the aircraft. Its important clause was not about clouds. “The entire flight program shall be conducted by the government, using exclusively government personnel and equipment, and shall be under the exclusive direction and control of such government personnel.” The Research Laboratory told everyone concerned “that it is essential that all of the G.E. employees who are working on this project refrain from asserting any control or direction over the flight program. The G. E. Research Laboratory responsibility is confined strictly to laboratory work and reports.” A Steering Committee of the three services, with Langmuir and Schaefer as consultants, limited the General Electric staff to laboratory work and analysis; an Operations Group under a Navy officer, Lieutenant Commander Daniel F. Rex, took the flying. Until June the project’s flights were made by a Weather Squadron aircraft from Olmsted Field in Pennsylvania, which came to Schenectady six times; in the summer the operations moved into the General Electric hangar at the Schenectady County Airport. The name Project Cirrus was adopted on 25 August 1947.

So when the hurricane season came, the arrangement was already what it would be on 13 October: a government crew in government aeroplanes, dispensing dry ice that a General Electric machine metered out, with a General Electric scientist aboard who was not permitted to direct anything, on the strength of a calculation made in Schenectady in August 1946 and a cold box lined with black velvet. In the Senate four years later Suits would ask for a law to take the liability off any contractor, and Schaefer would add, for the record, “It so happens that we did not conduct the experiments.” Both statements were true, and the clause in the contract is why.

Sources

• Barrington S. Havens, History of Project Cirrus, General Electric Research Laboratory Report RL-756, Schenectady, July 1952, chapters II and III: the gas masks and smokes, precipitation static and Mount Washington, the cold box, Vonnegut’s silver iodide, Langmuir’s August 1946 calculations, the flights of 13 November, 21, 23 and 29 November and 20 December 1946, the snowstorm, the general counsel, the contract and its flight clause, and the organisation of the project. archive.org
• United States Senate, Weather Control and Augmented Potable Water Supply, joint hearings, 14 March 1951, for Suits’s and Schaefer’s statements on liability. archive.org