SNOW AND EXPERIMENTAL METEOROLOGY 
the wind velocity is high have features in common with 
the more turbulent portions of large cumulus clouds. 
The Development of Experimental Meteorology 
Control of the Weather. For many years man has 
dreamed of the possibilities of exercising control over 
the weather. To a limited degree he has accomplished 
this by heating, air conditioning, and artificial illumina- 
tion of homes, workshops, and places of culture and 
amusement, as well as by the use of flood-control 
techniques, irrigation systems, and similar technical 
improvements in limited areas of his countryside. 
One of the first suggestions that scientific research 
would eventually lead to the modification of weather 
systems is contained in an important paper by Findei- 
sen [5]. Others, such as Gathman [7], had previously 
proposed the use of dry ice to produce clouds and rain, 
a method subsequently tried by Veraart [386]. Their 
idea was to use dry ice to chill the air below its dew 
point to produce clouds and rain. Unfortunately, count- 
less tons of dry ice would be required to produce even 
a small rainfall by this method. The insignificant results 
produced by this method and others, such as the forma- 
tion of rain by dumping electrified sand into clouds 
(tried by Warren and Bancroft), led in 1926 to the 
conclusion (Humphreys [8]) that none of the proposed 
methods were of any consequence and that they had no 
scientific importance. 
Recent Advances. In 1941 Langmuir [12] and Schaefer 
devised a method for producing an artificial fog which 
eventually was used to blanket sixty miles of the Rhine 
Valley as well as other large areas during the latter part 
of World War II. Subsequent research following the 
production of artificial fog involved basic studies of 
precipitation static on aircraft and of aircraft icing [20]. 
This latter study, which involved research with super- 
cooled clouds, led the writer to discover a method of 
changing a supercooled cloud to ice crystals in a simple 
but highly effective manner [21]. 
When this discovery was made, it was immediately 
apparent that the time had arrived to attempt the 
modification of clouds on a large scale. Unlike all pre- 
vious attempts, the methods proposed would utilize dry 
ice as a triggering device to release the energy stored 
within supercooled clouds by causing a change in phase 
of the unstable supercooled cloud. By utilizing the 
inherent instability of such clouds it was planned to 
introduce ice crystals in such numbers that large effects 
would be produced according to the Bergeron mecha- 
nism. 
The first experiments designed to transform a super- 
cooled cloud to snow crystals were carried out by 
Schaefer in the fall of 1946 [24] and the early months of 
1947. In March 1947 cloud-seeding experiments using 
dry ice in supercooled clouds were undertaken on a 
larger scale with government support. This activity, 
called Project Cirrus, was sponsored by the Army Signal 
Corps and the Office of Naval Research, using planes 
and crews supplied by the Air Forces, with technical 
consultation provided by Langmuir and Schaefer of the 
229 
General Electric Research Laboratory at Schenectady, 
New York. 
The laboratory, field, and flight studies of this group 
have been described in more than a score of reports 
[15]. Flight studies up to July 1950 have included ex- 
perimental investigations of clouds in the northeastern 
United States, New Mexico, Puerto Rico, and Florida. 
In addition, detailed observations of natural clouds 
have been undertaken in various parts of the country 
and many field and laboratory projects have been 
completed or are actively under way at the present 
time. 
Cloud Modification Methods 
The Use of Dry Ice to Modify Clouds. As mentioned 
earlier, whenever any object with a temperature less 
than —39C is introduced into air colder than OC and 
supersaturated with respect to ice, tremendous numbers 
of ice crystals are generated. The simplest method for 
achieving this effect is to introduce small fragments of 
dry ice (solid carbon dioxide) into the air. This may be 
accomplished in the free atmosphere by using free 
balloons, projectiles, or airplanes, or by placing the dry 
ice or expanding liquid CO: in orographic clouds on 
mountains or in supercooled ground fogs. 
Dry ice may be introduced ito various parts of 
clouds with differing effects. The ice crystals it generates 
will immediately disappear unless the air is below 
freezing and supersaturated with respect to ice. Lang- 
muir has summarized our general methods [14]. Briefly, 
the judicious use of dry ice may (1) dissipate clouds, 
(2) precipitate clouds, or (8) make existing clouds more 
persistent. 
Clouds may be dissipated by seeding their tops from 
an airplane before they develop a deep supercooled 
layer or by seeding along a line from the side, using 
liquid carbon dioxide. It is much easier to dissipate 
clouds after they have passed their maximum vertical 
development than when they are actively growing, since 
seeding a growing cloud may trigger off a chain reaction. 
Since it requires considerable experience and judg- 
ment to select a growing cumulus, experiments planned 
to cause precipitation may fail because of improper 
selection of the area to be seeded. If the larger of the 
towers in a cumulus system are selected for seeding 
while the airplane is in a position to see the general 
cloud system, the selected cloud will probably be in a 
subsiding state before seeding takes place. Under such 
conditions, dissipation is likely to result. The life cycle 
of such clouds is so short [43] that an intimate knowl- 
edge of their dynamic properties is essential if effective 
seeding is to be achieved. 
The effects are generally spectacular when the seeding 
is done to release effectively the heat of sublimation 
stored in supercooled clouds. In unstable air, a tremen- 
dous upheaval may occur if the seeded cloud is actively 
growing and well supercooled. This increase in convec- 
tion occurs because of the heat liberated by the shift 
from the water to the ice phase. In stratus clouds, this 
increase in temperature may amount to 0.5C, while in 
