Numerous estimates of local fallout have been preparec -om previous operations, mainly from analyses of radiation intensity data obtained in aerial and surface monitoring surveys. However, the uncertainties in converting from dose rate measurements to fission products deposited per unit area are so great that the results cannot be regarded with a great deal of confidence. More reliable values are evidently needed, and in planning for Operation Hardtack, the Atomic Energy Commission examined possibie ways of obtaining such information (Reference 1). After consideration of the difficulties inherent in additional refinement of surface measurement techniques, this approach was abandoned. An alternative program based on further devel- opment of existing cloud-sampling procedures was formulated (Reference 2), and this culminated tn Project 2.8. A knowledge of fallout partition and how it is influenced by shot environment may contribute to reduction in worldwide fallout during future tests and to a better understanding of the military implications of local fallout. It will algo assist in extrapolation to previously untried shot condi- tions and yields. 1.2.1 Formation and Nature of Fallout Particles. When a surface burst is detonated, great quantities of the adjacent environment are swept up and mixed with the incandescent air in the fireball. There igs sufficient thermal energy in the hot gas to completely vaporizeall the material in the immediate vicinity, but the flow of heat into a massive object, such as a shot tower, shieid, or coral rock, will be comparatively slow even with a high temperature gradient. Consequently, the interior portions of large structures in the neighborhood may not receive enough heat to evaporate and will be melted only. Later, when the fireball has risen above the surface, the material carried into it by the vertical air currents around ground zero will not be heated to the melting point. Asa result, the fireball in its later stages will contain the environmental components as a mixture of solid particles, molten drops, and vapor. The extraneous materialin the Pacific shots will consist of coral and ocean water salts plus the components of the device, Shield, and tower or barge. The preponderance of oxygen and of the environmental material in the fireball ts of outstand- ing importance in the formation of the fallout particles. As the hot air cools through the range 3,500° to 1,000° K, it becomes saturated with respect to the vaporized constituents, and they condense out as an aggregate of liquid drops (Reference 3), most of which are very small (References 4 and 5). These are mixed with the larger drops formed by melting the environmental material and with the solid particles. The radionuclide atoms present will collide frequently with oxygen atoms or molecules and, because the majority of them are electron donors, metallic oxide molecules will be formed, which become thermodynamically stable as the temperature falls. The oxide molecules, or free radionuclide atoms, also have frequent collisions with the liquid drops of environmental material (silica, alumina, iron oxide or calcium oxide), and these collisions may be inelastic, because in some cases the incoming molecules will be held by strong attractive forces. The radioactive oxide molecules that condense at the liquid surface will spread into the Interior of the drops and become more or less uniformly distributed throughout. Later, after the liquid drops have frozen, the incoming radionuclide molecules may be held by surface forces. Be- cause of the very low concentrations of the radionuclide oxide molecules, collisions with one another will be relatively infrequent, and it appears that the aggregation of enough molecules of this type to form a drop or crystal will be a rare event, if !t occurs at all. Another way in which the radionuclide molecules may become associated with the environ- mental material is by participation in the structure of the cluster embryos, which are the precursors of the liquid dropa (References 4 and 8). The isobaric radionuclide chains formed in the explosion are known to be distributed on a mass scale in a way generally similar to the products of asymmetric fission of U*** by thermal neutrons, but with some important differences. The experimental yield curve for slow neutron fission has a broad minimum for mass numbers approximately half that of the original nucleus and maxima on either side at mass numbers in the neighborhood of 95 and 139 (Reference 7). Comparing the chain yields for megaton-range detonations with this curve, it is noted that there 12