rapid calculation and ease of plotting, as well as involving an acceptable degree of accuracy. The model of the Armed Forces Special Weapons Project is probably the best of this type, basically using a conservation of material scaling approach, i.e., it assumes that a fixed percentage of material will fall out from all shots. As the yield is in- creased, the sizes of the contours are correspondingly increased. The Navy method is considered second in usefulness for this purpose to the AFSWP method, its main difficulty being its dependence upon an initial survey reading or series of readings made post-shot. If this normali- zation reading is in error, the entire plot is erroneous. The methods developed by the Air Weather Service and the Los Alamos Scientific Laboratory also fall into this category, the former furnishing little more than direction and time-of-arrival, and the latter having partic~ular application to the radsafe oroblem. The second use to which fall-out models are put is thet of postshot analysis. For this purpose, the information must be as accurate as can be obtained end the yield, height of the stabilized cloud, and exact wind history at all levels for the duration cf the fall-out period must be utilized. The contours so calculated are then compared in detail with accurate survey data of the area of the fall-out fields and the intensity within the field. The RAND model, used in conjunction with an electronic computer programmed to handle the voluminous calculations involved, appears to furnish the most accurate analysis of the mechanisms and patterns of fall-out deposition at the detonations studies. Much of the information used in describing atomic cloud be- havior, particle distribution and meteorological action on the cloud and quoted in this report was also used in the preparation of the RAND model. The Naval Radiological Defense Laboratory {fs considered second to the RAND model in supplying useful post-test data largely because it has not yet been programmed for a computer. The large amount of manual calculation made necessary in using the NRDL method makes the RAND method preferable. The Army Signal Corps has a useful and fairly accu- 54