in cloud models and deposition mechanisms that are introduced by a change in yield. Nevertheless, the method is useful and fairly accu- rate if applied over a limited yield range, not to exceed about two orders of magnitude. Wind velocity scaling is even more subject to error than is yield scaling. Based upon limited high explosive experimental data, using dye tracers, it has been postulated that total areas of effect of fallout for a particular detonation are essentially wind-independent, although the specific regions which these areas cover are of course determined in detail by the wind pattern. The experimental data ex- tends only to winds up to 25 knots; however, in the absence of other information, extrapolation to higher wind velocities is not unreason- able. On this basis, then, contour dimensions in a downwind direction would be scaled in direct proportion to the cube root of the ratio of the winds involved, while crosswind dimensions would be scaled inversely by the same factor, leaving the area of effect essentially constant. The effect of winds upon the direction of the fall-out and the area involved can be shown graphically by setting forth certain assumed meteorological conditions. in Figure 9. One example of such assumptions is presented Superimposed upon the wind vector plot is a family of idealized isodose contour lines for a 20 KT ground burst. It will be noted that the shaded area, which is due to diffusion, and the A", B” eeeee F" due to 50 micron particles, do not contribute apprecisbly to the radiation intensity within the area. The highest radiation read- ings in the contour areas are close in to the burst point and the radioactivity deposited there is carried by the larger particles. For purposes of the illustration, the size assumed for the large particles was 150 microns and the greatest effect by the wind is indicated on the line A', B’ ..... F'. The area covered by superweapon detonations would be much larger and wind shears at high altitude would have a greater effect upon the true pattern, but the method of illustration is the same. Such a superposition is useful in indicating where the KL