Once the basic data, consisting of the number of particles in each arbitrary size interval between 50 and 2,600 microns, were obtained for the selected trays, they were normalized to a l-micron interval and smoothed, to compensate in part for sample sparsity, by successive applications of a standard smoothing function on a digital computer. These, with appropriate unit conversions, are the results listed in Tables B.3 and B.5: the numbers of particles, within a 1-micron interval centered at the indicated sizes, collected per hour for each square foot of surface. Figures 3.6 through 3.9 show how the concentration of each particle size varied over the buildup period by providing, in effect, successive frequency distributions on time-line sections. The curves representing the 92.5- and 195-micron particles have been emphasized to bring out overall trends and make the figures easier to use. Measures of central tendency have been avoided, because the largest particles which make the most-significant contribution to the activity are not significantly represented in the calculation of the mean particle size, while the small particles which make the greatest contribution in the calculation of the mean particle size are most subject to errors from counting and background dust deposits. It should also be remembered that sampling bias is present and probably assumes its greatest importance for the smal! particles. Plots of pure background collections for the ship and barge stations resemble theplotof the YAG 39 data for Shot Zuni, but without the marked peaks in the small particles or the intrusions of the large particles from below, both of which are characteristic of fallout arrival. This is not necessarily true for the Howland station, however, where such features may result from disturbances of the surface dust; the series of peaks at about 4 hours during Shot Zuni, for ex- ample, appears to be the result of too close an approach by a survey helicopter. 3.2.5 Ocean Penetration. Figure 3.10 shows the general penetration behavior of fallout ac- tivity in the ocean for Shot Navajo, a water-surface shot, and Shot Tewa, resembling a landsurface shot. These simplified curves show a number of successive activity profiles measured during and after the fallout period with the oceanographic probe (SIO-P) aboard the YAG 39 and demonstrate the changing and variable nature of the basic phenomena. The best estimates of the rate at which the main body of activity penetrated at the YAG 39 and YAG 40 locations during Shots Flathead, Navajo, and Tewa are summarized in Table 3.3, and the depths at which this penetration was observed to cease are listed in Table 3.4. The data from which the results were obtained are presented in graphical form in Figure B.1; reduced-activity profiles similar to those shown in Figure 3.10 were used in the preparation of the plots. Estimates of the maximum pene- tration rates observed for Shots Zuni, Navajo, and Tewa appear in Table 3.5. The values tabulated in Reference 20 represent the result of a systematic study of measured Profiles for features indicative of penetration rate. Various shape characteristics, such as the depth of the first increase in activity level above normal background and the depth of the juncture of the gross body of activity with the thin body of activity below, were considered; but none was found to be applicable in every case. The concept of equivalent depth was devised so that: (1) all the profile data (i.e., all the Curves giving activity concentration as a function of depth) could be used, and (2) the results of the Project 2.63 water-sampling effort could be related to other Program 2 studies, in which the determination of activity per unit volume of water near the surface (surface concentration) was a prime measurement. The equivalent depth is defined as the factor which must be applied to the Surface concentration to give the total activity per unit water surface area as represented by the measured profile. Because the equivalent depth may be determined by dividing the pla- timetered area of any profile by the appropriate surface concentration, it is relatively independent of profile shape andactivity level and, in addition, can utilize any measure of surface concentration which can be adjusted to the time when the profile was taken and expressed in the Same units of activity measurement. Obviously, if the appropriate equivalent depth can be determined, it may be applied to any measurement of the surface concentration to produce an es- timate of the activity per unit area when no other data are available. The penetration rates in Table 3.3 were obtained by plotting all equivalent-depth points avail- 47