while a value more nearly applicable for material deposited from 1 to 3 hours after detonation was used for the barge and island stations. It is to be noted that the insoluble solids of the slurry particles (Section 3.3.2) were not included in the conversion of grams of NaCl to gramsof fallout. Even though highly active, they constituted less than 2 to 4 percent of the total mass and were neglected in view of measurement errors up to +5 percent for sodium chloride, +15 percent for specific activity, and +25 percent for water content. 3.2.4 Particle-Size Variation. The way in which the distribution of solid-particle sizes _ varied over the fallout buildup period at each of the major stations during Shots Zuni and Tewa is shown in Figures 3.6 through 3.9. The data from which the plots were derived are tabuiated in Table B.3, and similar data for a numberof intermediate collection intervals are listed in Table B.5. AL of the slurry particles collected over a single time interval at a particular lo- cation during Shots Flathead and Navajo tended to fall in one narrow size range; representative values are included in Table 3.12. The information contained in Tables B.3 through B.6 and plottec in the figures represents the results of studies described in detail in Reference 19. All IC trays were inserted in a fixed setup employing an 8-by-10-inch-view camera and photographed with a magnification of 2., soon after being returned to NRDL. Backlighting and low-contrast film were used to achieve maximum particle visibility. A transparent grid of 16 equal rectangular areas was then superim- posed on the negative and each area, enlarged five times, printed on 8-by-10-inch paper ata combined linear magnification of 10. Since time-consuming manual methods had to be used in sizing and counting the photographed particles, three things were done to keep the total number as small as possible, consistent with good statistical practice and the degree of definition required. (1) The total number of trays available from each collector was reduced by selecting a representative number spaced at more or less equal intervals over the fallout-buildup period. Reference was made to the TIR and IC curves (Figures 3.1 to 3.4) during the selection process, and additional trays were included in time intervals where sharp changes were indicated. (2) Instead of counting the particles in all areas of heavily loaded trays, a diagonal line was drawn from the most dense to the least dense edge and only those areas Selected which were intersected by the line. (3) No particles smaller than 50 microns in diameter were counted, this being arbitrarily established as the size defining the lower limit of significant local fallout. (The lower limit was determined from a fallout model, using particle size as a basic input parameter (Section 4.3.1). Particles down to ~ 20 microns in diameter will be present, although the majority of particles between 20 and 50 microns will be deposited at greater distances than those considered. ) Actual sizing and counting of the particles on the selected ten times enlargements was accomplished by the use of a series of gages consisting of four sets of black circular spots of the same magnification, graduated in equal-diameter increments of 5, 10, 30, and 100 microns. These were printed on a sheet of clear plastic so that the largest spot which could be completely inscribed in a given particle area could be determined by superimposition. Thus, all of the par- ticle sizes listed refer to the diameter of the maximum circle which could be inscribed in the projected area of the particle. A preliminary test established that more-consistent results could be achieved using this parameter than the projected diameter, or diameter of the circle equal to the projected area of the particle. A number of problems arose in connection with the counting procedure: touching particles were difficult to distinguish from single aggregates; particles which were small, thin, translucent, or out of focus were difficult to see against the background; particles falling on area bor- derlines could not be accurately sized and often had to be eliminated; some elongated particles, for which the inscribed-circle method was of questionable validity, were observed; a strong tendency existed to overlook particles smaller than about 60 microns, because of the graininess of the print and natural human error. Most of these problems werealleviated, however, by havying each print processed in advance by a specially trained editor. All particles to be counted were first marked by the editor, then sized by the counter. 46