SMALL BOY SHOT FALLOUT RESEARCH PROGRAM 37 presence of particles with diameters as small as 1 to 3u inall samples;° therefore the midrange diameter of the 0- to 44- fraction was taken as 22 u. The mean values of i(100) were computed for each particle-size fraction on the assumption that the average radionuclide composition carried by particles of a given size is the same everywhere and is independent of where the particles land. After correlation the two sets of ion-chamber data were generally in very good agreement; therefore most of the spread in the i(100) values for each particle size is due mainly to spread in the radiochemical results. The spread in the values is largest for the smallest and largest particle diameters where the analytical sample sizes were smallest. In the midrange of the curve, the minimum-to-maximum spreadin i(100) is a factor of about 3. The major significance of the data is in the variation of i(100) and therefore of r,,(100) and r;, with particle size. From the average particle diameter of 22 uy up to an average particle diameter of about 400 y, i(100) decreases by a factor of about 5.5. At 400 » diameter, a minimum in the curve exists. A Sin..iar variation with particle size in the gross solubility of the radionuclides is observed.® The K, values for the collecting-station locations in the fallout area from the Small Boy shot were calculated with adjusted and averaged A; values; correlations of the variation in i’(100) and i(100) with the median particle size of the fallout particles at each station were used to make the adjustments in A,;. Initial correlations of i’(100) indicated a consistent difference in value depending on the time at which the samples were allocated for chemical analysis; the median value of the correlation factor for this difference was calculated to be 1.56. Since evidence for locating the cause of this difference is presently not available and since the calibrated gamma-scintillationcounter measurements were a highly unlikely source of error, the reported A, values? for the samples with the higher values of i’(100) were increased by a factor of 1.25, and the A; values for the remainder of the samples were decreased by a factor of 1.25. If the same type of correlation, including comparisons with the observed I, values,’ are applied to a third set of samples collected at offsite locations, the reported A; values’ are found to be more or less consistently high by a factor of 10. The data for the fallout samples leading to the calculation of Ky for each location are summarized in Table 7. Included are the activity median particle size, dsj; the ion-chamber measurements on the gross samples, I,, I,, and I; the values of i,(100), i,(100), and i7(100), as calculated from the initially adjusted values of A, (given as A,); the smoothed-curve values of i(100) and i’(100), as taken from the curves in Figs. 2 and 3; the recomputed values of A;, where A, is I,/i(100), A; is 1,/i(100), and A, is I’/i’(100); the average value of A,;; I,; and Ky.