8 FREILING, CROCKER, AND ADAMS tained from the radiochemical analysis of fractionated fallout samples) to predict fractionation effects.4 When so used, with a lognormal particle-size distribution, mass-balance requirements are automatically fulfilled for each mass chain. The assumptions of spherical particles and radial power distribution guarantee logarithmic correla- tions for monodisperse samples, and mass-balance fulfillment can be expected for any overall particle-size distribution. The thermodynamic equilibrium model can be utilized in a semi- bo — Q oa UNFRACTIONATED EXPOSURE-DOSE RATE, % empirical manner by using it as originally described to calculate the CONTRIBUTION OF NEGLECTED GROUP sy oO an oOo 80 L/ I ALL VOLATILE GROUP Se etSon ALL REFRACTORY GROUP OSTaso 0 pees 0.1 po pepeee pepe pe pe 0.2 O38 04 O05 06 O07 0.8 0.9 1.0 Ta9,95 Fig. 3—Semiempirical model prediction of the effect of fractionation on the 1.12-hy dose vate for a burst in which solidification occurs in 6 sec. degree of fractionation for each sample and then attaching to that sam- ple the effect empirically expected for that degree of fractionation. It has not yet been demonstrated, however, that mass-balance requirements will then be fulfilled. The seriousness of this defect would be less for dose-rate estimates than for radionuclide partition and may not be serious in either case, compared to the magnitude of other un- certainties. Figure 3 shows the percentage of unfractionated exposure-dose rate expected for different degrees of fractionation at 1.12 hr from a burst in which solidification occurred in 6 sec. According to Miller’s estimates’ this would correspond to a yield of 25 kt. The prediction indicates a maximum depletion of a factor of 5 in the dose rate from local fallout. We have data from the 1962 test series in Nevada which indicate that depletion may be much greater in actual situations. The observations require further substantiation and verification, however.