120 LOCKHART, PATTERSON, AND SAUNDERS INTRODUCTION All nuclear explosions produce a wide spectrum of particles of various sizes, but the conditions under which an explosion is carried out markedly affect the shape of the distribution curve. Fractionation of the various nuclides relative to one another can occur during the cooling and condensation phases immediately following the detonation with the result that ratios of nuclides may be expected to vary with particle size. The use of a series of filters of different retentivities arranged in a pack so that the same air parcel passes through them in succession has been under consideration for a number of years as a possible means for determining the relation of radioactivity to particle size. A few measurements made at close-in sites during the Operation Greenhouse tests! indicated that size discrimination could be obtained by this method and, furthermore, that fractionation of nuclides with particle size occurred to a measurable extent in fresh nuclear debris. More recently, measurements made of the retentivity of filters to fission products and to the short-lived natural radioactive aerosols have shown gross differences in the behavior of the two types of radio- active particles as well as day-to-day differences infilter retentivity.?.? It was obvious that a pack of filters of different retention character- istics could effect at least a partial separation of radioactive particles into different size groupings. As a consequence of the preceding findings, aprogram wasinitiated to study the feasibility of the filter-pack technique for particle-size determinations, to make measurements of the temporal changes in the particle size of fission debris following cessation of the nuclear-test programs, and to determine the extent of fractionation of fissionproduct radionuclides in relation to particle size. EXPERIMENTS Collections of Gross Beta Activity Routine collections of gross fission-product radioactivity in the air at ground level have been made since the end of 1962 through use of three different filter media arranged inathree-filter or a four-filter pack, as shown in Fig, 1. Filters 4 in. in diameter were employed; the pumping rate was about 18.5 cu ft/min. Since filter efficiency is a func- tion of air velocity through the filter, positive displacement blowers have been used to give more reproducible and constant flow rates. The air flow was calibrated against a 55 cu ft/min Fischer and Porter Flowrater and a working graph of pressure vs. flow used for routine flow determinations.