Chains 144 and 96 were not fractionated. Still more exte 2 ouclide separation was found for the underground shot, with all the above chains showin, vepletion in the crater area (Ref~ erence 65). From Shot 6 of Operation Tumbler-Snapper, the gross decay exponent decreased steadily with distance up to 70 miles from ground zero (Reference 65). Radiochemical data from Shot Bravo of Operation Castle showed fractionation of Sr™ and Ba‘** with respect to Mo”, but none for Ce!“ (Reference 65). In the land shots, Zuni and Tewa, of Operation Redwing, depletion of Cs'*’, Sr®, and Te? was found in the close-in fallout with maximum factors of 100, 13, and 7 (Reference 66). These de~ pletion factors became smaller with increasing distance from the shot point. Fractionation of the fallout from the barge shots, Flathead and Navajo, was much legs, and variations in abundance were not greater than a factor of 2 (Reference 68). Analytical data on cloud samples from these four events corroborated the fallout results (References 62 and 63). Some radiochemical analyses have been performed on particles of differem sizes from certain balloon shots (Reference 64). For Shot Boltzmann of Operation Plumbbob, both the Sr**/Mo" and Sr"/Mo” ratios were a factor of 2 greater in 22-micron particles than in 137micron particles. Enrichment of Sr®* in smaller particles was also found in two other balloon shots, Hood and Wilson. . 1.2.8 Fractionation Effects-—— Relations among the R-vValues for Several Radionuclides. As noted above, some scattered observations on fractionation were reported from the earlier tests, but it was not until Operation Redwing that enough data became available to investigate the separation of various nuclides from one another in any detail. During Shot Tewa of Opera- tion Redwing, six particle samples were collected from different locations in the cloud and subsequently analyzed for about 30 nuclides. From this work, relations among the R-values for the products became apparent, which seem to be of significance for understanding the fallout formation process (Reference 67). The R-values for the substances studied (normalized to give unit intercept on the axis of ordinates) were plotted against the R-value for Eu and a series of straight linea resulted with slopes ranging from positive to negative values. Posi- tive slopes indicated a simultaneous enrichment of the cloud particles tn europium and the product nuclide, whereas negative slopes showed that as the particles became richer in europium they were more and more depieted in the product nuclide. Products having rare-gas and alkali metal precursors had the steepest negative slopes, whereas U, Np and Pb had small negative slopes. The more refractory oxide elements— neodymium, beryllium, zirconium, and niobium— had positive slopes, and those elements such as calcium, which showed no fractionation with respected to europium, had infinite positive slopes. The results are consistent with the view that those products having rare-gas or alkali metal ancestors at the time of condensation will concentrate in the smaller particles, which have a larger surface-to-volume ratio. Similar relationships have been found for several high-yield airbursts, using Ba‘? as the secondary reference nuclide and Mo" as the primary reference nuclide (the primary reference nuclide is the substance used ag reference in calculating the R-values; the secondary reference nuclide is the substance used as abscissa in the R-value pilots}. In this reference system, Ag'!!, v3" cal§) cg) Np? vy! and Sr®® had approximately unit positive slopes, whereas Zr?’, Ce'“, pu’and the rare earths had average negative slopes of 1.5. For these shots, there was evidence that the nuclides in the larger particles (3 to 12 u) were fractionated, but thoge in particles smaller than 1 were not (Reference 68). This method of data analysis has been shown to be valid regardless of the secondary refer- ence nuclide, the primary reference nuclide, and the reference event (Reference 8). 1.3 EXPERIMENTAL PROGRAM 1.3.1 Outline of the Program. The foregoing discussion indicates that further progress in the development of a realistic fallout model will require an improved knowledge of the structure of nuclear clouds with respect to the vertical and radial distribution of particle size and 18