34 FREILING, CROCKER, AND ADAMS FRACTIONATION CORRELATIONS As previously explained, correlation parameters for fractionated radiochemical compositions provide useful and realistic input for the semiempirical prediction of fractionation effects. Moreover, the re- sults of statistical analyses of fractionation data furnish insightinto the overall fractionation process. We are presently completing the correlation of thousands of analytical results on fractionated debris from 15 air bursts covering a yield variation of over 300-fold. Details will appear in a forthcoming USNRDL technical report. For a more effective comparison of the correlation parameters for air bursts with those obtained from surface bursts, the air-burst data will be presented in a following paper.* Some observations on air-burst data of interest, however, are appropriately given here: 1, Although the majority of rgy 9, values observed fell in the range from 0.1 to 3.0, several shots showed a hundredfold variation, and the greatest was over 900-fold. 2. The preference for linear or logarithmic correlations was variable and depended upon the “goodness-of-fit” criteria. 3. There was little, if any, dependence of correlation parameters on device yield. 4, Tests are still in progress to determine whether the values of given nuclide ratios (as determined on different shots) are all from the Same population. Upon completion of air-burst data correlations, we plan to cor- relate data from underwater bursts and additional high-yield surface bursts. Another project under way in this area is the retrieval of information from old fallout samples. There are extant debris samples from most nuclear detonations. Many of these detonations were poorly documented with regard to fractionation data. In addition, some large gaps exist in our knowledge of fractionation phenomena (e.g., tower- burst fractionation, fractionation of soil-induced activities, and frac- tionation as a function of particle type). There remain in these samples both volatile and refractory fission products to use as standards (e.g., 387Cs and ‘“Ce). Thus the present moratorium on weaponstesting of- fers an opportunity to recover much pertinent data, and this project has already yielded a few surprises. LABORATORY STUDIES Our laboratory studies to date have been concerned with the equilibrium properties of systems containing volatile oxides. These studies *See paper by Glenn R. Crocker, Freiling, this volume, Francis K. Kawahara, and Edward C,