The the and for correlation coefficients show significance of the correlations at 95% level (or higher) for three sets of data: Shots 2, 4 (Set b), 5. However, for these, m is significantly different from zero only Shots 2 and 4 (Set b), and of these, m < 0.1 for Shot 2. The corre- lation in Shot 4 (Set b) is not duplicated in Set tre "a" tt from this shot. The correlation coefficients are significant at the 90% level for Shots 1 (239/147), 3, and 8. The value of m is significantly different from zero only for Shot 3, although we see 88% significance levels for Shot l. It is our conclusion that these results confirm the qualitative statement made by Russell (1966): Plutonium, and by inference uranium, behaves like a refractory nuclide in shots on or over coral. The behavior is not entirely clear in the shot detonated at shallow depth (Johnie Boy). The apparent behavior of plutonium is a little less refractory than that of the rare earths, as evidenced by the calculated negative slope of the correlation. The significance of the slope has not been established definitely. The remaining three shots over alluvium all had scaled heights of burst of about 200 feet or more and approached free-air bursts. In all of these, the apparent plutonium behavior was about the same as the behavior of the 144ce, which is refractory. It would appear, then, that no large errors are made when it is assumed that the plutonium and refractory fission products are distributed similarly, and that, therefore, plutonium behavior may be inferred from gross beta or gross gamma measurements. Any errors caused by this inference would be such that the plutonium concentrations would be overestimated in fallout samples if such estimates are made on the basis of cloud-sample analyses. REVIEW OF RADIOACTIVITY DATA Airbursts The objectives of this work do not relate to free-air bursts. Yet a capsule overview of radionuclide behavior in such shots is appropriate, not only because airbursts comprise the limiting case of height-ofburst phenomena, but also because radionuclide behavior in all or most balloon shots resembles radionuclide behavior in airbursts. Some of the work in this country has been published (Benson and Leventhal, 1965; Nathans, 1971), but most of it remains as yet unpublished (primarily work by Hicks, Stevenson, Nathans and Benson, and Nathans). The specific beta and gamma activities of particles larger than about 2 or 3 um of medium- and high-yield shots is usually independent or only very slightly dependent on the particle size. Figure 1, where the points represent grouped data, shows an example. Regression analysis 567