CORRELATIONS ON DEBRIS FROM SILICATE BURSTS 1¢ — T [~ | 77 , & LABORATORY A | & LABORATORY B a, a . io a 10° a a SLOPE = 0.509 a — tA Ld 107! 107? ea INTERCEPT = 5.75 COEFFICIENT OF CORRELATION = 0.439 | | iy 10° a995 L Ludi 10! Fig. 3——Fractionation plot for “Ce in Small Boy. All the data were reinvestigated by plotting and fitting the data points of each laboratory separately. Differences in the degree of correlation appeared which were sufficient to cast suspicion on part of the data for '*Te, “cs, and ‘Cs. Calibration differences among the laboratories also appeared to be obscuring the correlation in some instances. Figure 4 illustrates this effect. When the data from the three laboratories are treated separately, the three lines shown are obtained with a reasonable degree of confidence. Note that their slopes do not differ greatly. However, when the data are lumped together, the correlation is much less clear. In spite of the difficulties involved in correlating the data, some facts emerge fairly clearly, as can be seen in Table 1. First, the same group of nuclides previously mentioned (®sr, *°sr, *y, !Ru, 16Ru, 1347, pe bog Mics MB, Mice, and ?%Np) fractionated from *°Zr in all three shots, whereas °*Mo, '“4Ce, and ***Pu did not. Second, '33Cs appears to fractionate about as severely as “Sr. Beyond these basic points, the Sedan data display some differences from the data obtained in the other two shots. This is not surprising in view of the radical differences in the Sedan shot conditions. For Sedan the frac- tionating nuclides other than '"Cs all showed a moreorless intermediate degree of fractionation—the slopes of the log—log plots were between about 0.3 and 0.6—whereas, for Johnie Boy and Small Boy,