RADIOCHEMICAL FRACTIONATION CHARACTERISTICS 115 Similar graphs were prepared for ‘4Ce and **"U. The plot of the particle data for ‘4Ce also had to be extended a decade higher. The uranium activity per particle was too low to get accurate data, anda linear fractionation plot could not be developed from these data. Analyses were madefor "sr, °8cs, !*Te, !#3ce, and ?**Np, but they were not detected in the samples. Satisfactory values were obtained for 29mpe@ Mice and Na. GAMMA-RAY SPECTRA The individual radionuclides that are susceptible to gamma-spectra analysis in particles were determined. Limits of detection were ob- tained for the radionuclides of interest by assuming 10° thermal-neutron fissions. This was done by correction of the estimated gamma rays per minute for the efficiency of the 3- by 3-in. Nal crystal at a particular energy. Further selection was made by determining if interfering gamma energies were present that produced unresolvable spectrain the region of interest. Of the heavy elements, 7°Np was the only nuclide high enough in yield for analysis. However, because of its short halflife, no attempt was made to resolve this nuclide. The induced activities were too low for spectral analysis. Amongthe fission products listed in Table 2, **Mo also was too low for spectral analysis. The radionuclides Ru, 1, and "Cs were not detected, but Zr and '!°Ba were resolved. Also '*Te, ‘“!ce, 4Ce, and ‘4"Nd can be resolved. The ‘Ba and »Zr fissions for two high-yield shots have been cal- culated‘ from gamma spectra obtained from 21 particles from shot A and 45 particles from shot B. Peak resolution was confirmed by decay of the respective peaks. Three of the particles were sacrificed after gammastripping and were radiochemically analyzed for ‘Ba and **Zr. Agreement between physical and radiochemical measurements was + 5%. The logarithm of ‘!°Ba fissions was plotted as a function of the logarithm of the particle diameter (Figs. 4 and 5), and a linear regression calculation was performed. The slopes of the regression line obtained for the shots A andBdata were 2.63 + 0.24 and 2.32 + 0.50, respectively. The hypotheses that the populate slopes are square or cubic (equal 2.0 or 3.0) were tested. Because ofthe scatter of the points about the regression line, both hypotheses could notbe rejected except for the hypothesis that the shot A slope was two, which was rejected at the 95% confidence interval. The intermediate value may be explained as due to the decaying of volatile precursors before particle condensation. The slopes obtained for the regression plots for Zr (Figs. 6 and 7) for shots A and B were 3.41 + 0.18 and 3.09 + 0.37. For this nuclide a cubic relation appears to hold. An explanation for the large value for 7 r for shot A is not apparent. Spectra stripping of additional particles may reduce this value.