It is apparent, as would be expected, that the means for results for the top 1-cm samples are less than those at the respective sites for the top Thus, it was questioned whether the standard deviations 5-cm samples. Based on graphical analysis and the would be dependent on the means. Pearson product correlation coefficient (see Appendix 3), it was concluded that the means and variances were not independent. Therefore, the sample coefficient of variation (C, standard deviation divided by the mean; i.e., normalizes for differences in the means) was used for testing the variability of the techniques. Variability of Deposition in a 2- by2-m Plot Versus a Site Table 3 indicates results for the top l1-cm samples for Site I and results for five samples all taken from within plot 21 of Site I (i.e., I-21). The coefficient of variation for the five samples taken from throughout the site is numerically almost twice that for the samples taken within the plot (the difference in the coefficients of variation are not signifi- cant at the 95 percent confidence level). Although this is only one point of comparison, it appears there is more variability within the 12by 14-m site than within the 2- by 2-m plot for the l-cm sampling depth. Table 3. Comparison of Sampling 2- by 2-m Plot Versus 12- by 14-m Site Plutonium-239; Top 1-cm Depth Standard Mean Sample (£Ci/m2) Coefficient Deviation (f£Ci/m?) of Variation Plot I-21° 0.28 0.09 0.31 site 1° 0.42 0.29 0.69 Basis (each sample composed of fifteen 30-cm2 subsamples) a. Average of five samples from Plot I-21 (2 by 2 m). b. Average of five samples from five separate plots (one sample per plot) within Site I (12 by 14 m). Compositing of Subsamples Figure 3 presents data from the 0- to 5-cm subsamples for Sites I and III. The sample results are based on analyses of aliquots from the individual subsamples prior to compositing for the plot sample (1-1 and III-14). The C's for both groups of subsamples are about 35 percent larger (40 and 28, respectively) than the C's for the total plots (the differences are not statistically significant). It should be noted that 656