71 = Cy = [(5.37 x 0.12565) + ... + (1230.01 x 0.02513)]; (Equation 6, the arithmetic strata means in Table 4 and the percentage area values in Table 3), 236 =1 ana exp [3.9387 + 0.5(1. 7464)2], 579 =C [(35.97 x 0.12565) + ... + (7827.83 x 0.02513) ] based on Equation 6, the arithmetic strata means of Table 4 and the percentage area values of Table 3. The mean value of I u indicated by Equation 6, i.e., 571 nCi/day, is about 1 percent more than Smith's empirical estimate, 565 nCi/day, based on the rumen contents of fistulated steers. (See the article by D. D. Smith in this volume.) It could be concluded at this point that Smith's estimate has been resoundingly confirmed by an independent estimate based on (1) a theoretical model of I_ correctly adjusted to sitespecific conditions, (2) an uncertain’estimate of I_ based on but three site-specific data points, and (3) estimates of mea plutonium concentrations in composite random samples of vegetation and soil, C andC, based on the means and relative areas (Tables 3 and 4) of the Sampling strata. It could also be concluded that the excellent agreement between the two almost independent estimates of I, confirms the validity of the random grazing assumption in spite of the Clear evidence, provided by the fistulated steers (D. D. Smith, this volume), of marked seasonal preferences for different plant species. The Problem of Estimating Digestibility While these conclusions are probably valid, "Catch 22" which must be reconsidered. there is at least one The principal adjustment of I model, Equation 2, was to set D = 0.48 + 0.12. The mean of D (digesti- bility) was selected to make it possible for a 410-kg cow to obtain sufficient food energy (kcal) to meet its digestible energy requirement on a daily ration of about 6 kg of vegetation. The actual value of D for Area 13 vegetation is not known. The large standard deviation was assumed to simulate variation of forage quality. If the mean were 0.36 as suggested by one study of desert vegetation digestibility (McKell, 1975), the daily ration required for maintenance of a 410-kg cow would be over 8 kg of vegetation and the corresponding estimate of I w Equation 4, would be 713 nCi/day instead of 571 nCi/day, an increase of 25 percent. Thercfore, practical application of the I_ model to other sites requires development of a dependable method of eStimating the mean and standard deviation of D for a given area based perhaps on the species composition and biomass of its vegetation. Simulation of Random Grazing Another question of more than passing interest is whether or not the results of vegetation, soil, and grazing studies conducted in Area 13 are reproducible. It would be impractical and expensive to actually repeat vegetation and soil studies conducted in Area 13 and also time 497