Simulations based on arithmetic means and standard deviation and simulations based on the means and standard deviations of logarithms both result in estimates of I, which are quite close to the best empirical estimates available, i.e, ¥Yhe estimate based on the rumen contents of fistulated steers (Smith, this volume). The good agreement between these almost independent estimates of Ip tends to confirm the usefulness of the simulation model, Equations Y and 2, and the basic assumption that grazing can be represented as a random process. It does not, however, confirm the structure of the model. It could be, for example, the the soil ingestion rate is actually much lower than indicated, by this study, on the basis of but three samples. Indeed, the soil, or sediment, recovered from the gastrointestinal tracts of cattle may have been ingested not as soil per sé but rather in the form of dust deposited loosely on vegetation. As the digestibility of the vegetation consumed by grazing cattle in Area 13 is unknown, it is theoretically possible, as pointed out earlier, that the daily intake of vegetation could be closer to 8 kg than to 6 kg. If that were the case, most of the observed plutonium ingestion, 565 nCi/day, might be due to vegetation intake (i.e., 8 kg/day x 70 pCi/g = 560 nCi/day), leaving only a small amount that might be due to soil ingestion. The model can, of course, be applied to other contaminated areas at or near the Nevada Test Site, but the results of such applications will remain uncertain until methods are devised for estimating (a) the digest- ibility of vegetation available to grazing cattle in a given area and (b) the soil ingestion rate. The soil ingestion rate is difficult to measure directly, but it seems probable that some sort of estimate could be made by sampling the sediment (soil) content of fecal materials. As estimates of I, _, based on the simulation model, are obviously less expensive and less Eime-consuming than grazing studies conducted by means of fistulated steers, efforts to implement the model via independent field studies could prove to be quite rewarding. 504