To generate synthetic random samples of ly based on Equation 2, body weight (W) and digestibility (D) were assumed to vary normally with means and standard deviations of 410 + 82 kg and 0.48 + 0.12, respec- tively. Random variation of body weight simulates variations in apparent metabolic weight and appetite while random variation of digestibility simulates variations in quality of vegetation ingested. The mean body weight was chosen to match that of the 409-kg cow left in the inner compound for 177 days. A mean digestibility of 0.48 was chosen to match the Gilbert et al. (1977) estimate of I, ~ 6 kg/day (dry weight). When D is held constant and W is allowed to vary normally, synthetic samples of I_ based on Equation 2 had means close to 6,000 g/day and exhibited apparently normal distributions. When both W and D were varied normally, as indicated above, synthetic samples of I, based on Equation 2 exhibited lognormal distributions. A histogram of one such synthetic sample is shown in Figure 1.* The arithmetic mean of this sample (n = 100) was 6,760 + 2,439 g/day. The median, indicated by exp (xg), was 6,400 g/day. Both of these values are higher than the value obtained (6,006 g/day) by substituting the assumed means of W and D in Equation 2. Soil Ingestion Rate Cattle are known to ingest quantities of soil and a variety of other nonfood items such as flagging material, rope, and rubber boots. In some regions, cattle occasionally ingest so much soil that their gastro- intestinal tracts are blocked and massive doses of castor oil are required to relieve the situation. In earlier modeling studies (Martin and Bloom, 1976 and 1977), soil ingestion rates were conservatively assumed to be as much as 2,000 g/day, a quantity which would probably result in the problem mentioned above. The only site-specific soil ingestion data presently available were provided by Smith (1977), who reported the weights of soil recovered from the rumen and reticula of three cows which had been grazing the outer compound of Area 13 just prior to sacrifice in January, 1976. The quantities reported were 8.5 g, 57.3 g, and 278 g. As Smith aptly observed, "These data suggest that the total amount of soil ingested is much less than 2 kg per day and that a reasonable estimate would be between 0.25 and 0.5 ke." The arithmetic mean and standard deviation of 8.5, 57.3, and 278 are 115 + 144 g/day. Since the coefficient of variation (144/115) is greater than 1, a lognormal distribution is assumed. (Note also that the numbers *All the histograms, Figures 1-8, have been normalized and the "normality" of distribution has been "confirmed" by chi square tests comparing observed cell counts with frequencies expected for a normal distribution. 487