Both the concentrations and the stocks of soil organic carbon vary across the landscape. An estimate of the quantity of the soil organic carbon stock and the precision of that estimate are both important for many reasons but especially to improve the accuracy and precision of dynamic carbon models - that must eventually be calibrated on real carbon values measured at the same scales for which predictions are being made. In this part of our project, our objective was to compare the values of total stock of organic carbon in agriculturally managed fields and the precision of the estimated values as estimated by alternative methods. We sampled two sites in Iowa, representative of till-derived soils and loess-derived soils. Using grid-based sampling design, we sampled 14- and 17-ha fields with 7.5-cm-diameter cores to a 30-cm depth and calculated carbon stocks based on the carbon concentration and the bulk densities. For the loess-derived soils, the estimates of organic carbon in 17 ha to 30 cm depth were (1) using a soil survey database (ISPAID) - 1636 +/- 229 Mg; (2) using an estimate from the global mean of all sampling points in each field - 1430 +/- -17 Mg; (3) using an estimate after stratification of values by mapping unit - 1442 +/- 14 Mg; and (4) using geostatistical interpolation - 1614 +/- 5 Mg. How much carbon is in that field? It depends on how one makes the estimate and how much uncertainty one can tolerate. Both the estimated quantity and the precision of the estimate depend on choices that we make. Since those choices depend mainly on cost (how many points we can sample), the amount of carbon measured depends on how valuable carbon is to us. That is not really a scientific question but an ethical question about human values.