Natural ecosystems are characterized by variability in space and, to a lesser extent, in time. Soil types and properties vary on landscape and field scales. Such variability is either inherent or human-induced and influences crop management. Variability at the plot level impinges on field experimentation. In this study at the main station, Tel Hadya, of the International Center for Agricultural Research in the Dry Areas (ICARDA), we intensively sampled four large field plots and analyzed the samples for soil organic matter (SOM), available Olsen phosphorus (P), and extractable potassium (K). Variability parameters were established for each property, as well as semivariograms. Contour maps for each plot were drawn by kriging. Despite the history of uniform management in each plot, large residual spatial variability was evident for each parameter. Although such variability is generally disregarded by agronomists and crop scientists and is generally believed to be eliminated by an alternate year of uniform cropping (i.e., "homogenization"), the residual persistent variability in these parameters cause variation in crop growth and need to be considered in designing experimental treatments, especially with regard to plot size and orientation. Although P and K variability can be masked by uniform fertilizer application, the effects of SOM persist for several years and thus can influence nutrient availability and soil moisture under such dryland conditions. Systematic soil sampling can help in reducing variability at trial sites, but background variability cannot be eliminated completely in natural bodies such as soils.