Food web models and biogeochemical cycling models have played an important role in the study of agricultural ecosystems by framing questions, interpreting empirical outcomes, and conveying in numerical and visual form the complexity of interactions among species, soil organic matter and nutrients (Parton et al. 1983; Hendrix et al. 1986; Andrén et al. 1990; Moore and de Ruiter 1991; de Ruiter et al. 1993, 1994). Three important observations that have emerged from these studies serve as the basis for this chapter. First, conventional agricultural management practices have resulted in significant losses of soil organic matter (Metherell 1992; Beare et al. 1994; Vitousek et al. 1997). Second, the patterning of nutrient flow and trophic interactions within the soil food web under conventional management practices differ from their counterparts in native or conservation management practices, in ways that favor the bacteria and their consumers over the fungi and their consumers. Third, coincident with the losses in organic matter and changes in food web interactions and nutrient flow is a decrease in stability as measured by mathematical representations derived from the interactions (Moore et al. 1993; de Ruiter et al. 1995; McCann et al. 1998; Neutel et al. 2002). This chapter focuses on these observations and with models offers propositions on how food web interactions, nutrient dynamics, and stability are related.