Three isolates of nonpathogenic Fusarium spp. (CS-1, CS-20, and Fo47), previously shown to reduce the incidence of Fusarium wilt diseases of multiple crops, were evaluated to determine their mechanisms of action and antagonist-pathogen inoculum density relationships. Competition for nutrients, as represented by a reduction in pathogen saprophytic growth in the presence of the biocontrol isolates, was observed to be an important mechanism of action for isolate Fo47, but not for isolates CS-1 and CS-20. All three biocontrol isolates demonstrated some degree of induced systemic resistance in tomato (Lycopersicon esculentum) and watermelon (Citrullus lanatus) plants, as determined by split-root tests, but varied in their relative abilities to reduce disease. Isolate CS-20 provided the most effective control (39 to 53% disease reduction), while Fo47 provided the least effective control (23 to 25% reduction) in split-root tests. Dose-response relationships also differed considerably among the three biocon-trol isolates, with CS-20 significantly reducing disease incidence at antagonist doses as low as 100 chlamydospores per g of soil (cgs) and at pathogen densities up to 105 cgs. Isolate CS-1 also was generally effective at antagonist densities of 100 to 5,000 cgs, but only when pathogen densities were below 104 cgs. Isolate Fo47 was effective only at antagonist densities of 104 to 105 cgs, regardless of pathogen density. Epidemiological dose-response models (described by linear, negative exponential, hyperbolic saturation [HS], and logistic [LG] functions) fit to the observed data were used to quantify differences among the biocontrol isolates and establish biocontrol characteristics. Each isolate required a different model to best describe its dose-response characteristics, with the HS/HS, LG/HS, and LG/LG models (pathogen/biocontrol components) providing the best fit for isolates CS-1, CS-20, and Fo47, respectively. Model parameters (defining effective biocontrol dose (ED50) indicated an ED50 of 2.6, 36.3, and 2.1 × 106 cgs and estimates of biocontrol efficiency of 0.229, 0.539, and 0.774 for isolates CS-1, CS-20, and Fo47, respectively. Differences in dose-response relationships among the biocontrol isolates were attributed to differences in their mechanisms of action, with CS-20 and CS-1 functioning primarily by induced resistance and Fo47 functioning primarily by competition for nutrients.