Gaeumannomyces graminis var. triciti (Ggt), the causal agent of take-all in wheat, is difficult to detect accurately and rapidly due to its similarity to fungi in the Gaeumannomyces-Phialophora complex. My objectives are to detect the fungus in infested plants and soil, and to predict effective combinations of bacteria as biological control agents. Detection was based on avenacinase-based primers and polymerase chain reaction (PCR) conditions specified by earlier research. PCR conditions were modified to effect detection. The annealing temperature was lowered from 68 to 62°C for plant and soil extracts, and the concentration of Taq polymerase was doubled for soil extracts. The lowest detection limit for plant extraction was with plant grown on 4 g Ggt-infested millet seed per kg soil, and that for soil extraction was 16 mg of purified Ggt DNA per g soil. Chemical and cultural control methods are currently inadequate. Biological control using bacteria is an alternative. Combinations of several bacterial strains are expected to work better than a single strain, but they may be less effective if bacteria antagonize each other or compete for the same rhizosphere habitat. Antagonism of potential biological control agents were assessed using a Petri plate assay. To estimate possible habitat competition, nutritional profiles of the strains were evaluated using the BIOLOG system. I hypothesized that bacteria not antagonistic to each other and having low coefficients of nutritional similarity would make better biological control combinations. Six bacterial combinations gave better mean root weight in the greenhouse experiment but not in the field.