Selective media without pentachloronitrobenzene were developed for quantitative assays of Fusariumoxysporum in soils. Media Fo-G1 and Fo-G2 were effective for naturally infested soils, Fo-W1 and Fo-W2 for wild-type isolates in soils containing a nitrate-nonutilizing (nit) mutant, and Fo-N1 and Fo-N2 for nit mutants. Selective media were made using ammonium citrate dibasic, l-sorbose, econazole nitrate, 25% iminoctadine triacetate solution and 50% tolclofos-methyl wettable powder for soil dilutions of 100-fold or more (Fo-G1, FoW1 and Fo-N1) and 10-fold (Fo-G2, Fo-W2 and Fo-N2). Potassium chlorate was added to Fo-N1 and Fo-N2. The efficacy for selectively isolating F. oxysporum was confirmed using six soils naturally infested with one of six formae speciales of F. oxysporum and with soil dilutions containing conidia of wild-type strains or nit mutants from the six formae speciales. On Fo-G1 and Fo-G2, most colonies of F. oxysporum were compact and round with purplish or reddish pigment in the reverse. Cylindrocarpon sp. formed colonies as large as those of F. oxysporum but were distinguishable by their colony morphology. Other contaminants such as F. solani, F. moniliforme, and Trichoderma were suppressed by medium ingredients and colonies of F. oxysporum. On Fo-W1 and Fo-W2, colony morphology of F. oxysporum and contaminants corresponded to that on Fo-G1 and Fo-G2, although F. oxysporum failed to produce the pigment. On Fo-N1 and Fo-N2, nit mutants formed clear colonies from 100- and 10-fold soil dilutions, respectively, and contaminants seldom formed large colonies.