Thirty isolates of Fusarium oxysporum were recovered from six substrates: (i) tap roots of Amaranthus hybridus showing symptoms of root rot, (ii) side roots of A. hybridus showing symptoms of root rot, (iii) soil surrounding plants of A. hybridus, (iv) pigweed weevils (Hypolixus haerens) that feed on A. hybridus, (v) rhizosphere of maize plants growing adjacent to a field of amaranth, and (vi) rhizosphere of dry bean plants growing adjacent to a field of amaranth. The isolates were characterized by means of pathogenicity tests, isozyme analysis, and vegetative compatibility group (VCG) tests. In the pathogenicity tests, toothpick tips were infested with F. oxysporum and inserted into amaranth stems. All 30 isolates were pathogenic on A. hybridus, with significant differences in pathogenicity based on lesion length measured 4 weeks after inoculation. The isolates were grouped into nine VCGs by complementation tests using nitrate nonutilizing mutants. Self-incompatibility was not observed for any of the isolates. The most common VCG was VCG1, which comprised 20 of the 30 isolates tested. The second most common group was VCG3, which included three isolates, while the remaining seven VCGs each consisted of a single isolate. The results indicate that the population of the amaranth root rot pathogen examined in this study is relatively homogeneous. Results of the isozyme analysis supported the results of VCG tests. Three major groups were delineated within the 30 isolates of F. oxysporum following cluster analysis of electrophoretic phenotypic values for seven isozymes (EST, IDH, G6PDH, ACP, PEP1, PEP2, and PEP3) tested. No relationship was found between isozyme phenotype and the substrate from which the isolates were recovered.