[Synopsis] The correlation of deoxynivalenol (DON) and its incidence with Fusarium headblight of wheat and barley is very strong. However, the correlation with disease severity is weak. The focus of this presentation is on the role of DON in pathogenesis. Single-kernel analysis is the model system developed for the study of this disease. In natural infection, DON is found in spikelets other than those showing symptoms and is most likely due to spread of the fungus. In wheat and barley heads inoculated with DON, no translocation of DON out of the treated spikelets was noted after 12, 24 and 48 hrs. Macroconidia of F. graminearum isolates do not contain DON although these same isolates when grown on rice or wheat produce the toxin. In inoculated wheat and barley heads, DON as well as 15-ADON are produced 48 hrs after inoculation. Trichodiene mutants that do not produce DON did not do so when inoculated into single spikelets on wheat heads although they did show mild symptoms of infection. The wild type (DON producer) caused more severe symptoms of infection than the mutant. This manuscript attempts to review some of the pertinent literature relative to pathogenesis of Fusarium species on cereal grains as well as include some original experiments designed to address the role of deoxynivalenol in pathogenesis.