Knowledge of the movement of Gibberella zeae (Fusarium graminearum) from a local source of inoculum in infested cereal debris is critical to the management of Fusarium head blight (FHB) of wheat. Previous spatial dissemination and infection studies were unable to completely distinguish the contributions of released inocula from those of background inocula. Clones of G. zeae were released and recaptured in five wheat fields in New York and Virginia in 2007 and 2008. Amplified fragment length polymorphisms were used to track and unambiguously identify the released clones in heterogeneous populations of the fungus recovered from infected wheat spikes collected at 0, 3. 6, and >= 24 m from small-area sources of infested corn residues. The percent recovery of the released clones decreased significantly at fairly short distances from the inoculum sources. Isolates of G. zeae recovered at 0. 3, 6, and >= 24 m from the center of source areas shared 65, 19, 13, and 5% of the genotypes of the released clones, respectively. More importantly, the incidence of spike infection attributable to released clones averaged 15, 2, and <1% at 0, 3, 6, and >= 24 m from source areas, respectively. Spike infection attributable to released clones decreased an average of 90% between 3 and 6 in from area sources of inoculum, and the spike infection potential of inocula dispersed at this range did not differ significantly from background sources. Our data suggest that FHB field experiments including a cereal debris variable should incorporate debris-free borders and interplots of at least 3 m and preferably 6 m to avoid significant interplot interference from spores originating from within-field debris.