A popular pre-harvest strategy to mitigate aflatoxin contamination of corn involves field application of non-aflatoxigenic strains of Aspergillus flavus. The basis of this biological control may involve multiple factors, but competitive displacement of aflatoxigenic strains by the biocontrol strains is a likely mechanism. Three biocontrol strains (NRRL 21882, 18543, and 30797) were applied annually, over a 4-year period, to the same 3.2-ha commercial corn field in the Mississippi Delta, where we monitored their post-release establishment, spread, and persistence. Within 2 months of the first biocontrol application, the percentage of soil-inhabiting aflatoxigenic A. flavus strains in some plots was reduced from 48 to 9% of the population. The frequency of aflatoxigenic A. flavus strains was also significantly reduced in the rest of field. After 4 years, neighboring plots that had never received a biocontrol treatment, and distanced from our treatment plots by at least 20 meters, had less than 20% aflatoxigenic isolates. This significant halo effect might be attributed to movement of soil through tillage operations, but the aflatoxigenicity shift could be detected in the untreated plots within 2 months of the initial applications, at a time when there was no tillage. The A. flavus populations that colonized the grain were also monitored and found to be less than 15% toxigenic in the fourth year for all treatments. Over all treatments and years, less than 2 ppb of aflatoxin was detected, which could be a consequence of the field-wide shift of the inherent A. flavus population to predominately non-aflatoxigenic strains. This study supports the efficacy of using non-aflatoxigenic A. flavus strains as pre-harvest biocontrol, and shows that most of its effectiveness occurs with the first application.