A strain of the fungus Beauveria bassiana (Balsamo) Vuillemin (Deuteromycota: Hyphomycetes) isolated from varroa mites, Varroa destructor Anderson & Trueman (Acari: Varroidae), was used to treat honey bees, Apis mellifera L. (Hymenoptera: Apidae), against varroa mites in southern France. Fungal treatment caused a significant increase in the percentage of infected varroa mites compared with control treatments in two field experiments. In the first experiment, hives were treated with a formulation containing 0.37 g of B. bassiana conidia per hive and in the second experiment with a dose of 1.0 g of conidia per hive. The percentage of infected varroa mites also increased in the nontreated (control) hives, suggesting a movement of conidia, probably via bee drift, among the hives. Mite fall was significantly higher among treated hives compared with control hives on the sixth and eighth days after treatment in the first experiment. These days correspond to previously published data on the median survivorship of mites exposed to that fungal solate. The interaction of treatment and date was significant in the second experiment with respect to mite fall. Increases in colony-forming unit (cfu) density per bee were observed in all treatments but were significantly higher among bees from treated hives than control hives for at least a week after treatment. The relationship between cfu density per bee and proportion infected was modeled using a sigmoid curve. High levels of infection (>80%) were observed for cfu density per bee as low as 5 × 10² per bee, but the cfu density in hives treated with 0.37 g generally dropped below this level less than a week after treatment.