The outcome of host-parasite interactions may depend not only on the genotypes of the species involved but also on environmental factors. We used the fungus Cryphonectria parasitica, the causal agent of chestnut blight, and its hyperparasitic virus, Cryphonectria hypovirus-1 (CHV1), to test for genotype-by-genotype-by-environment interactions in a host-parasite system. In C. parasitica, infection with CHV1 induces a hypovirulent phenotype with reduced virulence toward the chestnut tree (Castanea spp.) and thus controls chestnut blight in many European regions. In contrast, uninfected virulent C. parasitica have nearly eradicated the American chestnut in North America. We applied a full factorial design and assessed the fungal growth and sporulation of four C. parasitica strains, uninfected and infected with each of the four known CHV1 subtypes, at 12°, 18°, 24°, and 30°C. We found a significant (P ≤= .0001) genotype-by-genotype-by-environment interaction, demonstrating the potential for a selection mosaic. As a consequence, different host and parasite genotypes would be selected under different climatic conditions, affecting the coevolutionary dynamics of the host-parasite interaction and the course of chestnut blight epidemics. Genotype-by-genotype-by-environment interactions are essential to take into account when designing biological control strategies.