Flowering plants exhibit a wide variation in the resources they invest in dispersal structures (dispersal effort), but the environmental correlates remain unclear in many plants. Canopy openness is predicted to be negatively correlated with dispersal effort, because selective pressures on increased dispersal effort, including host-specific natural enemies and the paucity and/or ephemerality of safe sites, will be more prevalent in shady sites. Here, we tested this prediction using an ant-dispersed sedge, Carex lanceolata (Cyperaceae). By comparing seven populations of C. lanceolata in central Japan, we found a negative correlation between dispersal effort and canopy openness, thus supporting the prediction. A subsequent cafeteria experiment showed that a large species of ant (Formica japonica) with relatively long seed-dispersal distances tended to prefer diaspores with greater dispersal efforts, while smaller ant species with shorter dispersal distances preferred diaspores with lower dispersal efforts. These results correspond with the assumption that greater dispersal efforts result in a greater dispersal ability and indicate that among-site variations in dispersal efforts actually reflect differentiation in seed dispersal strategy rather than a neutral variation. There were more signs of rust infection caused by Puccinia spp., one of the main natural enemies of C. lanceolata, on adult plant leaves in more shady sites. A seedling transplant experiment was also performed and revealed that rust severity in sedge offspring was considerably reduced by the typical seed dispersal distance (ca. 4 m) afforded by the large ant species, F. japonica. The increased rust severity at shady sites, combined with the narrow dispersal ranges of rusts, can partially explain the negative correlation of dispersal effort with canopy openness. These results support the importance of canopy openness as a factor underlying the variations in dispersal effort seen among flowering plants.