We examined rodent dispersal of Dioscorea japonica bulbils. Bulbils dispersed by rodents may travel farther from their parent plants than those dispersed by gravity alone. To confirm the presence of secondary dispersal and the types of vector, we used an automatic camera to photograph animals that approached bulbils. To clarify secondary dispersal distance, we set marked bulbils at eight sites and searched for the bulbils once they were removed from their original locations by a disperser. We took photographs of mice (Apodemus speciosus) approaching the bulbils. In June, 10 of 333 bulbils that were moved germinated. We also measured the primary dispersal distance at four sites by dropping bulbils from heights of 0.5, 3.0, and 12.0 m. We then compared primary and secondary dispersal distances. The mean secondary dispersal distance was significantly greater than the primary dispersal distance for dispersal heights of 0.5 and 3.0 m, but less than that for a 12.0-m dispersal height. This indicates that rodents may disperse bulbils farther than gravity alone when bulbils drop from lower heights but they contribute little to dispersal for bulbils that drop from greater heights (e.g., from the top of the canopy). Thus, we confirmed the occurrence of rodent dispersal of bulbils and rodents might to some extent contribute to the dispersal and establishment of bulbils.