Experimental Echinococcus multilocularis infection and deworming was repeated three or five times in nine dogs at various re-infection schedules. The mean number of worms decreased more than 91% in dogs with repeated infection, compared to first infection controls (n= 6). The copro-antigen assay and the egg count in the faeces suggested that the worm burden gradually decreased each time the dogs were re-infected. To examine whether such worm exclusion was a non-specific response, five dogs were sequentially infected with the parasite four times and subsequently fed freely for 6 months. Even after the 6-month interval, the five dogs that were infected five times with the parasite were still able largely to exclude the adult worms. The results suggested that the ability of worm exclusion in dogs that developed a resistance did not become rapidly extinct. Observation of the condition of faeces and the excretion of hooks in the faeces of repeatedly infected dogs revealed that the exclusion of worms started at the first week after the re-infection, and it continued during the patent period. Serum antibodies specific to the parasite antigen increased gradually until the third infection and significantly decreased during the 6-month interval. There was little enhancement of serum antibodies after the fifth infection in most dogs, although no clear correlation was observed between the antibody response and the worm burden. These findings suggested the possibility of developing a vaccine.