Identification of a definitive host for Neospora caninum has been inhibited by lack of an efficient method for producing bradyzoites, needed for oral infectivity trials. An improved protocol for producing bradyzoite-containing tissue cysts in mouse brains is described. Six variables, including mouse strain (Balb/C, CBA/Ca, and ICR), sex, N. caninum isolate (NC-2 and NC-Liverpool), tachyzoite inoculum dose, immunosuppression with methylprednisolone acetate (MPA), and sulfadiazine treatment were tested. Tissue cyst numbers were estimated using an immunohistologic staining procedure specific for bradyzoites. Male ICR mice (> or = 30 g) that were immunosuppressed with 2 mg MPA 7 days prior to and 2.5 mg MPA at the time of subcutaneous inoculation with 400,000 N. caninum tachyzoites produced the highest numbers of tissue cysts. Significant numbers were produced by methods using the NC-2 strain of N. caninum; however, protocols using NC-Liverpool produced greater numbers of tissue cysts. Sulfadiazine treatment did not appear to contribute to tissue cyst production. The procedure described is superior to previously described methods with regard to numbers of tissue cysts produced, protocol reproducibility, and survival of mice until tissue cyst formation.