The time interval required to develop immunity after vaccination, in the event of a bioterrorist attack using variola virus, is yet to be clarified. In this article, a historical vaccination study conducted in Japan in 1929 was re-examined. Forty-four previously vaccinated and 44 unvaccinated children were involved. After successful first round primary (or re-) vaccination, all children underwent revaccination at variable intervals. Absence of a major reaction (vaccine 'take') after revaccination was taken as a sign of immunity conferred by first round primary (or re-) vaccination. Univariate analysis was employed to examine the relationship between vaccine 'take' and the exposure variables. Maximum likelihood estimates of the time period required to develop immunity were obtained using a simple logit model. The interval between vaccinations was significantly associated with vaccine 'take' in both the previously unvaccinated (p < 0.01) and vaccinated (p < 0.01) groups, and the median interval required for immunity after vaccination was estimated to be 6.4 [95% Confidence Interval (CI): 5.8, 7.1] and 4.3 days (95 % CI: 4.1, 4.7), respectively. Obtained estimates were consistent with previous observations, and the logistic fits reasonably explained the discrepancy among previous suggestions. The findings suggest that it is necessary to vaccinate exposed susceptible individuals within 3 days after exposure to ensure disease protection, and within at least 5 days (for those previously unvaccinated) to provide a certain level of protection; the probability shows a dramatic decline hereafter.