Three papers by Hoskins and collaborators published in The Lancet in the 70s, have been challenging the common policy to annually vaccinate people at risk with inactivated influenza virus vaccine. From an analysis of a vaccination campaign in adolescent pupils of a boarding school and four influenza outbreaks in the period 1970-76, Hoskins et al. concluded that annually repeated vaccinations would not confer protection against epidemic influenza in the long-term ('Hoskins' Paradox'). A review of the papers revealed, however, that most of the study subjects were not consequently vaccinated every year and that most of the presented data were, therefore, not relevant for the problem of annually repeated influenza vaccination. When applying strict definitions of single vaccination (immunised immediately prior to the epidemic, but not in the years before) and multiple vaccination (immunised immediately prior to the epidemic, and also in the year(s) before), only two of four epidemics (A/England/42/72 (H3N2) in 1972/73 and A/Port Chalmers/1/73 (H3N2) in 1973/74) could be evaluated: in one case, no negative effect of repeated vaccination could be detected, in the second case, the attack rate difference between groups with single and multiple vaccination was of borderline significance. Data on two other epidemics (B/Hong Kong/8/73 in 1973/74 and A/Victoria/3/75 (H3N2) in 1975/76) could not be interpreted because of incomplete vaccination strategies. In conclusion, Hoskins' Paradox cannot be substantiated by Hoskins' own data. Considering other published data on the subject, it is suggested that no negative effect of annually repeated vaccination on protection against epidemic influenza exists.