p s a t M i i Last year the rotavirus (RV) field sustained a profound surprise and setback when the first licensed RV vaccine (Rotashield–Wyeth Ayert) was withdrawn from the market because of its temporal association with very rare cases of intestinal intussusception which mostly occurred in the first week after administration of the first vaccine dose [reviewed in Nakagomi (2000)]. The impact of the RV vaccine on the total incidence of intussusception (attributable risk of Rotashield vaccine) has not yet been thoroughly evaluated but could be very small (Nakagomi, 2000). Nonetheless, given the reported strength of the temporal association, it seems unlikely that this vaccine will ever be reconsidered for widespread use in the United States. Hence, the need for a new vaccine or vaccines is great. Since the Rotashield vaccine also had minor side effects (low-grade fever in children, primarily after the first dose) (Vesikari et al., 1999) and induced incomplete protection (Bresee et al., 1999; Conner et al., 996), the obstacles to the development of an ideal RV accine are not trivial (Table 1). In this brief review, we ill address selected immunological problems associted with the development of a successful RV vaccine. Studies in naturally infected children and in animals ave shown that intestinal antibodies are pivotal in proection from RV reinfection (Conner et al., 1996; Coulson t al., 1992; Feng et al., 1994; Franco and Greenberg, 999; Matson et al., 1993). As opposed to systemic memry antibody responses, antiviral intestinal IgA reponses in people are generally of short duration (Coulon et al., 1992; Murphy, 1999) and this is one of the likely explanations for the high rate of reinfection and disease in young children and some adults (Conner et al., 1996; elazques et al., 1996). Because T-cell-independent B ell responses are short lived, it is tempting to speculate