Previous field trials of the rhesus rotavirus (RRV) vaccine in young infants in Venezuela, Rochester, and Arizona have indicated that serotype specific immunity is necessary for protection against rotavirus diarrhea. This goal could be achieved by incorporating into a candidate vaccine (i) a strain which is highly cross-reactive with a wide range of wild type human rotaviruses, or (ii) strains that collectively represent the VP7 serotype repertoire of epidemiologically important rotaviruses. The second approach was investigated in phase 1 studies designed to evaluate the antigenicity and reactogenicity of human rotavirus-RRV reassortants with the VP7 specificity of human serotypes 1 (D x RRV), or 2 (DSI x RRV). These studies were carried out in Venezuela and Peru. The 2 reassortant vaccines were compared with the candidate RRV vaccine. A quadrivalent vaccine composed of RRV, D x RRV, DSI x RRV, and a serotype 4 x RRV reassortant (ST3 x RRV) has also been tested in Venezuela. In addition, we have tested a new vaccine candidate developed in LID, the M37 strain, originally isolated from an asymptomatic newborn infant.