A double-blind, placebo-controlled trial of oral rhesus rotavirus vaccine at a dose of 10(4) plaque-forming units was performed in 114 young infants in Maryland. Significantly more vaccinees than controls had fever and vomiting during the week after vaccination, but these reactions were mild. Of the vaccinees, 83% had a fourfold or greater rise in neutralizing antibody to rhesus rotavirus vaccine and 69% shed vaccine virus. Seventeen percent of the vaccinees and 24% of controls had rotavirus-positive diarrhea during the 2 years of surveillance. Vaccine efficacy was therefore 29% (95% confidence limits, -31% to +66%). Stools from 12 of 13 episodes containing sufficient antigen to type were serotype 1. We conclude that rhesus rotavirus vaccine was infective, immunogenic, and probably acceptably attenuated but that this serotype 3 vaccine provided little heterotypic protection during serotype 1 outbreaks in the community.