Oral streptomycin-dependent shigella vaccines were evaluated for the control of endemic shigellosis caused by Shigella sonnei. In February 1972, 101 institutionalized children received four doses of S. sonnei vaccine, and 106 control cohorts received Shigella flexneri 2a vaccine. Unexpectedly, evidence of in vivo reversion of S. sonnei vaccine and of transmission of both vaccines appeared; 18 of the 101 recipients of S. sonnei vaccine asyrnptomatically excreted streptomycinindependent revertant vaccine organisms, and 6% of vaccinees excreted the heterologous strain of Shigella. Booster vaccination was given in August 1972. The February vaccine lots were used for the first dose; new lots were used for the subsequent doses. Four recipients of S. sonnei vaccine excreted streptomycinindependent S. sonnei, all after the first dose (February lot). Forty of 208 vaccinees and 18 of 68 unvaccinated cohorts excreted the heterologous vaccine strain; this excretion documented child-to-child transfer of vaccine.