After control measures were initiated to stop an outbreak of shigellosis in an institution for the developmentally disabled, there was a sharp decline in the number of cases of Shigella sonnei infection. Among ill residents, those treated with antibiotics had shorter mean duration of diarrhea (2.4 vs. 4.5 days, P < .01) and were less likely to have stool cultures positive for shigellae 2-4 weeks after onset of diarrhea (0/25 vs. 5/19; relative risk [RR] = undefined; P = .02). The attack rate was higher in villages where segregation of ill residents was not practiced (46/73 vs. 53/155; RR = 1.8; 95% confidence limits [CL], 1.4, 2.4). In individual housing units where ill residents were not segregated (preintervention), a correlation was found between mean duration of diarrhea and unit attack rates (r = .88; 95% CL, 0.29, 0.99). A study of all 305 residents 10 weeks after the intervention began revealed no positive stool cultures.