To evaluate ceftriaxone for the treatment of shigellosis, 94 adult males with acute dysentery were randomly assigned to receive ceftriaxone (1 g), ampicillin (4 g), or saline placebo intravenously in single doses in a double-blind design. Stool cultures were positive for Shigella dysenteriae in 52 patients, S. flexneri in 38 patients, and other species in 4 patients. Both ceftriaxone and ampicillin caused reductions in the mean duration of fever and the means of daily stool frequency 2 to 4 days after therapy versus placebo (P less than 0.05). The ability of ceftriaxone to reduce stool frequency during 6 days after treatment was significant in patients with S. flexneri infections (P less than 0.05), whereas S. dysenteriae infections were relatively refractory to improvement by both antibiotics. Neither drug had a significant effect on overall duration of diarrhea, blood in stool, or tenesmus. Ampicillin reduced the mean duration of positive stool cultures after treatment from 2.6 days in the placebo group to 1.1 days (P less than 0.05), whereas ceftriaxone did not affect the duration of Shigella sp. excretion. These results indicate that single intravenous doses of ceftriaxone and ampicillin caused some clinical improvement in acute shigellosis but only ampicillin exerted a bacteriological effect on Shigella sp. excretion.