A double-blind treatment study comparing ampicillin trihydrate and placebo was performed in 373 outpatients with acute diarrhea. No pathogens were isolated from 58.4% of patients. Shigella strains were recovered from 27.1% of patients, enteropathogenic serotypes of Escherichia coli were isolated from 9.1% of cases, and 3.2% of patients excreted Salmonella. Two pathogens were present in the stools of 2.1% of patients. Bacteriologic failure occurred in 8% of patients with shigellosis treated with ampicillin compared with a 70% failure rate in placebo-treated patients. Clinical failure developed in 5% of ampicillin-treated patients and in 40% of those who received placebo. Ampicillin is as effective in outpatients with milder shigellosis as it is in patients hospitalized with severe disease.