We compared nalidixic acid, 55 mg/kg/day, with ampicillin, 100 mg/kg/day, both given orally for 5 days, in the treatment of children with dysentery caused by shigellosis. All patients entered into the study had illness of less than 72 hours' duration and no prior allopathic drug therapy. Treatment was randomized and administered in double-blind fashion. Patients initially treated with ampicillin who were infected with a Shigella strain resistant to ampicillin were considered as a separate group (ampicillin-R). All isolates were susceptible to nalidixic acid. Similar percentages of patients treated with nalidixic acid (26/32, 81%) and with ampicillin (17/22, 77%) were clinically cured by the end of therapy; the rate in ampicillin-R (3/14, 21%) patients was significantly lower (p less than 0.001). Stool frequency in patients treated with nalidixic acid was significantly less than for ampicillin-treated or ampicillin-R patients during the final 3 study days. All patients treated with nalidixic acid and ampicillin had Shigella eradicated from their stool by day 3, compared with 77% (10/13) of ampicillin-R patients (p less than 0.05, ampicillin-R vs nalidixic acid or ampicillin). We conclude that nalidixic acid is an effective alternative to ampicillin in the treatment of shigellosis caused by nalidixic acid-susceptible strains.