Dietary restriction in cases of mild infantile diarrhea is often advocated but has not been shown to be effective. We enrolled 176 healthy infants less than 1 year of age in a randomized controlled trial to determine the effects of diet on the course of mild diarrhea. When diarrhea occurred (56 episodes), infants were randomly assigned to a treatment diet (24 hours of electrolyte solution then dilute soy formula, dilute cow's milk formula, or undiluted soy formula) or their usual formula. Parents recorded daily weights, stool losses, and oral intake. The difference in effects of unrestricted and treatment diets was small and not clinically significant. Patients on an unrestricted diet averaged 0.7 fewer days of diarrhea (95% confidence interval [CI], -1.43 to 0.02), 5.0 fewer total stools (95% CI, -10.26 to 0.33), and 1% less weight loss (95% CI, 2% to -1%) compared with those receiving a treatment diet. Treatment failures were similar in both groups. An unrestricted diet does not appear to affect the course or symptoms of mild diarrhea.