The use of oral rehydration solutions containing essential electrolytes and either glucose or sucrose of equal osmolality was compared in a double-blind sequential trial of 784 children with rotavirus-associated diarrhea treated at a center in rural Bangladesh. The oral fluid failure rate was 11.5% for the sucrose-containing solution group and 7.3% for the glucose-containing group (P = NS). Vomiting was a significantly more common cause of failure for the group treated with sucrose-containing oral rehydration solution and was associated with an increased rate of intake of the sweeter sucrose-containing solution. The purging rate was not different for the two groups. The oral fluid failure rates for children in the most underweight category (<60% of expected weight for age) were not different from those for other groups, although, as assessed by purging rate and initial dehydration, the stool losses of members of this group constituted a greater proportion of their body weight. Glucose is the preferred carbohydrate for oral electrolyte solutions, although sucrose can be substituted with only minimum loss of efficacy.