The effects of a nucleotide-supplemented formula on diarrhoeal disease was studied in 141 infants (group 1) who belonged to the low socioeconomic stratum; 148 controls (group 2) received the same formula but unsupplemented. Group 1 experienced less episodes of diarrhoea (109 versus 140), including less first episodes (74 versus 102; chi-square = 8.19, p < 0.004; odds ratio 2.01) and for a lesser number of days (807 versus 996 days); 45.0% and 31.1% of infants in groups 1 and 2, respectively, never developed episodes of diarrhoea. There were no differences in the clinical characteristics of the episodes or in the enteropathogens isolated from symptomatic or asymptomatic infants. The mechanisms through which nucleotides decrease the incidence of diarrhoeal disease in infants remain unclear.