To investigate the role of Helicobacter pylori in childhood diarrhoea, specific IgG antibodies to H. pylori (determined by an ELISA) were sought in 119 infants aged 3-36 months in Peru. Thirty one of the infants had acute diarrhoea (defined as lasting < 72 h and not present in the previous 3 weeks), 67 had persistent diarrhoea (lasting > or = 14 days with no more than 1 intervening diarrhoea-free day) and the remaining 21 had not had diarrhoea in the previous 3 weeks. The children with diarrhoea had been admitted to hospital in Lima for diarrhoea treatment, and the diarrhoea-free children for investigation of possible tuberculosis. Aspirates of duodenal contents and duplicate stool samples were investigated for the presence of bacterial overgrowth and of pathogenic bacteria, viruses and parasites. Anthropometric measurements were also made. There were no statistically significant differences between the prevalence rates of IgG against H. pylori in the children with acute diarrhoea, persistent diarrhoea and without diarrhoea (32%, 43% and 29%, respectively). In addition, H. pylori infection (as evidenced by specific antibodies) had no apparent influence on the presence of small-bowel overgrowth (in 20% of seropositive children compared with 18% of seronegative children) or of pathogens in the stool (in 53% of seropositive children compared with 49% of seronegative children) or on the occurrence of malnutrition in the groups of children considered as a whole. We conclude that H. pylori infection is not associated with acute or persistent diarrhoeal disease, small-bowel overgrowth, stool pathogens or malnutrition in Peruvian children.