Entamoeba histolytica is the predominant causative agent of human amoebiasis, a significant and common diarrhoeal disease among children of developing countries. Diagnosis of this illness by the microscopical detection of the parasites in stool samples is insensitive and often incapable of differentiating the pathogenic Entamoeba histolytica from the commensal parasite E. dispar. In this study, the results of testing stool samples in an ELISA, based on a monoclonal antibody, that detects E. histolytica-specific galactose adhesin were compared against the results of the microscopical examination of the same samples. The samples investigated came from 131 children (aged < 15 years) with diarrhoea, who lived in the provinces of Adana and Mersin, in southern Turkey. Overall 22 cases of E. histolytica infection, including eight that appeared negative by microscopy, were detected using the ELISA. The 16 patients considered positive for E. histolyticalE. dispar by microscopy included two who were ELISA-negative. With the ELISA results used as the 'gold standard', and assuming that all the E. histolyticalE. dispar cysts seen by microscopy were E. histolytica, microscopy had a specificity of 98.2%, a positive predictive value of 87.5% and a negative predictive value of 93.1% but a sensitivity of only 63.6%. Compared with microscopy, culture and PCR-based assays, the antigen-detection ELISA appears to be easier, faster and probably more cost-effective, with high sensitivity, specificity and predictive values.