This paper reports the first survey in Australia to use faecal culture to detect Dientamoeba fragilis in patients with diarrhoea. Of 3 different protozoal culture media evaluated on a case of known infection, modified Boeck & Drbohlav's medium was the most suitable. The organism could be grown from faeces stored for up to 24 h at room temperature, but for only 10 h at 4 degrees C. Culture was then used, in combination with microscopy of smears fixed with polyvinyl alcohol and trichrome-stained, to examine single stool specimens from 260 consecutive patients with diarrhoea in the city of Brisbane. D. fragilis was detected in 4 (1.5%) specimens, only 2 of which were positive by microscopy. Other protozoa were found in 36 (13.8%) specimens: Blastocystis hominis in 28 (10.8%), Giardia duodenalis in 4 (1.5%), Endolimax nana in 3 (1.2%), and Entamoeba coli in one (0.4%). One strain of Dientamoeba was cryopreserved in liquid nitrogen and recultured successfully 10 d later. Culture was more sensitive than microscopy in diagnosing D. fragilis infection and the organism, of dubious pathogenicity, was not common in patients with diarrhoea in this community. Prevalence surveys of intestinal protozoan infections should use faecal culture, and specimens should be less than 12 h old and not refrigerated. Dientamoeba strains isolated by culture can be cryopreserved.