Faecal microscopical diagnosis of Strongyloides and hookworm infections is insensitive. We have therefore compared duodenal fluid and faecal microscopy for detection of these parasites in a group of 292 patients being investigated for gastrointestinal symptoms who were examined by both techniques. Thirty-three of these patients (8%) were infected with Strongyloides stercoralis and 88 (30%) had hookworm infections. Microscopical examination of up to 3 faecal specimens detected only 33% and 65% of patients with Strongyloides and hookworm infections, respectively. Microscopical examination of a single specimen of duodenal fluid was more sensitive for detection of strongyloidiasis, identifying 76% of patients; the parasite was found exclusively in duodenal fluid (and not in faeces) in 67% of patients. For hookworm, the diagnostic sensitivity was similar with both techniques but duodenal fluid microscopy detected some patients (35%) who had not been identified by faecal microscopy. This study confirms previous work indicating the insensitivity of faecal microscopy in these infections and emphasizes the need to consider routine examination of duodenal fluid to exclude chronic strongyloidiasis. This may have particular relevance for south-east Asian war veterans and immunocompromised patients.