During a deworming study in northern Bangladesh, the prevalence of Giardia duodenalis increased significantly among three groups of children receiving periodic treatment with mebendazole. This was first observed in the treatment group (Group A) of a 12-month study. Giardia increased from 5 to 31 per cent in Group A (n = 49 chi 2 = 10.42, P < 0.005), but did not increase significantly in the placebo group (Group B) (11 to 18 per cent; chi 2 = 0.4, P = NS). The prevalence of Giardia in the treatment and placebo groups was not significantly different at the beginning of the study (n = 89, chi 2 = 0.8, P = NS). During a 6-month follow-up study, children in Groups A and B received 2-monthly treatment with mebendazole. Within 6 months the prevalence of Giardia increased significantly in Group B from 18 to 44 per cent (chi 2(1) = 6.14, P < 0.05). The prevalence of Giardia was also examined in 265 children at the end of an 18-month deworming study in the same area and found to be significantly higher among the treatment group (Group C) compared with the placebo group (Group D) (38 v. 21 per cent, chi 2 = 8.87, P < 0.005). Finally, the prevalence of Giardia was significantly higher in Group B after being treated than in Group D which had not been treated (44 and 21 per cent, respectively, chi 2 = 9.08, P < 0.005). These children were from the same villages and measured during the same month of the monsoon season.(ABSTRACT TRUNCATED AT 250 WORDS)