Urine specimens from 61 women with symptoms of cystitis who are infected with amoxicillin-sensitive organisms were examined by the antibody-coated bacteria assay. Patients with negative assays were randomized to receive either a single 3-g oral dose of amoxicillin or 10 days of amoxicillin, 250 mg, given by mouth four times per day (conventional therapy). Patients with positive assays received conventional therapy. All 43 patients without antibody-coated bacteria in the urine, 22 given single-dose therapy and 21 treated conventionally, were cured of their infection. Of 18 patients with antibody-coated bacteria, nine relapsed within one week of completion of conventional therapy. The results of the antibody-coated bacteria assay appear to predict the therapeutic response: both single-dose and conventional amoxicillin therapy are completely successful in patients with negative assays; in contrast, conventional therapy is ineffective in 50 per cent of patients with positive assays.