Thirty-eight patients with bronchiectasis and daily expectoration of purulent sputum despite conventional antibiotic courses were randomly allocated to receive a sachet of amoxycillin (3 g) or matched placebo twice daily for 32 weeks in a double-blind study. Nine patients (four amoxycillin, five placebo) were withdrawn from the study treatment; the response of the two patients (both on amoxycillin) withdrawn within the first six weeks was not assessed. The pretreatment characteristics of the two groups were similar. Independent assessment of overall response based on patients' diary cards showed that a higher proportion improved in the amoxycillin group (11 of 17) than in the placebo group (four of 19; p = 0.02). Patients in the amoxycillin group spent significantly less time confined to bed and away from work during treatment. The frequency of exacerbations during the study treatment phase was similar in the two groups but they were less severe than before study treatment in the amoxycillin group. There was a greater reduction in purulent sputum volume between exacerbations during the study treatment in the amoxycillin group to 20 per cent of pretreatment volume than in the placebo group (88 per cent of pretreatment volume, p = 0.008), although the concentrations of Haemophilus spp. in sputum between exacerbations was similar in the two groups. Adverse effects experienced were minor except in one patient (amoxycillin) withdrawn after developing a rash and in six patients (three amoxycillin, three placebo) who had diarrhoea lasting more than one week necessitating withdrawal of two patients (one amoxycillin, one placebo) from study treatment. Sputum and stool cultures collected regularly during the study showed no important changes in the bacterial flora in either group. Prolonged higher-dose antibiotic therapy in these patients with severe purulent bronchiectasis significantly reduced the host (patient) inflammatory response to colonizing microorganisms and reduced morbidity.