Although xerostomia in older adults has received substantial research attention, there have been few longitudinal studies of non-patient populations. Consequently, little is known about the incidence or course of this condition among this group. This paper reports the results of a longitudinal study designed to address these issues. In 1989, data on xerostomia were collected from 907 randomly-selected community dwelling adults aged 50 years and over. Three years later, 611 (71% of those presumed to be alive) were followed-up and data on xerostomia collected again. At baseline, 15.5% of these 611 subjects reported xerostomia, while at follow-up this had risen to 29.5%. The majority of the latter (115/180) were incident cases, reporting xerostomia only at follow-up, while the remainder were chronic cases, reporting xerostomia at both baseline and follow-up. A crude estimate of the three-year incidence rate was 22.5%. In a logistic regression analysis, three baseline variables were associated with incidence; older subjects, those with one or more chronic medical conditions and those reporting their general health as poor were more likely to develop xerostomia. In a similar analysis, age was the only variable associated with chronicity, with older subjects more likely to be chronic cases. The data also suggest that the onset of xerostomia was associated with an increase in other oral symptoms and problems with eating, communication and social interaction.