To determine incidence and remission rates of insomnia in older adults and associated risk factors. Three-year longitudinal study, 1982-198--East Boston, MA; New Haven, CT; Iowa and Washington counties, IA. Participants were 6,899 men and women aged 65 years and older. Self-reported difficulty falling asleep or early morning arousal (insomnia), along with physician diagnosis of heart disease, stroke, cancer, diabetes, or hip-fracture, self-report of physical disability, depressive symptomatology, perceived health status, and use of medications ascertained at both baseline and three-year follow-up. Nearly 15% of the 4,956 participants without symptoms of insomnia at baseline reported chronic difficulty falling asleep or early morning arousal at follow-up, suggesting an annual incidence rate of approximately 5%. Incident insomnia was associated with depressed mood, respiratory symptoms, fair to poor perceived health, and physical disability. In multivariate analyses, these risk factors explained the higher incidence of insomnia among those with medical conditions such as heart disease, stroke, and diabetes. Other factors associated with an increased risk of insomnia included use of prescribed sedatives, and widowhood. Only 7% of the incident cases of insomnia occurred in the absence of associated risk factors. Of the nearly 2,000 survivors with chronic insomnia at baseline, almost half no longer reported symptoms upon follow-up and were more likely to report improved self-perceived health compared to those who continued to report symptoms. Chronic disease, depressed mood, physical disability, poor perceived health, widowhood, and use of sedatives are associated with development and remission of insomnia symptoms. Because the vast majority of incident cases of insomnia were among persons with one or more of these risk factors, these data do not support a model of incident insomnia caused by the aging process per se.