Previous research on the association between illness and depression in older people has relied on self-reported diagnoses with their inherent limitations in scope and reliability. This longitudinal study examined the association between depressive symptoms and medically-diagnosed chronic physical and neurodegenerative diseases and disability in community-living older people. In 1992, a random sample of 299 people aged 75 years and older underwent a clinical interview and medical examination by a physician experienced in geriatric medicine. This was repeated in 1995. The examination included diagnoses of chronic active physical illness, a standardized neurological examination, an assessment of functional disability and an abbreviated neuropsychological assessment. Depressive symptoms were assessed by the Centre for Epidemiological Studies-Depression scale. Multivariate analyses of concurrent associations, longitudinal associations and the impact of incident disease showed very few independent connections between individual diseases or the total burden of disease and depressive symptoms. However, in every circumstance, disability had a marked independent impact on depressive symptoms. Insofar as the diseases examined in this study are concerned, the association between physical disease and depressive symptoms in older people appears to be mediated through disability; disease may or may not result in disability but disability is likely to lead to increased depressive symptoms.