Observational study examining the retrospective, cross-sectional and prospective associations between vitamin D concentration and depressed mood in a community-derived sample of 3105 older men living in metropolitan Perth, Western Australia. We measured the plasma concentration of 25-hydroxyvitamin D using standard procedures. Past depression was ascertained by direct questioning and through the use of administrative health data linkage. A geriatric depression scale score equal or greater 7/15 established the presence of current depression. Incident depression was established by a patient health questionnaire (PHQ-9) score ≥10 or by administrative health data linkage during the 6-year follow up (range 0.1–10.9 years)