The purpose of this work is to identify risk markers of mortality in a cohort of 645 people aged 60 and over. The study was carried out in rural areas in south west France. Data were collected by questionnaire in 1982. Mortality was determined 4 years later; 111 deaths were registered. The analysis of age-adjusted odds ratios (OR) showed strong relationships between mortality and disability (OR = 7.75), compared health (OR = 3.94), self-rated health (OR = 2.47), home comfort (OR = 0.52), physical activity (OR = 0.32), sociability (OR = 0.43) and two subjective well-being items: the feeling of uselessness (OR = 3.51), and the lack of projects for the future (OR = 2.35). By contrast, no significant association was observed with reported morbidity and social support. Two multivariate analyses were performed: the first on longevity using Cox's regression model, the second on mortality using a linear discriminant analysis. The results of these analyses were translated into a simple set of 8 independent risk markers for the identification of a "high risk group" of mortality within 4 years. The sensitivity of this mortality risk indicator was 73% and its specificity 77%.