The Australian Longitudinal Study of Ageing (ALSA) aims to identify factors that contribute to & predict the health & social well-being of older Australians. Analyses were performed to determine the predictive value of anthropometric measurements in older Australians for four-year mortality. Weight, height, skinfolds (triceps, abdominal, supra-spinale, sub-scapular, medial calf, and front thigh) & girth (arm, waist, hip, calf) measurements were performed on a randomly selected community-living sample of 772 men & 624 women aged>70 years. Waist: Hip, % weight loss, corrected-arm-muscle area (CAMA) & BMI were calculated. These measures were categorised into quartiles & also according to commonly adopted definitions of nutritional status. Cox regression analysis was undertaken to assess the predictive value of the independent anthropometric variables for four-year mortality, adjusting for potential confounders (age, gender, marital status, smoking, alcohol status, self-rated health, basic activities of daily living & co-morbidity). Risk of four-year mortality increased with weight loss >10% over two years (HR=2.53, CI=1.37-4.67) & CAMA <21.4cm2(M) & <21.6cm2(F) (HR=1.93, CI=1.03-3.60) independent of confounding variables. These results confirm that selected anthropometric indices (weight loss, CAMA) independently increase the risk of four-year mortality & highlights their potential use in the nutrition screening and assessment of community-living older adults.