Anthropometric data were collected in a sample of healthy adults (23 men and 21 women), 54 to 85 years of age. Compared to groups of younger individuals, these older adults have more frequent and larger inter-observer errors for most body measurements. Repeated measurements reduce the incidence of large inter-observer differences for measures of body size. Accurate anthropometry is important because of relationships to body stores of fat and muscle. In recent years, there has been an increase in the use of anthropometry and more direct measures of body composition in studies of nutritional status in the elderly (Steen et al. 1977, 1979; Fisher et al. 1978; Yearick, 1978; Calloway and Zanni, 1980; MacLenman et al. 1980; Lee et al. 1981; Garry et al. 1982; Mitchell and Lipschitz, 1982). The anthropometry used in these studies and considered essential at younger ages, such as upper arm circumference, waist circumference and triceps skinfolds become increasingly difficult in older adults. As a result of this difficulty, it is possible that the errors of these and other measurements are greater among the elderly than among younger adults. However, a consideration of measurement errors in studies of the elderly using anthropometry has received little or no attention (Friedlaender et al. 1977). The present study reports the errors of measurement for anthropometry collected from a sample of older adults.