Abstract
Famine early warning systems benefit from a variety of indicators which together signal the initial stages of food stress for particular population groups. Anthropometry has been used as an indicator in early warning systems, but there are inherent problems in its use which should be understood. Using data from Turkana pastoralists of northwest Kenya, this paper discusses the problems of: time lag between food shortages and changes in body size and composition; use of reference points; accurate age assessment; and establishment of baseline data. Diet composition data are suggested to be an additional nutrition-oriented indicator of impending food stress and one in which problems associated with anthropometry are not inherent. Both measures may be useful in monitoring a population, but their strengths and weaknesses should be appreciated.