OBJECTIVE The purpose of this study was to determine the presence of malnutrition in indian children of Alto Xingu region in Brazil, and to evaluate the usefulness of arm circumference and QUAC index in the diagnosis of protein-energy malnutrition (PEM). SITE: The indian population of Alto Xingu maintains most of its traditional customs. METHODS 335 children were studied prospectively in at least one of seven consecutive years between 1974 and 1980, and 1278 measurements were obtained of weight, height, and arm circumference. RESULTS The annual prevalence of PEM (weight-for-height) less than 90% according to Brazilian reference tables was 4.1% in children aged under 1 year, 7.3% in children from 1 to 5 years of age, and 7.5% in children older than 5. According NCHS tables, however, values of 1.0%, 2.8%, and 1.1% were obtained for the respective age groups. The sensitivity and specificity of arm circumference and QUAC index were calculated in relation to the weight-for-height relation. According to Brazilian reference tables, arm circumference had a sensitivity of 8.9% and specificity of 99.3%, whereas the QUAC index had a sensitivity of 11.1% and specificity of 98.3%. However, according to NCHS tables, slightly different values were obtained. The sensitivity and specificity in this case were 23.5% and 99.3%, respectively, for arm circumference and 35.3% and 98.3%, respectively, for the QUAC index. CONCLUSIONS The authors concluded that the prevalence of PEM was low in the Alto Xingu and that arm circumference and the QUAC index should only be used in situations suited to their sensitivity and specificity.