Low mid-upper-arm circumference (MUAC), determined on the basis of a fixed cut-off value, has commonly been used as a proxy for low weight-for-height (wasting). The use of a fixed cut-off value was based on the observation that MUAC showed small age- and sex-specific differences. However, in 1993, a WHO Expert Committee concluded that age independence is not reflected in the true pattern of mid-upper arm growth, recommended the use of MUAC-for-age, and presented age- and sex-specific MUAC reference data developed with observations obtained from a representative sample of children in the USA aged 6-59 months. In this article, we explain the methodology for the development of these data, present age- and sex-specific growth curves and tables and discuss the applications and limitations of MUAC as a nutritional indicator. To develop the reference data, estimates were first obtained for the mean and standard deviation of MUAC for each month of age using 7-month segmental regression equations; a 5th-degree and a 3rd-degree polynomial in age was then used to describe the mean and standard deviation, respectively, of MUAC-for age. These curves show important age-specific differences, and significant sex-specific differences for boys and girls < 24 months of age. Correct interpretation of MUAC with regard to nutritional status requires the use of MUAC-for-age reference data such as those presented here.