Length, height, weight and mid-upper arm circumference (MUAC) were measured in 4320 children aged between 0 and 59 months, and their socio-economic status was assessed, in 31 villages in Southwest Uganda during March-April 1988. A follow-up survey assessed the mortality of the children during the 12 months following anthropometry. Mortality rates were higher in those with low anthropometric indices at the first survey. MUAC was the most sensitive predictor of mortality followed by weight-for-age, height-for-age and weight-for-height. MUAC increased the predictive power of other parameters whereas the other parameters did not increase the predictive power of MUAC. MUAC below 12.5, 11.5 and 10.5 cm predicted 10.9%, 18.7% and 36.5% of the deaths respectively. Nutritional status was worse in the low socio-economic group but the predictive power of anthropometry for mortality was not influenced by socio-economic status. This suggests that nutrition per se has an influence on mortality which is independent of socio-economic status.