Summary: The purpose of this paper is to examine potential advantages and disadvantages of anthropometric indicators of high risk of delivering low birth weight babies (LBW ≤ 2.5 kg). Indicators and high risk limits selected for this analysis were: weight gain during pregnancy (equal or less than 16th percentile for gestational age equivalent to 8 kg at term); arm circumference (equal or less than 23.5 cm at any point in pregnancy); uterine height (equal or less than 10th percentile for gestational age); and weight gain expressed as percentage of standard weight for height (equal or less than 90 per cent at the beginning of pregnancy).The results suggest that:. 1. Arm circumference, independently of gestational age, has similar sensitivity and specificity as weight gain during pregnancy for gestational age. Uterine height and weight gain for standard height may perform a little better.At the community level combinations of arm circumference either with uterine height or with weight for height (at health services level) would produce better results. Such a combination of measures would help to avoid four to five times more LBW babies for a given input of resources as compared with no use of risk indicators. Therefore, the hypothesis is proposed that this last combination of risk indicators is the best feasible option to use in poor, malnourished populations with high prevalence of LBW babies. © 1988 Oxford University Press.