This review aims to quantify the risks of mortality and morbidity associated with intrauterine growth retardation (IUGR). Twenty-nine data sets with birth-weight-specific mortalities are examined to determine whether consistent patterns of risk emerge when data from different populations are compared. Measures of mortality risk are also made with birth weight as a dichotomous variable. Twelve data sets are presented. From the data available, it is estimated that for term infants weighing 2000-2499 g at birth, the risk of neonatal death is 4 times higher than for infants weighing 2500-2999 g, and 10 times higher than for infants weighing 3000-3499 g. The risk of postneonatal death in term infants weighing 2000-2499 g is estimated to be 2 times higher than for infants 2500-2999 g, and 4 times that of infants weighing 3000-3499 g. Estimates of risk for IUGR infants are less consistent than for preterm infants. This could be due to methodological differences, particularly smaller sample sizes in the studies in developing countries, or may reflect real variation in risk. The latter may be associated with the heterogeneity of IUGR across populations, or to varying risks depending, for example, on which infections predominate or infant age at peak prevalence. IUGR is most prevalent in developing countries and the review therefore focuses on morbidity from diarrhoeal and respiratory infections. Data from nine studies are presented. There is an increased risk of diarrhoea in term infants < 2500 g and an increased risk of pneumonia. The risks of morbidity and mortality appear to differ depending on whether infants are wasted or stunted at birth. Stunted infants of low birth weight have higher neonatal mortality than wasted newborns, but this could be due to inclusion of infants with congenital anomalies who are often stunted. Wasted infants are more prone than stunted infants to neonatal morbidity. No comparative postneonatal data were located.