This study was conducted to provide community-based data on maternal health and predictors of newborn weight in rural Malawi. Data were obtained prospectively from a community-based cohort of 581 pregnant women who attended an antenatal clinic and delivered a term, live-born, singleton infant in Lungwena, rural Malawi. Morbidity from infectious diseases and anaemia was common. Maternal weight gain in rural Malawi was slower but fundal height gain was comparable to that of an affluent western population. The mean +/- SD weight of term newborns was 3.2 +/- 0.5 kilograms. A regression model including data from all routine investigations explained only 24% of the variance in newborn weights, suggesting that routine antenatal measurements had a limited power to predict the size of term live-born babies. Maternal parity, initial weight, the duration of pregnancy and gestational weight gain were associated with newborn weights and should, therefore, be systematically recorded in rural Malawian antenatal clinics.