To study the effects on the fetus of variations in maternal glucose tolerance, a 25 g rapid intravenous glucose tolerance test was performed at or about 32 weeks gestation in 917 randomly selected nondiabetic women with singleton pregnancies. The results were withheld from the patients and their obstetricians and paediatricians, and no treatment or advice was offered. Fasting plasma glucose and indices of glucose disposal (including a new index which we have termed "summed glucose") were distributed unimodally, with no evidence of a separate pathological group towards the diabetic end of the distributions. Significant associations were found between maternal glucose metabolism and various measures of neonatal nutrition and morbidity, including the incidence of congenital malformations and morbidity related to asphyxia, suggesting that variations within the normal range in maternal glucose metabolism can influence growth and development in the fetus. These relationships were continuous throughout the range of maternal glucose tolerance and were not of predictive value in individual cases.