Maternal and fetal magnesium homeostasis is reviewed. Current evidence suggests that pregnancy-associated growth is unlikely to cause maternal magnesium deficiency and that the case for magnesium supplementation during pregnancy is unproven. Similarly, data do not support the use of magnesium in prevention of preterm labor or its preferential use in tocolysis. Magnesium might have a role to play in the prevention/treatment of eclamptic seizures. Hypomagnesaemia occurs in poorly controlled diabetic pregnancy and may be part of the cause of the hypomagnesaemia in the infants of such mothers. Fetal magnesium homeostasis is poorly understood as is the mechanism of placental transfer. Data from the rat suggest that the bulk of maternofetal placental magnesium transfer occurs via a transcellular route utilising a Na+/Mg2+ exchanger and that maternofetal flux of magnesium is reduced in the presence of maternal diabetes mellitus. Further study of diabetic pregnancy will increase our understanding of magnesium homeostasis both in normal and abnormal pregnancy.