Serum fructosamine was assessed in 41 previously diagnosed diabetic patients throughout pregnancy and compared with other standard measures of glycaemic control. Fructosamine was measured by a precise automated method using the reduction of nitro-blue tetrazolium at alkaline pH. Corrected serum fructosamine was calculated from serum fructosamine and serum protein concentrations in view of the falling protein concentrations during pregnancy. Serum fructosamine corrected for total serum protein was significantly lower at the end of pregnancy than at the initial visit. Clinic blood glucose, HbA1c and fructosamine corrected for serum albumin did not fall significantly. Fructosamine corrected for total protein concentration at all points through pregnancy showed positive correlations with birth weight ratio (r = 0.51-0.67). HbA1c did not show the same consistent pattern of results. The correlation of first and second trimester fructosamine values with birth weight ratio suggests that glucose control is a determinant of birth weight throughout pregnancy rather than just in the last trimester.