To investigate the relationship of hemoglobin A1c (HbA1c) to average blood glucose concentration and to birth weight of infants of diabetic mothers, HbA1c was determined in 42 consecutive insulin-dependent pregnant diabetic women in the third trimester. HbA1c correlated significantly to the average blood glucose levels in the preceding 8 wk (r = 0.73, P less than 0.001). No correlation was found between HbA1c and the relative birth weight ratio (RBWR) for all newborn infants. However, in the major subgroups of pregnancies, White class B and C without prognostically bad signs in pregnancy (PBSP), HbA1c in the third trimester significantly correlated to RBWR (r = 0.59, P less than 0.01). In a subgroup of six pregnant diabetic women in whom HbA1c and blood glucose concentrations frequently were determined from the 13th to the 33rd gestational week, a relationship between HbA1c and the average blood glucose concentration of the preceding 8 and 12 wk in the individual pregnant subject was established. In spite of this correlation, HbA1c was found to be a poor predictor of the average blood glucose concentration in the individual patient. This indicates that HbA1c cannot satisfactorily describe the degree of diabetes control without simultaneous determinations of blood glucose. HbA1c determinations were found to be of value as an additional indicator of the quality of regulation during ambulatory control in diabetic pregnancy because an increase in HbA1c suggested an impairment of diabetic control. The perinatal mortality and morbidity of the infants of diabetic mothers were satisfactory in this series, as only one perinatal death, one nonserious malformation, and two cases of mild respiratory distress syndrome occurred among the 43 infants.