Hypovitaminosis D is common in India. In the present prospective partially randomised study of vitamin D (D3) supplementation during pregnancy, subjects were randomised in the second trimester to receive either one oral dose of 1500 μg vitamin D3 (group 1, n 48) or two doses of 3000 μg vitamin D3 each in the second and third trimesters (group 2, n 49). Maternal 25-hydroxyvitamin D (25(OH)D) at term, cord blood (CB) alkaline phosphatase (ALP), neonatal serum Ca and anthropometry were measured in these subjects and in forty-three non-supplemented mother–infant pairs (usual care). Median maternal 25(OH)D at term was higher in group 2 (58·7, interquartile range (IQR) 38·4–89·4 nmol/l) v. group 1 (26·2, IQR 17·7–57·7 nmol/l) and usual-care group (39·2, IQR 21·2–73·4 nmol/l) (P = 0·000). CB ALP was increased (>8.02 μkat/l or >480 IU/l) in 66·7 % of the usual-care group v. 41·9 % of group 1 and 38·9 % of group 2 (P = 0·03). Neonatal Ca and CB 25(OH)D did not differ significantly in the three groups. Birth weight, length and head circumference were greater and the anterior fontanelle was smaller in groups 1 and 2 (3·08 and 3·03 kg, 50·3 and 50·1 cm, 34·5 and 34·4 cm, 2·6 and 2·5 cm, respectively) v. usual care (2·77 kg, 49·4, 33·6, 3·3 cm; P = 0·000 for length, head circumference and fontanelle and P = 0·003 for weight). These differences were still evident at 9 months. We conclude that both 1500 μg and two doses of 3000 μg vitamin D3 had a beneficial effect on infant anthropometry, the larger dose also improving CB ALP and maternal 25(OH)D.