The effects of maternal postpartum vitamin A or beta-carotene supplementation on maternal and infant serum retinol concentrations, modified relative dose-response (MRDR) ratios and breast milk vitamin A concentrations were assessed during a community-based trial in Matlab, Bangladesh. At 1-3 wk postpartum, women were randomly assigned to receive either (1) a single dose of 200,000 international units [60,000 retinol equivalents (RE)] vitamin A followed by daily placebos (n = 74), (2) daily doses of beta-carotene [7.8 mg (1300 RE)] (n = 73) or (3) daily placebos (n = 73) until 9 mo postpartum. Compared to placebos, vitamin A supplementation resulted in lower maternal MRDR ratios (i.e., increased liver stores) and higher milk vitamin A concentrations at 3 mo, but these improvements were not sustained. The beta-carotene supplementation acted more slowly, resulting in milk vitamin A concentrations higher than the placebo group only at 9 mo. Irrespective of treatment group, over 50% of women produced milk with low vitamin A concentrations (</=1.05 micromol/L or </=0.28 micromol/g fat) throughout the study. Overall, mean maternal serum retinol concentrations were not affected by supplementation. Compared to the placebo group, the mean MRDR ratio of 6-mo-old infants was higher in the vitamin A group. Infants (33%) had serum retinol concentrations <0.70 micromol/L and 88% had MRDR ratios >/=0. 06. We conclude that while both interventions were beneficial, neither was sufficient to correct the underlying subclinical vitamin A deficiency in these women nor to bring their infants into adequate vitamin A status.