A randomized trial tested whether a priming dose of vitamin A would extend the protection of a subsequent 60,000-micrograms retinol equivalent (RE) oral dose. Seventy-five xerophthalmic and 74 age- and neighborhood-matched non-xerophthalmic preschool children were randomized to one of three oral regimens of vitamin A, receiving peanut oil only (Group A), 7500 micrograms RE (Group B) or 60,000 micrograms RE (Group C), followed in all instances by 60,000 micrograms RE 1 wk later. Serum retinol was measured 2, 4, 6 and 12 mo following the second dose by technicians unaware of the children's treatment status. Among xerophthalmic children, mean values differed across treatment groups at 2 mo (C > A) and tended to be different at 12 mo (C > A and B > A). Among non-xerophthalmic children mean retinol concentrations differed across treatment groups at 6 mo, but not in a consistent way (A > C > B), and at 12 mo (C > A and B > A). Xerophthalmic children reverted to biochemical deficiency faster than non-xerophthalmic children. A small or large priming dose may extend the protection conferred by a 60,000-micrograms RE dose, supporting the use of repeated, spaced doses of vitamin A for treating xerophthalmia. Similar retinol concentrations in Groups B and C at 12 mo suggest the 60,000-micrograms RE prophylactic dose currently recommended by the World Health Organization need not be increased.