Two experiments were conducted at separate commercial farms. In Exp. 1, all primiparous sows and an equal number of multiparous sows weaned each week were randomly assigned to one of four treatment groups. Treatments consisted of one i.m. injection on the day of weaning of 0, 50, 100, or 200 mg of beta-carotene. Sows were checked for estrus with boars once daily and mated at first estrus after weaning. A sample of sows (n = 100) was selected for determination of plasma beta-carotene and vitamin A. Blood samples were obtained from another group (n = 120) 14 d after injection to determine plasma progesterone. In Exp. 2, treatments consisted of i.m. injection of 200 mg of beta-carotene, 50,000 IU of vitamin A, or vehicle on the day of weaning, on the day of mating, and on d 7 after mating. In both experiments, the sow diet was supplemented with 11,000 IU/kg of vitamin A. In Exp. 1, there was no effect of dose of beta-carotene on the interval from weaning to estrus or on the repeat service rate. There was a dose x parity interaction on the number of pigs born dead (P < .01) and born alive (P < .10), because treatment with beta-carotene did not affect reproduction in primiparous sows, but litters subsequently farrowed by multiparous sows had more pigs born alive and fewer pigs born dead. There was a dose x day interaction (P < .05) on plasma beta-carotene; beta-carotene was elevated on d 6 and 13 after injection.(ABSTRACT TRUNCATED AT 250 WORDS)