A model of marginal vitamin A deficiency was developed in young rats by limiting the vitamin A intake of dams and of offspring postweaning. Prior to and throughout pregnancy and lactation, female rats were fed diets in which the vitamin A concentration was either low marginal [0.18 retinol equivalent (RE)/g diet], high marginal (0.4 RE/g) or sufficient (4 RE/g). Vitamin A restriction had no effect on dams' reproduction or litter sizes, but total retinol in liver was depleted by the end of lactation. Pups fed all diets grew steadily from birth through 35 d. The milk curd total retinol concentration of 9-d-old pups' stomachs was significantly different among the low marginal, high marginal and vitamin A-sufficient groups. By 35 d, plasma retinol concentrations of pups in the low marginal group were less than half of those in the high marginal and vitamin A-sufficient groups. The liver total retinol concentrations and lecithin retinol:acyltransferase activity of 35-d-old pups in the low and high marginal groups were much lower than those of vitamin A-sufficient pups. When pups from dams fed low marginal diet were weaned onto vitamin A-free diet, frank vitamin A deficiency was evident by 33 d of age as judged by physical signs and biochemical alterations in vitamin A status. Thus, these dietary protocols are useful in inducing early onset of either marginal or frank vitamin A deficiency.