In three experiments, we examined the ontogeny of latent inhibition in a conditioned taste-aversion paradigm. In the first experiment, 18-, 25-, and 32-day-old rats received a pairing of a 0.6% Sanka solution with an i.p. injection of LiCl (0.75% b.w., 0.4 M) or physiological saline either after four preexposures to the taste or control treatment without taste preexposure. Impairment of conditioning by taste preexposure (latent inhibition) was evident at 32 days of age, but not at 18 or 25 days of age. In Experiment 2, we investigated the possibility that our failure to observe latent inhibition in 18-day-olds resulted from the particular experimental parameters employed in Experiment 1. In Experiment 2, taste-exposure durations were longer, the retention interval between conditioning and testing was shorter, and deprivation conditions were more typical of those used with preweanling subjects. Again, no evidence of latent inhibition was found. In Experiment 3, we examined the effect of sham or fornix lesions at 18 days of age on latent inhibition at 32 days of age. Lesioned subjects exhibited equivalent conditioned taste aversions but failed to show latent inhibition. Thus, latent inhibition of taste-aversion learning develops between 18 and 32 days of age in the rat, and this development is disrupted by fornix transection during infancy. These findings suggest a role for septohippocampal maturation in the ontogeny of latent inhibition.