When a caloric substance is followed by a flavored substance, preference for the flavor that followed the calories can increase because of a flavor-nutrient association. We showed here that this effect is opposed by a fullness effect: consuming the caloric substance itself reduces consumption of the flavor that follows. Because consumption of this flavor was less than consumption of a flavor that was given alone, there was a reduced preference for the flavor that followed the calories-an effect opposing flavor-nutrient learning. The preference for the flavor that was given alone and consumed in greater amounts seems to be due to exposure per se and is not a result of contrast. When the amount of the flavor given alone was directly controlled so that it was one half the amount of the flavor that followed the calories, preference for the flavor following the calories was no higher than when consumption of the two flavors was equal. Thus, exposure can interfere with flavor-nutrient learning, but does not enhance it.Capaldi and Sheffer (1992) showed that a flavor of saccharin given after chocolate milk was later preferred to a flavor given alone. They suggested that the flavor given after the chocolate milk was associated with the postingestive effects (nutrients) of the chocolate milk, and was therefore later preferred on the basis of flavor-nutrient learning. Boakes and Lubart (1988) suggestedthis hypothesis to account for their finding that a flavor of saccharin given shortly after glucose was preferred to a flavor given alone. When flavored saccharin followsglucose, the flavor of saccharin is more closely associated in time with the postingestive effects of glucose than is the flavor of glucose itself. Consistent with this hypothesis, Boakes and Lubart (1988) showed that a flavor of saccharin given 60 min after glucose, long enough for the postingestive effects of glucose to have dissipated, was not preferred to a flavor given alone.In Experiment I, we were concerned with whether exposure to flavored saccharin or the nutrient prior to flavor-nutrient learning would interfere with that learning. Holder (1991) showed that a cue (taste or odor) that had been paired with sucrose was preferred to one that had been paired with saccharin. When a second cue was added that was associated with sucrose, the second cue was ignored in favor of the cue that was learned first-a blocking effect (Kamin, 1969). We hypothesized that as a result of preexposure to chocolate milk, rats would form an association between the taste of chocolate milk and the resulting calories, and that when saccharin followed chocolate milk in training, the saccharin cue would be ig-