During sensory preconditioning, rats were given two distinct flavors (saccharin and coffee) in sequence with 0, 9, or 27 sec between the two flavors. A control group received the flavors unpaired. In subsequent training, the second flavor preceded sucrose by 5 min. Later, the subjects that had had 0- or 9-sec delays between the two flavors showed a significant preference for the flavor not directly reinforced, whereas those that had had a 27-sec delay showed only a marginally significant preference. Although taste aversions have been produced using these methods, this is the first demonstration of conditioned preferences for flavors not directly associated with the reinforcers. These results offer an alternative way to study flavor-flavor learning.