

Relatively similar and relatively dissimilar aircraft were studied at canonical or noncanonical orientations. Half of the participants studied images of the original aircraft; the other half studied morphed images that enhanced the distinctiveness of each aircraft. At test, participants were required to identify the aircraft in a variety of novel orientations. All conditions were fully counterbalanced. The relatively similar aircraft were more difficult to learn than the relatively dissimilar ones, but both were easier to learn when aircraft distinctiveness was enhanced. During testing on the novel orientations, the similar aircraft were harder to identify than the dissimilar ones. The canonical presentation during learning affected performance during testing on novel orientation, but this effect disappeared when the images with enhanced distinctiveness were used during learning. The results are discussed in terms object identification, aircraft recognition training, categorization, learning, and mental representations.