Four pigeons previously trained to home to the roof of the University of British Columbia psychology building and 4 nonhoming pigeons were trained to discriminate between two sets of color slides projected onto a pecking panel of a Skinner box. One slide set consisted of photographs taken in the vicinity of the psychology building; the other set consisted of similar views taken at locations not previously visited by the homing subjects. All subjects were rewarded for pecking during slides from the first but not the second set. Every few sessions, new "Home" and "Away" slides were introduced during transfer tests. In a final transfer test, a completely new tray of Home and Away slides was introduced. The homing pigeons were slightly (but not statistically significantly) better at discriminating Home from Away slides. The implications of these results for understanding pigeons' homing behavior, concept attainment, and spatial memory are discussed.Pigeons appear to be able to form conceptual discriminations based on photographs of such natural objects as people, fish, water, trees, and leaves (see Herrnstein, 1984, for a review). In these studies, pigeons were rewarded for pecking at projected images that contained the target object (e.g., a tree), and were not rewarded for pecking at images in which the target was absent. Typically, pigeons learned these discriminations rapidly, showed good transfer to novel instances of the target category, and remembered these discriminations over periods of many months (e.g., Vaughan & Greene, 1984). Pigeons seemed to treat these two-dimensional projected images as if they represented "real" three-dimensional objects. Both Cabe (1976) and Lumsden (1977) have shown that pigeons transfer discriminations of actual objects to photographs of the same objects. In addition, Looney and Cohen (1974) have shown that pigeons will attack photographs of conspecifics in situations in which aggressive responding is normally elicited. Perhaps the most convincing evidence that birds treat photographic projections as if these images were actual objects comes from a study of Roberts and Weigl (1984). In that study, investigators tested wild-eaught dark-eyedjuncos in a test chamber that contained two perches mounted on end walls containing projection screens. Images of this species' distinctive summer and winter habitats were projected onto the two screens, and the subjects' preference for an image was measured by recording how much time they spent This research was supported by the Natural Sciences and Engineering Research Council of Canada. John Zahradnik kindly donated racing homing pigeons whose offspring served as subjects in this research. C. Von Borman greatly assisted with the data analysis. Send correspondence regarding this paper to Donald M. Wilkie, Department of Psychology, University of British Columbia, Vancouver, British Columbia V6T IV7, Canada, or the following E-Mail address: userdonw@mtsg.ubc.ca. on the perch in front of the respective image. Roberts and Weigl found that winter-...