Four experiments are reported which demonstrate the ability of an unconditioned stimulus (UCS) presentation following extinction to partially reinstate the conditioned response. These experiments are interpreted in terms of the strengthening of an extinction-reduced UCS representation. The first two experiments address alternative interpretations in terms of sensitization, reinstating the stimulus conditions of acquisition, conditioning of background cues, and stimulus generalization. Experiment 3 suggests that reinstatement is possible with a UCS qualitatively different from that used in conditioning. Experiment 4 explores an alternative extinction procedure which especially preserves the conditioned stimulus-unconditioned stimulus association while encouraging modification of the UCS representation. The results are discussed both in terms of related empirical phenomena, such as spontaneous recovery and sensory preconditioning, and in relation to the general role of the UCS representation in conditioning.

Pigeons (Columba livia) searched for hidden food in a rectangular environment constructed to eliminate external orientation cues. A feature group was initially trained with distinct features in each corner. A geometric group was initially trained with no featural information. Tests revealed that both groups encoded the geometry of the apparatus. The geometric group was then retrained with features, and a series of tests was administered to both groups. Transformation tests revealed that the groups differed in reliance on features versus geometry. Pigeons in the feature group followed the positive feature even when it was placed in a geometrically incorrect corner, whereas pigeons in the geometric group showed shared control by features and geometry. Both groups were able to use features in corners distant to the goal to orient themselves, and both groups relied more on the color than on the shape of the features. Survival within an environment frequently requires efficient processing of spatial information. Spatial abilities underlie activities that are critical for the individual (e.g., establishment of lodging, avoidance of predation, and attainment of nourishment) and for a species (e.g., migratory behavior or reproduction); these activities may involve a variety of mechanisms. Navigation, for example, may be achieved through inertial guidance, orientation to a beacon, piloting by use of landmarks, or developing a spatial representation of the environment (Gallistel, 1990). Questions concerning which aspects of an environment are encoded and used in navigation have been addressed in recent research (for reviews, see Cheng & Spetch, 1998; Gallistel, 1990; and Poucet, 1993). Many studies have shown that animals can encode and use multiple sources of information to locate a goal (e.g., Spetch & Edwards, 1988) and that the primacy of control by different sources of information may differ according to context (e.g., Strasser & Bingman, 1996) or species (e.g., Brodbeck, 1994). One particularly interesting set of results has emerged from studies that have controlled and manipulated the information available for encoding by restricting access to navigational cues in an enclosed environment and by disrupting other positional cues through disorientation techniques (Cheng,

Four experiments examined the possibility that the outcome of simultaneous and backward fear conditioning procedures might depend upon the number of CS-UCS pairings. A punishment procedure will rats as subjects and shock as the UCS was used; the amount of suppression produced by response-contingent CS presentations indexed the strength of acquired fear. Experiments 1, 3, and 4 examined the suppressive tendencies of simultaneous-and backward-trained CSs after 0, 10, 20, 40, 80, and 160 pairings. The pattern of data suggested that initial pairings have the effect of increasing suppressive tendencies of the CS, while subsequent pairings decrease them. In addition, evidence of fear inhibition was found after 160 pairings in the case of the backward paradigm. Experiment 2 examined several nonassociative accounts based upon differential shock exposure. Groups given 10 pairings or 80 pairings were compared to ggroups given 10 pairings plus 70 shock-alone presentations. The results indicated that number of pairings, rather than number of UCS occurrences, is the important factor in decreasing the initial suppression. The evidence for eventual inhibition in the backward paradigm suggests that this occurs through the acquisition of inhibitory tendencies which are antagonistic to the previously conditioned excitation.

The paths taken by 6-and 12-year-old children who were asked to lead the way back after their first walk acrossa university campus were recorded. Prior to the return, children were either uninformed of the requirement to lead the way back, generally informed, or shown specific near or far landmarks during the walk. In comparison with uninformed children, neither age group prepared adequately in response to the general admonition, "\bu will have to lead us back here, so pay good attention." Analyses of distance traveled on route and choices at important intersections indicated that both age groups benefited when two landmarks on the path (a telephone booth and a mailbox) had been pointed out. Analyses of distance traveled off route indicated that 12-year-old, but not 6-year-old children, wandered less when distant orientation cues (a smokestack and a building on the skyline) had been pointed out. The results suggest a prudent prescription for parents who are concerned about their child's independent travel: Specify route features where the child should continue or change heading.Several strategies may develop to allow successful navigation in unfamiliar territory. For example ', Allen, Kirasic, Siegel, and Herman (1979) found an" age-related trend in the ability to select scenes that were high in potential landmark value along a walk. The walk was simulated by a sequence of slides photographed along a commercial street. Seven-year-old children tended to pick colorful awnings and window displays as reference points, despite the fact that there were many similar awnings and displays along the street. They tended to overlook pictures of intersections that provided distinctive distal and proximal cues for location. Ten-year-old children were more likely than 7-year-old children to select scenes with landmarks close to changes in heading. They recognized that certain cues mark where changes in direction of travel are necessary.In addition to judicious selection of landmarks, older children may have a variety of sophisticated spatial reasoning abilities (Cohen, 1985;Liben, Patterson, & Newcombe, 1981;Stiles-Davis, Kritchevsky, & Bellugi, 1988). By 10 years of age, they may be able to infer shortcuts, estimate distance traveled, and use procedures to discover and correct navigation errors. These cognitive abilities would facilitate wayfinding in unfamiliar environments only if the child has first attended to landmarks and remembered relations between landmarks that indicate the distance and direction of travel. Accordingly, the purpose of the This research was supported by grants to Edward H. Cornell from the Natural Sciences and Engineering Research Council of Canada. We gratefully acknowledge the cooperation of the principals, parents, and students of Crawford Plains, Ekota. Greenview, Lee Ridge, and Sakaw Elementary Schools. We thank Linda Walley, Lee Paskar, and Wanda Rowat for their help with the data collection and transcription.

Children and adults were escorted on their first walk across our university campus and were periodically led off the original route during the return trip. During the return, we stopped prior to intersections on and off the original route to obtain estimates of place recognition accuracy and confidence. The subjects were then asked to point to the path that led back to the start and were corrected if wrong. Accuracy of place recognition was intermediate in a way-finding task requiring reversal of an incidentally learned novel route. However, accuracy increased as subjects were farther from the original route, indicating that the presence of novel landmarks boosted the discrimination of old and new places. Eight-year-old children were less accurate than 12-year-old children and 25-year-old adults, who did not differ in accuracy. There was a similar age difference in the ability to point to the direction to return when subjects correctly recognized that they were off route. The results are used to develop a model of way finding by place recognition.