Spatial reorientation in a rectangular array of four landmarks located in the center of a circular enclosure was investigated in domestic chicks (Gallus gallus). One of the landmark possessed unique visual features, indicating the location of a food reward. After training, chicks were tested (a) with the same array as during the training; (b) with four identical landmarks of the type previously nonrewarded, of the type previously rewarded, or of a new type; (c) after having transformed one of the landmarks located at the geometric incorrect location into the type of landmark previously rewarded; or (d) with a fifth landmark of the rewarded type at a new location. Chicks encoded information provided by local featural cues but not the geometric information provided by the shape of the array. Moreover, when trained in a rectangular array of identical landmarks chicks failed to reorient. In a second series of experiments, the array was located in correspondence to the corners of a rectangular enclosure. This time chicks successfully learned to locate the reward using geometric information. However, when the rectangular array was located in the center of a larger rectangular enclosure, chicks failed to reorient, indicating that the geometric information given by the macroscopic layout of arena surfaces was not used to specify different locations. These results suggest that chicks reorient on the basis of a local representation of single landmarks and that encoding of the global aspects of geometry only occurs with respect to the large, extended surfaces of an enclosure.
Read full abstract