Perceived line orientation changes during prolonged inspection time. Gibson and Radner (1937) attribute it to the normalization effect (perceptually adapting line shifted towards horizontal or vertical one). According to Gibson orientation of presented line is evaluated with respect to horizontal and vertical ones, which are called "norms". Later investigations had shown that this normalization couldn't explain other phenomena related with adaptation (Mitchell, Muir, 1976; Fomin et al., 1979). The hypothesis had been presented that also should be 45° orientation norm (Fomin et al., 1979). However, this orientation has not been investigated. Our task was to investigate the influence of line orientation on perceived tilt in the course of adaptation. Four subjects were presented with adapting stimulus consisting of 3 parallel lines 2 deg of arc in length and distance between lines was 20 min of arc. After initial adaptation (30 s) the test line oriented at F + + n×0.5° to the adapting line appeared at 8 deg of arc distance from the center of the adapting stimulus. The integer n was chosen randomly from interval [-10 ... 10]. Subjects were required to make a 2AFC judgment whether test line appeared to be tilted clockwise or anticlockwise relative to the adapting lines. The inclinations of adapting lines were from: -10°, 0°, 10°, 35°, 45°, 55°, 80°, 90° and 100° to vertical. The results for all subjects show that three orientations: 0°, 45°, and 90° appear to be "stable norms". The lines with orientation deviations in 10° from "norms" were perceptually tilted towards the closest of "norms" in the course of adaptation. The experimental results are compared with those predicted by the proposed model.