i6 5 6 



HANDBOOK OF l'HYSIOI ( >GY 



M I \<i H'llVli 'I ' i<;1 111 



fig. 35. Orbison's patterns. 

 The regularly patterned back- 

 grounds induce systematic and 

 predictable distortions in the 

 line patterns (perfect circle in 

 I, perfect square in B) that have 

 bem superimposed on these 

 special backgrounds. [After Orbi- 

 son (370).] 



illusion, viz. of Fonzo's design (shown in fig. 36) 

 where two or more lines are seen to converge; figures 

 inscribed in those regions where the converging lines 

 approach each other are overestimated. The illusion 

 here would be due to a normal tendency to see more 

 distant tracks as larger than their geometrical projec- 

 tion. In fact, Tausch (462) has attempted to find, for 

 every illusion, a corresponding prototypical constancy 

 in real life (fig. 37). 



This approach is particularly attractive because it 

 would account, if true, for the curious anisotropics of 

 apparent extents in the visual field. In a normal 

 (monocular) field of vision, targets of equal visual 

 angle appear relatively smaller to indirect (periph- 

 eral) gaze, as compared with direct fixation; the) 

 also seem smaller in temporal and lower portions of 

 the field. These asymmetries can be demonstrated 

 quantitatively on monocular bisection of lines; the 

 nasal (and upper) portions are usually made shorter 

 than the temporal (or lower) portions, since the 

 latter 'look smaller.'* 1 Such results imply a relative 

 micropsia of temporal and of lower halves of the 

 visual field, and relative maeropsia of nasal and upper 

 hakes. According to the views just developed, these 

 physiologic micropsiae and macropsiae are under- 

 standable Normally, objects in the nasal and upper 

 parts subtend smaller visual angles, since the) are 

 near the vanishing point, and constancy of size would 

 therefore tend to enlarge these objects, As the moon 

 illusion shows, however, this effect is not restricted to 

 retinal areas, but is also associated with particular 



"Brown 69) has shown thai such ninniniii.il asymmetries 

 omewhat unstable; they change progressively on pro- 

 1 testing, especially it ni.ils extend over several days, 



positions of the eyes; the moon is maximally enlarged 

 as we look toward the horizon and it "shrinks' on ele- 

 vation (as well as deflection) of binocular regard. 



PERCEPTUAL HABITUATION; DECREMENT OF THE 

 MULLER-LYER ILLUSION ON REPEATED TRIALS. The 



Miiller-Lyer effect and those induced by similar 

 patterns ( Poggendorff, Zollner) diminish on repeated 

 exposure (244). Particularly when trials are massed 

 [rather than distributed over longer periods (353)], 

 an observer will set the two segments of the pattern 

 with decreasing 'errors,' sometimes to the point 

 where the illusion disappears and gives way to a 

 slight effect in the opposite direction. This decrement 

 of the illusion on repeated trials occurs in the absence 

 of any knowledge of results, and thus differs from 

 ordinary forms of learning (276). In this respect, the 

 decrement is quite analogous to the progressive 

 diminution in the errors of localization induced bv 

 prisms (194) or pseudophones (192). The decrement 

 of the Miiller-Lyer effect disappears when trials are 

 resumed alter the pattern has been turned around 

 end to end. Kohler & Fishback (276) propose there- 

 lore that the decrement should be subsumed under 

 the paradigm of those figural aftereffects which are 

 specific for certain positions in the visual field. In all 

 these respects, the decrement resembles the phenom- 

 enon ol specific habituation to originally arousing 



stimuli described b) Sharpless \ Jasper !-4-.'| see 

 also Sokolo\ (432)]. 



[NTERMODA1 1K\NSFER OF MULLER-LYER DECREMENT; 

 'HAPTIC' ILLUSIONS, It cm also be shown that the 

 decrement takes place lor a tactile analogue of the 

 Miiller-Lyer pattern (Rudel, unpublished obscrva- 



