THE SENSE OF VISION. 351 



D seems higher than it is broad, and the square E broader than it is high, the 

 illusion being more marked in the case of D than in the case of E, because, as 

 above explained, vertical distances are, as a rule, over-estimated. 



The explanation of this illusion seems to be that the eye in passing over a 

 subdivided line or area recognizes the number and size of the subdivisions, 



Fig. 163.— Zollner's lines. 



and thus gets an impression of greater total size than when no subdivisions 

 are present. 



A good example of this phenomenon is afforded by the apparently increased 

 extent of a meadow when the grass growing on it is cut and arranged in hay- 

 cocks. 1 



The relations of lines to each other gives rise to numerous illusions of 

 spatial perception, among the most striking of which are those afforded by the 

 so-called " Zollner's lines," an example of which is given in Figure 163. Here 

 the horizontal lines, though strictly parallel to each ^ j * 



other, seem to diverge and converge alternately, their 

 apparent direction being changed toward greater per- 

 pendicularity to the short oblique lines crossing them. 

 This illusion is to be explained in part by the tendency 

 of the eye to over-estimate the size of acute and to 

 under-estimate that of obtuse angles — a tendency which, 

 according to Filehne, 2 is due to the fact that we are 

 constantly surrounded by square-cornered objects 

 (houses, furniture, etc.), the right angles of which, 

 being seen obliquely, arc projected onto our retinas 

 as acute or obtuse angles. Knowing these angles ton. 



. . . i i • ,i Fig. 164.— To illustrate illusion 



be right angles, we are constantly applying mental of Bpace-perception. 



corrections to our visual data, and the habit thus 



acquired forces us to regard all acute and obtuse angles as nearer to right 

 angles than they really are. The illusion in Zollner's lines is more marked 

 when the figure is so held that the long parallel lines make an angle of about 



1 It is interesting to note that a similar illusion has been observed when an interval of time 

 subdivided by audible signals is compared with an equal interval not so subdivided (Hall and 

 Jastrow : Mind, xi. 62). 



2 Zeitschrift fur Psychologie und Physiologic drr Sinncsorgane, xvii. S. 16. 



