VISUAL JUDGMENTS 



661 



ILLUSIONS OF SIZE 



The distance between two points appears longer if a number of 

 points be interposed between the two. Thus if two equal quadri- 

 lateral figures be divided one by horizontal and the other by vertical 

 l mes the one divided by horizontal lines will appear elongated ver- 

 tically, and that divided 

 by vertical lines elongated 

 horizontally (Fig. 301). 

 Apparently it requires a 

 somewhat less effort to 

 pass directly from one 

 point to the other than Fia. 301. 



when the gaze has to 



follow an interrupted line. The eye muscles probably make a separate 

 effort of movement at each interruption of the line. To these move- 

 ments is due the illusion in Fig. 302, where parallel lines seem to diverge 

 or converge. Of two equal lines, one of which is vertical and one hori- 

 zontal, the vertical line seems the longer. When the eye is moved 

 upwards, the tendency of the superior rectus to move the eye inwards 

 has to be counteracted by a pull of the inferior oblique muscle turning 



the eye outwards. A greater effort 

 is therefore required than when 

 the eye is moved either inwards or 

 ^^^^^^B^^^^^^I^I^^I^ outwards, and it has been sug- 

 gested that it is this greater 

 muscular effort which is respon- 

 sible for our over- valuation of the 

 length of vertical as compared 

 with that of horizontal lines. 



It must not, however, be 

 ^j^j^X^X^X^^^^^^^i^ imagined that an actual move- 



V V ment of the eyes is necessary in 



order to judge of the size and 

 302 distance of any object. If a white 



thread be hung up in a dark 

 room and be illuminated for an instant of time by an electric 

 spark it is impossible for an observer in the dark room to move 

 his eyes so that the image of the thread shall fall on the corre- 

 sponding points of the two retinse before the illumination has 

 disappeared. In spite of the fact, however, that the image of 

 the thread falls on non-corresponding points, as will be seen from 

 the diagram (Fig. 303), the thread is seen, not as double, but as 

 single, and a very correct impression may be obtained of its size and 



i\XXXXXXX\XXXX\XXX 



