ILLUSION. 473 ILLUSION. then a group of horizontal lines, filling the same area as the square; and then a similar group of vertical lines. The horizontal group looks I 1 — I 1 E II II oAoO Fig. 9. absociative illusions (from Sanford, Course in Experimental Psychology). For A and B see text. C and D show the same an^Ie differently estimated. E and F are the same kind of illusion as A and B. In G the circle surrounded by small circles is the same in size as that surrounded by large circles, though it appears larger. higher, the vertical group broader, than the ob- jectively equal square, for the reason, again, that the grouped lines offer more obstacles to the passage of the eye over the area of the figure. One of the most disputed figures of this class is the illusion of Miiller-Lyer, or the arrow-head and feather illusion. Draw two vertical or hori- zontal lines of equal length. Tip the two ends D TfT^' with oblique lines of the same length directed outward (arrow feathers) ; the latter appears considerably the longer of the two. Wundt's explanation is that the inward-going obliques offer a check to the passage of the eye along the principal line, while the outward-going oblique* favor a continuance of eye-movement. The con- stant illusions of extent (due to asymmetries of muscular action about the eyeball) are summed up in a right-angled and equal-armed cross. The vertical line in this figure seems to be longer than the horizontal; the upper vertical arm longer than the lower; and, in monocular vision, the outer horizontal arm looks longer than the inner, (c) Thirdly, we have variable and con- stant illusions of direction. Draw a large out- line square, marking in one of the diagonals, and three or four lines parallel to the diagonal, at equal distances on either side of it. Now pro- ceed to cross-hatch the oblique lines, cutting the diagonal with short vertical pieces, the two ad- joining obliques with short horizontal pieces, the two oblique lines next beyond these with verticais again, the two next with horizontals, and so forth. Notice that the parallel obliques no longer look parallel, but appear to converge and diverge alternately. The explanation of this and similar illusions is that small angles are over- estimated, and large angles relatively underesti- mated; and the reason for such overestimation, again, is that it takes more muscular effort and energy to start a movement of the eye than it does to continue a movement already begun. The constant illusion of direction is that, under cer- tain conditions of fixation, hyperbolas are per- ceived as straight lines, owing to the concavity of the retina. (d) There remains a group of illusions which Wundt clas.ses as 'associative,' i.e. as psychological, not physiological, in origin. Flo. 10. THE STROBOSCOPE OR ZofiTROPE. Looking at the figures through the slits in the rim. you get. when the instrument is rotated, an effect similar to and depending ou the same causes as that of the kineto- acope. of the one with oblique lines directed inward (arrow-heads), and the two ends of the other Fig. 11. AN ILLUSION OF MOVEMENT. Move the figure around in a circle in its own plane, and observe the apparent motion of the six out*'rdisk8 and the central co^r-wheel. (From Sanford, Course in Experimental Psycbology.) If equal short horizontal lines are drawn, some giouped closely together, and some widely spaced, the latter appear longer than the former; the length of the line is 'assimilated' to the size of the interspaces. In the same way, if two broad
