184 UNIVERSITY OF COLORADO STUDIES 



dimensional effects, and at about the point of intersection the object 

 seems to be. Suppose A and B were made to recede laterally along the 

 line of incidence, the third dimensional image would nevertheless main- 

 tain its position at AB, but would become smaller and smaller. The 

 image would maintain its position because the convergence would remain 

 the same, but the decrease in apparent size would be due to the decrease 

 of size of retinal images. In Fig. 6, where there is another set of mirrors 

 by distancing A and B along the respective incidence lines the position 

 of the referred third dimensional image at AB again does not change, 

 but its size changes with the recession of A and B. Figs. 7 and 8 show 

 two arrangements with double sets of mirrors whereby actual objects 

 in space at A can be made to appear in the position of A'. In the first 

 case it seems more remote than it really is and in the second nearer. 

 Here again the distance is inferred from the amount of convergence, 

 and as in all the other cases, there is an illusion. In Figs. 5 and 6, if 

 A and B could be made to compensate, when receding, by increasing 

 in size so that the size of the retinal image remained unchanged, the eye 

 would not be conscious of any movement or change whatever. In 

 Figs. 7 and 8 this would not be so. If A, the real object, should be 

 made to approach or recede from the observer, and mirrors A" should 

 be kept turning on their pivots so that the angles on mirrors A'" would 

 be kept constant, no movement of the object could be detected, but it 

 would be observed as changing its form — longer or shorter or inverted. 

 The reasons for this can be seen by manipulating the figures, and 

 verifying by experiment. 



So far we have not made any experiments with the refractive stereo- 

 scope of Brewster which usually has two double convex (six-inch) 

 lenses. The lenses add nothing essential to the effects, only they serve 

 to concentrate the rays, and consequently there is an advantage, as 

 larger pictures can be used. Fig. 9 shows the function of convex 

 lenses. A and B are corresponding points on the two pictures of the 

 stereograph. Without their aid these points would probably be farther 

 apart than the interocular distance. The fines of reference of the two 

 eyes would never meet, and there could be no superposition of the two 

 images and therefore no third-dimensional effects. 



