JULIUS RHEINBERG ON STEREOSCOPIC EFFECT AND A 



vertical side on the right hand (Fig. 6). We may suppose this 

 side to be coloured red, and we will further suppose the top of 

 the cube to be in true focus. Then it is evident that no light 

 rays from the red side of the cube can reach the left half of 



the objective all the Light from the red side of the cube im- 



pj, ,. fche right half of the objective only, and it would not 



be rether out of place, therefore, in such a case to say that 



the righi half of the objective "sees" what the left half does 

 ,,,,! "see." But how is the image formed in such a case? 

 Clearly all points of the red side of the cube, except those in the 

 surface plane F, which is supposed to be in true focus on the 

 image plane I, are represented by semicircular diffusion discs, 

 which increase in size as the distance increases from the plane F. 

 Tin- shaded part of Fig. 6 shows the course of the rays from a 



int G on the lowest plane of the cube; the diffusion disc G" 

 lies on the left-hand side of F', the image of F. It will be 

 evident that the diffusion discs from points intermediate between 

 I' and G will all overlap, and the resultant appearance in the 

 image plane, therefore, is that of a broadened red patch, brightest 

 at one edge and fading away towards the other. 



Suppose next the cube to be slightly shifted towards one side. 

 The same reasoning as before holds good, only the sector of the 

 objective which " sees " the red side will either be less or greater 

 than half the objective, and so the diffusion discswill be corre- 

 spondingly greater or less. But as pointed out in the earlier part 

 of this paper, the central point of a diffusion disc stands for the 

 position of the disc as a whole, and the particular thing to which 

 I wish to draw your attention is that here we have cases of 

 lateral shifting of image points, notwithstanding the fact that the 

 u hole objective is given free play. In other words, when certain 

 parts of an object block out the passage of the rays from other 

 parts we are liable to get parallactic displacements, which, so far 

 as tin- resultant image is concerned, simulate to a certain extent 

 the appearance which would be presented if some isolated part 

 objective were "looking at" the object, round the corner 

 it were. When we use the two halves of an objective 

 separately, as in binocular microscopes, of course, the same thing 

 holds good; and theoretically, at least, it is possible that certain 

 points of an object represented in the one picture are not 

 represented at all in the other. 



