STEREOPSIS IN MAN 317 



case are absolutely identical.* An approach to this situation is obtained 

 when we look at objects farther and farther away. We can judge their 

 distances binocularly, with convergence as the chief clue, up to about 

 one hundred feet; but even far short of that distance all solidity — 

 where it really depends upon disparate retinal images and not upon our 

 familiarity with the object — is lost. Such distant objects appear flat 

 simply because, with the lines of sight making so slight an angle with 

 each other, the two images we have of the object are not different enough 

 to yield any rotundity when fused. 



Stereopsis, then, results from the fusion or unification of two views 

 which differ slightly in a particular way and within certain limits. The 

 object must be seen singly, and this is where the matter of the location 

 of the images on the two retinae comes in : 



The two images of any object-point must fall upon 'corresponding 

 points' of the two retinae if they are to be fused. The two foveae are cor- 

 responding points, and if identical small images are falling upon them, 

 no matter whether they emanate from a single object or not and no 

 matter in what direction each eye is pointing, those images will be fused. 



When a point at any given distance is fixated by both eyes and is 

 seen singly, there is at the same instant an infinity of other points in 

 space which are seen singly along with it. Their images are falling upon 

 corresponding retinal points other than the foveae; and the external, 

 spatial points themselves determine a complexly shaped hypothetical 

 surface hanging out in space, called the horopter. There is a different 

 horoptral surface for every point of binocular fixation, at every distance 

 and direction. Obviously, the whole matter of horopters can become 

 hideously complicated, and it is as well for the reader (and the writer!) 

 that we shall not need to worry much more about them. Suffice it to say 

 that when you fixate a point across the room, and raise a finger into the 

 line of fire, you see the finger doubled, because the right- and left-eyed 

 images of it are not falling upon points in one retina which 'correspond' 

 with the stimulated points in the other. They could be made to do so if 

 the finger were amputated and carried out and glued onto the horopter- 

 of-the-moment. But by merely looking at the finger the two images are 

 made to slide together into one, for the change of convergence and 

 accommodation has created a new horopter on which the finger now lies. 



*Unless the object is quite small and quite near — say, a calling-card at the near point — 

 when the images are of course appreciably 'keystoned': for each eye, the card tapers in 

 the direction of the other eye. 



