are simply the projections of the P. through the ocular center O (Fig. 1). Al 
| | The objective angles A. =} Po à Pr determine the retinal relation of the fi 
| image points. In fact, ay = Q, 0 Qy The relative positions of the image re 
| points on the retina is constant through all rotations about 0. In other words, - 
as the eye rotates and the retina scans the optical image, the correspondence m 
of points of the image to points on the retina changes, but the relative positions 
| | of the image points on the retina do not. This constant relation of the points in al 
| the retinal image is the most fundamental property of the information supplied SE 
to the brain. If an entirely different physical configuration is provided which es 
stimulates the retina in the same way, then the perception of the new stimulus oc 
is identical with that of the old one. So, for example, the points P. may be st 
displaced arbitrarily along the lines 0 B,, or the entire fan of lines 0 P, in 
rotated rigidly about 0, without distorting the visual impression. The point of re 
view we adopt here is that the essential geometric information relayed monocu- 
larly to the higher centers consists only of the relative positions of points in the L« 
retinal image and nothing else. These relative positions, then, constitute the la 
I effective monocular stimulus. ax 
The corresponding binocular principle is simply the monocular principle CC 
compounded: the essential geometric information relayed binocularly to the m 
| higher centers consists of the separate monocular components. Hence we may di 
| reproduce the binocular visual effect of a stimulus merely by reproducing the ca 
ll separate monocular components. This is the principle of the stereoscope. It m 
| follows from the stated monocular principle that there are infinitely many stimuli Ti 
| which are capable of producing the same binocular impression. This conclusion Or 
| provides an explanation of Helmholtz's observations on visual straightness [6, Vol.3, tr 
p. 318]. and of the related binocularly equivalent rooms of Ames [7]. Helmholtz noted (1 
that if two points are placed symmetrically to the median on a transverse line, then a = 
third point placed on the median and aligned visually between the two does not neces- 
sarily fall on the objective straight line joining them. In fact, there is just one dis- a 
tance sagittally for which the visually aligned point is physically aligned. If the 9 
experiment is performed closer to the observer then the visually aligned point is farther 
from the observer than the objective straight line; a reverse effect is obtained if the 3. 
experiment is performed at a greater distance from the observer (Fig. 2). 
To explain Helmholtz's result we consider three points A B C in the one place 
where they are aligned both visually and physically. If theocular images are rotated to 
about R and L by equal amounts nasally, there is determined by projection a stimulis A 
A', B', C' binocularly equivalent to A,B,C, closer to the observer and convex away | i 
from him (Fig.2). Similarly, if the rotations are made temporally we find a configurati E: 
= 4 —