198 Mr. S. Rowley on a New Theory of Vision. 



C B, fig. 2. Then by my proposition the left eye's visible image 

 of the luminous point at A should be seen at ul in a surface cut- 

 ting the axis F 13 at right angles in B (and the plane of the axes 

 in the line E G), and the right eye's image at a. 



But if the left eye's image of the point at A is seen at a 1 , then 

 by the same proposition the right eye's image of the point at E' 

 opposite A in a line between F' and e, a position in I K directly 

 beyond a' should be seen in the same line of no variation a little 

 beyond a' at e. If the experiment be made, it will be found that 

 the right eye's visible image of an object placed as a point at E', 

 nearly lj- - inch from A, will be seen a little behind the place 

 where the left eye's image of an object placed as a point at A is 

 seen. 



So the two middle images of the pins placed as a test of dis- 

 tance in the first of these experiments are seen united at the in- 

 tersection of the surfaces of vision. 



If the surfaces of vision were planes, since their line of inter- 

 section would then coincide with the line of intersection of the 

 vertical planes of the axes, it is evident that the images of these 

 pins longitudinally bisected by the latter planes would exactly 

 coincide. But, owing to the slight curvature of the surfaces of 

 vision, the line of intersection of these surfaces immediately 

 begins very gradually to recede from that of the vertical planes 

 of the axes, and the longitudinal axes of the images of the pins 

 to cross over and fall on opposite sides, thereby preventing a 

 mathematical coincidence of the images. 



This effect of the departure of the intersection of the surfaces 

 of vision (supposing the optic axes parallel to the horizon, to 

 which the pins are perpendicular, to be directed to the bottoms of 

 the images) may be counteracted and a mathematical coincidence 

 of the images nearly produced by inclining the pin whose image 

 is seen by the left eye a very little to the left, and the other duly 

 to the right. Should this small inclination be given with such a 

 curve that a line drawn from any point of the axis of the pin to 

 the focal centre of the eye would intersect the line of intersection 

 of the surfaces of vision, then an exact mathematical coincidence 

 of the images would be secured. 



The approximate result may be obtained if two rulers*, or the 

 legs of a pair of compassesf, be so placed that their longitudinal 

 axes shall lie in the vertical planes of the optical axes. 



It is evident that for the two objects in the vertical planes of 

 the axes, we may put a single object at the intersection of these 

 planes J. 



But lo make the accurate performance of these experiments 



* See Wells, p. 36. f Smith's ' Onties/ art. 977> 



X Wells, p. 49. 



