J. LeConte — Phenomena of Binocular Vision. 101 



Standing before this and combining by extreme convergence, 

 on looking obliquely upward the phantom plane slopes away 

 from the observer ; on looking obliquely downward it slopes 

 away downward. So that sweeping the point of sight upward 

 and downward alternately, the phantom plane dips away fore 

 and aft from a transverse anticline forming a kind of arch. The 

 explanation of this is of course the same as that already given 

 for the floor. 



It is important to state that this slope was seen only in looking 

 obliquely upward or downward, and not at all in looking stead- 

 ily at right angles to the plane. I said at right angles. In real- 

 ity the neutral point is not exactly on a level with the eyes, 

 but about 7° above. This is the result of the rotation of the 

 eyes on the optic axes in convergence, as shown in one of my 

 previous papers, and is a beautiful demonstration of such rota- 

 tion. 



Experiment 3. — We have thus far experimented only by look- 

 ing obliquely upward or downward on the plane. If now we 

 look directly at right angles (or in reality a trifle above the 

 perpendicular) on the plane, whether floor, or vertical wall, or 

 wire screen, then, on extreme convergence, the phantom plane 

 is seen to slope away on each side, so as to form a transverse 

 arch. On sweeping the eyes upward and downward, the fore- 

 and-aft sloping combined with the side ways sloping, gives the 

 appearance of a mound sloping in all directions. But if the 

 eyes are steady, only the transverse arching is seen. In fact, 

 under these conditions I seem to see a fore and aft concavity, 

 which, combined with the transverse arching, gives a saddle- 

 shape surface. 



Explanation. — Principles. All the phenomena of binocular 

 vision, as already said, may be, and in last resort can only be, 

 explained by the law of corresponding points. In the explana- 

 tion of the phenomena described under experiments 1 and 2, it 

 was only necessary to trace the cause back to the behavior of 

 the double images of a rod viewed binocularly in various posi- 

 tions ; but this behavior is itself explained only by the law of 

 corresponding points. But the phenomena of the last experi- 

 ment can best, perhaps only, be explained by direct reference to 

 that law. It is necessary, however, to have a clear conception 

 of the law in order to understand the explanation. 



Corresponding points are points in the two retinas exactly 

 similarly situated, i. e. taking the central spot as pole, having 

 the same latitude and longitude. Or to express it differently, if 

 the two retinae were laid one on the other so as to coincide in 

 the manner of geometric figures, then the coincident points are 

 corresponding points. Or again, the distance between all corres- 

 ponding points is the same. If therefore we open a pair of di- 



