104 J. LeConte — Phenomena of Binocular Vision. 



the stereoscope, and thus also producing reverse effects. But 

 it has also the great, the fatal disadvantage for our purposes, 

 that the phenomena of curvature, as already explained, are 

 complicated by a curvature developed even by geometric con- 

 struction. 



For example, figs. 3 and 4, represent cards with regular fig- 

 ures bent in the middle, backward in fig. 3 and forward in fig. 

 4. The reason of the stronger curvature is obvious on inspec- 



3. 4. 



a 



tion of the figures. By geometric construction the phantom in 

 fig. 3 is already convexly curved and by the law of correspond- 

 ing points, as already explained, would be still more curved, as 

 shown in the dotted line. Similarly in fig. 4 by geometric 

 construction the phantom would be concave, and by the law of 

 corresponding points still more concave, as shown in the dotted 

 line. In this latter case inspection of the figure shows that 

 with the point of sight at c, points right and left as a, b, d, e, 

 of phantom, would fall on non-corresponding points of the two 

 retinae a a, bb, do 7 , ee farther apart than cc, and therefore the 

 impressing object will seem nearer than the point of sight and 

 nearer also than geometric construction would make it. This 

 would increase the apparent curve. In other words, even with 

 flat retinas the phantom surface would in both cases be curved, 

 but the curve of the retinas in both cases exaggerates the curva- 

 ture of the phantom. 



It is needless to say, that in combination beyond the plane 

 of the cards, whether by naked eyes or by stereoscope, the phe- 

 nomena would be exactly the same, only reversed, i. e. in 

 fig. 3, it would be concave and fig. 4 convex. 



Another phenomenon I have observed in these last experi- 

 ments of cross combination. In fig. 3 the phantom is not only 

 convex from side to side, but concave fore and aft, i. e. it is 



