1882. 119 Trans. N. ¥. Ac. Séz. 
glasses or for binocular combination of images by direct vision, and to re- 
duce the difficulty usually attendant upon stereoscopic vision by the 
latter method. 
VI. To secure the means of producing stereoscopy from perfectly 
similar pictures by making the retinal images of these dissimilar. 
The construction and manipulation of the instrument was illustrated 
in full; but a description of this is not given in the present abstract, 
because already furnished to the Amerzcan Journal of Sctence. A 
printed description also accompanies each instrument, as issued by the 
manufacturers, Messrs. E. & H. T. ANTHONY & CoO., 591 Broadway. 
The feature in binocular vision, mentioned in paragraph VI. above, 
has not been explained hitherto, and has been but rarely perceived. If 
an object possessing three dimensions in space be held within a short 
distance and viewed alternately by the right and left eyes, the retinal 
images of it at these different standpoints are necessarily dissimilar, 
On this principle depends the whole art of stereoscopy, as illustrated 
with the instruments and stereographs in ordinary use. Each of the 
latter consists of two pictures of an object, taken from different points, 
so as to secure dissimilarity. The binocular combination of their 
retinal images hence presents the appearance of solidity, independently 
of any perspective effect secured by art. 
If a large plane surface, on which are drawn similar figures regularly 
recurring at equal distances apart, be appropriately viewed with very 
strong cross vision, so that a phantom image, reduced in size, is seen in 
mid-air, the latter appears slightly curved, with the convexity toward 
the observer. This mere fact was first noticed by Sir DAVID BREw- 
STER, but on his theory of visual triangulation, the phantom image 
should be a plane, parallel to the given plane. The appearance of any 
convexity would disprove his theory, and BREWSTER undertook no ex- 
planation of what was to him, under the circumstances, probably nota 
striking phenomenon. This long-forgotten peculiagity of the phantom 
image, or at least its curvature in one direction, has lately been redis- 
covered by Proféssor LE CONTE, andits curvature in all directions by 
myself. It was at first regarded as one of the consequences of strong 
convergence of visual lines. By using, instead of a large plane surface, a 
card on which the two pictures are perfectly similar, such as a pair of series 
of concentric circles, and cutting the two halves apart, the horizontal vis- 
ual lines may be made parallel while the two small cards are oppositely 
inclined to them, at any desired angle, by rotation about a pair of verti- 
cal axes. Each plane may thus bear to the visual line that meets it the 
same relation that would result from strong convergence or divergence 
in viewing a single continuous plane. The binocular effect is that the 
combined surface appears convex or concave, at will, by varying the 
