ABBE ON STEREOSCOPIC VISION 95 



aperture, good delineation with these must be confined to thinner 

 objects than can be successfully employed with objectives of narrow 

 apertures. 



Stereoscopic vision with the microscope, therefore, is due solely 

 to difference of projection exhibited by the different parallactic dis- 

 placements of the images of successive layers on the common ground 

 plane and to the perception of depth, not to the delineation of the 

 plane layers themselves. For, if there were dissimilar images per- 

 ceptible at different planes, the ouf-of-focus layers must appear con- 

 fu.M-d and no vision of depth would be possible. 



Now stereoscope vision requires, as shown by Dr. Carpenter, that 

 the delineating pencils shall be so divided that one portion of the 

 admitted cone of light is conducted to one eye and another portion to 

 the other eye. If this division of the image is effected in a symmetri- 

 cal way. the cross section of, e.g., a circle must be reduced to two 

 semicircles representing one of these two arrangements seen in 

 o and P,fig. 70. 







FIG. 70. 



Dr. Abbe's theory is that the only condition necessary for ortho- 

 scopic effect in any binocular system is that these semicircles oi- 

 l-heir equivalents should be depicted according to diagram O, fig. 70, 

 and for pseudoscopic effect according to diagram P in the same figure ; 

 and he demonstrates that all other circumstances, such, e.g., as the 

 crossing of the images, are wholly immaterial. 



Orthoscopic vision is always obtained when the right half of the 

 right pupil and the left half of the left pupil only are employed ; 

 pseudoscopic vision in the opposite conditions. ' It is quite indif- 

 ferent whether the effect is obtained with crossing or non-crossing 

 ray>. whether the image be erect, or inverted, or semi-inverted, and 

 whatever may be components of the optical arrangement.' 



The observant reader will perceive that it is at this point that 

 there is a radical divergence from the interpretation given by Dr. 

 Carpenter, who, as we have seen above, insisted that orthoscopic 

 vision is not to be obtained in a binocular with non-erecting eye-pieces 

 unless the axes of the two halves of the admitted cone cross each 

 other. 



Of course we must keep clearly before us the fact that in micro- 

 scopic vision it is not the object but its virtual image only that we 

 see. This apparently solid image is placed in the binocular micro- 

 scope under circumstances similar to those of common objects in 

 ordinary vision. Clearly, then, it is the perspective projections of 

 this image which require to be compared to the projections of solid 

 objects in ordinary vision, in respect to which the criteria of ortho- 

 scopic and pseudoscopic vision have been defined. But it can be 

 geometrically demonstrated that right-eye perspective of the ap- 

 parently solid image is always obtained from the right-hand portion of 

 the emergent pencils, left-eye perspective from the left-hand portion ; 



