438 Mr. W. Le Conte Stevens on Wheatstone and 



or implicitly in every explanation of the stereoscope to which 

 I have had access, where a diagram is employed in tracing the 

 course of rays entering the eyes. That it is misleading and 

 should be abandoned, is shown by the fact that the perception 

 of binocular relief and the judgment of apparent distance in 

 the stereoscope are easy when conditions are such as to make 

 the visual lines slightly divergent. If their intersection deter- 

 mines the point of sight, this would then be behind the ob- 

 server's head. In taking the pictures composing the stereo- 

 graph, each point in the field of view is at the intersection of 

 a pair of secondary camera axes; and thus far, but no further, 

 Wheatstone's formula is strictly applicable. No diagram can 

 give more than an approximation to the truth when the pic- 

 tures are viewed in the stereoscope ; and not one hitherto pub- 

 lished has exhibited any provision for the possibility of optic 

 divergence. I have elsewhere* shown how such a diagram 

 may be constructed by application of Hering's theory of a 

 binocular eye {ceil de cyclope imaginaive)\ , assuming that we 

 know the angle of convergence between the camera axes for 

 a given point in the field of view. 



The fact that optic divergence is possible with distinct bin- 

 ocular fusion of retinal images is not new. The use of prisms 

 for testing the rectus muscles of the eyeballs has long been 

 known; and Helmholtz not only mentions the use of stereo- 

 graphs for the same purpose, but observes the increase in ap- 

 parent distance of the point of sight when divergence of 

 visual lines is induced J. No analysis of the visual pheno- 

 mena seems to have been made, however ; and their important 

 bearing upon the theory of stereoscopic perspective seems to 

 have escaped notice. Helmholtz says§ in regard to such illu- 

 sions of sight, " If these images are such as could not be pro- 

 duced by any normal kind of observation, we judge them 

 according to their nearest resemblance ;" but comparison is 

 scarcely satisfactory if direct analysis is possible. 



I have examined many stereoscopes and stereographs in 

 testing Brewster's theory ||, and have found that optic diver- 

 gence is often necessary; hence it is unconsciously practised 

 by nearly every one who uses them. What is really requisite 

 is not that the visual lines shall converge, but that the eyes 

 shall receive upon corresponding retinal points the images of 



* American Journal of Science, Nov. and Dec. 1881. 

 t Hering, Beitrdge zur Physiologie : or Helmholtz, Opt. Plnisioloqiqne, 

 p. 777 (edit. 1867). 



\ Opt. Phys. pp. 616 & 828. 



§ ' Popular Lectures on Scientific Subjects/ p. 307 (tr. 1873). 



|| American Journal of Science, Nov. and Dec. 1881. 



