374 PROCEEDINGS OF THE AMERICAN ACADEMY 



4. Trace a circle round the secondary image with a lead pencil. 

 Fasten a wafer on the end of a narrow strip of card. Push this wafer 

 into the circle just traced. The two wafers will coincide, and appear 

 as one. Draw the wafer gradually away, and it will carry the second- 

 ary image with it, leaving the circle blank. Continue drawing it away, 

 and at a certain distance (about one diameter of the wafer, for in- 

 stance) the secondary image will separate from the moving wafer, and 

 sail very slowly back into the traced circle. 



5. Close one eye and look at a window with the other. Now shut 

 both eyes, and there will be a spectrum of the window in the eye which 

 was opened, and in that only. 



6. Repeat the experiment, keeping both eyes open, but in such a 

 position that only one shall see the window. Close both eyes, and 

 there will be a spectrum of the window in both eyes, most distinct, fre- 

 quently, in the eye to which the window was not visible. 



These experiments appear to show that an image formed on one 

 retina produces a retinal spectrum undistinguishable in many cases from 

 a direct retinal image. That the seat of this secondary spectrum is the 

 retina, is shown by the fact that the eye must be opened in order that 

 it shall be perceived. The retina seems to require the -stimulus of 

 light in order to repeat the impression. Again, in the sixth experi- 

 ment, the spectrum in the eye which has not seen the object is like 

 that in the eye which has seen it ; and this is always recognized as a 

 retinal spectrum. 



If the conclusion from these experiments is correct, the transfer of 

 impressions from one retina to the other falls into the great category of 

 reflected nervous actions, and is properly called Rejlex Vision. The 

 recognized connection of the retina by looped fibres, the decussation of 

 the optic nerves, the connection of the optic ganglia, afford abundant 

 anatomical confirmation of the probability of the suggestions offered. 



To state the general result of the experiments briefly : — 



If an object, A, is seen by one eye only, both being open, there will 

 be a direct image, o, and a reflex image, a'. The retinal impressions 

 will be represented by a -j- a'. 



If ^ is seen by both eyes, the retinal expression will he a a' -\- a a'. 



If A by one eye, and B by the other, a h' -\-b a'. 



The direct and reflex impressions exactly coincide in the normal 

 state, except so far as ocular parallax makes a difference between 



