CHAP, in.] SIGHT. 1287 



turn our eyes from something near to something farther off. In 

 a large number of instances this change to convergence from 

 parallelism, or this increase or decrease of convergence takes 

 place without any change in the visual plane, without any raising 

 or lowering of the visual axes ; in such instances the movement is 

 carried out in convergence by the two internal rectus muscles, or 

 in decrease of convergence by the two external rectus muscles ; 

 and the only coordination necessary is one which secures that the 

 muscle of one eye should work in harmony with the muscle of 

 the other eye. But even this relatively simple movement is 

 limited in a very marked way. We can bring the visual axes of 

 the two eyes from a condition of parallelism to one of almost any 

 degree of convergence, but we cannot, without artificial assistance, 

 bring them from a condition of parallelism to one of divergence. 

 The stereoscope will enable us to create such a divergence. If 

 in a stereoscope the distance between the pictures be increased 

 very gradually so as carefully to maintain the impression of a 

 single object, the visual axes may be brought to diverge ; and the 

 subject of the experiment may himself be made aware of the 

 divergence, by the sudden removal of the instrument from his 

 eyes ; his vision of external objects is for a moment double, but 

 for a moment only. This experiment shews the reason of the 

 limitation of which we are speaking. So long as the visual axes 

 are parallel or appropriately convergent the images of external 

 objects fall on corresponding parts of the two retinas, and single 

 vision results; when the visual axes are carried beyond parallelism, 

 the images on the two retinas are not on corresponding parts and 

 vision is double. Thus, as regards convergence or divergence of 

 the visual axes, the movements of the two eyes are governed by 

 the principle that the will can of itself only carry out those move- 

 ments which are consistent with images of external objects falling 

 on corresponding parts of the two retinas. There is an exception 

 to this in the case of extreme convergence ; we can as in 

 squinting make the visual axes converge too much, and in conse- 

 quence by a simple effort of the will can obtain double vision ; but 

 this is probably in order to leave a margin which shall secure our 

 being able to use to the utmost our accommodation mechanism 

 for near objects ; otherwise the rule holds good. Not only 

 so, but as the above experiment also shews, when by artificial 

 assistance, which is in itself directed towards securing single 

 vision with the two eyes, we obtain divergence of the visual axes, 

 immediately that the assistance is done away with the axes 

 return, by an involuntary movement, to parallelism ; the double 

 vision occurring at the moment of removal of the instrument 

 rapidly gives way to normal single vision. Other illustrations 

 of the same principle may be met with. For instance, if a 

 distant object be looked at with both eyes, but with a prism 

 held horizontally before one eye, and if the image of the object 



