of Binocular Vision. 



189 



10' for each Hue with the vertical, or 20' with one another ; for 

 this angle I can distinctly detect under these circumstances. I 

 next constructed a diagram like Professor Helmholtz's, except 

 that the outward inclination of the verticals was only 40' in- 

 stead of 1}°. On combining the two halves of this diagram by 

 means of a stereoscope, there really seemed to be perfect coin- 

 cidence of both verticals and horizontals ; but I soon found, by 

 trying several, that stereoscopes differ much in this respect. I 

 therefore discarded them as unreliable. On combining the same 

 diagram with the naked eye in the manner of a stereoscope, i. e. 

 beyond the plane of the diagram, the verticals coincided per- 

 fectly when the point of sight was about twelve inches distant, 

 but the horizontals very perceptibly crossed, though certainly, 1 

 think, at an angle less than 40' (it seemed about 20'). On com- 

 bining the two halves by squinting (of course turning the dia- 

 gram upside down), I found the result precisely the same when 

 the point of sight was at the same distance, viz. 12 inches. In 

 the next diagram which I constructed the verticals inclined 1£° 

 and the horizontals 50', the difference being therefore 25'. In 

 this case both seemed to combine perfectly when the point of 

 sight was distant 1\ inches. The next diagram tried had the 

 verticals inclined 5° and the horizontals 3° 45', the difference 

 being 1\°. In this case both verticals and horizontals combined 

 perfectly at the distance of 2*2 inches. I then tried one in 

 which the verticals inclined 10°. In this case I could not make 

 perfect coincidence of both verticals and horizontals until the dif- 

 ference of inclination was made as great as 5°. The diagram 

 used is shown reduced in the figure (fig. 10). The point of 



Fig. 10. 



sight in this experiment was only 1£ 



ich from the line joining 



