216 H. M. JOHNSON 



the unaccommodated eye of Dog 1. An object 133 cm. distant 

 may be imaged on at least some parts of the retina if the animal 

 accommodates or rotates the eye properly. In the discrimina- 

 tion work just described, the animal had to choose with the eye 

 less than a meter from the test-fields, and he actually chose, as 

 a rule, with the eye at the minimal distance, which in part of 

 the work was 60 cm. and in the rest, 40 cm. There was some 

 room for doubt, therefore, whether a clear image had been 

 formed on the retina under the prescribed conditions. Further 

 work was necessary to determine whether the dog's inability to 

 discriminate was due to retinal insensitivity or to other causes. 

 I proceeded at once to modify the method so as to insure 

 that a sharp image of the patterns to be discriminated would 

 be formed on the retina of the animal's unaccommodated eye. 

 There are several means by which this could have been accom- 

 plished. We could have mounted the test-objects 20 feet away, 

 so that rays proceeding from each point on them would have 

 reached the dog's eye as nearly parallel rays. But the test- 

 objects are small, and might not have attracted the dog's atten- 

 tion. The largest striae which these particular gratings will 

 form subtend a very small angle at 20 feet. Negative results 

 thus obtained would have been ambiguous. Of course it would 

 have been possible to project an enlarged image of the patterns 

 on to a distant screen, but in such case modification of the Yerkes 

 experiment-box would have been necessary. It would also 

 have been possible to equip the dog with spectacles. Had this 

 course been adopted it would have been necessary to use extreme 

 care to insure that the animal inspected the test-object at a 

 constant distance. This procedure involved some practical 

 difficulties. Another course seemed most nearly free from ob- 

 jectionable features. I took two lenses, each having a free open- 

 ing of 6 cm. and an equivalent focus of 40.8 cm., and mounted 

 a test-object in the first principal focus of each of these lenses. 

 Each of the two optical systems thus formed was mounted on 

 a photometer bench, so that the face of the lens remote from the 

 test-object was presented at the window (W 1? W 2 , fig. 6, p. 338 

 of the first article of this series) of the Yerkes box. The axis 

 of each optical system was normal to the plane of the window at 

 which it was presented, and coincided with the center of the 

 latter. The lenses are achromatised for two points in the visible 



