METHODS OF STUDYING VISION IN ANIMALS 85 



make loo per cent, of correct choices, on the basis of the differ- 

 ence in illumination of the two compartments. 



This defect in the method of presenting chromatic stimuli 

 is not pecuHar to spectral light. Any transmission method 

 offers the same difficulty. So far as we know this defect has 

 not hitherto been called to the attention of experimenters. 

 The difference in illumination after 10-15 minutes of darkness 

 adaptation is so marked that one must notice it. The phenom- 

 enon does not depend upon the difference in energy (or " bright- 

 ness ") of the two colors. It is apparent even when the green 

 is but one-twentieth the energy of the red. It is still more 

 apparent when the two colors have been judged " equal in 

 brightness " by direct comparison. It is due chiefly to the 

 fact that black paints offer a surface of reflection for rays of 

 the green-blue- violet regions of the spectrum. 



After long experimentation we have found two ways of wholly 

 eliminating it. The first method calls for the abandonment of 

 tests with complete darkness adaptation. One mounts a 40 

 watt tungsten lamp in a light-tight metal cylinder, in the front 

 end of which is fitted a graduated iris diaphragm with a disc 

 of ground glass behind it. The diaphragm can thus be opened 

 and closed to regulate the intensity of the Hght. Since the 

 light is perfectly diffused by the ground glass, the size of the 

 illuminated field does ■ not appreciably change when the dia- 

 phragm is opened or closed. This light is mounted centrally 

 over the experiment box at a height of 1.5 meters. Care muct 

 be taken that the white light offered by this source does not 

 enter the cabinet (p. 57) and fall on the plaster of paris surfaces. 

 With this illuminating device the two food compartments may 

 be so highly illuminated that the relatively slight difference in 

 illumination, arising from the source discussed above, is com- 

 pletely eliminated. Entirely apart from its uses in this con- 

 nection, the device gives us the means of studying the formation 

 of associations under standard conditions of light adaptation. 



The second method consists (a) in mounting a series of dia- 

 phragms in front of the windows in the experiment box which 

 admit the colors; (b) in making the floor of the experiment 

 box non-reflecting, by means of a metal grill; (c) in fastening 

 dead-black velvet to the sides of the box and to both sides of 



