156 The Animal Mind 



animation (231). Lubbock's tests with wasps gave nega- 

 tive results. Lovell (439) and Turner (725) also infer 

 color vision in the honey-bee from its ability to pick out 

 objects of the same color as that on which it has recently 

 found food : the former takes no account of the brightness 

 error, while the latter holds that it has been sufficiently 

 eliminated by the fact that the color identifications were 

 made by the bees under varying lights and shades out of 

 doors. This, however, is probably an inadequate pre- 

 caution. Von Frisch (246) offers more convincing evi- 

 dence of color vision in the bee, and thinks he has indica- 

 tions that bees are red-green color-blind. His experiments 

 were performed in the open air. Having trained the bees 

 to come to strips of yellow paper, on which food was placed, 

 he mingled such strips, without food, among strips of thirty 

 different shades of gray. The bees, he reports, were able 

 to make the discrimination, and to do equally well when 

 blue was used : they failed, however, with red, confusing 

 red-violet with blue, and dark red with dark gray. In 

 another article (247) he says that a certain bluish green 

 also was confused with gray. This general method, where 

 a large range of grays is used and an animal proves capable 

 of discriminating a color from any or all of them, is the 

 best way of eliminating the brightness error. The use of 

 colored papers in experiments on color vision in animals 

 is open to criticism unless some precaution is taken against 

 the possibility that something in the surface texture or 

 grain of the papers may help the animal to distinguish. 

 This possibility Von Frisch guarded against by varnishing 

 the papers, a proceeding which did not affect the behavior 

 of the bees. 



Hess (312, 315), anxious to defend his theory of the total 

 color-blindness of all invertebrates, repeated Von Frisch's 



