The only experimental evidence hitherto existing, so far as we have been able to learn, that fish possess the power to discriminate colors is contained in the work by VITUS GRABER, published over twenty years ago and entitled Grundlinien zur Erforschung des Helligkeits-und Farbensinnes der Tiere.1 GRABER experimented on a large number of animals, including two species of fish, Cobitis barbatula and Alburnus spectabilis. His method was to offer the animals the choice between two compartments differently illuminated, and at the end of a given period to count the number in each compartment. T h e results thus tested light-preferences rather than light-discrimination merely. GRABER himself points out that the two do not coincide, inasmuch as an animal may be quite capable of distinguishing between two colors and yet find them so nearly equal in feeling-value that it seeks them equally often. There must also be reckoned with the possibility that apparent color-preferences are really brightness-preferences, due to the difference in brightness between the two stimuli employed. This latter difficulty GRABER avoided in the following manner:
If an animal showed itself to be, in our modern phrase, positively 'Rough experiments performed some years later by W. BATESON (Jour. of the Marine B i d . Assoc. o f t h e UnitedKingdom,N. S.,Vol. I, 1889-90, p. 225) gave, as the author says, "chiefly negative results." BATESON fed young mullet with minced worms sprinkled on tiles of various colors and noticed that the light-colored tiles were first cleared of food. It is to be remarked that BATESON was working for pteference and not for discrimination of color and also that he neglected to control the element of brightness.