148 M F. WASHBURN AND EDWINA ABBOTT 



compartments of the box ; but the door carrying the gray paper 

 was fastened on the inside. A rabbit approaching the red door 

 and pushing the lower part of it could get food from the inside; 

 the gray door would not open. The places of the red and gray 

 papers were exchanged in successive experiments (see figures 

 2 and 3). 



One of Watson's objections to the use of colored papers is 

 that if pasted on supports they cannot fail to present inequali- 

 ties of surface, wrinkles, which will make them easily distin- 

 guishable by this means alone. Evidently here, where the papers 

 were not pasted but pinned on the doors in each succeeding ex- 

 periment, the irregularities of surface could not possibly serve 

 as means of identification, since they were probably different 

 each time the papers were fastened on the doors. Another pos- 

 sibility of error pointed out by Watson as involved in the use 

 of colored papers is that papers of different colors may have 

 differences of surface texture that would render them discrim- 

 inable even by a color-blind subject. As will be seen, our rabbits 

 proved able to learn the discrimination between red and various 

 grays, but unable to acquire that between red and the black 

 papers used : yet the surface texture of the red paper to ordinary 

 human vision differed more from that of the black papers used 

 than it did from that of the gray papers. A special test of the 

 "texture" possibility of error was made with two of our rabbits. 

 One of them, "Abednego," had made a perfect record in dis- 

 criminating red and gray number 1 5 for fifteen consecutive days, 

 eight tests being performed a day. On the sixteenth day pieces 

 of red and grey velvet, matching the papers as closely as pos- 

 sible, were substituted for them. The character of the surfaces 

 here should have been the same, and should have been sufficiently 

 unlike that of the paper surfaces to confuse the rabbit if he had 

 been depending on surface texture. He made a perfect record 

 in the ten tests made, although he seemed much puzzled by the 

 unaccustomed "feel" of the velvet on his nose, and twice after 

 touching the red drew back and seemed to look at it. The other 

 rabbit, Polly, gave exactly similar results: an interpolated series 

 of ten velvet experiments did not cause her to make any errors 

 in the discrimination of red and gray number 15. 



The smell error in its various forms was dealt with as follows. 

 Since food was always, and in equal amounts, in both com- 



