LABYRINTH HABITS OF THE PIGEON 299 



co-ordinations used sixteen days before may have persisted 

 unimpaired by the intervening training. In this case, the in- 

 stance would be one of (organic) memory. (2) The pigeons 

 may each have succeeded, during the intervening training, in 

 acquiring a system of cues which would be undisturbed by any 

 degree of rotation. These two theories should have been tested 

 by noting the effects on the birds of degrees of rotation as yet 

 untried, as well as by repeating tests for 90° and 270°. 



The two hypotheses just advanced cover the case of pigeon 

 no. 8, who was confused at 90° and 270°, but who was perfect 

 at 360°. Comparative psychologists are already familiar with the 

 view that an animal may use one set of cues in learning a maze 

 and another in running it after learning is completed. But so far 

 as the present writer knows facts have never before been presented 

 in support of the hypothesis here presented, viz., that an animal 

 may change its system of cues from time to time in response to 

 the varying demands of a situation. This view of the matter 

 is of more than passing interest in that it at least suggests the 

 complexity of the animal mind for types as high as the pigeon. 

 Most investigations in animal behavior (particularly in maze 

 problems) lay chief stress upon the simplicity of animal con- 

 sciousness. The other alternative should not be overlooked 



and neglected. 



HABITS IN LABYRINTH C 



Tests for kinaesthetic control. Four pigeons, nos. i, 2, 5 and 7 

 were tested in this labyrinth. All had gone through the tests 

 in labyrinth B, and with the exception of no. 5, they had all 

 been trained in A as well. Constant training had made the 

 birds quite docile so that a maximum efficiency record might 

 now be looked for. The results are given in Table VII and 

 fig. 9. The long alleys in this maze were very confusing at first 

 — the birds were constantly turning back before the ends of the 

 alleys were reached. All were perfect in the 54th, 55th and 56th 

 trials. Nos. 5 and 7 both made errors in the 57th trial. The 

 cause of the former's errors is unknown. The latter apparently 

 blundered because he was going too fast. The learning of this 

 maze was accomplished a few trials earlier than in the case of B. 

 However, the learning curve possesses essentially the same form 

 as in the former case. This, as was pointed out above, lends 

 much additional support to the hypothesis which would make 



