LABYRINTH HABITS OF THE PIGEON 295 



error was avoided, save twice. But this is not all. As the 

 reader has doubtless observed, it is difficult to say whether or 

 not pigeons nos. i, 2 and 7 were confused by the rotation. An 

 explanation of this will be offered below. It is not to be assumed 

 that the visual cues here in question were acting alone. They 

 might be modified by kinaesthetic factors. The extent to which 

 it is probable that these latter were present will be touched upon 

 later in this paper. 



Circumstances did not permit the writer to rotate the 

 environment with the maze. Had this been done the dis- 

 turbance due to the change of external visual cues in relation 

 to the maze would have been avoided. What the pigeon's be- 

 havior would have been under those circumstances cannot be 

 stated with certainty. It is probable that there would have 

 still been individual variations depending upon the cues in use 

 by the several birds. But certainly if the analysis given above 

 and continued later in this paper is valid, those pigeons de- 

 pendent upon external visual cues would not have been disturbed 

 in their reactions. 



Table V gives the results for a rotation of 270° to the 

 left from the position in which the normal learning record 

 was obtained. The confusion was greater in this case than in 

 the one above. A normal error record was not reached until 

 after the eighteenth trial; the time record was reduced until 

 the last. If the explanation ventured in the above case of the 

 influence of the external visual factors were true, it would only be 

 in evidence here as a tendency to go slowly or even to turn back. 

 It is difficult to say whether such tendencies were present in 

 unusual force or not. However, the diary record— kept before 

 this explanation was thought of — contains much reference to 

 "going slowly" and to constantly "turning back." A marked 

 peculiarity of this table is that pigeon no. 5 made only one error 

 (in the fifteenth trial), and it could hardly have been due to 

 the rotation. 



After this series of tests the maze was rotated to 360°, Table 

 VI. Nine tests were made, but no confusion was present. 

 As will be seen later {vide infra, pp. 298, 299), this may indicate 

 either that the original co-ordination persisted for sixteen days or 

 that between the two tests the birds had acquired a system of 

 cues that would save them from anv future confusion. 



