282 WALTER S. HUNTER 



The climate in this portion of the country is ideal for outdoor 

 work. The regular experiments were never interfered with by 

 bad weather. 



The preliminary tests, which were given to all of the birds, 

 trained them to go directly from E to F, — no partitions had as 

 yet been placed in the maze. This training resulted in the 

 establishment of an association between the maze and getting 

 food. As a result, the first trials in the regular experiments 

 were influenced by as strong a motive to thread the maze as 

 were the immediately subsequent ones. Again, the habit which 

 was formed of going directly across the box was one that had 

 to be broken up when the partitions were placed in the maze. 

 As soon as a pigeon had entered F, the door separating F from the 

 maze was dropped in order to prevent retracing. This retracing 

 was not prevented in Rouse's work." What the exact result 

 may be, it is impossible to say. At the close of these preliminary 

 tests, work with pigeons nos. 5 and 8 was discontinued. How- 

 ever they were handled regularly every night and morning in 

 order to prevent them from becoming wild again. 



Throughout both the preliminary and the regular tests, the 

 birds were given three trials daily. The exact pathways fol- 

 lowed were represented upon a small plan of the maze. 



HABITS IN LABYRINTH A 



/. First series, — normal learning records. The results for this 

 series are recorded in table I. An inspection of this will reveal 

 a typical learning curve. Both times and errors gradually dimi- 

 nish until the final values are reached.^ There is but slight 

 variation in the time records tow^ard the last. This is due to 

 the simplicity of the maze and to the shortness of the runs. The 

 pigeons never became automata. Although they made the runs 

 rapidly and without a pause, the}^ were constantly on the alert 

 in passing alleys. The number of trials, however, was hardly 

 sufihcient to produce automatism. It is a point worth noting 



* Loc. cit., p. 592. 



* In this paper the term " error " designates, perhaps as in Porter's paper:* 

 (1) every entrance into a cul-de-sac; (2) every turning back from its exit, when 

 within a bUnd alley: and (3) all returns toward E over the true pathway. These 

 three classes of errors are counted of equal value. No distinctions are made between 

 errors on the basis of the distance covered in making them. 



« Porter, J. P. Further Study of the English Sparrow and Other Birds. Amer 

 Jour, of Psychol., 1906, vol. 17, p. 253. 



