LABYRINTH HABITS OF THE PIGEON 



289 



learning records. The conventional graphs show almost an 

 immediate descent. What then is the explanation of the present 

 results and what are their imjDlications ? 



«^> ■>(■ a so 



Trialt 



^ ^) > "o Co 



Figure 7 — Error graphs for the first twenty-one trials of pigeons nos. 1, 2, and 7 

 (A) and nos. 5 and 8 (B) in labyrinth B. 



Two reasons might be advanced in general for the mainte- 

 nance of the learning curve at or about a certain height in ex- 

 periments with mazes related as A and B. (i) The blind alleys 

 in the two labyrinths are such that the habits acquired in A 

 interfere \^ith the learning of B. The persistence of acquired 



