338 J- F. DASHIELL 



The one outstanding difference in the records is that of the 

 initial trials of groups I and II on the one hand and of groups 

 III and IV on the other. Groups I and II show for the first trial, 

 in time, 149.8 and 130 seconds, respectively, in errors, 11.6 and 

 11, respectively. Groups III and IV show for the same, 49.2 and 

 59.5 seconds, respectively, and 5.6 and 5.7 errors, respectively. 

 A note made just after the first test trial reads: " Conspicuous 

 feature of rats of III and IV on first trial in Q = absence of hes- 

 itancy in general behavior." In the further course of learning, 

 however, no marked differences emerge, the averages of groups 

 III and IV being only slightly better than those for I and II. 

 Now the one difference between the methods of preliminary train- 

 ing for groups I and II taken together and for groups III and IV 

 taken together is in the non-presentation and the presentation, 

 respectively, of blind alleys or culs-de-sac. That an animal is 

 made previously acquainted with the physical materials of which 

 his test maze is to be constructed, seems, then, to be decidedly 

 less important than that he be given acquaintance with some 

 blind alleys; and furthermore, experience in gradually eliminating 

 blind alleys seems to be decidedly less important than mere 

 acquaintance with the alleys. (Compare group III with group 

 IV.) The latter point seems further borne out by the observa- 

 tion that if training in actual elimination of blind alleys were 

 highly important it would surely show itself in an accelerated im- 

 provement after the first trial and throughout the learning pro- 

 cess for group III which is only to a slight degree the case. 



A conclusion to be drawn from this part of our study would be : 

 When training of white rats in one maze is found to transfer or 

 carry over to the learning of another maze, the principal factor 

 that is transferred is adjustment to maze-situation-in-general. 11 

 Possible secondary factors are adjustment to physical materials 

 and adjustment in the form of elimination of constant errors in 

 successive runs in one maze. A practical corollary that follows 

 is that for the learning of a maze de novo no rat should be used that 

 has previously been run in any kind of maze whatever. 



11 Wiltbank (op. cit., p. 35) would explain this positive transfer in early trials 

 in terms of the recession of the instinct of timidity. This would seem more 

 relevant to the elimination of surplus time than of errors. 



