68 



FREDERICK S. BREED 



39 to black-blue. The preference tests are not representative 

 of our chicks in general, for those chicks that already had a 

 preference for black were rejected because they were not suitable 

 for the special purposes of the experiment of which this black- 

 blue training formed a part. There were a few of these. But 

 this does not affect in the least the point that I wish to make 

 in connection with the effects of training. The state in which 

 the black-blue modification persisted at the time of the later 

 color preference tests is shown in the columns under " Per- 

 sistence." The tests on blue-white, black-white, blue-yellow, 



TABLE 24 



Some Effects of Training 



blue (tint no. i) -yellow, blue-blue (tint no. i) were given in 

 the order named and, in all but two cases, immediately fol- 

 lowing the persistence tests. The most striking feature of these 

 results is again the character of the reactions to blue. If bright- 

 ness were the only factor, a chick trained to black-blue would 

 be expected thereafter to react positively to the less bright of 

 two stimuli, other things being equal, when a different combina- 

 tion were presented. But chicks that one minute accepted 

 black and rejected blue, that is, selected the card of lower bright- 

 ness value, the next minute accepted white and rejected blue. 

 How shall we explain this behavior if we consider merely the 



