PATTERN-DISCRIMINATION IN VERTEBRATES 351 



Usually only a very short fraction of a second would elapse 

 between the noise of his rubbing against the window and the 

 flicker of the signal lamp. But as his responses became slower 

 I began watching him through a crack in the top of the box, and 

 continued from May 7 to the end of the tests. Between the 7th 

 and the 13th of May this irregularity never occurred; between 

 the 13th and the 31st, in a total of 360 trials, it occurred five times. 

 That is, the animal would extend his head and sometimes part 

 of his body through the window opening into the wrong alley, 

 and then reverse the choice without dropping on to the grill. 

 Each time I threw out the choice so made and repeated the trial 

 elsewhere in the series; four times the repetition resulted in a 

 correct choice and once incorrect. This shows that the distance- 

 factor is very constant and reliably controlled even in the case 

 of the monkey. 



I found that the monkey was not disturbed by occasional 

 visitors when the latter were seated in an inconspicuous place. 

 For the purpose of getting suggestions as to possible secondary 

 sources of error I conducted several series of tests in the presence 

 separately of Dr. Cobb, Dr. Worthing, Mr. Cady and Mr. 

 Luckiesh of the research staff of this laboratory, and Dr. Breiten- 

 becker of the Western Reserve University. 



Chick 1 is an unsatisfactory animal for this work because of 

 extreme excitability. It was difficult to control the current 

 used as punishment so as to affect him at all without so dis- 

 turbing him as to render him unfit for work. The learning 

 record in Table 3 shows that I did not always succeed in this 

 particular. Several times I had to remove him from the box 

 and leave him for 24 hours without food, in order to obtain 

 responses. At other times a rest of several days seemed necessary. 

 However, it was possible, by re-training him at larger stimulus- 

 values, to overcome such disturbances for the time, and he 

 actually yielded a slightly lower threshold than did Chick 2, 

 who was a much better subject. The learning record of the 

 latter I consider quite good. 



Reduced to terms of visual angle subtended by individual 

 striae, the values taken as the stimulus-threshold are for Chick 1, 

 4' 04"; for Chick 2, 4' 14"; for Monkey 2, 57". Generally 

 speaking, such " threshold-values " obtained by the discrimina- 

 tion should not be construed too rigidly. They certainly are not 



