

90 HAROLD C. BINGHAM 



23+, (November 27), then to 19+ and finally to 15+ before the 

 chick again began to react correctly to the stimuli. When the 

 o 28+ — o 19+ discrimination was again presented, the subject 

 was able to choose correctly in little more than half of the tests. 

 The physical condition of No. 3 on December 1 made it neces- 

 sary to abandon further work. Up to that date it had appeared 

 in all respects normal, but the final series was conducted with 

 great difficulty on account of the common "leg-weakness" 

 which had suddenly developed. With the appearance of this 

 disease, the experiment was abruptly closed. 



The chick, then, had detected a faint streak of light, not 

 more than 6 mm. long and less than 1 mm. wide, — an illumina- 

 tion so small that not only I but others who were working in 

 the laboratory had quite overlooked it. By considering the 

 amount of light that came from the upper illumination, the 

 poor reflecting surface of a steel strap worn only moderately 

 bright, and the narrow surface from which the light was re- 

 flected, (not more than 1 cm. in width), the reader can judge 

 how insignificant this strip of light would appear. It was clearly 

 a case of the chick "outguessing" the experimenter. 



III. SIZE PERCEPTION 



1. Discrimination between unequal circles 



The foregoing discussion of method suggests the nature of the 

 difficulties which were encountered in the early experimental 

 work. No trustworthy results were secured until after the 

 crack of light was discovered. By this time No. 3 was the 

 only subject of the first group that was still working. The 

 later experiments with this chick indicated that the early dis- 

 crimination had been made on the basis of size difference, but 

 as the variable became larger and the discrimination corres- 

 pondingly harder, No. 3 began to look for other cues and hap- 

 pened upon the one which has been described. 



A glance at table 2 suggests the point where No. 3 ceased to 

 discriminate between the two size stimuli. After the control 

 tests, o 28+ — o 28+ discrimination, during which the crack 

 was stopped, the chick's response to 28+ — o 23+ was tested. 

 The results were decidedly negative since the right stimulus was 

 chosen no more than chance would allow. The response to 



