THE COLOR VISION OF BIRDS 



2. Cues from the noise and shifting 

 position of the apparatus. 



3. Unequal illumination of the pas- 

 sages due to the reflection of the shorter 

 wave-lengths from black surfaces. 



4. Differences in the pattern of the 

 stimulus plates. 



5. Light leakage in the apparatus 

 and dark room. 



1. Cues from the movements of the a. The experimenter was probably 



experimenter. never visible to the birds, certainly 



not in control experiments. 



b. White lights were substituted for 

 the monochromatic ones. 



a. White lights as above. 



b. Interchanging the monochromatic 

 bands so that all movements of adjust- 

 ment were reversed. 



c. Pretense of shifts during training. 



a. The apparatus was built to avoid 

 this. (Yerkes and Watson 1. c.) 



b. White light was substituted alter- 

 nately for each of the monochromatic 

 bands. 



a. Plates interchanged. 



b. Lights interchanged. 



c. White lights as above. 



a. Tests under 2 above. 



b. The stimulus plates were illumi- 

 nated directly by small Tungsten 

 lamps, with and without interposed 

 color niters. 



These tests were all employed in the study of red-green dis- 

 crimination and the more important also in all other tests except 

 those upon the form of the spectrum. They gave always nega- 

 tive results: occasionally the chicks were detected in attempts 

 to follow the movements of the experiment box, but they never 

 learned to distinguish between a real shift in its position and 

 the pretended shift which was made at every fourth or fifth 

 trial. The results of the other controls limit the reactions 

 definitely to the illumination of the stimulus plates in every 

 important test. They will not be given in detail except where 

 they bear upon brightness vision. 



Adaptation was controlled in the following way. All tests 

 were carried out in a completely darkened room with the ex- 

 periment box enclosed in a black hood. When the chick made 

 a correct choice a 4 ca. p. lamp 18 inches above the food dish 

 was turned on until the chick picked up a bit of food, usually 

 for less than one second. The chick was then returned to the 

 starting box in darkness. Where light adaptation was desired, 

 a 40-watt lamp with etched globe was held in the starting box 

 for 30 seconds before each trial. That this produced a thorough 

 brightness adaptation was shown by the following comparison 



