THE COLOR VISION OF BIRDS 7 



motive than hunger the apparent limen might be much lessened. 

 The statements of other investigators imply that the chick is 

 very strongly photo-positive and will choose the brighter of 

 two lights of almost equal intensity. As this was found to be 

 untrue, long training in brightness discrimination was necessary 

 in order to prepare the animals for the following experiments 

 upon the Purkinje phenomenon. The training was continued 

 for 500 trials after fairly accurate brightness discrimination had 

 been established. Many changes in the brightness of the lights, 

 position of the experiment box, noise of rotating sectors, illumi- 

 nation of the experiment box, etc. were introduced to accustom 

 the animals to changing experimental conditions. 



When reactions to brightness had become automatic mono- 

 chromatic lights were substituted for the white lights. Those 

 first introduced were Red 650/^ and Green 520^/*. The inten- 

 sities of the two beams were equated with each other by the 

 method and with the standard of Pfund 1 ('12, Yerkes and Wat- 

 son, '11). 



a. The fowls, light adapted, were offered a choice of the 

 stimulus plates with these lights at standard intensity, food 

 being given at every trial. They chose — 



Green Red 



Chick A 12 9 



Chick C 19 11 



Average 63% 37% 



The green had perhaps the greater stimulating value under 

 these conditions but the difference in brightness for the chick 

 was evidently not very great. 



b. The chicks were next dark adapted for one hour and the 



illumination of the stimulus plates was cut down to 1/90 of the 



standard intensity. Chick C was tested under these conditions. 



He chose — 



Green Red 



13 1 



93% 7% 



At this energy the green had a much greater stimulating value. 



^his standard has not been expressed in other terms. The stimulus patches 

 illuminated at the standard energy have a luminosity which I estimate for 580 

 ufi. at 2 candles per square meter. In the tests for the Purkinje phenomenon the 

 lights were equated with the Pfund standard after reflection from the stimulus 

 patches. * In all other tests the energies of the beams were equated before reflection 

 from the diffusing surface. 



