THE COLOR VISION OF BIRDS 25 



comparable, but some unrecorded experiments by Dr. Middleton 

 and myself upon the fowl's difference limen for brightness indi- 

 cate that where threshold intensities are not involved habits of 

 reaction to brightness are established with much greater difficulty 

 than are those of reaction to wave-length. 



From the evidence even of a. alone I believe that the field 

 experimenter may feel confident that, if the birds show a differ- 

 ential reaction to colored objects not differing enormously in 

 brightness for him, their reactions are made upon the basis 



of wave-length. 



DISCUSSION 



One fact the experiments seem to have brought out clearly: 

 The fowl is sensitive to differences of wave-length irrespective 

 of intensity and its sensitivity is greater than, or at least different 

 from, that of any form of partial color blindness in man. Less 

 certainly it appears that for the fowl the spectrum is divided 

 into areas of widely different reaction value, of which there are 

 not less than five, probably more. It is not probable that all 

 these occupy areas of the spectrum corresponding to the more 

 conspicuous divisions for man. There is also some slight evi- 

 dence for a Purkinje effect but this needs verification. 



The proof of color vision in the fowl should make possible 

 the more ready correlation of the so-called physiological methods 

 of studying sensitivity with other behavior methods. The prob- 

 ability that the specific action-currents of the retina following 

 stimulation with different wave-lengths are an index to a more 

 general color sensitivity is increased. The lack of such specific 

 action-currents in night birds furnishes a final test of the inter- 

 pretation of retinal action currents. Experiments upon the color 

 vision of owls are in progress. The value of such methods as 

 those of Rouse and Babak is also enhanced by a verification of 

 some of their results by other methods. The reflex responses offer 

 an index of sensitivity more accurate and convenient than the 

 discrimination method, but they must be correlated with other 

 activities of the animal before they can furnish a basis for attack 

 upon more general problems. 



The probability of a high degree of sensitivity to wave-length 

 in other day birds suggests anew the problems of organic and 

 sexual selection in birds, problems which have largely fallen 

 into disrepute of late. The prominent role played by sensitivity 



