COLOR VISION IN AMPHIBIANS 491 



The presence of a Purkinje phenomenon, however, is no evidence for 

 color vision, but only demonstrates the presence of two types of receptors 

 with different absorption spectra. A Purkinje phenomenon of the iso- 

 chromatic type could exist in an animal with a duplex retina and achrom- 

 atic vision. Hess even denied the existence of the phenomenon in the 

 frog, on the basis of pupilloscopic findings, claiming the frog pupil to 

 be most responsive to green light both scotopically and photopically. 

 We have seen how devoid of any certain meaning such findings are, 

 particularly when gained in an animal whose iris muscles are wholly 

 or largely autonomous. Pupillometry is scarcely more trustworthy as a 

 means of determining accurately the limits of the spectrum; but it is 

 a convenient means, and with it Hess determined the spectral limits of 

 the frog and other amphibians to be practically the same as those of man. 



Nearly forty years ago, Yerkes studied the learning ability of the 

 green frog (Rana clamitans) in very simple mazes, employing red and 

 white cards as parts of the stimulus patterns offered the animal as cues 

 to true path and blind alley. The frogs were guided partly by these 

 grossly different visual stimuli; but neither Yerkes nor anyone else has 

 since gone further than this in attempts to train frogs to discriminate 

 hues. Their learning ability, which is next door to zero, makes this quite 

 out of the question. Hess and others got nowhere with the color-pref- 

 erence method in frogs; and even the conditioned-reflex technique, 

 which obviates any need of a conscious choice by the animal, gave no 

 results when Bajandurow and Pegel tried to apply it to the frog in 1932. 



Promising leads have come lately from the electroretinograms picked 

 up from the excised eye under monochromatic stimulations. The Helsinki 

 group found that the form of the gram is different for colored stimuli, 

 when differences in intensity are ruled out. They have decided that there 

 must be three systems in the frog eye : 



A. The rods, with their rhodopsin. 



B. Rods, or cones, containing a substance absorbing light maximally 

 in the blue and violet ('green rods' [p. 58]? cone oil-droplets?) . 



C. Cones of at least two types (the singles and doubles?) overlapping 

 with respect to the distribution of their sensitivities to spectral lights. 



They conclude that "the selective effect of wavelength on the retina 

 represents a mechanism that can be used for color differentiation." But 

 when they plotted the spectral distribution of the effects of strong mono- 

 chromatic lights upon the subsequent electrical response to stimulations 



