COLOR VISION IN MAMMALS 513 



The two green-trained animals were now investigated further, with one 

 green and three blue stimuli offered simultaneously in varied positions. 

 Assuming that the animal could not discriminate the stimuli, it should 

 have gone 25% of the time to the green stimulus and 75% to the various 

 identical blue ones. One of the mice went to green ten times and to blues 

 38 times (21%-79%). The other went to green 16 times and to blue 32 

 (33.3%-66.6%). Salzle felt confirmed in his judgment that for this 

 animal green and blue are qualitatively identical. He attempted, quite 

 unsuccessfully, to fit this into the framework of either the Hering or the 

 Young-Helmholtz theory of color vision. But the strong probability is 

 that the animal is achromatic and that the particular green and blue were 

 a match for it in brightness — Salzle is most vague concerning his alter- 

 ations of intensity, and his text gives no assurance that this factor was 

 controlled. 



Lastly, for nocturnal rodents, may be mentioned Sackett's (1913) 

 negative results with the porcupine. The absence of color vision in 

 such rodents, all of which have few cones (or even none), is no sur- 

 prise. But in the diurnal squirrels, whose retinae contain no visible amount 

 of rhodopsin and appear to contain only cones (this being certain in the 

 case of the ground-squirrel and prairie-dog), color vision would be ex- 

 pected — indeed, a color vision about as rich as that of our own foveal 

 region, though of course affected in the short-wave realm by the presence 

 of a yellow filter, the lens (see p. 199). In the light of this expectation, 

 the results of experiments on squirrels are most interesting : 



Colvin and Burford, in 1909, were able to train a native squirrel posi- 

 tive to a pigmentary red and negative to either another color or a gray, 

 all the stimuli having the same brightness to man. They drew the un- 

 warranted conclusion that the squirrel, like their dogs, discriminated the 

 hues as such. Salzle also worked on one specimen of the European 

 squirrel iSciurus vulgaris) , training it first positive to green, then to red, 

 and offering three negative color-stimuli with each. The animal had no 

 trouble in making all discriminations. Salzle states that with each pair of 

 lights, one or the other could be made brighter or darker, or the two 

 about the same brightness (for his own eye) without it making any differ- 

 ence to the animal's ability to tell them apart. But no attempt was made 

 to match their brightnesses for the animal, and no details are given as to 

 just how intensity was varied. 



In contrast to these imperfect and inconclusive studies we have the 

 extremely careful work of Charlotte Locher which, though offered as a 



