512 ADAPTATIONS TO PHOTIC QUALITY 



spot a point of matched brightnesses. Brown's technique is promising, but 

 in his hands it has yielded no evidence of color vision in the rabbit. 



Another piece of work which would bear careful repetition is that of 

 Sgonina, in 1936, on the guinea-pig. This animal's retina is even more 

 certainly a pure-rod one than that of the rabbit; yet Sgonina claims that 

 it is able to discriminate between colored papers whose difference in 

 brightness is less than that which must exist between two gray papers, if 

 the guinea-pig is to discriminate the latter. He found that two grays 

 could be told apart only when one was about one-third brighter than the 

 other. The validity of his conclusion obviously hinges upon the correct- 

 ness of the assumption that a guinea-pig sees the brightnesses of colored 

 papers as Sgonina himself did — and we have seen, ad nauseam, that such 

 an assumption must never be made. 



Salzle, also in 1936, studied two species of wild mice. He found that 

 despite its excellent learning capacity, the European long-tailed field 

 mouse (Apodemus sylvaticus) was hopelessly confused when a red light 

 was offered it alongside a yellow or green to which the animal had been 

 trained to go. The red-backed mouse (Clethrionomys glareolus) told a 

 different — and unique — story : 



Animals trained to filtered red light readily learned to distinguish it 

 from green, blue, and yellow. Animals trained to green quickly learned 

 to discriminate it from red; but when offered green versus yellow their 

 discrimination was poor and, though it improved rapidly, never became 

 perfect. Offered green versus blue, they failed completely. Salzle then 

 trained two animals to each of the four colors, and offered each of the 

 animals all four colors at once, their positions being changed from trial 

 to trial to avoid position habits. The result was that the animals trained 

 to red or to yellow went mostly to red or to yellow as the case was, but 

 the animals trained to green went about equally to green and blue, and 

 the animals trained to blue went equal numbers of times to blue and to 

 green. Salzle was sure that the apparent equivalence of the green and 

 blue stimuli was not due to their being matched in brightness for the 

 animal — but his evidence for this was that the green light was much 

 brighter than the blue one for his own eye. We would expect a human 

 green-blue brightness-match to be no match to achromatic rodents; for 

 their scotopic and photopic brightness curves appear to rise from zero in 

 the red to a maximum in the blue-green or blue. A green which matched 

 a blue for them would look brighter than the blue to a human. 



