624 



THE EYE IN EVOLUTION 



Such a technique is, of covirse, full of pit-falls and would lead to the con- 

 clusion, for example, that the bull recognizes and dislikes red — which has 

 been proved untrue. In more recent years it has been superseded by the training 

 techniques. The first to apply these was Zolotnitzky (1901) who fed fishes on 

 red larvae and then, when they had been trained to respond to this stimulus, 

 offered them pieces of wool of different colours ; they continued to choose the 

 red, the presumption being that they ajapreciated it as such. Subsequently 

 more adequately controlled techniques have been employed involving the use 

 of T- or Y-maze experiments such as we have already described ^ ; their value 

 and their limitations should again be stressed. 



The COLOUR VISION of cyclostomes is entirely unexplored. 



Syngnathus 



THE COLOUR VISION OF FISHES 



The colour vision of Fishes has received much attention, but none 

 has been given to types other than Teleosteans. It is unlikely that the 

 Selachians have colour vision with their pure-rod retinae, ^ but as 

 Walls (1942) suggested, it is conceivable that among the Holosteans, 

 A^nia, with its duplex retina, may have been the first vertebrate type 

 to develop colour vision. However that may be, no fish has been 

 proved not to have colour vision, and those Teleosteans which have 

 been investigated certainly exhibit this faculty in a considerable degree 

 of development. 



We have already noted that Graber (1884—85) first showed that 

 the teleostean fish with which he experimented (both fresh-water, 

 Barbatula and Albtiryms, and marine, Spinachia and Syngnathus) 

 showed a preference for certain colours, swimming towards red in 

 preference to green and green in preference to blue, while Zolotnitzky 

 (1901) confirmed that fish could be trained to come to red. The 

 possibility of establishing a similar association of red with food despite 

 variations in brightness was established by Washburn and Bentley 

 (1906) in the dace, Semotilus, while Reighard (1908) found that the 

 snapper, Lutianus, despite confusional variations in brightness, 

 avoided red and preferred the shorter waves of the spectrum. This 

 suggestion that fish were able to discriminate hues excited a consider- 

 able amount of research and not a little controversy.^ On the one 

 hand, v. Hess (1909-22), applying the same methods of colour preference 

 and the observation of the degree of pupillary contraction to different 

 spectral bands, found the greatest response to the green region of the 

 spectrum while red light elicited a poor or negative reaction ; since 

 this was typical for scotopic vision or total colour-blindness in man, 

 he argued that fish were colour-blind, an argument fortified by his 

 contention that, if sufficiently intense illuminations were used, an equal 



1 p. .569. 



^ Except Myliobatis and Mustelus. 



^ For summary, see Warner, 1931. 



