246 THE SENSE-ORGANS AND PERCEPTION OF FISHES. 



leave red foi' yellow, and yellow for green, preferring blue-green to 

 violet and to all the coloured glasses witli wliicli they were tested. 

 As these animals seek the light it may be supposed that the blue- 

 green seems to them the most intense. Whether, as happened in 

 the case of Sir John Lubbock's experiments with ants, any of these 

 animals perceive vividly the ultra-violet rays I cannot say, as I had 

 not the necessary appliances. 



As it has been suggested that the bright colours of animals may 

 have a protective value, which suggestion has recently been extended 

 by Garstang [Journ. M. B. A., N. S., i, 2, p. 175) to marine animals, 

 I endeavoured to ascertain in the case of mullet whether bright 

 colours have any such deterrent power. The mullet in question were 

 a shoal of small fry about one and a half inches long. They were 

 accustomed to eat minced worms off a slate slab. Upon the slate 

 slab I arranged a number of brightly coloured tiles, some having 

 plain and others mottled surfaces, and the minced worm was laid 

 on these. The tiles were dark red, white, pale blue, dark blue, 

 and mottled greenish brown. On several occasions the food was 

 first cleaned off the pale blue and the white tiles upon which it was 

 most conspicuous, and next off the mottled ones. The food on the 

 dark blue and dark red tiles generally remained the longest, but was 

 eventually eaten. On the whole it seemed to me that the fish dis- 

 tinguished between the tiles, but there was nothing to suggest that 

 they were afraid of any of them. Certainly the bright colour of the 

 pale blue tile did not seem to trouble them. It would perhaps be 

 worth while to make a similar experiment with some of the glass 

 models of anemones, &c., which are now obtainable, in order to test 

 whether the colours, per se, have any deterrent effect. 



It has been stated above that pollack will snatch at a wire curved 

 into the shape of a worm when it is held up outside the tank. The 

 same wire when painted white, or bright yellow, or blue, proved 

 equally attractive. 



Smell. — It was stated above that the great majority of fishes hunt 

 their food by sight, and there is a good deal of evidence that it is 

 sought for by sight alone. None of the fishes mentioned on pp. 241-2 

 show symptoms of interest when the juice of food-substances is put 

 into the water. They will attempt to take worms, shrimps, pieces 

 of fish, &c., which are lowered into the water inside a glass tube, or 

 which are simply sticking to the glass window of the tank. When 

 hungry they are unable to find food in the dark, while by day they 

 will seize uneatable substances which are quickly moving in the 

 water. This evidence goes to show that the sense of smell plays 

 little or no part in helping them to discover their food. On the 

 other hand, both pollack and whiting, when their first hunger is satis- 



