550 EEPOET— 1893. 



also recognise tlie smell of food, apart from seeing it, was demonstrated 

 in various ways. First, if instead of a closed tube, as in the last-men- 

 tioned experiment, one open at the bottom was used, after a short interval 

 the nosing at the part where the worm was seen ceased, and the lower 

 end of the tube, from which, doubtless, worm-juice was diffusing, was 

 vigorously nosed. If, again, instead of putting worms into a tube I placed 

 a number of them in a closed wooden box with minute apertures to let 

 water pass in and out, there was a similar excitement produced, and the 

 dabs hunted eagerly in eveiy direction. When water in which many 

 worms had lain for some time was simply poured into the tank through 

 a tube that had been in position for several days, and by a person who 

 was out of sight of the dabs, the results were most marked. In a few 

 seconds hunting began, and in their excitement the dabs frequently leapt 

 out of the water, apparently at air-bubbles, ard on one occasion one 

 even cleared the side of the tank, which was about two inches above the 

 water, and fell on to the floor of the aquarium. Yet there was nothing 

 visible to stimulate this quest. 



II. Smell- feeders. — In the case of the smell-feeders I was also led to 

 doubt the exclusive dedication of the one sense to the task of food- 

 finding. Congers {Conger vulgaris) certainly do all that Bateson's paper 

 says of them : they hunt by night ; they do not go at once or direct to 

 food that is near them, but after it has been in their neighbourhood for a 

 short time grope around for it, till they gradually approach so near that 

 they touch it ; and then they snap at it greedily. They cannot find food 

 that has been washed in ordinary salt-water so as to remove the smell ; 

 they hunt when good-smelling fluids are poured into their tank ; and they 

 swallow corks or even stones that are suitably flavoured. But I came to 

 the conclusion that the congers in Plymouth Aquarium — and I have no 

 reason for supposing them to be different from other congers — were 

 'practically hlind. I have seen two when hunting come into direct colli- 

 sion with such violence as to produce a loud thud ; and once I saw one 

 start back in alarm on coming in contact with a crab that was in its 

 way. I have seen one lose its prey, and then hunt for it again in the same 

 indirect way as at first, though it certainly had discovered that the fish 

 was a thing good to eat. Often, too, a conger when close to its food 

 makes a snap in the wrong direction : it is when it touches that it snaps 

 successfully. So the sense of sight cannot be very good ! There is, 

 however, at least a perception of the difference between light and 

 darkness ; and a conger will retreat from a lighted match even by day, 

 and at night will seek shelter when a lantern is exposed. 



Unfortunately, though there were plenty of congers in the tanks 

 while I was at Plymouth, other smell-feeders were not well represented. 

 The rockling and sole, however, were available, and I repeatedly made 

 experiments with them. The rockling (Motella tricirratd) seems to be a. 

 true smell- feeder : it ' hunts ' by night ; it becomes highly excited when 

 an extract of any of its favourite foods is poured into the water in which 

 it is. Unlike the conger, however, it sees very well, as Bateson himself 

 pointed out. And, more than that, it can find its food by sight, though for 

 some reason — perhaps because of the timidity of the fish — this is difficult 

 to observe. Several times I saw appearances that almost demonstrated 

 the fact, and on one occasion the behaviour of a rockling left no room 

 for further doubt. For ten days it had had almost no food, and when, at 

 the end of that time, worms were thrown into its tank, it quite plainly 



