BEHAVIOUR OF A CAPTIVE ROCKLING. 193 



then back again. It will be observed that the place it had taken 

 is on a part of the old route from H to J. There are places 

 amongst the rocks covering the greater part of the back of the 

 tank in which the Eockling would apparently have been safer 

 than, and not so conspicuous as, it was in the position which it 

 had taken up, though there is no hiding-place which seems so 

 suitable for it as the one it had at first. It still pursued parts 

 of its old courses, say from C to E, or G to J, but I do not think 

 it did so with quite its former confidence. I observed no sign 

 that the Eockling was making fresh courses to M of a kind likely 

 to mislead possible enemies as to the real position of M, 



The loss of its first hiding-place had one curious, and some- 

 what ludicrous, result. When it had seized beef, which I had 

 induced it to follow, the Eockling, for the first few days, would 

 sometimes dart back along one of its old courses to its former 

 resting-place, and then instantly rush out again in terror when 

 its snout touched the Goby in possession. It appears that the 

 lesson learned so thoroughly in sixteen weeks required some few 

 days to unlearn. 



On the morning of April 17th I became aware that the Eock- 

 ling had formed a new habit. When the food of the Gobies, 

 which consists of fragments of raw beef, was dropped into the 

 aquarium, the Gobies began excitedly to dart out of, and into, 

 their hiding-places, and at first the Eockling shrunk timidly 

 against the wall when a Goby dashed past it. The Eockling was 

 fed last, because unsatisfied Gobies were apt to snatch the food 

 from its mouth. The sight of the Eockling is very poor, and 

 food was therefore placed close to its head. On the 17th I noticed 

 that when the Gobies began to scurry about the aquarium the 

 Eockling showed unmistakable signs of interest, and after a 

 moment or two began to swim as though seeking food. Its 

 behaviour at feeding time (then and since) suggested that in the 

 twenty-five days during which the Gobies had shared its aquarium 

 the Eockling had come gradually to associate unwonted activity 

 of the Gobies with the advent of food. By May 1st the Eockling 

 had become so used to the Gobies that it had begun to lie, along 

 with one or two Gobies, in or close to its old resting-place (but 

 with its tail directed often to the right), and to resume its old 

 courses, though not with its former precision. 



Zool. 4th ser. vol. XIX., May, 1915. q 



