38 POPULAR SCIENCE MONTHLY. 



to be due to defective eyesight, for even though a moving worm were 

 put in its immediate neighborhood, the fish did not appear to detect 

 its presence through the sense of sight. At first, stimulated probably 

 through its lateral line system, the fish seemed to feel the movements 

 of the earthworm; it would then turn in the direction of the food, 

 move towards it with apparently increasing enthusiasm, but when only 

 at close range did it seem actually to see the prey. The fish's move- 

 ments in feeding reminded one rather of a turtle than of a fish, or, 

 best of all, of its kindred salamanders. Eyeing the moving worm 

 steadily, it would make a sharp snap at it. If this movement failed, 

 it would appear to deliberate, gaze fixedly at the object, and snap again. 

 If more successful this time, it would pause with the food in its mouth, 

 then with a series of accelerating snaps, the entire worm would be 

 ingested. Occasionally a worm would be cut entirely in two by the 

 quick snap of the fish's powerful jaws, and this would result in the 



Fig. 7 a. 



Fig. 7 6. 



Fig. 7 c. 



loss of the worm and in the feeding beginning anew. During this 

 entire process the fish's arms would be spread widely apart, so as to 

 support the weight of the head. 



In later years the fish became quite tame, and would feed out of 

 the hand of the laboratory attendant, who always maintained that the 

 fish distinguished him from other visitors. Certain it was that he 

 finally accustomed the fish to a diet of raw meat, and this substitute 

 for earthworms proved a convenient one during the cold season. A 

 finger thrust into the aquarium and stirred vigorously would be enough 

 to attract the fish's rather sluggish attention: it would slowly leave 

 its resting place, 'walk' toward the region of the disturbance, rise to 

 the surface and after giving the usual evidence of bad eyesight would 

 finally get its mouthful of food. The fecal material of the fish, one 

 might mention, showed the cast of the spiral intestinal valve which in 

 lung-fishes is almost as well developed as in sharks. Possibly, there- 



