MUSEUM OF COMPARATIVE ZOOLOGY. 9 



the head the larva is snapped up. Occasionally it is bitten through, 

 and, whether struggling or not, is allowed to escape ; but the reason of 

 this I did not discover. Usually the larva is not snapped at until by 

 some slight movement it shows signs of being alive, and then only a 

 single snap is made. If unsuccessful, the fish remains quite motionless, 

 or glides slowly away by means of the same motion of the pectoral fins. 

 Fishes, especially when less than an inch in length and when well fed, 

 rarely make a lunge forward toward their prey when they snap at it ; 

 the motion is usually only a sudden sidewise bend in the neck region. 

 So cautious are the fishes not to snap at dead larvae, that, after they 

 have advanced so that the insect is opposite the middle of the bill, if it 

 does not move, they begin to swim obliquely forward, pushing the larva 

 before them without allowing it to glide backward along the bill. In 

 this way a larva may be pushed about several inches before it makes 

 the fatal movement which betrays its condition ; or it may, if it remain 

 entirely passive, escape altogether, since the fish, failing to discover evi- 

 dences of life, leaves it and glides off in search of other food. 



The number of insects caught by a single gar-pike is undoubtedly 

 large. I was accustomed to feed the fishes twice a day. In July, 1883, 

 I had the curiosity to watch the largest of them, then a month old and 

 about 2\ inches (62 mm.) long, during its feeding. In the course of 

 ten minutes it caught twenty-one large mosquito larvae and made nine 

 ineffectual attempts at seizure. But the voracity of young fishes was 

 still more forcibly exhibited when, early in August, they were given 

 small minnows (Fundulus) for food ; for then they did not hesitate to 

 catch the minnows even when the latter were considerably larger than 

 themselves; but they succeeded in swallowing only those which were 

 of their own diameter or smaller. Since the minnows have much 

 thicker heads than the gar-pikes, the total weight of the former was 

 doubtless always somewhat less than that of the latter. 



The minnows were caught in the same way as the mosquito larva), 

 by a sudden sidewise motion of the head ; but being too large to be 

 swallowed at a single gulp, they were at first impaled on the sharp 

 teeth, and then by a series of deliberate movements put in a position 

 to be swallowed. Almost without exception, the minnows were swal- 

 lowed head first, irrespective of the region of the body first seized. If 

 caught near its tail, as usually happens, then the prey is moved be- 

 tween the jaws — to w 7 hich it lies crosswise — by successive shiftings, 

 until it is held near the base of its head. A few movements usually 

 suffice to make it take a direction parallel to the jaws, and the head 



