6 BULLETIN OF THE 



the water has not been too much disturbed, they make no motion while 

 sinking until they have nearly reached the bottom. Before they actu- 

 ally touch the bottom of the dish they appear to recognize their prox- 

 imity to it, and then begin to make vigorous efforts to swim up to the 

 surface again. This is apparently a very laborious undertaking, and, if 

 they fail to attach themselves at once, they again begin to sink slowly ; 

 they seldom attach themselves at the bottom, — especially if the water 

 has remained for some time unchanged, — but always as near the sur- 

 face of the water as possible. If there are too many to be accommodated 

 in a single row, those last to come crowd in between the tails of those 

 already attached, thus forming a second row ; but if there are still others, 

 they usually attach themselves to other fishes rather than take a lower 

 position on the sides of the dish. During the period of yolk absorption 

 they hang pendent and nearly motionless, except for the respiratory 

 movements ; those which hang from the surface of the water are verti- 

 cal, and any sudden motion in the water shows that their bodies are 

 quite limp. When the absorption of the yolk is well advanced, the 

 flexibility of the body is shown in a striking way by the snake-like 

 motions which the animal slowly executes while remaining attached. 



The disinclination to swim lasts about as long as the yolk-sac per- 

 sists. With the gradual disappearance of the latter, the fishes show an 

 increasing tendency to swim about. When at length the sac is nearly 

 absorbed they rest in quite another way. They float near the surface, 

 taking a horizontal position, and remaining perfectly straight and mo- 

 tionless until disturbed ; whereupon, by vigorous strokes of the tail, 

 they swim away with remarkable celerity. In transferring the fishes 

 from one dish to another I was accustomed to use a small spoon, but 

 after the absorption of the yolk-sac I found it exceedingly difficult to 

 capture them in that way, so rapid were their movements. The stage 

 at which the fishes begin to swim and float is reached in eight or ten 

 days after hatching. Even at this early age locomotion is accomplished 

 by two distinct methods. The rapid motions are executed by vigorous 

 strokes of the whole caudal region. A slower, gliding motion is main- 

 tained by means of the very rapid vibrations of the extreme end of the 

 tail, which are so characteristic of the caudal filament at a later stage, 

 and by the still more rapid motions of the pectoral fins. Not only do 

 the pectorals vibrate when the tip of the tail is motionless, and vice versa, 

 but either of the pectorals may be in rapid motion while the other is at 

 rest. This second method of locomotion is apparently very serviceable 

 to the fish, in allowing it to approach its prey unobserved. 



