MUSEUM OF COMPARATIVE ZOOLOGY. 7 



At a later stage of development, when the upper lobe of the tail is re- 

 duced to a caudal filament, this gliding motion is accomplished, princi- 

 pally at least, by the action of the pectorals. When the fish is advancing, 

 these fins are directed obliquely backward ; but when, as often happens, 

 the motion becomes retrogressive, they are directed more nearly at right 

 angles to the body. The motion of the fins is so rapid, that I have been 

 unable to determine by observation if, as is probable, the direction of 

 the stroke is reversed in the two cases. Not only the direction of the 

 long axis of the fin, but also the inclination of its transverse axis to the 

 horizon, is conspicuously changed at such times. 



The vibrating movement of the caudal filament perhaps assists the 

 forward motion of the fish, but it cannot be considered essential to it, 

 since the filament often remains motionless while the animal is gliding 

 by means of the pectorals. The amplitude of the vibrations made by 

 the filament is not great in any case, — about 15°, — and the terminal 

 half alone is vibratory. When in motion the direction of its axis 

 is usually continuous with that of the spinal column, although it may 

 droop more or less while in motion, and is quite liable to do so when at 

 rest ; it then presents an even curvature, as seen from the side, and 

 often inclines a little either to the right or left. 



When the fish is stemming a current, or, in swimming, is directing 

 the head downward, the caudal filament is kept in rapid vibration ; it 

 then takes a dorsal turn, and the curvature is rather abrupt at its base. 

 The whole curvature may amount to enough to make the extreme tip 

 of the filament perpendicular to the axis of the body, but usually it is 

 much less. 



During the night of August 6-7, 1883, one of the individuals hatched 

 in June of that year escaped from the tank, and was found in the' morn- 

 ing lying in only sufficient water to keep the body moist. Upon being 

 returned to the tank, though still able to swim, it showed evident signs 

 of weakness. The body was considerably arched, just as it has been fig- 

 ured for somewhat younger fishes by A. Agassiz ('78 a , Plate IV. Fig. 39, 

 and Plate V.). I think this case suggests an explanation of the peculiar 

 curved shape exhibited by the fishes reared by Mr. Agassiz. I had 

 already in the previous year imagined that the arched condition was not 

 common, for all my fishes were quite straight, at least so long as they were 

 well nourished. The curved condition of the escaped fish was apparently 

 due to muscular weakness ; the curvature was also accompanied by a 

 slight distortion from the sagittal plane. Inasmuch as it subsequently 

 regained its normal condition and became straight, I have no doubt that 



