12 BULLETIN OF THE 



the end of July, 1883, I found an individual in the process of swallow- 

 ing a somewhat smaller mate. The bill of the victim and part of its 

 head were still protruding from the distended jaws of the captor ; so 

 that this individual was swallowed tail first, contrary to the more usual 

 method. 



The movements of the gill covers vary in rapidity at different times, 

 but they are always executed with considerable promptness. Their 

 adduction is quickly followed by their abduction, but the interval that 

 follows before another adduction is usually rather prolonged ; it is the 

 more variable element. It may be so short as to make the abduction 

 and adduction separated by equal intervals, or it may be prolonged to 

 several seconds. These respiratory movements seldom exceed sixty per 

 minute, and may diminish during the torpor of winter to scarcely more 

 than one a minute. 



The emission of bubbles of gas, which begins soon after the young 

 fishes detach themselves from their fixed supports, at first takes place 

 through the gills of one side. It is usually preceded by a forward 

 lurch of the body, accompanied by a slight rolling to one side and the 

 elevation of the gill covers. The bubble usually emerges before the fish 

 regains its normal position, and consequently comes through the gill 

 slits of the side which happens to be uppermost. Occasionally two 

 smaller bubbles escape from beneath the gill covei-s, one from each side. 

 When the fishes have become much older, the amount of gas is so great 

 that the bubbles often escape not only from beneath both gill covers, 

 but also through the mouth opening, and the rolling of the body does 

 not always accompany the escape of gas. 



In the earlier stages of their growth the gar-pikes remain most of the 

 time very near the surface of the water in a horizontal position. In 

 such cases the only premonitory symptoms of the escape of gas are the 

 motions just described ; but as they grow older they gradually habituate 

 themselves to lying in deeper water, and then they almost invariably 

 ascend to the surface before emitting gas. The ascent is nearly always 

 accomplished by a slow forward and obliquely upward motion, the body 

 being at an ancfle of about 45° with the horizon. The motion is usuallv 

 deliberate, and at a uniform rate. After reaching the surface the body 

 is allowed to assume the horizontal position before any effort is made to 

 expel the gas. This motion of ascent is so characteristic, that after 

 studying their habits one may predict with tolerable certainty whether 

 a given fish is about to emit gas. 



I believe the slight rolling of the body to one side is for the purpose 



