THE AMERICAN NATURALIST. [Vol. XXXVIII. 



like knife. The chief danger in this operation is in cutting too 

 deep, in which case excessive bleeding may follow. When this 

 was avoided, the fishes usually recovered, and the success of the 

 operation could be judged by their subsequent movements. 

 When they first recovered from the ether they seemed to have 

 lost all power of equilibration, swimming now one side up, and 

 now the other, or resting with their long axes vertical. After 

 about a day, however, they usually acquired and kept their 

 normal position, at least while resting or swimming slowly ; and 

 this ability increases, until after two or three weeks they were, 

 in all their ordinary movements, indistinguishable from normal 

 fishes. If, however, such a fish be placed in a large body of 

 open water and made to swim rapidly, it soon loses all power of 

 orientation and darts about, turning over and over until 

 exhausted. This condition, so far as I know, is permanent, 

 for in the case of one fish which lived for over three months 

 after the operation these reactions showed no tendency to dis- 

 appear, but persisted till death. The partial recovery of equili- 

 brium noticeable soon after the operation is probably due to a 

 successful attempt on the part of the fish to retain its normal 

 position through sight. 



Earless fishes are usually more quiet than normal ones, and 

 hence they can very easily be tested. In all, I made 73 tests 

 on 7 fishes, and in no instance did I get an undoubted response 

 to sound. This is in strong contrast with the reactions of the 

 same fishes before their eighth nerves had been cut, and points 

 beyond question to the ear as an organ of hearing. 



I supplemented the foregoing experiments by another series 

 in which two normal fishes that I found to respond well to 

 sound, were etherized, and the eighth nerve of each cut on the 

 right side. After an interval of twenty-four hours, they were 

 both tested again, and found to respond about as well as they 

 did before the operation. In 20 tests on the two, there were 

 19 responses. They were then etherized again, and the eighth 



nerves of the left sides 



tested 



were cut. After recovery, they were 



more, and, although the experiments were 



ducted with the greatest possible care, not a single response 

 was obserx^ed in 20 trials. This experiment shows that the 



