No. 448.] CARASSIUS AURATUS L. 



rod like, lying for the most part in the saccular portion, the 

 other lenticular in outline, and lodged in the lagena projicr. 

 The wall of this sac is supplied with branches from ilu- eighth 

 nerve and is so surrounded by bone that all attempts to icmcw 

 it by pulling out the semicircular canals were complete laihuvs. 

 In the four fishes on which I operated by Kreidl's method, 

 subsequent dissection showed these structures intact. Thus, liis 

 operation leaves uninjured a large part of the internal ear, m 

 fact, just that part which, from comparison with the ears ot 

 higher vertebrates, would be expected to be concenKnl with 

 hearing. I believe, therefore, that Kreidl's method ot oi)eratin- 

 is defective, and the reason that the fishes upon which he liad 

 operated responded to sounds much as normal ones did, was not 

 because in both cases the skin was stimulated, as he believed, 

 but because his so-called "earless" fishes still retained intact 

 a part of the ear which, as I have already shown, acts as an 

 organ of hearing. That it is such an organ follows from the 

 fact that when its nerve connections are cut, the responses t( 



Normal goldfishes usually respond in a definit< 

 sound-vibrations in water. 



Goldfishes in which most of the skin has bee 

 insensitive by cutting the nerves, and specimens 

 the ears, except the saccular portic 

 still respond in an essenti 



Goldfishes in which the eighth nerves have been cut^^ 

 sides, thus eliminating the sacculi and lagena- as uej 

 rest of the ear, seldom or never respond to soun( \i 

 in water. 



Goldfishes possess the sense of hearing, and the p. 

 the ear concerned with this sense is the sac which i 

 represents the sacculus and the lagena of highe 



