ACCESSORY ORGANS OF HEARING 53 



base of the vesicle is a small blind rudimentary duct ; tliis has been 

 sho^^^l by a study of the development of Gymnarchus, one of the 

 jMormp-ida\ to be the vestigial remains of the original swim-bladder 

 coiuiection. 



We shall find when we describe the spherical air vesicles of the 

 herring that the posterior vesicle is surrounded by the semicircular 

 canals, just as we have described in Mormyrus ; this seems to point 

 to the probability that the posterior vesicle is functionally similar 

 to the air-vesicle of the latter fish. It may be asked why has the 

 swim-bladder connection been lost in Gymnarchus but remains in 

 the herring ; the explanation is that the herring has a wide range 

 of movement, at times swimming rapidly from the depths to the 

 surface ; if there was no connection, the gas in the vesicles would 

 expand to such an extent as to put the mechanism out of gear 

 when the fish comes to the surface ; moreover, there is a special 

 orifice in the swim bladder of the herring which allows of a free dis- 

 charge of gas, directly near the vent. We know little of the habits 

 of the IMormyi'idse, except that they are bottom -feeding fish, and 

 keep approximately at the same depth, so that there is no necessity 

 for any contrivance to release the pressure in the air-vesicle. 



The accessory organ of hearing in the herring, which we shall now 

 attempt to describe, has never been quite satisfactorily investigated. 

 In an endeavour to tlu-ow more light on this very complicated 

 mechanism, we have dissected a very large number of fresh specimens 

 and have been struck by the number of details that can be made by 

 simple anatomical methods. But the most important contribution 

 to the interpretation of its structiu*e has been made by the study of 

 a series of horizontal sections which were given us by Dr. Hillier, 

 who has made a detailed study of the bones of the cranium of this 

 fish. These sections are very beautiful and seem to solve many of 

 the difficulties that have arisen in the correct understanding of the 

 function of this organ. 



In order to simplify our description, it is proposed to describe 

 this complex mechanism under tliree headings, the ear proper, the 

 swim-bladder, and the accessory air-vesicles by which the first two 

 organs are connected. The two drawings of the ear proper (Fig. i) 

 are made from the dissections of a fresh specimen, and show the 

 usual type of internal ear that is found in fish, namely, the utricle 

 with its three semicircular canals, and the saccule with its 

 otolith. There are two other facts or relations that do not 

 appear in the drawings : the bony cavity of the saccule has a specially 

 modified external wall which is known as the auditory fenestra ; 

 this is closed by a membrane which has the characteristics of a 



