THE CRANIAL NEKVES AND LATERAL SBN8E ORGANS OF FISHES. 193 



sense organ, but became an obvious necessity when the superfical organs, increasing in 

 number, size, and sensitiveness, were sinking below tlie surface. The later process, whilst 

 it afforded the desirable protection to the sense organs, necessarily removed them further 

 from the sphere of their activity and demanded the physiological compensation supplied by 

 tlie introduction of mucus into tlie canals. 



That the superficial type of sense organ preceded the central canal organ, or, in other 

 words, that tlie sensory canals arose jirimitively on the surface and then sank into their 

 usual subdernial position as definite canals, is to my mind amply proved both by fossil 

 Ichthyology and Emlnyology. The former shows us that in the most ancient fishes known, 

 such as the Ostracoderms, Artlirodira, and the older Sharks {Cladoselache, Cladodus), the 

 lateral canals existed as superficial grooves, which must therefore be held to be the most 

 primitive known condition. The development of the canals iu existing fishes is further 

 evidence in the same direction, and tells us that superficial grooves arise first by a 

 process of decentralisation, whilst the canals themselves are formed in segments after- 

 ■viards /i/ situ. These two considerations enable us to form a general idea as to the 

 lines ujion which the lateral canals were laid down. Turning to the comparative 

 anatomy of the system, we can not only confirm the view suggested by the above evidence, 

 but are enabled to follow the process iu some detail. Whilst I fully concur in Lsydig's 

 contention previously mentioned, we may for the purposes of convenience classify the 

 lateral sense organs under three heads, as follows : — 



/ (a) Pit organs sensu striclo, as iu Tcleosts. Most superficial. 



(1) Superficial ij/pc ■! (b) Saviun vesicles [Torpedo] and nerve-sacks (Ganoids). lutermeJiati! 



^ between 1 and 2. 



(2) Intermediate type. AmpuUffi of Lorcnzini (Elasniohi'anchs). 



(3) Central type. The canal sense organs (almost all fishes). 



Here we have a continuous series connecting the superficial sense organs on the one 

 hand with the canal organs on the otlier, and indicating the manner in which the canals 

 and their organs may have been produced. In nearly all fishes the lateral line system is 

 in a very unstable condition. In Elasmobranchs the pit organs have sunk below the 

 surface, and have been converted either into Savian vesicles or ampulkB of L u-enzini, 

 principally the lattei', with the result that pit organs sensu stricto are largely absent. 

 In Teleosts, where Lorenzmian ampullae are absent, we find, as we should expect, 

 the pit organs on the surface, and existing in large numbers, perhaps as com- 

 pensation for the more or less reduction of the canal organs themselves. In Ganoids, 

 where the canal system, as far as our limited information goes, still flourishes, 

 the pit organs are comparatively few and have in some cases Ijcen converted into what 

 Merkel calls nerve-sacks. Iu many cases we find undoubted evidence of a canal in one 

 form being represented by a line of pit organs in another, showing tliat the latter 

 represent the possibility of a canal. Thus there seems to be a constant migration of 

 sense organs from the surface to the interior — a constant stream of superficial organs 

 inwards forming canals and organs in some cases and reinforcing them in others. It is 

 easy to see, for example, from what we know of the development of the canal organs, how 

 a line of ampuU-ae might form a canal. The insinking would only have to form a furrow 



