82 PROFESSOR J. C. EWART ON THE 



on a level with the anterior part of the pectoral fin ; while in Acanthias, the canal is 

 closed except for a short distance in the caudal region. In as far as these species have 

 furrows, they agree with Chimsera, in which grooves take the place of canals in the head 

 as well as in the trunk, but differ from Callorhynchus, which has canals like the skates 

 and rays, and the majority of the sharks. The difference, however, between canals and 

 furrows is, in most cases, more apparent than real, as the sense organs which lie in the 

 furrows are protected by overlapping scales or folds of skin, which practically convert the 

 furrows into canals. 



In some sharks, the arrangement of the canals is simpler and the tubules less 

 numerous than in others. In Laemargus and Heptanchus, the arrangement of the canals 

 is very simple. Their length has not been greatly increased by the formation of loops, 

 and the tubules are short and never break up into branches before opening on the 

 surface. But in the extremely active thresher shark Alopias, there is a loop on the infra- 

 orbital under the eyeball ; and the mandibular part of the hyomandibular is well 

 developed and provided with long branching tubules. But Alopias is especially noticeable 

 for the remarkable development of the lateral canal and its tubules, which are not only 

 extremely abundant, but are of unusual length, and give off numerous branches. 



It is extremely probable that the remarkable development of the tubules in Alopias is 

 related to the active pelagic habits of the fish, which in this respect differs strikingly 

 from Lgemargus, which, from its being frequently taken in the beam trawl, seems to move 

 about in a leisurely manner, near the bottom. 



Further evidence of a relation between the development of the sensory canals and the 

 habits of the fish is especially found among the skates and rays (Batoidei). In the 

 common skate [Raia batis) the expansion of the pectoral fins has been accompanied with 

 a great extension of the hyomandibular canal ; and the appearance of long, dorsal offshoots 

 or branches from the lateral canal. 



That there is a relation between the extent of the sensory canals and the size of the 

 pectorals is placed beyond doubt by comparing the condition of the canals in the more 

 specialised Batoidei with such shark-like forms as Ehinobatis and Pristis. In these, 

 although the ventral loop so characteristic of the Batoidei has made its appearance, it is 

 very short and narrow, not reaching as far as the first branchial cleft. 



As fully explained in a paper in process of preparation, the hyomandibular in the 

 skate has extended backwards external to the branchial region to form a long, wide, 

 ventral loop, one end of which passes to the upper surface to form, with a branch from 

 the lateral, a long and still wider dorsal loop over the pectoral fin ; while a second branch 

 from the lateral extends over the fin behind the loop. 



By these extensions and folds an extremely sensitive apparatus has been arranged 

 over the surface of the pectoral fin. Though, in the meantime, it is impossible to 

 account for these complex arrangements of the sensory canals, it may be taken for granted 

 that they are far from meaningless. 



In the rays there seems to be, as in the sharks, a relation between the development 



