480 ON THE CANAL-SYSTEM OF PTERASFIDIAN FISHES. [May 17, 



of an extensive canal-system, which indicates a more highly 

 specialized development of the " lateral-line " structures than has 

 hitherto been suspected. 



As shown by the drawing (p. 479), there are four longitudinal canals, 

 two marginal, and two situated close together in the median area of 

 the plate ; and the latter pair diverge in front, perhaps meeting the 

 former at the edge, while more posteriorly they are all connected at 

 irregular intervals by similar transversely-directed passages. The 

 median commissural branches extend directly across the space 

 between the two longitudinal canals they unite ; but those proceed- 

 ing to the lateral canals take a slightly more devious course, being 

 inclined either backwards or forwards. Throughout their extent 

 the tubular excavations give rise to short diverticula, alternately left 

 and right, which place them in connection with the external pores ; 

 and these are most numerous in the median portion of the shield 

 and the great marginal trunks, being relatively further apart in the 

 lateral commissures. 



The complexity and widely-spread character of the system is still 

 more satisfactorily demonstrated in the almost perfect specimen of 

 Holaspis already referred to ; and here, it will be observed, there is 

 an essentially similar arrangement. Our fossil corresponds to the 

 central portion of this shield, which appears to consist of the homo- 

 logues of the seven plates of Pteraspis wholly fused together. 

 And it is perhaps worthy of note that neither here nor in any other 

 Pteraspidian have distinct traces of the pores been detected upon 

 the rostral region in advance of the " orbital " notches or apertures. 

 Finally, it is interesting to institute a comparison between tliese 

 ancient traces of a " lateral-line " system and the various structures 

 adapted for the protection of the corresponding sense-organs in 

 existing fishes. Chimseroids possess merely the primitive open 

 groove ; but in Selachians the canal is complete, and there are also 

 present the short secondary diverticula leading to the external pores. 

 The last-named branches, however, are all directed to one side 

 (outwards or downwards) of the main canal in the Kays and upon 

 the trunk of the Sharks ; and there is no very close agreement with 

 Pteraspis even in the cephalic region of the latter group, where the 

 branches are given off in both directions, but are not completely 

 closed, being peiforated by a series of small orifices in addition to 

 the terminal one '. Moreover, it is scarcely likely that these canals 

 in the old Devonian fish had the Selachian mode of development. 

 In bony fishes, where the structures bear a similar relation to the 

 hard skeletal parts, in most cases excavating them, there is naturally 

 a much greater resemblance ; and some of the most specialized 

 Teleostei (e. y. the Pleuronectidse) exhibit an almost precisely 

 corresponding "feather-barb" arrangement ^ Unfortunately, how- 

 ever, the character of the sense-organs themselves necessarily 

 remains unknown, for the palaeontologist can rarely give much sure 



' P. C. Sappey, ' Etudes sur I'Apparoil Mucipare et sur le Systemc Lj lu- 

 pbatique des Poissoiis,' 1880, p. 55, pi. ii. fig. 2, pi. x. fig. I. 

 ■^ P. C. Sappe\ , op. cit. p. 44, pi. xi. figs. 3, 4. 



