ON CEPHALASPIS AND PTERASPIS SI3 



centre of the basal half of the elevation and then suddenly ends in a 

 number of excessively minute branches, which pass towards the surface, 

 ramifying as they go, and closely resembling the canals of dentine or 

 cosmine. They appear to terminate on the surface, on which I have 

 been unable to discover any trace of laminated structureless ganoin. 

 The central canals of the elevations open internally into the network 

 of vascular canals which lies in the reticular layer. These canals rarely 

 exceed T-J-o-th to -e-J-Trth of an inch in diameter, and they are rendered 

 particularly obvious by the dark red granules with which their walls 

 are dotted. 



Internally they open directly into the interspaces of the septa which 

 connect the reticular with the inner layer, and the granules are con- 

 tinued on to the walls of the septa, which are themselves occasionally 

 traversed by short canals. The interspaces (<?) are full of a more or 

 less transparent inorganic matter, identical with that of the matrix. 

 It follows, therefore, that the " bony prisms " or " granules which 

 have been described have no existence, these so-called prisms being 

 nothing but the matrix which has filled up the cavities of the poly- 

 gonal cells, visible in their natural empty condition in Pt. Batiksii. 

 Canals resembling those of the reticular layer, as I have said, traverse 

 some of the septa and put their chambers in communication. 



In the section under description, the inner layer is for the most 

 part devoid of canals ; but one (/) is exhibited very beautifully. It 

 has in the middle a diameter of about -g-io-th of an inch, but is wider 

 at both ends, and traverses the inner layer almost perpendicularly. 

 The laminae are bent outwards for a certain distance, where they 

 impinge upon its walls. 



The structure just described is that of the central part of the 

 section. At one of its ends, near the margin of the disk, the arrange- 

 ment of the vascular channels is more like that in Cephalaspis, — the 

 reticular layer assuming a much greater development, and the areolar 

 character of the sinuses of the second layer becoming greatly obscured. 



On comparing together the appearance of a section with those 

 presented by the internal and external views of Pteraspis, there can 

 be no doubt that the elevations of the outer layer of the one are the 

 sections of the ridges of the other ; and it is remarkable that there 

 should be so striking a difference in the form of these ridges in Pf. 

 Banksii and Pt. Lloydii. The ridges seen in concave casts probably 

 always correspond with the whole interspaces between the ridges of 

 the outer layer in Pt. Banksii; but it is quite conceivable that in 

 Pt. Lloydii the ridges, in consequence of their peculiar form, might 

 sometimes be held by the matrix and sometimes not ; so that at one 



L L 



