CEPHALASPID^. 9 



none that I measured exceeded ^o'^gth of an inch in this direction. The form of the 

 lacunae is very irregular, in consequence of the long branching and anastomosing 

 canaliculi which are given off, not only from their ends, but from their sides. In some 

 parts the innermost layer appears almost black when viewed by transmitted light, in 

 consequence of the quantity of air retained in the multitudinous lacunae and canaliculi. 

 Large vascular canals, measuring from ^i^th to ^th of an inch in diameter, whose inner 

 openings correspond with brown spots distinctly visible on the inner surface of the 

 shield, traverse the innermost layer very obliquely in their course towards the middle 

 layer. Their branches are few, and for the most part run parallel with the main trunk ; 

 but they give off a great number of minute canaliculi which anastomose with those of 

 the nearest lacunae. Such of these canals as I have seen in section were oval, their long 

 diameters being parallel with the planes of the lamellae The middle layer is distin- 

 guished from the inner by the rarity or entire absence of the lacunae, and by the indis- 

 tinctness of the lamination as compared with that of the deep layer. Such striations of the 

 nearly homogeneous base as seem to indicate lamination are, in the middle and inner parts 

 of the middle layer, so disposed as to be nearly perpendicular to those of the deep layer, 

 appearing to follow the course of the vascular canals. The latter are continuous with 

 the large vascular canals of the deep layer, but they are smaller, and form a close 

 network. Each of the large canals, on reaching the middle layer, gives off several 

 branches, which run nearly parallel with the surface (and therefore greatly inclined to the 

 course of the great canals), and anastomose with those around, above, and below them. 

 In this particular part of the disc, in fact, a large canal gives off as many as three tiers of 

 these lateral branches, separated from one another by not much more than their own 

 diameter, and all ramifying and anastomosing with one another. These lateral vascular 

 canals have at first a diameter of about yi^th of an inch ; but many of their anastomotic 

 branches are much smaller. Sooner or later all these branches appear to end in a close 

 ' superficial network,' vi'hich lies in the boundary between the middle and the superficial 

 layers. The latter or third layer of the disc sometimes appears structureless, at others 

 presents an obscure vertical striation, as if it were, like enamel, made up of minute fibres. 

 The superficial vascular network sends into it a great number of minute short processes, 

 which branch out abruptly at their ends, like a thorn-bush or standard rose-tree, and end 

 in excessively fine tubuli, like those of dentine.^ .... This substance, it will be observed, 

 corresponds very closely in structure with the ' cosmine ' of Professor Williamson. I have 

 been unable to find any trace of a 'ganoin ' layer external to it." 



A structure in every essential respect similar to this is found in all the other regions 

 of the cephalic shield, the three layers varying somewhat in relative thickness. When 

 flakes of the inner layer are detached and examined, the axes of the lacunae of each lamina 

 are seen to be directed nearly at right angles to those of the laminae above and below ; 



' In tliose species in which the tubercles are well marked, it is seen that a " bush " of fine tubuli 

 forms the centre of each tubercle. 



2 



