10 OLD RED SANDSTOxNE FISHES. 



so that under a low power the section appears to be cross-hatched by a series of dark 

 hnes. " In flakes of the disc similarly treated, but containing more of the middle and 

 outer layers, it is obvious that the great canals divide into the branches of the middle 

 layer, which have already been seen in the vertical section, chiefly, if not only, along lines 

 corresponding with the apparent sutures of the so-called 'polygonal scales.' The canals 

 of the middle layer are very singularly arranged, passing from their origin, across these 

 sutm"al lines and nearly parallel with one another, towards the centre of the adjacent 

 ' scales.' "^ This, then, is the cause of the appearance of a division of the external surface 

 into distinct scales ; the minute canals, with their reddish infiltration of matrix, 

 showing very distinctly against the lighter general substance, or even allowing the 

 shield to crack along these lines. 



It is worth noting, in addition to this, that the more superficial parts of the hard 

 calcareous matter in each of these apparently polygonal scales has a radiated fibrous 

 arrangement, the fibres diverging from the centre. This fibrous structure is visible in 

 many specimens of Cephalaspis from the Cornstones of Herefordshire. 



((5.) We now turn to the other members of the family Cephalaspida, typified by 

 Agassiz's C. rostratus, C. Lloydii, and C. Lewisii, embraced by Kner's genus Pteraspis, and 

 forming my Section Heterostraci. 



A naked-eye examination of suitable specimens of any of these shields shows the 

 exterior surface to have a very peculiar ornamentation : instead of enamel-like tubercles, 

 minute ridges and grooves, running parallel to and concentrically with one another, 

 everywhere mark the convex superficies. These ridges are exceedingly small in some 

 species, and vary in different forms from j^tli inch or larger to ^^th of an inch in breadth: 

 they are usually crenated at the margins, and give the notion, in some species, of a linear 

 series of minute tubercles fused together. They may not unfittingly be compared 

 to the markings of the epidermis on the palms of the hand and fingers in Man 

 and other Primates. The inner surface of the shield is seen to be quite smooth and 

 polished, and exhibits no vascular channels, as in Cephalaspis Lyellii, but a few 

 irregular ridges of some size, diverging from the centre of the scute or running parallel to 

 the margins, having the aspect of lines of growth. Scattered rounded apertures are to be 

 detected on this smooth inner surface, not above ^i^th of an inch in diameter. If a 

 vertical section is now made through the shield, or if it be judiciously broken, the 

 structure of the interior becomes obvious. Firstly, the thickness of the ' shell ' in 

 most places is greater than in specimens of Cephalaspis of the same size, being frequently 

 more than ^'^th of an inch, and in terminal portions (such as the margins, spine, or rostrum) 

 much thicker. The most striking feature, however, which comes to view is the separation of 

 the material of the inner and outer surface of the shield by a stratum of polygonal cavities, 

 sometimes filled up by the infiltration of carbonate of lime, and having the appearance of 



1 Op. cit., p, 2/3. 



