ON CEPHALASPIS AND PTERASPIS 507 



fibres" of the one correspond with the grooves and furrows of the 

 other. The surface of the cast is remarkably darker than the sur- 

 rounding matrix, and might not unreasonably at first be supposed 

 to be of a different nature. When the inner surface of the disk is 

 carefully examined with a magnifying glass, a number of reddish- 

 brown minute dots appear scattered irregularly over its surface. It 

 will be seen immediately that these are the internal openings of vas- 

 cular canals which enter the substance of the disk. 



If a vertical section of the cephalic shield of Cephalaspis Lyellii 

 is carried through the orbits and perpendicularly to the axis of the 

 body, it will be seen that the disk is exceedingly thin, hardly anywhere 

 attaining ^th of an inch in thickness, except at the margins and the 

 spine, which are thicker. At the lateral margin the thin lamella is 

 bent abruptly and almost horizontally inwards for about a quarter of 

 an inch. It then suddenly thins so much as to be little more than 

 a flexible membrane, which in the specimen now under description 

 is pressed up into close proximity with the dorsal part of the shield 



(fig- 4). 



The thinness and fragility of the disk of Cephalaspis render it 

 difficult to obtain good sections for microscopical examination. The 

 best I have seen (PL XIV. [XXXI.] fig. i.) is taken at an angle of about 

 45° to the longitudinal axis of the head, and intersects the occipital 

 spine just beyond its origin. The section of the spine is in the best 

 condition, and may be described first. 



It is about xffth of an inch thick in its thickest part, which corre- 

 sponds with the median ridge of the spine, and presents three regions 

 or layers, distinguishable from one another partly by their minute 

 structure, and partly by the different mode of distribution of the vas- 

 cular canals by which the tissue is permeated in each. The innermost 

 or deep layer {d) is made up of superimposed lamellae not more than 

 2-xnr th of an inch thick, each of which sometimes appeared to be still 

 more finely laminated. 



Interspersed among these, at greater or less distances, are numerous 

 osseous lacunae, whose long axes are parallel with the planes of the 

 laminae (fig. 3). The length of these lacunae varies greatly, but 

 may be taken at arnro-th of an inch on the average ; some, however, 

 are twice or three times this length, while others are much less. The 

 transverse diameter is equally variable ; but none that I measured 

 exceeded i^oV-oth 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. 



