5o6 ON CEPHALASPIS AND PTERASPIS 



without convolutions and rarely branching, towards the external 

 sharp angle." (/. c. p. 46.) 



I am not aware of the existence of any other account of the minute 

 structure of Cephalaspis and Pteraspis beyond these ; and I will 

 therefore now proceed to the immediate subject of this paper, which 

 is, to describe that structure more fully and, I hope, more accurately 

 than previous observers have done, — to compare Pteraspis and Cepha- 

 laspis, pointing out their real differences and resemblances, — and 

 finally to consider the bearing of the structural facts upon the ques- 

 tion of the zoological position of these ancient fishes. 



Cephalaspis. (PI. XIV. [XXXI.]) 



In but few of the specimens of Cephalaspis Lyellii which I have 

 had the opportunity of observing, has the external surface of the 

 cephalic shield been well exhibited, or preserved over any considerable 

 surface. Where best shown it is somewhat uneven, and presents that 

 curious apparent division into polygonal (usually hexagonal) are^ 

 which has been described by Professor Agassiz. On examining the 

 apparent sutures closely, however, they have not presented to my ob- 

 servation precisely the appearance figured in the pi. \b. fig. 2. of the 

 ' Recherches.' They appear rather as if short, delicate, reddish- 

 brown lines had been ruled across the line of junction of the sides of 

 the hexagons, for some way towards the centre of each hexagon ; and 

 these lines are so gently convergent as to seem nearly parallel. 

 Neither do I remember to have met with such strongly marked 

 central elevations as those represented in the figure cited. 



The inner surface of the disk has presented itself well preserved 

 in more than one specimen. It never exhibits any trace of the 

 apparent sutures of the outer surface (compare Agassiz's ' Recherches,' 

 pi. I b. fig. 3, where this fact is clearly shown), but appears whitish,, 

 enamel-like, and very smooth where it is not furrowed by certain 

 shallow and narrow depressions which radiate from the region of the 

 orbits and occiput towards the margin, before reaching which they 

 repeatedly subdivide and anastomose. I do not doubt that these are 

 the impressions of the vessels which ramified under the disk during 

 life. Sometimes, by the elevation of the substance of the disk into a 

 wall on each side of one of these depressions, the latter may become 

 almost converted into a canal, so as to retain a portion of the matrix. 

 This however is a rare occurrence. 



When the concave inner surface of a disk and the convex cast of 

 another specimen are compared, it is at once seen that the ''radiating 



