116 OLD EED SANDSTONE FISHES. 



termination it gives off on each side, at an acute angle, a forwardly directed branch, 

 which is, however, rather faintly exhibited in the photograph. These ridges are 

 not seen either in Pterichthys or Asterolepis. We observe in this figure on the left 

 side (right hand side of the reader) the extensive antero-lateral area for over- 

 lapping the anterior dorso-lateral plate, while on the other side a portion of the 

 bone is broken away at the corner, showing the impression of the external sculpture 

 below. 



The outer surface of the posterior median dorsal plate (cast) is shown in PI. 

 XXV, Fig. 1, where, like the anterior median dorsal, it is seen to be entirely 

 destitute of a keel along the middle line. The internal surface is represented by 

 another specimen, Fig. 2, which is perfect except as regards a small piece broken 

 away from the hinder part of the right margin. In the middle line and nearer the 

 back than the front, is a rough, longitudinal crest, no doubt for muscular attach- 

 ments ; quite in front is a transversely elongated triangular area for overlapping 

 the posterior margin of the anterior median dorsal, while on each side is an exten- 

 sive surface for overlapping the posterior dorso-lateral element. Both of these 

 plates are represented one half the natural size. 



The outer aspect of the anterior dorso-lateral plate (cast) is shown in PI. 

 XXIV, Fig. 3, and is complete anteriorly and posteriorly, but not quite so at the 

 sides. The exposed surface of the dorsal portion is two and a half times as 

 long as broad ; part, but only part, of the area overlapped by the anterior median 

 dorsal is seen along its inner border. The lateral aspect of the plate is imperfect, 

 yet shows very distinctly the longitudinal lateral sensory groove running along 

 from front to back. The internal aspect of the plate, cast from the counterpart 

 of the same specimen is shown in Fig. 4, the broad area for overlapping the posterior 

 dorso-lateral being conspicuous at the posterior end. The lines distinctly radiating 

 from a j^oint near the front show perfectly plainly that the plate was ossified from 

 one centre, and does not represent two fixed elements, as supposed by Jaekel to be 

 the case, at least in Pterichthys. 



Of the posterior dorso-lateral element, no perfect specimen has turned up, the 

 most perfect I have seen being represented in PL XXV, Fig. 3 (cast) ; it is, 

 unfortunately, deficient along the inner margin of its dorsal portion. The lateral 

 portion is, however, pretty well intact, and shows a large surface or area, deep and 

 broad behind, which is overlapped by the corresponding posterior ventro-lateral, and 

 over which the backward continuation of the lateral line groove is observable. In 

 front of both the dorsal and lateral portions of the plate is seen the area overlapped 

 by the anterior dorso-lateral, while behind we see part of the surface overlapped by 

 the posterior median dorsal. 



The right anterior ventro-lateral (cast) is shown in PL XXV, Fig. 4, the specimen 

 being nearly complete except that the brachial process is broken. The plate of the 



