﻿30 



GANOID FISHES OF THE CARBONIFEROUS FORMATION. 



ornamented with punctures, furrows, ridges, or even with raised tubercles ; the patterns, 

 often of great beauty, are, however, not in every case available for the determination of 

 species. On the inner surface near the middle there is seen a raised keel, parallel with the 

 anterior and posterior margins, and which ends above just in front of the base of the pointed 

 articular spine which projects from the upper margin, and is received in a corresponding- 

 pointed depression on the inner surface of the scale above just behind the lower commence- 

 ment of its own keel. The anterior superior angle of the scale is more or less produced 

 upwards, sometimes considerably so, the produced part being also overlapped partly 

 by the scale in front, partly by that above, so that, with the proper articular spine, 

 there may seem to be two projections from the upper margin, as in Acrolepis, 

 Pygopterus, &c. In Nematopfychius the form of the scales of the flank is very 

 peculiar. They are much higher than broad, the overlapped area is very narrow, 

 the exposed one is indeed rhomboidal, but the acute angles are here the posterior 

 superior and the anterior inferior. The anterior superior angle is not produced into a 

 point distinct from the proper articular spine, which, broad, flat, and triangular, here 

 arises from the whole or nearly the Avhole of the upper margin of the scale. On the under 

 surface of the scale the keel is obsolete, but the usual socket for the spine of the scale 

 next below is seen at the inferior margin. Some approximation in general form is here 

 shown to the form of scale characteristic of the Platysomida, excepting of course the 

 want of the strong rib on the inferior surface of the anterior margin found in that 

 family. The scales of Centrolepis have also a very similar form, but are not so high 

 proportionally. 



But the most aberrant condition of the squamation in the entire family is that of 

 C/teirolepis, and which led to considerable error as to the systematic position of the genus, 

 before the osteology of its head was specially studied. The scales are here extremely 

 minute, quadrangular, thick in proportion to their small size, and the amount of overlap 

 is so small as to be hardly perceptible, so that the squamation reminds us of the 

 shagreen-like covering of the Acanthodei with which Cheirofajpis was for a long time 

 classed. These minute scales are, however, most markedly arranged in oblique dorso- 

 ventral bands, which, on the caudal body-prolongation, undergo a change of direction 

 exactly as in other Palaoniscida. The outer surface is ganoid and marked with diagonal 

 striae, the inner one shows a well-marked keel parallel with the anterior and posterior 

 margins, but which is not prolonged upwards into an articular spine ; this keel was, 

 in my opinion, mistaken by Agassiz in C. Tra'dlii for an ornament of the outer 

 surface. 



Returning now to the scales of the more typical genera of Palaoniscida {Palceoniscus, 

 fflonichthys, &c), we find that these vary somewhat in form on different parts of the 

 body. On the side of the flank they are usually higher than broad, but above, below, 

 and towards the tail they get more equilateral ; on the ventral aspect, indeed, they 

 usually become very low and narrow, the anterior superior angle becoming more 



