﻿20 



GANOID FISHES OP THE CARBONIFEROUS FORMATION. 



other. The operculum preserves the same general form with slightly varying proportions 

 in most of the other genera, such as Elonichthys, Nematoptychius, Pyyopterus, Acrolepis, 

 Oxyynathus, &c. In Cosmoptychius (PI. II, fig. 7) and Bhabdolepis (PI. II, fig. 6) it is 

 very narrow and acutely pointed inferiorly ; in the latter genus it is also rather small. 

 I rather suspect it has a similar form in Cheirolepis, though in that remarkable genus I 

 have never obtained a satisfactory demonstration of its contour. In Amblypterus (PI. II, 

 fig. I) it is rather of a more quadrate form, being less deep and less obliquely placed on 

 the side of the head. Below the inferior margin of the operculum, and overlapped by it, 

 we find in Palceoniscus another plate (i. op.), of a somewhat quadrate form, but broader 

 behind than in front, the posterior margin being still a little oblique, the inferior one 

 horizontal. A corresponding plate, always more or less square-shaped in aspect, in 

 Jmblypterus, however, a little higher proportionally, is found in all the genera of 

 Palaoniscidce. At first sight one is inclined to set down this plate unhesitatingly as 

 " suboperculum," and in my own previous essays on fishes of this family I have hitherto 

 lettered it as such, but a recent examination of specimens of the Lower Permian genus 

 Bhabdolepis has induced me to take another view of its homologies. In Bhabdolepis 

 (PI. II, fig. 6) there occurs below the narrow and inferiorly acutely pointed operculum, and 

 between it and the quadrate plate {i. op.) last under consideration, tending also anteriorly 

 to pass obliquely upwards for a little distance between the operculum and praeoperculum 

 (p. op.), another plate (s. op.), of a horizontally narrower form. An attenuated repre- 

 sentative of this additional plate is found also in the beautiful Cosmoptychius striatus 

 (PI. II, fig. 7) of the Wardie shales ; here it is triangular, placed above the anterior 

 superior angle of the quadrate plate (i. op.), which it thus does not entirely exclude from 

 contact with the operculum ; its upper angle is produced into a slender tapering process, 

 which extends upwards for a little distance along the anterior margin of the operculum. 

 Now, it is clear that this plate {s. op.) is a part of the opercular apparatus, and as such 

 must either correspond to the suboperculum or be considered as additional and supple- 

 mentary. I have preferred the former alternative, and have lettered it accordingly, 

 though in the vast majority of Palceoniscidce it seems to be either entirely absent or so 

 small as to escape notice. Consequently the quadrate-shaped plate (i. op.) which, on the 

 other hand, occurs conspicuously throughout the entire family, must be the equivalent of 

 the inter op ercid urn, and, in truth, its relation to the mandible bears out this interpretation, 

 its anterior inferior angle coming close to the articular extremity of the lower jaw ; that 

 relation being, however, somewhat concealed externally by the posterior overlapping 

 angle of the maxilla. Covering the hyomandibular and a portion of the cheek in front of 

 the last-described bones is a plate of a peculiar form {p. op), which may best be considered 

 asprteoperculum. It consists of two portions or " limbs," upper and lower, set at an obtuse 

 angle to each other, the centre of ossification being placed at the point of divergence. 

 The upper part is narrowly triangular ; the lower margin, nearly horizontal, is in contact 

 with the maxilla ; the upper margin, sloping obliquely downwards and backwards, is in 



