CLASSIFICATION OF DEVONIAN FISHES. 



19 



All the structural characters, in fact, which have been enumerated 

 above among the peculiarities of Undina, are equally well marked 

 in Macropoma, except that, hitherto, I have been unable to meet 

 with the caudal appendage in the latter, and that the teeth are 

 more distinct and cylindrical. But further than this, as Dr. Mantell 

 originally suspected, and as Professor Williamson has since demon- 

 strated, MacTojpoma exhibits the peculiarity, without a parallel, 

 so far as I know, among fishes of other families, of having the 

 walls of its air bladder ossified. Now, I find good evidence of the 

 existence of a similarly ossified air bladder, not only in Undina, 

 but in a well-preserved specimen of a new genus of Coelacanth 

 from the Lias (described in the subjoined note by Sir Philip 

 Egerton), in the Museum of Practical Geology.* 



* Holophagus Gulo. 



Mr. Harrison's specimen wants the anterior portion from the dorsal and pectoral fins 

 forwards. From the insertion of the dorsal fin to the extremity of the tail it measures 

 inches, and 4|- inches in depth. The stomach is distended with a recently swallowed 

 Dapedius, and a large coprolite occupies the rectum. The first dorsal fin springs (as 

 in Macropoma) from a single disc, resulting from the coalescence of the interneural 

 spines. It contains eight long, thick, undivided, and multiarticulate rays. They are beset 

 with numerous short spines or tubercles. The second dorsal is situated 4 inches behind 

 the first. Between the two is seen a strong bifurcate interneural ossicle, which has been 

 displaced forwards from its proper position at the base of the fin. The second dorsal fin 

 contains sixteen rays. The anterior ones are short and slender. The succeeding ones 

 are long, broad, and multiarticulate, but not tuberculate. Tlie base of the fin is obtusely 

 lobate, with a scaly investment. The pectoral fins are much mutilated. Judging from 

 what remains of them, and from some indistinct impressions, they seem to have been of 

 great size. The anal fin occurs imimediately below the second dorsal fin, with which it 

 corresponds in form and structure, but contains many more rays. The ventral fins are 

 mutilated, but their position below the first dorsal fin is indicated by the preservation of 

 a pair of strong T-shaped pelvic bones, having their longer limbs directed forwards, and 

 nearly reaching the base of the pectoral fins. The caudal fin is of great size, and presents 

 in an eminent degree the most special and characteristic feature of the Coelacanthus 

 family, namely, the interposition, in the caudal region, of an interneural between the 

 neural and dermo-neural spines. The base of this spine abuts upon the extremity of the 

 neural spine, and unites with the true fin-ray by an overlap or splice. This structure 

 coincides with that observed in Undina. In Macropoma the bone of the interneural spine 

 is bifurcate for the reception of the distal extremity of the neurapophysis. A small 

 supplemental fin extends an inch beyond the larger caudal fin, as in Undina and Ccela- 

 canthus. The notochord is unossified. The apophyses, both above and below, have 

 very wide bases. The scales are curvilinear, and covered with a vermiculate pattern on 

 the upper surface, occasionally broken up into small tubercles. 



In the Woodwardian Museum at Cambridge there is the head and part of the trunk of 

 a CccJacantlms, from the Kimmeridge Clay at Cottenham, The head shows the frontals, 

 prefrontals, and lower jaw, with the tympanic attachments. The glossohyal plate is double, 

 as in Holoptychius. The scales are roughly undulate, coarser in pattern than in Undina, 

 Codacanilius, and Holophagus, but not absolutely tuberculate, as in Macropoma. One 

 fin is preserved, probably the left pectoral. It is lobate, broad, and strong. The 

 operculum is triangular, the frontals short, and the prefrontals descend at an abrupt 

 inclination. 



