12 



BRITISH FOSSILS. 



respect my genus Cmlacanthus very nearly approaches the type of a 

 fish from the Lithographic limestone of Kehlheim, for which Count 

 Milnster has proposed the generic name of Undina. But notwithstanding 

 this analogy, and the altogether similar disposition of the other fins, the 

 fish in question is distinguished by many peculiarities which do not 

 permit it to be confounded with Ccelacanthus. The most important 

 difference is presented by the dentition. The genus Undina has, accord- 

 ing to Miinster, pavement-like teeth, very similar to those of certain 

 Pycnodonts. Ccelacanthus^ on the other hand, has conical teeth, like the 

 Sauroids, and everything leads to the belief that it is a carnivorous fish, so 

 that, far from belonging to the same genus, it is doubtful whether it 

 belongs to the same family. Leaving the caudal aside, the other fins of 

 the genus Ccelacanthus present a very simple structure, composed of 

 slender but not dichotomous rays. The first dorsal corresponds to the 

 extremity of the pectorals. The second is opposite the space between 

 the ventrals and the anal. The anal itself is very closely approximated 

 to the caudal. This last fin (comprising in it the bundle of articulated 

 rays which fringes the extremity of the vertebral column) nearly equals 

 one-third of the total length of the fish. The vertebrae are much higher 

 than they are long towards the anterior part of the trunk, but they 

 become sensibly elongated posteriorly. It is the same with the apophy- 

 ses, which, very slender in the abdominal region, take on a much greater 

 development in the caudal region. The scales, to judge by the fragment 

 of C. granulosus^ are large, elongated, and have their posterior edge 

 rounded. I have not been able to ascertain whether they are enamelled 

 or not, but the fact that they are found in strata older than the Jura 

 leads me to suppose that, as in all the fishes of that age, they were 

 enamelled. Their extreme thinness, no doubt, has made them too 

 fragile to be often preserved. I conclude from this description that the 

 genus Ccelacanthus, although near the genus Undina of Count Miinster, 

 is nevertheless different from it, and that the latter should therefore form 

 a separate type in the Ccelacanth family. Consequently it will be 

 necessary to exclude from Ccelacanthus, and to place in Undina, the 

 remarkable species which Count Munster has described and figured in 

 the fifth part of his "Beitrage," under the names of Ccelacanthus 

 striolaris and Kdhleri. The true genus Ccelacanthus is at present 

 restricted to the coal, the Zechstein and the Muschelkalk. I am ac- 

 quainted with six species. 



" Ccelacanthus granulosus, Agass. (Vol. 2, Tab. 62). 



" The species to which I give this name was for a long time the only 

 one known, and the two fragments which are figured were the only 

 representatives of this remarkable family. Both represent the posterior 

 part of a fish of very large size, which, to judge by the relative positipn 



of its fins, ought to have been at least two feet long 



As a general rule the apophyses (neural spines and arches) and the 

 inter-apophysial bones are equal in length. The rays, on the other 

 hand, are a little longer, but they are never jointed down to their bases. 

 The cleft of the ray into which the point of the inter-apophysial bone is 

 inserted is much narrower than that of the apophysis (neural arch), 

 which embraces the vertebral column. It is probable, from all I have 

 been able to see, that, in reality, these anomalous rays are composed of 

 filaments [Jilets~\, as in most other fishes, only these filaments do not 

 become separate. The rays at the extremity of the caudal fin are 

 exceptions to the general rule, inasmuch as they are directly attached 

 to the vertebral column, without being borne either by an apophysis, or 



