38 ME. E. LEONARD GILL ON THE PERMIAN 



name of Acentrophorus cMcopensis. The caudal lobe, which his 

 figures show to be as short as in Semionotics, is enough, however, 

 to prove that the fish is not an Acentrophorus. In addition, the 

 scales are too uniform in size, and the general outline does not 

 suggest Acentrophorus either in head or body. It should also be 

 said that Newberry's discussion of the genus and of its relation- 

 ship to '■'■ Ischypterus'''' (pp. 67-69) is on the whole misleading, 

 since it is founded only on the obscure figures accompanying 

 Kirkby's paper of 1864. 



Another fish assigned to the genus is Acentropihorus dispersus 

 Eritsch (1894, p. 81, pis. 113, "ll4; from the Lower Permian 

 (Schwartenkohle) of Kounova, Bohemia. Its remains are very 

 imperfect, but Fritsch's figures show enough of the upper and 

 lower jaws, the opercular apparatus and the pectoral girdle to 

 prove that in all these essential elements of its structure the fish 

 was fundamentally different from any Semionotid. Indeed, at the 

 end of his description of it, Fritsch himself appears to have 

 concluded that it was a Palfeoniscid. 



Summary. 



The propriety of placing Acentrophorus in the family Semiono- 

 tidag was not in doubt, but it is amply confirmed by the additional 

 information now brought forward. In the exact correspondence 

 of the dorsal and anal fin-rays with their endoskeletal supports, 

 in the reduced maxilla, the narroAv preoperculum and the whole 

 plan of the opercular apparatus, in the absence of an infra-clavicle 

 as in all the details of the pectoral arch and the pelvic bones, 

 AcentrophoTUiS is a characteristic representative of the Proto- 

 spondyli and of the family Semionotidse. 



At the same time, Acentrophor%is does in certain respects retain 

 primitive characters. It seems, in fact, to provide us with some- 

 thing that is comparatively rarely found among fossil fishes, 

 namely a really early representative of a new group which still 

 shows definite traces of its origin from older types. The most 

 striking character of this kind in Acentrophorus is undoubtedly 

 the upper caudal lobe. As compared with the caudal lobe of the 

 contemporary Palaeoniscids it is certainly greatly reduced, but the 

 reduction is only in depth; the caudal lobe of Acentrophorus is 

 still of the maximum length. It is interesting to find in another 

 contemporary fish, the extraordinary form Dorypterus hoffmanoii, 

 a caudal lobe in precisely the same stage of reduction, except 

 tha.t, in place of the single line of scales found in Acentrophorus, 

 the lobe is marked externally to the very tip by two lines, the 

 scales as they are traced backwards becoming linear and almost 

 microscopic. That a stage such as this — the caudal lobe 

 extremely attenuated in its hinder portion but still more or less 

 of the full length — may have been passed through pretty regularly 

 in the production of a " hemi-heterocercal "' tail is suggested not 

 only by the parallelism between AcentrojjJwrtis and Bcrypterus, but 



