1889.] ON SO.MK EOCENE SILUROID FISHES. 201 



3. A Contribution to the History of Eocene Slluroid 

 Fishes. By E. T. Newton, F.G.S., E.Z.S. 



[Received March 14, 1889.] 

 (Plate XXI.) 



The pectoral and dorsal spines of Siluroid fishes from Brackle- 

 sham, wliich were referred to the genus Silurus by Dixon (2'), have 

 recently been studied by Mr. A. Smith Woodward (12), who has 

 shown the improbability of these remains belonging to the temperate 

 genus Silurus and the close relationship existing between them and 

 the widely distributed tropical genus Arius. In addition to the 

 spines and pectoral arch, named by Dixon Silurus egertoni, Mr. S. 

 Woodward has called attention to several other specimens, some 

 from the Upper Eocene of Barton, preserved with the types in the 

 British Museum, among which are bones of the skull and notably 

 some large and characteristic supraoccipitals, one of which he 

 figures ; these he also refers to Arius ecjertoni. Some smaller spines 

 with a double curvature, from Barton, he places in a new species, 

 Arius (f.) bnrtonensis. 



The Museum of Practical Geology now possesses the greater part of 

 a skull from the Eocene beds of Barton (Plate XXI. figs. 1, 2, 3), which 

 confirms in a most satisfactory manner Mr. S. Woodward's reference 

 of the Eocene Siluroids to the genus Arius. The skull is somewhat 

 crushed, but the bones are still in position, and by careful manipu- 

 lation both the upper and under surfaces have been exposed. The 

 ethmoid, prefrontals, and part of the supraoccipitiil are wanting, 

 and on the right side the temporal region is broken, but on the left 

 only one of the temporal plates is lost. 



All the bones of the ujiper surface, which are preserved, are 

 ornamentfd with rounded granules, and these in nearly all cases 

 radiate from an ill-defined centre towards the margins of the bone. 

 No distinct s^utures can be seen, but the ornamentation being less 

 strongly marked towards the edges of the bones, the boundaries can 

 be fairly well made out ; the dark lines in the figure indicate these 

 boundaries, which agree in the main with the positions of the sutures 

 in the recent specimen with wliich it has been compared. 



The frontals {/r.) occupy the anterior part of the specimen ; they 

 are narrow posteriorly and meet each other in the middle line for 

 about half their length. The median point of the supraoccipital 

 projects for a short distance between their hinder extremities. An- 

 terioily a wide and deep depression occupies the median portion of the 

 frontals, and at the bottom of this depression a long cleft separates 

 their inner margins. Each bone is in front divided into two parts, 

 the outer of them no doubt joined the prefrontal and the inner the 

 ethmoid, as in the recent Arius. 



Behind and on the outer side, each frontal joins a plate (sji.ot.) 



^ These numbei's refer to a list uf work,« giveu p. 206. 



