306 A. Smith Woodward — Eocene Siluroid Fishes. 



arch ; and (iii.) by the ornamentation of the upper cranial bones, 

 which are evidently not covered with more than a very thin skin. 

 In all these respects, the present species differs widely from the 

 /S«7MrMS-type, and the first two characters alone, which were originally 

 indicated by Dixon, are quite sufficient to exclude it from the genus 

 just mentioned. Sir Philip Egerton, indeed, seems to have already 

 recognized the fact, and has labelled some of his specimens as 

 Pimelodus. But a careful comparison of the supraoccipital bone 

 with the corresponding element in the large series of recent Siluroids 

 in the British Museum can leave little doubt that the fossil really 

 belongs to the well-known genus Arius, or to some closely allied 

 form which cannot be distinguished upon present evidence. There 

 is the most striking similarity in the few points as yet known, and 

 it is scarcely likely that any great divergence will be noted in the 

 other structures still to be revealed. 



If such a conclusion be substantiated by future discoveries, the 

 fact will become of considerable interest, as showing that this 

 Siluroid fish belonged to a type now characteristic of tropical waters, 

 instead of representing an existing form (Silurus) that rarely trans- 

 gresses beyond the temperate zone ; and this circumstance, of course, 

 is in harmony with all the indications of the associated fauna. 

 Nothing, however, can be said as to its having its nearest ally in 

 the forms now inhabiting any particular region, for the living 

 species of Arius have an extraordinarily wide distribution, and are 

 scattered throughout the freshwaters and littoral areas in almost all 

 tropical parts both of the Old World and the New. 



Formation and Locality. — Middle Eocene : Bracklesham Bay, 

 Sussex. Probably also Upper Eocene, Barton Cliff, Hampshire. 

 [Egerton Collection, B. M. p. 1894a.] 



Arius? bartonensis, sp. nov. 



The second species of Siluroid is indicated by some small spines 

 from the Barton Clay of High Clifi", near Christchurch, Hampshire, 

 but in this case it is impossible to determine the genus even approxi- 

 mately. It may belong to Arius, or may represent some other 

 generic type ; and the name by which we venture to designate it 

 is, therefore, as provisional as those already applied to detached 

 examples elsewhere. The dorsal spine has a very characteristic 

 curvature, being gradually arched backwards for three-quarters of 

 its length, and assuming a more upward direction in the distal 

 fourth, — a peculiarity shown in the accompanying Fig. 3, though 

 still more marked in two other less perfect specimens : and the 

 largest example is almost twice the size of the one here represented. 

 The sides are ornamented with irregular, delicate longitudinal ridges, 

 exhibiting but the slightest traces of the thickenings or nodose 

 expansions so conspicuous in the rugee of A. Egertoni. The anterior 

 edge is provided in its lower part with a short series of small, blunt 

 tubercles ; and towards the distal extremity, which becomes much 

 compressed and sharp, there are little downwardly-directed denticles 

 both in front and behind, imparting to the tip somewhat of a 



