1889.] EOCENE SILUROID FISHES. 205 



sented by figure 10, and it will be seen that all the points mentioned 

 as characteristic of the fossil otolith are repeated in this, and it is 

 only in outline that there is any real difference. The projection of 

 the inner margin marked x (fig. 10 b) is in a depression and does not 

 form a prominent angle as in the fossil (fig. 3 b), and the swelling of 

 the outer margin {y) is, in A. gagorides, placed further backwards 

 than in tiie fossil. 



Unfortunately, the otoliths of nearly allied recent species or genera 

 are not available for comparison, and consequently we know nothing 

 of their specific differences. In the collection of Fish otoliths pre- 

 served in the Ilunterian Museum of the Royal College of Surgeons 

 there are a few belonging to Siluroids, but none of them to genera 

 nearly allied to Arius, and they all differ widely from the otolith of 

 Arius gagorides. 



The series of otoliths from the Upper Eocene of Barton, pre- 

 served in the British Museum, includes many which agree with 

 A. gagorides in these main characters which seem to me to be generic, 

 and these, therefore, 1 also refer to the genus Arius. Besides differ- 

 ences of size, which in part no doubt are due to age, these otoliths 

 {)resent several distinct forms, which I believe will be found to 

 represent at least three species, in addition to the skull above 

 described. The largest of these (fig. 4) is a little longer and more 

 regularly oval than that found in the Barton skull ; its lower surface 

 is also flatter, and its uj)per surface is raised into an almost conical 

 boss. 



The second form to be noticed (fig. 5) is smaller, flatter, and more 

 rounded in outline, having the hinder point only slightly produced. 



The third form (fig. 6) is likewise flat and about the same size as 

 tlie une last noticed ; in outline, however, it more resembles that of 

 A. gagorides, but the swelling of the outer side (ij) is not thrown so 

 much backwards as in that species. 



There is still another form of Arivs otolith to which I should 

 like to call attention. Among the fossils brought from Madagascar 

 by the Rev. R. Baron, and noticed in his p;iper read before the 

 Geological Society (Mar. 6, 1889), were some small otoliths (fig. 7) 

 which he had collected in the village of Ankoala, where they 

 occurred in some numbers scattered over the surface of the ground. 

 These otoliths bear such a close resemblance to some of those from 

 the Eocene beds of Barton, that they not unnaturally led to the 

 supposition that they also were of Eocene age; but both these 

 forms are referable to ihe living genus Arius, which is a widelv dis- 

 tributed tropical form, and it seems very probable, therefore, that 

 the Ankoala specimens may prove to be of much more recent origin, 

 and the peculiar conditions under which they were found seem to 

 point to their belonging to a living species. 



We have now to consider the relation which the Barton skull and 

 the otoliths above described bear to the specimens referred to Arius 

 egertoni and to A. ? bartonetisis ; and before doing so I may say that 

 1 quite agree with Mr. Smith Woodw;ard's reference of the cephalic 

 plates from Bracklesliam to the species A. egertoni ; for their 



