540 MR. A. SMITH-WOODWARD ON THE [June 23, 



are short and of equal size, and the vertical extent of the operculum 

 is about twice its autero-posterior measure. The first soft ray of the 

 dorsal fin is very robust and divided by numerous, closely approx- 

 imated, transverse sutures. 



Systematic Position, 



Finally, it remains to determine the systematic position of the 

 genus under consideration. By Agassiz', Rhacolepis was regarded 

 as a Percoid, probably because the scales had the appearance of 

 being ctenoid, for he had already observed the abdominal situation of 

 the pelvic fins, which would rather point towards a relationship with 

 otber types. More recently, Dr. Giinther ^ has quoted the genus as 

 one of the Berycidse ; and these are the only two expressions of 

 opinion in regard to the affinities of Rhacolepis that I have succeeded 

 in discovering. A glance at the fossils now made known, however, 

 renders it obvious that we are here concerned with a truly physosto- 

 mous fish ; and it is in this primitive division of the Teleostei that we 

 must look for its nearest living representatives. 



As kindly pointed out to me by Dr. Giinther, some features 

 displayed by these fossils are curiously similar to those of certain 

 Characinoids still inhabiting the fresh waters of Brazil. The scales, 

 for example, have an especially Characinoid aspect, and the large 

 size of the circumorbital bones is also a prominent character of the 

 fishes of this family. But the great number of the branchiostegal 

 rays, the peculiarities of the tail, and the fact that these fossils are 

 accompanied mostly by marine forms, are circumstances that seem 

 to point in another direction. 



The discovery of an "axillary appendage" in some of the 

 specimens, indeed, suggests affinities with the Elopine and Chanine 

 sections of the Clupeidse ; and it is with the first of these groups 

 that I would venture to associate the genus. Flops and its allies are 

 marine types ; their bodies exhibit but little lateral compression ; 

 their posterior circumorbitals are very large ; their branchiostegal 

 rays are generally numerous ; and the tail in these forms almost 

 precisely parallels that of the ancient Rhacolepis. The correspondence 

 is thus so close that there can be no doubt as to the correctness of 

 the determination. 



It is, in fact, difficult to satisfactorily distinguish the Brazilian 

 fossil from some other Elopine genera already recognized ; for, iu 

 dealing with extinct forms, the imperfection of the palseontological 

 record often prevents any very precise comparisons. Taking first the 

 living genera, Megalops may be said to ditfer especially in possessing 

 a long anal fin, a distinct lateral line, and villiform teeth ; while 

 Elops seems to be separated by little beyond the conspicuous charac- 

 ter of the lateral line, and the absence of small scales on the dorsal 

 and caudal fins. Among fossil allies, Elopopsis ^ has a more power- 



' L. Agassiz, Edinb. New Phil. Journ. vol. xxx. (1841), p. 83 ; Eech. Poiss. 

 Foss., Synoptical Table, vol. i. p. xliv. 



^ A. Giiuthei-, ' Study of Fishes,' 1880, p. 421. 



^ J. J. Ileckel, ' Beitr. Keiiiit. foss. Fische Oesterreiehs,' 1856, p. 65. 



