Jordan : Fossil Fishes from Riacho Doce. 27 



This species appaa-ntly l^elongs to the same ^roiip as Diplomystiis 

 longicostatiis Cope, but (he latter is readil>' distinjiuished by the much 

 elevated back, on which the dorsal fin is perched, and by the much longer 

 ribs, as well as by the great enlargenu nt of the posterior ventral plate. 



In both species the ventral fins are very small and median, but in 

 Ellipcs lnfi<^icoslatiis they are inserted well before the dorsal under the 

 eleventh ril), the dorsal being tlirowii backward !)>■ the obiicjue setting 

 of the body. The insertion of the ventrals behind the front of the dorsal 

 separates Ellipes brauneri and E. riacensis from E. longico status, as well 

 as from Diplomystus dentatus of the North American Eocene. 



I have submitted specimens of Ellipes hranneri to Prof. T. D. A. Cock- 

 ercll, of the Unixersity of Colorado. Professor Cockerell says: 



"I have examined the scales of Ellipes with very great interest. In no 

 one case can I see the outline of a complete scale, but the transverse circuli 

 are very well developed, and the scales are evidently large. I cannot 

 distinctl\- determine that they differ in any respect from those of the 

 common herring, except perhaps that the sculpture is not quite so uniform. 

 I cannot see the transverse radii, but these might not be visible in pre- 

 served material, and possibly the radii did not exist in these ancient 

 herrings." 



We refer to Ellipes branneri, with some doubt, two other specimens 

 well preserved, Number 3 described below, with another specimen of the 

 same size, well preserved, but broken into two (Number 24 and Number 

 25). These agree with the type of Ellipes branneri in all technical char- 

 acters, but the body is more elongate, especially posteriorly, and the 

 ventral region less gibbous and more angulated. The belly is deeper, and 

 the mouth larger than in Ellipes riacensis. 



Number 3 is a little fish 2f inches long, very perfectly preserved in 

 black shale, collected at Riacho Doce, state of Alagoas, Brazil. (Plate 

 VIII, fig. 5-) 



Head 3^ in length to base of caudal; depth 2}; length of longest rib 

 considerably more than length of head. Distance from snout to nape 

 about equal to distance from nape to dorsal fin Length of caudal part 

 of vertebral column equal to greatest depth, and half more than head. 

 Body deep mesially, the caudal region produced. Belly very convex and 

 sharply keeled, the plates strong and sharp, about 17 in number; trace of 

 four or five scutes before dorsal. These are subacute, smooth, and entire. 

 Caudal peduncle longer than deep in front; head deeper than long- 

 maxillary about 2 in its length, reaching to below middle of eye; lower 



