30 Annals of the Carnegie Museum. 



moderately acute, the mouth not vevy large, oblique, the jaws subequal. 

 Opercle very large, longer than eye, smooth and convex; subopercle dis- 

 tinct; preopercle well developed but not expanded; scales small, even, 

 represented by depressions or pits; about 50 in lateral series. Dorsal 

 fin median, short and high, of about 12 rays. Anal short and rather low, 

 of about 12 rays; 2nd interhaemal strong. Ribs short. Ventrals well 

 developed, about as large as pectorals, inserted under front of dorsal. 

 Pectorals moderate. Vertebrae about 17 + 13 = 30. Caudal deeply forked. 

 The opercle, nearly round, very large, smooth or slightly striated, shining 

 black as preserved, is a most conspicuous feature of the broken specimens 

 of this species. We are not able to place this fish in any of the recognized 

 genera. Assuming that it is a herring, which is most probable, its nearest 

 relative would seem to be the genus Halecopsis of the Eocene of Europe. 

 But Halecopsis has the preopercle greatly expanded, while the opercle 

 is moderately developed. The reverse is true in Dastilbe, in which the 

 large opercle shining black in the fossil state is a most conspicuous feature 

 of the fragments in the rocks. 



We name our species of this group Dastilbe crandalli. We are reasonably 

 certain that it is distinct from a Brazilian fish described by Woodward as 

 Scomhroclnpea scutata, as that species has forty vertebrae, a lower dorsal, 

 and no traces of the very conspicuous opercle characteristic of Dastilbe, 

 In any event, our fish cannot be a Scombroclupea. The generic traits of 

 Scombroclupea, the finlets and scutes behind the anal, and the strong short 

 plates along the ventral line are wanting in Dastilbe crandalli. 



3. Dastilbe crandalli Jordan (sp. nov.) 



? Scombroclupea scuiata Woodward, Quart. Journ. Geol. Society, LXIV, No. 255, 

 1908, 360, pi. XLIII, fig. 3, 4. (Ilheos, Brazil.) 



The following account is drawn particularly from No. 91, 2f inches 

 in length, collected by Dr. J. C. Branner, at Riacho Doce (Plate IX, 



fig. 9)- 



Body moderately elongate, compressed, the depth 3^ in length to 

 base of caudal. Head 3^ in length, Head badly crushed in all speci- 

 mens, especially anteriorly, the eye apparently about as long as the snout; 

 snout moderately acute, but crushed; mouth not very large, oblique, its 

 structure apparently as usual in herrings. Opercle unusually large, form- 

 ing nearly two-fifths length of head; about as deep as long, convex, and 

 nearly smooth. It is shining black as seen in the rocks, and is recognizable, 

 however crushed the specimen may be. Subopercle rather large, evident 



