428 Bulletin American Museum of Natural History. [Vol. XIX, 



been pointed at both ends. It resembles a similar ray seen 

 in the same situation in Microccelia dayi. 



Another example, No. 45206 (3696), shows the ventrals 

 placed slightly farther backward than in the type. The pec- 

 toral fin has ID rays. Other specimens, referred to the same 

 species, present the pectoral fin as placed considerably above 

 the ventral border. 



It is probable that normally the ventral fins of this species 

 are inserted opposite the middle of the dorsal fin. 



This species differs from A. hoops Agassiz and A. lihanicus 

 A. S. Woodward in having a shorter head and a smaller orbit. 

 All the specimens, about 10 in number, are from Hajula, ex- 

 cept No. 4520c (3836), No. 45200^ (3767), No. 4520^ (3730). 

 which are from Hakel. 



Named in honor of Rev. D. Stuart Dodge, the donor of the 

 collection here described. 



Nematonotus longispinus {Davis). 

 Plate XXXIV, Figure 2; Plate XXXV. 

 Psetidoheryx longispina Davis (J. W.), Trans. Roy. Dublin Soc. 



(2), III, 1887, p. 511, pi. XXV, fig. 2. 



Nematonotus hottcB (in part) Woodward (A. S.), Cat. Foss. Fishes, 

 IV, 1901, p. 250. 



Among the fishes collected by Professor Day at Hajula there 

 is a considerable number which appear to me to be different 

 from Nematonotus bottcB, and which I have no means of dis- 

 tinguishing from the species described by Davis, as above 

 cited. The latter has been identified by Dr. A. S. Woodward 

 as N. bottcs; and his procedure may prove to be correct. If 

 so, the form here described will require a new specific name. 



These Hajula specimens differ from N. bottcB especially in 

 the great length of one of the anterior dorsal fin-rays. It also 

 appears to attain a much greater size than does the species 

 from Hakel. Six specimens are especially to be mentioned, 

 as follows: Nos. 451°^-/ (3655. 3663, 3664, 3678, 3679, 3723). 



No. 4510a (3655) has a total length of 95 mm. The pec- 

 torals and ventrals appear to be larger than they are in speci- 



