614 A Revision of the American Cichlidse. 



7. RETROCULUS geii. dov. 



This genus is closely allied to Acaropsis, from which it differs to 

 a remarkable deo^ree in the shape of the head. The difference is 

 largely due to the high development of the snout and the pre- 

 orbitals. 



Retroculus boiilengeri sp. nov. 



Type: One specimen (No. 1922), 240 mm. long, Brazil. F. C. Hartt. 



D. XVI, 11 ; A. Ill, 7. Head about 3 ; depth 3. Lateral line 27 + 19 + 2 

 or 3 on caudal. Lateral line overlapping by 7 — 9 scales. Seven series of 

 scales from origin of dorsal to lateral line ; ten or eleven from lovrer limb to 

 vent. 



Anal margined with dusky, a black spot on base of the first five dorsal 

 rays. Upper lip dusky. 



Greneral shape of a Geophagus. Ventral outline nearly straight. Profile 

 steep to first dorsal spine, then gradually descending to caudal. Eye entirely 

 above the gill opening and in the posterior half of the head, 2 in preorbital, 

 If in interorbital, about 5 in head. Snout much produced, more than twice 

 the postorbital portion of the head. Mouth nearly horizontal, maxillary 

 reaching to nares, which are nearer the eye than the tip of the snout. Pre- 

 maxillary when protracted equals one- third of the distance from its tip to the 

 orbit. Four series of scales on the cheeks. A series of strong conical teeth, 

 those of the lower jaw smaller, a band of villiform teeth behind them. 



Gill-rakers small, compressed claw-shaped, the anterior border pectinate. 

 Soft dorsal and anal angular reaching the caudal. A few scales along the 

 rays. Caudal rounded, thickly scaled to its tip. 



We take pleasure in dedicating this species to Dr. G. A. Boulen- 

 ger, of the British Museum. 



8. PETEIVIA Giinther. 



Petenia Gunther, Cat. Fish. Brit. Mus., IV, 1862, p. 301 (splendida). Steind. 

 Beitr. Kenntn. Chrom. Amazon'str., 1875, p. 36 (redefined). Eigen- 

 mann & Eigenmann, Proc. U. S. Nat. Mus., XIV, 1891, p. 69 (redefined). 

 Type : Petenia splendida Giinther. 



This genus is known from three species. It is closely allied to 

 Acaropsis and Retroculus, from which it differs in its elongate anal. 

 There is no difficulty in finding intermediate numbers of spines, if 

 all the species which have hitherto been united under Acara are to 

 be retained as one lump? There are, however, characters aside 

 from the anal spines to warrant the separation of at least some of 

 the species as separate genera. As soon as this is done the remain- 



