4 PROCEEDINGS OF THE NATIONAL MUSEUM vol.84 



Pelvic fins united to each other and to the abdomen by a thin mem- 

 brane. A genital sheath of scales about the bases of the first few 

 anal rays of the female. Preorbital relatively wide and armed with 

 many irregular rows of very closely imbricated ctenoid scales. 



Premaxillaries not expanded; they are protractile and strongly 

 joined, the posterior processes fairly long. Dentaries firmly joined. 

 Teeth in both upper and lower jaws conical, in many rows, the outer 

 ones slightly enlarged and situated on the lips outside the mouth. 

 Vomerine teeth and pseudobranchiae absent. Near its widened distal 

 end, the maxillary is loosely connected to the outer skin of the pre- 

 orbital, allowing considerable freedom of movement. Vertebrae, 

 counting hypural, 40 or 41. Haemal arches expanded for the exten- 

 sion of the coelom and the large air bladder into the caudal region 

 as far as the twenty-fourth vertebra.. Hypocoracoid very large, ver- 

 tically expanded; actinosts set high on the arch, half on the hypo- 

 coracoid and half on the hypercoracoid. Flange of the cleithrum 

 extending far upward to upper end of pectoral arch. 



The peculiar atherinid-like habitus and color of this fish sets it 

 off as one of the most interesting specializations of the cyprinodonts. 

 Evidently these little fishes have the same habits in the sealike 

 expanse of Tanganyika as have the atherines in the ocean. 



Subfamily Fundulinae 



APLOCHEILICHTHYS PUMILUS (Boulenger) 



There are 48 specimens, U.S.N.M. no. 84158, of a small Aplocheil- 

 ichthys, the largest 25 mm in standard length, which are placed under 

 this name with strong doubt. They are labeled Nyanga, Lake Tan- 

 ganyika. 



I think there is a distinct probability that some confusion has oc- 

 curred in Boulenger's placement of the Victoria and Tanganyika 

 specimens of this genus (Boulenger, 1915, pp. 45-46). Aplocheilich- 

 thys dhonti (Boulenger, 1919, p. 17) is certainly different from what 

 he calls pumilus, chiefly because of the different dorsal and anal fin 

 contours, distinctly noticeable in the specimens of dhonti I have seen. 

 I have examined three specimens labeled H. pumilus from Boulen- 

 ger's material. One of them (U.S.N.M. no. 94327) from the Lufuko 

 River (collector, Stappers) has a long peduncle and the middle dor- 

 sal and anal rays longest; I believe it represents A. dhonti or a re- 

 lated form. The second (U.S.N.M. no. 86643) is from the Lukuga 

 River (collector, Dhont), and the third (U.S.N.M. no. 94297) is one 

 of Degon's specimens from Entebbe, Lake Victoria; both of these 

 have a deep peduncle and the posterior dorsal and anal rays longest ; 

 they fall into Boulenger's pumilus. The 48 specimens of Raven's 

 collection have the long posterior dorsal and anal rays of pumilus, 



