NO. 5 MYERS AND WADE : ATHERINID FISHES 137 



ti'on. To this alone maj^ be ascribed their reference of starksi to Menidia 

 and their failure to recognize that this species represents a very distinctive 

 genus which is only distantly related to Menidia. 



Anal fin elongate, with a single spine and 24-28 rays, its origin in 

 advance of that of first dorsal, midway between caudal base and hind 

 border of orbit. Dorsal fins widely separated, the height of the first being 

 less than the distance between the origins of the two fins. First dorsal with 

 III or IV flexible spines, its origin over base of fourth or fifth anal ray. 

 End of second dorsal over base of fourth from last anal ray. Caudal fin 

 forked. Pectoral fin set high, falcate, slightly longer than the head, its 

 tip extending beyond ventral base. Ventral fins posterior in position, their 

 insertion farther from the upper angle of the pectoral base than from the 

 anal origin. Anus normal in position, a short distance in front of anal fin. 



Premaxillaries broadly dilated posteriorly, the anterior part not 

 separated by a notch from the posterior, protractile, the skin covering 

 them separated by a deep fold from that of the head. Gape strongly 

 curved, restricted at the corners of the mouth by a membrane between 

 the jaws. Rami of mandibles scarcely elevated. Teeth well developed in 

 both jaws, in two rows of large, sharply pointed, slightly curved, widely 

 spaced teeth, somewhat smaller and more close-set laterally, the first row 

 of upper jaw enlarged. Lower jaw very slightly included at the tip of 

 the mouth when mouth is tightly closed. 



Sides of head and trunk scaled. Scales large, 39 or 40 in a lateral 

 series from above opercular opening to caudal base. Predorsal scales 

 about 18. Scales around caudal peduncle 12. 



Head not abruptly truncate posteriorly. The lower, posterior angle of 

 the preopercle is drawn out veiy slightly into a flat projection,^ but this is 

 not as much produced as the "spine" of Eury stole and Nectar ges, and we 

 do not believe that there is any close relationship with those genera. 

 Interorbital slightly convex. 



Scale borders entire, the nucleus centered, the ciiculi obsolete in vis- 

 ible sector of scale. Abdomen not greatly compressed, but this character 

 is not easily observable in the single specimen available to us, owing to 

 distortion. Air-bladder not produced backward into urosome. Peritoneum 

 brownish black. Silvery lateral band present, very well marked and 

 sharply delimited above and below, somewhat constricted on caudal pe- 

 duncle, and bordered above with a dark line. 



6 On one side only of our single specimen. There is no projection on the oppo- 

 site preopercle. 



