NO. 7 ginsburg: new species of gobioid fishes 111 



head ; a band, similar to the two on interorbital, from anterior lower 

 quadrant of eye, extending obliquely forward to upper lip; chin and 

 upper lip dark; caudal light, crossed by four very dark, broad, curved 

 bands, the anterior one, at its base, very broad, concave posteriorly, the 

 others gradually decreasing in width, the second and third somewhat 

 sinuous, the fourth convex posteriorly; second dorsal with very narrow 

 dark streaks running downward and backward ; anal dusky, the pigment 

 increasing in intensity from its base distad, the margin hyaline; a dark, 

 diffuse blotch at lower part of fleshy pectoral base; a diffuse, obliquely 

 elongate blotch on upper rays, near their base; rest of pectoral, ventral 

 and first dorsal light, devoid of chromatophores. 



Aleasurements. — Male 27 mm (the only specimen examined). Cau- 

 dal 31, ventral 26.5, pectoral 26.5, depth 18.5, least depth of caudal 

 peduncle 13.5, head 30, postorbital part of head 16, maxillary 10.5, 

 snout 7, eye 7.5, interorbital 3, antedorsal distance 40.5. 



Holotypc. — In Hancock Coll. San Gabriel Bay, Espiritu Santo 

 Island, Gulf of California; shore collecting; March 7, 1937. 



Distinctive characters and relationship. — This species is easily dis- 

 tinguished from all known American eleotrids by its striking, sharply 

 marked, banded color pattern. It is rather intermediate in its squama- 

 tion between Chriolepis minutillus Gilbert and Gymneleotris scminudus 

 (Giinther), and largely bridges the gap between those two genera with 

 respect to the squamation, the only substantial character hitherto em- 

 ployed for their separation. It has the characteristic row of four scales 

 on the caudal and some large ctenoid scales at the posterior part of the 

 caudal peduncle, like C. minutillus, and it also has smaller cycloid scales 

 anteriorly, like G. seminudus. Unlike the latter species, it lacks mucous 

 pores and the interorbital is not notably wide. In this respect it agrees 

 with C. minutillus and of the known American eleotrids it is structurally 

 nearest to that species. The difference in the squamation together with 

 the strikingly distinctive color pattern shows that minutillus and zebra 

 are widely divergent, to an extent to be placed in distinct subgenera 

 at least, and a new subgenus is therefore established for zebra. 



Eleotriculus, new subgenus 



Genotype. — Chriolepis zebra, new species. 



A definition of the subgenus is included in the description of the 

 genotype. How it differs from the typical subgenus is discussed above 

 in comparing the two genotypes. 



