GILBERT AND STARRS— FISHES OP PANAMA BAY 193 



and between the anterior dorsal rays is black, contrasting sharply with the white- 

 tipped spines and rays. A well-defined elliptical spot of jet-black occupies the mem- 

 brane between the first and second dorsal spines; it is nearly as large as the eye, and 

 is margined with a whitish ring. The blackish blotch, mentioned by Steindachner, 

 between third and fifth dorsal spines is diffuse and ill-defined, and sometimes 

 extends beyond the limits assigned. There is a black spot at the base of each anal 

 ray. The ventrals have the basal portion dusky, the distal third or half of each ray 

 white. The supraorbital tentacle is crossed with red and whitish bars. 



The usual fin-formula is dorsal XII, 10; anal 20: occasionally dorsal XII, 15; 

 anal 19. The dorsal spines are constantly twelve, and the dorsal rays never as numer- 

 ous as seventeen. Steindachner's type, described with seventeen dorsal rays, is figured 

 with sixteen rays. The pectoral fin contains constantly fourteen rays, as figured, not 

 fifteen, as stated in the description. The ventrals are described with one spine and 

 two rays. They have constantly one spine and three rays. The spine is short but 

 strong, and the inner ray is usually hidden in the integument enveloping the second 

 ray, but is sometimes distinguishable externally. The ventral fins have been examined 

 by us in H. gilberti, scrutator, 2'>unctatus and gentUls, and are found to contain con- 

 stantly one spine and three rays. 



The interorbital space is rather deeply grooved. The anterior nostril is in a 

 short tube, the posterior rim provided with a cirrus. The long supraorbital tentacles 

 are unfortunately omitted in Steindachner's drawing. The length of the gill-slit 

 equals the distance from the tip of the snout to the front of the pupil, scarcely extend- 

 ing below to the level of lower base of pectorals. 



A short lateral line extends to or into the second dark cross-band behind the 

 head, its posterior portion more or less broken up into detached fragments. It gives 

 off a few short and irregular branches above and below, each ending in an open pore. 



We have examined the mandibular teeth in twenty specimens without finding 

 a posterior canine, such as was reported by Steindachner in one of his typical 

 examples. 



339. Hypsoblennius brevipinnis {Gilnther). 



The species is apparently rare, as but one young specimen was obtained. 



Homesthes Gilbert. 



Homesthes Gilbert (Jordan Sl Evermann, 1898, p. 2394) (caiihpus). 



Differing from Hypsoblennius chiefly in the presence of four articulated ventral 

 rays instead of tliree as is usual in Blenniiwe. We have examined the ventrals of 

 Hypsoblennius striatus, punctaius, ionihas, cjentilis, and gilberti, and have found them 

 to consist constantly of one short strong spine and three simple articulated rays. In 

 Homesthes caulopus there is one strong short spine and four well-developed simple 

 jointed rays. 



