(lILBEItT AND STAUKS — FISHES OF PANAMA HAV 



195 



pMiiiily CKKDALID.K. 

 344. Microdesmus dipus (liniUicr. 



Not seen by us. In addition to llic tyiic, tliis species is known only from a 

 specimen recordeil by Lockinj^^ton (18.S1, p. 114), from La Taz, L. C. 



Dr. G. A. Boiilciii^er has kindly re-examined for ns tiie type of the species, 

 and writes: " There are four ventral lays. TIhj (ifteen anterior dorsal rays are 

 simple and inarticulate; further back they j;radually become branched and articulate, 

 and are distinctly so from the eighteenth. The anal rays are all articulate and 

 branched." 



345. Microdesmus retropinais Jordan iL- G'dberl. 



I'l.ATi. XX\I, Fi.:. r)9. 



Seven specimens were obtained in rock-pools on the Tanama reef. 



The genus Microdesraus — witii the two species M. dipus and M. retropinnis — 

 has been described as having the ventrals reduced each to a single i-ay. < )iir 

 material has sliown, however, that in M. retropinnis a serious error was committed, for 

 each ventral lin consists of a short slender spine and three slender unbranched rays, 

 tlie inner of which is the longest. Througli the great kindness of Dr. G. A. 

 Boulenger, who has examined for us the type of M. dipns, we learn that that species 

 also has "four ventral rays" (undoubtedly one spine and three rays). 



In .1/. retropinnis, the dorsal fin contains 15 slender s])ines and 32 to 34 rays. 

 Each of the rays is definitely articulated, and the majority of them are many times 

 forked. The anal rays are all articulate and all but the first one forked. 



We are informed by Dr. Boulenger that in .1/. dipus also tlie fifteen anteiior 

 dorsal rays are simple and inarticulate, while further back they gradually become 

 branched and articulate, being distinctly so from the eighteenth back. In this species 

 the anal rays are all articulate and branched. 



In one specimen of M. retropinnis, we enumerate 58 vertebrae in addition to 

 the hypural element. The latter is assisted by one spine in forming the basis for 

 attachment of the caudal fin. 



In six specimens the fins count as follows: 



Dorsal., 

 Anal ... 



47 

 29 



The scales are circular in outline, attached by their entire margin, and are 

 non-imbricate. On the head and the anterior part of the body, they are closely 

 crowded and are arranged in definite rows. Toward the tail, they are more widely 

 spaced, and on the abdomen they are partially or wholly embedded and difiicult to 

 detect. They cover the entire head, including snout and branchiostegal membranes. 



The margin of the upper jaw is formed by the premaxillaries, the broad max- 



