168 
PACIFIC SCIENCE, VoL IX, April, 1955 
ily Trichonotidae by Fowler (1938: 206) as 
well as to the gobioid Taeniodidae (Hora, 
1924: 162). It was supposedly redescribed 
from Line Island specimens by Fowler (1938: 
206), but since Fowler’s redescription states 
in part: "gill openings wide, extend well 
forward,” there seems some doubt that Fow- 
ler had Kendall and Goldsborough’s species. 
In 1951 Smith described Paragobioides cop- 
leyi and erected for the genus the family 
Paragobioididae, which he places near the 
Ammodytidae and Trichonotidae. As Dr. 
Strasburg first pointed out to me, Smith’s 
description and figure of Paragobioides copleyi 
differ in only minor respects from those given 
by de Beaufort {loc. cit.) for Gunnellichthys 
pleurotaenia. (In Gunnellichthys pleurotaenia the 
dorsal and anal fins are said to be attached 
by a low membrane to the caudal; in Para- 
gobioides copleyi the dorsal and anal are said to 
end before the caudal base. In G. pleurotaenia 
the pelvic rays are said to be 2; in P. copleyi 
they, "vary from 4-6.” In P. copleyi the chin 
protrudes; in G. pleurotaenia , the figure does 
not show a protruding lower jaw.) It would 
seem therefore that Paragobioides copleyi and 
Gunnellichthys pleurotaenia may well be con- 
generic. 
Another species with a banded color pat- 
tern similar to those of Gunnellichthys pleuro- 
taenia and Paragobioides copleyi is " Cerdale ” 
bilineatus, rather briefly described but not fig- 
ured by Clark (1935: 394) from the Gala- 
pagos. About all that can be said from Clark’s 
description is that his species most probably 
belongs in the Gunnellichthys-Paragobioides- 
Microdesmus group. 
The differences between Paragobioides cop- 
leyi and P. grandoculis would seem to be of 
about the same magnitude as those separating 
the former species from Gunnellichthys pleuro- 
taenia. Smith has already noted {op. cit.: 521) 
the more anterior origin of the pectoral and 
the different coloration of P. copleyi . In addi- 
tion, unlike P. copleyi , the dorsal rays of P. 
grandoculis are said to be "somewhat produced 
beyond membrane.” 
Without specimens at hand it seems futile 
to try to determine whether Microdesmus , 
Gunnellichthys , and Paragobioides are separate 
and valid genera. It may be that all three will 
have to be combined into a single genus 
{Gunnellichthys') but in order to make as few 
nomenclatorial changes as possible the three 
genera may provisionally be maintained as 
separate. They may be separated, probably 
artificially and certainly unsatisfactorily, by 
the following key: 
la. Pelvic soft rays 2 or 3. 
2a. Body without a well marked longitu- 
dinal stripe. Dorsal and anal united to 
the caudal by a membrane 
Microdesmus 
2b. Body with a well-marked longitudinal 
stripe. 
(3a. Dorsal and anal free from the caudal 
"Cerdale” bilineatus Clark) 
3b. Dorsal and anal connected with the 
caudal by a low membrane 
Gunnellichthys 
jf 
lb. Pelvic soft rays 4 to 6. Caudal free from 
the dorsal and anal . . . Paragobioides 
The osteology of Microdesmus multiradiatus 
has been dealt with above. That of Para- 
gobioides copleyi has been described by Smith 
(1951). The main differences I can find be- 
tween these two accounts are as follows. In 
P. copleyi the radials of the first few dorsal rays 
are said, and shown (Smith, 1951: 521 and 
fig. IB), to be quite isolated and subhorizon- 
tal; in Microdesmus they are well developed 
and interdigitate with the tips of the neural 
spines. In P. copleyi the pectoral girdle is well 
ossified and exactly as described and shown 
by Regan for the Gobiidae; I am unable to 
determine the details of the pectoral girdle i 
structure in Microdesmus. The skull is said to 
be highly cartilaginous in P. copleyi (in which 
the total length of the fish was only 2 to 3 
inches), whereas in a skeleton of M. multi- 
radiatus (5.5 inches in length) the skull is 
well ossified. In P. copleyi there is said to be 
a large mesopterygoid, but I believe that this 
