Hawaiian Lava-Flow Fishes, Part IV 
Snyderidia canina Gilbert, with Notes on the 
Osteology of Ophidioid Families 1 
William A. Gosline 2 
Although the genus Snyderidia is well 
marked and well described, it has been assigned 
at one time or another to four different families. 
In the attempt to determine the proper family 
allocation of this genus, it has been necessary 
to reinvestigate the nature of the families of the 
suborder Ophidioidea to which it belongs. 
The history of Snyderidia is as follows. The 
genus and species ( Snyderidia canina ) were 
described by Gilbert (1905: 654, 655) from a 
single specimen 309 mm. in length taken off 
Kauai of the Hawaiian Islands at a depth of 
385-500 fathoms. Another specimen 122 mm. 
in length was taken by Gosline, Hayes, and 
Keen from among the fishes killed by the lava 
flow off Hawaii on June 6, 1950. 3 Gilbert’s 
original description is excellent and applies well 
to the lava flow specimen. Indeed there is noth- 
ing worth adding, so far as external features are 
concerned. Parr (1933: 49), because of the 
squamation in Brotulotaenia , has suggested that 
small, embedded scales might be present in 
Snyderidia, but Gilbert is correct in stating that 
scales in the latter genus are absent. 
Snyderidia was placed by Gilbert in the Lyco- 
dapidae, a family hitherto represented by the 
single genus Lycodapus from the west coast of 
North America. However, Regan places Lyco- 
dapus in the blennioid family Zoarcidae ( 1912: 
276) with the statement: 
Lycodapus, Gilbert, includes small deep-sea 
fishes of the North Pacific, and has been made 
1 Contribution No. 135, Hawaii Marine Laboratory. 
Previous papers on these same collections of fishes 
were: I, The Origin and Nature of the Collections; 
II, Brotulidae; and III, Sternoptychidae. All of these 
previous sections were published in this journal. 
2 Department of Zoology, University of Hawaii. 
Manuscript received April 29, 1958. 
3 A third specimen of this species is mentioned by 
Radcliffe (1913: 176), but he does not state where it 
was taken. Schultz and Kanazawa (in lift.) provide the 
information that this third specimen is USNM 99230, 
collected in the Philippines. 
the type of a distinct family and placed near 
the Fierasferidae [Carapidae] . But the head and 
mouth recall those of Lycodopsis or Bothrocara, 
the gill-membranes join the isthmus between 
the rami of the lower jaw (at least in L. fieras- 
fer), and the dorsal and anal rays correspond 
in number to the myotomes. 
C. L. Hubbs ( in litt.) concurs with Regan’s 
placement of Lycodapus in the Zoarcidae. On 
the basis of these allocations I will assume that 
Lycodapus is a blennioid fish and dismiss it from 
further consideration. Inasmuch as the several 
dorsal and anal rays per vertebra indicate im- 
mediately that Snyderidia is not a blennioid, 
it cannot be placed in the same family with 
Lycodapus. 
Smith and Radcliffe (in Radcliffe, 1913: 
175 ) described Pyramodon, a genus very similar 
in appearance to Snyderidia, and placed both 
genera in the Carapidae. 
However, Regan (1914: 20) included these 
genera in the Brotulidae with the following 
explanation: 
In many respects these two genera seem to 
connect the Brotulidae with the Fierasferidae, 
but I have ascertained that Pyramodon agrees 
with the Brotulidae in the structure of the upper 
surface of the skull, the supraoccipital separat- 
ing the rather small parietals. 
Smith (1955: 546) tentatively assigns Sny- 
deridia to his family Pyramodontidae, which is 
distinguished as follows: 
I. Parietals small, separated by supraoccipital. 
Maxilla well expanded posteriorly. Usually 
pyloric caeca. Pelvics usually present. Pec- 
torals normally with more than 20 rays. 
A. Dorsal origin well in advance of anal 
origin and vent, the latter behind 
pectoral origin. Scales usually present. 
Dentition feeble. Spines on opercular 
bones usually present. Preopercle mar- 
gin usually free Brotulidae 
373 
