70 
PACIFIC SCIENCE, Vol. VIII, January, 1954 
ment seems desirable is the caudal fin struc- 
ture. I have shown elsewhere (in press) that 
the caudal fin in brotulids varies at least from 
a fairly normal percoid type {Dinematichthys) 
to one with a single "hypural” plate to which 
only four rays are attached {Brotula and Micro- 
hrotuld). The species here considered all fall 
between these two extremes in that there are 
never less than 10 or more than 12 rays ar- 
ticulating with the last vertebra. These 10 to 
12 rays are counted as the caudal rays in the 
present paper. 
The following characters are held in com- 
mon by all the brotulids treated here and will 
not be redescribed for each genus and species. 
Body completely covered with overlapping, 
elliptical scales with eccentric axes. No en- 
larged lateral line scales. Dorsal and anal con- 
tinuous with caudal. Pectoral fin undivided, 
with no rays greatly produced. No barbels or 
crests, but at least a few papillae on top and 
sides of head. No spines on head except on 
opercle and (in one species) on preopercle. 
Mouth subterminal. A flap of skin above 
opercle connected with the pectoral base. 
Opening behind fourth gill arch restricted. 
Three separate pairs of pharyngeals above and 
one pair below, all with small, granular teeth. 
Peritoneum black. 
Counts and measurements for these five 
species are given in Table 1 rather than in the 
specific descriptions. The major differences 
between the species are summarized in Table 
2. Their interrelationships and the recorded 
distributions of the genera to which they be- 
long are dealt with at the end of the paper. 
Genus Diplacanthopoma Gunther 
Diplacanthopoma Gunther, 1887: 115. (Type 
species: D. brachysoma Gunther, by mono- 
typy-) 
Sarcocara Smith and Radcliffe, in Radcliffe, 
1913: 167. (Type species: Diplacanthopoma 
{Sarcocara) brunnea Smith and Radcliffe, by 
monotypy.) 
Head with large muciferous channels, com- 
pletely scaleless. Anterior and posterior nos- 
trils well separated, the former opening on 
the snout rim, the posterior just ahead of eye. 
Teeth in villiform bands on jaws, vomer, and 
palatines; the vomerine and palatine bands 
very narrow. Posterior preopercular border 
without spines. Opercle with a spine above, 
ending in a sharply angulated point below. 
Three developed gill rakers on the first arch. 
Pelvic fins each of a simple filament that 
originates below the opercle. Pseudobran- 
chiae represented by a pair of minute filaments 
or absent. Pyloric caeca absent. 
The reasons for placing Sarcocara in syno- 
nymy are explained by Norman (1939: 79). 
Myxocephalus and Saccogaster appear to be 
closely related to Diplacanthopoma and are 
perhaps not genetically distinct. 
Most of the seven known species of Di- 
placanthopoma^ have been inadequately de- 
scribed, probably due in part to the poor 
condition of the specimens upon which the 
descriptions were based. Scale count, for ex- 
ample, is given for only one of the seven 
species and fin counts for only two. The 
described species have been differentiated 
from one another (where they have been dif- 
ferentiated at all) by such features as the shape 
and size of the head, the size of the eye, and 
the standard length. As a result, it is impossi- 
ble to tell from existing descriptions (and 
probably from re-examination of the type 
specimens) how many of these seven species 
are valid. Furthermore, all the species have 
been described from one or a few specimens, 
so that the nature of the individual variation 
within species in this genus remains un- 
known. Until abundant and topotypic mate- 
rial of these species becomes available, there 
seems no way of determining with any as- 
3 Diplacanthopoma jordani, described from near the 
Galapagos Islands by Garman (1899: 160), is not men- 
tioned in Norman’s list of deep-water brotulid species 
(1939: 84-92), nor can I find other mention of it in 
the literature. In any event, D. jordant is described in 
hopelessly unrecognizable fashion; no figure, no gen- 
eric characters, no fin-ray or scale counts, not even the 
length of the type is given. 
