378 
PACIFIC SCIENCE, Vol. XIV, October I960 
fishes. In addition to the features noted by Regan 
(Table 1, nos. 1 1-14), various authors have noted 
the peculiar first vertebra of the Ophidiidae, and 
Arnold (1956) has recently discussed several 
aspects of the vertebral column in the carapids. 
Posteriorly, the vertebral column of carapids 
and Snyderidia tapers to nothing, and there is 
no hypural fan (Emery, 1880, pi. 2). In bro- 
tulids, the caudal skeleton is quite variable, some 
species having a moderately well-developed hy- 
pural structure (Gosline, 1953, fig. 2b, which 
unfortunately is upside down ) ; in others ( Gos- 
line, 1953, fig. 2a, c) no hypural fan whatever 
seems present. In Otophidium the caudal skele- 
ton is about as in Dinematichthys (cf. Gosline, 
1953, fig. 2b), i.e., moderately well developed. 
At the anterior end the first vertebra shows 
considerable variation in the different ophidi- 
oids. For example, among the skeletons exam- 
ined Otophidium has a rather more specialized 
method of articulation between the skull and 
the vertebral column than the others, and the 
neural arch of the first vertebra is heavy, with 
the two sides fused above, but low and rounded 
(Fig. 4a ) . In Snyderidia the two sides of the first 
neural arch extend up and back but are short 
and barely meet above. In the three brotulids 
(see Fig. 4b) the first neural arch has about the 
same inclination but is higher and is closely 
appressed against the second. In the carapids ac- 
cording to Emery (1880, pi. 5) and Arnold 
(1956) the first neural arch projects upward and 
forward. 
One or more of the first few pairs of ribs in 
Brotula, Otophidium (cf. Muller, 1843; Fac- 
ciola, 1933), and Carapus (cf. Emery, 1880) is 
closely associated with the air bladder. The 
nature of these first ribs has been used by Regan 
(Table 1, no. 13) to differentiate ophidioid fam- 
ilies. In Brotula (Fig. 4b) and Dinematichthys 
the ribs on vertebrae one and two are sessile and 
unexpanded; in Snyderidia the first two ribs 
are sessile but somewhat expanded; in Oto- 
phidium the first vertebra and its ribs (Fig. 
4a) are exceedingly specialized. In Brotula and 
Dinematichthys the ribs on the third and fourth 
vertebrae are separate, sessile and more or less 
expanded. In Snyderidia the rib on the third 
vertebra is sessile and considerably expanded but 
that of the fourth is minute and attached to a 
long parapophysis. In Otophidium the second 
vertebra seems to lack ribs; the third, fourth, 
and fifth have sessile ribs, of which that from 
the third runs medially to the air bladder. 
As to the number of transverse processes 
(Table 1, no. 14), there is in Otophidium , at 
least, the probability that "transverse processes” 
and "sessile ribs” are actually the same structures, 
the distinction being merely one of definition, 
i.e., it would appear that in this fish a "transverse 
process” that has a basal articulation with the 
centrum becomes a "sessile rib” (cf. Emery, 
1880, pi. 5, fig. 50, c 4 ). Nevertheless the old 
FlG. 3. Right pectoral girdles of, a, Otophidium 
and, b, Dinematichthys, from outside, ac, Actinost; cl, 
cleithrum; co, coracoid; pe, pelvic girdle; pm, post- 
temporal; sa, scapula; sc, supracleithrum. 
