TRIASSIC FISHES FROM SPITZBERGEN 
227 
is occupied in Amia, Lepidosteus and Teleosts by the so-called epihyal, according to the 
rather inconsistent but most usual terminology (cf. Gaupp, 1904, pp. 902—904). 
The operculum (Op, PI. 3 1, fig. 1 ; PI. 32 , fig. 4) is a wide and high bone-plate. Its 
ventral margin is straight and meets the posterior one almost at a right angle. The 
posterior margin has a weak concavity immediately dorsally of the postero-ventral corner; 
otherwise it is convex. 
The suboperculum (Sop, PL 3 i, fig. 1; PL 32 , fig. 4) is incomplete, but it is obvious, how¬ 
ever, that it was wide and rather low. Its upper margin is straight, its posterior one convex. 
A couple of radii branchiostegi are partly preserved, but no details can be ascer¬ 
tained as to their shape. 
The sculpture on the operculum, suboperculum and radii branchiostegi consists 
of short and rather dense striae, mainly arranged parallel to the axis of length of the 
body. On the lower part of the operculum, however, they deviate somewhat from this 
arrangement as they run obliquely backwards and downwards. 
The branchial arches (Pl. 3 i, fig. 2; PL 32 , fig. 3 , 5) are strongly ossified and appa¬ 
rently subdivided into the usual segments. Thus I can distinguish, at least in the three 
anterior arches, a hypo- (Hbrj, cerato- (Cbr), and an epibranchial (Ebr), of which the two 
latter are long and have a rather deep longitudinal furrow on the posterior side. The 
basibranchials (Bbr) are also strongly ossified but no details can be established either 
as tho the shape or number of the ossifications. On the dorsal side they are covered 
with tooth-bearing bone-plates. Both the branchial arches and the basibranchials consist 
of the same sort of bone substance as the substitution bones in the neurocranium. 
Along the posterior surface of each of the gill-arches present there is, mostly in 
situ, as is seen in several of my figures, a double row of long slender comb-shaped 
supporting- gill-rays (Goodrich 1909, p. 94, fig. 57; filamentary cartilages Danforth 1912, 
p. 422; fig. 9; gill filaments Woodward 1895 b, pp. 80, 219; Kiemenstabchen Gaupp 1904, 
p. 909). These supporting gill-rays may attain a length of as much as 15 mm, and are 
almost always somewhat arched, the processes that cause the comb-like shape issuing 
on the concave side (Pl. 3 i, fig. 4 ). Shorter processes may sometimes also occur on the 
convex side, so that under these circumstances we may speak of a double-sided comb- 
shape, but this is not general. The supporting gill-rays are all somewhat thickened 
at their proximal ends and bent towards the concave side, forming a condyle for the 
articulation with the branchial arches. In each double row they have been so situated 
that they had the convex sides turned towards each other, i. e. towards the septum. 1 ) 
A single row of supporting gill-rays occupies such a position in relation to the hyoid 
arch (Pl. 3 i, figs. 1, 3 ) that it decidedly indicates the presence of a large hyoid gill, which 
has extended from the dorsal end of the operculum downwards and forwards on to the 
radii branchiostegi. The most ventral of these supporting gill-rays are situated, to put 
it more exactly, behind the posterior end of the ceratohyal ossification and like those 
situated farthest dorsally, they are weak and short. 
*) Comb-shaped supporting gill-rays are not uncommon among the Teleostomes. Woodward (1895 b, pp. 80, 
219) has described them in Lepidotus and Mesturus, Riess (188 r, p. 522) in Salmo and Cyprimis, and I have 
myself observed that they are also developed in this way in Clupea. According to Goete (1901, p. 557) and 
Stannius (1846, pp. 114—115) this is also the case in Acipenser (cf. also Polyodon Danforth 1912, p. 421, fig. 9). 
29* 
