TRIASSIC FISHES FROM SPITZBERGEN 
245 
by the preopercular sensory canal in a longitudinal direction. If this were the case, it 
would, in the species in which it is not present as an independent element, be fused 
with the bone elements lying in front of it on the cheek, as in such forms the pre¬ 
operculum, according to what we now know, should have been pierced by the pre¬ 
opercular canal along the posterior margin. 
Assuming that the above-mentioned narrow skeletal element was pierced by the pre¬ 
opercular canal, we should also have a very strong argument in favour of the view that apart 
of the forward directed shank in the bone previously called the preoperculum should corres¬ 
pond to certain of the squamosal elements of the primitive Rhipidistids (cf. pp. 76—79, i 3 g). 
The operculum and the suboperculum vary a great deal in shape and size in 
different genera, as Traquair pointed out in detail already in 1877 (1877 a). Sometimes 
(Boreosomus, Acrorhabdus) the suboperculum may be so low that it shows almost the 
same conditions as a branchiostegal ray. 
The radii branchiostegi vary in number and are sometimes (Boreosomus, Acrorhabdus) 
very few, probably not more than 5. Anterior to them between the two mandibulae 
there is, at least in most forms, a median jugular plate, which, curiously enough, in 
certain forms is very large (Boreosomus, Acrorhabdus). 
The gill-arches are at least in several forms strongly ossified (Traquair, 1877 a, 
p. 22; Fritsch, 1895, vol. 3 , p. 79). Four pairs have so far been observed, but it 
is probable the a fifth existed too. At least in Birgeria and Acrorhabdus the hypo-, 
cerato- and epibranchial are found ossified in the first three pairs. In Birgeria pharyngo- 
branchials have also been found to exist in the two anterior pairs of arches, although 
with the exception of infrapharyngobranchial II, which is represented by a comparatively 
large ossified element, they were cartilaginous. In the fourth pair of arches only a 
ceratobranchial is known so far. All the cerato- and epibranchials have a comparatively 
deep longitudinal furrow on their posterior surface. 
The basibranchials have formed a long row between the ventral ends of the gill- 
arches in the usual way and are occupied partly by ossifications (cf. Fritsch, 1895, 
vol. 3 , p. 79). In Birgeria three of these ossifications are shown to exist, but it is not 
impossible that an additional one may have been present. On the dorsal surface the 
basibranchials are covered with tooth-bearing bone plates of membranous origin. 
Supporting gill-rays are semetimes calcified or ossified and could therefore be 
preserved (Acrorhabdus, Birgeria). In Acrorhabdus a number of them situated in a single 
row occupy such a position in relation to the opercular bones and the hyoid arch that 
they must have belonged to a hyoid gill, which has been large. 
Dentition. 
With regard to this I have very little here, to add to our knowledge, and I can 
therefore merely refer to Traquair’s description of 1877 (1877 a, pp. 32 — 33 ; cf. Allis, 
1919 a, pp. 374—377). 
The sensory canals of the head. 
The sensory canals of the head are fine and, apart from the lines of the pit organs, 
situated within the bones. They open outwards with a large number of fine tubuli, 
which seem in most cases to be straight and unramified. 
