TRIASSIC FISHES FROM SPITZBERGEN 
75 
truncated, its dorsal one has posteriorly a rather strong and pointed process. Interiorly 
all the ceratobranchials were partly cartilaginous (text fig. 26 B) and their two ends also 
undoubtedly had cartilaginous epiphyses. Their posterior surface is grooved longitudinally. 
The anterior edge, on the other hand, is rounded. 
The dorsal ossified segment (text fig. 28 C) is small and bifurcate. It is un¬ 
certain how it was attached to the ceratobranchial and similarly it is not clear 
how it is to be oriented in other respects. It is still difficult to take any definite view 
as to its interpretation. But if one starts out from the conditions found in Polyp ter us 
and if the ventral element really represents a ceratobranchial, as I have assumed, 
it does not seem improbable that the dorsal element corresponds to a pharyngo- 
branchial or possibly this element together with a strongly reduced epibranchial 
(van Wijhe, 1882, pp. 257—258; G-aupp, 1904, pp. 907—gi 3 ; Allis, 1918b, pp. 258—261). 
The gill-arches of the two sides 
were joined ventrally by means of a ft 
cartilaginous copula, but as nothing is 
preserved of this it is impossible to 
decide whether it formed a single piece 
or was sub-divided into several seg¬ 
ments. It was covered dorsally by one 
or a few tooth-bearing dermal plates, 
which are thin and consequently much 
crushed. The bone hitherto generally 
interpreted as a copula is, as pointed 
out by Reis already in 1892 (pp. 26—27), 
a dermal bone that has been situated 
ventrally of the real copula and in 
other respects shows a complete corres- 
spondence with the urohyal of the Teleostei (cf. Goodrich, 1909, p. 350). In the species 
in question it is very large and long, and has, from the ventral side, the shape shown 
in text fig. 29 and PI. 4, fig. 2 and PI. 7, fig. 3 . It consists, like other plates of membrane 
bone, of a very compact substance. The centre of ossification is at about the middle, 
and corresponds to the narrowest part of the bone. A cross-section at this place (text 
fig. 29 B) shows that the lateral margins were turned somewhat upwards, so that the 
upper surface became grooved longitudinally and moreover it is clear that there was 
developed on this surface a median longitudinal crista in the same way as on the 
corresponding bone in a number of the Teleosts. Just as in these the anterior end is 
bipartite in certain of the Coelacanthids studied by me, e. g. Coelacanthus elegans New¬ 
berry and Undina penicillata Munster, but it is impossible to decide with certainty if 
this was the case in W. sinuosa too, as the anterior part of the bone in this species is 
incompletely preserved. 
The ossifications of the gill-arches. From P. 25 7. 3 / 4 . 
.. Ceratobranchial. B. Cross-section through a ceratobranchial. 
C. Pharyngo or epipharyngo-branchial. 
Dentition. 
Larger, acutely conical teeth, bent slightly backwards, are found on the precoronoids 
and the supposed maxillary bone; smaller ones of the same shape occur partly on 
10* 
