TRIASSIC FISHES FROM SPITZBERGEN 
69 
It is very probable that in respect of the membrane bones of the cheek, some of 
the previously known Coelacanthids may show the same tendency to reduction as 
W. sinuosa, although nothing is yet known with certainty in this matter. The fact that 
the bones in question often are very poorly preserved or even quite missing in most 
Coelacanthids seems to me to speak in favour of such a suspicion being correct. 
The visceral skeleton. 
The cartilage of the palatoquadrate persisted presumably to a rather considerable extent, 
although one cannot prove that such was the case as plainly as in certain of the forms 
I shall describe below. The three replacing bones in the palatoquadrate cartilage, the auto¬ 
palatine the quadrate and the metapterygoid are all well developed and preserved in situ. 
The autopalatine ( Aup , text figs. 23 , 26; PI, 4, figs. 1, 2; PI. 6, figs. 1, 4; PL 7, figs. 1, 
5; PI. 8, fig. 1) is a rather small, solid bone, shaped something like a three-sided pyramid, 
with the point lying forwards and the basal surface straight caudally. Is has a dorso¬ 
lateral, a dorso-medial, and a ventro-medial side surface, of which the two last-mentioned 
ones pass into each other without any distinct boundary. A rather thick but short process 
issues from the ventral part of the caudal surface backwards. The ventro-medial side- 
surface lies against the anterior and of the pterygoid, the dorso-medial one has articulated, 
as far as one can judge, with the ethmoidal region. — It ought to be mentioned in 
connection with this that Huxley (1866, p. 37, PI. 8, fig. 3 ) has already described an 
autopalatine in Macropoma , calling it the palatine, and that Reis (1888, pp. 85—87) has 
described the same bone in Jurassic genera, though he uses the term «propterygoid» for it. 
The quadrate (Qm, text figs. 26 A, B, PI. 4, fig. 2) is of rather considerable size. It 
is broadest in its dorsal part, and narrowest a little above the ventral end, which is 
developed as a condyle for the articulation with the mandibula. In the species in question 
it is imperfectly preserved and its contours in text fig. 26 A and B are therefore partly 
restored according to what is known in other species. That its appearance did not differ 
essentially from my restoration is, however, tolerably certain. The dorsal part of the 
medial surface, of the quadrate lies against the ventro-caudal corner of the pterygoid, 
but stretches at the same time a little behind the posterior margin of that bone. 
The metapterygoid ( Mptg , text figs. 26 A, B; PI. 4, figs. 1, 2; PI. 5, fig. 2; PI. 6, 
figs. 1, 3 , 4; PI. 7, fig. 5; PL 8, fig. 1) lies at the dorsal end of the palatoquadrate. It 
is a rather thick plate, rhomboidal in shape, situated on the lateral side of the pterygoid 
bone, but united firmly with that bone only with its anterior margin. Besides this margin, 
which is slightly concave, a dorsal one can also be distinguished, which is likewise 
concave, as well as a posterior and a ventral margin, which are both convex and merge 
into each other in a postero-ventral corner that is much rounded off and thus is not 
distinct. The antero-dorsal corner is extended to a point, while the postero-dorsal one 
is rather rounded. The anterior lower corner forms almost a right angle. The centre of 
ossification lies near the dorsal margin of the bone, a little in front of the middle of 
that margin. The bone substance is clearly radiate. 
With its dorsal, rather broad surface the metapterygoid articulates with the basi- 
pterygoid process (e) and because of this articulation between the palatoquadrate and neuro¬ 
cranium most authors (Huxley 1866, pp. 35—37; Woodward, 1891b, p.395; 1898a, p. 79; 
