TRIASSIC FISHES FROM SPITZBERGEN 
117 
In the Jurassic and Cretaceous forms the parasphenoid is in most cases rather narrow 
between the orbits, while — at least in the majority of Palaeozoic forms and in the Triassic 
forms from Spitzbergen \— it is distinguished on the whole by a considerable breadth. 
The question of the vomeres ought at present to be considered an open one. The 
material so far available scarcely permits of any certain conclusions. The plates in 
Undina called vomeres by Heineke (1907) may very well be remains of detached tooth¬ 
bearing plates from the copula or the branchial arches. Such plates, as we have seen, 
are often found scattered about in great numbers everywhere among the head bones 
of Coelacanthids. The unpaired tooth-bearing bone found in the mouth of Macropoma 
and Coelacanthus (Huxley, 1866, p. 36 ; Wellburn 1902, p. 477; Woodward 1909, p. 174, 
PI. XXXVII, figs. 3 , 3 a) may, as Huxley pointed out, be interpreted either as an 
unpaired vomer or a premaxillary, with about the same degree of correctness in either 
case. A third possibility with regard to this unpaired 
bone also seems to me worthy of attention, namely that 
it might represent both vomeres and premaxillae fused 
into one bone. In any case the view put forward by 
Goodrich 1909 in his text figure 261 B is untenable, as 
he figures both vomeres and premaxillae in Macropoma as 
paired, independent elements. On the other hand nothing 
certain can be said about the parts of the skeleton that 
Reis (1888, pp. 6, 7, i 3 , 43, 44, 56, 59) has described as 
vomeres and premaxillae in certain Jurassic forms. 
The membrane bones of the cheek. — In the 
majority of Coelacanthids the membrane bones of the 
cheek are incompletely preserved or quite unknown. Coela¬ 
canthus, Undina, Libys and Macropoma are, as we know, 
the forms that, from this point of v 1 ew, are the best preserved. Dean has also given 
a restoration of Diplurus, according to which the plates in question seem to bear the 
closest resemblance to those of the Jurassic forms (Dean, 1895, text fig. 156). In the 
species Coelacanthus trelleri from the Carboniferous, Eastman (1908, pp. 251—-252) has 
described three postorbital cheek plates situated in a vertical row, the ventral one of 
which he considers to be the Huxleian «postmaxillary» (coronoid, according to our 
modern opinion). As far as one can judge from his exposition there seems, however, 
not to be the slightest doubt that this plate is identical with my preoperculo-quadrato- 
jugal, and that in this respect Coelacanthus welleri closely agrees with the Carboniferous 
Coelacanthid type. 
A number of peculiarities with regard to the shape and degree of development 
of the cheek plates in different forms have already been pointed out in connection 
with the descriptions of the Spitzbergen material above. 
Visceral skeleton. — Both in the Carboniferous and Permian Coelacanthids the 
pterygoid is, as far as we know, developed in a fairly uniform way. It seems to be 
more variable in the Triassic, Jurassic and Cretaceous forms. 
The metapterygoid, quadrate and autopalatine are not infrequently preserved and 
fairly often easy to observe. The metapterygoid has, for instance, been found by me 
Text fig. 54. 
Mawsonia gigas. A. S. Woodward. 
Sketched from Woodward’s type specimen 
1907. About 3 /„. B. M. P. 10355. 
Mplg, metaptervgoid; PI, pterygoid. 
