TRIASSIC FISHES FROM SPITZBERGEN 
27 
In any case it is clear from my account above that the Catopterids with the 
definition I have given to them, are closely connected phylogenetically with the 
Palaeoniscids. They must therefore be assigned from a systematic point of view to the 
sub-order Chondrostei, in which, as we shall see from my account below (Part II), they 
form together with Palaeoniscids and Platysomids a group that is well defined both 
from the Saurichthyids and the sturgeons. 
It is of course impossible to decide with certaintly whether the Catopterids form 
an intermediate link between the Chrondrostei and the higher ganoids, but this does not 
seem probable. As we have found and as will be shown still further below, even the 
Palaeoniscids appear to be more specialized in certain respects than we should expect 
of forms from which, for instance, Protospondyli and Teleosts could be supposed to have 
descended, and it seems as if the same thing is true of the Catopterids. It is, above 
all, the development of the membrane bones of the cheek and the ethmoidal region and 
the course of the supraorbital canal that form arguments against the suppositions that 
Catopterids or the Palaeoniscid-like forms in general could have been the ancestors 
of the Protospondyli and other higher Ganoids and Teleost. 
It ought also to be noted here that it seems as if a specialization of the skeleton 
of the unpaired fins in the direction of the Teleosts had taken place independently in 
Palaeoniscids. ( Coccolepis ), Platysomids (Lambe, 1914, p. 21), and Catopterids. We thus 
see that this specialization can take place rather easily among Chondrosteans, and that 
it must not be considered to be such an important morphological feature as has hitherto 
been thought. 
Palaeoniscids, Platysomids and Catopterids compared with 
recent Ganoids and Teleost 
Palaeoniscids, Platysomids and Catopterids compared with sturgeons. 
Traquair (1877 a, pp. 34—42; 1879, pp. 377—389; 1887, pp. 248—257) and Woodward, 
(1889 c, pp. 24—44; 1890 a, pp. 6—7, 15—16; 1895 b, pp. V—VIII; 1898a, pp. 81—94) 
have clearly shown that Palaeoniscids, Platysomids and Catopterids offer many resem¬ 
blances to the sturgeons, and that they ought to be placed together with these in a 
common larger group. In the description given above I have been able to state some 
additional common features, as e. g., with regard to the extension of the chorda far 
forward in the basis cranii, the relations and shape of the inter- and basidorsals, the 
endoskeleton of the ventral fins, the pelvic girdle, and the membrane bones of the 
shoulder girdle. 
On the other hand my investigations above also show the existence of considerable 
differences between the fossil forms under discussion under the sturgeons, and I intend 
here to pay attention to these differences so as to try to show their morphological 
importance. 
A character that strikingly separates the Palaeoniscids and the allied families among 
the fossil forms from the sturgeons is the considerably greater degree of ossification in 
the primordial skeleton of the former. There is, however, a good deel of evidence to 
