TRIASSIC FISHES FROM SPITZBERGEN 
I 4 3 
that there were considerable resemblances in many respects between the shape of 
the cranial cavity in Temnospondylous and Stereospondylous Stegocephalians on the 
one hand and Coelacanthids and Rhipidistids on the other. Thus in both cases we 
find, inter alia, a deep fossa hypophyseos with a well-developed dorsum., which is 
formed by the basisphenoid corpus. In addition there follows in the cranial base, 
caudally of the basisphenoid corpus, another deeper fossa, which was presumably filled 
with cartilage to a great extent or even entirely, and the farthest posteriorly, the 
basioccipital. Finally it is also interesting to find that the portion of orbitotemporal region 
that is situated between the canalis transversus and the foramen opticum is rather long-. 
The nerve exits in the Crossopterygians and Stegocephalians in question, at least 
in the orbito-temporal region and the anterior part of the labyrinth region, have also 
shown rather similar conditions. In this connection it is of special importance that the 
exits of the trigeminal branches were situated immediately above the basisphenoid 
corpus in front of the prootic. While all the trigeminus branches were probably trans¬ 
mitted at this place in the Stegocephalians, this does not seem to have been the case 
in the, Rhipidistids and the Coelacanthids. For in these the ramus ophthalmicus profun¬ 
dus, as far profundus as we can judge, has pierced the cranial wall considerably farther 
forward and has emerged into the orbit antero-medially of the basipterygoid process. 
In several Rhipidistids there is undoubtedly an inner nasal aperture developed as 
in the Tetrapods, while this is lacking, as we have seen, in the Coelacanthids. 
A characteristic common feature of the parasphenoid of Stegocephalians, Rhipidistids 
and Coelacanthids is the absence of processus ascendentes, while as we shall see, such 
processus are, on the contrary, strongly developed in the primitive Actinopterygians. 
One of the grounds for considering the Rhipidistids as ancestral forms of the 
Stegocephalians is, as is known the resemblances of their membrane bones above the 
labyrinth and orbitotemporal regions. As is clear from my exposition in this work the 
Coelacanthids are in this respect rather specialized and thus cannot be directly compared 
with the Stegocephalians. 
The parietal, supratemporal, intertemporal, tabular, postparietal, frontal and post¬ 
frontal (dermosphenotic) of Rhipidistids usually seem to have been taken as homologues 
or approximate homologues of the similarly denoted bones in Rhipidistids (cf. Allis, 1899, 
pp. 58, 67—70; 1919b, pp. 82—86; Baur, 1896, pp. 658—662; Goodrich, 1919, pp. r86—187; 
Gregory, 1915, pp. 325— 338 ; Moodie, 1908, pp. 511 —538; Pollard, 1892a, pp. 408—412; 
Watson and .Day, 1916, text fig. 9; Woodward, 1898 a, pp. 123—125, fig. 81). 
Among the Stegocephalians the intertemporal occurs, however, as an independent 
element only in certain primitive forms; as a rule in the other forms it seems most 
commonly to form part of the bone that is called according to the current terminology 
the supratemporal, which, under these circumstances, would be a supratemporo-inter- 
temporal as in the Actinopterygii (cf. Allis, 1919 b, loc. cit.). In certain forms, however, 
it may have fused with some other of the bones situated close to it. The postparietal 
is paired in the Stegocephalians, unpaired, on the other hand, in the Rhipidistids. Finally 
the postfrontal of the Stegocephalians has practically always a considerable extension 
above the orbit. As anteriorly it is attached to the bone that is sometimes called the 
lacrymal, sometimes the prefrontral and independent supraorbital 'elements are lacking, 
