XXVIII 
ERIK A : SON STENSIO 
Finally if we turn to the fish fauna we are immediately struck by the fact that it 
lacks almost entirely or even perhaps quite entirely forms belonging to the Protospondyli 
and Isospondyli, while the Coelacanthidae and Chondrostei play a very important part in 
it. This gives it a character that is on the whole rather ancient. 
In the fauna of the fish horizon this ancient character is especially striking owing 
to the appearance of the genera Platysomus and Pygopterus, both of which are late 
Palaezoic. Their appearance seems to me an argument in favour of the view put forward 
by Stolley in 1911 and adopted by me, namely that the fish horizon (Posidonomya bed) 
ought to be assigned to the lower Triassic. With regard to the Platysomids it is noteworthy 
in this connection that they are now known from lower Triassic in South Africa (Broom, 
igi 3 a, pp. 4—5) and from formations that are either of late Palaeozoic or lower Triassic 
age bouth at Madagascar (Lemoine 1911, p. 5) and in Canada (Lambe 1914, pp. 17—23). 
On the other hand they are hitherto not found in the middle Triassic. Pygopterus is, as 
we know, a form that really belongs to upper Permian in Europe, although, according 
to recent observations, it seems to occur in the Bunter Sandstone of Germany as well 
(cf. Stolley, 1920, p. 81). 
In its relation to other known Triassic fish faunae the one described here from 
Spitzbergen undoubtedly shows the closest relationship to that of the Alpine and conti¬ 
nental Trias of Europe, a fact that is shown especially by such common forms as 
Palaeobates (cf. Jaekel 1889, pp. 327— 33 o; Wilckens 1909, pp. 187—188), Polyacrodus 
(cf. Jaekel 1889, pp. 321—326), Pygopterus (Stolley 1920, p. 81), Birgeria (cf. Stensio, 
1919, pp. 177—181) and Perleidus (cf. De Allessandri 1910, pp. 49—51). The genera Hybo- 
dus, Acrodus, Ceratodus and Saurichthys are certainly also common to the Triassic strata 
of Spitzbergen and Europe, but we cannot ascribe to them the same value in deciding 
the relationship between the faunae as to the abowe-mentioned genera, since, as is 
known, they have a very extensive geographical distribution (cf. Wemple 1906; Jordan 
1907; Broom 1909 a, b; Stromer 1910; Woodward 1890 a; 1895 b). It is obvious, however 
that all the Hybodus species, certain of the Acrodus species, and the Ceratodus species 
from the Triassic of Spitzbergen are most closely related to the species from the Triassic 
deposits of Europe. 
On the other hand the number of fish genera that are known exclusively from the 
Trias of Spitzbergen is considerable, and the fauna has on the whole a rather special 
character. The most characteristic of these genera are Axelia, Mylacanthus and Sclera- 
canthus among the Coelacanthids and Boreosomus and Acrorhabdus among the Palaeoniscids. 
As will easily be seen from the facts given, the fish fauna in the lower Trias of 
Spitzbergen forms a very strong argument in favour of Diener’s view (1916) mentioned 
above, that during the Skytic time the Arctic Ocean was in communication with the 
ocean in which the Alpine and continental Triassic strata of Europe were deposited. 
That at the same time the Arctic Ocean was also in communication with the Triassic 
Ocean covering the west of America seems to be beyond doubt from certain elements 
in the evertebrate fauna ( Gyronites aplanatum, etc.). In opposition to Diener I suppose 
in accordance whit Merriam (1911) that both communications persisted also during the 
Anisic time. 
