TRIASSIC FISHES FROM SPITZBERGEN 
H5 
In certain Stegocephalians we find a distinctly developed ethmoidal commissure 
between the infraorbital canals ( Tertrema , Wiman, 1914 b, PL VII; Loxomma, Embleton 
and Atthey, 1874, PI. IV). This commissure is situated there on the premaxillae. In a 
number of forms the supraorbital canal also extends into the premaxillary on each 
side and may anastomose there with the ethmoidal commissure as in Tertrema (Wiman, 
loc. cit.). Now since, as we know, the premaxillae of the Stegocephalians often have a 
rather considerable extension backwards over the dorsal surface of the ethmoidal region, 
and also always form part of the dorso-medial boundary of the nasal apertures, it can 
of course scarcely be doubted that these bones are not homologous with the similarly 
termed bones in Dictyonosteus and Eustlienopteron. For, as far as one can see, they seem 
in the first place to correspond not only to the premaxillae of the Rhipidistids but also 
to the rostrals of these fishes (cf. Allis, 1898, pp. 450—456), and it is also very probable 
that in their posterior part, medially of the external nasal apertures, the most anterior 
elements of the nasal series are included. In certain Stegocephalians, as e. g. Keraterpeton 
(Fritsch, i 883, vol. I, p. i3g, fig. 83), it seems, at least to judge from their extension 
backwards, that even the homologues of the interrostrals might occasionally be fused 
with the premaxillae. On the other hand the homologues of the postrostral elements on 
each side, by fusion with the homologues of the middle element of the nasal series of 
its side and often probably with the homologues of the interrostral as well, has given 
rise to the large paired bone-plate, called «nasal» in the Stegocephalians. The most 
posterior nasal plate in Stegocephalians is, to judge from the course of the supraorbital 
canal, probably fused with the bone that is sometimes called the adlacrymal, sometimes 
the lacrymal. 
The conditions just described also seem to explain the appearance of certain un¬ 
paired elements in the cranial roof of Stegocephalians. Thus Watson’s internasal (cf. 
Wiman, 1917, pp. 232 —- 236 ; Broili, 1917, pp.571—576; Watson, igi 3 a, p. 340; Gregory, 1917, 
pp. 977, 979, 980) seems apparently to correspond to a rostral element, and the so-called 
interfrontal might eventually correspond to a pair of fused postrostral plates. It is more 
probable, however, that the interfrontal is homologous with a bone of Dipterus valen- 
ciennesi lettered by Pander 1858 in his PI. 3 , fig. 1 with the figure 3 . 
With regard to the membrane bones of the cheek I must first of all strongly con¬ 
trovert the view of Baur (1896) ,and a number of succeeding authors (Moodie, 1908; 
Allis, 1898, pp. 439—445; 1919b, pp. 81—86) that the squamosal (paraquadrate, Gaupp, 
1895) of the Stegocephalians corresponds to a preopercular and their quadratojugal to a 
subopercular or preopercular. 1 ) The knowledge we now possess of the skeletal elements in 
the cheek of Rhipidistids and Coelacanthids and the course of the sensory canals, shows us 
that Baur’s homoligization cannot be correct. Similarly it is evident that the quadratojugal 
of the Stegocephalians cannot correspond to an interopercular, as Gregory (1915, p. 335) 
has maintained. It seems instead to be fairly certain that the bone termed the squamosal 
in the Stegocephalians is a squamoso-preopercular and the quadratojugal a preoperculo- 
quadratojugal in the same sense as in the Coelacanthids. It is also probable that in the 
same way as in the Coelacanthids the jugal of Stegocephalians includes both the homo- 
1 ) That Pollard’s view 1892 (1892 a, p. 410) in this case is still more wrong is easy to understand from 
the facts put forward above. 
S ten si 6, Triassic Fishes from Spitzbergen. 19 
