TRIASSIC FISHES FROM SPITZBERGEN 
129 
In Libys, Axelia ,. Mylacantlius and Scleracanthus the sensory canals are strikingly 
powerfully developed and the pores are large. Little is known of the number and position 
of the pores in many of the forms. It is clear, however, that there is much variation in 
this respect in different genera. In Coelacanthus, Undina and Cocoderma, for instance, the 
pores are probably very numerous; in the forms that have the sensory canals strongly 
developed they are on the other hand, relatively few and much enlarged. 
Axial skeleton. — As I have nothing fresh to add to our knowledge with regard to 
the axial skeleton I merely refer to earlier accounts of it (Reis, 1888, 1892; Schmalhaussen, 
191 3 a). As to the caudal region I mention below, in connection with the description of 
de caudal fin, the views that have recently been put forward by Schmalhaussen (igi3). 
Skeleton of girdles and fins. —The primordial skeleton of the shoulder girdle was, 
at least partly, ossified (Reis, 1892). Parts of its ossifications are known in Undina, Libys, Macro- 
potna. The dermal skeleton consists ofthe clavicle («infraclavicula»), cleithral, supracleithral 
and suprascapular («posttemporal»). It is uncertain whether a postcleithral was developed. 
The clavicle was probably on the whole rather high and narrow. In certain genera, 
however, as shown by Wimania, it was fairly wide in the ventral horizontal limb and 
connected to that of the opposite side in a median suture. The cleithral is higher than 
the clavicle and generally narrower too. The supracleithral is short and seems to 
exhibit a rather great variety in shape. Finally the suprascapular is known in detail 
only in Axelia, where it is represented by a little scale-like plate. 
The pectoral fins are bluntly lobed and the inner skeleton in each (Woodward, 1895 a, 
pp. 3 —4; Wellburn, 1901, pp. 71—72) consists, at least in certain genera ( Coelacanthus, 
etc.) of a number of ossified endoskeletal radials arranged radially in relation to a centre 
situated in the primary girdle skeleton, with which as a matter of fact they articulated. 
The lepidotrichia are usually unjointed for a longer or shorter distance proximally. 
The pelvic girdle is represented by a paired bone-plate, closely resembling in 
shape the so-called basale metapterygii of the Teleostei. This plate is generally loosely 
attached to the one on the opposite side and is sometimes, as, e. g. in Coelacanthus, 
developed with spines on the medial margin. At the postero-caudal part it has on the 
lateral side a wide postero-lateral projection with which the inner skeleton of the ventral 
fin has articulated. The homologues of the plate are of course not yet quite clear, but, 
as far as one can venture to judge, it seems on the whole to correspond well to the 
pelvis in Eusthenopteron (cf. Goodrich, 1901). 
The ventral fins) like the pectoral ones, are bluntly lobed. Their inner skeleton was 
cartilaginous throughout and nothing is known about it so far. The lepidotrichia show 
the same conditions as in the pectoral fins, but they are generally somewhat weaker and 
shorter than in these. 
As in the Rhipidistids so in the Coelacanthids one finds an ossified supporting 
plate both in the two dorsal fins and the anal fin. In the anterior dorsal fin the lepidotrichia 
are in direct articulation with the plate, while in the posterior dorsal and anal fin the 
articulation must have been by means of cartilaginous elements. J ) The supporting plates 
l ) According to Dean the lepidotrichia of the second dorsal fin of Diplurus were directly attached to the 
supporting plate (Dean, 1895, fig. 156, p. 154), but this must certainly be a mistake, as is also confirmed by 
Newberry’s figure 1, pi. 20, in his monograph on the Triassic fishes from New Jersey and Connecticut Valley 1888. 
Stensio, Triassic Fishes from Spitzbergen. 17 
