TRIASSIC FISHES FROM SPITZBERGEN 
207 
by Schauinsland (1905, p. 261) for Amia , Lepidosteus and Teleosts. We shall see that we 
have an additional support for this view in the fact that in certain fossil Actinopterygii 
both the inter- and basidorsals are probably developed similarly and produced into a 
long spine (cf. Saurichthys, Part II). Under these circumstances unpaired neural spines 
are of course lacking (cf. in addition Schmalhaussen igi 3 a, pp. i— 7). 
The elements of the haemal arches ( Hb , PI. 26, fig. 2) are poorly preserved, so 
that nothing can be etablished about them in detail. 
That there is to a certain extent a commencement of the formation of vertebral 
centra in this species is not impossible, according to what I have found with regard 
to the interdorsals. It is evident, however, that, under any circumstances, these centra 
were weak and that the chorda cannot have been much influenced by then. 
Traquair (1877 a, p. 23 ) is of the opinion that he found indications of vertebral 
centra in P. humboldti Agassiz. His statement is as follows: «Only in Pygopterus I have 
seen anything resembling a vertebral body. In this genus, in which partly owing to the 
comperative thinness and smallness of the scales, the internal skeleton is more comple¬ 
tely seen than in any other, a series of nodular-looking ossifications may generally be 
observed, though much crushed and obscured, extending backwards from the head along 
the abdominal region of the body in the position of the vertebral axis, and wlych cer¬ 
tainly do remind us of vertebral centra, as which they were considered by Agassiz. That 
these were completely developed osseous vertebrae like those of Teleostei and certain 
Ganoids is not probable from their general appearance; it is more likely that they had 
only attained the stage of «hemi-vertebrae» or at most of «ring-vertebrae*, and that in 
any case the notochord passed on uninterruptedly through them». 
Later Sauvage too (1890, p. 26, PI. IV, fig. 1) tought that he had found in the so- 
called P. bonnardi the beginning of formations of vertebral bodies, which he describes 
as «renflements nodulaires a la base des apophyses epineuses». To judge from his figure, 
howewer, it is only the proximal and somewhat widened part of the basidorsals that 
he interprets in this way. He does not show that the elements of the neural arches are 
lengthened in a ventral direction (cf. also p. 203 above). 
With regard to P. de geeri we may mention finally that ossified ribs were not 
developed. 
The skeleton of the unpaired fins. 
The anal fin (A, PI. 25, fig. 3 ) has the shape characteristic of Pygopterus rather 
distinctly developed. It thus has a rather long base and is in the anterior part relatively 
high, while its posterior one is low and weak. The lepidotrichia are about 50 in number, 
are situated close together and are all jointed throughout their length. They are also 
probably dichotomically ramified at their distal parts. Only a few fragments of the endo- 
skeletal radials of the anal fin are preserved ( Ra , PI. 26, fig. 2). 
The dorsal fin has, as far as one can judge, had at least its posterior half opposite 
the anal fin. It is very incompletely preserved and its position is only indicated by a 
few fragments of the endoskeletal radials ( Rd , PI. 26, fig. 2), 
The caudal fin (PI. 25, fig. 3 ) is large and robust and deeply cleft. It cannot be 
decided whith certainty whether it was equilobate as the dorsal lobe is not completely 
