TRIASSIC FISHES FROM SPITZBERGEN 
247 
In each neural arch are found two paired skeletal elements, a small and weak 
interdorsal, and behind this a large and high basidorsal, both of which were ossified in 
Birgeria. In other forms the basidorsals alone seem as a rule to be ossified. The slender 
unpaired neural spines are always ossified and seem as a rule to be developed in the 
whole abdominal region, so that the most anterior of them are often found under the 
anterior part of the dorsal fin. It is true that they enter into contact with the distal 
parts of the basidorsals, but they are so situated in relation to these that they seem to 
represent the distal parts of the interdorsals, or perhaps, to express it better, they belong, 
like the interdorsals, to the anterior halves of the sclerotomes. 
Ventrally of the chorda we find, at least in the caudal region of Birgeria, a large 
paired bone-plate for each segment. This bone-plate has probably arisen from a fusion 
of the interventral and a proximal part of the basiventral. The distal part of the basiventral, 
on the other hand, is probably represented by the long haemal spine. 
It is very noteworthy that from each of the large haemal bone-plates of Birgeria 
a broad structure of very cancelleous bone issues in a dorsal direction, forming with 
that of the opposite side a semicircle around the ventral half of the chorda. The for¬ 
mation of vertebral bodies is thus begun and, like Traquair (1877 a), I consider it probable 
that this is the case in Pygopterus, but it does not seem impossible that we have there 
instead the interdorsals extended downwards. Finally vertebral body structures are also 
described by Fritsch (1895, pp. 92, 94) in Sceletophorus Fritsch and Phanerosteon Traquair, 
but at least in the case of the latter genus it seems as if Fritsch has made a mistake, 
as Traquair (1911 a, p. 166) states that he did not find any traces of them there. 
In the majority of forms indications of vertrebral bodies seem to be quite lacking 
as in Palaeoniscus (Traquair 1877 a, p. 23 ), Phanerosteon (Traquair 1911 a, p. 166), Psilichthys 
(Hall 1900, p. 147), Coccolepis (Woodward 1891 b, p. 426; 189b a, p. 6, PI. 1, fig. 1 ; PI. 2, 
fig. 4; 1915, p. 24; Traquair 1911b, p. i 3 ), Apateolepis (Woodward 1890 a, pp. i 3 —14) 
and Elpisopholis (Woodward 1908 a, p. 19). 
Ribs have not been ossified in any of the forms so far known. 
The skeleton of the unpaired fins. 
The skeleton of the unpaired fins is well known on the whole (Traquair, 1877 a, 
pp. 24, 26—28; Woodward, 1891b, pp. 426, 427; 1895 a, pp. 6, 7; 1908 a, pp. 19, 20; Hall, 
1900, pp. 146, 147; Abel, 1919, pp. 184, 185; etc.) and in this connection I need only 
point out a few details. 
In most Palaeoniscids the dorsal fin is situated far back. In Boreosomus and Acro- 
rhabdus, however, we have also become acquainted with forms in which it is opposite 
the ventral fins or even has its anterior half extended in front of these. Its endoskeletal 
radials are represented as a rule by two longitudinal rows of ossifications, of which 
the proximal ones are known as axonosts, the distal ones as baseosts. As far as we 
know, the only exception to this is Coccolepis, in which so far it has not been possible 
to show any baseosts with certainty. 1 ) 
z ) In a specimen of Coccolepis bucklandi Agassiz belonging to the Alte Akademie of Munich I think, 
however, I was able to find traces of baseosts as well. 
