ERIK A : SON STENSIO 
248 
The axonosts probably correspond to the proximal and the baseosts to the middle 
segments of the endoskeletal radials. The distal segments must evidently have been 
small and cartilaginous. 
In Phcinerosteon (Traquair, 1911a, p. 167) and Birgeria the axonosts and baseosts 
of the dorsal fin are of about the same length, which seems to mean that the proximal 
and distal segments of the endoskeletal radials were equally long, a condition that, 
according to Schmalhaussen’s investigations (1912; igi 3 a; igi 3 b), ought to be considered 
as rather primitive. In other forms, on the contrary, the axonosts are more or less 
longer than the baseosts ( Psilichthys Hall, 1900; Elpisopliolis Woodward, 1908 a, etc.), 
i. e. the proximal and the distal segments in the endoskeletal radials show conditions 
more like those in higher ganoids and Teleosts. 
In the endoskeletal radials of the anal fin as a rule only one long ossification 
seems to be present in each radial. In Elpisopliolis, however, Woodward (1908 a) has 
shown both axeosts and baseosts in all the radials of the fin in question and in Bir¬ 
geria too I found both these elements, though only in a number of the most caudal 
radials. The remaining endoskeletal radials of this fin have a single long ossification 
in the usual way. 
Both in Elpisopliolis and Birgeria the baseosts of the anal fin are rather short 
compared with the axonosts. Whether the baseosts have fused with the axonosts to 
the single long element of bone usually found in the endoskeleton of the fin or whether 
they are more or less reduced and cartilaginous, it is of course not easy to say with 
certainty. It appears to me, however, at present most conceivable that the former alter¬ 
native is the more probable one. 
The ventral lobe of the caudal fin is always supported by a number of haemal 
spines. Under the large fulcra of the dorsal lobe is regularly found a series of endo¬ 
skeletal radials (Schmalhaussen, 1912, igi 3 a, igi 3 b), which are in an unusually distinct 
state of preservation in Coccolepis australis Woodward (Woodward, 1895 a, PI. 2, fig. 4). 
The lepidotrichia in all the unpaired fins are always densely placed and are 
generally jointed throughout their length, except in Birgeria, in which a proximal, 
shorter part is unjointed. 
The skeleton of the girdles and the paired fins. 
The skeleton of the primary shoulder girdle is only known in Acrorhabdus, in 
which it is represented by a plate, ossified in one piece. This plate in pierced in a 
dorsi-ventral direction by a nerve canal. No canal for the dorsal musculature of the 
pectoral fin has been developed and the plate as a whole seems in this respect to 
show the closest resemblance to the girdle in the Selachians. As in the substitution 
bones of the primordial neurocranium the bone substance is cancellous throughout. 
' The membrane bones of the shoulder girdle are developed in the usual way. With 
regard to the claviculae (infraclaviculae, Traquair, Woodward and other palaeoichthylogues) 
it is, however, noteworthy that they may be developed almost in the same way as in 
Acipenser ruthenus and Acipenser sturio. We find, inter alia, in Acrorhabdus, a sinus on 
their dorso-medial margin situated at exactly the same place as in the two sturgeon 
species mentioned, in which, according to Vetter (1878), te sinus is related to a posterior 
