274 
ERIK A : SON STENSIO 
canal running downwards and forwards laterally of the nasal aperture. Whether the 
continuation of the supraorbital canal forwards and medially of the nasal aperture is 
totally lacking or was developed as a «pit-line» is unknown. If we could verify the 
absence of this anterior part of the supraorbital canal, this would mean that Palaeoniscids, 
Platysomids and Catopterids belonged to a special type, from which neither the sturgeons 
nor the higher Ganoids could be thought to descend. With regard to the anterior end 
of the supraorbital canal the Saurichthyids show the conditions of typical sturgeons, 
while, on the other hand, with regard to the posterior end of this canal they resemble 
the Palaeoniscids. 
At least in certain Palaeoniscids (Acrorhabdus) the primary skeleton of the shoulder 
girdle is, apart from its strong degree of ossification, considerably more primitive than 
in the recent sturgeons, as the canal for the dorsal musculature of the pectoral fin is 
quite absent. On the other hand, however, it appears to a certain extent to be specia¬ 
lized by the strong development of the nerve canal. 
As ought to be clearly seen, most of the characters in which the sturgeons differ 
from the Palaeoniscid type are to be considered as regressive specializations. The sturgeons 
are closely related to the Palaeoniscids and forms nearest to them, but the idea that 
they should be directly descended from the Palaeoniscids, as Woodward (1889 c, pp. 42, 43; 
1895b, pp. V—VIII; 1898a, pp. 82—94; 1915a, p. LXXII) especially and Goodrich (1909, 
pp. 307—32i) have tried to show, seems, however, to be contradicted both by the 
condition of the ossifications of the primordial neurocranium, the extension and course 
of the supraorbital canal, and possibly also by the development of the shoulder girdle. 
Many of the better known Palaeoniscid forms are otherwise too strongly specialized for 
us to be able to imagine that the sturgeons had arisen directly from any of them. 
Palaeoniscids, Platysomids and Catopterids compared with Amia. 
On pp. 152—182 above I have been able to make detailed comparisons between 
the neurocranium of Birgeria among the Palaeoniscids and that of Amia among higher 
ganoids. I merely refer here to this account, but must, however, emphasize, as I have 
done just previously, the fact that a number of Palaeoniscids may possibly in certain 
respects, e. g. the fenestration of the orbitotemporal region, have shown more primitive 
conditions than Birgeria. 
Apart from this I shall only in this connection enter upon some questions con¬ 
cerning the homologues of the membrane bones on the dorsal side of the ethmoi¬ 
dal region. 
First with regard to the bone in Amia previously known as the ethmoid (Allis 1889, 
PI. XLI; 1898, pp. 433—436; 1909a, pp. i 3 —14, 17—22, 23—28^, this apparently corres¬ 
ponds to the middle rostral in Colobodus and a couple of median rostrals in primitive 
Rhipidistids. The homologue of the paired lateral rostral in Colobodus is in Amia probably 
fused with the antorbital. Whether the middle rostral in Amia — a term that I now 
introduce instead of «ethmoid» — also includes within itself elements corresponding to 
the interrostrals of the Crossopterygians, is of course impossible to decide with certainty. 
To judge from the extension of the bone backwards between the nasals, this does not, 
however, seem improbable. 
