TRIASSIC FISHES FROM SPITZBERGEN 
277 
Allis 1898, pp. 434—436; 1905, PL 14, fig. 22; 1909 a, pp. 27, 28) be paired or show a 
certain indication of having had a paired origin, as in Megalops { Ridewood 1904 a, p. 43), 
Chatoesus and Dussumieria (Ridewood 1904 b, pp. 464, 469). As far as is known, it is 
only in Elopsaurus that it is pierced by an etmoidal commissure between the infraorbital 
sensory canals. In certain cases a pit-line corresponding to the ethmoidal commissure may 
be situated above its anterior part, as in Salmo, Esox, Silurus ' etc. (cf. Allis 1905; 
1909 a). Usually, however, this pit-line or indications of it are quite absent, and in at 
least the majority of the forms investigated the premaxillaries seem, as Allis has pointed 
out, to have such conditions that the anterior parts of the mesethmoid, which was related 
to the ethmoidal commissure, should be incorporated with them and form their pro¬ 
cessus ascendentes (Allis 1898, pp. 450—456; 1909 a, pp. 23—28).*) 
My observations in the fossil forms dealt with in this work seem to render possible 
a closer homologization of the so-called mesethmoid, at the same time also affording an 
argument in favour of the view maintained by Allis. It thus seems very probable that 
at least in such a form as Elops and possibly even in Salmo, Silurus and Esox, the anterior 
part of the so-called mesethmoid is homologous to the middle rostral in Colobodus or 
one or more of the medial rostral elements of the Rhipidistids, while its posterior part 
includes the homologues of the inter- and postrostrals of the latter. * 2 ) It would thus be most 
closely homologous to the so-called ethmoid of Polypterus. 
Under these circumstances it is not difficult to imagine either that in certain Teleosts 
the homologues of certain of the medial rostral elements of the Rhipidistids might, 
together with the interrostrals, fuse with the corresponding premaxillaries and form a 
processus ascendens, as Allis considered to be the case. As we have seen (pp. 144—145) 
an analogous phenomenon may be observed in Tetrapods. 
In certain Teleosts the mesethmoid has, as we know, entered into the cartilage 
situated beneath it and is thus partly developed as cartilaginous bone. 
The homologue of the antorbital (cf. Allis 1898, pp. 436 — 438) seems in Teleosts 
generally either to be reduced or to form part of the so-called lacrimal or possibly the ect- 
ethmoid as well. Only in certain Siluroids may an anterior part of it perhaps be represented 
(cf. Juge 1898, PI. 1; Me Murrich, 1884, p. 278; PI. 2, fig. 1; PL 3 , figs. 1, 2). 
As Allis (1909 a, pp. 17—22) has shown, the preethmoids seem to be fused with 
the vomere(s) in several Teleosts. 
q In 1917, however, Allis (1917 a, pp. 111 —112) came to the conclusion that he must relinquish this view. 
2 ) In Esox, however, it is uncertain whether some of the middle rostrale elements are not absent, as in 
his form the two so-called ethmoids have their anterior ends separated. 
