TRIASSIC FISHES FROM SPITZBERGEN 
243 
quadrate are well developed and seem, at least in some cases, to be fused with each 
other to a single large bone (Boreosomus). The metapterygoid has also, at least in 
certain forms (Boreosomus), a blunt antero-dorsally directed process, so directed that it 
may possibly have articulated with the neurocranium in the posterior basal part of the 
orbitotemporal region. The autopalatine, metapterygoid and quadrate consist all three, 
at least in Boreosomus and Acrorhabdus, of the same cancellous bone substance as the 
substitution bones of the neurocranium. 
Of the membrane bones of the palatoquadrate only the ento- and ectopterygoid 
are so far known. Of these the entopterygoid often shows the same conditions as in 
Crossopterygians, its postorbital part being high and covering the whole metapterygoid 
and a large part of the quadrate on their medial surfaces (Boreosomus). The postorbital 
part of the entopterygoid may sometimes, however, be represented by a number of 
mutually independent plates (Acrorhabdus), and then the transition between this part 
and the anterior part takes place gradually. In my opinion we are concerned in the 
latter case with a secondary phenomenon, the posterior part of the bone being reduced. 
If this reduction process were continued so that the posterior part disapeared, the re¬ 
maining anterior part would correspond fairly well to the metapterygoid in Teleosts 
and higher ganoids. 
The ectopterygoid is in most cases a long and narrow bone reaching backwards 
in a caudal direction to and covering the medial side of the ventral part of the quadrate. 
In Birgeria it has had, at least in its anterior part, a lateral lamella covering another 
lamella from the medial side of the maxillary, a character that possibly occurs in Palaeo- 
niscids in general (Traquair, 1877 a, p. 18). The maxillary is always high behind and low 
beneath the orbital entrance. As I have pointed out above (p. i 3 g), and as I shall show 
still further below (se my account of the Saurichthyids in Part II), it probably comprises 
not only the homologue of the maxillary of the Rhipidistids but at least also the homologue 
of their quadratojugal, so that it ought really to be called the maxillo-quadratojugal. It 
is also conceivable that the homologues of certain of the squamosal elements of the Rhipi¬ 
distids have been fused with it, which could naturally be decided with a somewhat great 
degree of probability if one know the position of the «pit-lines» of sensory canal organs 
that were presumably developed on the cheek (cf. also Allis, 1919 a, pp. 374—382, 385— 386 ); 
The mandible is always long and has a straight upper margin. Its cartilage may, 
at least in certain forms (Boreosomus), have been occupied by a single long ossification 
of a cancellous structure. This ossification, which extends from the anterior to the 
posterior end of the mandible, has on its dorsal surface the articulatory fossa for the 
quadrate, and on the lateral side in front of this another fossa in which the musculus 
adductor mandibulae was inserted; according to its extension and relations it thus seems 
to comprise the homologues of the articular, autangular, mentomandibular and possibly 
the autocoronal as w r ell. 
The bone described by Traq.uair (1877 a) as the articular is apparently a supra- 
angular according to the prevalent terminology. 
Besides this supraangular the lateral side of the mandible is covered by an an¬ 
gular and a large dentalo-splenial. The latter, as indicated by its name, is, as far as 
can be judged, a complex bone (cf. also Schleip, 1904, pp. 3 g 8 —899). It is pierced ven- 
