TRIASSIC FISHES FROM SPITZBERGEN 
7 
about the corresponding place as in Megaliclitliys (cf. Watson, 1912 a, pp. 9—10). I 
consider it very probable too that the palatoquadrate of Dictyonosteus articulated with 
or was in some way attached to the basipterygoid process. A connection of this sort 
between the palatoquadrate and the neurocranium thus seems to have been developed 
in several forms among the Crossopterygians and there may be every reason to ask 
whether this was not a character common to all of them. 
Although the palato-basal articulation of the Elasmobranchii is rather variable in 
its position (Gegenbaur, 1872, pp. 186—192; Luther, 1909, pp. 105 —ii 3 ) it seems, however, 
in essentials to correspond to that between the methapterygoid and the orbitotemporal 
region in the Coelacanthids and other Crossopterygians. It seems also to be probable 
that in both cases we are concerned with the primary articulation of the mandibular 
arch against the neurocranium, a view that has already been put forward by Watson 
(19x2 a) in the case of Megaliclitliys and primitive Tetrapods. Moreover in the more 
recent literature there is a fair degree of unanimity as to the palato-basal articulation, 
even in sharks, being an old formation that seems to represent the primaiy connection 
of the mandibular arch with the- neurocranium, or that this articulation in any case was 
situated near the present palato-basal articulation (Allis, 19x5; Huxley, 1876, pp. 43—45; 
Goodrich, 1909, p. 98; Luther, 1909, pp. 112— ii 3 ; Sewertzoff, 1899, pp. 299 — 3 oo; Veit, 
1911, pp. 217—218); 1 ) Swinnerton, (1902, pp. 568—573) also believes that he has found 
that the palatoquadrate in the ancestors of the Actinopterygii must have articulated against 
the posterior part of the trabecular section of the neurocranium between the exits of 
the optic and the trigeminus nerves. 
To return now to W. sinuosa, the medial side of the palatoquadrate is covered in 
this species in the' way characteristic for the Coelacanthids, by a large pterygoid ( Pt , 
text figs. 26 A, B; PI. 4, figs. 1, 2; PI. 6, figs. 1, 3 , 4; PI. 7, figs. 2, 5; PI. 8, fig. 1). On 
this bone one can distinguish a posterior vertical and an anterior horizontal limb, which 
are so connected that an angle, open antero-dorsally, is formed between them. 
The posterior limb is strikingly broad and reminds one in this respect of the 
corresponding limb of the Palaeozoic Coelacanthids. It is broadest at its dorsal part. 
At the ventral part the breadth decreases considerably. The ventral corner is bluntly 
rounded off. The posterior margin is high and very convex, the dorsal one on the 
other hand slightly concave. The antero-dorsal corner is acutely pointed, the 
postero-dorsal one forms an obtuse angle and is not very conspicuous. From the 
antero-dorsal corner there extends downwards and somewhat backwards on the 
lateral surface a rounded ridge, which is strongest in its. dorsal parts, from where 
it becomes gradually lower ventrally and finally disappears near the upper end 
of the quadrate. Posteriorly it is not sharply marked off from the other lateral 
surface of the bone, whereas its anterior border is, along the greater part of its 
length; it is accordingly convenient to take the latter border as the boundary between 
the posterior and anterior limbs. 
x ) Gegenbaur (1872, p. 186), as we know, was of an opposite view; he maintained that the primary articulation 
of the mandibular arch against the neurocranium was to be seen in the postorbital articulation between the palato¬ 
quadrate and the auditory capsule in Notidanids. Jaekel (1899 a, pp. 249—258; 1899 b, pp. 56 — 60) also came 
to a similar conclusion. 
