i34 
ERIK A : SON STENSIO 
however, that the postrostral of each side ot the median line is represented by two 
plates, an anterior one (Ptr i) and a posterior one (Ptr 3 ), and the nasal by three plates 
(Na 7 -Na s ), all pierced by the anterior section of the supraorbital canal. The antorbital 
is probably represented by two plates, of which the anterior one (Ant j) is situated 
immediately behind the external nasal aperture (na), while the posterior one (Ant s ) 
which also seems to be the larger one, forms part of the boundary of the aditus 
orbitae. — I am practically certain that the considerable number of independent elements 
in the dermal cranial roof of the ethmoidal region is a primitive feature (cf. Watson 
and Day 1916, pp. 21— 23 ). 
We can thus, in the present state of our 
knowledge, refer all the Rhipidistids and Coela- 
canthids with regard to the cranial roof to a 
common type, which they still rather closely 
resemble. The same is also true of the bone- 
plates of the cheek, as I have been able to show 
above in the description of W. sinuosa. 
In at least Tristichopterus, Rhizodopsis and 
Glyptopomus among the Rhipidistids we find a 
large pterygoid element, which probably corres¬ 
ponds most closely to an entopterygoid. An ecto- 
pterygoid is also shown to exist at least in 
certain forms (cf. Traquair, 1875 a, pp. 38 y— 388 ; 
1881, p. 171; Watson and Day, 1916, pp. n, 14, 
23 ). The strong development of the entoptery¬ 
goid posteriorly in Rhizodopsis is especially 
noteworthy, as the bone almost comes into contact 
with the cranial roof (cf. Watson and Day, 
p. 14) A similar state of affairs is also seen in 
the Coelacanthids, where, as I have shown above, 
the posterior limb of the pterygoid is in most 
cases high and strongly developed. On the other 
hand, as i have also pointed out, the Coelacan¬ 
thids have no independent ectopterygoid. 
We have seen that in the Coelacanthids the 
hyoid arch is weakly developed in its dorsal part and it is even uncertain wheter a hyo- 
mandibular was present. Among the Rhipidistids, on the other hand, according to Traquair’s 
1881, p. 171) and Watson and Day’s (1916, p. 16) investigations, there seems to have been 
a rather powerful hyomandibular, at least in certain forms. Like the hyoid arch, the oper¬ 
cular apparatus of the Coelacanthids, unlike that of the Rhipidistids, is also somewhat 
reduced, since, as we know, only one large opercular plate is preserved in the former 
The teeth show, as far as we know, a simpler structure in the Coelacanthids than 
in the Rhipidistids, but it seems at present difficult to decide whether this difference is 
to be considered as a primary or a secondary one. In this case it would be of interest 
to undertake a microscopical investigation of the teeth of the Devonian Coelacanthids. 
Pm« 
Text fig. 58. Dictyonosteus arcticus Stensio. 
The anterior part of the cranial roof. Sensory canals 
with dotted lines and shading. From P. j>55. */ 2 . 
Ant 1 , anterior antorbital element; Ant 2 , posterior ant¬ 
orbital element; Dsph, dermosphenotic (dermal post¬ 
frontal); Fr, frontal; Ir, interrostral; Na t -Na y nasal 
elements; Pmx, premaxillary ; Ptr v anterior postrostral 
element; Ptr , posterior postrostral element; R, rostrals; 
So, supraorbitals; na , external nasal aperture; orb, 
orbita; y, accessory bone-plate. 
