146 
ERIK A : SON STENSIO 
logues of the jugal and the lacrymal of the Rhipidistids, according to the terminology 
used by Watson and Day (1916) and Goodrich (1919). Finally with regard to the postorbital 
of the Stegocephalians it is not possible to establish whether it corresponds to that of 
the Rhipidistids alone or whether it has other components as well. 
As is seen, the cheek plates of the Coelacanthids thus practically all show a similar 
composition to those in the Stegocephalians. It appears most probable,, however, that 
this is not a primary agreement but a secondary one, originating in both cases from a 
Rhipidistid-like type. 
We have found that the palatoquadrate of the Coelacanthids and probably of the 
Rhipidistids is attached to the posterior part of the orbitotemporal region in about the 
same way as in the primitive Stegocephalians. In addition it is also striking that in the 
majority of Coelacanthids and at least in certain Rhipidistids the entopterygoid has been 
exceedingly high and wide at its posterior part, owing to which it has a considerable 
resemblance to that of the Stegocephalians. 
The mandible of the Coelacanthids is, as we have seen, more specialized than that 
of the Rhipidistids and therefore different from that of the Stegocephalians in various 
respects. It still preserves certain primitive features, however, and such bones as the 
coronoid and the compound splenial (infradental) show relations somewhat like those 
of the similarly termed bones in the Stegocephalians. 
With regard to the hyoid arch the Coelacanthids and the primitive Tetrapods 
show at least a superficial resemblance, inasmuch as its dorsal part is considerably 
reduced. In Rhipidistids, on the other hand, according to Traquair’s (1881, p. 171) and 
Watson and Day’s (1916, p. 16) statements, a hyomandibula which, if not strongly ossi¬ 
fied, is still somewhat large is developed, and it seems very probable that under such 
circumstances the reduction of the hyoid arch in the Coelacanthids ought to be con¬ 
sidered as an independent specialization not implying any relationship with the Tetrapods. 
Finally the sensory canals in Stegocephalians shows in their course considerable 
resemblances in several respects with those both of Coelacanthids and Rhipidistids, as 
is clearly shown by Goodrich’s (1919) figures of Osteolepis and my own figures in this 
work of Dictyonosteus, Eusthenopteron, Wimania and Axelia. 
* 
The facts mentioned above ought to show that the Coelacanthids too have contri¬ 
buted essentially to a correct understanding of the fossil Crossopterygii and of the affi¬ 
nities of the latter with the Tetrapods. We can now see more clearly than before that 
the Stegocephalians most closely resemble in several respects the fossil Crossopterygii 
and among them specially the Rhipidistids. I consider it improbable, however, that the 
Tetrapods originate from Ripidistids of any known type. It is true that among the 
primitive Teleostomi the ancestral form of the Stegocephalians has apparently resembled 
most closely the Rhipidistian type, but it must, however, have been still more primi¬ 
tive than this. We must suppose, for instance after what we have seen with regard to 
the relations of the membrane bones in the roof of the ethmoidal region, that this primi¬ 
tive ancestral form had the dermal skeleton of the cranial roof over the labyrinth and 
orbitotemporal region developed somewhat as is found in Dipterus valenciennesi, and 
