TRIASSIC FISHES FROM SPITZBERGEN 
135 
With regard to the course of the sensory canals Osteolepis (Goodrich, 1919) and 
Ensthenopteron (cf. text fig. 57) agree completely with the Coelacanthids and probably 
this is also the case with the majority of the other Rhipidistids as well. 
Finally, as has already been known for a long time, the Coelacanthids are most 
closely related to the Rhipidistids with regard to the fin skeleton as well. This is 
especially true of the dorsal fins and the anal fin, and to some extent of the paired 
fins too, while the caudal fin apparently differs considerably. But Schmalhaussen’s 
investigations (igi 3 a), as mentioned above, seem to show that this fin too is probably 
more closely related to that of the Rhipidistids than has been known. 
As the result of this comparison we thus find that the Coelacanthids and the 
Rhipidistids are really closely related. It is true that there are a number of differences 
between them, but these cannot be regarded as being of a profound nature and in any 
case are of little significance in comparison with the series of obvious agreements that 
are observed in other respects. Accordingly, as far as we can decide from the facts 
now known, the Coelacanthids ought to be taken as a highly specialized group, the 
ancestors of which are to be sought for among the primitive Rhipidistids. 
Coelacanthids and Tarassids. — Tarassius in still too incompletely known for 
auything in detail to be discovered as to its relationship with the Coelacanthids. As far 
as we know at present, the only common feature are the bluntly lobed form of the 
pectoral fins. 
Coelacanthids and Polypterids. — The account given above ought to show 
that neither the Coelacanthids nor the Rhipidistids are so closely related to the Poly¬ 
pterids as has been believed previously. The Polypterids represent rather, as Goodrich 
(1909, pp. 298— 3 oo) has pointed out, a type that is more closely allied to the Actinopterygii, 
although in my opinion they cannot be grouped with these either. 
Coelacanthids and Dipnoans. 
As has already been pointed out several times before in this work, the Coelacanthids 
present several similiarities to the Dipnoans. This fact was observed early, as we have 
seen from the historical introduction, and the Coelacanthids have even for a long time 
been considered as being closely related to forms which we now definitely place among 
the Dipnoi. 
The most important resemblances between the Coelacanthidae and Dipnoi are briefly 
the following - : the general shape of the mandible and the development of its membrane 
bones (cf. Huxley 1861, p. 20), the reductions in the hyoid arch and in the opercular 
aparatus, the development of the axial skeleton and the diphycercal shape of the caudal 
fin as well as the character of the endoskeleton of this fin. — In addition it has been 
considered that in Coelacanthids as in Dipnoans the air bladder has had a respiratory 
function (Reis 1888, p. 33 ) and finally the branchial arches of the Coelacanthids also 
show in their segmentation similarities to those of the Dipnoans (cf. Gaupp, 1904, p. 920—g 3 i; 
Furbringer, 1904, pp. 483—488; van Wijhe, 1882, pp. 295—297). As, however, we 
do not know the conditions of the other Crossopterygians with regard to the last- 
mentioned character, it is still uncertain whether this peculiarity is not also charac¬ 
teristic for them. 
