126 
ERIK A : SON STENSIO 
of reduction. In at least certain species of Axelia they are still more reduced, and are 
represented merely by a number of small scale-like plates. 
Visceral skeleton. — The palatoquadrate cartilage was present to a rather 
considerable extent. In its anterior end an autopalatine is developed. In its ventro- 
caudal corner we always find a powerful quadrate, and in its dorso-caudal end a 
metapterygoid, generally also powerful. 
The autopalatine is a rather small bone, which, at least in a number of forms, 
has articulated against the preethmoid (Macropoma). The quadrate is always large and 
powerful and generally also thick, with a well developed condyle in the ventral end. 
The metapterygoid usually consists of a fairly large bone-plate, which has without 
doubt been continuously attached ventrally to the quadrate by a persisting part of the 
palatoquadrate cartilage. While the metapterygoid is large and well developed, especially 
in the earlier forms (Coelacanthus , Wimania, Axelia, Mylacanihus, Scleracanthus), it seems 
in a number of younger ones to be more or less reduced and more intimately connected 
with the pterygoid (Undina, Mawsonia). 
The palatoquadrate articulates against the basipterygoid process of the basisphenoid 
by means of the dorsal border of the metapterygoid and we have here apparently a 
correspondence to the so called palato-basal articulation of the sharks. This articulation 
must certainly be looked upon as a very old formation, since, as far as can be discovered, 
it is also present in other fossil Crossopterygians (Megalichthys, Watson, 1912 a; Dictyonosteus, 
Stensio 1918) besides the Coelacanthids. We probably have in this the primitive connection 
of the mandibular arch with the neurocranium. 
The palatoquadrate has on the medial side a very large pterygoid which probably 
corresponds both to the ecto- and the entopterygoid of the Rhipidistia (Watson and 
Day 1916) and probably sometimes also to the dermopalatine of those fishes. In shape 
it is an angular bone with a posterior limb situated vertically and an anterior horizontal 
one issuing from this. They both show in detail in different genera and species considerable 
variation with regard to width, height, the course of the margins, etc. 
The dorsal part of the posterior limb is covered externally by the metapterygoid, 
which, as pointed out, is more or less intimately connected to it. The usual conditions 
of this connection are that the anterior border of the metapterygoid is firmly attached 
to only the dorsal part of a ridge along the anterior margin of the posterior limb 
( Coelacanthus, Wimania, Axelia, Mylacanthus , Scler acanthus, etc.). It seems to be more rare 
for the firm connection to be extended to the ventral border of the metapterygoid as 
well, or for a fusion between the two bones in question to have begun (Undina, 
Mawsonia). — The lower corner of the posterior limb is covered from the lateral .side 
by the dorsal part of the quadrate, but no firmer connection between the two has 
existed. The anterior limb, which is always turned so that its medial surface faces 
somewhat downwards and its lateral one somewhat upwards, is anteriorly in connection 
with the lower surface of the autopalatine. 
In certain forms the dermopalatine is present as an independent narrow bone 
along the lateral margin of the anterior pterygoid limb in the anterior part of the 
latter (Coccoderma, Libys, Undina). In other forms, on the contrary, it may perhaps, as 
already mentioned, have been fused with the pterygoid ( Wimania, Axelia, etc.) or be reduced. 
