6 ON THE STRUCTURE OF EUSTHENOPTERON 
is a lower layer of bony lamellae of the type known as isopedin, 
penetrated by horizontal canals; a middle layer of vascular bone, 
and an upper layer of very dense isopedin, in which the matrix is 
deposited in successive layers with bone cells in parallel planes. 
At the extreme end of the scales the vascular layer disappears and 
the upper layer slopes downward to fuse with the lower. This 
condition is almost exactly paralleled in the dermal scutes of the 
Antiarchi; for instance in Bothriolepis canadensis from the same 
formation, the magnified section of whose shield figured by Good- 
rich (1909, p. 206) might well serve to illustrate that of Eusthen- 
opteron. It seems to me remarkable that fishes so closely related 
as Eusthenopteron and Megalicthys should differ so remarkably 
in the structure of their dermal plates; the former resembling that 
of the Antiarchi, while the latter has the complex structure of the 
_Pteraspidae. 
In this simple lammelar, bony structure of the outer layer of 
the dermal defenses of Eusthenopteron is probably to be found tne 
reason for the fusion of the outer surface of certain head plates, 
for instance, those covering the snout and cheeks, where such 
fusion is complete externally but internally the sutures may still 
be detected. Owing to this fusion of the external surfaces of 
many of the dermal bones of the skull their true outlines are more 
readily made out in their inner aspect. 
This fact has a somewhat important bearing upon the question 
of the origin of the dermo-cranium in fishes. Dr. Gregory (1915, 
p. 327) in support of his “fragmentation” theory, instances the 
continuous rostro-frontal shield of Osteolepis as more primitive 
than that of types in which paircd nasals and other plates 
appear in that region. Now as Dr. Broom has remarked, the 
test that must be applied to all theories of this sort, is how did the 
intermediate stages work? The conditions in Eusthenopteron, 
however, indicate that whatever may have been their origin there 
was in this family a progressive or secondary fusion of elements ; 
that this fusion took place at first in the outer layer; and that the 
more primitive members of the family were those in which the 
dermal elements were the most numerous. The development of 
the exoskeleton thus parallels that of the endoskeleton as shown 
for instance in the progressive fusion of fin elements. 
Text fig. 2 shows the skull of Eusthenopteron in dorsal, ventral 
and lateral views. At the posterior end of the skull in its dorsal 
