THE DEVONIAN EOCKS OP CANADA, 89 



second dorsal, and like it, forms part of a comparatively large and probably interspinous 

 bono, which narrows abruptly inwards at about the midleugth, especially in front, but 

 which seems to be directly attached, by a transverse articulation, to the oiiter and 

 slightly expanded end of one of the modified hjemal spines. 



The Internal Skeleton of the Tkunk. — Very little if any additional information 

 on this point is afforded by the specimens collected by Mr. Foord since the original 

 description of the genus was written. The statement therein made, that the vertebral 

 centres of Eusthenopleron are not ossified, appears to be fully corroborated by the new 

 material obtained. In si^ecimen No. 4 and others, as well as in one of the original 

 types, the neural and hœmal spines of the posterior half of the vertebral column are 

 beautifully preserved in place, but in none of them can a trace be detected of the verte- 

 bral centres of which they must originally have formed parts. Not a vestige of a rib or 

 of any other part of the anterior half of the endoskeleton can be observed in any of the 

 specimens in the Survey collection. A\'ith the exception of those to which the inter- 

 spinous bones of the azygos fins and tail are articulated, the neural and hœmal spines 

 of the posterior half of the vertebral column, are slightly (nirved, and diverge obliquely 

 backward and outward. Their inner extremiti(»s are slightly expanded and each sj^ine 

 narrows outward into an acute point. In the upper lobe of the tail, the fin-rays spring 

 from the outer terminations of the neural spines, but those of the central lobe seem to 

 have originally abutted directly on the vertebral axis. The nine or ten "osselets " from 

 which the fin-rays of the lower lobe of the tail spring, and whose truncated inner ends 

 are articulated to as many modified hœmal spines, are doubtless true interspinous bones, 

 Avhich are not essentially different in their nature from the corresponding bones in the 

 tail of Trislichnpterus &s they were first supposed to be. 



Stjpposed Affinities of the GtENUS. — From the foregoing description it will be 

 seen that there are so many points of resemblance between Eusthenopteron and Tristichop- 

 terus that it may possibly be doubted whether the distinction between the two genera 

 can be maintained. The principal differences between them, as they now appear to the 

 writer, may be thus briefly summarized. 



(1). So far as can be ascertained at present, the vertebral centres of Eusthenopteron 

 do not appear to have been ossified at all, and they are therefore presumed to have been 

 originally notochordal or cartilaginous in their structure, as is known to have been the case 

 with many other ganoids of the Devonian rocks. In Trislinhopterus, on the other hand, 

 the vertebral centres are stated to be completely ossified and in this respect the genus 

 is said by its author to stand " alone among the contemporaneous fishes." 



(2). In Eusthenopteron the interspinous bones to which are attached the three os- 

 selets from which the component rays of the second doi"sal and anal spring, appear to 

 have been articulated, the one to one of the modified neural and the other to one of the 

 modified heenial spines. The corresponding bones in Tristichopterus are represented as 

 spinous rather than interspinous in their character and as each abutting directly on 

 the vertebral axis. Moreover, in Eusthenopteron, as originally pointed out. the outer 

 extremity or apophysis of the interspinous bone in each of these fins is much more 

 broadly expanded than the apophyses of the corresponding bones of Tristichopterus are 

 represented to be in Sir Philip Egerton's figures of the types of that genus. 



Sec. IV. 1888. 12. 



