[87] THE EVOLUTION OF THE FINS OF FISHES. 1067 



hollow bony bar, and the union of this to the lateral elements of the 

 pelvic arch, together with the union of both by suture with the car- 

 apace, rendered any lateral bending of the trunk impossible, so that 

 an almost universal union of the trunk or body segments ensued, owing 

 to this structurally enforced loss of mobility between the vertebral ele- 

 ments. As a consequence, the centra or bodies of the segments disap- 

 peared or were atrophied, leaving only their trough-like plates about 

 one-fourth of an inch thick, formed of the degenerate united central 

 bodies. This trough, with the united rib-bearing arches which arose from 

 its edges, formed a tube for the lodgment and i)rotection of the spinal 

 or nervous cord. Unlike idl other Vertebrates except Turtles, this tube, 

 in that portion over the lungs, is perforated at intervals on each side, at 

 points about midway of the length of each one of the united segments, 

 to give egress to the spinal nerves.* The points of egress for the spinal 

 nerves are usually between the spinous processes in other orders of 

 Vertebrates. 



" In living Armadillos the centra of the trunk vertebrie still remain 

 as more or less depressed cylinders of bone, or, at least, they are distin- 

 guishable as centra, from which arise the rib-bearing arches, which do 

 not completely unite, leaving lateral interspinous openings so as not to 

 entirely close over the nervous cord, as hapj)ens in fossil forms. The 

 reason why the vertebrtiB remained separated in recent species is un- 

 doubtedly because of the mechanical conditions to which these parts 

 of their skeletons were subjected. Here the carapace was jointed and 

 flexible, hence the need of flexibility in the spinal column. In the ex- 

 tinct species, as in Turtles, the degeneration of the centra into mere con- 

 duits for the nervous cord is one of the many contrivances the origin 

 and teleological significance of which can only be explained by a mechan- 

 ical theory. The vertebral column in both was similarly conditioned 

 with respect to strains, mostly transverse — hence the similarity of struct- 

 ure, which it must be borne in mind is, however, no indication of zoolog- 

 ical affinity. 



" Beginning with the homogeneous notochord or rod-like axis of some 

 such form as Amphioxus, Mr. Spencer points out how, as this axis be- 

 came bony with the assumption of the characters of the higher fishes, 

 the alternate pressure and tension incident to the flexures of this axis 

 during locomotive acts would tend to differentiate the vertebral seg- 

 ments ; for it is obvious that, in order to be flexible and at the same time 

 bony, the vertebral axis must become segmented. The mechanical con- 

 ditions under which vertebral axes are placed would indicate that the 

 segmentation took place from within outwards, which is in accordance 

 with observed facts. It is also obvious, in view of the premises, that, 

 in the absence of flexures or bendings of the vertebral axis, we should 

 have a return to the homogeneous structure [now replaced by a continuum 



' In birds, as, e. g., the common fowl, the first segments of the sacrum, the centra 

 of which are similarly atrophied, are perforated laterally in the same situation. 



