386 MR. ST. GEORGE MIVART ON THE VERTEBRATE SKELETON. 
ation; and thus it is that in the Sole (Plate LIII. fig. 7) the subcaudal arches are hyp- 
axial in nature, while in the Flounder (Plate LIII. fig. 6) they are certainly in part 
paraxial, being, in the latter case (as pointed out by Dr. Cleland), to a certain extent in 
series with the transverse processes of the trunk, but in the former case with the aortic 
investment, as they are in Lophius piscatorius (Plate LILI. fig. 8), where they are also, 
to all appearance, simply hypaxial parts. ! 
That the subeaudal arches of the Sole should be reckoned as belonging to a more or 
less different category from the subcaudal arches of the Flounder, may by some be deemed 
paradoxical. It is, nevertheless, simply a short statement of facts—namely, that they 
succeed serially in a certain different manner in one case from what they do in another. 
We have already met with a parallel instance in the caudal transverse processes of 
certain Cetaceans; and the instances are numerous, indeed, in which skeletal parts in one : 
animal closely resemble, in external appearance and ‘function, skeletal parts in another 
animal, which parts nevertheless have a different origin. "Thus the different ways in 
whieh the cranial walls are formed in different animals, now by an enlargement of a 
lateral part, e. g. the alisphenoid, now by the greater extension downward of an upper 
part, e. y. the parietal, will supply an example of functional resemblance and genetic . 
divergence. 
We have now reviewed, in connexion with the questions proposed, the more important 
modifications of the axial skeleton as it exists in its fullest development in the five ver- 
tebrate classes. And it seems to me that the view thence derivable à posteriori, appears 
to coincide tolerably well with the à priori view derived from a consideration of the 
mode of development of the vertebrate embryo with its three pairs of laminæ ; for we 
have found, beside the neural arches (or epaxial parts), a series of more or less seg- 
mented hard parts embracing externally the pleuro-peritoneal cavity (paraxial parts) and 
also a series of hard parts (hypaxial parts) either situated within the pleuro-peritoneal 
cavity and its prolongation, e. g. the branchial arches, or else placed within the paraxial 
elements, and related to the dorsal vessels, e. g. chevron bones and hypapophyses. 
Moreover, as the laminse start from a common longitudinal axis, so we have found that 
coalescence or connation sometimes takes place between adjacent series, whether between 
paraxial parts and epaxial parts, on the one hand, or between paraxial parts and hyp- 
axial parts on the other. 
Adopting then the view before mentioned as to the nature of the trabeculæ cranii, 
the axial skeleton of vertebrates may be shortly described as follows (Plate LIII. 
fig. 1):— 
1. A central axis, terminating anteriorly behind the sella turcica, and posteriorly at 
the end of the spine itself, and segmented antero-posteriorly or not, but generally so, the 
segments having a numerical relation to the spinal nerves. 
2. A superior cylinder, enlarged (except in Amphioxus) at its anterior end to embrace 
the brain. The cylinder is generally segmented, the segments (except at the anterior 
end) generally corresponding in number to those of the axis, though sometimes much 
more numerous— Epaxial parts. 
3. An inferior cylinder external to the pleuro-peritoneal cavity (or continuous vith | 
