Moreover, the great salamander of Japan is much more inert and sluggish than either, and yet it has a well-developed, bony spine.
I can learn nothing of the habits of the sharks _Hexanchus_, _Heptanchus_, and _Echinorhinus_, but Mueller describes them as possessing a persistent _chorda dorsalis_.[174] It may be they have the habits of the tope, but other sharks are amongst the very swiftest and most active of fishes.
In the bony pike (_lepidosteus_), the rigidity of the bony scales by which it is completely enclosed must prevent any excessive flexion of the body, and yet its vertebral column presents a degree of ossification and vertebral completeness greater than that found in any other fish whatever.
Mr.Spencer supports his argument by the non-segmentation of the anterior end of the skeletal axis, _i.e._ by the non-segmentation of the skull.

But in fact the skull _is_ segmented, and, according to the quasi-vertebral theory of the skull put forward by Professor Huxley,[175] is probably formed of a number of coalesced segments, of some of which the trabeculae cranii and the mandibular and hyoidean arches are indications.

What is, perhaps, most remarkable however is, that the segmentation of the skull--its separation into the three occipital, parietal, and frontal elements--is most complete and distinct in the highest class, and this can have nothing, however remotely, to do with the cause suggested by Mr.
Spencer.
Thus, then, there is something to be said in opposition to both the aggregational and the mechanical explanations of serial homology.

The explanations suggested are very ingenious, yet repose upon a very {173} small basis of fact.

Not but that the process of vertebral segmentation may have been sometimes assisted by the mechanical action suggested.
It remains now to consider what are the evidences in support of the existence of an internal power, by the action of which these homological manifestations are evolved.