Of the four great Orders into which Agassiz[41] subdivided the class _Pisces_, the Placoids and Ganoids, agreeing on the whole with the cartilaginous fishes of Cuvier, occur abundantly in the palæozoic rocks, while the Cycloids and Ctenoids, answering to Cuvier's osseous fishes, began in the Secondary formations, and are found in all subsequent deposits. The two former reached their maximum in the earlier geological ages, and have been gradually dwindling down ever since, till now they are represented by comparatively few genera; the two latter are emphatically modern orders; they have been constantly increasing in numbers since their creation, and swarm in every sea at the present day. The carboniferous ichthyolites belong, of course, only to the two first-mentioned orders the placoids and ganoids. [Footnote 41: The classification of Agassiz, which is certainly not a little arbitrary and artificial, has been altered by Müller, a distinguished German anatomist, whose arrangement has been modified again by Professor Owen. See Owen's _Lectures on Comparative Anatomy_, vol. ii. p. 47. There is far from anything like unanimity on the subject. Every naturalist thinks himself at liberty to modify and restrict the groupings of his predecessors or contemporaries, sometimes without condescending to give synonyms or any clue by which one may compare the rival classifications. The geological student cannot engage in a more sickening task than that of ranging through these various arrangements, and he must possess some self-command who can refrain from throwing up the search in disgust. The best way of progressing is to select some standard work and keep to it, until the characteristics of the genera and families have been mastered, and as far as possible, verified from actual observation. After such preliminary training, the student will be more able to grope his way through the "chaos and dark night" of synonyms and systems.]