ntra flattened or
biconcave, while the modern members of the same group have them
procoelous. But the most remarkable examples of progressive
modification of the vertebral column, in correspondence with geological
age, are those afforded by the Pycnodonts among fish, and the
Labyrinthodonts among Amphibia.

The late able ichthyologist Heckel pointed out the fact, that, while the
Pycnodonts never possess true vertebral centra, they differ in the
degree of expansion and extension of the ends of the bony arches of the
vertebrae upon the sheath of the notochord; the Carboniferous forms
exhibiting hardly any such expansion, while the Mesozoic genera present
a greater and greater development, until, in the Tertiary forms, the
expanded ends become suturally united so as to form a sort of false
vertebra. Hermann von Meyer, again, to whose luminous researches we are
indebted for our present large knowledge of the organization of the
older Labyrinthodonts, has proved that the Carboniferous _Archegosaurus_
had very imperfectly developed vertebral centra, while the Triassic
_Mastodonsaurus_ had the same parts completely ossified.[38]

The regularity and evenness of the dentition of the _Anoplotherium_, as
contrasted with that of existing Artiodactyles, and the assumed nearer
approach of the dentition of certain ancient Carnivores to the typical
arrangement, have also been cited as exemplifications of a law of
progressive development, but I know of no other cases based on positive
evidence which are worthy of particular notice.

What then does an impartial survey of the positively ascertained truths
of palaeontology testify in relation to the common doctrines of
progressive modification, which suppose that modification to have taken
place by a necessary progress from more to less embryonic forms, or from
more to less generalized types, within the limits of the period
represented by the fossiliferous rocks?

It negatives those doctrines; for it either shows us no evidence of any
such modification, or demonstrates it to have been very slight; and as
to the nature of that modification, it yields no evidence whatsoever
that the earlier members of any long-continued group were more
generalized in structure than the later ones. To a certain extent,
indeed, it may be said that imperfect ossification of the vertebral
column is an embryonic character; but, on the other hand, it would be
extremely incorrect to suppose that the vertebral c