rt of the pre-maxillary, the maxillary,
the dentary and certain bones of the hyo-mandibular skeleton of
Teleosts. All the investing bones (_Deckknochen_) of the skull were of
common origin, and could be traced back to integumentary skeletal
plates, which in the ancestral fish formed a dense carapace.

These conclusions were accepted by Koelliker himself, who wrote in his
_Entwickelungsgeschichte_ (1879)--"The distinction between the primary
or primordial, and the investing or secondary bones is from the
morphological standpoint sharp and definite. The former are
ossifications of the (cartilaginous) primordial skeleton, the latter are
formed outside this skeleton, and are probably all ossifications of the
skin or the mucous membrane" (p. 464).

Gegenbaur[459] consistently upheld the phylogenetic derivation of
investing bones from dermal ossifications, and even went further and
derived substitutionary bones as well from the integument, thus
establishing a direct comparison between the skeletal formations of
Vertebrates and Invertebrates. Investing bones were actual integumentary
ossifications which had gradually sunk beneath the skin to become part
of the internal skeleton; substitutionary bones were produced by cells
(osteoblasts) which were ultimately derived from the integument.[460]

A further instance of the historical interpretation of animal structure,
taken from quite a different field, is afforded by the speculations of
Dollo[461] on the ancestral history of the Marsupials. In a brilliant
paper of 1880[462] Huxley made the suggestion that the ancestors of
Marsupials were arboreal forms. "I think it probable," he wrote, "from
the character of the pes, that the primitive forms, whence the existing
Marsupialia have been derived, were arboreal animals; and it is not
difficult, I conceive, to see that, with such habits, it may have been
highly advantageous to an animal to get rid of its young from the
interior of its body at as early a period of development as possible,
and to supply it with nourishment during the later periods through the
lacteal glands, rather than through an imperfect form of placenta" (p.
655). Dollo followed up this suggestion, which had in the meantime been
strengthened by Hill's discovery of a true allantoic placenta in
_Perameles_, by demonstrating in the foot of present-day Marsupials
certain features which could only be interpreted as inherited from a
time when the ancestors of Marsupials