whose shell (fig. 19)
consists of a straight conical phragmacone covered posteriorly by a
very thick rostrum, and produced anteriorly into a thin long
proostracum which is only occasionally preserved. In certain cases
remains of the arms provided with hooks, and of the ink-sac, have been
recognized. The _Belemnitidae_ appear first in the Upper Trias, attain
their maximum development in the Jurassic rocks, and are not continued
into the Tertiary period, though represented in the Eocene by a few
allied forms.

There is no difficulty in deriving the typical existing Decapoda from
_Belemnitidae_, and many of the extinct forms may have been directly
ancestral. Chitinous "pens" like that of _Loligo_, however, begin to
appear in the Jurassic and Cretaceous rocks, so that in this case as
in many others the parent form and the modified form existed
contemporaneously, and the latter alone has survived. The oldest
shells of the _Sepia_ type are from the Eocene, and it is perhaps
possible that the _Sepiidae_ arose separately from the Belemnites.

It is a curious fact that no fossil specimens of the genus _Spirula_
have been found, but this may be due to the fact that it occurs only
in deep water. At any rate there is no evidence that the shell of
_Spirula_ has lost a rostrum and a proostracum; its characters must be
regarded as primitive, not secondary. In the characters of the
protoconch and of the commencement of the siphuncle, the shell of
_Spirula_ agrees with that of the Ammonoids, and in both its position
is ventral, although in most Ammonoids the shell being exogastric the
ventral side is the convex or external, while in _Spirula_ the shell
is endogastric and the siphuncle internal. The fact that the shell is
not completely enclosed by the mantle is also a primitive character.

With regard to the general morphology of the Cephalopoda, it is
difficult to reconcile the existence of two pairs of renal tubes as
well as a pair of genital ducts in _Nautilus_ with the view that the
original Mollusc was unsegmented and had only one pair of
coelomoducts. Considering the great specialization, however, and high
degree of organization of the Cephalopods, it is evident that the
earliest Nautiloid whose remains are known to us must have had a long
evolutionary history behind it, and such metamerism as exists may have
been developed in the course of its own history.