orm mesenterial filaments, phacellae
or gastric filaments, &c.); the external musculature of the body-wall is
circular (except in _Cerianthus_); the internal, longitudinal; and the
sexual cells probably always arise in the endoderm.

The SCYPHOMEDUSAE, like the Hydromedusae, typically present a
metagenesis, the non-sexual scyphistomoid (corresponding to the hydroid)
alternating with the sexual medusoid. In other cases the medusoid is
hypogenetic, medusoid producing medusoid. The sexual cells of the
medusoid lie in the endoderm on interradii, that is, on the second set
of radii accentuated in the course of development. The medusoids have no
true velum; in some cases a structure more or less resembling this
organ, termed a velarium, is present, permeated by endodermal canals.

The ANTHOZOA differ from the Scyphomedusae in having no medusoid form;
they all more or less resemble a sea-anemone, and may be termed
actinioid. They are (with rare exceptions, probably secondarily
acquired) hypogenetic, the offspring resembling the parent, and both
being sexual. The sexual cells are borne on the mesenteries in positions
irrespective of obvious developmental radii.

The CTENOPHORA are so aberrant in structure that it has been proposed to
separate them from the Coelentera altogether: they are, however,
theoretically deducible from an ancestor common to other Coelentera, but
their extreme specialization precludes the idea of any close
relationship with the rest.

As regards the other three groups, however, it is easy to conceive of
them as derived from an ancestor, represented to-day to some extent by
the planula-larva, which was Coelenterate in so far as it was composed
of an ectoderm and endoderm, and had an internal digestive cavity (I. of
the table).

At the point of divergence between Scyphozoa and Hydromedusae (II. of
the table of hypothetical descent), we may conceive of its descendant as
tentaculate, capable of either floating (swimming) or fixation at will
like Lucernaria to-day; and exhibiting incipient differentiation of
myoepithelial cells (formerly termed neuro-muscular cells). At the
parting of the ways which led, on the one hand, to modern Scyphomedusae,
on the other to Anthozoa (III.), it is probable that the common ancestor
was marked by incipient mesenteries and by the limitation of the sexual
cells to endoderm. The lines of descent--II. to Hydromedusae, and III.
to Scyphomedusae--represent periods during which